From Bones to Behavior
From Bones to Behavior • Center for Archaeological Investigations Southern Illinois University at Carbondale Visiting
Views 216 Downloads 2 File size 31MB
Recommend stories
- Author / Uploaded
- jdspeth
- Categories
- Teoría
- Paleolítico
- Humano
- Cazador recolector
- Método científico
Citation preview
From Bones to Behavior
•
Center for Archaeological Investigations Southern Illinois University at Carbondale Visiting Scholar Conference Volumes Lithic Resource Procurement: Proceedings from the Second Conference on Prehistoric Chert Exploitation (Occasional Paper No. 4)
edited by Susan C. Vehik
Foraging, Collecting, and Harvesting: Archaic Period Subsistence and Settlement in the Eastern Woodlands (Occasional Paper No.6)
edited by Sarah W. Neusius
Emergent Horticultural Economies of the Eastern Woodlands (Occasional Paper No. 7)
edited by William F. Keegan
Tracing Archaeology's Past: The Historiography of Archaeology (Southern Illinois University Press, Publications in Archaeology)
edited by Andrew L. Christenson
Between Bands and States (Occasional Paper No. 9)
edited by Susan A. Gregg
Processual and Postprocessual Archaeologies: Multiple Ways of Knowing the Past (Occasional Paper No. 10)
edited by Robert W. Preucel
Quandaries and Quests: Visions of Archaeology/ s Future (Occasional Paper No. 20)
edited by LuAnn Wandsnider
From Bones to Behavior Ethnoarchaeological and Experimental , Contributions to the Interpretation of Faunal Remains Edited by Jean Hudson
Center for Archaeological Investigations Southern illinois Uiliv~rsity at Carbondale Occasional Paper No. 21
Copyright© 1993 by the Board of Trustees, Southern Illinois University All rights reserved Printed in the United States of America Second Printing
Figur Table Ackn
ISBN: 0-88104-076-2 Library of Congress Catalog Card Number: 91-78336 Edited by Ruth N .- Kissell . Designed by Linda Jorgensen-Buhman Production supervised by Donna-Whitfield ButlerWord processing by Brenda Blythe Wells Graphics layout by Thomas Gatlin Frontispiece by J. Massey
1. Introc Jean I I. S1
c.
2.
Multi D. Ge
3.
TheP Comr
Mary 4.
Unco' Strate
Anne_ 5.
Ethnc Recog
]ames
IJ.
s 0
6.
TheA
Kevin 7.
Persp fromJ
Laurer
Contents
rsity
Figures Tables Acknowledgments
ix XV
xvii
1. Introduction
jean Hudson
1
I. Subsistence Strategies: Meat Acquisition and
Carcass Utilization 2. Multiple Predation: A Definitive Human Hunting Strategy D. Gentry Steele and Barry W. Baker 3. The Place of Hominids among Predators: Interspecific Comparisons of Food Procurement and Transport Mary C. Stiner
9
38
4. Uncovering Technological, Organizational, and Seasonal Strategies of Paleolithic Hunting: Experimental Contributions
Anne Pike-Tay and Heidi Knecht
62
5. Ethnoarchaeology of Marrow Cracking: Implications for the Recognition of Prehistoric Subsistence Organization
]ames G. Enloe
82
IJ. Settlement Patterns: Site Function and Duration of Occupatiol} 6.
The Archaeological Structure of a Short-Term Camp
Kevin T. ]ones
101
7. Perspectives on Skeletal Part Profiles and Utility Curves from Eastern Kalahari Ethnoarchaeology
Laurence E. Bartram, Jr.
115
v
I
vi Contents 8. The Role of Body Part Utility in Small-scale Hunting.under Two Strategies of Carcass Recovery Alice M. Emerson
18. Densi Insigl 138
19. Discu
9. Bone Assemblages at Base Camps: A Further Consideration of Carcass Transport and Bone Destruction by the Hadza
Anne 156
Henry T. Bunn 10. Discussion: Subsistence and Settlement Interpretations
fames F. O'Connell
·
169
III. Social Interaction: Food Sharing, Preparation and Consumption, and Questions of Gender and Ethnicity
181
12. Carcass Processing by the Hadza: Bone Breakage from Butchery to Consumption
James S. Oliver
200
13. Food Sharing and the Faunal Record
Fiona Marshall
228
14. Foragers and Farmers: Material Expressions of Interaction at Elephant Processing Sites in the Ituri Forest, Zaire John W. Fisher, Jr.
247
15. Discussion: Social Interaction Bonnie W. Styles
263
IV. Noncultural Processes: Carnivore Scavenging and Density-Dependent Attrition 16. A Carnivore's View of Archaeological Bone Assemblages Robert J. Blumenschine and Cu'rtis W. Marean
273
17. The Impacts of Domestic Dogs on Bone in Forager Camps
Jean Hudson
20. Concl
D.K.' S.Ken
Contr
11. Gaps in the Zooarchaeological Analyses of Butchery: Is Gender an Issue?
Diane Gifford-Gonzalez
R. Lee
301
I
Contents vii 18. Density-Mediated Attrition of Bone Assemblages: New Insights 138
R. Lee Lyman
324
19. Discussion: Noncultural Processes
Anne K. Behrensmeyer 156
20. Concluding Discussion: The Role of Actualistic Studies D. K. Grayson, P. ]. Watson, D. Gifford-Gonzalez,]. E. Yellen, S. Kent, and A. K. Behrensmeyer
342
349
169
Contributors
181
200
228
247
263
273
301
353
Figures 2-1. 3-1. 3-2.
3-3. 3-4.
Predatory strategies concerning number of predators participating in hunt and n~mber of prey taken
11
Geographic distribution of four Middle Paleolithic (Mousterian) cave sites in west-central Italy
40
Examples of the two anatomical representation patterns for medium ungulates in Italian Middle Paleolithic cave sites
41
/Anatomical patterns for medium ungulate remains in carnivore- and hominid-generated shelter faunas
47
The proportion of head and horn parts to limbs, (H+H) /L, for medium ungulate remains by predator
48
The relative frequenCies of horn parts (HORN /L) to head parts (HEAD /L) for medium ungulate remains in predatorcollected faunas
49
Comparison of (H+H)/L medians and ranges for predator series using a linear ratio axis
51
Comparison of CH+H)/L medians and ranges for predator series using a logged ratio axis
51
Comparative data on total fat to total protein content in soft cranial organs and postcranial muscle mass of adult cattle and white-tailed deer
55
4-1.
Map of southwest France
64
4-2.
Single-beveled points from the site of Laugerie-Haute
66
4-3.
Proposed technique for hafting single-beveled points
72
4-4.
Distribution of bevel angles of single-beveled points from Laugerie-Haute
73
3-5.
3-6. 3-7. 3-8.
ix
xI Figures
4-5. 4-6. 5-1.
Seasons.:.of-death of red deer as determined by ~ementum annuli analysis
74
Mixed age profile (an overlay of attritional and prime age patterns?) of red deer
75
Relative representation of all elements
88
5-2.
Relative representation of selected element$
90
5-3.
Average fragment length by element
92
6-1.
Ache camp, 9 January 1985
105
6-2.
The Buchu site
110
7-1.
The Kua research area in the Republic of B'otswana
119
7-2.
Four bar graphs illustrating complementary kill/butchery site and campsite bone assemblages
124
Average element abundances plotted against Standardized Food Utility index for Kua gemsbok kill sites and base camp assemblages
7-3.
7-4. 8-1.
8-8.
E n
8-9.
F
p 8-10.
F
p 9-1.'
F
e
9-2.
R e
9-3.
R
9-4.
R
9-5.
R e.
129
11-1.
A v
Relationship of skeletal completeness index to kill site processing time
130
11-2.
1v a·
Comparative general utility model values for bison with proximal and distal limb units evaluated separately
141
11-3.
1v 0·
8-2. 8-3.
Comparative general utility model values for bison with limb units evaluated as whole elements
142
12-1.
Total fat utility models for bison
143
12-2.
A
c li
8-4.
Skeletal fat models for bison
143
8-5.
Marrow fat models for bison
144
8-6.
Frequency of unit transport as indicated by scalogram analyses for (a) alcelaphine antelope and (b) buffalo
146
Frequency of unit transport as indicated by scalogram analysis for giraffe
147
13-1.
8-7.
A
a
13-2.
SJ m
14-1.
L1
14-2.
P: cc
Figures Relationship between bison skeletal fat yield and a proxy measure of relative transport cost
148
Relationship between bison nonskeletal fat yield and a proxy measure of relative transpprt cost
149
Relationship between bison bone grease fat yield and the proportion of grease to skeletal fat yield
151
Representation in percent (%MNU) of different skeletal elements of all taxa transported to base camps by the Badza
159
Representation in percent (%MNU) of the different skeletal elements of zebra transported to base camps by the Badza
160
9-3.
Representation in percent (%MNU) of zebra bones at SN-A
163
9-4.
Representation in percent (%MNU) of impala bones at SN-A
164
8-8. 74 8-9. 75 88
I xi
8-10.
90 92 105
9-1. 9-2.
110 119
y 124
?d mp
9-5.
1 Representation in percent (%MNU) of different skeletal
elements of all taxa at SN-A
165
A visual model of the movement of faunal remains through various sites
191
Model showing animal processing activities likely to occur at each locality type
192
Model showing bone modifications and discards likely to occur at each locality type
193
129
11-1.
130
11-2.
141
11-3.
142
12-1.
Axial and limb bone breakage frequencies
208
143
12-2.
Comparison of breakage frequencies for cooked and uncooked limb bones by animal size class
210
Animal body parts accumulated by Okiek households over a six-month period
236
Spatial distribution of the faunal remains of a widely shared animal and a little-shared animal from a !Kung camp
240
14-1.
Location of the study area in the Ituri Forest, Zaire
252
14-2.
Plan map of an Efe campsite showing the primary structural components
253
143 13-1. 144 13-2. 146 147
I
xii Figures 14-3. 14-4. 16-1.
16-2.
17-5.
Plan map of a Lese village showing the primaty ~tructural components
254
Plan map of an elephant processing site showing the primary structural components and discarded bones
257
Comparison of the rank order of 29 skeletal parts based on Binford's modified general utility index and standardized grease index
17-6. 17-7. 277 17-8.
Proportion of long bone fragments that bear at least one tooth mark
281
16-3.
Size-specific destruction/ deletion of long bone epiphyses
284
16-4.
Relationship between two-measures of epiphyseal fragment deletion and destruction
285
Relationship between the extent of epiphyseal removal and the proportion of shaft £ragmen ts that bear at least one tooth mark
287
16-5.
16-6.
16-7.
16-8.
Qualitative model of the degree of competition among scavengers of assemblages of hammerstone-broken long bones
17-9. 17-10. 17-11. 18-1.
288
19-1.
...c
~
Relationship between bulk density and the sequence by which skeletal parts of sheep were observed to be removed/ from 17 Berkeley simulated sites by spotted hyenas
289
Pre- and post-ravaging MNE estimates for 33 experimental assemblages at Berkeley
290
19-2.
E
l
19-3.
~
c c
17-1.
17-2.
17-3.
17-4.
Plot of the relationship between MNI and the actual number . of individuals observed ethnographically
306
Comparison of percentage contribution of each taxon based on MNI and actual number of individuals observed ethnographically
308
Plot of the relationship between MNI survival and carcass size as measured by live ~eight
308
Plot of the relationship between MNI survival and sample size as measured by the actual number of individuals observed ethnographically for each taxon
309
Tables Examples of Hunting Strategies for Selected Social Mammalian Predators
14
Selected Examples of Taxa Taken by Humans in Multiple Predation Episodes
20
3-1.
Expected MNE Values
42
3-2.
SummaryStatistics for (H+H)/L by Predator Type
50
, Results of Rank-Ordered Comparisons of Predators
52
2-1. 2-2.
3-3. I
3-4.
Summary Data on Fat and Protein Content
54
5-1.
Bone Splinters from Three Nunamiut Samples
87
5-2.
Adjusted Relative Percentages of Identifiable Elements
89
5-3.
Fragment Lengths
91
5-4.
Summary of Impact Cones and Average Length
93
6-1.
Site Area and Population for Six Ache Foraging Sites
103
6-2.
Distance from Fire for Activities in Three Ache Camps
106
6-3.
Buchu Site Faunal Remains
111
7-1.
Some Mammals Hunted by the Kua
120
7-2.
%MAU Values for Major Gemsbok Skeletal Elements
125
7-3.
Gemsbok Kill/Butchery Sites
127
12-1.
General Carcass Procurement Data
204
12-2.
Frequency of Breakage Techniques
206
12-3.
Frequency of Roasting, Boiling, and Direct Consumption of Marrow I Cancellous Tissue
209
XV
xvi !Tables 13-1.
13-2.
13-3.
16-1.
Variation in the Degree to Which Meat Is Shared among Contemporary Hunter-Gatherers
229
Distribution of Animal Carcasses Through Transport and · Sharing
234
Food Sharing and Contrasts Between Subsistence and Social Organization in Contemporary Hunter-Gatherers
242
Composition of the Experimental Sample
278 sponsore1
17-1.
Comparison of :MNI and Actual Number of Individuals
307
17-2.
Comparison of NISP and Actual Number of Individuals
311
17-3.
Actual Number of Elements Versus MNE for Four Taxa
315
17-4.
Survival per Element and Part for Four Taxa
316
18-1.
Correlation Coefficients Between Skeletal Part Frequencies and Bone Density and the MGUI
329
Correlation Coefficients Between Percent Weight Loss of Carnivore-gnawed Long Bone Ends and Bone Density
332
Summary Distribution of Correlation Coefficients for 20 New Assemblages and 67 Previously Analyzed Assemblages
333
Correlation Coefficients Between Bone Frequencies for Bqvid Size Classes and Bone Density
335
Classification of 20 Correlation Coefficients Between Bone Density and Small and Large African Bovids
336
Variables Known to Affect Skeletal Part Frequencies
345
18-2.
18-3.
18-4.
18-5.
19-1.
Investiga~
have con· tion with are Georl Center; B director c tion; Tom Joan Con: The staff the logist other cor willingne 1
229 234
Acknowledgments
cial
242 278 307 311 315 316 es
329
This volume, and the conference from which it developed, is sponsored by the Visiting Scholar Program of the Center for Archaeological Investigations, Southern Illinois University at Carbondale. A great many people have contributed to the smooth operation of the program during my association with it, and I gratefully acknowledge their support. Included among them are George Gumerman and Don Rice, the past and present directors of the Center; Brian Butler, associate director; Kim Smiley, curator; Patrice Teltser, director of publications; Donna Butler and Brenda Wells, publications production; Tom Gatlin, graphics layout; Carolyn Taylor, administrative assistant; and Joan Corse, Center secretary. Ruth Kissell did an exceptional job as copy editor. The staff.bf the Division of Continuing Education helped in essential ways with the logi~tics of the conference itself. My special thanks go to the authors and other conference participants for their thoughtful contributions and their willingness to adapt to a demanding production schedule.
332 New
333 lovid
335 1e
336 345
xvii
1.
Introduction Jean Hudson
There are two commonalties that link all of the papers in this volume. One is the use of zooarchaeological remains, specifically vertebrate remains, to address behavioral questions of relevance to the understanding of hominid evolution and hunter-gatherer adaptations. The other is the use of controlled data sets, where the causal variables are known either experimentally or ethnoarchaeologically, to better understand the links between particular variables and the patterning they produce in bone assemblages. Faunal remains are an important constituent of many archaeological sites. They ar~ used to address a wide variety of behavioral questions, ranging from characteristics of subsistence and settlement strategies to aspects of social interaction within and between sites. The papers in this volume examine some of the ways that animal bone can be analyzed to yield inferences about human behavior. The papers vary in the particular aspects of zooarchaeological analysis they use to address behavioral questions, incorporating studies of dentition, population age structure, body part frequency, taxonomic richness, fragment size, breakage morphology, tooth marks, butchering marks, spatial distribution, tool manufacture, bone density, and standard quantitative measures such as MNI, NISP, MNE, and various utility indices. The geographic and temporal interests represented are also diverse, ranging from early hominid sites in East Africa, to the archaic Homo sapiens of the European Middle and Upper Paleolithic, to modern hunter-gatherers from a variety of ecological environments. These papers were originally collected as part of the Eighth Annual Visiting Scholar Conference, an ongoing tradition of the Center for Archaeological Investigations, Southern Illinois University at Carbondale. The theme of the Visiting Scholar Conference changes from year to year, reflecting the interests of the individual holding the fellowship. One of the great benefits of the conference -is its focused, friendly atmosphere and the opportunity that it provides for dialogue among scholars working on related problems. This potential was realized at the 1991 conference. From Bones to BehaT?ior: Ethnoarchaeological and Experimental Contributions to the Interpretation of Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights reserved. ISBN 0-88104-076-2.
1
I
L I
I j
~-
21 ]. Hudson
~ lil I
This volume, like the conference from which it _came, is structured around four broad themes. The first three concern aspects-of cultural behavior of interest to archaeologists: subsistence strategies, settlement patterns, and patterns of social interaction. The fourth theme concerns the impacts of certain nonhuman behaviors and processes that can affect our ability to see the · cultural behaviors that interest us as archaeologists. Part I focuses on several aspects of the interpretation of subsistence strategies, including methods of meat acquisition, such as scavenging and a variety of hunting techniques, differences between foraging and collecting strategies, the role of cooperation in certain types of hunting, considerations of the skills represented by different hunting strategies operating at various stages of hominid evolution, adaptations to temperate climates and seasonal variations in prey resources, and issues of anatomical utility. Steele and Baker raise the question of what aspects of human hunting are uniquely human and of how they might relate to the evolution of a successful adaptive strategy for hominids. Stiner looks at models for the meat acquisition strategies practiced by Neandertals and anatomically modern humans during the Middle and Upper Paleolithic in Italy, suggesting seasonal variation in the role of scavenging in a temperate climate. Pike-Tay and Knecht combine two different types of experimental data, one concerning bone tool manufacture and the other concerning dental evidence for season of death and age structure of the prey population, to model hunting strategies in the Upper Paleolithic of France.~ with the aim of locating these strategies on the foragercollector continuum. Enloe takes a different approach to a similar problem, using controlled data on long bone fragmentation to differentiate between forager-like processing for immediate return and collector-like bulk process· ing of accumulated reserves. Part IT deals with the interpretation of settlement patterns from skeletal part profiles, seasonality measures based on tooth eruption and wear/ and characteristics of site size and spatial distributions ·of bone. These archaeological clues are evaluated as evidence for site function and duration of occupation. Jones uses ethnoarchaeological data from Ache foraging camps to examine patterns associated with very short-term occupations, including small site area, clustering of discarded bone near hearth areas where almost all activities took place, and comprehensive representation of skeletal elements reflecting · the lack of logistical transport. Bartram reviews Kua transport patterns for gemsbok kills, noting that utility models alone do not account for variability in carcass treatment and that processing strategies, particularly those that remove meat from bone prior to transport, must be considered when using skeletal element profiles to interpret site function. Emerson examines the use of various types of anatomical utility indices to interpret transport strategies from skeletal part profiles, focusing on single kills of large game and the role of fat in processing and transport decisions. Bunn reviews the varying transport patterns documented for the Badza, particularly as they relate to the relative abundance of axial and appendicular elements in base camps, suggesting that quantity and quality of marrow may play an important role in decision making.
~
I
!
I
_,~
I
Part Ill between r field, the s action of· ·patterns o has been a less attent ·conzal~z.
Giffordpreparatio likely to h that a foct earlier stal port, wouJ impacts o: terns, sugl influence i breakage I and cookii sharing or the distanc tion with duce patt between k data on an them fora~ between st in the patt the faunal Part IV influence 1 both wild taphonom: Blum en: captive hy carnivores importanc epiphyses and midsh related pr1 camps, sul :rvt:l\TEs que tically relic can be dra tic dogs S< ethnoarch' frequency 1
~~
]
l
:Ji,
'il·
Jj
,_,il:
jjl
Introduction I 3 tured around Ll behavior of patterns, and 1acts of certain ity to see the istence strate; and a variety ing strategies, 1s of the skills ~ous stages of mal variations n hunting are )f a successful ~at acquisition umans during ariation in the t combine two I manufacture :md age strucin the Upper m the forager1ilar problem, ~tiate between bulk processn skeletal part tr, and characuchaeological occupation. ps to examine .ing small site st all activities ents reflecting rt patterns for for variability .rly those that d when using Lmines the use port strategies Le and the role varying transIf relate to the e camps, sugJortan t role in
Part III deals with aspects of social interaction, including the relationship between meat processing in residential camps and meat processing in the field, the sharing of meat between members of different camps, and the interaction of two ethnic groups at a kill site. These papers focus attention on patterns of behavior that act to modify a~d redistribute bone after the game has been acquired and transported. Such patterns have traditionally received less attention in interpretations of faunal remains, as pointed out by Gifford·Gonzalez. Gifford-Gonzalez questions our apparent neglect of the final stages of meal preparation and consumption, suggesting that they are the behaviors that are likely to have the greatest effect on patterns recovered archaeologically and that a focus on end products and how they are produced, rather than on the earlier stages of decision making concerned with meat acquisition and transport, would be productive. Oliver addresses this topic in his discussion of the impacts of Badza butchering and cooking practices on bone breakage patterns, suggesting that carcass size and the cooking techniques available will influence the degree to which bone is fragmented and that greater attention to breakage patterns may allow us to trace the development of both butchering and cook,ing techniques through time. Marshall considers the impacts of meat sharing 16n skeletal part distribution between Okiek households, noting that the distances between households, typically a kilometer or more, in combination with differences in hunting success and patterns of sharing, act to produce patterns similar to those traditionally associated with differences between kill sites and residential camps. Fisher. presents ethnoarchaeological data on an elephant butchery camp used jointly by two ethnic groups, one of them foragers, the other hortkulturalists; material evidence of the interaction between societies with different patterns of mobility and food storage is seen in the patterning of hearths and structural remains as well as the structure of the faunal assemblage. Part IV focuses on two particular types of nonhuman phenomena that influence the composition of faunal assemblages: destruction by carnivores, both wild and domesti~, and density-mediated attrition. Both represent taphonomic processes of major importance in many archaeological sites. Blumenschine and Marean use experimental observations on wild and captive hyenas to improve our understanding of how ·scavenging by wild carnivores impacts bone that has been processed by hominids; they stress the importance of collecting data on the placement of tooth marks, the ratio of epiphyses to shafts, and the distinction, in skeletal part profiles, of epiphyseal and midshaft fragments. Hudson uses ethnoarchaeological data to address a related problem, the impacts of resident domestic dogs in modern forager can1ps, suggesting that destruction of bone by Aka dogs renders the use of MJ\JEs questionable, that taxonomic ranking by MJ\JI and NISP remains statistically reliable, and that prey species with live weights of less than a kilogram can be dramatically underrepresented archaeologically in sites where domestic dogs scavenge bone. Lyman uses a collection of 87 archaeological and ethnoarchaeological assemblages to assess the extent to which skeletal part frequency is predicted by bone density alone, finding density a reasonable
41 J. Hudsonpredictor in almost half the cases. Several interpretive concerns and analytic-approaches crosscut the thematic groupings of the papers. Those concerns operate at two levels. One is methodological-how to confidently read the behavior of interest from the archaeological remains. The other is theoretical-how did early hominids successfully evolve into modern humans? Analytically, three aspects receive attention by several authors: the identification of limb bone shaft fragments; the interpretation of skeletal part profiles; and the choice of appropriate scales of analysis. The importanc~ of identifying shaft fragments is brought out in the papers by Bartram, Blumenschine and Marean, Bunn, Hudson, Lyman, and Oliver. Multiple causes, both human and nonhuman, act to preferentially destroy articular ends. Faunal analyses that ignore shaft fragments are likely to underrepresent the importance of appendicular elements, biasing skeletal part profiles and potentially influencing measures of taxonomic abundance as well. In the papers presented here, skeletal part profiles are used to interpret a wide range of cultural behaviors, including meat transport, site function, sharing, trade, and hunting success, and to document noncultural agents of bone destruction such as carnivores and density-dependent attrition. The current diversity of opinions about associations between particular profiles and particular behaviors and destructive agents suggests that we are still exploring and defining the critical variables and that problems of equifinality have yet to be resolved. The importance of choosing a scale of analysis appropriate to the research question is discussed by several participants, including Blumenschine and Marean, Bunn, Hudson, Lyman, O'Connell, Oliver, and Stiner, although there is no consensus about which scale might be the most appropriate. Blumenschine, Marean, and Lyrrtan advocate more detailed approaches to skeletal part profiles, with attention to fine-scale variations in bone' density, as they may represent the real cause of differences in preservation ahd identifiability. Stiner and Hudson, on the other hand, suggest that more robust measures, produced by lumping elements into broader body part categories, might be more useful archaeologically. Comments by Bunn, O'Connell, and Oliver raise questions about when body part analyses should focus on particular species and when the analyses should group species into size classes. The Hadza data suggest that the use of size classes rather than taxa as the analytic ~ategory is advantageous because it increases sample size but can obscure potentially significant differences in the way particular species are treated. Sample size, in terms of the number of controlled assemblages available for any given study, is a concern for many authors, particularly those working with ethnoarchaeological data. Case studies are presented as opportunities for insight, and tests of the hypotheses generated against additional data sets are recommended. Stiner points out that a logical next step in the application of small-scale, fine-grained ethnoarchaeological studies to archaeological problems will be the combination of multiple assemblages to simulate the coarser scale of archaeological signatures.
;j "I I
:,I :! ':
;[
;I :t
H
:J
I m
m
m
II:!:
Theoret best differ ent types c particular differenti.: sometime: although: both anci1 yielded m Throug: motivate( served to concerns, points ou1 optimal f( being opt] and bene1 conferenc collector ' opment oj ing, as we: Graysm whether o highly pa: made exr hypothesi: tales that~ part of sci often fails very simi] case of as: Behrensm, science is hope for u logically. One of 1 insights be The pape1 ideas abot became cl which to: experimen better sen frameworl extra step can be tes1 situations.
Introduction I 5 .t the thematic ~vels.
One is from the trly hominids ~rest
s: the identifi1 part profiles; of identifying tens chine and thhuman and analyses that mce of appen.y influencing to interpret a site function, :ural agents of attrition. The icular profiles at we are still of equifinali ty o the research 1enschine and :tlthough there appropriate. approaches to Jne density, as L and identifia~e robust meaut categories, >'Connell, and >uld focus on ~cies into size er than taxa as 1le size but can lar species are ~s available for those working portunities for l1 data sets are application of ~ological probate the coarser
Theoretically, these papers are characterized by repeated interest in how to best differentiate the roles of scavenging and hunting, how to identify different types of hunting and the kinds of cultural behaviors associated with them, particularly as they relate to cooperation and planning skills, and how to differentiate foragers from collectors. There is also repeated reference, albeit sometimes oblique, to theoretical models 'of optimization. Common to many, although not all, papers is an underlying assumption that hunter-gatherers, both ancient and modern, used acquisition and processing strategies that yielded maximum meat or fat while minimizing labor effort or risk. Throughout the conference's discussion sessions, comments were often motivated by ethnoarchaeological experience, and such comments often served to push theoretical frameworks beyond existing assumptions and concerns, advocating more critical evaluations of popular models. O'Connell points out in his written commentary that while many of us do lip service to optimal foraging theory, we often stop short of actually evaluating what is being optimized and we typically lack the data to measure the relevant costs and benefits. Both O'Connell and Jones question, as did Yellen during the conference's final roundtable discussion, the usefulness of the foragercollector distinction when applied to evolutionary models about the development pf human abilities to plan and cooperate, noting that modern foraging, as well as modern collecting, regularly incorporates those skills. Grayson raises questions about what the aims of ethnoarchaeology are and whether or not they are of long-term value. He notes that many studies seem highly particularistic and that frequently their theoretical importance is not made explicit. Watson and others respond that the scientific method of hypothesis testing requires particularistic studies and that even the cautionary tales that seem to characterize much ethnoarchaeological work are an essential part of scientific progress. Styles. points out that ethnoarchaeological research often fails to take into account the fact that different processes can produce very similar results and that ethnoarchaeology, when it stops with the first case of association of behavior with material remains, is stopping too soon. Behrensmeyer also notes this habit of oversimplification and observes that the science is still young. She charts the trends we can expect as we balance our hope for understanding against the reality of the data available to us archaeologically. One of the ai:rns of the conference and of the resulting volume was to share insights based on the richness of ethnoarchaeological and experimental data. The papers presented here do indeed represent a wealth of both data and ideas about how to get behavioral meaning out of archaeological bone. It also became clear in the periods of open discussion that we see many ways in which to improve our use of the faunal record and the· types of data that experimental and ethnoarchaeological research methods can produce. We can better serve the archaeological community by making explicit the theoretical frameworks and the behavioral questions that concern us and by taking the extra step beyond particularistic descriptions to build analytic methods that can be tested in other actualistic studies and applied in purely archaeological situations. While we work to refine particular analytic approaches, we should
6J J.Hudson also expand our efforts to combine different types of analysis, both those concerned with other aspects of faunal remains and those involving other types of archaeological data. There is as well. a clear need to continue research aimed at integrating multiple stages in the sequence of cultural processes and in the interactions between cultural and noncultural processes .. These papers, and the discussions they inspired, are rich in data and ideas and suggest many productive avenues for future research on the best ways to get from archaeological bones to cultural behavior.
I.
1oth those conother types of ~arch aimed at ses and in the ;e papers, and suggest many . from. archaeo-
I.
Subsistence Strategies: .Meat Acquisition and Carcass Utilization
2.
Multiple Predation: A Definitive Human Hunting Strategy D. Gentry Steele and Barry W. Baker Abstract: Previous attempts to identify the diagnostic patterns of human hunting have emphasized its social aspect, comparing the· human pattern of predation with that seen in other social hunters such as wolves, lions, and nonhuman primates. The traditional assumption has been that the significant advantage of social hunting is that it has allowed the human hunter to prey upon larger species or to enhance the success rate of hunting. However, there has not been a clear distinction in the anthropological literature between the number of predators participating in a hunting event and the number of prey taken. We view multiple predation as a predatory strategy distinct from social hunting, both being strategies used to increase the harvest of animal resources. Though the traditional assumption has been that social hunting among humans coevolved with big-game hunting, it is more likely that social hunting coevolved within a broader hunting repertoire that included multiple predation, as well as hunting of single large prey. The purpose of social hunting and multiple predation is to efficiently increase the harvesting of animal resources so that fewer hunters can provide resources under a wide variety of environmental conditions for groups containing a disproportionate number of nonhunters.
Introduction An examination of the vast amount of archaeological research pertaining to human hunting and predatory strategies quickly reveals two directions in research.t The first is concerned with determining the antiquity of the human dietary regime. This line of r'esearch has established that the incorporation into the diet of significant amounts of animal protein can be traced back 1.8 million years ago, if not earlier. What has proven more difficult to establish is whether early hominids acquired these animals predominantly by hunting From Bones to Behavior: Ethnoarchaeological and Experimental Contributions to the Interpretation of Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights reserved. ISBN 0-88104-076-2.
9
10 I D. G. Steele cmd B.
W. Baker
live game (Bartholemew and Birdsell 1953; Dart 1953; Isaac 1976, 1978, 1980, 1983 Leakey 1971; Lee and DeVore 1968) or by scavenging (Binford 1981, 1984; Blumenschine 1986a, 1986b, 1987, l988a, 1988b, 1989; Bunn 1982; Bunn and Kroll1986; Potts 1982, 1984a, 1984b; Shipman 1983, 1986). The second direction of research, and the one that is. addressed in this paper, concerns identifying what is unusual and significant about human hunting and which features, if any, are uniquely characteristic of the human predatory strategy and understanding the logic behind human foraging behavior. For this paper we are restricting our, discussion to the issue of distinguishing hunting strategies and understanding how they function and are not considering the question of the evolutionary and ecological relationships of predation and scavenging. To date, analysts have emphasized that human predation involves tool using, social hunting, differential labor, cooperation, elaborate communication systems, sharing, beliefsystems, utilization .of large ho:t;ne ranges, and transportation and delayed consumption of prey (Hall and Sharp 1978; Isaac 1976, 1978, 1980; Lee and DeVore 1968; Marks 1976:204-205; Oswalt 1973; Peters and Mech 1975; Potts 1982, 1984a, 1984b; Suzuki 1975; Washburn and Lancaster 1968; Wolpoff 1980). Additionally, researchers have emphasized that humans,·unlike many predators, have an omnivorous diet consisting of both plant and animal matter that makes subsistence under marginal conditions possible and of value (Wolpoff 1980). Of the behavioral features that have been taken as hallmarks of human predation, the social aspect of human hunting has been one of the first recognized by scholars, the most intriguing, but at the same time one of the most difficult features to clearly recognize in the archaeological record or characterize as distinctly human. The r~ason for the difficulties in distinguishing human social hunting patterns from the social hunting patterns of other predators is that two distinct hunting strategies are commonly being considered simultaneously and referred to under a variety of names Stich as social hunting, communal hunting, communal drives, and cooperative hunting. The two hunting strategies we consider separately are social hunting and multiple predation. Of these two strategies, multiple predation is the less understood and is potentially one of the most distinctive hallmarks of human predation. Figure 2-1 illustrates the distinctiveness of the two strategies concerning how many predators are involved versus how many prey specimens are taken. The figure also indicates the relationship of the two phenomena.
Early Views on Social Hunting One of the most complete reviews of predatory behavior, which unfortunately does not include an analysis of human predatory behavior, is that of Curio (1976). While most of the nonhuman predators reviewed by Curio hunt as single predators, 'he also reviewed an array of social predators as diverse as spiders and lions. Among the social hunters, hunting of single prey specimens during one hunting episode was the rule. The hunting of single prey provided social predators with certain advantages: social hunters
Numt
improved t: and theyw As an ex, ing from 0 had succes; behavior o 1970, 1972J contention · Curio, fa: and attack€ ally. In thj gazelles we: zebra was: groups rest made it po: supported most likely by itself." · In reviev human sod the primar· prey could. sees the evi horses, bisc larger than major facto Since few o of larger p considered, among hun
Multiple Predation 111 76, 1978, 1980, )rd 1981, 1984; 182; Bunn and
Number of Predators Hunting
Number of Prey Taken
Single Hunter ~
.ressed in this about human . , of the human man foraging o the issue of y function and )gical relationlphasized that .allabor, coop!ms, utilization nption of prey ·e 1968; Marks , 1984a, 1984b; Additionally, lators, have an 1at makes sub,off 1980). 1r ks of human the first recog'ne of the most ~d or characterdistinguishing tterns of other { being consid; such as social e hunting. al hunting and tion is the less 1arks of human strategies conJrey specimens )henomena.
ehavior, which )ry behavior, is rs reviewed by ;ocial predators 1nting of single fhe hunting of ;: social hunters
Social Hunting Group
Single Specimens
Multiple Spec1mens
Sequential Predation of Single Specimens in Separate but Associated Episodes
Mass Predation in Singl·e Episode
Figure 2-1. Predatory strategies concerning number ofpredators participating in hunt and number of prey taken. improved their success ratio when hunting in groups rather than individually, and they ,were able to hunt bigger prey. As an1example, single hyenas hunting wildebeests had success rates varying from 0 to 18% of the time, while two or more hyenas hunting wildebeest had success rates varying from 32% to 100% of the time, depending on the behavior of the wildebeest upon confrontation with the predators (Kruuk 1970, 1972). Bertram (1979) provided similar evidence that supported the contention that social predators tend to have higher success rates. Curio, following Kruuk (1970), also pointed out that spotted hyenas hunted and attacked larger prey when hunting socially than when hunting individually. In this example, the average hyena group size hunting Thomson's gazelles was 1.2 individuals, while the average.hyena hunting group size for zebra was 10.8 individuals. Thus, for nonhuman social predators, hunting in groups resulted in more efficient capture of individual prey specimens and/ or made it possible to captu,re specimens of larger prey. This view was recently supported by Packer and Ruttan (1988:189) who stated that "cooperation is most likely when an individual has a low probability of capturing a large prey by itself." In reviewing the anthropological literature concerning the evolution of human social hunting, we feel many anthropologists also have assumed that· the primary advantage to social hunting is that single specimens of larger prey could be more succes.sfully taken by early humans. Certainly, when one sees the evidence of early humans' having killed animals such as mammoths, horses, bison, bear and a wide variety of ungulates (all animals substantially larger than humans), it is impossible to deny that social hunting has been a major factor in making it possible for humans to prey upon those animals. Since few other advantages to social hunting in humans other than the taking of larger prey species or the enhancement of hunting success rates were considered, the traditional conclusion was that the evolution of social hunting among humans coevolved with the evolution of hunting big game.
12 I D. G. Steele and B. W. Baker The basic scenario, then, was. that humans evolved from an early hominoid that may have been·a scavenger and/or an opportunistic predator analogous to the modern common chimpanze~ (Wolpoff 1980). With the adaptation to a more open savanna grassland environment, a segment of this lineage evolved into the Hominidae and developed a greater reliance on ~eat, acquired either through predation (Isaac 1978, 1983; Wolpoff 1980) or scavenging (Binford 1981, 1984; Blumenschine 1986a, 1986b, 1987, 1988a, 1988b, 1989; Potts 1982, 1984a,.1984b; Shipman 1983, 1986). With the evolution of Homo erectus, the adaptive shift was assumed to be toward a more.structured social hunting of single individuals of larger prey (Binford 1981; Shipman et al. 1981; Shipman and Walker 1989; Thompson 1979:135-179; Washburn and Lancaster 1968:297; Wolpoff 1980). To recapitulate the traditional sequence of hypotheses and correlated assumptions concerning human predatory patterns: (1) social hunting has .involved predominantly the hunting of solitary prey, and (2) as social hunting evolved, or the social hominids undertook social hunting, it primarily provided the advantage of preying on larger species or enhancing the success rate · of the hunt. In a large measure, because both of the hypotheses and assumptions either were based on at least some eviden·ce or followed a logical sequence of assumptions, it was, until recently, difficult to consider alternative or complementary hypotheses. Within recent years, researchers began to consider ecological models and optimal foraging theory as an explanation why human and nonhuman predators hunt the way they do and how such practices may have affected human evolution (cf. Driver 1990:21-28; Earle and Christenson 1980; Hawkes 1990, 1991; Hawkes et al. 1982; K. Hill 1982, 1988; Hill et al. 1987; Jochim 1976:19-45; Jones 1990:70-71; Kamil et al. 1987; Kamil and Sargent 1981; Packer and Ruttan 1988; Shipman and Walker 1989; Slobodchikoff 1988; Smith 1980, 1983; Stephens and Krebs 1986; 'furner 1989; Webster and Webster 1984; Winterhalder 1987; Winterhalder andiSmith 1981). Such models emphasize the maximization of benefits versus costs in the procurement of food.
Human Predatory Strategies By separating the question of how many predators participated in the hunt from the question of how many prey were taken during a hunting episode, it permits us to focus on the issue of how adequate animal resources are acquired rather than on the social behavior of the hunt. Anthropologists have commonly hinted at the distinction between the number of prey versus the number of predators involved in a hunting episode, though none have examined this distinction in detail or considered the differences in an interpretive framework. Saunders (1977) for example, noted that entire families of mammoth may have been hunted prehistorically, in addition to the taking of single individuals by communal efforts. Driver (1990:12) stated that communal hunting is
not defined prey and pr game sped where a nu number of' haps with o To faciliti cooperative l act of two o emphasis is no refereno (1976:199) c predators. In additic added emp communal] hunting," iJ hunting grc cooperative two ormon Research relatively 1 apparently Examples c fishes (Cur: Murie 1944 (Teleki 1972 Lee and De human sod In Curio' except one, single prey predators h social hunti: hunting its' killed. The ' single hunt practice, thE instances. 1 predation.
social organ as a separa1
qr··
Multiple Predation j13 ~arly
hominoid ator analogous :tdaptation to a lneage evolved :tcquired either 1ging (Binford . '89; Potts 1982, mo erectus, the ,cial hunting of 1981; Shipman :aster 1968:297; :md correlated al hunting has ; social hunting primarily prothe success rate 1mptions either al sequence of 'native or com~an to consider on why human 1 practices may nd Christenson L988; Hill et al. '87; Kamil and Walker 1989; :6; Turner 1989; nd Smith 1981). ~osts in the pro-
: participated in 1ring a hunting nimal resources ~nthropologists
not defined by the number of prey taken but may involve varying numbers of prey and predators. In addition, Kooyman (1990:330) noted: "Hunting any big game species can be accomplished using two basic techniques: communal, where a number of hunters coordinate their efforts (and usually procure a number of animals), and individual, where hunters work in isolation (or perhaps with one or two other hunters), usually to capture a single prey animal." To facilitate the separation of these phenomena, the phrases social hunting, cooperative hunting, and communal hunting are used here strictly to refer to the act of two or more predators actively pursuing or killing prey. In this vein, the emphasis is on the number of predators involved in the act of hunting, with no reference to the number of prey taken. This use is very similar to Curio's (1976:199) definition of "communal hunting'' in his discussion of nonhuman predators. In addition to the number of hunting participants, Frison (1987:182), placed added emphasis on aspects of planning and temporary authority in defining communal hunting. Hayden (1981:421) distinguished "small-scale cooperative hunting," involving two to four individuals, from a "large-scale communal hunting group," involving five or more individuals hunting in a coordinated, cooperative fashion. Driver (1990:12) viewed communal hunting involving two or mbre cooperating hunters with a preconceived plan. Reseafch on nonhuman predators has revealed that there are in fact a relatively large number of social hunters and that social predation has apparently evolved independently in a wide variety of animals (Table 2-1). Examples of social hunting can be seen among spiders (Wilson 1972) and fishes (Curio 1976:210) and among mammals such as canids (Mech 1970; Murie 1944), hyenaids (Kruuk 1970, 1972)~, felids (Schaller 1972), pongids (Teleki 1973), and hominids (Coon 1971; Davis and Reeves 1990; Forbis 1978; Lee and Devore 1968; and Nitecki and Nitecki 1987 are general sources on human social hunting with extensive bibliographies). In Curio's (1976) discussion of social hunting, all examples presented, except one, discussed how the hunting groups under consideration attacked single prey specimens. Thus, while his definition considered only the act of predators hunting in gro·ups, his discussion suggested a close correlation of social hunting with killing of single animals. Here we consider only the act of hunting itself, divorced from the question of how many specimens were killed. The actual number of prey specimens killed by social hunters during a single hunting episode either may be a single animal or many animals. In practice, the majority of social hunters do kill single animals, but not in all instances. These exceptions are highly significant to understanding human predation.
r of prey versus
mgh none have sin an interpremammoth•.may f single indi vidmnal hunting is
Multiple Predation By restricting the definition of social hunting to the question of the social organization of the predator, it i~ possible to more thoroughly examine as a separate issue the question of how many prey specimens are taken by
Table 2-1. Examples of Hunting Strategies for Selected Social Mammalian Predators as Related to Prey
Number and Prey Size
,........ ~
~
0
Multiple Prey Single Prey
Taxa
Sequential Predationa
Mass Predationa
~
("I:) ("I:)
Prey Sizeb
Author Labelc
Reference
~ ~
;:::: :;.::;...
Order Carnivora Family Canidae Canids Canis latrans Coyote
S,M,L
*
Cooperative hunting
Macdonald (1984:62)
tJj
;E. tJj ~
~ -....:
Canis lupus Gray wolf
*
*
Canis rufus Red wolf
*
*
Canis dingo Dingo
*
Canis aureus Golden jackal
*
Canis mesomelas Silverbacked jackal
*
Cuon alpinus Dhole, red dog
*
Lycaon pictus · African, or cape hunting dog
*
Table 2-l.-continued
-...___
S,M,L
Pack hunting
Anderson (1982); Macdonald (1984:58-61); Mech (1970); Sharp (1978).
S,M,L
Pack hunting
Anderson (1982); Macdonald (1984:5&-61); Mech (1970); Sharp (1978)
S,M,L
Pack or cooperative hunting ·
Anderson (1982); Macdonald (1984:84)
S,M,L
Cooperative hunting
Macdonald (1984:64)
S,M,L
Cooperative · hunting
Macdonald (1984:64)
S,M,L
Cooperative hunting
Anderson (1982); Macdonald (1984:80)
S,M,L
Cooperative or ·organized hunting
Anderson (1982); Macdonald (1984:76-79)
~~.
~~ ~~~~-~~
)~-~-·
S,M,L
Cuon alpinus Dhokred dog
S,M,L
Lycaon pictus African, or cape hunting dog
Cooperative
Anderson (1982);
hunting
Macdonald (1984:80)
Cooperative or organized hunting
Anderson (1982); Macdonald (1984:76--79)
Table 2-1.-Continued Multiple Prey Taxa
Single Freya
Sequential Predationa
Mass Predationa
Speothos venaticus Bush dog
Author Labelc
s;M
Pack or cooperative hunting
*
Family Hyaenidae
Crocuta crocuta
Prey Sizeb
*
*
*
Anderson (1982); Macdonald (1984:85)
S,M,L
Kruuk (1972:89, 204); Group hunting and mass killing Macdonald (1984:154158)
S,M,L
Group hunting
*
Spotted, or laughing hyena
Hyaena hyaena
Reference
Macdonald (1984: 154158)
Striped hyena S,M
Family Felidae
None
Anderson (1982)
Communal, pride or multiple hunting
Macdonald (1984:29-35); Schaller (1972:251, 437)
Cooperative hunting
Macdonald (1984:169)
Felis pardalis Ocelot S,M,L
Panthera leo Lion
Order Cetacea Family Delphinidae
Delphinus delphis
S,M *
*
*
Killer whale
1--'
::r.
~
ro1-d '"1 ro 0.. Pl
::r.
Common dolphin
Orcinus orca
a:::
c
0
::s
'>f-
S,M
*
*
Cooperative hunting
Macdonald (1984: 190191)
-......... (.11
Table 2-1.-Continued
~
0\
>:J
Multiple Prey Taxa
Sequential Predationa
Single Prey1
0
Mass Predationa
Prey Sizeh
Author Labezc
t:.l'l ....... ('I)
Reference
~
('I)
Family Monodontidae
S,M
l':l
;::s
Cooperative hunting, group herding
Macdonald (1984:201)
S,M
Simple cooperative hunting, coordinated hunting
Macdonald (1984:377); Strum (1981 :262, 274-275)
S,M
Simple cooperation, cooperative hunting
Lawick-Goodall (1968); Macdonald (1984:424); Teleki (1973, 1981:323, 331-333); Wrangham
Communal, cooperative, social, and group hunting
Coon (1971); Davis et al. 1986; Davis and Reeves (1990); Forbis (1978); Lee and Devore (1968); Nitecki and Nitecki (1987), and references therein
Delphinaptems leucas Beluga, belukha, or white whale
~
lJ:!
;E tci
Order Primates Family Cercopithecidae
Papio cynocephalus Savanna baboon Family Pongidae
Pan troglodytes Chimpanzee
l':l
~ """'!
(1975)
S,M,L
Family Hominidae
Homo sapiens Humans
a Asterisk = denotes practice of this hunting strategy. bS =small (0-100 g); M =medium (100 g-25 kg); L =large (greater than 25 kg).
cRefers to the labels used by the authors cited to define social hunting patterns associated with the taxon.
:s
re
as
s sr s
s· s
s !:I0 &. s o.. ~- ~ s=- --o o s· o.. ~ 0 n 0 ~~~o~oo@~OQ h3 2. 2· p ::t' ~ ~ 0.. PJ 8 p;· 0.. :=; n~ ~ 11> n- "" ~ 0.. H .......
~ ~ t:L s=- ~ t;;· '"d rt 8 ~- i:'r ~ ~ s=- 0' B fti ~ ~ s=< OQ
~ ~ ~ ~ 2.. --g § ~,.J
•s ...
~ ~ ..... ~
s:-
oo
s· < ~ ...ro~
1-'•0Q
00
!:Is·
0
.......
~ ro • ...... !4. 1-j Q...
~
.............. ::::M ~ ~ 11>-t-,0• ..
s:- g- % § .g. e ~ 0' ffi
0
~
~
rt o ::r s 0.. 11>
......
~
.......
o.. ?\
u '"d
;::l.
1-'4
.-J
>-+.
(t
Pl ::=:
1-f
0
;::::>
~ ~· Pl
~-!:Is~~!:::!
~ §..~ st- ~ ~ ~ ~- ~ ~ : ~ ~ -.
o..
~ '"M
(i)
::;::'
OO
e ~ s::r. ::::· ::r. o ;r ~ '"""'"', ~- ~ ~ ~ ~ ::s: ~ ~ s -::tro.-J ft ~ ~ E. g :! §- ~ _; ~ ri o.. ffi"
Ft ~ o.. Q. .r> ft g ro ~ - ~ ........ 0
'" ..... 0 !T.-::::
..... ('"\ -
>-1->00
11>
>-!->
......
.-J
~ ~ OQ
-
11> . ,.J "'" 11> vuoo>-1"-\~
,.J
11>
0
P>'
~ ....
,.J
Multiple Predation 117
.s QJ
1-
eople congre(5) existing in ploi ted by the ave practiced 1ammals such ey have been their remains rtise of prehism, become all ·esource of the ~ricas and the ts, subsistence ple predation ! as a response (musk-ox and 1ups. Multiple e responses. 1unting of the ties that were l of the welley include the tgico-religious d, and the use r, it is the most the only form. those societies considered as :ies emphasize schooling fish. lment may be ganized effort .at in northern :essful because mals and fish, nge of species ~ prey, and the an be effective taking of more g the harvest.
I>
;~\, ;, , ,-
1-
z
0:::
Figure 3-3 summarizes variation in anatomical representation for medium-sized ungulates in faunas generated by the six pret?-ators at shelters. Predators are listed on the x-axis, and anatomical profile types appear on the y-axis. The profiles range from being head- and/ or horn-dominated to more or less anatomically complete, and then on to slightly "leggy." The __chart illustrates the range of patterns for each predator, but not the total number of cases. At the high left corner are cases generated by known obligate scavengers (brown and striped hyenas), and cases generated by known obligate hunters (wolves) appear at the opposite corner in the bottom right. The generally diagonal arrangement of cases across the chart suggests that some larger relationship exists between procurement strategies and part transport choices. Experimental situations in which obligate scavengers are provisioned with whole medium-sized ungulate carcasses near their dens (outlier case in lower left corner of Figure 3-3; Skinner et al. 1980) indicate that everybody likes to carry away legs when presented with the choice. In foraging situations subject to natural constraints, however, scavengers tend to move heads and/ or horns more often than other items due, at least in part, to what is usually l~ft over at death sites. Predators that primarily hunt move relatively more 1parts, and especially more legs, to shelters. The Upper Paleolithic human series is packed tightly together; it is relatively invariant at this scale of comparison and generally resembles the wolf series. Predators that do a lot of both hunting and scavenging, such as spotted hyenas, show good representation throughout the possible range. The same is true for the Middle Paleolithic series. A surprisingly simple anatomical ratio, (H+H)/L, describes much of the variation among cases (Figure 3-4), not only in the kinds of patterns typically produced by each predator but also in-the modes and ranges of variation. The (H+H)/L index is obtained by combining unstandardized horn or antler plus head counts and then dividing the sum by the total unstandardized number of limb elements (except the phalanges; see Table 3-1). It should be noted that the relative amounts of leg parts vary most, whereas heads and/ or horns are well represented in nearly all cases. The ranges of variation in the {H+H)/L ratio are especially important for seeing niche differences among the predators because the general tacticshunting and scavenging-are not unique to any of the species considered and certainly will not, in and of themselves, distinguish subspecies of Homo
[ [
H
w
11-
H E A D L
.
E G
s
r points, multile points or other ture of particular :he collector'stratleS of aggregation ~er I caribou), or if
Experimental manufacture and use of single-beveled bone points has demonstrated the efficacy of the hafting mechanism first proposed by D. and E. Peyrony (1938; Figure 4-3). The distal end of a wood shaft was whittled to a bevel with an angle that matched that of the beveled bases of the points. During experimentation, resin was applied to the beveled surfaces to increase adhesion. The irregular sm::face of the bevel created by the natural openings of . the cancellous bone from which it was manufactured (Figure 4-2a), or from intentionally made striae, allowed for increased tenacity between the base and the end of the handle. During the Perigordian, the shafts of single-beveled points were sometimes scored with incisions perpendicular to the axis at the distal end of the bevel (Figure 4-2d). It is likely that these incisions reduced slippage ·of the ligature with which th~ point was bound to the shaft. The superior surface opposite the distal end of the bevel of a number of
72IA. Pike-Tay and H. Knecht 25
Figure 4-3. Proposed technique for hafting single-beveled points.
Perigordian single-beveled points is flattened and marked with oblique or perpendicular incisions (Figure 4-2c). This combination of features would allow for additional cohesion of lashing and/or adhesive. With the production of textured bevels, incised shafts, and flattened surfaces, there appears to have been an emphasis on maintaining the attachment of the haft in the design of Perigordian single-beveled points. This may reflect that the spear was highly stressed at the level of the haft during use. Results of the projectile experiments suggest that armed with spears tipped with single-beveled bone points Perigordian hunters would h~ve been capable of dispatching both large and small animals. With forces well within the range of spears projected with spearthrowers, the spears were found to easily enter and even pass through the animal. The projectile trajectory was unimpeded by the streamlined contours of the single-beveled points attached to single-beveled handles. Given the range of single-beveled point lengths (Knecht 1991), it is probable that the points were resharpened following breakage during manufacture, handling, or use. Among the assemblage from Laugerie-Haute are several single-beveled points that appear to have been resharpened (Figure 4-2b). It would be possible to resharpen a broken point in the field, that is, during a hunting episode, without detaching it from the spear shaft. The relatively unrefined appearance of the retooled points may indicate ad hoc resharpening of the projectile points by a hunter. The tight range of variation of bevel angles (Figure 4-4; Knecht 1991) suggests the possibility of reusing the same spear shaft for the hafting of different single-beveled points. Spare projectile points could therefore have been carried by the hunter to serve as ready-to-use extra components.
(n=64)
4-!
Figure
Laugeri.
Season1 The ba surrounding the n because all past 1 Dordogne has beeJ noted differences from the region ar nomenon. Howe\ important Upper First, since our ain that additional·pn small portion of o: that of reindeer. Season-of.:.death and 4-6) are from I for study in part b a fair amount of J sample was camp teeth recovered w accounted for. Cor ber of individuals Appendix C) proh:
Paleolithic Hunting Strategies I 73 25
(n=64)
201---····'"---·---------------------··------······-··--···--···----·...............
15 .......................................
5
0 ~led
points.
4-5
6-7
8-9
10-11
12-13
14-15
Bevel Angle (0 ) ~ with oblique or f features would With the produc' there appears to 1f the haft in the ~ct that the spear
rith spears tipped rould have been ·orces well within .rs were found to ile trajectory was d points attached )91), it is probable ing manufacture, -Iaute are several :i (Figure 4-2b). It , that is, during a ft. The relatively hoc resharpening on of bevel angles .g the same spear e projectile. points ·eady-to-use extra
Figure 4-4. Distribution of bevel angles of single-beveled points from Laugerie-Haute.
Seasonality and Age Profiles of Upper Perigord ian Prey: Red Deer The background discussion above pertained primarily to issues surrounding the nature of reindeer hunting during the EUP in the Dordogne because all past research concerning Upper Paleolithic subsistence in the Dordogne has been concerned with a single species emphasis on reindeer. We noted differences of opinion as to why reindeer dominates the faunal lists frbm the region and even.derived our main research questions from this phenomenon. However, most of the data presented here pertain to another important Upper Perigordian cervid prey-the red deer-for two reasons. First, since our aim is to illuminate the nature of EUP hunting, it is important that additional prey species be considered. Second, the data represent just a small portion of ongoing research in which red deer analyses have preceded that of reindeer. Season-of-death determinations q.nd age profiles of the red deer (Figures 4-5 and 4-6) are from Pike-Tay (1989, 1991b). The faunal assemblages were chosen for study in part because of their excellent preservation. Although there were a fair amount of hemi-mandibles and partial maxillae, the majority of the sample was composed of isolated teeth. All DP4, P4, Ml, and M2 red deer teeth recovered were analyzed to insure that every individual red deer was accounted for. Conservative linking of teeth to determine the minimum number of individuals (:MNI) represented by teeth (detailed in Pike-Tay 1991b: Appendix C) prohib~ted problems of overrepresentation by a single animal.
741 A. Pike-Tay_ and H. Knecht
LA FERRASSIE, level D2 spring, summer
LE FLAGEOLET I, level 7 late fall through late spring season of death
season of death
Fall
Fall
late F/early W
late F/early W
Winter
Winter
late W/early Sp
late W/early Sp
early Spring
early Spring
Spring
Spring
late Spring
late Spring
late Sp/early S
late Sp/early S
Summer
Summer late Sull}mer
late Summer
0
4
4
Number of teeth sectioned =18 Dental MNI = 12
Number of teeth sectioned =17 Dental MNI = 12
RED DEER HUNT
RED DEER HUNT
LES BATIUTS, levelS winter
ROC DE COMBE, 1 (a,b,c) late summer through early winter'
season of death
season of death
F
Fall
Fall
pt
late F/early W
late F/early W
Ct
Winter
Winter
late W/early Sp
late W/early Sp
early Spring
early Spring
Spring
Spring
late Spring
late Spring
le B.
late Sp/early S
late Sp/early S
Summer
Summer
In sum, se< 1. (l,
le b< ca
late Summer
late Summer 0.2
0.4
0.6
0.8
Dental MNI = 2
RED DEER HUNT
0.2
0.4
0.6
0 ..8
1.2
nc 2. St
Number of teeth sectioned = 5 Dental MNI = 3
RED DEER HUNT
Figure 4-5. Seasons-of-death of red deer as determined by cementum annuli analysis by Pike-Tay (1989).
8(
3. p< d 4. 1) d~
Paleolithic Hunting Strategies 175 PERIGORDIAN SAMPLE 1RASSIE, level D2 ,ring, summer
age profile of RED DEER based on individual teeth 16
number of teeth sectioned
14 !-························
12
!--·······················
10
8 6
4 of teeth sectioned =17 )ental MNI = 12
D DEER HUNT
2 0
.5-2
3-4
5-6
7-9
10+
age classes >E COMBE, 1(a,b,c) 1er through early winter
MIXED-AGE-PATTERN (n=29)
Figure 4-6. Mixed age profile (an overlay of attritional and -prime age patterns?) of red deer. This graph is a composite of data derived from cementum annuli counts of individual teeth from the Upper Perigord ian levels specified in the text at the sites of Le Flageolet I, La Ferrassie, Les Battuts, and Roc de Combe. In sum, season-of-death determinations of red deer are as follows:
'r of teeth sectioned = 5 Dental MNI = 3
::D DEER HUNT
ined by cementum
1. Fall through spring (cold season) hunting is indicated at Le Flageolet I (level· 7). The number of identified specimens (NrSP) counts from this level-compose over 50% red deer, followed by about 36% reindeer, with bovines (Bas/Bison sp.), horse (Equus caballus), chamois (Rupicapra rupi- · capra), roe deer (Capreolus capreolus), and ibex (Capra ibex or Capra pyrenaica) in decreasing frequencies (Delpech 1983:352). 2: Red deer hunting at La Ferrassie (level· D2) occurs spring through summer (warm season). Here this taxon dominates the NISP count at 80.1% (Delpech 1983:348). 3. Les Battuts (level 5) shows winter hunting of red deer. Red deer compose !).early 50% of the NISP count, followed by ibex, bovines, horse, and chamois in decreasing frequencies (Delpech 1983:358). 4. Red deer were taken summ~r through winter at Roc de Combe (level 1). At this site, reindeer dominate the NISP count at over 80% with red deer reaching about 5% (Delpech 1983:343).
76IA.Pike-TayandH. Knecht The age profiles of the red deer .from the Upper Perigordian levels sampled at Le Flageolet I, La Ferrassie, Les Battuts, and Roc de Combe show a broad mix of all age classes (an overlay of prime age and attritional patterns? [Figure 4-6]) derived from relatively small numbers of individuals in each level. Age profiles were constructed for each individual site, but as they are detailed elsewhere (Pike-Tay 1989, 1991b) and differences among them are slight, they are combined here.
Seasonality and Age Profiles of Upper Perigordian Prey: Reindeer Analysis of the reindeer from level4 of Abri Pataud (Spiess 1979), which is roughly contemporaneous with the assemblages from which red deer were sampled, offers data that complement the findings outlined above. Spiess (1979:185-208) claims that the reindeer-dominated fauna from all Upper Perigordian levels of Abri Pataud suggest fall, winter, and early spring (cold season) hunting. He estimates an MNI of 26 reindeer for level4 (Spiess 1979). . As regards the demography of hunted reindeer during the Upper Perigordian at Abri Pataud, Spiess (1979:185) notes the consistently small MNis, sees the same demographic pa,ttern throughout all Upper Paleolithic levels, and concludes: 1. The total demography of all bands of caribou wintering within range of the Abri Pataud was not significantly different in composition from the total "natural" population. . 2. The hunters from the Abri Pataud randomly selected animals to be killed, or used a hunting technique that did not allow selectivity (and was not a large-scale-drive, either).
Whilereindeer was the dominant taxon at Abri Pataud, Spiess's a/~alysis of faunal data from other species present led him to generalize for all levels: "Animals were killed singly or in small numbers, including everything from chamois to 1500-kg cattle. There is no evidence of large-scale drives or slaughters" (Spiess 1979:185).
Conclusions Which approach, then, was taken by the Upper Perigordian hunters at the sites considered here? Are the data presented above, limited as they are at present, consistent with specialization of hunting strategy or with a more general response to local availability of resources? As noted previously, the behavioral correlates of the former, the collector type of approach, would include decision making based 011 the costs and benefits of particular taxa. The second situation is more consistent with a generalized and opportunistic foraging approach. In light of the Upper Perigordian data, we now return to the set of hypotheses pertaining to specialized collectors and generalized foragers. For the
reasons previot (1) the age pro (reflecting high tive fuscavengin~ aggregation for for a substantia profile (indicat forces would 1 hunting techni redundant par1 nized by the pr ing special typE H a generaliz (1) the age prof and composed ' numbers that v and deviCes sw small-scale sm (however, as la not expected, tl given season, 1 being suggesti' versatile, with would be char recognized arcl We recognizE egy other relev ence and absen remains, estim2 levels concerne Our purpose he hand and to su~ In sum, our F are: 1. A~ and attril all a seco: 2. Ba .dian no tic sing: evid,
Ther spec: tile t
Paleolithic Hunting Strategies 177 n levels sampled Je show a broad ?atterns? [Figure L each level. Age hey are detailed n are slight, they
Prey: Reindeer ud (Spiess 1979), n which red deer outlined above. fauna from all and early spring Jr level 4 (Spiess ~ing the Upper 1nsistently small Tpper Paleolithic
~ring
within range composition from
:ted animals to be w selectivity (and
less's analysis of ze for all levels: everything from drives or slaugh-
per Perigordian above, limited as ;trategy or with a. toted previously, approach, would f particular taxa. nd opportvnistic Le set of hypotheoragers. For the
reasons previously elaborated, we would expect that for specialized collectors (1) the age profiles of the prey would be prime-dominated or catastrophic (reflecting highly skilled individual hunters, successful game drives, or selective scavenging); (2) the season-of-death would be consistent with times of aggregation for migratory species or, if game drives were employed, the same for a substantial number of animals who together constitute a catastrophic age profile (indicative of mass kills), or to at least cluster seasonally since task forces would be responsible for procurement; and (3) the technology and hunting techniques would be "reliable;" that is, overly specialized with redundant parts, essentially failure-free. Such technologies would be recognized by the presence of special purpose projectile points designed for hunting special types of game or for use under special hunting conditions. If a generalized, opportunistic approach were taken, then we would expect (1) the age profile to be mixed (i.e., a combination of attrition and prime-age) and composed of relatively small numbers of individual animals (the ages and numbers that would be taken by sufficiently skilled hunters with techniques and deviCes such as snares, traps, deadfalls, encounter ambush, stalking, and small-scale surrounds); (2) the season-of-death to also cluster seasonally (however, as large-scale drives or intensive high-yield hunting episodes are not expficted, the season-of-death for a few animals should be spread across a given season, rather than indicating many simultaneous deaths, the latter being suggestive of mass kills); and (3) technologies to be maintainable and versatile, with generalized application to a variety of circumstances. They would be characterized by spare or interchangeable parts that would be recognized archaeologically by morphological standardization. We recognize that to ultimately distinguish foraging from a collector strategy other relevant site data must be considered as well. Of interest are presence and absence of full residential residues in sites, body part bias in prey remains, estimates of group size per site, degree of temporal resolution of the levels concerned, and season~of-death and age profiles for all species present. Our purpose here, however, is to offer a preliminary assessment of the data in hand and to suggest a vi~ble methodology for future research. In sum, our preliminary conclusions concerning the Upper Perigordian data are: 1. Age profiles for red deer from Le Flageolet I, La Ferrassie, Les Battuts, and Roc de Combe and the reindeer from Abri Pataud do not suggest attrition. Therefore, either the hunters were skilled enough to pick from all ages or they cooperated to procure mixed age herds. However, the second approa5=h is not supported by the low numbers of individuals. 2. Based on the analysis of bone single-beveled points, Upper Perigordian projectile technology appears quite adequate to accommodate the notion of fairly skilled hunters. Several characteristics of the Perigordian single-beveled points, including standardization of bevel angle and evidence of retooling, suggest a portable, "serviceable" weapons system. There does not seem to be evidence of special-purpose projectiles for special game or hunting conditions. In sum, Perigordian organic projectile techl).ology appears to be, in Bleed's terqlinology, "maintainable"
78j A. Pike-Tay and H. Knecht
and thereby indicative of a foraging hunting strategy rather than logistical collecting. 3. Although seasons-of-death o.f the red deer from four sites and of reindeer from the Abri Pataud during the Upper Perigordian appear· to cluster seasonally (with the exception of Roc de Combe), detailed breakdowns (Pike-Tay 1989, 1991b) show small numbers of animals taken across a given cold or warm season during the Upper Perigordian rather . than many animals taken simultaneously. The spread of kills at La Ferrassie (Figure 4-5b) across the warm SE?ason illustrates this well. Thus, the seasonal distribution of red deer kills is consistent with Spiess's (1979:185) interpretation of cold-season reindeer hunting at Abri Pataud: "Animals were killed singly or in small numbers, ... There is no evidence of large-scale drives or slaughters."
In conclusion, our preliminary data regarding age structure and season-ofdeath of prey do not appear to be consistent with a collector strategy at least as concerns the procurement of red deer during the Upper Perigordian at Le Flageolet I, La Ferrassie, Les Battuts, and Roc de Combe and of reindeer at Abri Pataud. Upper Perigordian organic projectile technology seems to be lacking several optimal design features characteristic of the weapons systems of logistical collectors. Much future work remains to be conducted to elicit the exact nature of EUP hunting techniques and strategies. However, the limited evidence available at present suggests that a generalized and opportunistic foraging approach was operationalized for the capture of game resources · during the Upper Perigordian.
Acknowledgments We wish to thank Jean Hudson, Jean-Michel Genestek James O'Connell, Randy White, and an anonymous reviewer for valuable c~mp1ents on an earlier version of this paper. We thank Francessing of such processing for ~cting has been d Upper Paleo. The ability to aid us tremenld culture. hat are responquestion of the 10rtant food re.unal resources 1sitional or non'80; Brain 1981; ssing is the best ites are usually d Rigaud 1974; tion of marrow :ollecting? One of remains that uired food verlput and subsemer case, meat ;articulated just rocessing might Meat would be racked by each ~ logistical case, o be dried and e to obtain the lt is a matter of onsumption or tption. ·een intentional ~dental to meat 1rable residues ;ical resource is
often overlooked because it is difficult to deal with and the conventional rewards are frequently meager. Bone splinters rarely yield information applicable to such conventional goals as determining past environmental conditions from species representation or counting minimum numbers of individuals; who knows how many splinters come from a single bone? That obviously depends on how many times the bone is broken, and to what ends. The goals here are to examine some attributes that can be observed in archaeological samples and to relate them to subsistence organization behavior known through ethnoarchaeological research.
Data Base Observations were made on assemblages of caribou (Rangifer tarandus) bone splinters, which had been recovered from ethnographically
known situations among the Nunamiut Eskimos of the Brooks Range in the interior of Alaska by Lewis R. Binford (1978:428-447, 1983:176-184). Two samples came from distinct activity areas at a site known as Palangana's House. This winter residence was occupied in the late nineteenth century. Stone tools figure significantly in the processing of faunal materials from the site. Sample P1 (N = 181) consists of material from debris associated with mass processing fqr marrow. The bones had not been previously cooked or processed, except for dismemberment and removal of meat (if there had been any). A large number of bones were broken open at one time, and the raw marrow was removed. Sample P2 (N = 624) consists of material from a domestic cooking dump, that is, debris resulting from consumption of meals. Usually, joints ofmeat were cooked in a stew pot. After the meat was eaten, the bone were cracked open and the marrow was eaten. The dump, therefore, was the product of an accumulation of repea.ted consumption events. The third sample comes from a site known as the Bear site. This winter house was occupied in 1948. Steel tools were used during the occupation. Sample B3 (N = 528) consists of materials from a kitchen midden similar to that of P2. These samples were derived from restricted areas (1m2), in order to avoid mixing of the contents with debris from other activities. They do not represent the range of activities that might have been carried out at either site but were selected to highlight the differences between mass processing of ni.arrow and marrow cracking incidental to meat consumption at meals.
Analysis A model for differentiating between breakage during mass processing of raw m.arrow and breakage for immediate consumption can be proposed. In the first case, elements present should primarily reflect their high marrow utility (Binford 1978). When a large quantity of the same elements are processed at the same time, there should be a standardized procedure for breaking them, that is, the points of impact should be in the same anatomical
86j
J. G. Enloe
locations on each,bone. One might expect that this activity would result in relatively few impact cones per element and relatively long fragments, reflecting maximum efficiency in breakage (Binford 1978:155, Table 4.6). In the second case, element representation should reflect high meat utility. Since the bones would be broken incidental to m'eat consumption, they would be broken one at a time by various consumers. Thus, one would not expect that any standardized pattern of breakage would be evident. Impact fractures would be more variably located and more numerous, and fragments would be relatively short. In order to test those expectations, observations were made on each fragment for element, length, and number of impact cones. Table 5-1 presents the element representation of the three samples. It must be noted that not all fragments could be identified as to skeletal element. That is a very common problem in dealing with bone splinters. The unidentified fragments most likely come from the shafts of the major upper long bones, that is, humerus, radius, femur, and tibia. Recognizability is a critical criterion here. The major landmarks on the bones are mostly located at the articular ends, and there are long spans of what are essentially tubes between them. Various fragmentation processes result in the formation of bone splinters separated from their articular ends; those fragments were put into a category labeled Unspecified. The metacarpal and metatarsal are much more recognizable than the upper limb bones owing to the deep groove and crests on their posterior surfaces, so a larger percentage of them could be correctly identified. The anterior surface presents another problem, however. The metatarsal is very distinctive and easily recognized, but the anterior portion of the metacarpal shaft is much less distinctive. Therefore, anterior metacarpal fragments lacking any portions of the posterior surface could have been placed in the unspecified category. Posterior portions of either of those elements lacking proximal, cJistal,_ or cranial portions had to be placed into the unidentified metapodial category .. The final category (total metapodials) combines all of the fragments of metacarpals, metatarsals, and unidentified metapodials.
Element Representation The first test of the expectations deals with element representation in each assemblage (see Table 5-1). Strong preference for marrow-bearing elements in assemblage composition i~ argued to reflect mass-processing events, whereas no preference or preference for meat-bearing elements is argued for consumption incidental to meat consumption. Figure 5-1 displays the relative percentages of each element for all three cases. It is clear from this figure that assemblages P2 and B3 are almost identical to one another, while assemblage P1 appears radically different. All three track very closely through the upper limb elements. Where .they diverge is in the unspecified and metapodial categories, with relative differences of at least 36% and 43%, respectively. If, as I suspect, the unspecified fragments really do come from the upper limb bones, then what we are seeing is a divergence between
Tab]
Element Humerus
Radius. Metacarpal Femur Tibia Metatarsal Phalanges Unidentified metapodial Unspecified Total metapodialsa aIncludes all metac;
assemblages de dominated by rr To clarify the; percentages, di1 were omitted ft threeassemblag assemblage, wE Figure 5-2, whE indices (Binforc (Binford 1978:81 residue assemh element represe assemblage, lie elements of the crosses the lines both meat and: ences in elemen1 bearing element both of the cons lines. Elements 1 the line for the has relatively 1· important marrc similarity of th1
Subsistence Organization I 87 vould result in ~ents, reflect- >le 4.6). In the tility. Since the hey would be not expect that· tpact fractures aen ts would be ~
on each frag;-1 presents the
i as to skeletal ~splinters. The te major upper snizability is a ostly located at sentially tubes e formation of nents were put than the upper ·ior surfaces, so mterior surface distinctive and aft is much less a.ny portions of :ified category. imal, distal, or odial category. ~ fragments of
; representation tarrow-bearing Lass-processing ng elements is 1re 5-1 displays ; clear from this another, while closely through nspecified , and 36% and 43%, · do come from gence between
Table 5-1. Bone Splinters from Three Nunamiut Samples Pl Mass Processing NISP %
Element
Humerus Radius Metacarpal Femur Tibia Metatarsal Phalanges Unidentified metapodial Unspecified Total metapodialsa I
P2 Meal Midden NISP %
B3
Meal Midden NISP %
267 81 5 63 52 43 1
11
24 37 0
4 10 1 6 13 20 0
13 1 10 8 7 0
50 52 7 55 70 23 0
9 10 1 10 13 4 0
62 19
34 10
34 278
5 46
25 246
5 47
101
56
82
13
55
10
8 18 2 11
alncludes all metacarpals, metatarsals, and unidentified metapodials. assemblages dominated by the meat-bearing elements and the assemblage dominated by marrow-bearing elements. To clarify that difference, a second calculation was made of the relative percentages, disregarding the unspecified fragments _(Table 5-2). Phalanges were omitted from consideration .since there was only one identified in all three assemblages. Similarly, metacarpals, which constituted only 1% of each assemblage, were omitted. The percentages are graphically displayed in Figure 5-2, where the el~ments are ordered with the highest meat utility indices (Binford 1978:23) to the left and the highest marrow utility indices (Binford 1978:81) to the right. The lines for P2 and B3, the meat consumption residue assemblages, remain quite similar to one another in proportions of element representation. The line for assemblage P1, the marrow processing assemblage, lies substantially below the other two for the meat-bearing elements of the upper limbs but climbs sharply to the right of the graph. It crosses the lines for the other two at the tibia, which has significant values for both meat ahd marrow. In this figure the contrast demonstrates the differences in element representation between the assemblages biased toward meatbearing elements and the one biased toward marrow-bearing elements. Here, both of the consumption debris assemblages, P2 and B3, exhibit relatively flat lines. Elements vary between 9% and 25% of the total assemblage. In contrast, the line for the marrow-processing assemblage exhibits a significant rise. It has relatively low values for meat-bearing elements but high values for important marrow,.bearing elements. The contrast of t:t¥s assemblage with the similarity of the other. two suggests that significantly different things are
BBI ]. G. Enloe
Element representation
Tab]
Element
Q)
Ol
co
§
30+-----------------------~--------~--4-~------4
()
05
a..
Humerus Radiocubitus Femur Tibia Metatarsal Unidentified metapodial Total metapodials Note: Adjusted rela rnents.
HM
RD
MC
FM
, _ P1
TA
MT Element
-1-
PH
US
MP
MPs
P2-*- 83
Figure 5-1. Relative representation of all elements. happening in the two situations and that we will be able to see the difference in the material residues. We can further test the differences to see if they are statistically d,ifferen t. First, the P2 and B3 assemblages were subjected to a chi-square an~lysis for the relative proportions of element representations, using the raw counts. Low frequency cells for metacarpal and phalanx were eliminated, so only the total metapodials were used to represent metacarpals, metatarsals, and unidentified metapodials. The resultant X2 = 9.99 is not significant at the .05 level, so we cannot reject the null hypothesis that the samples were drawn from the same population. This finding supports the similarity of their makeup, and we can therefore infer that they are likely to be derived from the same formation processes. The same test was performed between Pl and P2, and between Pl and B3. In each of the tests the results are quite different from those of the first test, and they are very similar to one another. The X2 values are very high (257.01 and 244.29), particularly in contrast to that of the first test (9.99), and are highly significant at less than the 0.001level. This strongly confirms that both times the samples being compared:,:'.
•.· ..;. '•..... ··,..,c.·r.::l
.....
-...; ~
......
o
._, en ::;:::
('~)
m~···~
~·?·.·g
g" (") "'1::3 :::::: r:,en~~~
B ....
~«('~)
::l
I
3 .·. . 6
"0 ·• ::0
en n.
(o)
'·'
II\ I I
-:::
:.
1•·.·
...
..
• ,
I
...
..a.
0
!
0 C;
qwnaJo:l
~ QW!IPU!H
~~~i>::l~
:
3 ...
g I
Q) 1·. 5: "0
~ ·.•· . g
efficients were of site can also .nation of each 1ery and trans; the NINEs for ones from four the residential It may be best eflected in the ~letal elements, as indicated by ·iginating from ~lements recovide of the chart ansported and lu I, complete Jar vertebrae is . On the other :licate that the ite. Gaps in the .e case of camp 1 one similarly
Base Camp (Carcass No.) KH (ORYXG2)
KH (ORYXG3) MP/EB/KX-8E
ay =evidence for scavenging, most likely by brown hyenas, N = unscavenged,? =scavenging uncertain. bproduct of previous two columns. 'Entire suppurating hindlimb intentionally abandoned at kill site (see Figure 7-20). Later visit to site revealed evidence of the bones; removed by scavengers.
70
sized carcass, they reflect my inability to relate some specimens conclusively to a specific carcass. In four cases, complete destruction or removal of elements by scavenging carnivores may be the cause of the gap . Three additional complementary distributions of bones are shown in Figure 7-2 (B to D). Each graph represents a single carcass and where its bones were recovered. Once again, the charts illustrate the extreme variability between functionally identical kinds of site assemblages. The plaid pattern in Figure 72B, C, and D represents areas where the kill site and campsite bars overlap, for the same elements were represented in. both the kill site and the campsite assemblages as a result of the fragmentation of these parts during butchery followed by partial transport of the fragments. In fact, I have found a number of cross mends between kill sites and campsites. The archaeological potential for intersite conjoining, although daunting, is thus convincingly demonstrated. Axial elements apparently present the likeliest candidates for conjoining because they are generally fragmented during primary butchery. The Kua conduct limb butchery at the joints, and limb elements normally remain whole until cracked for marrow. Occasional exceptions include incompletely fused epiphyses, broken off in joint-directed chopping. Part Abundance and Food Utility To examine the extent to which variability in element frequencies was explained by nutritional utility, I next plotted the abundance values for each assemblage against Metcalfe and Jones's Whole Bone Standardized Food Utility Index.
1281 L. E. Bartram,]~. It is here that the influence of food utility considerations would be expected to emerge. Kill sites would be expected to appear as reverse utility curves, while transported camp site assemblages would have positive slopes. However, no compelling relationships could be discerned in these scatterplots. Based on the reasoning thatsample size problems at a si~gle carcass site may be misrepresenting central tendencies that would appear in a larger sample, plots were constructed for the average element abundances from both kill sites and their complementary base camp assemblages (Figure 7-3). Once again, no convincing relationships were apparent; ev~n the "average" assemblages of each type seem equally amorphous.
Behavioral Data
80 70
•
/
What Explains Kua Assemblage Composition? If nutritional utility is so effective in explaining the composition of Nunamuit assemblages and apparently so ineffective in the Kua case, what accounts for the difference? It cannot be merely the use of inappropriate utility indices. Despite the reasonable expectation that functionally identical sites would exhibit positive linear relationships and high correlation coefficients, inspection of the scatterplot matrix generated for all pairs of kill/butchery sites revealed virtually no systematic correlations in part frequencies. It is extremely unlikely that this lack of similarity among assemblages of the same type would be suddenly rescued by the refinements conferred by the use of utility data derived from gemsbok. Both the overall lack of correlation between the very same types of camps and the widely variable abundance rankings between assemblages implies that factors other than nutritional utility are conditioning the content of the assemblages.
•
CARP
•
~50
eMAN
~
~
RAI
CRA
60
MCM
•
40
TA THOR
•
30
c
c
20 10
Data of a more behavioral nature (Table 7-3) help. to explain the observed variability in the gemsbok bone assemblages. The carcass completeness statistic described earlier was plotted against the number of butchers, number of carriers, distance to camp, and processing time of each carcass. No significant relationships were evident between any of these variables and carcass completeness. However, because the number of butchers and the time spent processing a carcass can vary independently, their product, the number of person-hours of processing time, is a :tnore meaningful figure. It was here (Figure 7-4) that a highly significant correlation (p = .005) was observed. As one would expect, the longer that processing activities were conducted at a kill site, the more bones were abandoned there. The fact that this strong correlation exists in the absence of compelling nutritional relationships strongly suggests that factors other than nutritional utility determine the composition of Kua bone assemblages.
Discussion
B
90
PHA
•
0 0
0 ,....
0
C\1
Standa1
Figu Food Why do the a: that the camps t an assumption explain the obsulk" transport strategy, but one without the bones to evidence its operation because of biltong production at the primary butchery location. The lack of assemblage similarity between functionally homologous camps stands in sharp contrast to the highly significant correlation between kill site carcass completeness and processing intensity. The probability a given bone would be discarded at the kill site or transported seems to be strongly related to available processing time. With this in mind, the question for archaeologists turns to, What determines available processing time? Unfortunately for the interpretation of individual archaeofaunal assemblages, the answer is highly variable situationally. The variables mentioned earlier such as distance to kill, available carriers, and carcass size all play a role. Yet other factors, including
I
130 L. E. Bartram, Jr.
Kua Gernsbok Kill/Primary Butchery Sites 90
2
r :0.695 r = 0.812
80
5
J
8
HZ /
@ Scavenged e Unscavenged
70
@
60
8
,
KO
K2
50 8
40
©l
8
30
KM ©l
SY MP
TL
20 10
SG
0~----~--~--~----~------~------~----~
0
5
10
15
20
25
30
Processing Time (Person-Hours)
Figure 7-4. Relationship of skeletal completeness index to kill site processing time (in person-hours).
motivation born of hunger and the requirements of social obligations, influence the decisions made at specific sites as well.
Transport Costs In a recent paper on kangaroo bone transport, O'Connell and Marshall (1989) listed three factors that determine the selection of body parts for transport: (1) the weight or nutriti9nal value of edible tissue attached to various body parts; (2) the kill site processing costs; and (3) the relative benefits of consumption at the kill site versus transport and consumption at the residential base. There is support for the suggestion that these same factors operate in other areas as well. Indeed, Binford (1978:90) identified "transport difticulties, time limitations on labor activities, and location of the kill site relative to residential or consumer locations" as factors that may influence which parts of a carcass are chosen for transport, even though the desire to maximize nutritional benefits in transport was the overarching concern among the Nunamuit. The
Kua examples i We have see: importance, bu fact that the Kt: seems that the of the Kalahari the only real f value of each c; whole carcass played an.impc ranking of spe lowest-ranked threshold." For the Kua, carcass someti1 filleting and su the limb bones, elements; they breaking them the meat is dry carrying the ca enticement to 1 gemsbok). Thus, kill sit1 the extra proce limitations of tl occasionally im in order not to now add the be ering transpor1 geographically skeletal part fre
Arc) WhE them) expresse1 should we expE shown that the reflect transpor but reflect inst influence that n The fact tha1 .hunters pursue kinds of surplu be obtained wi transport strate conditioning as.
Skeletal Part Profiles 1131 r
Kua examples illustrate these points and expand the list. We have seen that the first of the factors, nutritional value, is of enormous importance, but with a twist. Decision making at a carcass is dictated by the fact that the Kua want it all, and they do whatever it takes in order· to get it. It seems that the notion of "abandonment" of usable food is foreign to the Kua of the Kalahari. In fact, we were told repeatedly by informants that "meat is the only real food." In each example presented here the entire nutritional Value of each carcass was exhausted; either at the kill site or back at camp, the whole carcass was consumed by camp me.mbers. Thus, nutritional utility played an important role in influencing decision making, but not in terms of a ranking of specific carcass parts for transport. Stated differently, even the lowest-ranked part on the carcass was still above the "abandonment threshold." For the Kua, doing whatever it takes to get the entire nutritional benefit of a carcass sometimes means processing the heavy, wet meat at the kill site by filleting and sun-drying it. Whentransport considerations dictate defleshing the limb bones, Kua butchers see no point in then transporting defleshed limb elements; they do not have to. Instead, they solve the transport problem by breaking them and eating the marrow at the kill site. This can be done while the meat·'is drying. The fatty marrow provides extra energy for the work of carrying/ the carcass home, and the prospect of eating some is an additional enticement to enlist in tracking parties (there is little subcutaneous fat in a gemsbok). Thus, kill site processing costs increase with filleting and meat drying, but the extra processing directly lowers transport costs by accommodating the limitations of the available transport pooL It is clearly worth the effort, for it occasionally inspires small groups of Kua to risk spending the night with lions in order not to abandon meat. Thus, to O'Connell and Marshall's list we must now add the benefits of processing at the kill site as a method of directly lowering transport costs by reducing weight. This option is obviously variable geographically, but it has important implications for the interpretation of skeletal part frequencies from archaeological sites.
Sites
HZ
25
30
to kill site proces-
Mgations, influ-
Archaeological Implications Where else should we expect to see the kinds of patterns (or lack of them) expressed in the Kua data? In which archaeological subsistence contexts should we expect utility analyses to come up short for similar reasons? I have shown that the interassemblage differences in skeletal part frequencies do not reflect transport strategies based upon some ranked scale of carcass part value but reflect instead a variety of situational constraints that overwhelm any influence that nutritional ranking may have otherwise exerted. The fact that utility indices are at their best as explanatory tools where hunters pursue large herds of gregarious prey is probably no accident. The kinds of surpluses available at mass kills, or in contexts where carcasses may be obtained with almost predictable regularity, lend themselves to selective transport strategies, and thus, the appearance of part utility as a key factor conditioning assemblage composition.
0' Connell and 'n of body parts ;sue attached to Le relative bene;umption at the operate in other iifticulties, time ve to residential arts of a ca,rcass 1ize nutritional Nunamuit. The
I
l
1321 L. E. Bartram~Jr.
By contrast, these conditions qearly do. not obtain .in the modern Kalahari environment, where transport costs are reduced by increasing processing costs at the kill site. The Kua example~ discussed here all involve low population densities of both hunters and prey; infrequent, solitary, medium-to-large animals; limited transport capabilities; and semiarid enviro~ents with high insolation. It is precisely in these kinds of contexts that we should expect to find similar strategies pursued prehistorically. Interestingly, it is from broadly similar paleoenvironmental contexts that much of the best ~vidence for hominid subsistence behavior is derived. This issue demands additional ethnoarchaeological consideration and much wider attention froni. archaeologists working in areas where meat drying is a potentially effective way of reducing transport costs. Importantly, there are no technological restrictions of this effective way of reducing transport costs, even with Oldowan assemblages. All that is required for biltong production with lithic-based technology is sharp-edged flakes to dismember carcasses and to cut meat into strips and a cooperative climate. It would be surprising if meat drying were not a widespread practice with enormous antiquity; the benefits of meat drying are simply too great to be ignored. The implications of this simple and dramatically effective cost-saving strategy for current debates on the meaning of skeletal part frequencies are farreaching. Until we are able to demonstrate conclusively in a given archaeological context that defleshing and meat drying were not taking place and that meat and bone were indeed traveling together, we are on increasingly thin ice regarding statements about the meaning of assemblage composition and carcass part transport. What is the archaeological"signature" of kill site defleshing and meat drying? As I have shown, the Kua provide quantitative support for t;he notion that processing time and kill site carcass completeness will exhibit a significant positive relationship. This is not, however, what most archaeologists wish for when they think of an archaeological signature! The answer must almost certainly include other lines of evidence. For example, if large assemblage size and high artifact or feature density suggest a residential site, relatively complete skeletal inventories may be reflecting low processing times at kill sites. This, in turn, may be a function of relatively large prehistoric group sizes to provide transport, high prey densities, or other factors that reduce transport costs. For some Oldowan assemblages judged to be accumulated by hominids, the apparent lack of robust skeletal part patterning on food utility and the presence of relatively complete skeletal inventories for large mammals may either be telling us about a reluctance to process carcasses at the death sites of animals or may be suggesting the lack of a need to do so. The latter case would imply relatively effective carcass acquisition strategies. In any case, further consideration is called for. It is clear from the Kua ex~mple that skeletal part frequencies may be a poor reflection· of the whole nutritional story. Skeletal part frequencies must be considered an important, but likely incomplete, line of evidence in archaeo-
logical inferenc ethnoarchaeolo: of kill site meat well as a searcl clearly warrant meat drying or think in genera: to see utility in' the con texts in applied researc recognizi11g me record.
Cor SevE 1.lr1
tran sets
byr logi· rom ofn 2.0 botr tran inso 3.C not' 4. T Old, 5. Y\ ratic rate: 6.A tal. 1 othe and
Ac1 ·I am .the From Bone~ anonymous re\ Wisconsin Kala National Scienc Fulbright Progr
Skeletal Part Profiles j133 10dern Kalahari 5ing processing lve low popula:ledi urn -to-large nents with high :t to find simila~ broad! y similar or hominid subhnoarchaeologi~ists working in .ucing transport effective way of . that is required ·edged flakes to ·ative climate. It d practice with too great to be ost-saving stratuencies are farven archaeolog; place and that ~asingl y thin ice •osition and car-
g and meat dryt for the notion exhibit a signift archaeologists lle answer must ample, if large residential site, low processing i vel y large preor other factors ~s judged to be l part patterning 1 inventories for mce to process Le lack of a need cass acquisition
may be a'poor tencies must be ~nee in archaeo-
'S
logical inference about subsistenc~ behavior and site function. Additional ethnoarchaeological consideration of the factors that determine the likelihood of kill site meat drying in particular kinds of paleoenvironmental contexts, as well as a search for unambiguous, "diagnostic" evidence of meat drying, is clearly warranted. At this early stage in our consideration of the effects of meat drying on bone assemblage composition, it would be most helpful to think in general terms about the types of contexts in which we should expect to see utility indices influencing transport decisions. If we better understand the contexts in which food utility plays a role, we face smaller risks of misapplied research effort and erroneous conclusions, and a better chance of recognizing meat drying as a viable processing strategy in the archaeological record .
Conclusions Several conclusions may be reached from the above discussion: 1. In tropical and subtropical Africa, the sun is a potent tool in reducing transport costs. Here, and in similar regions, drying meat more than offsets the additional processing costs associated with biltong producing by reducing transport costs. This issue invites additional ethnoarchaeological and archaeological consideration, especially where paleoenvironmental data suggest that meat drying was a potentially effective way of reducing transport costs. 2. Conditions that favor this strategy include low population densities of both hunters and prey, infrequent and single carcass availability, limited transport capabilities, and arid or semiarid environments with high insolation. 3. Current interpretations of variability in skeletal part frequency have not considered the effects of defleshing and drying. 4. The technology required for. efficient meat drying is present in Oldowan assemblages. 5. Where bone assemblages are likely to have been scavenged, incorporation of limb ·shaft fragments into MNE estimates is critical to accurately represent the contents of the assemblage. 6. Archaeologists should not attempt to identify site function from skeletal part frequencies alone but should identify it in. conjunction with other evidence, such as number of hearths, site size, and assemblage size and diversity.
Acknowledgments I am grateful to Jean Hudson for the opportunity to participate in the From Bones to Behavior conference and this volume. I thank Jean and an anonymous reviewer for their comments on this paper. The research of the Wisconsin Kalahari Project was made possible by the financial support of the National Science Foundation and the Institute for International Education's Fulbright Program. I thank the Office of the President, .Republic of Botswana,
1341 L. E. Bartram, ]r.
and C. Mogotse for permission to conduct fieldwork in Central District, and Robert K. Hitchcock f0r introducing us to thefield area. In Gaborone, Alec Campbell, James Denbow, Angier Pe~vey, and Ruby Apsler provided much needed advice and assistance. Special gratitude is extended to Keith aRd Irene Whitelock and Debswana Mining Pty., Ltd. (Orapa), for e;ndless logistical support and extraordinary kindness. I thank Iron Gabatswane and Keikantsemang "Shakes" Tshikhinya for translations between English and Kua and project members Lisanne Bartram, Henry Bunn, Christian Gurney, Marty Jakobs, and ·Ellen Kroll for their conscientio~s and enthusiastic assistance in the field. I especially thank Henry Bunn for countless discussions and for his patient supervision of my graduate research. Finally, I thank the Kua for their tolerance, good humor, and friendship.
References Bartram, L. E., Jr, . 1993 An Ethnoarchaeological Analysis of Kua San (Botswana) Bone Food Refuse. Unpublished Ph.D. dissertation, Department of Anthropology, University of Wisconsin-Madison, in preparation. Bartram, L. E., and H. T. Bunn n.d. An Ethnoarchaeological Perspective on the Klasies Pattern, Limb Bone Fragmentation, and Zooarchaeological Analyses. Ms. in possession of authors. Bartram, L. E., E. M. Kroll, and H. T. Bunn 1991 Variability in Camp Structure and Bone Food Refuse Patterning at Kua San Camps. In The Interpretation of Archaeological Spatial Patterning, edited by E. M. Kroll and T. D. Price, pp. 77-148. Plenum Press, New York. Binford, L. R. 1978 Nunamiut Ethnoarchaeology. Academic Press, New York. 1981 Bones: Ancient Men and Modern Myths. Academic Press, New Yorky 1984 Faunal Remains from Klasies River Mouth. Academic Press, New Yqrk. Binford, L. R., and S. R. Binford · 1966 A Preliminary Analysis of Functional Variability in the Mousterian of Levallois Facies. American Anthropologist 68:238-295. Blumenschine, R. J. 1988 An Experimental Model of the Timing of Hominid and Carnivore Influence on Archaeological Bone Assemblages. Journal of Archaeological Science 15:483502. Blumenschine, R., and T. Caro 1986 Unit Flesh Weights of Some East African Bovids. Journal of African Ecology 24:273-286. Borrero, L.A. 1990 Fuego-Patagonian Bone Assemblages and the Problem of Communal Guanaco Hunting. In Hunters of the Recent Past, edited by L. B. Davis and B. 0. K. Reeves, pp. 373-399. Unwin Hyman, London. Bunn,H.T. . 1982 Meat-Eating and Human Evolution: Studies on the Diet and Subsistence Practices of Plio-Pleistocene Hominids in, East Africa. Unpublished Ph.D. dissertation, Department of Anthropology, University of California, Berkeley. 1986 Patterns of Skeletal Representation and Hominid Subsistence Activities at
Olduvai Gm 15:673-690. Bunn, H. T., and I n.d. Dutara: authors. Bunn, H. T., L. E. 1988 Variabi Scavenging,, 457. Bunn, H. T., J. W. K. Behrensm 1980 FxJj 50: 12:109-136. Bunn, H. T., and I 1986 System;: Tanzania. Cu 1988 Reply tc and Interpret Chase, P. G. 1985 On the· Sites. P. A. C. Elphick, R. 1977 Khoikhoi Emerson, A. M. 1990a Archaeol bison. Unpul ton State Uni 199Gb Carcass ed at the Sixt Institution, V\ Frison, G. C. 1974 The Cas1 New York. Gifford, D.P., and 1977 A Camp Antiquity 42:2 Grayson, D. K. 1988 Danger ' Anthropolog~
York 1989 Bone Tn
Archaeological Hitchcock, R. K. 1978 Kalahari
Agriculturalist of Local Gove Jones, K. T., and D 1988 Bare Bor Archaeological Klein, R. G. 1975 Paleoant:
Skeletal Part Profiles 1135 ~al District, and ::;aborone, Alec provided much Keith and Irene .dless logistical batswane and ~n English and ristian Gurney, :husiastic assis:iiscussions and : thank the Kua
:ma) Bone Food )pology, Univer:ern, Limb Bone ion of authors. rning at Kua San edited by E. M.
York. few York.
N
e Mousterian of nivore Influence I Science 15:483-
i African Ecology of Communal Davis and B. 0.
1
bsistence Practices D. dissertation, ~nee
Activities at
Olduvai Gorge, Tanzania, and Koobi Fora, Kenya. Journal of Human Evolution 15:673-690. Bunn, H. T., and L. E. Bartram n.d. Dutara: An Eland Kill Site in the Kalahari, Botswana. Ms. in possession of authors . Bunn, H. T., L. E. Bartram, and E. M. Kroll 1988 Variability in Bone Assemblage Formation from Hadza Hunting,, Scavenging, and Carcass Processing. Journal of Anthropological Archaeology 7:412457. Bunn, H. T., J. W. K Harris, G. Isaac, Z. Kaufulu, E. Kroll, K. Schick, N. Toth, and A. K. Behrensmeyer 1980 FxJj 50: An Early Pleistocene Site in Northern Kenya. World Archaeology 12:109-136. . Bunn, H. T., and E. M. Kroll 1986 Systematic Butchery by Plio-Pleistocene Hominids at Olduvai Gorge, Tanzania. Current Anthropology 27:431-452. 1988 Reply to "Fact and Fiction about the Zinjanthropus Floor: Data, Arguments, and Interpretation," by Lewis Binford. Current Anthropology 29:135-149. Chase, P. G. 1985 On the Use of Binford's Utility Indices in the Analysis of Archaeological Sites. I?. A. C. T. 11:287-302. Elphick, R/ 1977 Khoikhoi and the Founding of White South Africa. Ravan Press, Johannesburg. Emerson, A. M. 1990a Archaeological Implications of Variability in the Economic Anatomy of Bison bison. Unpublished Ph.D. dissertation, Department of Anthropology, Washington State University, Pullman. 1990b Carcass Product Yields ii1 Bison Bison and Hunter Selection. Paper presented at the Sixth International Council for Archaeozoology Meeting. Smithsonian Institution, Washington, D. C. Frison, G. C. 1974 The Casper Site: A Hell Gap Bison Kill on the High Plains. Academic Press, New York. Gifford, D. P., and D. C. Crader 1977 A Computer Coding System for Archaeological Faunal Remains. American Antiquity 42:225-238. Grayson, D. K. 1988 Danger Cave, Last Supper Cave, and Hanging Rock Shelter: The Faunas. Anthropological Papers 66, pt. 1. American Museum of Natural History, New York. 1989 Bone Transport, Bone Destruction, and Reverse Utility Curves. Journal of Archaeological Science 16:643-652.· Hitchcock, R. K. 1978 Kalahari Cattle Posts: A Regional Study of Hunter-Gathers, Pastoralists, and Agriculturalists in the Western Sandveld Region, Central District, Botswana.· Ministry of Local Government and Lands, Republic of Botswana, Gaborone. Jones, K T., and D. Metcalfe 1988 Bare Bones Archaeology: Bone Marrow Indices and Efficiency. Journal of Archaeological Science 15:415-423. Klein, R. G. 1975 Paleoanthropplogical Implications of the Non-Archaeological Bone Assem-
1361 L. E. Bartram,Jr. blage for Swartklip I, Southwe.stem Cape Province,.South Africa. Quaternary Research 5:275-288. · Landals, A. 1990 The Maple Leaf Site: Implications of the Analysis of Small-scale Bi$on Kills. In Hunters of the Recent Past, edited by L. B. Davis and B. 0. K. Reeves, pp. 122151. Unwin Hyman, London. Lartet, E., and H. Christy (editors) 1865-1875 Reliquae Acquitanicae: Being Contributions to the Archaeology and Paleontology of Perigord and Adjoining Provinces of Southern France. Williams and N orgate, London. Lee, R. B. 1979 The !Kung San: Men, Women, and Work in a· Foraging Society. Cambridge University Press, Cambridge. Lyman,R. L. 1985 Bone Frequencies, Differential Transport, and the MGUI. Journal of Archaeological Science 12:221-236. · Lyruan,R. L. 1984 Bone Density and Differential Survivorship .of Fossil Classes. Journal of Anthropological Archaeology 3:259-299. Marean, C. L., and L. Spencer 1991 Impact of Carnivore Ravaging on Zooarchaeological Measures of Element Abundance. American Antiquity 56:645-658. Metcalfe, D., and K. T. Jones 1988 A Reconsideration of Animal Body-Part Utility Indices. American Antiquity 53:486-504.
O'Connell, J. F., and B. Marshall 1989 Analysis of Kangaroo Body Part Transport among the Alyawara of Central Australia. Journal of Archaeological Science 16:393-405. Perkins, D., and P. Daly 1968 A Hunter's Village in Neolithic Turkey. Scientific American 219:97-106. Silberbauer, G. B. 1 1981 Hunter and Habitat in the Central Kalahari Desert. Cambridge University Press, Cambridge. Smithers, R. H. N. 1971 The Mammals of Botswana. Mus. Mem. Natl. Mus. Manum. Rhodesia 4:1-340. 1983 The Mammals of the Southern African Subregion. University of Pretoria, Pretoria. Speth,J. D. 1983 Bison Kills and Bone Counts: Decision Making by Ancient Hunters. University of Chicago Press, Chicago. Speth, J. D., and K. A. Speilmann 1983 Energy Source, Protein Metabolism, and Hunter-Gatherer Subsistence Strategies. Journal of Anthropological Archaeology 2:1-31. Tanaka, J. 1980 The San: Hunter-Gatherers of the Kalahari, A Study in Ecological Anthropology. University of Tokyo Press, Tokyo. Thomas, D. H., and D. Mayer 1983 Behavioral Faunal Analysis of Selected Horizons. In The Archaeology of Monitor Valley: 2. Gatecliff Shelter, edited by D. H. Thomas, pp. 353-391. Anthropological Papers 59. American Museum of Natural History, New York.
Tobias, P. V. (edit 1978 The Bu: Rousseau, C: Wheat, J. B. 1972 The Ols Society for A White, T. E. 1954 Observ; Nos. 3, 4, 5, c
Skeletal Part Profiles j137 frica. Quaternary
scale Bison Kills. Reeves, pp. 122-
Archaeology and ce. Williams and :iety. Cambridge Jurnal of Archaeolasses. Journal of
sures of Element
merican Antiquity ·awara of Central
219:97-106.
ridge University
Rhodesia 4:1-340. :sity of Pretoria, mters. University erer Subsistence
ical Anthropology. he Archaeology of 353-391. AnthrowYork.
Tobias, P. V. (editor) 1978 The Bushmen: San Hunters and Herders of Southern Africa. Human and Rousseau, Cape Town. Wheat, J. B. 1972 The Olsen Chubbuck Site: A Paleo Indian Bison Kill. SAA Memoirs No. 26. Society for American Archaeology, Washington, D.C. White, T. E. 1954 Observations on the Butchering Technique of Some Aboriginal Peoples, Nos. 3, 4, 5, and 6. American Antiquity 19:254-264.
8.
The Role of Body Part Utility in Small-scale Hunti~g under Two Strategies of Carcass Recovery Alice M. Emerson Abstract: Data provided by O'Connell and colleagues (1990) on the · transport of body parts to base camps. are examined for Hadza hunting of large game animals under conditions of small-scale kills. Axial and appendicular unit transport is considered with respect to different measures of bison body part utility and discussed in light of the expectations of th~ "schlepp effect" (Perkins and Daly 1968; White 1952). The greater appendicular utility suggested by that model is found·to apply only when skeletal fat or marrow is the focus of utility assessments. The variability in carcass unit transport to base camps by Hadza hunters is examined with respect to two strategies of carcass transport-a Maximum Carcass Recovery strategy, which over time yields assemblages dominated by axial elements, and a Limited Carcass Recovery strategy, which yields assemblages dominated by appendicular and some axial units. Transport and processing costs are evaluated with respect to body part utility and suggested to be important factors influencing elimination of elements from the transport assemblage to meet in-transit food needs.
Introduction Recent observations by Bunn and colleagues (1988) and O'Connell and associates (1988, 1990) of modern Hadza hunter-gatherers suggest that small-scale hunting of large game (the taking of one or two individual animals) permits greater flexibility in carcass transport practices than is possible with large-scale kills. This flexibility is reflected in what first appears to be inconsistent relationships between the transport of body parts and models of general body part utility, such as the Modified General Utility Index (MGUI) as developed by Binford (1978) and simplified by Metcalfe and Jones (1988), yielding various standardized or unstandardized versions of the Food Utility
Index (FUJ scalogram Hadza hu "schlepp e rather thar the hunte: between SJ of axial un cited by OJ interpretat abundance also leads greatly infl transport a Analyse~
although < appendicu the "schleJ caloric yiel and marrO' ity does pt they are m· sit food ne1 based upo1 and the otl Expectatio: ported assE dicular pal · small-scale The rem Versus Ap] the expectc body part section, Ha ed by O'C Strategies ences in thE high utili!) Modificati< nated from
From Bones to Behavior: Ethnoarchaeological and Experimental Contributions to the Interpretation of Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional Paper No. 21. © 1993 by the Board of Trustees, Southern illinois University. All rights reserved. ISBN 0-88104-076-2.
138
earlier by 'Y
Strategies of Carcass Recovery 1139
ltyin ·Two rery
es (1990) on the )r Hadza hunting e kills. Axial and to different meaf the expectations l952). The greater nd to apply only :sments. The varil hunters is examart-a Maximum :semblages domi':'1 strategy, which some axial units. pect to body part ng elimination of sit food needs.
) and O'Connell ers suggest that individual anithan is possible st appears to be :sand models of y Index (MGUI) . nd Jones (1988), the Food Utility ' the Interpretqtion of gations, Occasional iversity. All rights
Index (FUI) for complete bone or proximal/ distal end models. Specifically, scalogram analyses (O'Connell et al. 1990) of carcass element transport by Hadza hunters indicate that, contrary to expectations grounded in the "schlepp effect" of Perkins and Daly (1968) and earlier in White (1952), axial rather than appendicular units are most frequently returned to base camps by the hunters. Furthermore, their analyses indicate that both within and between species differences in the transport of body parts occur. The contrast of axial unit transport with the expectations of the "schlepp effect'' has been cited by O'Connell and colleagues (1990) as reason to question archaeological interpretations of hunting versus scavenging that are based upon the relative abundance of limb versus axial elements. The variability in transport practices also leads one to question whether differences in body part utility, which greatly influences transport decisions at large-scale kills, is pertinent to carcass transport at small-scale kills. Analyses presented here of carcass composition data for bison (Bison bison), although clearly not an African species, suggest that the expectation that appendicular elements are of greater value than axial units, as suggested by the "schlepp effect," is unfounded unless the evaluation is based upon the caloric yield of skeletal fat (determined from the fat content of bone grease and marrow) or marrow alone. The analyses also suggest that body part utility does~play an important role in transport decisions at small-scale kills, but they are mediated by evaluations of transport costs and immediate or in-transit food needs. Two possible strategies of carcass recovery are recognized: one based upon an initial expectation of complete carcass recovery and transport and the other based upon limited recovery and transport of high utility parts. Expectations of how carcass units are subsequently eliminated from transported assemblages through use or abandonment clarify why axial or appendicular parts might dominateassemblages transported to base camps from small-scale kills. The remainder of this paper is divided into five sections. The first, Axial Versus Appendicular Utility and the "Schlepp Effect," includes a discussion of the expectations of this model and considerations of how different models of body part utility rank the major subdivisions of the bison carcass. The next section, Hadza Transport Practices, is a brief review of transport data presented by O'Connell and colleagues (1990). It is followed by Carcass Transport Strategies and Assemblage Compositions, which focuses upon how differences in the initial decision of whether to transport the complete carcass or the high utility parts result in assemblages that are distinctive in composition. Modifications to Transport Assemblages examines how body parts are eliminated from the initially selected transport assemblage. Conclusions follow.
Axial Versus Appendicular Utility and the ~"'Schlepp Effect" The "schlepp effect," as presented by Perkins and Daly (1968) and earlier by White (1952), predicts that long distances to camp or large prey size
I
140 A.M. Emerson may prevent complete carcass recovery. Under .those circumstances the least useful elements may be discarded and a partial carcass assemblage returned to base camp. Traditionally, it .has been assumed that limb units. are more useful than axial units, and differences in the relative abundance of these skeletal elements at archaeological sites have been us~d to suggest different modes of acquisition (scavenging versus hunting). White's (1952) interpretation of bison use reflects this expectation. He suggests that axial units of hunted bison were often stripped of meat and the associated skeletal elements discarded at· the kill site, while the appendicular skeleton was transported back to camp. He attributes this practice to differential transport costs and marrow yields. White's work suggests that assessments of value may be made on the utility of component products of a carcass unit rather than on general utility. Analyses presented here demonstrate that the relative utility of axial and appendicular units varies depending upon which carcass products are included in the assessment. Four types of bison body part utility models (Emerson 1990b) are evaluated with respect to their relative axial anq appendicular yields: (1) two general utility models-one (Emerson 1990b) calculated from the caloric yields of skeletal fat, muscle protein, intramuscular and other dissectible fat (the Total Products model [S]MAVGTP) and the other calculated from weights of body parts minus their associated dry bone weight (the Food Utility Index [S]AVGFUI) following Metcalfe and Jones (1988); (2) a Total Fat model ([S]MAVGTF) calculated from caloric yields of skeletal, intramuscular, and dissectible fat (excluding the stomach and intestinal depots, which were not evaluated); (3) two Skeletal Fat models ([S]MAVGSKF and [S]AVGSKF), each calculated from the caloric yields of fat present in bone grease and marrow; and (4) two Marrow Fat models ([S]MAVGMAR and [S]AVGMAR) based upon the caloric yield of marrow alone.l Carcass composition data (Emerson 1990b) from four range-fed bison (yearHhg and 4-yearold males and 7-year-old and 16-1/2-year-old females) provide the data for the models. It is expected that generation of similar data for African species would reflect similar relationships between major carcass divisions. Figure 8-1 compares two models of general utility for bison in which proximal and distal limb values are determined. Before discussing relative axialappendicular division utility differences, one must first account for the differences evident in the models that are based upon data from the same animals. The Total Products nlodel is based upon caloric yields of carcass products that are standardized so that each body part value is expressed as a percentage of the highest yielding unit. The standardized values of the limb units are then modified to account for the effects of riders. In essence, the rider averaging routine (developed by Binford 1978) raises the value of most limb units to account for the fact that lower-valued parts are sometimes transported because of their location next to, or between, higher-valued units. The bison Food Utility Index model (also modified to account for riders) is based upon Metcalfe and Jones's (1988) simplified model construction. Although body part values are based upon weight rather than caloric yield, in bison this has little effect on the comparative rankings of the elements
..J
.o;
x.o;
CERVICA THO
ll Cl
PR PRO XI DIS
a:
:5:::>
PRO XII
0
ANT
u
~.o;
DIS
PRO XI Dl~
POST
(Emerso of axial differen1 Like I units rel rider a-v nouncec limb un lated. Tl whole u complet units) ar the valu for each 8-1a, wl Product units. Tl femur,tr Ifwhc the majc (S)AVG: have hi~
Strategies of Carcass Recovery 1141 tances the least tblage returned units are more ldance of these tggest different 952) interpreta -. t axial units of celetal elements ras transported :;port costs and te may be made :han on general
a. TOTAL PRODUCTS MODEL ((SIMAVGTPI 0 SKULL ATLAS ·AXIS
_,
x 200 kg) and the parts left are more likely to be axial than appendicular; and (3) where appendicular parts are left in the field, the explanation involves a hunter's or carrier's monopolization of marrow for personal consumption, not his or her attempt to reduce transport costs. Bunn does not present his data in sufficient detail for one to determine if the patterns he describes are really different from those O'Connell and colleagues identified or whether he controlled for the variables they found to be important determinants of Badza transport decisions. He and his associates made similar claims before (Bunn et al. 1988), but subsequent analysis of the combined. data sets (O'Connell et al. 1990) showed them to be unsupported. The key issue here is analytic approach. As Bunn's discussion makes clear,
1721 J. F. O'Connel~ his immediate objective is to characterize the "typical Hadza pattern" of body part transport. O'Connell and his colleagues' analysis indtcates that transport patterns vary from time to time, place to place, sample to sample, depending on such factors as the number and type(s) of prey taken, the number· of carriers available, and the distance(s) from kill(s) to base c~p(s). In short, there is no "typical" pattern. This is an asset, not an obstacle, to analysis. It enables one to explore alternate explanations for observed variation in carcass processing and transport within a single ethnographic case, in this instance, among the Hadza. Attempting to identify the normal_ or modal Hadza pattern in one or another sample of cases effectively precludes this. Bartram's paper on gemsbok butchery and transport by the Kua makes three main points: (1), varying, sometimes high, numbers of bones are stripped of meat and other edible tissue and left at kill sites, probably as a function of weight reduction for transport (cf. Bunn, this volume); (2) as a result, simple utility indices based on the amount of meat, marrow, and grease originally associated with various elements do not accurately predict their probability of transport; and (3) element transport patterns do not-reflect a body part ranking of any kind. Though the first two observations are not unprecedented, the addition of another case to the small body of data on carcass processing among traditional hunters is very useful. On the other hand, Bartram's third point is surprising, especially in view of the evidence from other settings that element rank on some index (not necessarily simple utility) determines or at least strongly conditions transport (e.g., Binford 1978; Emerson 1990, this volume; Metcalfe and Jones 1988; O'Connell et aL 1990). If Bartram were right, the prospects for interpreting archaeologically observed differences in body part representation in terms of transport by hunters might well be quite limited. The situation he reports is complicated in that some assemblages in his sample were ravaged by carnivores prior to collection. Until this effect is controlled by analyzing whatwas actually transported rather than rEj'covered ethnoarchaeologically, one cannot be certain whether the absence oH~~vidence for a transport scale is real or a function of secondary attrition. Emerson's report on bison body part utility is an important contribution. Her observation that parts may be scaled quite differently depending on the currency used to measure nutritional benefit is particularly instructive. It suggests that very different patterns in body part assemblage composition might result from differential transport, depending on the hunter's goals and the condition of the carcass on encounter. It would be interesting to know how processing costs might affect those decisions, but Emerson's experiments were undertaken before it became apparent such costs might be critical. Her attempt to apply her bison data to the empirical problem of variability in carcass transport by the Badza, specifically in cases involving giraffe, is intriguing; but one may be skeptical about its results, partly because parts of animals as morphologically different as bison and giraffe are unlikely to be ranked in sufficiently similar ways on any scale and partly because of the absence of data on processing costs. . Jones describes small marnmal transport and its implications for bone assemblage composition at a base camp created by Ache foragers. Though
interesting ( among other makes it diff has observec assemblages role in huma predict from mals should resulting wa how they n (including tl possible. Fe· among modE Hudson.1991 achievement tant in that t driven work 1991; Hill et'
s T
I
research on pioneering v refuse in an Gonzalez 191 the investig1 archaeologic. to transport i a well-founc aspects of si analogy, an E Enloe and do not addn because of j determinanb age morpho criteria that c (Binford 198 set and pro' archaeologic which forag: those strateg associated w representatic examples pn
Subsistence and Settlement 1173 !rn" of body .at transport , depending number of :s). In short, , analysis. It min carcass 1is instance, tdza pattern Kua makes are stripped L function of !Sult, simple ;e originally robability of y part rankedented, the . processing tram's third settings that rmines or at ~90, this velwere right, 1ces in body ell be quite emblages in this effect is n recovered of evidence ontribution. tding on the ctive. It sugsition might )als and the ) know how iments were ~ritical. Her ariability in ;iraffe, is inuse parts of 1likely to be cause of the ns for bone ers. Though
interesting (not least because of the contrast with well-documented cases among other groups involving larger prey), the anecdotal quality of the work makes it difficult to apply beyond the Ache context. As Jones (1984) himself has observed, small mammals are an important part of many archaeological assemblages from the late Pliocene onward, yet uncertainty persists over their role in human diets (e.g., Grayson 1991). It would be useful if one were able to predict from general principles the circumstances under which small mammals should be taken, how they should be processed for consumption and the resulting waste discarded, what archaeological patterns would result, and how they might be distinguished from the effects of other processes (including the actions of nonhuman bone accumulators); but this is not yet possible. Few comprehensive descr"iptions of small mammal exploitation among modern hunters have been undertaken with these goals in mind (cf. Hudson 1990, this volume; Yellen 1991a, 1991b), and none yet enable their achievement. A detailed analysis of the Ache data could be especially important in that they can be placed in the context of a large body of theoretically driven work on foraging, prey selection and food sharing (e.g., Hawkes 1990, 1991; Hill et al. 1987; Kaplan and Hill1985) .
Site Structure The past 15. years have witnessed many reports of actualistic research on hunter-gatherer site structure, beginning with Yellen's (1977) pioneering work on the !Kung. Few, however, treat the distribution of bone refuse in any detail (cf. Bartram et al. 1991; Binford 1983, 1987; GiffordGonzalez 1989). Even they involve little analysis of the kind needed to guide the investigation of patterns in bone distributi~n that might be observed archaeologically, other than in general terms. Unlike the situation with respect to transport and processing, archaeologists currently lack even the promise of a well-founded theoretical basis for predicting form and variation in most aspects of site structure. Their only interpretive tool is direct ethnographic · analogy, an essentially nonexplanatory basis for argument. Enloe and Jones make important contributions to the growing literature but do not address the central problem. Again, Jones's single case is interesting because of its. simplicity but difficult to employ analytically because its determinants are unclear. Enloe analyzes body part representation and dam-. age morphology in three Nunamiut middens with the goal of identifying criteria that distinguish storage-dependent collectors from nonstoring foragers (Binford 1980). He deserves credit for tackling an initially unpromising data set and producing some intriguing results. In order to make use of them archaeologically, one needs to know more about the circumstances under which foraging or collecting might be expected, how sites associated with those strategies are likely to be organized internally, and why faunal remains associated with various areas within sites display the patterns in body part representation and damage morphology they do, at least in the Nunamiut examples presented.
174 I J. F. O'Connell
Archaeological Applications Ultimately, the payoff for. actualistic research comes with the application of its results to substantive problems in prehistory. Two papers in the sections under review tackle this difficult challenge. . The first, by Pike-Tay and Knecht, deals with aspects of subsistence and technology in the Upper Paleolithic of southwestern France, the goal being to determine· whether they represent the activities of "opportunistic" foragers or "specialized" collectors. The exercise is ambitious ~d potentially important but open to criticism on two grounds. First, as Pike-Tay and Knecht see it, foragers choose prey opportunistically as a function of their local abundance, while collectors specialize in taking particular species from among an available array, essentially as a function of cost-benefit considerations. The distinction differs from that originally proposed by Binford (1980)~ who defined reliance upon stored foods as the key criterion. It is also inconsistent with standard ecological models of prey selection, which emphasize the importance of cost/benefit considerations for all predators and downplay the role of abundance alone as a predictor of selection under most circumstances (e.g., Bettinger 1991; Stephens and Krebs 1986). A second problem involves the archaeological correlates of foraging and collecting. Among collectors, Pike-Tay and Knecht expect "reliable" technology, defined by "the presence of such weapons as barbed harpoons or points, or a multiplicity of points"; for foragers; "maintainable" technology and the absence of evidence for "large [scale] game drives or intensive high-yield hunting episodes." The validity of those and other criteria mentioned are open to question. For example, both Hadza and Alyawarra foragers use barbed projectile weapons; Hadza hunters routinely carry as many as three different types of arrows. Similarly, Great Basin ethnography and archaeology are rich in evidence of large-scale game drives, all by people acting as foragers, at least in the seasons when the drives took place (e.g., Egan 1917:238-239; I Steward 1938). Though Pike-Tay and Knecht do not develop the argument, thete are at least two reasons to be interested in the forager-collector contrast in the European Paleolithic, one good, the other not. The latter, mentioned at the conference and elsewhere, entails the view that collecting might be a critical indicator of what Binford (1989) calls "planning depth," the ability to conceptualize and meet subsistence needs over long periods of time through food storage. Binford himself has appealed to this notion in distinguishing Middle from Upper Paleolithic subsistence patterns and, by extension, archaic sapiens' cognitive abilities from those of modern humans. A moment's reflection undercuts the argument. The fact that modern human foragers do not store food does not imply they lack the ability to think far ahead: witness the time depth in marriage arrangements among traditional Australian Aborigines, most of whom were foragers. Conversely, the habit of storing food, found in many nonhuman organisms (Vander Wal11990), does not necessarily mean they think like we do. The other reason turns on the observation that Europe was unoccupied during periods of full glacial climate prior to the most recent cycle (Gamble 1986). Following Binford's argument, the difference might well involve reliance on storage. Whether it also involves some new conceptual ability is
another que circumstano define the lil and Knecht': implication. determinant ologists have In the last body part're tals at sites control data and nonhurr ing ambush prey. Thee important. Impressiv first involvi ethnoarchae an importan for any one more than c habitat-sped meaning to 1 transport ta' nonhuman I modern hm Nunamiut i sample relat though the I ical contexts data here an one might a~ poorly unde (at least as F patterns, eitl The secon' Despite the exercise. Var tral tendenc: paints for in grounds of s likely to be reasonable a tal foraging ance, both v Upper Paleo this issue m central tend~ predator-sea
Subsistence and Settlement 1175
es with the vo papers in ;istence and ~oal being to foragers or y important necht see it, abundance, ng an availThe distincrho defined sistent with ! the imporly the role of :tances (e.g., I
:>raging and >le" technolrrs or points, ogy and the ~ high-yield !n tioned are ~rs use barbthree differtaeology are ; foragers, at 117:238-239;
there are at .trast in the loned at the be a critical y to conceplrough food hing Middle trchaic sapit' s reflection do not store ,ess the time Aborigines, :>d, found in :saril y mean unoccupied de (Gamble . involve retal ability is
another question. In any case, one should be able to predict the ecological circumstances under which storage becomes the optimal subsistence strategy, define the likely archaeological indicators, and run the test. In short, Pike-Tay and Knecht's question has an interesting and extremely important substantive implication. Addressing it successfully will require more knowledge of the determinants of storage and their archa~ological implications than archaeologists have at present. In the last piece under consideration, Stiner reports two distinct patterns in body part representation for red deer and aurochs accumulated by Neandertals at sites in west-central Italy. She interprets them by comparison with control data from bone assemblages accumulated by other predators (human and nonhuman) as the product of two distinct foraging strategies, one involving ambush hunting, the other scavenging of selected parts from winter-killed prey. The exercise is original; the results intuitively appealing and very important. Impressive as it is, Stiner's analysis can be challenged on two points, the first involving internal consistency. Though one might have expected the ethnoarchaeologically observed bone accumulations of modern hunters to be an important reference dimension, Stiner rejects them on grounds that data for any one case are too few to be informative. Lumping information from more t~an one J/would obscure a diverse array of ecological contexts and habitat..specific foraging strategies across the world without providing ... any meaning to the madness it would create." Hunter-gatherer prey selection and transport tactics are also said to be less well understood than those of the nonhuman predators she employs for comparison. While many data sets on modern hunter-gatherers are indeed quite small, Binford's (1978) on the Nunamiut includes information on more than 275 prey animals, a large sample relative to most in the nonhuman suite used in the analysis. Further, though the point about the problems of lumping data from different ecological contexts is well taken, this is exactly what Stiner does with her nonhuman data here and elsewhere (e.g., Stiner 1991: Figures 8.4, 8.7, 8.8). Finally, while one might agree that variati01,1in human bone accumulation practices remains poorly understood, it seems unlikely to make much difference to the analysis (at least as presented here) since explanations of variation in the nonhuman patterns, either between or within taxa, play little role in the argument. The second objection is essentially the one raised above about Bunn's paper. Despite the occasional reference to explanation, this is a pattern recognition exercise. Variance observed in the modem world is reduced to a series of central tendencies· "typical" of each bone accumulator; these serve as reference points for interpreting the Neandertal samples. The approach is justified on grounds of scale: the fundamental niche is the target of interest, and all that is likely to be reflected archaeologically, at least in this time range. This is a reasonable argument as far as it goes. But if Stiner's inferences about Neandertal foraging are correct, the problem then becomes accounting for the variance, both within the Middle Paleolithic sample, and between that and the Upper Paleolithic. Again we confront the problem of explanation. Addressing this issue must attend to the variance subsumed in the characterization of central tendencies in prey selection and bone accumulation among modern predator-scavengers, human and nonhuman.
176
I ]. F. O'Connell Summary
Recent research on the composition of faunal assemblages represents some of the best work now being done in archaeological method and theory. The papers reviewed here reflect some of the strength _of this research, especially in their reliance on the living world as a source of information about the past. The key problem entails their use of that information. Analysts here and elsewhere have generally been concerned with identifying "typical" patterns in faunal assemblage composition, as observed ethnographically or /l.ethologically," and using them as the basis for interpreting variation in assemblage composition in the past. Stiner's work shows some of the gains to be made from this approach; the ongoing argument about bone transport among the Hadza suggests its potential pitfalls. However successful it may be, it has a limit: some patterns represented in ~he past are not matched in the present, and even the "matches" (as in the Italian Mousterian case) create problems of explanation. Either way, one confronts the problem of accounting for archaeologically observed variability. As most contributors to this volume would agree, the living world is the place to develop the necessary frameworks for explanation, primarily because potentially pertinent variables can be observed there directly. Questions of prey selection, processing and transport, and their implications for variation in assemblage composition s,eem especially tractable, partly because they lend themselvE~s to quantitative cost/benefit analysis and partly because they can be tackled yvith the use of theoretical models developed in other fields, notably behavioral ecology. Site structure presents a more challenging problem, primarily because of the current absence of any such models. It is important that analysts pay more attention to those issues and that they undertake a more theoretically driven approach to ethnography and ethnoarchaeology. Without that, archaeologists will be in no position to address the problems in prehistory that attracted their interest in the first place. /
Acknowledgments Jennifer Graves, Donald Grayson, Kristen Hawkes, and Duncan Metcalfe provided useful comments on earlier versions of this paper.
References Bartram, L., H. Bunn, and E. Kroll 1991 Variability in Camp Structure and Bone Food Refuse Patterning at Kua San Hunter-Gatherer Camps. In The Interpretation of Archaeological Spatial Patterning, edited by E. Kroll and T. D. Price, pp. 77-148. Plenum Press, New York Bettinger, R. 1991 Hunter-Gatherers: Archaeological and Evolutionary Theory. Plenum Press, New York Binford, L. 1978 Nunamiut Ethnoarchaeology. Academic Press, New York. 1980 Willow Smoke and Dogs' Tails: Hunter-Gatherer Settlement Systems and Archaeological Site Formation. American Antiquity 45:4-20.
In F Res and Theo; Kent, pp 1989 Isol preach.: Pleistoce1 Cambrid
1983 1987
Bunn, H., L. I 1988 Var
· ing and 1 Egan, H. 1917 Pi01 Estate, R Emerson, A. 1990 Arc bison. Pl versity,: Gamble, C. 1986 The Cambric Gifford-Goru 1989 Mo cation, e• the First 1991 Bm in Zooa1 Grayson, D. 1988
Dal
pologicc= 1991 Th1 Reamers. edited 1 Museun Hawkes,K. 1990
Wl
Tribal ar. Boulder 1991 She and Sod Hawkes, K.,: 1991
Hu
Fora gin of the Ro Hill, K., H. K 1987 Fm
cations. Hudson,J. 1990 Ad Aka Py~ Califorr
Subsistence and Settlement 1177 1983 1987
ages reprelethod and .is research, nformation n. Analysts .g "typical" tphically or ·ariation in the gains to e transport 1 it may be, :hed in the ~ase) create accounting rorld is the ·ily because uestions of >r variation :e they lend se they can ther fields, tging probIt is imporlndertake a ·chaeology. >roblems in
In Pursuit of the Past. Thames and Hudson, London.
Researching Ambiguity: Frames of Reference and Site Structure. In Method and Theory for Activity Area Research: An Ethnoarchaeological Approach, edited by S.
Kent, pp. 449-512. Columbia University Press, New York. 1989 Isolating the Transition to Cultural Adaptations: An Organizational Approach. In The Emergence of Modern Humans: Biocultural Adaptations in the Later Pleistocene, edited by E. Trinkaus, pp. 18-41. Cambridge University Press, Cambridge. Bunn, H., L. Bartram, and E. Kroll 1988 Variability in Bone Assemblage Formation from Badza Hunting, Scavenging and Carcass Processing. Journal of Anthropological Archaeology 7:412-457. Egan, H. 1917 Pioneering the West, 1846-1878. Major Howard Egan's Diary. Howard R. Egan Estate, Richmond, Utah. Emerson, A. 1990 Archaeological Implications of Variability in the Economic Anatomy of Bison bison. Ph.D. dissertation, Department of Anthropology, Washington State University, Pullman. University Microfilms, Ann Arbor. Gamble, C. 1986 The Palaeolithic Settlement of Europe. Cambridge University Press, Cambridge. Gifford-Gonzalez, D. 1989 Modem Analogues: Developing an Interpretive Framework. In Bone Modification, edited by R. Bonnichsen and M. Sorg, pp. 43-52. Center for the Study of the First Americans, Orono, Maine. 1991 Bones Are Not Enough: Analogues, Knowledge, and Interpretive Strategies in Zooarchaeology. Journal of Anthropological Archaeology 10:215-254. · Grayson, D. 1988 Danger Cave, Last Supper Cave, and Hanging Rock Shelter: The Faunas. Anthropological Papers of the American Museum of Natural History 66:1. 1991 The Small Mammals of Gatecliff Shelter: Did People Make a Difference? In
Reamers, Bobwhites, and Blue-points: T.ributes to the Career of Paul W. Parmalee, 1d Duncan
:; at Kua San
rl Patterning, >rk. t
Press, New
)ystems and
edited by R. Purdue, W. Klippel, and B. W. Styles, pp. 99-109. Illinois State Museum Science Papers No. 23. Springfield. Hawkes,K. 1990 Why Do Men Hunt? Benefits for Risky Choices. In Risk and Uncertainty in Tribal and Peasant Economies, edited by E. Cashdan, pp. 145-166. Westview Press, Boulder, Colo. 1991 Showing Off: Tests of an Hypothesis about Men's Foraging Goals. Ethology and Sociobiology 12:29-54. Hawkes, K., J. O'Connell, and N. Blurton Jones 1991 Hunting Income Patterns among the Badza: Big Game, Common Goods, Foraging Goals and the Evolution of the Human Diet. Philosophical Transactions of the Royal Society, London B 334:243-251 ~ Hill, K., H. Kaplan, K. Hawkes, and A. M. Hurtado 1987 Foraging Decisions among Ache Hunter-Gatherers: New Data and Implications for Qptimal Foraging Models. Ethology and Sociobiology 8:1-36. Hudson,J. 1990 Advancing Methods in Zooarchaeology, An Ethnoarchaeological Study among the Aka Pygmies. Ph.D. dissertation, Department of Anthropology, University of California, Santa. Barbara. University Microfilms, Ann Arbor.
T I
178
I f. F. O'Connell
j(mes,K. 1984 Hunting arid Scavenging by Early Hominids: A Study in Archaeological Method and Theory. Ph.D. dissertation, Department of Anthropology, University of Utah, Salt Lake City. University Microfilms, Ann Arbor. }1mes, K., and D. Metcalfe · 1988 Bare Bones Archaeology: Bone Marrow Indices and Effitiency. Journal of Archaeological Science 15:415-423. I differentiate ·om those ere. 1989; Binford , 1989; Oliver
1e Interpretation of tions, Occasional All rights
~rsity.
1986, 1989, 1991; Potts 1988; Potts and Shipman 1981; Shipman and Rose 1983). Few data are available, however, on how cultural behaviors may be reflected in bone accumulation and damage patterns, particularly fractured bone. To a large degree this is due to the episodic treatment of human behavior (Binford 1987), where otherwise related activities are artificially partitioned, resulting in a patchwork of data but no appreciation of how several behaviors may be integrated into a decision-making strategy. With a focus on differentiating between hominid- and nonhominid-induced patterns, most actualistic work has examined only segments of carcass processing behavior. That is, ethnoarchaeological and experimental studies have restricted their focus to select events including (1) those occurring in the early stages of carcass processing, such as transport (e.g., Bunn et al. 1988; Bunn and Bartram 1990; O'Connell et al. 1988a, 1988b, 1990) and cut mark patterns (e.g., Bunn 1981, 1983; Potts and Shipman 1981; Shipman and Rose 1983) and (2) marrow extraction from fresh bone (e.g., Binford 1981;/ Bonnichsen 1979; Bunn 1989; Johnson 1985). Carcass processing as a whole, a set of related behaviors from field butchery and transport to cooking and consumption, has not been given explicit treatment (but see Binford 1978). The episodic treatment of carcass processing has created a number of difficulties. Ffrst, it has not provided archaeologists with the referential tools necessary to develop expectations about, much less to identify, the various behaviors behind bone frequency and damage patterns in most zooarchaeological assemblages where post-transport processing associated with cooking and consumption rri.ay have occurred. We do not know how different elements are cooked, which bones are broken prior to cooking, whether roasting or boiling of bone creates· visible damages, how cooking- and consumptionrelated bone breakage varies by element, taxon, and cooking technique. Should we, for example, expect bones of smaller ungulates to enter the postprocessing taphonomic system more fragmented than those of larger ungulates because they are less durable? A related problem this treatment engenders is difficulty in understanding how one set of processing options (e.g., dry meat for storage or trade vs. immediate consumption) inay influence earlier or later transport and processing decisions. How do variables like cooking technology influence earlier field butchery and transport decisions? Yellen (1977), Bunn and colleagues (1988), and particularly Binford (1978) have noted that final processing goals may structure earlier butchery and transport decisions. Do variables like limb bone size and cooking options force bone breakage decisions that create greater fragmentation for large or small ungulate limb bones? Finally, because the organization-of carcass processing behavior has not been defined, archaeologists are poorly prepared to develop evolutionary carcass processing rp.odels that explain how cultural and technological innovations might influence processing behaviors and how those changes in processing behavior maybe archaeologically -defined. vVe need to develop models, for example, that will define how the invention and use of fire for cooking may have altered previous carcass processing patterns established in the Plio-Pleistocene when bone breakage and gnawing were the only nutrient extraction technologies.
I
202 f. S. Oliver
Similarly, we need to understand how later innovations in bone boiling might have altered previous ·carcass processing strategies where the only cooking technique was roasting. . Experimental approaches may provide some solutions to the first problem (e.g., definition of fire-related damages and consumption-rel~ted bone breakage). Clearly, however, the most fruitful way to assess carcass processing as a nutrient extraction strategy is with ethnoarchaeological observations. Bone modification behaviors have been discussed with varying degrees of detail for a few groups including the Australian Aborigines (~inford 1983; Gould 1967; O'Connell and Marshall 1989), San (Bartram 1993; Bunn 1989; Yellen 1977), Nunamuit (Binford 1978, 1981), Cree (Bonnichsen 1973; Zierhut 1967), and Dassanetch (Gifford-Gonzalez 1989). With few exceptions, however, those studies have not approached carcass processing from an organizational perspective, and nm1e provide frequencies of variability in bone breakage behavior, tools employed, or the resultant damages so that archaeological expectations can be developed. Here I present observations on butchery-, cooking-, and consumptio~ related bone breakage behavior of the Hadza, a small group of huntergatherers living near Lake Eyasi, Tanzania. There are three objectives to presenting these observations. First, the bone breakage data for each behavioral context are presented by the animal size class, the frequency of element breakage, the location of break on the bone, the tools employed, and the type of site where the activity occurs as a way to develop an understanding of the organization of carcass processing. Transport distance is known to be a major variable in field processing and transport decisions (Bunn et al. 1988; O'Connell et al. 1990). Data presented here suggest that two other variables guide processing and transport decisions, carcass size and later processing options, particularly cooking (see also Bunn et al. 1988). The second objective is to show that the ultimate goal of many Hadza carcass processing decisions (guided as they are by carcass size and later processing options) may be to maximize nutrient returns while minimizing field processing and transport costs. Note, however, that I am not explicitly testing an optimization model; alone, data presented here are insufficient to test such a model. Rather, I merely observe that nutrient maximization and reduction of field processing costs are reasonable (though perhaps ad hoc) explanations of some carcass ~rocessi~? behavi.ors. These observation.s on the. full range of carcass processing deciSions are 1mportant, not because 1n sorrw 1nstances they are at odds with previous optimization explanations of Hadza carcass processing that dealt only with transport (i.e., O'Connell et al. 1990) but because they establish a broader behavioral context for transport decisions. Nor do I assume all Hadza carcass processing behavior to be rational and economically based. Several observations suggest that Hadza treatment of dikdik (Rhynchotragus kirki) and impala (Aepyceros melampus) carcasses may be mitigated by social costs associated with the inability to adequately share such small animals. Other observations and informant data indicate that taste may also be an important variable in explaining treatment of some animals, particularly zebra (Equus burchelli) and wildebeest (Connochaetes taurinus).
The third obj Hadza carcass tions of some 1 ative of partie carcass proces cooking (boilir sions, serves a: cessing for prE nologies. Expe' extraction tech bone breakagE Pleistocene hm elements becm smashing and parts. Increase' fire-based innc underscore thE help understm organization o face value, the Pleistocene hor
Me The rugged semim south and east pological stud and Bartram 1 Woodburn 196 bone breakagE through Octob archaeology Pr Direct obser (Table 12-1).3 1 more easily tr• camp, or (3) re: of the distal lin that provided consumption a behavior wher . to facilitate pic (hearth-sizing; sumption if th breakage. Parts of mas
From Butchery to Consumption I 203 ~ boiling might ! only cooking ~first
problem bone breakJrocessing as a . rvations. Bone !es of detail for 3; Gould 1967; ; Yellen 1977), 1ut 1967), and owever, those 1izational per·eakage behav)gical expecta~d
consumptionup of hunter! objectives to )r each behavncy of element i, and the type ,tanding of. the 1 to be a major n et al. 1988; )ther variables ter processing many Hadza and later prolnimizing field plicitly testing ~n t to test such Ld reduction of xplanations of ~ full range of instances they -Iadza carcass · al. 1990) but Jort decisions. :ional and ecoltmentof dik:asses may be :ely share such that taste rhay nilnals, particzus).
The third objective is to point out the obvious archaeological implications of Hadza carcass processing behavior. In part, this objective is met with descriptions of some bone modifications and frequency patterns that may be indicative of particular processing behaviors. Additionally, the organization of carcass processing by the Hadza, particularly the apparent importance of cooking (boiling and roasting) in structuring earlier carcass processing decisions, serves as a basis for a preliminary evolutionary model of carcass processing for pre-fire, post-fire, pre-bone-boiling, and post-bone-boiling technologies. Expectations derived from this model suggest that as new nutrient extraction techniques developed, changes occurred in carcass transport and bone breakage patterns. In particular, the expectations suggest that PlioPleistocene hominids lacking fire would have preferentially transported limb elements because the only available nutrient extraction techniques of bone smashing and gnawing would have precluded efficient processing of axial parts. Increased transport of axial parts is expected for hominid groups after fire-based innovations in nutrient extraction technology. These expectations underscore the difficulties in using carcass processing by modern groups to help understand the past without defining the variables that structure the organization of a particular behavior. Thus, the data suggest that, if taken at face value, the Hadza carcass transport decisions are a poor analogue for PlioPleistocdne hominids (see also Bunn et al. 1988; O'Connell et al. 1988a, 1990).
Methods The Hadza are a small group of hunter-gatherers occupying the rugged semiarid savannas and thorn scrub habitat in the hills and plains south and east of Lake Eyasi, Tanzania, 1 and have been the subject of anthropological study for many years (Bunn, this volume; Bennet al. 1988; Bunn and Bartram 1990; O'Connell et al. 1988a, 1988b, 1989, 1990; Vincent 1985; Woodburn 1964, 1968a,1968b,1970, 1972). Observations on carcass processing bpne breakage reported here were made during the dry season (August through October) as part of the 1988 University of Wisconsin-Madison Ethnoarchaeology Project directed by H. T. Bunn.2 Direct observations of bone breakage were made for parts of 24 animals (Table 12-1).3 Butchery is defined as the sectioning of a carcass to (1) create more easily transportable units, or (2) facilitate carcass part distribution in camp, or (3) remove less desirable units from a larger body part (e.g., severing of the distal limb from the upper limb). This definition is more restricted than that provided by Lyman. (1987:252) in that I regard cooking preparation and consumption as separate behaviors. Cooking preparation breakage is a limited behavior where the bone is broken immediately prior to cooking specifically to facilitate placement in a cooking pot (pot-sizing) or on a fire for roasting (hearth-sizing) ..Finally, a bone is defined as having been broken for consumption if the marrow and/ or cancellous tissue is eaten immediately after breakage. . Parts of most carcasses (N = 15) were followed from primary field butchery,
204
I J. S. Oliver Table 12-1. General Carcass Procurement Data
Taxa Dik-dik Dik-dik Impala Zebra Zebra Wildebeast Zebra Zebra Zebra Zebra Impala Zebra Buffalo Zepra Zebra Buffalo Zebra Warthog Dik-dik Impala Impala Cow Baboon Impala
Size Classa la la 2 3b 3b 3b 3b 3b 3b 3b 2 3b 4
3b 3b
Sex
Age
pb
F F M M M M
Adult Juv Juv Adult Juv Adult Adult
pb
Adult
F M
Adult Adult Adult' Adult Juv Adult Adult Adult
M F M
4
3b 2 la 2 2
M M M
Juv Adult
4
2C 2
F F
Adult Adult
Procurement Method Evening hunt Day scavenge Day scavenge Night hunt Night hunt Night hunt Night hunt Trade w I Isanzu Isanzu night hunt Trade w I Isanzu Day hunt Night hunt Night hunt Night hunt Day scavenge Day scavenge Night hunt Day hunt Day scavenge Evening hunt Day hunt Trade w I Tatoga Evening hunt Day hunt
.· I asize class definitions follow those provided by Bunn et aL (1988: Table 1). la = 7-9 kg; lb = 9-23 kg; 2 = 23-113 kg; 3a = 113-204 kg; 3b = 204-340 kg.; 4 = 340-907 kg; 5 = 907-2722 kg; 6 = > 2722 kg. bWith fetus. cThough technically baboons belong in size class lb, this baboon was placed in size class 2 because baboon bones are significantly stronger than those of other size 1 animals.
to secondary butchery and defleshing,in camp, to preparation for cooking, to cooking, and to final consumption. Breakage of bones from parts of other carcasses were observed at various stages from butchery to consumption (N = 8). All parts of only one carcass, an adult female dik-dik, were "tracked". completely from initial butchery through complete consumption. Breakage events occurred at residence camps and kill sites. Snack sites like those reported by Bunn and colle9-gues (1988) were not observed. However, a few bone breakage events (N = 3) were observed after successful begging for food at a neighboring Tatoga (a pastoralist group) camp and at an Isanzu (agriculturalists, most of whom live in camps at the southern end of Lake
Eyasi) carca HadzahelpE The residenc the bones aJ those preset Is anzu procE Bone brea ahammerst< panga or axE knife, (5) WJ before. and ; technique, a quent fractu initial break and site ofb in addition struck to bn of cooking removal fro: times were I
First, the ge carcass pro' sumption, a during butc dence camp description~
ous breakag Hadza m1 hunting frc encounter l scribed by I 1990), Addi
services for sharing wit Hadza area quently pro N = 10, 41 (Phacochoen and baboon only other (Rhynchotra 8.3%; Table
From Butchery to Consumption I 205
rement Method
lng hunt :cavenge :cavenge thunt thunt t hunt thunt ~ w/ Isanzu unighthunt ~ w/ Isanzu mnt thunt thunt thunt scavenge ;cavenge thunt mnt :cavenge lng hunt mnt ~ w/ Tatoga lng hunt mnt l
= 2-9 kg; lb = = 907-2722 kg;
:i in size class 2 mals.
or cooking, to :Jarts of other mmption (N = tracked" comnack sites like d. However, a Jl begging ,for . at an Isanzu 1 end of Lake
Eyasi) carcass processing station4 where nine limb bones from a zebra the Hadza helped butcher and transport were broken and the marrow consumed. The residence camp sample includes these events because (1) the Hadza broke the bones and consumed the marrow, (2) the tools at the sites included all those present at Hadza residence camps, and (3) the broken bones from _the Isanzu processing camp were transported to a Hadza camp. Bone breakage techniques include (1) striking a bone lying on an anvil with a hammerstone, (2) clubbing bone onto an anvil, (3) chopping at a bone with a pariga or axe, (4) hacking at a bone (with or without the use of an anvil) with a knife, (5) wrenching a bone, and (6) gnawing. Bone breakage occurred both before and after cooking. Many limb bones were initially broken, using one technique, and later the exposure of medullary bone was increased by subsequent fracture, sometimes using a different technique. Direct observations of initial breakage techniques for 879 axial and limb bones by animal size class and site of breakage are reported here (Table 12-2; Figure 12-1). Data recorded, in addition to breakage technique, are (1) the number of times a bone was struck to break it for marrow I cancellous tissue consumption, (2) the method of cooking (Table 12-3; Figure 12-2), and (3) the cooking time until first removal from fire or boiling pot for initial consumption (subsequent cooking times w~te not recorded).
Results Observations of bone breakage are divided into three sections. First, the general features of the sample are presented (Table 12-1). The three carcass processing behaviors, butchery, preparation for cooking, and consumption, are then discussed. The frequencies of axial and limb bones broken during butchery, cooking preparation, and consumption at kill sites and residence camps are provided in Table 12-2 and Figure 12-1. Finally, general field descriptions of breakage and cooking-induced damages created by the various breakage techniques are presented. Hadza meat procurement methods include hunting (day and night ambush hunting from hunting blinds near game trails and water holes, and day encounter hunting) and scavenging. These hunting methods have been described by Bunn and colleagues (1988) and O'Connell and colleagues (1988a, 1990). Additional procurement methods observed in 1988 include performing services for the Tatoga (a pastoralist group; one observation) and trading and sharing with the Isanzu (agriculturalists living to the south who enter the Hadza area in the dry season to hunt; three observations). The most frequently procured taxa for the breakage observations are zebra (Equus burchelll; N = 10, 41.7%) and impala (Aepyceros melampus; N = 5, 20.8%) .. Warthog (Phacochoerus aethiopicus), wildebeest (Connochaetes taurinus), cow (Bas sp.), and baboon (Papio cynocephalus) are each represented by one individual. The only other taxa represe?ted by more than one individual are dik-dik (Rhynchotragus kirki; N = 3, 12.5%) an.d cape buffalo (Syncerus caffer; N = 2, 8.3%; Table 12-1).
Wrnch Knife
Panga I Ax
Table 12-2.-Continued KILL SITE Hstn& Club& Anvil Anvil Gnaw
No Site Data Total
Wrnch
Knife
RESIDENCE CAMP Panga Hstri& Club& Anvil Anvil Gnaw I Ax
No Site Data Total
TOTAL
r-
--
LlffiD
0 0
u
0
0
u
0
2 6 0
0 0
0 0
0
0
u
6
185 0
30 1 31
0
0
0
199
0
199
14 0
0
2
2
0
0 0
0 0
0
0
0
0 2
0
0 0
0
0
0
2
0
0 6 2 8
0
0
0
0
0
0
28 1
0 0
0
0
0
0
2
0
0 0
0
0
0
0
0
0
0
0
2
0 0
0 0
0
-0 0
0
0
0 0
0
0
1 -
0
0 0
0 0
0
/Ax
0
0
u
KILL SITE Hstn& Club& Anvil Anvil Gnaw
0
0
u
0
Wrnch Knife
Panga
Table 12-2.-Continued
0 0
u
CONSUMPTION Size Class I Axial 0 Limb 0 Size Class II Axial 0 Limb 0 Size Class III Axial 0 Limb 0 Size Class IV Axial 0 Limb 0 BEHAVIOR SUM Size Class I Axial 0 limb 0 Size Class II Axial 0 Limb 1 Size Class III Axial 30 Limb 0 Size Class IV Axial 0 limb 0 All Classes Axial 30 1 Limb TOTAL 31
:Limb
Size Class IV Axial
0
0
u
0
75 0 75
0
0
0
75
0 0
0 0
0
28 4
342 4 346
28
0
28
0 0
14 0
0
0 0
298
0
0
0
28
2 0
0
0
0
0
0
0 0
0
2
Wrnch
0
0
u
0
0 0
0
0 0
8
3
0 0
0 0
0 0
No Site Data Total
0
0
u
178 7 185
0
0
0 0
0
85
93 7
0 0
0 0
0
0
0 7
Knife
------
0
0
u
0
14
;;
0
0
u
0
0
u
109 10 119
0 0
67 1
42 9
0
0
0
0
1 1
0
0
0 0
0
0
121
83
38
14 19
11 42
11 20
2 2
19
0
0 4 4
54 2 56
0
0
0
1 0 0
0
0 0
54 2
0
3
0
0
0
0
o· 0
0 0
0
1 37
0
0
54 2
1
0 1
0 11 17
2 2
RESIDENCE CAMP Panga Hstn& Club& Anvil Anvil Gnaw /Ax
0
0
u
0
14
::J
18 2 20
0
0
0 0
0
0
18 _2
0
0
0
0
0 0
2
0
425 108 533
14 19
44
78
138 32
195 13
19
0
39
2
11 18
58 13
No Site Data Total
0
0
u
767 112 879
28 19
44
376
166 36
197 13
19
0
10 39
11 21
58 13
TOTAL
0
14
J
~ 0
"'
0
-N:::::s
;:t 0
'ij
sc
fJJ
:::::s
0
n
0
r-1-
'--
ast the femur hite, solidified r time I saw a ~ason
r markedly for fammerstones :txial and limb size I animals ro of the size I tpodials) were nching action.
Only two limb bones were broken with a hammerstone, and seven were broken with knives.
Bone Breakage Damage Morphologies The following descriptions of bone breakage morphologies are based on field observations of bones from both the breakage events reported above and the inspection of bone in refuse piles of occupied camps. Although damage to axial parts is mentioned, damage to limbs is emphasized here. Fracture features described for a given breakage technique are not necessarily present on every bone processed in a similar manner. Rather, damages described for each butchery, cooking, and marrow consumption breakage technique represent a composite signature of observations of a number of bones processed in a similar manner. Most of the observed breakage techniques involved dynamic loading (panga and axe chops, knife hacks, hammerstone impacts and clubbing bone on anvils), one involved static loading (gnawing), while a third involved wrenching. Hammerstone and anvil breakage events produced fracture features typical of those noted in the experimental (e.g., Blumenschine and Selvaggio1988; Bonnichsen 1979; Bunn 1989; Johnson 1985) and ethnoarchaeological/(e.g., Binford 1981; Bonnichsen 1973) literature: (1) concentric flake scars at the loading and rebound points, (2) broad impact notches, (3} negative flake scars at the loading points, (4) hackle marks or other stress features, (5) a smooth fracture surface interrupted only by (6) stress features indicative of fracture front direction, (7) a smooth, curvilinear (sometimes "spiral") fracture outline, and (8) impact pits and scratches near the loading points. Though it is well known that carnivore activity (e.g., Binford 1981; Haynes 1983; Hil11989; Morlan 1980; Potts 1988) and some geological processes and trampling (e.g., Behrensmeyer et al. 1989; Fiorillo 1989; Oliver 1984, 1986, 1989) are capable of producing some impact-like features, they are not created with thf£·+egularity or frequency of hammerstone and anvil bone breakage (Oliver 1986, 1989, 1,991). That is;: it is the association of impact features and their context in relation to oth~F damages that defines hammerstone impact damage or any other fracture agent (Oliver 19.91). Because chopping and hacking of bone with metal tools are essentially dynamic loading actions, damages similar to those produced by hammerstone and anvil breakage are created, including negative flake scars and impact notches (which, if present, tend to be linear rather than oval in appearance). Instead of the irregular crushed area with a concentric ring of flakes produced by hammerstones, metal-tool-induced breakage produces a characteristic notch or flat-sided area where the blade cut through the bone prior to bone fracture (see also Lyman 1987). Of the 112 limb bone breakage events, 47 (42%) were struck at more than one location. The mean number of blows struck to dik-dik, Lrnpala, zebra, and buffalo limb bones is 1.7, 7.1, 9.9, and 14.6, respectively, clearly indicating a positive relationship between body (and therefore bone) size and the average number of impacts per bone. (Baboon and warthog bone breakage events may
---------------------~-
214
---~-~------,_----------
I]. S. Oliver
be excluded from this discussion because of their proportionally greater density and because several bone"s were pounded on not to obtain marrow, but simply to occupy time.) Repeated blows to different locations on a bone has a number of results. First, it damages the bone to a greater extent than a single impact at one location. More impact points, fracture lines, and ultimately bone fragments are created by repeated impact events (see also Bunn 1989). Repeated blows serve not only to straighten the fracture front but also create many more impact notches than suggested by the early experimental literature (e.g., Bonnichsen 1979). Finally, repeated blows after breakage often split limbs across the articular surface. , Although most larger limb bones were struck so that diaphyses fragmented into numerous pieces, bone tubes were created by striking both the proximal and distal diaphyses of several size I and IT limb bones. Aft~r marrow was sucked out, the exposed ends were chewed; without careful inspection of fracture features, these bones_ could easily be interpreted as the result of chewing by a small carnivore. Further, as stated above, pot-sizing of size I and II limbs required only a few blows to the midshaft, thereby reducing fragmentation. Post-boiling processing of these limbs for cancellous tissue may create additional impact damages. Boiling bone does not yield visible damage, nor does breaking previously boiled bone result in a diagnostic breakage morphology. Observed boiling times are short, however, usually less than 11 minutes. Observation of boiling ended after the bone was first removed from the pot and some tissue consumed. Many bones were subsequently returned to the pot, but the boiling times were not recorded. Longer boiling times may result in visible damage. Roasting slightly chars some limb bones, but many show no charring at all. Further, it should be noted that the Hadza typically do not discard bone into hearths. Breakage of roasted limb bones does, however, yield breakage feayures not at all like those created when green, uncooked bone is broken. Rather than the smooth, curvilinear fractures typical of hammerstone breakage of fresh limb bones, breakage of roasted limb bones produces jagged fracture lines that are frequently oriented transverse to the bone's long axis; associated damage includes hammerstone and anvil dents and scratches, and impact flakes. Overall, broken roasted bones are somewhat reminiscent of bone fractured after significant desiccation or fossilization. Structural alteration of bone due to burning (Shipii'.an 1981; Shiprnan et al. 1984) as v.rell as simple desiccation of outer compact bone may partially explain this fracture pattern. Microscopic examination of bones fractured after roasting and boiling would be informative. Gifford-Gonzalez (1989, this volume) has also noted similar bone fractures in faunal assemblages from abandoned Dassanetch camps, which, she suggests, were created by cooking. Finally, marrow and cancellous tissue extraction from broken bones created diagnostic damage. Marrow removal was accomplished in a variety of ways including (1) sucking, (2) pounding th~ exposed end on a rock, and (3) using a knife, stick, or bone fragment as a scraping tool. Binford (1981) reports the same marrow extraction techniques for the Nunamuit. Though cut marks
could not be row undoubt, Cancellous consisted of 1 hcrmmerstone knife or pang This action n pieces wi tl1 V nique was to the cortical t impact dents and scratches This suggesh much tissue a
D
n ior-specific br set, those bel the Hadza. T. vs. prehistori1 in all other ca 1981; Gifford Yellen 1977 < theless, the r goals serve a processing he tion of carcas from initial fi tion-may h extraction is r minimized. ~ define those structuring c B.unn et al. 1 available nut carcass proce:
0; H; consideration organization maximizing r smaller carca~
From Butchery to Consumption I 215 ·nall y greater 1tain marrow, ms on a bone extent than a nes, and ulti:;ee also Bunn front but also experimental reakage often ~s
fragmented the proximal marrow was ection of fracllt of chewing I and II limbs ~agmentation.
'f create addi-
1g previously ~rved boiling :ion of boiling 1e tissue conut the boiling sible damage. 1arring at all. ard bone into e features not ::tther than the of fresh limb lines that are d damage inflakes. Over~actured after :bone due to desiccation of . Microscopic d be informalar bone £raeIS, which, she bones created riety of ways nd (3) using a 1) reports the ~h cut marks
could not be identified in the field, _the intense use of knives to remove marrow undoubtedly produced cut marks on the medullary wall. Cancellous bone removal for consumption (both cooked and uncooked) consisted of repeatedly striking the diaphysis near the epiphyseal end with hammerstone on an anvil until cancelldus bone was exposed. Once exposed, a knife or panga was used to cut off pieces of cancellous tissue to eat or suck on. This action resulted in an area of consumption littered with cancellous limb pieces with V-shaped notches and/ or planed cancellous tissue. Another technique was to strike the exposed cancellous tissue with a hammerstone (while the cortical bone rested on an anvil). This technique created hammerstone impact dents and crushed areas of cancellous bone; the anvil left pits, dents, and scratches on the cortical surface. Generally, little edible tissue is ignored. This suggests bones may be transported to camps to cook and consume as much tissue as possible.
Discussion The above data show that Hadza carcass processing yields behavior-specific breakage patterns related to carcass size and cooking options. As a set, thos~ behaviors and the resultant patterns are probably specific only to the Hadza. That is, we may not conclude that correspondence in one (Hadza vs. prehistoric pattern) breakage or frequency pattern implies correspondence in all other carcass processing or cultural behaviors (see, for example, Binford 1981; Gifford 1980; Gifford-Gonzalez 1991; Oliver 1989; Saunders 1990; and Yellen 1977 on how analogues may be used to interpret the past). Nevertheless, the manner in which carcass size and ultimate nutrient extraction goals serve as boundary conditions to structure the organization of carcass processing has a number of archaeological implications. First, this examination of carcass processing behaviors has elucidated how several behaviorsfrom initial field butchery, to transport, to cooking-preparation, to consumption-may be integrated into a decision-making strategy where nutrient extraction is maximized, while transport, field processing, and social costs are minimized. Specific bone modification and frequency patterns may help define those behaviors. Furthermore, the important role cooking plays in structuring carcass processing behavior for modern people (Binford 1978; Bunn et al. 1988; Gifford-Gonzalez, this volume; Yellen 1977) suggests that available nutrient extraction techniques also helped determine prehistoric carcass processing and transport strategies.
Organization of Carcass Processing Hadza carcass processing decisions are clear! y made with due consideration of .carcass size and later nutrient extraction options. Further, the organization of carcass processing appears to be structured by the goal of maximizing nutrient returns while min~mizing transport, processing, and, for smaller carcasses, social costs.
216
I ]. S. Oliver
Processing and transport CO$tS for smaller carcasses are minimized by almost always transportingthe complete carcass to residence camps (five of six dik-dik and impala carcasses for .which I have transport data). These data are at odds with those provided by O'Connell et al. (1990, 1991), who observed frequent processing, consumption-of tissue, and eleme11t abandonment at kill sites. Nevertheless, situational information for six impala and dik-dik carcasses reported here may help to explain the conflicting observations. Specifically, the Hadza decision whether or not to transport a small carcass to camp seems to be made by evaluating the ability. to adequately share small carcasses with camp members. Three of the small antelope carcasses were transported complete by hunters to relatively small camps where they were easily shared. One of the three was found dead by a Tatoga herder and the location reported to a Hadza boy who clearly could not have consumed all of the carcass. Another scavenged dik-dik was found after the entire camp population was alerted to vultures circling just outside camp. The remains of this animal (all but the right hindlimb, pelvis, sacrum, and four lumbar vertebrae) were returned to camp compl~te, but not all camp members partook of this food. Most parts of the- animal were eaten by women and small children. One dik-dik was killed at night very near camp and transported complete to the camp edge where it was processed and totally consumed by two hunters without the knowledge of other camp members. Finally, one impala was largely consumed by several teenage boys at the kill site; The teenage boys' decision to consume most of a female impala at the kill site was apparently made because (1) the adult hunter had given up the animal as lost, (2) the animal was in poor nutritional condition, and (3) they would have been unable to adequately share meat from such a small carcass with the large Hadza camp. 5 Whether or not these situational variables also explain the impala transport pattern reported by 0' Connell and colleagues (1990) is unknown at this time, but the observations do suggest thlcit evaluation of the ability to share the kill may play a determining role in s1ze I and II carcass transport patterns. With some notable exceptions (wildebeest, giraffe, and apparently eland; O'Connell et al. 1988a, 1990, 1992) most carcass parts of larger animals, particularly zebra, are transported to camp (Bunn, this volume; Bunn et al. 1988; Bunn and Bartram 1990). Again, the overriding consideration seems to be maximizing nutrient returns by transporting most of the animal back to camp where nutrient extraction (i.e., through bone smashing and boiling) could proceed at leisure. Carcasses were butchered only to the extent that it facilitated transport of most carcass parts back to camp using available carriers. With regards to the single wildebeest killed in 1988, consumption of most limb marrow at the kill site and transport of axial parts are consistent with earlier observations of the unique treatment of wildebeest (Bunn et al. 1988; O'Connell et al. 1988b, 1990). According to O'Connell and colleagues (1990, citing Metcalfe 1989), the transport pattern is explicable in that the probability that a particular element will be transpbrted is related to the ratio of edible to inedible tissue and the processing costs. Carcass parts with high edible to inedible tissue ratios are likely to be transported, while bones from those with
low ratios ar1 transport sh may be relat, O'Connell a that they do taste good (I commented sweet. 6 The i transport of Isanzu procE bones in the ing options i The argur while other ' diate desire: .nature, a risl tiona! basis. social costs c. bers and thE Similarly, tb wildebeest N evertheles~ satisfying of Maximizil seen in the c. axial and lin whereas axic: from size I a but is not n1 limbs that a1 marrow and subject to co: O'Connel1 erentially trc ing higher-u Perkins and Bunn 1986; I strategy wa~ carcasses w Bunn, this v however, it' strongly infl choose to t: defleshing c high field f scraps, can t bral column:
From Butchery to Consumption 1217 ninimized by ::amps (five of :a). These data ), who observandonment at a and dik-dik obsE;rvations. nail carcass to ly share small ete by hunters : the three was adza boy who ·enged dik-dik 1tures circling ght hindlimb, tmp complete, .e animal were Light very near processed and ercampmemtge boys at the ~impala at the i given up the 1, and (3) they L small carcass variables also .nd colleagues st that evaluain size I and II arently eland; nimals, particnn et al. 1988; rr seems to be l back to camp ng) could pro3.t it facilitated carriers. With of most limb nt with earlier 1 et al. 1988; leagues (1990, :he probabi.lity tio of edible to high edible to om those with
low ratios are more likely to be abandoned at the kill site. Processing costs and transport should be inversely related. Thus, wildebeest transport patterns may be related to taxa-specific inedible to edible tissue ratios as suggested by O'Connell and colleagues (1990). Nevertheless, Hadza informants indicate that they do not carry wildebeest limbs to camp because the marrow does not taste good (H. T. Bunn, personal communication). Conversely, they frequently commented on their preference for zebra meat and marrow because it is so sweet. 6 The importance of taste in transport decisions is also suggested by the transport of about 10-13 broken zebra limb bones (lacking marrow) from the Isanzu processing camp to a Hadza residence camp. The boiling of the zebra bones in the Hadza camp further underscores the importance of later processing options in transport decisions. The arguments that some processing decisions are economically based while other decisions may be guided by considerations of social costs, immediate desires, and taste may ~eem incongruent, but sharing is, by its very nature, a risk reduction strategy whose solution must be evaluated on a situational basis. As such, the little meat a dik-dik or impala offers may not offset social costs associated with dividing such a carcass fairly among camp members and the nutritional gain on the part of those who decide not to share. SimilarlY;~ the "good taste//, of zebra meat and marrow compared to that of wildebe~st may in fact have an as yet unrecognized nutritional basis. Nevertheless, I do not believe all human behavior is totally rational, and some satisfying of immediate desires certainly occurs. Maximizing nutrient extraction while reducing processing costs may be seen in the animal-size-dependent cooking preparation and consumption of axial and limb parts. Pot- and hearth-sizing is minimal for smaller carcasses, whereas axial parts of larger carcasses are heavily processed. Cooking of limbs from size I and II carcasses apparently increases marrow extraction efficiency but is not necessary for limbs of larger animals. In cor.trast to size I and II limbs that are often minimally broken for marrow, the increased quantities of marrow and cancellous tissue for size class ill and IV limb bones make them subject to considerable processing and fragmentation. O'Connell and colleagues (1988a; 1990)·argue that because the Hadza preferentially transport axial and other low-meat utility parts to camp (abandoning higher-utility wildebeest and impala limb bones at kill sites) the use of the Perkins and Daly model (1968) to reconstruct early hominid foraging (e.g., Bunn 1986; Bunn and Kroll 1986) behavior is open to question. This transport . strategy was not observed in the 1988 work with the Hadza where almost all carcasses were transported nearly complete to residence camps (see also Bunn, this volume; Bunn and Bartram 1990). With either transport pattern, however, it appears that later nutrient extraction options, in this case cooking, strongly influence Hadza transport decisions. Thus, it appears that the Hadza choose to transport axial parts because (1) meat scraps remaining after det1eshing cannot be readily removed without cooking, that is, they have a high field processing cost, and (2) the available nutrients, including meat scraps, can be readily extracted through boiling (a process to which all vertebral columns were subjected) in residence camps where boiling pots a~e kept.
218
I]. S. Oliver
Similarly, complete limbs of most taxa may be transported because (1) the field processing time is reduced by foregoing time-consuming defleshing and transporting entire limbs to camps w.here defleshing and limb dismemberment can take place at leisure and (2) the nutrient value of limb bones is increased by boiling and by intensive breakage for cancellous tissue. Hadza transport strategies are thus a poor analogue for pre-fire hominids. Therefore, O'Connell and colleagues' (1990) criticism of carcass transport strategies of early hominids reconstructed by Bunn (Bunn 1986; Bunn and Kroll1986; Potts 1988) because they do not match Ha~za transport decisions is an example of interpretative risks arising from the episodic treatment of human behavior. There is no reason to suspect·that early hominids who lacked cooking technology would· make the same economic decisions as groups with more efficient nutrient extraction techniques (see below). This perspective is lost if only one component of an overall carcass processing strategy-in this case, transport-is the sole focus of attention. Furthermore, the stated Hadza concern with taste (poor tasting wildebeest marrow vs. sweet zebra meat and marrow) and the apparent concern with ability to adequately share small carcasses demonstrate the difficulties in applying economic models of human decision making. Economic concerns are obviously intertwined with social concerns and the need to satisfy personal tastes. This does not mean that economic models such as optimal foraging seriously misrepresent human behavior and should not be used. Economic models are useful heuristic devices that serve to structure our observations about human subsistence behavior and may sometimes explain much of it. Many of the observations reported here, for example, seem to fit predictions recently made by Metcalfe and Barlow (1992) in their model of the optimal trade-off between field processing and transport. Significantly, however, in cases where human subsistence behavior does not seemxational (in this case, the Hadza treatment of size I and II carcasses and wildebeest), cult4tal variability or other previously unconsidered variables may be highlight~d. Thus, Hadza carcass transport behavior appears to be mitigated by cultural attitudes or perceptions of sharing and taste and cooking options, in addition to a desire to optimize energy procurement while reducing field processing and transport costs. Further, it should be noted that although variable cultural attitudes cannot be effectively incorporated into economic models, processing options like cooking technology could be used (with appropriate experimental data on nutrient yields of various cooking technologies) to calculate the utility of transported loads discussed by Metcalfe and Barlow (1992).
Bone Modification and Frequency Patterns Generally, data presented above suggest that past carcass processing behaviors may be more fully understood through examination of bone fracture damage. Significant breakage observations include those related to (1) animal size and behavior dependent use of heavy- and light-duty tools, (2) multiple blows struck on large limb bones to expose areas of cancellous bone, (3) use of knives to scrape marrow and cut cancellous tissue from broken
bones, (4) che, breakage of b si ve breakage pots, and (7) u E>ifferential of smaller ani marrow. or pc fragmented t] bones, particu of larger. anin better techniq limb shaft fra 1991). The re· bones from li carcasses furt: and boiling rr cessing of sm limited mids] light gnaw da The intensi· ed by thedeg des, and dar ological asse1 points, a hig: damage to ca marrow cons1 by numerous pieces, and p: ing intensity £ragmen ted 1 animal axial f The diagno the potential1 of fire has h Swartkrans 0 Analysis of b animal food ' some concerr again that thE ior and the 1 outward sigr calcined. bon' cooking in tl about using 1987). Differentia1 is also of con~
From Butchery to Consumption j219 ~cause
(1) the efleshing and ~ dismemberimb bones is :issue. Badza ls. ass transport ~6; Blinn and rt decisions is treatment of )minids who decisions as below). This :;s processing Furthermore, t marrow vs. ith ability to applying ecoue obviously tastes. :oraging seri~d. Economic observations n much of it. it predictions f the optimal , however, in l (in this case, cultural variighted. Thus, :ural attitudes 1ddition to a ·ocessing and ~able cultural ls, processing experimental ate the utility
rcass processltion of bone : related to (1) luty tools, {2) tcellous bone, fron1 broken
bones, (4) chewing of small carcass ribs and articular ends of limbs, (5) limited breakage of bones from small carcasses in preparation for boiling, (6) extensive breakage of large carcass axial parts. to facilitate placement in boiling pots, and (7) unique breakage of roasted bones. Differential treatment of large and small carcasses suggests that limb bones of smaller animals (for which minimal br~akage is required to either extract marrow or pot-size) may enter the post-processing taphonomic system less fragmented than those of larger animals. Conversely, small mammal axial bones, particularly vertebrae, may be more highly fragmented than vertebrae of larger animals. This observation underscores the necessity of developing better techniques for estimating element frequencies (11NE) such as counting limb shaft fragments (Bunn 1982; Bunn and Krolll986; Marean and Spencer 1991). The repeated blows to limb bones and extensive pot-sizing of axial bones from larger animals compared to the minimal processing of smaller carcasses further suggests that archaeological visibility of marrow processing and boiling may be biased against small animals. Nevertheless, marrow processing of small ungulate limb bones may be indicated by (1) a pattern of limited midshaft impact breakage and (2) the presence of bone tubes with light gnaw damage near articular ends. The intensity of limb marrow I cancellous tissue processing may be indicated by th~ degree of diaphyseal fragmentation, complete articular end frequencies, and damage to the medullary wall and cancellous bone. Zooarchaeological assemblages with a large limb bones with few diaphyseal impact points, a high frequency of intact articular pieces, and a lack of crushing damage to cancellous tissue may indicate minimal processing such as only marrow consumption or bone boiling. In contrast, assemblages characterized by numerous diaphyseal fragments with impact marks, fragmented articular pieces, and pitted/ crushed/ sliced cancellous bone would indicate a processing intensity closer to that of the Hadz;a. Similarly, assemblages with highly fragmented large ·animal axial parts, but containing relatively intact small animal axial parts, likely indicate bone boiling. The diagnostic damages created by breakage of roasted limb bones suggests the potential to directly identify this behavior from sites where the earliest use of fire has been reported, for example, Chesowanja (Gowlett et al. 1981), Swartkrans (Brain and Sillen 1988), and FxJj20 at Koobi Fora (Bellomo 1992). Analysis of bone breakage from those sites could determine whether or not animal food was being roasted, broken, and consumed, thereby eliminating some concern about naturally calcined and charred bone. It should be noted again that the Hadza typically do not discard bone into the fires. This behavior and the -fact that most bones- are boiled and -few roasted bones show outward signs of exposure to fire suggests that frequencies of charred and calcined bone may not be very good indicators of either the use of fire or cooking in the archaeological record. Others have offered similar cautions about using burned bone as an indicator of cultural activity (e.g., Lyman 1987). Differentiation between heavy- and light-duty tool use in carcass processing is also of concern. Binford (1981) and others have commented on the peculiar i
220 I J. S. Oliver nature of chopping/hacking butcllery with heavy-duty tools, suggesting that their presence may alter transport strategies. Others (e.g., Bunn et al. 1980; Jones 1980; Toth 1985), however, hav~ argued that there may have been functionally equivalent stone tools in the past. Detailed examination of zooarchaeological assemblages for heavy- versus light-duty-tool:-induced damage patterns should help (1) refine our understanding of carcass processing, (2) determine whether or not heavy-duty metal tools significantly affected carcass segmentation or transport strategies, and (3) compliment Toth's tool use, curation, and transport data for Plio-Pleistocene hqminids. That is, definition of the tool used to create the cut marks at sites that lack stone tools (e.g., GaJiS; Bunn 1981, 1982) may help define early hominid tool transport strategies in areas distant from lithic resources (e.g., the Karari escarpment). Finally, bonebreakage and consumption observations reported here indicate that the Hadza do not frequently create archaeologically visible kill sites. Thus, for technologically similar human groups, few kill sites would be found in the archaeological record. Sites are created where hominids repeatedly gather to process and consume food. When group size permits adequate sharing, size I and II carcasses are typically transported complete to residence camps where butchery is undertaken; nothing remains at the kill site. Though processing activities are significantly greater at kills of size III and IV animals, most parts are transported to camp; only a few roasted ribs and possibly a smashed skull and mandibles near an ash scatter remain as evidence of the butchery site. Zebra kills at which informants report roasting of skulls and numerous ribs lacked all but two or three ribs a few weeks after the kill. Archaeological visibility of kill sites reported by 0' Connell and colleagues (1988a, 1988b, 1990, 1992) would be greater, particularly for large animals.
Innovations in Nutrient Extraction Technology and the Evolution of Carcass Processing The organization of Hadza carcass processing behavior thus (1) provides a baseline with which to compare zooarchaeological assemblages and (2) suggests new ways to approach prehistoric carcass processing behavior. Specifically, the role late-stage carcass processing options (i.e., nutrient extraction techniques like bone smashing, roasting, and boiling) play in structuring carcass processing strategies suggests that the organization of carcass processing evolved with innovations in nutrient extraction techniques. As new nutrient extraction technologies developed (e.g., stone tools for cutting meat and smashing bone, fire for roasting and bone-boiling facilities or implements), the relationship between processing and transport costs and nutrient value changed. All things being equal, a number of expectations about the evolution of carcass processing may be derived from the Hadza data. Nutrient extraction techniques available to Plio-Pleistocene hominids were limited to bone smashing and gnawing (but see Bellomo 1992). Thus, the difficulty in efficiently extracting nutrients from defleshed vertebrae and ribs without cooking suggests that, without fire, hominids should have preferentially transported
marrow-rich marrow and hammers ton marks, and J Plio-PleistocE Chewing 2 technique av inid tooth .m present) to b blages create to be a. comr (Figure 12-1E er carcasses r As fire be: axial parts ' removal of r be evidencec hearth-sizin~ promote adc the advent o axial parts, ~ cessed as is availability o Data from meet the pn 1986, 1988) ] Pleistocene s Marean, this fracture patb of the identi: 1991), suggE available tee increase in n gies may ha~ time, but the
s B has been do< and animal~ guide carcas~ to be organi returns whilE size class I a1 transport dis data suggest
From Butchery to Consumption I 221 .uggesting that nn et al. 1980; ave been funcJion of zooariuced damage processing, (2) . Lffected carcass :)th' s tool use, tt is, definition ols (e.g., GaJiS; rt strategies in rted here indiisible kill sites. rould be found ids repeatedly :1dequate shar:e to residence lll site. Though nd IV animals, :md possibly a vidence of the ~ of skulls and : after the kill. md colleagues ;e animals. ~he
tavior thus (1) tl assemblages cessing behav; (i.e., nutrient >iling) play in rganization of on techniques. tone tools for ing facilities or Jort costs and e evolution of ient extraction ~o bone srrlash' in efficiently :hout cooking 1y transported
marrow-rich limb bones (see also Bunn et al. 1988). Further, I expect that marrow and cancellous tissue extraction efforts were maximized via intensive hammerstone breakage. Thus, extreme fragmentation, numerous impact marks, and few complete articular ends may characterize limb bones from Plio-Pleistocene zooarchaeological assemblages. Chewing and gnawing of bone remains the only other nutrient extraction technique available to Plio-Pleistocene hominids. Thus, I would expect hominid tooth marks (which, to my knowledge, cannot be adequately defined at present) to be more frequent in Plio-Pleistocene assemblages than in assemblages created after the invention of fire. Furthermore, because chewing seems to be a common Hadza nutrient extraction technique for small carcass ribs (Figure 12-1e), I might expect early hominids to transport axial parts of smaller carcasses more frequently than those of larger carcasses. As fire began to be utilized for roasting, I expect additional transport of axial parts and less nutrient-rich limb bones because roasting facilitates removal of remaining meat scraps. More efficient nutrient extraction should be evidenced by increased modification of axial bones resulting from both hearth-sizing and consumption practices. Development of roasting pits might promote additional body part transport and processing of axial parts. With the adv7ht of bone-boiling technology, all elements, and especially, perhaps, axial parts, should show a tendency to be transported and intensively processed as is the case with the Hadza. That pattern would be mitigated by availability of firewood and other processing goals such as storage and trade. Data from Olduvai Gorge, Tanzania, and Ko.obi Fora, Kenya, appear to meet the predictions for Plio-Pleistocene sites. Bunn (1986; Bunn and Kroll 1986, 1988) has demonstrated preferential transport of limb bones to PlioPleistocene sites like FLK Zinjanthropus and FxJjSO (but see Blumenschine and Marean, this volume). Furthermore, recent preliminary analysis of limb bone fracture patterns in the FLK Zinjanthropus assemblage reveals that about 48% of the identifiable limb bones were intensively hammerstone broken (Oliver 1991), suggesting that nutrient extraction was being maximized with the available technology. Exactly how the introduction of fire and the resultant increase in nutrient extraction capabilities and subsequent boiling technologies may have altered transport and processing strategies is unknown at this time, but the Hadza example suggests that it may have had a significant effect.
Summary Bone breakage by the Hadza from butchery through consumption has been documented according to the site of breakage, breakage technique, and animal size class. Carcass size and later cooking options were.shown to guide carcass processing decisions. Further, Hadza carcass processing appears to be organized around an optimization principle of maximizing nutrient returns while minimizing transport and field processing costs. Variciliility (i.e., size class I and II animals and wildebeest) in transport decisions not related to transport distances may be related to considerations of sharing and taste. The data suggest that a great deal of behavioral information (certainly more than
222
If. S. Oliver
Identification of the modifying ag~nt as human or nonhuman) may be derived from more detailed examination of breakage damages. In general, the data suggest that because animal size play~ such a major role in Hadza decisions regarding bone breakage a large part of archaeological variability in breakage patterns may also be due to differences in body size. Finally, the importance of cooking in structuring initial butchery and transport decisions suggests that prehistoric innovations in nutrient extraction technologies (e.g, fire, roasting pits, stone boiling, and ceramic boiling vessels) may have driven the evolution (){ carcass transport and processing strategies. As su_ch, cooking options need to be incorporated into models of carcass transport and processing strategies a.nd explanations of bone frequency and modification patterns in zooarchae()logical assemblages.
Acknowledgments I thank J. Hudson and the Center for ~rchaeological Investigations, Southern Illinois University at Carbondale, for organizing a truly enjoyable; intense conference. I thank H. T. Bunn for inviting me to participate in his Hadza ethnoarchaeology project and for granting permission to publish some data before the overview of our 1986 and 1988 work is published. I thank the Tanzania Commission for Science and Technology for permission to conduct this research. I thank Dr. R. B. McMillan, Director of the Illinois State Museum, for administrative support and research space at the Research and Collections Center. Finally, I thank the Hadza without whose warmth and patience none of it would have been possible. Expedition support was provided by grants to H. T. Bunn from University of Wisconsin-Madison; data analysis was supported by dissertation improvement grants from the WennerGren Foundation for Anthropological Research and the National,'Science Foundation. This paper has benefited from discussions with H. T. Bujm, B. W. Styles, M.D. Wiant, D. Gifford-Gonzalez, and J. Hudson and from comments by R. L. Lyman and the anonymous reviewers. Any errors, omissions, or unfounded conclusions that remain are my own. This is contribution No. 89 of the Illinois State Museum Archaeological and Quaternary Studies Program.
Notes 1. The Hadza moved into agricultural settlements in 1989 but have since returned to the ''bush" and have resumed hunting and gathering. 2. Support for the 1988 fieldwork was provided by the Graduate School of the University of Wisconsin-Madison. 3. Other project members (H. T. Bunn, C. Gurney, and C. O'Brien) observed procurement of additional carcasses in 1988, which is the subject of a future report of carcass transport and bone breakage ,by the Hadza. A preliminary report of the transport patterns for 110 carcasses observed by project members in 1986 and 1988 was presented by Bunn and Bartram (1990). 4. In 1988 groups of Isanzu men came into the Hadza area to hunt, setting up at least one carcass processing station. They hunted for several weeks and processed
carcasses and dz Hadza visits to · taring" or const:J 5. The Hadza ed as people frOJ 6. In my limit that of some oth
Re: Bartram, L. E., J1 1993 An Et Refuse. U University Behrensmeyer, 1 1989 Nonh1 Modificatior Study ofth Bellomo, R. V. 1992 Early Koobi Fon Society of j Binford, L. R. 1978 Nunan 1981 Bones: 1983 In Pur: 1987 Resem and Theory demic Pres: Blurnenschine, F 1986 Carca: Scavengi!lg 1988 An Ex on Archaec 502. 1989 ALan
Opportunit Blurnenschine, F 1988 Percm haviour. Nt Bonnichsen, R. 1973 Millie' 291. 1979 Pleistc
Survej; of Ct · Brain, C. K., and 1988 Evide1 336:464-46(
Bunn,H. T. 1981 Archa
From Butchery to Consumption I 223 tay be derived Lerat the data .dza decisions ty in breakage 1e importance ; suggests that . fire, roasting Lthe evolution ; options need ;ing str'ategies in zooarchae-
cal Investigaa truly enjoyparticipate in on to publish ; published. I permission to e Illinois State Research and ~warmth and port was provfadison; data :1 the Wennerional Science r. Bunn, B. W. )m comments .ssions, or untion No. 89 of s Program.
since. returned ~
School of the
observed pro.Iture report of · report of the 1986 and 1988 t, setting up at and processed
carcasses and dried meat at a processing station near a hunting blind. The purpose of Hadza visits to Tatoga camps was to obtain maize or animal food by working e'doctoring'' or constructing huts and animal enclosures) or begging. 5. The Hadza surmised that we were leaving in a few days, and camp size increased as people from other nearby camps began to arrive. 6. In my limited experience, I also found zebra meat sweeter and less tough than that of some other animals including warthog, buffalo, and impala .
References Bartram, L. E., Jr. 1993 An Ethnoarchaeological Analysis of Kua San (Botswana) Bone Food Refuse. Unpublished Ph.D. dissertation, Department of Anthropology, University of Wisconsin-Madison, in preparation. Behrensmeyer, A. K:, K. D. Gordon, and G. T. Yanagi 1989 Nonhuman Bone Modification in Miocene Fossils from Pakistan. In Bone Modification, edited by R. Bonnichsen and M. Sorg, pp. 99-120. Center for the Study of the First Americans, Orono, Maine. Bellomo, R. V. 1992 ~arly Pleistocene Evidence of Hominid-Controlled Fire at FxJj20 Main, Koo~i Fora, Kenya. Paper presented at the 11th Biennial Conference of the Society of Africanist Archaeologists, University of California at Los Angeles. Binford, L. R. 1978 Nunamuit Ethnoarchaeology. Academic Press, New York. 1981 Bones: Ancient Men and Modern Myths. Academic Press, New York. 1983 In Pursuit of the Past. Thames & Hudson, London. 1987 Researching Ambiguity:· Frames of Reference and Site Structure. In Method and Theory for Activity Area Research, edited by Susan Kent, pp. 449-:-512. Academic Press, New York. Blumenschine, R. J. 1986 Carcass Consumption Sequences and the Archaeological Distinction of Scavenging and Hunting. Journal of Human Evolution 15:639-659. 1988 An Experimental Model of the Timing of Hominid and Carnivore Influence on Archaeological Bon_e Assemblages. Journal of Archaeological Science 15:483502. 1989 A Landscape Taphonomic Model of the Scale of Prehistoric Scavenging Opportunities. Journalof Human Evolution 18:345-371. Blumenschine, R. J., and M. M. Selvaggio 1988 Percussion Marks on Bone Surfaces as a New Diagnostic of Hominid Behaviour. Nature 333:763-765. Bonnichsen, R. 1973 Millie's Camp: An Experiment in Archaeology~ World Archaeology 4(3):277291. . 1979 Pleistocene Bone Technology in the Berengian Refugium. Archaeological Survey of Canada Paper No. 89. National Museum of Canada, Ottawa. Brain, C. K., and A. Sillen 1988 Evidence from the Swartkrans Cave for the Earliest Use of Fire. Nature 336:464-466. Bunn,H. T. 1981 Archaeological Evidence for Meat-eating by Pli?-Pleistocene Hominids
224
I f. S. Oliver
from Koobi F.ora and Olduvai Gorge. Nature 291:574-:577. 1982 Meat..,eating and Human Evolution: Studies on the Diet and Subsistence Patterns of Plio-Pleistocene Hominids in East Africa. Unpublished Ph.D. dissertation, Department of Anthropology, University of California, Berkeley. 1983 Evidence on the Diet and Subsistence Patterns of Plio-Pleistocene Hominids and Koobi Fora, Kenya and Olduvai Gorge, Tanzania. In Animals and Archaeology: Hunters and Their Prey, edited by J. Clutton-Brock and C. Grigson, pp. 21-30. BAR International Series 163. Oxford. 1986 · Patterns of Skeletal Representation and Hominid Subsistence Activities at Olduvai Gorge, Tanzania and Koobi Fora, Kenya. Journal of Human Evolution 15:673-690. 1989 Diagnosing Plio-Pleistocene Hominid Activity with Bone Fracture Evidence. In Bone Modification, edited by R. Bonnichsen and M. Sorg, pp. 299-316. Center for the Study of the First Americans, Orono, Maine. Bunn, H. T., and L. Bartram 1990 Carcass Processing and Bone Ass·emblage Formation by the Hadza HunterGatherers in Tanzania. Paper presented at the Sixth International Conference of the International Council for Archaeozoology, Wa~hington, D.C. Bunn, H. T., L. E. Bartram, and E. M. Kroll 1988 Variability in Bone Assemblage Formation from Hadza Hunting, Scavenging and Carcass Processing. Journal of Anthropological Archaeology 7:412-457. Bunn, H. T., J. W. K. Harris, G. Isaac, Z. Kaufulu, E. Kroll, K. Schick, N. Toth, and A. K. Behrensmeyer 1980 FxJj50: An Early Pleistocene Site in Northern Kenya. World Archaeology 12: 109-136. Bunn, H. T., and E. M. Kroll 1986 Systematic Butchery by Plio/Pleistocene Hominids at Olduvai Gorge, Tanzania. Current Anthropology 29:431-452. 1988 Fact and Fiction about the FLK Zinjanthropus Floor. Current Anthropology 29:135-149. . Fiorillo, A. R. 1989 An Experimental Study of Trampling: Implications for the Foss~{ Record. In Bone Modification, edited by R. Bonnichsen and M. Sorg, pp. 61-71. Center for the Study of the First Americans, Orono, Maine. Gifford, D. P. 1980 Taphonomy and Paleoecology: A Critical Review of Archaeology's Sister Disciplines. In Advances in Archaeological Method and Theory, val. 4, edited by M. B. Schiffer, pp. 365-438. Academic Press, New York. Gifford-Gonzalez, D. 1989 Ethnographic Analogues for Interpreting 1v1odified Bones: Some Cases from East Africa. In Bone Modification, edited by R. Bonnichsen and M. Sorg, pp. 179-246. Center for the Study of the First Americans, Orono, Maine. 1991 Bones Are Not Enough: Analogues, Knowledge, and Interpretative Strategies in Zooarchaeology. Journal of Anthropological Archaeology 10:255-254. Gould, R. 1967 Notes on the Hunting, Butchering, and Sharing Game among the Ngatatjara and Their Neighbors ,in the West Australian Desert. Kroeber Anthropological Society Papers 36:41-66. Gowlett, J. A.J., J. W. K. Harris, D. Walton, and B. A. Wood 1981 Early Archaeological Sites, Hominid Remains and Traces of Fire from Chesowanja, Kenya. Nature 294:125-129.
Haynes, G. 1983 Freq Modem~
Hill, A. 1989 Bon edite9- b) FirstAmt Johnso:q~ E. 1985 Cur:
Method a; ·NewYor Jones, P. 1980 Exp Paleo lith' Lyman,R. L. 1987 ArcJ
Advances 249-337. Marean, C. W 1991 Imp Abundar Marshall, F., a 1991 Mea Part Rep 18:149-H Metcalfe, D. 1989 A G Field Pro Americm Metcalfe, D., a 1992 A :M Transpor Morlan, R. E. 1980 Tap: Yukon T
Survey of O'Connell, J. I 1988a Had Implicati 1988b Had
Current I 1989 Dist Camps: I
pretation 61-76. Pl1 1990 Rea:
Journal of 1992 Pan HadzaK O'Connell, J. I 1989 Ana
From Butchery to Consumption I 225 ;istence Patterns ). dissertation, stocene Homi-
rn Animals and nd C. Grigson, tee Activities at
uman Evolution ~ Fracture Evi·g, pp. 299-316.
Hadza Hunter1 Conference of nting, Scaveng7:412-457. 'J. Toth, and A.
Archaeology 12: )lduvai Gorge,
nt Anthropology ossil Record. In .. Center for the teology's Sister 4, edited by M. -s: Some Cases nd M. Sorg, pp. :1e. pretative Strat255-254. :1e among the
roeber Anthropo~s
of Fire from
Haynes, G. 1983 Frequencies of Spiral and Green-Bone Fractures on Ungulate Limb Bones in Modern Surface Assemblages. American Antiquity 48:102-114. Hill, A. 1989 Bone Modification by Modern Spotted Hyaenas. In Bone Modification, edite9- by R. Bonnichsen and M. Sorg, pp, 169-178. Center for the Study of the First Americans, Orono, Maine. Johnson, E. 1985 Current Developments in Bone Technology. In Advances in Archaeological ·Method and Theory, vol. 5, edited by M. Schiffer, pp. 157-235. Academic Press, New York Jones, P. 1980 Experimental Butchery with Modern Stone Tools and Its Relevance for Paleolithic Archaeology. World Archaeology 12:153-165. Lyman, R. L. 1987 Archaeofaunas and Butchery Studies: A Taphonomic Perspective. In Advances in Archaeological Method and Theory, vol. 10, edited by M. B. Schiffer, pp. 249-337. Academic Press, New York Marean, C. W., and L. M. Spencer 1991 Impact of Carnivore Ravaging on Zooarchaeological Measures of Element Abundance. American Antiquity 56:645-658. Marshall, f., and T. Pilgram 1991 Meat Versus Within-Bone Nutrients: Another Look at the Meaning of Body Part Representation in Archaeological Sites. Journal of Archaeological Science 18:149-163. Metcalfe, D. 1989 A General Cost/Benefit Model of the Trade-off Between Transport and Field Processing. Paper presented at the 54th Annual Meeting of the Society for American Archaeology, Atlanta. Metcalfe, D., and K. R. Barlow 1992 A Model of Exploring the Optimal Trade-off Betwee:n Field Processing and Transport. American Anthropologist 94:340-356. Morlan, R. E. 1980 Taphonomy and Archaeology in the Upper Pleistocene of the Northern Yukon Territory: A Glimpse of the Peopling of the New World. Archaeological Survey of Canada Paper No. 94. National Museum of Canada, Ottawa. O'Connell, J. F., K. Hawkes, and N. Blurton Jones 1988a Badza Hunting, Butchering, and Bone Transport and Their Archaeological Implications . Journal of Anthropological Research 44(2):113-161. 1988b Hadza Scavenging: Implications for Plio/Pleistocene Hominid Subsistence. Current Anthropology 29(2):356-363. 1989 Distribution of Refuse Producing Activities at Hadza Residence Base Camps: Implications fo_r Analyses of Archaeological Site Structure. In The Interpretation of Archaeological Spatial Patterns, edited by E. Kroll and T. D. Price, pp. 61-76. Plenum Press, New York 1990 Reanalysis of Large Mammal Body Part Transport among the Hadza.
Journal of Archaeological Science 17:301-316. 1992 Patterns 1n the Distribution, Site Structure and Assemblage Composition of Hadza Kill-Butchering Sites. Journal of Archaeological Science 19:319-314. O'Connell, J. F., and B. Marshall 1989 Analysis of Kangaroo Body Part Transport among. the Alyawara of Central
226
I ]. S. Oliver
Australia. Jo~rnal of Archaeological Science 16:393-405 .. Oliver, J. S. 1984 Analogues and Site Context: Bone Damages from Shield Trap Cave (24CB91), Carbon County, Montana, U.S.A. Paper presented at the First futernational Conference on Bone Modification, Carson City, Nev. 1986 The Taphonomy and Paleoecology of Shield Trap Cave (24CB91), Carbon County, Montana. Unpublished Master's thesis, Institute for Quaternary Studies, University of Maine at Orono. 1989 Analogues and Site Context: Bone Damages from Shield Trap Cave (24CB91), Carbon County, Montana, U.S.A. In Bane Modification, edited by R. Bonnichsen and M. Sorg, pp. 73-98. Center for the Study of the First Americans, Orono, Maine. 1991 Damage Context Analysis: A Preliminary Analysis of the Amboseli Hyaena Den, FLK NN Level 2 and FLK Zinjanthropus Faunal Asse11.1blages. Paper presented at the 57th Annual Meeting of the Society for American Archaeology, New Orleans. Perkins, D., and P. Daly 1968 A Hunter's Village in Neolithic Turkey. Scient.ific American 219(11):97-106. Potts, R. 1988 Early Hominid Activities at Olduvai. Aldine, New York. Potts, R., and P. Shipman 1981 Cutmarks Made by Stone Tools from Olduvai Gorge, Tanzania. Nature 291:577-588. Saunders, J. J. 1990 Immanence, Configuration and the Discovery of America's Past. In Megafauna and Man: Discovery of America's Heartland, edited by L. D. Agenbroad, J. I. Mead, and L. W. Nelson, pp. 136-143. The Mammoth Site of Hot Springs, South Dakota, Inc., Scientific Papers 1. Hot Springs, S. D. Shipman, P. 1981 Applications of Scanning Electron Microscopy to Taphonomic Problems. In The Research Potential of Anthropological Museum Collections, edited by J- B. Griffin and N. Rothschild, pp. 357-385. Annals of the New York Academy~ of Sciences 276. Shipman, P., G. Foster, and M. Schoeninger 1984 Burnt Bones and Teeth: An Experimental Study of Color, Morphology Crystal Structure and Shrinkage. Journal of Archaeological Science 11:307-325. Shipman P., and J. Rose 1983 Early Hominid Hunting Butchering and Carcass-processing Behaviors: Approaches to the Fossil Record. Journal of Anthropological Archaeology 2:67-98. Toth, N. 1985 The Oldowan Reassessed: A Close Look at Early Stone Artifacts. Journal of Archaeological Science 12:101-120. Vincent, A. S. 1985 Plant Foods in Savanna Environments: A Preliminary Report of Tubers Eaten by the Hadza of Northern Tanzania. World Archaeology 17:132-148. Woodburn, J. C. 1964 The Social Organization of th? Hadza of North Tanganyika. Unpublished Ph.D. dissertation, Cambridge University. 1968a An Introduction to Hadza Ecology. fu Man the Hunter, edited by R. B. Lee and I. DeVore, pp. 49-55. Aldine, Chicago. 1968b Stability and Flexibility in Hadza Residential Groupings. In Man the Hunter,
edited b) 1970 Hur. Museum 1972 Eco: among :r.
Man, Set Dimbleb' Yellen, J. E. · 1977 Arc! Academi Zierhut, ·N. W 1~67' Bon 1:33-36.
From Butchery to Consumption I 227 edited by R. B. Lee and I. DeVore, pp. 103-110. Aldine, Chicago. Hunters and Gatherers: Material Culture of the Nomadic Hadza. British Museum, London. 1972 Ecology, Nomadic Movement and the Composition of the Local Group a111ong Hunters and Gatherers: An East African Example and Its Implications. In Man, Settlement, and Urbanism, edited. by P. Ucko, R. Tringham, and C. Dimbleby, pp. 193-206. Duckworth, London. Yellen, J. E. 1977 Archaeological Approaches to the Present: Models for Reconstructing the Past. ·Academic Press, New York. Zierhut, N. W. 1967 Bone Breaking Activities of the Calling Lake Cree. Alberta Anthropologist 1970
eld Trap Cave at the First
~d {.
, Carbon County, rnary Studies, eld Trap Cave n, edited by R. :4 6 7 3 3 3 4 3 3
>2 >7 1 1
>2 1
4
)Ur study area ~en neighbors. nwonek Okiek 1 with fathers, ~athers, sisters,
Jgs and spears n to be differ1 animals (less ~n households. 1ted with dogs mt the capture iistribution of Lt patterns and :en made up of t within settlenee of hunting lation, there is
great variability in faunal accumulation between households over time. This can be expressed in terms of species composition but is chiefly seen in variability in body part composition. In a community of nine households living on adjacent ridges in an approximately 10 km2 area of the high forest, one household (Household 1), that of the most successful hunter, accumulated inore bones than any other (Figure 13-1). Over a six-month period the assemblage accumulated by the household was made up of a wider range of body parts and more skins with carpals and tarsals, hindlimbs, and metapodials than that of any other household. There were also large numbers of forelimbs. By contrast Household 4 killed only one animal and accumulated only one hindlimb, one forelimb, one set of vertebrae, one head, and one set of ribs. Ho'wever, Household 9 also killed only one animal but, because the householder is very well liked in the community, accumulated not one but four forelimbs. Households 5, 6, 7, and 8 did not kill any animals and thus accumulated few bones and some restricted body parts, primarily ribs and portions offorelimb elements. Thus, over time the successful Okiek hunters have different access to meat than do less successful hunters and, as in the case of Household 1, accumulate many bones, a wide range of body parts, and a high proportion of high utility elementr Large numbers of bones are accumulated because successful hunters are typically involved in many butchery and sharing events and also because they retain a large share of the carcass. In time, because of the large share of and variation in portions of skeletons available (sometimes a limb is eaten by dogs or partially spoiled in a trap) and the varying portions of meat received from other people's kills, successful hunters will accumulate a great variety of skeletal elements. By contrast less successful hunters, such as Household 5, 6, 7, or 8, accumulate fewer bones and fewer body parts, typically of relatively low utility. In our examples the assemblages were composed mostly of ribs and sometimes forelimb or hindlimb elements. Up to now, most variability of this kind has largely been interpreted by archaeologists in terms of the logistics of carcass transport. It has been argued that bones associated with highly nutritious carcass parts that are easy to carry will be transported· more often than other bones (Bunn and Kroll 1986; Klein 1976; Perkins and Daly 1968). High proportions of axial elements have been taken to indicate kill sites, and high proportions of. limb elements have been taken to indicate camps. Lately those assumptions have been tested in ethnoarchaeological contexts and modified. Taking butchery costs into account, it has been suggested that vertebrae, scapulae, pelves, and upper limb bones are more likely to be transported from kill sites to camps than other bones (O'Connell et al. 1990, but see Bunn et al. 1988). By these criteria, and the number of bones present, the body part profiles offaunal assemblages accumulated by successful hunters, such as those from Household 1, resemble those interpreted by archaeologists as base camps, whereas those accumulated by less successful hunters, such as Household 5, 6, 7, or 8,look more like body part profiles that might be interpreted as kill sites, special purpose sites, or scavenged faunal assemblages. Fortunately, other features of site structure could be used to distinguish
I
236 F. Marshall Hou5ehold 5 (0 Kills)
Household 1 (5 Kills) Cranium Vertebrae
Ribs~ Forelimb Hindlimb
+---r--...---.....-.....--l Number
Number
Household 6 (0 Kills)
Household 2 (2 Kills) Cranium
Cranium t
!!I
a..
~
IJ)
t::
Vertebrae
le I am very
~
Study Mono-
tee Behaviour ;tence: Current
I 243
244j F. Marshall and Their Neighbors in the West Australian Desert. .The Kroeber Anthropological Society Papers 36:41-66. 1981 Comparative Ecology of Food-Sharing in Australia and Northwest California. In Omnivorous Primates; edited by Robert S. 0. Harding .and Geza Teleki, pp. 422-454. Columbia University Press, New York. Grayson, D. K. 1989 Bone Transport, Bone Destruction, and Reverse Utility Curves. Journal of Archaeological Science 16:643-652. Hayden,,B. 1990 Nimrods, Piscators, Pluckers, and Planters: The Emergence of Food Production. Journal of Anthropological Archaeology 9:31-69. Hewlett, B. S. 1988 Sexual Selection and Paternal Investment among Aka Pygmies. In Human Reproductive Behavior: A Darwinian Perspective, edited by Laura Betzig, Monique Borgerhoff Mulder, and Paul Turke, pp. 263-276. Cambridge University Press, . Cambridge. Hill, K., and A. Hurtaldo 1989 Hunter-Gatherers of the New World. American Scientist 77:5:437-443. Hill, K., and H. Kaplan 1988 Tradeoffs in Male and Female Reproductive Strategies among the Ache: Part 1. In Human Reproductive Behavior: A Dmwinian Perspective, edited by Laura Betzig, Monique Borgerhoff Mulder, and Paul Turke, pp. 277-289. Cambridge University Press, Cambridge. Hudson,J. 1990 Identifying Food Sharing Archaeologically: An Ethnoarchaeological Study among the Aka. Paper presented at the Sixth International Conference on Hunting and Gathering Societies, Fairbanks, Alaska. Huntingford, G. W. B. 1942 The Social Organization of the Dorobo. Anthropos 46:1-48. 1954 The Political Organization of the Dorobo. Anthropos 49:123-148. 1955 The Economic Life of the Dorobo. Anthropos 50:602-634 / Ingold, T. 1980 Hunters, Pastoralists, and Ranchers. Cambridge University Press, Cambridge. Isaac, G. L. 1977a The Food Sharing Behavior of Protohuman Hominids. Scientific American 238:90-108. 1977b Food Sharing and Human Evolution: Archaeological Evidence from the Plio-Pleistocene of East Africa. Journal of Anthropological Research 34:311-325. Kaplan, H., and K. Hill 1985 Food Sharing among Ache Foragers: Tests of Explanatory Hypotheses. Current Anthropology 26:223-246. · Klein, R.G. 1976 The Mammalian Fauna of the Klasies River Mouth Sites, Southern Cape Province. South Africa. South African Archaeological Bulletin 31:75-98. 1989 Why Does Skeletal Part Representation Differ Between Smaller and Larger Bovids at Klasies River Mouth and Other Archaeological Sites? Journal of Archaeological Science 16:363-381. · Kratz, C. A. 1977 The Liquors of Forest and Garden: Drinking in Okiek Life. Unpublished Master's thesis, Wesleyan University, Middletown, Conn. 1981 Are the Okiek Really Masai? Or Kipsigis? or Kikuyu? Cahiers d'Etudes Africaines. 79:355-368.
Lee, R. 1979 The Universil Lyman,R. L. ,1984 Bon Anthropo 1985 Bon MGUI.Jc Marshall, L. 1976 Sha: Kalahari, R. B. Lee Marshall, F., c: 1991 Mec Part Rer 18:149-11 Meillasoux, C 1973 On African E. Oxford. O'Connell, J. : 1990 Rea nal of An Perkins, D., aJ 1968 A :f. Potts, R. 1984 Hm 1988 Earl Potts, R., and 1981 Cut Nature 2~ Shipman, P 1983 Ear: ing? In 1 31-50. B. 1986a Sea· Tests. Ar 1986b Stw Evolutior. Speth, J.D. 1990 Sea: Foragin~
Steward, J. 1938 Basi Ethnolo~
Washburn, S. 1961 Soc edited b' Wiessner, P. 1982a Bey of Hunte
Food Sharing 1245 Anthropological nd Northwest ding and Geza
.rves. Journal of ·
:e of Food Pro-
nies. In Human etzig/ Monique niversity Press/
437-443. 10ng the Ache: ~dited by Laura 89. Cambridge
~ological
Study renee on Hunt-
48.
ss, Cambridge. ~ntific
American
lence from the :4:311-325. y Hypotheses. Southern Cape 98. ,er and Larger
urnal of Archae. Unpublished ~ahiers
d' Etudes
Lee, R. 1979 The !Kung San: Men, Women, and Work in a Foraging Society. Cambridge University Press/ Cambridge. Lyman, R. L. 1984 Bone Density and Differential Survivorship of Fossil Classes. Journal of Anthropological Archaeology 3:259-299. 1985 Bone Frequencies: Differential Transport/ In Situ Destruction, and the MGUI. Journal of Archaeological Science 12:221-236. Marshall/ L. 1976 Sharing, Talking and Giving: Relief of Social Tensions among the !Kung. In Kalahari Hunter-Gatherers: Studies of the !I\ung San and Their Neighbors, edited by R. B. Lee and I. DeVore/ pp. 349-371. Harvard University Press, Cambridge. , Marshall, F., and T. Pilgram 1991 Meat Versus Within-Bone Nutrients: Another Look at the Meaning of Body Part Representation in Archaeological Sites. Journal of Archaeological Science 18:149-163. Meillasoux, C. 1973 On the Mode of Production of the Hunting Band. In French Perspectives on African Studies/ edited by P. Alexandre/ pp. 187-204. Oxford University Press, Oxford. O'Connel1/ J. F., Hawkes, K./ and N. Blurton Jones 1990 ;Reanalysis of Large Mammal Body Part Transport among the Hadza. Journal of Archaeological Science 17:301-316; Perkins, D./ and P. Daly 1968 A Hunters Village in Neolithic Turkey. Scientific American 219:97-106. Potts, R. 1984 Home Bases and Early Hominids. American Scientist 72:338-347. 1988 Early Hominid Activities at Olduvai. Aldine, New York. Potts/ R., and P. Shipman 1981 Cutmarks Made by Stone Tools on Bones from Olduvai Gorge, Tanzania. Nature 291:574-577. Shipman/ P 1983 Early Hominid Lifestyle: Hunting and Gathering or Foraging and Scavenging? In Animals and Archaeology, edited by J. Clutton-Brock and C. Grigsori, pp. 31-50. BAR International Series 163. Oxford. 1986a Scavenging or Hunting in Early Hominids: Theoretical Framework and Tests. American Anthropologist 88:27-43. 1986b Studies of Hominid-Faunal Interactions at Olduvai Gorge. Journal of Human Evolution 15:691-706. Speth, J.D. 1990 Seasonality, Resource Stress and Food Sharing in So-Called "Egalitarian" Foraging Societies. Journal of Anthropological Archaeology 9:148-188. Steward, J. · 1938 Basin-Plateau Aboriginal Groups. Smithsonian Institution. Bureau of American Ethnology Bulletin 120. Washburn, S. L., and I. DeVore 1961 Social ~ehavior of Baboons and Early Man. In Social Life of Early Man, edited by S. Washburn/ pp. 91-105 Aldine, Chicago . Wiessner, P. 1982a Beyond Willow Smoke and Dogs' Tails: A Comment On Binford's Analysis of Hunter-Gathe.J;"er Settlement Systems. American Antiquity 47:171-178.
246jF. Marshall 1982b Risk, Reciprocity and Social Influence on !Kung San Economies. In Politics and History in Band Societies, edited by E. R. Leacock and R. B. Lee, pp. 61-84. · Cambridge University Press, London. Winterhalder, B. · 1986 Diet Choice, Risk and Food Sharing in a Stochastic Environment. Journal of Anthropological Archaeology 5:369-392. Woodburn, J. 1968 An Introduction to Hadza Ecology. In Man the Hunter, edited by R. B. Lee and I. DeVore, pp. 49-55. Aldine, New York. 1980 Hunter-Gatherers Today and Reconstruction of the Past. In Soviet and Western Anthropology, edited by Ernest Gellner, pp. 95-117. Columbia University Press, New York. · Yellen, J. E. 1977a Archaeological Approaches to the Present: Models for Reconstructing the Past. Academic Press, New York. 1977b Cultural Patterning in Faunal Remains: Evidence from the !Kung Bushmen. In Experimental Archaeology, edited by D. Ingersoll, J. Yellen, and W. Macdonald, pp. 271-331. Columbia University Press, New Yorlw-poor bones ·e of bones for ;e to boil. The 0 usee" this in
al evidence at te presence of :tat shape the nd situational, that archaeol~-farmer inter) differentiate to distinguish ributes of the ong indicators 3e sites. Bone fying the pres-
ence of Efe and Lese. The absence of distinguishing attributes reflects the use of similar processing techniques by both groups and unstructured discard behavior. The subsistence-settlement strategies of the Efe and Lese are key to elevating the archaeological utility of this Efe-Lese case from the particularistic to the more general. The contrasting subsistence strategies of the Efe and Lese regarding the accumulation and storage of surplus food represents an important determinant. The Lese are strongly concerned with acquiring large quantities of surplus meat during elephant butchery and processing, whereas the Efe have no interest in putting up large stores of meat. The differing attitudes toward surplus meat are understandable in the broader context of Efe and Lese subsistence adaptations. The Lese, whose subsistence derives primarily from domestic plants cultivated on a seasonal cycle, incorporate a strategy of delayed food returns and storage of food. The day-to-day acquisition and consumption of food by the Efe, on the other hand, is explainable with reference to the nature of wild food availability in tropical forest environments (see Binford 1980; Kelly 1983). We can expect variation as a function of increasing dependence on food storage by hunter-gatherer groups. Collectors might be expected to behave more like/ the Lese than the Efe. Nunamiut Eskimo subsistence is geared toward killing large numbers of caribou during brief periods of availability and putting up stores for consumption at other times (Binford 1978). Under those conditions, the material record of large mammal butchery and processing would differ from that produc~ by foragers such as the Efe and might resemble more closely that of farmers who are concerned with processing food in bulk for storage. The duration of site occupation stands out as an important determinant of certain site attributes. The casual spatial organization of the specialized butchery /processing sites, particularly the location and orientation of dwellings, departs from the structure and regularity at Efe residential camps. Site maintenance (cleanup, keeping debris out of the way) is desultory at the specialized, short-term sites when compared with the regularity of site maintenance at Efe residential camps and Lese villages.· The immense size of elephant carcasses dictates the very existence of the specialized processing sites. Smaller mammals can be transported whole or in segments from kill site to residential site, but intensive processing of elephants is more easily accomplished by moving the people to the carcass. Analogous conditions can arise if hunters slay multiple individuals of, say, bison or caribou. Thus, the Ituri sites are relevant for other settings. The environment conditions the nature of site activities. Drying is the only viable means of preserving meat in a tropical climate, whereas cold storage is an alternative option in suitable environments. This study has been concerned with attempting to link expectations for material patterning with the underlying determinants of the behavior that produced the material record. It has devoted less attention to developing the means to operationalize archaeological.criteria for identifying the patterns in the ground. For example, the spatial distribution of dwellings at these special-
260
I J. w. Fisher, Jr.
ized sites is an important attrib.ute, but how do we reconstruct the location of dwellings archaeologically? Reconstructing the location of dwellings at Efe residential campsites has been ad9.ressed by Fisher and Strickland (1991). · Further work of this nature is required.. The vagaries of preservation of material evidence in, the archaeological record presents a related issue. Almost all classes of evidence at the processing sites are organic and vulnerable to decay and obliteration. Preservation is contingent on regional, local, and microenvironmental conditions. The reddened, oxidized·earth left by past fires probably has the greatest potential for detection. Elaboration on the kinds of evidence that might or might not be preserved archaeologically will not be pursued here. Preservation obviously is crucial to archaeology, but I am more concerned in this presentation with learning to understand the material record to monitor human behavior. The constellation of multiple lines of evidence provides a more powerful indicator . of forager-farmer interactions than does any individual line of evidence in isolation.
Acknowledgments Helen Strickland's participation in the fieldwork reported here was vital to its success, and I am very grateful for her contributions. This fieldwork was funded by a grant from the National Science Foundation to Irven DeVore and Peter Ellison. I am very grateful to the Efe and Lese, among whom we lived and worked, for their cooperation. I thank Jean Hudson for inviting me to participate in the From Bones to Behavior conference and for helpful comments on an earlier draft of this paper.
References Bailey, R. C., G. Head, M. Jenike, B. Owen, R. Rechtman, and E. Zechenter 1989 Hunting and Gathering in Tropical Rain Forest: Is It Possible? American Anthropologist 91:59-82. Bailey, R. C., and I. DeVore 1989 Research on the Efe and Lese Populations of the Ituri Forest, Zaire. American Journal of Physical Anthropology 78:459-471. Binford, L. R. 1978 Nunamiut Ethnoarchaeology. Academic Press, New York. 1980 Willow Smoke and Dogs' Tails: Hunter-Gatherer Settlement Systems and Archaeological Site Formation. American Antiquity 45:4-20. 1987 Researching Ambiguity: Frames of Reference and Site Structure. In.Method and Theory for Activity Area Research: An Ethnoarchaeological Approach, edited by S. Kent, pp. 449-512. Columbia University Press, New York. Brooks, A. S., D. E. Gelburd, and J. E. Yellen 1984 Food Production and Cuiture Change among the !Kung San: Implications for Prehistoric Research. In From Hunters to Farmers: The Causes and Consequences of Food Production in Africa, edited by J.D. Clark and S. A. Brandt, pp. 293-310. University of California Press, Berkeley.
Bunn, H. T., I 1988 Var ing, and Ewers,J. C. '1954 ThE pretatim Fisher, J. W.,: 1992 Obt Spring~
andJ. D of Color, Fisher, J. W., . 1989 Eth Campsit 1991 Dv.
Ethnoarc Case Stu national Gargett, R., a 1991 SitE for Arcl byE.M. Hart, T. B., m 1986 Thi Forests: Headland, T. 1989 Hu
Current Hitchcock, R 1987 Sec Residen
archaeol1 Press, I\ Kelly, R. L. 1983 H1: 39:277-~
Milisauskas, 1978 Eu O'Connell, J. 1987 Al
Antiquit O'Connell, J. 1988 Ha Implica 1991. Di: Camps:
pretati01 pp. 61-: Smith, B. D. 1989 Or
Foragers and Farmers 1261 the location of rellings at Efe :kland (1991). Lrchaeological the processing reservation is lons.· The red:t potential for might not be n obviously is entation with behavior. The erful indicator >f evidence in
)rted here was rhis fieldwork Irven DeVore mg whom we Jr inviting me .d for helpful
nter sible? American
Bunn, H. T., L. E. Bartram, and E. M. Kroll 1988 Variability in Bone Assemblage Formation from Badza Hunting, Scavenging, and Carcass Processing. Journal of Anthropological Archaeology 7:412-457. Ewers, J. C. 1954 The Indian Trade of the Upper Missouri Before Lewis and Clark: An Inter· pretation. Missouri Historical Society Bulletin 10:429-446. Fisher, J. W., Jr. 1992 Observations on the Late Pleistocene Bone Assemblage from the Lamb Spring Site, Colorado. In Ice Age Hunters of the Rockies, edited by D. J. Stanford and J. Day, pp. 51-81. Denver Museum of Natural History and University Press of Colorado, Niwot. Fisher, J. W., Jr., and H. C. Strickland 1989 Ethnoarchaeology among the Ere Pygmies, Zaire: Spatial Organization of Campsites. American Journal of Physical Anthropology 78:473-484. 1991 Dwellings and Fireplaces: Keys to Efe Pygmy Campsite Structure. In
Ethnoarchaeological Approaches to Mobile Campsites: Hunter-Gatherer and Pastoralist Case Studies, edited by C. S. Gamble and W. A. Boismier, pp. 215-236. International Monographs in Prehistory, Ethnoarchaeological Series 1, Ann Arbor. Gargett, R., and B. Hayden 1991 Site Structure, Kinship, and Sharing in Aboriginal Australia: Implications for AFchaeology. In The Interpretation of Archaeological Spatial Patterning, edited by E./M. Kroll and T. D. Price, pp. 11-32. Plenum Press, New York. Hart, T. B., and J. A. Hart 1986 The Ecological Basis of Hunter-Gatherer Subsistence in African Rain Forests: The Mbuti of Eastern Zaire. Human Ecology 14:29-55. Headland, T. N., and L.A. Reid 1989 Hunter-Gatherers and Their Neighbors from Prehistory to the Present .
Current Anthropology 30:43-66. Hitchcock, R. K. 1987 Sedentism and Site Structure: Organizational Changes in Kalahari Basarwa Residential Locations. In Method and Theory for Activity Area Research: An Ethnoarchaeological Approach, edited by S. Kent, pp. 374-423. Columbia University Press, New York. K~lly, R. L. 1983 Hunter-Gatherer Mobility Strategies. Journal of Anthropological Research 39:277-306.
:t, Zaire. Ameri-
nt Systems and ture. In Method zch, edited by S. .n: Implications
rzd Consequences it, pp. 293-310.
Milisauskas, S. 1978 European Prehistory. Academic Press, New York. O'Connell, J. F. 1987 Alyawara Site Structure and Its Archaeological Implications. American
Antiquity 52:74-108. O'Connell, J. F., K. Hawkes, and N. Blurton Jones 1988 Badza Hunting, Butchering, and Bone Transport and Their Archaeological Implications. Journal of Anthropological Research 44:113-161. · 1991 Distribution of Refuse-Producing Activities at Badza Residential Base Camps: Implications for Analyses of Archaeological Site Structure. In The Interpretation of A~chaeological Spatial Patterning, edited by E. M. Kroll and T. D. Price, pp. 61-76. Plenum Press, New York. Smith, B.- D. 1989 Origins of Agriculture in Eastern North America. Science 246:1566-1571.
262 I J. W. Fisher~. Jr. Solway, J. S., and R. B. Lee . 1990 Foragers, Genuine or Spurious? Situating the Kalahari San in History. Current Anthropology 31:109-146. Spielmann, K. A. 1991a Coercion or Cooperation? Plains~Pueblo Interaction in the Protohistoric Period. In Farmers, Hunters, and Colonists: Interaction Between 'the Southwest and the Southern Plains, edited by K. A. Spielmann, pp. 36-50. University of Arizona Pre.ss, Tucson. Spielmann, K. A. (editor) 1991b Farmers, Hunters, and Colonists: Interaction Between the So.uthwest and the Southern Plains. University of Arizona Press, Tuc$on. Wilmsen, E. N., and J. R. Denbow 1990 Paradigmatic History of San-speaking Peoples and Current Attempts at Revision. Current Anthropology 31:489-524. Wood, W.R. 1974 Northern Plains Village Cultures: Internal Stability and External Relationships. Journal of Anthropological Research 30:1-16. Yellen, J. E. 1976 Settlement Patterns of the !Kung: An Archaeological Perspective.· In Kalahari Hunter-Gatherers: Studies of the !Kung San and Their Neighbors, edited by R. B. Lee and I. DeVore, pp. 47-72. Harvard University Press, Cambridge.
/
15.
I B It Fi
archaeologicc 20 years. My biased by m: Indian sites i1 ethnoarchaec Gould (1969, 1977) and Le debates ovez Binford (196: information 1 don "grab-ba rnation throu because, unli the behavion archaeologicc: Actualistic tal research (E if the researcl peoples and ' do, then the r logical reco~c nants of patt record in ord, and experirne the agents an blages. As C< actualistic res ideas about "' to flesh out < From Bones to Be
Faunal Remains, Paper No. 21. ~ reserved. ISBN C
t
in History.
Proto historic thwest and the y of Arizona
1west and the
15.
Discussion: Social Interaction Bonnie W. Styles
Attempts at
Introduction nal Relation-
·spective. In 'rs, edited by ridge.
First, I would like to confess that I have never done any ethnoarchaeological research, but I have been doing zooarchaeological research for 20 years. My interests are in paleoecology and subsistence, and my views are biased by my experiences with faunal remains from prehistoric American Indian sites in the Midwestern United States. I have certainly been inspired by ethnoarchaeological research since the days when I first read the works of Gould (19t9, 1971), Brain (1967, 1969), Isaac (1967), and Yellen (1974, 1976, 1977) and Lee and Devore's (1968) Man the Hunter volume. I had read the debates over the use of ethnographic analogy by archaeologists such as Binford (1967) and had agreed with Yellen (1977:xi) that to get the kind of information needed for archaeological research, archaeologists ,had to abandon "grab-bag analogy from the ethnographic present" and collect the information through their own research efforts. I appreciate that kind of research because, unlike most ethnographers, those ethnoarchaeologists are recording the behaviors and consequences that are of maximal interest to me from an archaeological perspective. Actualistic data, whether it is derived from ethnoarchaeology or experimental research (e.g., Binford 1981), is clearly important to archaeology. However, if the research is to serve archaeology and improve our understanding of past peoples and environments,· which ethnoarchaeologists say they are trying to do, then the research must be conducted with the "perspective of the archaeological record" in mind: that is, with a goal of understanding "the determinants of patterning and various structural properties of the archaeological record in order to. learn about their past" (Binford 1981:32). Ethnoarchaeology and experimental research have contributed a great deal to our assessments of the agents and processes that formed and acted on prehistoric faunal assemblages. As concisely stated by Gifford-Gonzalez in this volume, "we use actualistic research to our fullest advantage when it informs and re-forms our ideas about what might have happened in the past rather than simply 1-J.Sing it to flesh out our preexisting ideas about it." As a zooarchaeologist, I have From Bones to Behavior:. Ethnoarchaeological and Experimental Contributions to the Interpretation of Faunal Remains, edited by Jean Hudson. Center tor Archaeological Investigations, Occasional Paper No. 21. © 1993 by the Board of reserved. ISBN 0-88104-076-2.
Trustee~,
263
Southern Illinois University. All rights ·
2641 B. W. Styles
followed ethnoarchaeological research, although perhaps not as closely as I should have. Howev.er, I can guarantee that I have dohe a lot of reading since Jean Hudson called me to participate in the conference. I will focus my comments on the.papers and major topics for this session, that is, the recognition of ethnic differences, food sharing, and the impacts of food preparation and consumption on the record. I will attempt to synthesize the major points raised in the four papers in Part III and examine their applicability to archaeological research.
Bone 1\tlodification, Cooking Techniques, and Consumption The papers by Diane Gifford-Gonzalez and J. S. Oliver note the role that cooking practices play in the patterning of faunal assemblages. Oliver analyzes breakage patterns associated with butchering, preparation for cooking, and consumption. Based on his study of the Hadza, he notes that the decision whether or not to break a bone for cooking and the tool used for this "pot- or hearth-sizing" appeared to be primarily determined by carcass size and element type. Pot-sizing breakage of axial elements increased with body size. Pot-sizing of limb elements was mitigated by the amount of marrow and cancellous tissue and whether marrow was to beeaten raw or after boiling. He argues that ''behaviors-from initial field butchery, to transport, to cooking-preparation, to consumption-may be integrated into a decisionmaking strategy where nutrient extraction is maximized, while transport, field processing, and social costs are minimized." Oliver argues that the breakage and other bone modifications related to butchery and preparation for cooking and consumption, such as for the pot sizing and marrow extraction that he observed, would have archaeological signatures. The short period of boiling used by the Hadza left no visible damage and did not produce a diagnostic breakage pattern. However, 1roasting, as also observed by Gifford-Gonzalez, left breakage patterns (jagged edges) that differed from those noted for fresh bone. Marrow and cancellous tissue removal also created diagnostic damages. In addition, he argues that detailed analyses of.bone modifications are essential to the interpretation of patterns in the frequencies of bone elements because these frequencies can be the result of a number of different processes (e.g., transport by humans and/ or destruction of low density bones). I agree wholeheartedly with his assertion that more experimental research is needed to identify cooking-related damages because I am interested in the preparation of meat and how food preparation changed through time. Although post-consumption processes, such as disposal, scavenging by dogs, and trampling by humans, may obscure earlier n1odifications related to butchering, cooking, and consumption, my studies could certainly benefit from a more careful, albeit labor-intensive, look at bone breakage and modifications. Systematic coding of modifications and breakage patterns is no easy task but could certainly shed more light on the processing of faunal elements: One productive line of research for Midwestern zooarchaeologists would be to examine changes in food preparation related to the transition
from roastin Along the missing fror culinary enc ture of faun; ern Kenya, dismemberr jagged, trar mately link because th( uncooked bl that butche: about how t be stored m available at products· wi butchering;: Accordin~
reflects an E meat is like] tured raw, i1 the use off bias" within and butcher tions. At m< animals of; numbers, ar ed through: on the time kill sites do the search, interpretati< logical rec01 that considE storage, coo: Gifford-G analytical r ·Cooking pr. evolution c addressed i butchering c ]
1
sharing pla: notes that '
Social Interaction I 265 .s closely as I reading since r this session, he impacts of to synthesize 1e their appli-
liver note the assemblages. reparation for notes that the l used for this y carcass size ed with body fmarrow and after boiling. transport, to u a decision~ansport, field ms related to as for the pot .rchaeological J visible dam~r, roasting, as ~d edges) that :ellous tissue ; that detailed of patterns in 1e the result of ::>r destruction ~on that more 1ages because ition. changed lisposal, scavmodifications Juld certainly breakage and patterns is no ing of faunal trchaeologists the transition
from roasting to stone boiling and eventually to boiling in ceramic vessels. Along these same lines, Gifford-Gonzalez convincingly argues that "what is missing from recent ethnoarchaeological analyses is serious treatment of the culinary end of the processing spectrum and its major influence ori the structure of faunal assemblages." Based on her studies of the Dassanetch of northern Kenya, she found that culinary tactics were crucial to decisions about dismemberment, selective transport, and bone breakage. High frequencies of jagged, transverse fractures on bones from a Dassanetch camp were ultimately linked to roasting; however, the pattern was difficult to explain because the experimental and ethnographic literature had focused on uncooked bones. She indicates that many ethnoarchaeological accounts show that butchery decisions made at kilt sites are "influenced by expectations about how the meat, marrow, and other useful parts of the dead animals will be stored or processed for consumption." Thus, the "processing technology available at the destination site" and the "ultimate form or forms the animal products will take" should be added to the long list of factors affecting field butchering and transport. According to Gifford-Gonzalez, the research emphasis on uncooked bone reflects an early focus on 1nass kill assemblages and mass butchering where meat is likely to be removed prior to cooking and bones are likely to be fractured ra/v, if at all, and a focus on early hominid sites that presumably predate the use of fire. She also attributes the pattern to "unconscious androcentric bias" within the field of anthropology, which favored studies of male hunting and butchery of large animals as central to the understanding of past adaptations. At many Holocene archaeological sites, such as the ones I deal with, animals of all sizes entered the site and were processed in relatively small numbers, and the meat, the bones and the meat, and/ or the bones were cooked through roasting, stone boiling, and/ or boiling in ceramic pots depending on the time period under investigation. Clearly, early hm.ninid sites and mass kill sites do not provide the best analogues for many sites, and emphasis on the search, pursuit, and processing of large game will not lead to holistic interpretations of subsistence practices or of the formation of the archaeological record. Gifford-Gonzalez calls for a product-focused research strategy that considers all the tasks and factors that affect faunal remains, including storage, cooking, and refuse disposal. Gifford-Gonzalez offers more than a cautionary tale because she shows how analytical models can influence ethnographic research and observations. Cooking practices, although clearly important to our understanding of the evolution of human subsistence and nutrition, have been inadequately addressed in many studies because of an emphasis on search, pursuit, and butchering of large animals. Foo~
Sharing
The paper by Fiona Marshall emphasizes the important role food sharing plays in the dispersal of bones within and between sites. Marshall notes that with a few exceptions (e.g., Binford 1984). there have been few
2661 B. W. Styles
discussions "as to what might constitute archaeological evidence for food sharing." Yet Marshall found food sharing to be the the single major factor patterning early stages of the formation of assemblages on Okiek sites in the high altitude rain forests of Kenya. "In 12 out of 15 observed cases" animals were shared between 2 and 7 households separated by several kilometers." Secondary and tertiary distribution of meat often followed primary distribution. The factors governing food sharing included hunting success, social relationships, method of capture, size of the animal, and the context of consumption. There was a lot of variability between households in species composition, but primarily in body part composition. Notably a lot of the sharing took place between settlements. She argues that the archaeological evidence for food sharing would be variability in assemblage composition and spatial distributiOI\ of taxa between households within and between sites, presence of complementary element profiles at differenfhouseholds or different sites, bones from different households or camps that could be refitted, and paired bones from individual animals at different households or sites as recognized through osteometric analyses. She also points out that although many of the differences between assemblages were due to differences in hunting skills and social relationships they might have been erroneously attributed to differential transport of carcass parts or site fun~tion. , She rightfully notes that "recognition of food sharing in archaeological sites will depend on exceptional conditions of preservation." Although Marshall is optimistic that sharing can be detected in the archaeological record, I have my doubts that this will be easy to accomplish. Many zooarchaeologists assume that food sharing or at least some form of meat distribution was a major dispersant of bones across sites. However, it is rarely possible to attribute faunal remains, especially in secondary refuse deposits, to particular households, much less to link elements from a single animal across a site or bet~een sites. The effects of post-consumption processes, for example, refuse/ disposal, burning, scavenging by dogs, and weathering, serve to disperse and damage bones, making refitting of breaks or the recognition of the bones from an individual animal difficult. The relatively gross temporal resolution of the archaeological record compounds the problem. In highly structured archaeological sites, for example, late prehistoric villages where house basins, pits, and refuse deposits can be associated, it may be possible to reconstruct food-sharing units. However, even for highly structured sites where refuse can be tentatively associated with households, potential evidence cited by Marshall cannot be exclusively linked to food sharing. As she observes for the Okiek, "some of the more typical body part profiles resulting from sharing are similar to those resulting from the logistics of carcass transport, scavenging, and the destruction of bone during site formation processes." As noted by Marshall, complementary body part profiles at two camps could reflect e:ither sharing or differences in settlement function. Bone fragments that can be refitted suggest potential contemporaneity of deposits, but the agent of dispersal is not necessarily sharing. It will be difficult to find unequivocal evidence for the reconstruction of food-
sharing netw1 the other facl be :supportec animals am01 tial transport ered, as are t attention nee between site1 cause of pattE
E1 0:
ences. He de groups at sitE Lese horticul processing si the presence each other et: at cooperath include Efe c that that site · the spatialla~ tion, the num and the exter and element group; thus, 1 ethnic group their villageE carcass-procE be used to re( ties of dried sites. Thus, tl of the struct, sometimes d] the basis of tl Attempts 1 follow a coar. is, the struct differences in data. Howev1 presence of c interaction w would undo1 discrete grou: disperse bonE
j
j
Social Interaction I 267 :1ce for food major factor k sites in the tses, animals kilometers." ary distribuccess, social ntext of conspecies comf the sharing
sharing networks in archaeological contexts. However, if we can control for all the other factors cited above, then perhaps the food-sharing explanation can be supported. As Marshall argues, given "the ubiquity of sharing of large animals among contemporary hunter-gatherers, ... the possibility of differential transport of elements through food sharing should be routinely considered, as are transport based on anatomical·utility and transport costs." More attention needs to be paid to patterning in species and body parts within and between sites and food-sharing needs to be considered as a potential major cause of patterning in the record.
Ethnic Differentiation -
tg would be :axa between tary element =eren t house1 individual osteometric .ces between relationships sport of car-
Only the paper by John Fisher attempts to document ethnic differences. He develops criteria for recognizing the presence of different ethnic groups at site, in this case for seminomadic Efe Pygmy foragers arid sedentary Lese horticulturalists in the context of cooperative elephant butchering and processing sites. He addresses the basic question of whether we can identify the presence of foragers and horticulturalists who are widely different from each other, ethnically, culturally, and socially on the basis of material remains at cooper.~tive carcass-processing sites. Those short-term processing camps include Efe dwellings, Lese dwellings, drying racks, and fires. He concludes that that site type can be differentiated from residential sites of both groups in the spatial layout, the disposal areas for discarded bones, the species composition, the number of individual animals, the paucity of debris other than bones, and the extensive breakage of the bones. However, the distribution of bones and elements precludes their association with a single dwelling or ethnic group; thus, the faunal data do not contribute to the differentiation of the two ethnic groups. The horticultura1 Lese store large quantiHes of dried meat at their villages and consequently build large drying racks at the cooperative carcass-processing sites. The remains of the racks were detectable and could be Used to recognize their presence. The Efe foragers do not store large quantities of dried meat; therefore, they build small drying racks at the processing sites. Thus, the presence of the two ethnic groups was detectable on the basis of the structural features, for example, spatial organization of houses and sometimes different house forms and different-sized drying racks, but not on the basis of the distribution of bones or body parts. Attempts to recognize different groups at archaeological sites generally follow a coarse-grained approach and at le,ast initially use the other data, that is, the structural features and the ·artifacts rather than faunal data. Ethnic differences in prehistory are generally reconstructed from these other types of. data. However, subsistence data may play an important role. For example, the presence of cultigens would certainly argue for the presence of, or at least interaction with, ;horticulturalists. Archaeologists dealing with older records would undoubtedly face serious problems in linking faunal remains with discrete groups at short-term camps bec~use of the wide variety of agents that disperse bones .
ological sites l1 Marshall is ·d, I have my gists assume a major disribute faunal households, etween sites. .se disposal, and damage nes from an .ution of the ehistoric villated, it may 1 for highly households, ked to food al body part the logistics during site dy part pro:1 settlement 1 contempoaring. It will .on of food-
1
. 1 -
2681 B. W. Styles
For archaeologieal issues, a comparison of faunal records from residential camps associated with the foragers and farmers would also provide an interesting case study for comparing fau11al signatures at sites associated with foragers and farmers. However, these specialized cooperative sites do provide a good opportunity to look at (1) the evidence for forager-farmer interactions and (2) the underlying differences in subsistence-settlement strategies that ultimately shape the material record.
Conclusions These studies all provide potential insights into archaeological faunal assemblages. However, the degree to which it is possiple to deal with these "social issues" in the prehistoric archaeological record varies. Although Fisher could not discriminate the presence of Efe foragers and Lese horticulturalists at cooperative butchering sites on the basis of faunal remains alone because of the nature of disposal practices at such sites, faunal remains do have the potential to contribute to ethnic or at least social differentiation. B'y means of historic analogues, faunal remains from historic archaeological sites in the eastern United States have been used to interpret socioeconomic status mediated by settlement function (see Reitz and Zierden 1991) based on the presence of particular meat cuts. Such techniques have been less successful for prehistoric sites where it would be difficult to rank the value of cuts other than on the basis of food utility. Disproportionately high quantities of high food utility cuts of deer meat have been recorded for prehistoric ceremonial sites in the eastern United States, and it should be noted that in at least one case the body part representation does not correlate with bone density (e.g., Styles and Purdue 1991). Although I think there is little to no doubt that food sharing ~ccurred throughout prehistory, I am pessimistic about unequivocal demonst,tations of food-sharing or the reconstruction of food-sharing networks, given the relatively coarse temporal resolution of the archaeological record and the difficulty of linking food refuse with particular household units at many sites. Complementary body part profiles at two sites could reflect settlement function, for example, transport of parts from a hunting camp to a residential base, provisioning of a residential base by a logistical camp, or transport of high food utility parts from a village to a ritual site for ceremony or tribute. However, Marshall is correct in her assertion that food sharing, which is ubiquitous in the ethnographic record, should be explicitly considered as a possible explanation for patterns in bone distributions within and between archaeological sites. If sharing followed set rules and was patterned by hunting success and social relationships as noted for the Okiek, and if the rules and these relationships endured over a long enough period of time, and if bones were preserved and could be associated with particular households or social subsets within a site, then these patterns might be evident in the archaeological record. More detailed studies of within-site faunal distributions are needed to document potential evidence for food sharing.
I l l :!
:t
lI :1
j
Although: tiona! or inci< and weatheri sumption, I · tribute to ou patterning iJ ethnoarchaec pdst-transpOI see at many' uct we see at to place morE ravages of til mental conte: cooking,. and understand t us to worry a
R Binford, L. R. 1967 Smu Reasonin; 1981 Bone 1984 Butc logical Arc Brain, C. K. 1967 BonE of Science 1969 The AustralOJ Research t Gould, R. A. 1969 Subf Oceania 31 1971 The American Isaac, G. L. 1967 Tow Observat: Lee, R. B., and 1968 Man Reitz, E. L anc 1991 · Catt Bobwhites Purdue,' Scientific Styles, B. W., c 1991 Ritu
Social Interaction I 269 1m residential vide an inter3ociated with es do provide ~r interactions trategies that
Although post-consumption processes, such as garbage disposal, intentional or incidental burning, scavenging by carnivores, trampling by humans, and weathering, serve to obscure evidences of butchering, cooking, and consumption, I think more systematic studies of bone modifications can contribute to our understanding of the agents and processes responsible for patterning in the archaeological record. I agree that more emphasis in ethnoarchaeological and experimental research needs to be placed on the post-transport part of the continuum from cooking through disposal. What we see at many archaeological sites is patterning in disposal. Since the end product we see at most archaeological sites is really garbage, perhaps we do need to place more emphasis on how refuse disposal, post-disposal events, and the ravages of time affect the evidence of the earlier, more interesting environmental contexts and human activities such as hunting, transport, butchering, cooking, and consumption. You do not have to love garbage, but it helps to understand the agents and processes that made it, modified it, and left it for us to worry about.
rchaeological to deal with ·ies. Although Lese horticul~emains alone :~.1 remains do rentiation. By ~ological sites Jnomic status based on the successful for of cuts other 1tities of high ·ic ceremonial .n at least one density (e.g., ~
References Binford, L. R. 1967 Smudge Pits and Hide Smoking: The Use of Analogy in Archaeological Reasoning. American Antiquity 32(1):1-12. 1981 Bones: Ancient Men and Modern Myths. Academic Press, New York. 1984 Butchering, Sharing, and the Archaeological Record. Journal of Anthropo-
logical Archaeology 3:235-257._ Brain, C. K. 1967 Bone Weathering and Problem of Bone Pseudo-Tools. South African Journal
of Science 63(3):97-99. The Contribution of Namib Desert Hottentots to an Understanding of Australopithecine Bone Accumulations. Scientific Papers of the Namib Desert Research Station 39:13-22. Gould, R. A. 1969 Subsistence Behavior among the Western Desert Aborigines of Australia. Oceania 39(4):253-274. 1971 The Lithic Assemblage of the Western Desert Aborigines of Australia. American Antiquity 36(2):149-169. · Isaac, G. L. 1967 Towards the Interpretation of Occupation Debris: Some Experiments and· Observations. The Kroeber Anthropological Society Papers 37. Lee, R. B., and I. DeVore (editors) 1968 Man the Hunter. Aldine, Chicago. Reitz, E. J., and M.A. Zierden 1991 Cattle Bones and Status from Charleston, South Carolina. In Beamers, Bobwhites, and Blue-points: Tributes to the Career of Paul W. Parmalee, edited by J. R. Purdue, W..E. Klippel, and B. W. Styles, pp. 394-405. Illinois State Museum Scientific Papers 23. Springfield.. Styles, B. W., and J. R. Purdue . 1991 Ritual and Secular Use of Fauna by Middle Woodland Peoples of Western
·ing occurred onstrations of ;iven the relaand the diffiLt many sites. :tlement func'idential base, .sport of high 1y or tribute. ing, which is nsidered as a and between rned by hunttd if the rules f time, and if 1ouseholds or vident in the l distributions
1969
l
.
------·
270 IB. W. Styles Illinois. In Reamers, Bobwhites, and Blue-points: Tributes to the Career of Paul W. Parmalee, edited by J. R. Purdue, W. E. Klippel, and B. W. Styles, pp. 420-436. Illinois State Museum Scientific Papers 23. Springfield. Yellen, John E. · 1974 The !Kung Settlement Pattern: An Archaeological Perspective.~Pnpublished Ph.D. dissertation, Harvard University, Cambridge. 1976 Settlement Pattern of the !Kung Bushmen: An Archaeological Perspective. In Kalahari Hunter-Gatherers, edited by R. B. Lee and I. DeVore, pp. 47-72. Harvard University Press, Cambridge. 1977 Archaeological Approaches to the Present: Models for Reconstructing the Past. Academic Press, New York.
IV.
-Il I
I-
·eer of Paul W. ., pp. 420-436.
Unpublished :tl Perspective. re, pp. 47-72.
cting the Past.
IV.
Noncultural Processes: Carnivore Scavenging and Density-Dependent Attrition
16.
A Carnivore's View of Archaeological Bone Assemblages Robert J. Blumenschine and Curtis W. Marean Abstract: The residues of carcass butchery by horninids offer carnivores a different range of feeding options than available on whole carcasses or on those fed on previously by other carnivores. The differing feeding opportunities, in turn, will alter the intensity and location of gnawing , damage and the. types of skeletal parts and portions that are dispersed I from or destroyed at sites. Variable properties of carcasses (e.g., size, nutritional quality) and differing degrees of competition among carnivores may further modify these aspects of carnivore disturbance of butchered bone assemblages. The relationships are documented through experiments that simulate archaeological bone assemblages, using both captive and fre~-ranging spotted hyenas. The experimental results demand that the reconstruction of hominid behavior from dualpatterned bone assemblages can only proceed once the activities of site scavengers have been ascertained.
Perspectives and Issues More than two decades have passed since Brain's (1967, 1969) observations on Hottentot bone middens forced archaeologists to recogn.ize c:arnivores as powerful taphonomic agents in the formation of zooarchaeological assemblages. His work, along with Isaac's (1967) lesser-known experimentation, was the 'first to demonstrate the need for controlled pattern-recognition studies that would allow archaeologists to distinguish the effects of carnivores and hominids on the accumulation and modification of fossil bone assemblages. A host of taphqnomic studies ensued, the urgency of which wa$ underscored by Binford's (1977, 1981) insistence that a spatial association of stone artifacts and animal bones at archaeological sites was insufficient to From Bones to Beha~ior: Ethnoarchaeological and Experimental Contributions to the Interpretation of Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights . reserved. ISBN 0-88104-076-2.
273
274 I R. ]. Blumens~hine and C. W. Marean produced b~ colleagues (" of whole ar spotted hyeJ and Marear include axil hyenas, as l Selvaggio (J tooth-marki stone tools. archaeologit strably more The purpc published s~ can affect th patterns of c ary access t< and distribu sity and lo1 produced b: size, nutritic of assembla!
proclaim hominids as the dominant agent of bone assemblage formation. The research effort has produced a detailed understanding of the composition and condition of bone assemblages cre~ted by either carnivores or humans (Andrews and Cook 1985; Binford 1978, 1981, 1984; Binford and Bertram 1977; Blumenschine 1988; Blumenschine and Selvaggio 1991; Bonnichsen 1973, 1983; Brain 1967, 1969, 1981; Bunn 1981, 1983, 1990; Bunn et al. 1988; Capaldo 1990; Casteel1971; Crader 1983; Gifford 1980; Gifford-Gonzalez 1990; Haynes 1980, 1982a, 1982b, 1983; Henschel et al. 1979; Hill1979a, 1979b, 1983, 1990; Hill and Behrensmeyer 1984; Hughes 1954; Johnson 1985; Ke~t 1981; Klein 1975, 1980; Lyon 1970; Maguire et al. 1980; Marean and Spencer 1991; Mareah et al. 1992; Morlan 1984; Myers et al. 1980; Noe-Nygaard 1977, 1989; O'Connell et al. 1988, 1989; Potts and Shipman 1981; Read-Martin and Read 1975; Richardson 1980; Sadek-Kooros 1972, 1975; Scott and Klein 1981; Selvaggio 1990; Shipman and Phillips 1976; Shipman and Phillips-Conroy 1977; Shipman and Rose 1983; Simons 1966; Skinner et al. 1980; Sutcliffe 1970; White 1952; Yellen 1977). Archaeologists have met the "carnivore challenge" by applying those models of single-agent bone accumulations to fossil assemblages. Yet many of the Pleistocene archaeological bone assemblages that have been used to interpret hominid behavior preserve clear evidence of udual-patterning" (Capaldo 1990) by carnivores and hominids. This is seen most conspicuously in the presence of both tool marks and carnivore tooth marks on bones from the same assemblage and, occasionally, on the same bone specimen. Single-agent models of assemblage formation, however, are not sensitive probes for isolating the behavioral signatures and contributions of carnivores and hominids to dual-patterned assemblages. The feeding activities of one agent will influence the range of choices available to the other and, hence, the extent and nature of damage to and dispersal of remaining bones and bone portions. The premise is justified by several studies, subsumed under the rubric of the hunting-scavenging debate. They have begun to shpw how carnivores, in reducing feeding opportunities for hominids on a cavcass, will alter the condition and range of parts available for hominid butchery and transport (e.g., Binford 1984; Blumenschine 1986, 1991; Potts 1983; see also Bunn 1986; Gifford-Gonzalez 1990). Relatively little experimental attention has been given to carnivores scavenging from hominid-butchered assemblages. This is ironic, given both the contention and importance surrounding the sequence and degree of carnivore and hominid involvement with Pleistocene archaeological bone assemblages (e.g., Binford 1986, 1988; Bunn and Kroll 1986, 1988) and the fact that Brain's and Isaac's work that posed the "carnivore challenge" was, purposely or not, modeling a hominid-followed-by-carnivore sequence of taphonomic agents. Crader (1983) showed the effects of carnivores scavenging large mammals butchered by the Valley Bisa in an attempt to assess the validity of Pleistocene "butchery" sites. Blumenschine (1988) and Blumenschine and Selvaggio (1991) showed how the epiphyseal:shaft fragment ratios and segmental distribution of tooth-marking and percussion-marking on long bone assemblages experimentally broken with a hammerstone and subsequently scavenged by spotted hyenas (Crocuta crocuta) differed from those
I j
ened menu 1 on a whole hominids, t refuse woul< axial bones. to includes cranial proc posed marrc animals whE remaining f• bone-crushil hominid-bu attractive to A carnivc butcher.ed a~ carcass. Can decreasing r defleshed b) order nutriti nutritional '
I
.J
r A Carnivore's View I 275 ·mation. The position and or humans ertram 1977; n 1973, 1983;
:tpaldo 1990; !aynes 1980, 190; Hill and 1
1975, 1980;
n et al. 1992;
:11 et al. 1988, 1rdson 1980; hiprnan and l Rose 1983; i977).
>lying those Yet many of ~en used to ·patterning" mspicuously l bones from :n. 1ot sensitive )f carnivores vi ties of one d, hence, the es and bone d under the >show how carcass, will utchery and 183; see also > carnivores ·, given both ;ree of carnibone assemthe fact that lS, purposely taphonomic ~nging large le validity of nschine and t ratios and :ing on long ! and subsel from those
produced by carnivores alone and hammerstone breakage alone. Binford and colleagues (1988) similarly experimented with the dispersal and modification of whole and hammerstone-broken buffalo (Syncerus caffer) long bones by spotted hyenas. More recently, Marean (1991), Marean and colleagues (1992), and Marean and Spencer (1991) have expanded these later experiments to include axial and compact (tarsals, carpals) bones using captive spotted hyenas, as has Capaldo (1990), using a variety of free-ranging carnivores. Selvaggio (1990) also examines the frequency and anatomical patterning of tooth-marking on bones defleshed and fragmented by carnivores and/ or stone tools. All of these studies aim to improve the realism of models for archaeological bone assemblages whose predepositional history is demonstrably more complex than that described by single-agent taphonomic studies. The purpose of this paper is to integrate and expand upon our previously published studies that demonstrate the manner in which hominid butchery can affect the nutritional attractiven~ss of parts to carnivores, thus influencing patterns of damage and deletion that are unique to carnivores having secondary access to bones. Specifically, we will illustrate how the different amount and distribution of food on a hominid-butchered assemblage affects the intensity and location of damage and the skeletal part and portion profiles produced by carnivore ravaging. We will also explore the effect of carcass size, nutritional quality, and competition among carnivores on these measures of assemblage disturbance.
Premise A hominid-butchered bone assemblage offers a carnivore a shortened menu of parts and a reduced nutrient yield compared to that available on a whole carcass. Prior to· the advent of grease-rendering technology by hominids, the skimpiest Plio-Pleistocene menu offered by hominid food refuse would have featured mainly grease from defleshed limb epiphyses and axial bones. If butchery was incomplete, the menu may have been expanded to include some flesh, especially the near-bone flesh around vertebral and cranial processes that is hard to remove with a stone tool, and some unexposed marrow cavities from long bones and axial parts. Except for very large animals where prior hominid feeding might not have depleted all flesh, the remaining fare· would be marginal and would entice only a carnivore with bone-crushing abilities (mainly hyaenids and canids). Despite the low yield of hominid-butchered fare, our studies demonstrate that it is still extremely attractive to scavenging carnivores (contra Binford 1981:286). · A carnivore's usual selection of parts for consumption from a hominidbutchered assemblage should also differ from its preferred choices on a whole carcass. Carnivores can be expected to select parts in a sequence that follows decreasing nutr~ent yield. This is true for the sequence by which bones are defleshed by a variety of carnivores (Blumenschine 1986). However, the rank order nutritional value of bones on a whole carcass differs from the rank order nutritional value of bones retaining only grease. A comparison of the rank
276 I R. f. Blumenschine and C. W. Marean order of 29 skeletal units in Binford's (1978) Modified General Utility Index; which includes the total flesh, marrow, and grease value of bones, reveals little correspondence to that for his Standardized Grease Index (Figure 16-1; Spearman's rho= .13, p = .49). One would also expect a bone-cracking and bone-crushing (Werdelin 1989) carnivore's selection of parts from a hominid-butchered assemblage to differ from the selection chosen from a kill abandoned by a flesh-specialist felid. With felid kills, the defleshed limb bones are still very attractive, not so much for the grease in their epiphyses, but for the ma:r;row in their diaphyses. Marrow extraction by carnivores will nonetheless result in the incidental consumption of most limb epiphyses prior to the consumption of defleshed axial parts (Blumenschine 1988; Haynes 1982b), probably because most marrow parts outrank axial grease yields. Hammerstone breakage of long bones, on the other hand, leaves no marrow but consistently creates a nutrient package-grease-filled limb epiphyses dissociated from diaphyses-that a carnivore never encounters on its own kills and finds only infrequently on those scavenged from flesh-specialists. In summary, hominid butchery will (1) restrict the parts that still offer nutrients, (2) create nutrient packages otherwise infrequently encountered by carnivores, and (3) in the process rerank the nutritional utility of skeletal parts. These considerations suggest that the nature and intensity of bone modification, destruction, and dispersal by a carnivore from a hominidbutchered assemblage will differ from those in which hominids played no prior feeding role.
The Experimental Sample We will address the nature and extent of carnivore distu:rbance to hominid-butchered assemblages through reference to several sets of experimental assemblages that we have reported previously (Blumenschine 1988; Blumenschine and Selvaggio 1988, 1991; Marean et al. 1992; Marean and Spencer 1991). Results of those studies will be summarized and new analyses of the data sets presented. We will continue to refer to these experiments as "simulated sites" (Blumenschine 1988; Marean 1991), for they were designed with the explicit purpose of simulating assemblages of hominid-butchered bones that are subsequently disturbed by scavengers. We will also continue to refer to the single-agent control assemblages as "hammerstone-onl y" and "carnivore-only" (Blumenschine 1988). One group of simulated site experiments was created in the Serengeti National Park, Tanzania, in 1983-84 (Blumenschine 1988; see also Blumenschine and Selvaggio 1988, 1991). Twelve experiments were conducted, using fresh bovid limb bones, usually one hindlimb and one forelimb, of a single individual. The sample (Table 16-1) includes Thomson's gazelle (Gazella thomsoni; size 1), Grant's gazelle (Gazella granti) and impala (Aepyceros melampus; both size 2), wildebeest (Connochaetes taurinus) and topi (Damaliscus korrigum; both size 3) and buffalo (size 4). Individual limb elements were
I
A Carnivore's View 277
Jtili ty Index, )nes / reveals (Figure 16-1;
\
P FEMUR
RIB PELVIS D TIBIA THORACIC SCAPULA
I
P HUMERUS D HUMERUS CERVICAL
~-?0?~;~~
LUMBAR
lsturbance to ~ts of experitschine 1988; Marean and 1ew analyses perimen ts as ere designed ~d-butchered
continue to te-only" and
D
he Serengeti 8; see also s were con)ne forelimb, 3on' s gazelle la (Aepyceros i (Damaliscus ~ments
were
'=
ASTRAGALUS
I
CALCANEUM TARSALS P METATARSAL P RADIUS-ULNA D METATARSAL
-~
D RADIUS-ULNA
.~~
w~~~~~J!~fiJ
CARPALS
%
MANDl B LE
.. -
2D PHALANX 3D PHALANX 1ST PHALANX
tat still offer :oun tered by y of skeletal .si ty of bone . a hominidIs played no
1§11
~-~
P TIBIA STERNUM
erdelin 1989) lage to differ ecialist felid. not so much r diaphyses. te incidental of defleshed se most marf long bones, ~s a nutrient .yses-that a ~equen tl y on
Modified General Utility Index
Standardized Grease. Index D FEMUR
~~ ~4fi2~~.;i@'*tNflJJJS~~;;'I,J
P METACARPAL
~
D METACARPAL
~
AXIS ATLAS I
30
25
I
T
I
I
20
15
10
5
0
5
I
I
I
I
10
15
20
25
30
Figure 16-1. Comparison of the rank order of 29 skeletal parts based on Binford's (1978) modified general utility index (meat, marrow, and grease) and his standardized grease index (grease within cancellous bone). Highest utility parts have a rank of 1. For long bones, D =distal, P =proximal. ~
usually disarticulated after total defleshing. Carpals and tarsals were usually disarticulated as a unit and left attached by the posterior tendon bundle of a ~etapodial to the phalanges, which were not broken. All long bones were broken, using a hammerstone-on-anvil technique, and to an extent sufficient to remove all marrow with the aid of a thin, spatulate scoop. To the extent possible, like bones were broken in a similar manner by placing hammerstone blows at the same locations. On the day of breakage, all fragments produced by hammerstone breakage were placed in simulated sites in a variety of settings so they could be disturbed freely by bone-crushing carnivores. Most were placed in Acacia woodland or plains settings. Two that were placed in riparian woodlands were not disturbed after as many as 13 days; they were recovered .intact and are here included in the hammerstone-only sample described below. Most simulated site bones derived from animals in good condition prior to death. The marr?w (and grease) in these bones was fat-rich, as determined by visual inspection (Sinclair and Duncan 1972) and additionally by oven-drying experiments. Two simulated sites were composed of bones from a wildebeest (size 3) whose marrow and grease was severely fat-depleted. The Serengeti
278
I R. J. Blumenschine and c. w. Marean Table 16-1. Composition of the Experimental .Sample
Site Type Simulated sites Serengetia Fat-rich Fat-depleted Berkeley 1b Berkeley 2c,d
Size 1 & 2 N Assem NISP
TOTAL N Assein NISP
326 1162
4 2 4 0
267 44 98
10 2 13 33
554 44 424 1162
3
91
6
140
9
231
6
303
1
24
7
327
57
2169
15
573
74
2742
6 0 9 33
Carnivore onlya Hammerstone onlya
TOTALd
Size3 &4 ·N Assem NISP
287
Note: Carcass size groups are defined in the text. Number of assemblages and number of identified specimens (NISP) are given. anata reported by Blumenschine (1988). bNew analyses. cnata reported by Marean and Spencer (1991) and Marean et al. (1992). dFive sites (NISP = 229) overlap with Berkeley 1.
simulated sites are therefore divided into a fat-rich sample (10 assemblages) and a fat-depleted sample (2 assemblages; Table 16-1). The Serengeti simulated sites were probably disturbed mainly l]y spotted hyenas. This was not confirmed by direct observation since all d;isturbance occurred at night and sites were not monitored continuously. However, hyenas were heard at a number of the sites, and many preserved hyena spoor after disturbance were complete. Black-backed jackals (Canis mesomelas), common in the Seronera area, probably also contributed to assemblage disturbance, as Capaldo's (1990) ongoing experiments amply suggest. Disturbance was complete after the first night, as indicated by the usual total deletion of limb epiphyses. All bones remaining in the vicinity of the the original site (within a 50 m radius) were collected and cleaned by boiling for subsequent analysis. A second group of simulated sites was created at the Berkeley Spotted Hyena Colony, Berkeley, California,. between 1988 and 1989 (Marean 1991; Marean and Spencer 1991, Marean et al. 1992). Limb bones and some postcranial axial bones of sheep (Ovis aries; size 1 /2) and cow (Bas taurinus; size 3I 4) obtained from a local butcher. were used. The experiments at Berkeley afforded complete observations of hyena disturbance and strict control of their number and dominance rank. Such experimental control, which could not be achieved for the Serengeti sites, was
l
-!
I j
possible ow hyenas live ( simulation I simulated s collected aft. came to lean While the data analysi marking anc sites (refern Serengeti sat also summa previously 1: for 33 assen overlap wit: Marean and includes 21 whichhyenc Two sets c analyses. B conducted i assemhlageE bones (withe animal that stone-only ' composition The carni vm to be produt whole long[ Thus, our data sets th free-ranging stone-brokeJ hammers ton ed butcher 1 from carnivc no carnivore
•1
l
~-
I 1
c
portions. Er distal articu· ity' s sake, pr derive from
-l.,. -;.?.-.'
,.. r ~
'.r·,-·.'.-'.'.·'..· ·-
lI
possible owing to the physical layout of enclosures in which the Berkeley hyenas live (see Marean et al. 1992 for a plan of the experimental setting). Presimulation preparation of bones was similar to that used for the Serengeti simulated sites. All remaining fragments from the Berkeley sites were collected after 15 minutes of no activity by the hyenas, a duration which we came to learn as signaling lack of interest fat as much as several hours. While the attribute analysis of the Berkeley assemblages is complete, the data analysis is not. Here, we present new data on the incidence of toothmarking and epiphyseal:shaft fragment ratios for 13 of the Berkeley simulated sites (referred to as "Berkeley 1" in Table 16-1). Comparable data from the Serengeti samples is presented in this bivariate form for the first time here. We also summarize information on skeletal part and portion profiles reported previously by Marean and Spencer (1991) and Marean and colleagues (1992) for 33 assemblages from Berkeley ("Berkeley 2" in Table 16-1), 5 of which overlap with the above sample of 13. In Marean and Spencer (1991) and Marean and colleagues (1992), this sample of 33 hyena-ravaged assemblages includes 21 in which the bones were first hammerstone-broken, and 12 in which hyenas were fed unbroken, defleshed bones. Two sets of single-agent assemblages are included as controls in subsequent analyses. iBoth are currently limited to experiments and observations conducte'd in the Serengeti (Blumenschine 1988). The hammerstone-only assemblages consist of seven experiments (Table 16-1). Each includes long bones (without phalanges) from usually one forelimb and one hindlimb of an animal that contributed its other limbs to a simulated site. The hammerstone-only assemblages provide a pre-simulation control on the original composition of the simulated sites in a way described by Blumenschine (1988). The carnivore-only sample consists of nine assemblages (Table 16-1) observed to be produced by hyenas, and, in one case, a lion (Panthera leo), destroying whole long bones of animals they killed or scavenged fron1 other carnivores. Thus, our total experimental sample analyzed to date includes an array of data sets that provide the complementary benefits of observations in both free-ranging and captive settings. The sample comprises defleshed hammerstone-broken bones scaven.ged by carnivores in a natural setting, defleshed hammerstone-broken bones scavenged by hyenas in a captive setting, defleshed butcher bones scavenged by hyenas in a captive setting, bones derived from carnivore kills in a natural setting, and hammerstone-broken bones with no carnivore involvement.
TOTAL
ssem NISP
10 2 13 33
554 44 424 1162
9
231
7
327
74
2742
A Carnivore's View I 279
1d number of
ssemblages) by. spotted disturbance '. However, hyena spoor melas), com:tge disturbr
>y the usual ty of the the r boiling for ley Spotted 1991; some postzurinus; size
Definitions and Identification of Carnivore Tooth Marks
~arean
Our sample of fragmented long bones is divided into the following portions. Epiphyseal fragments preserve at least some of the proximal or distal articular surface, along with varying lengths of the shaft. For simplicity's sake, proximal metapodials are lumped with epiphyses even though they derive from the metaphyseal growth area of the bone's diaphysis. Near-
f hyena disSuch exper:i sites, was
l
1
, 280 I R. ]. Blumenschine and C. W. Marean epiphyseal fragments are those that preserve cancellqus bone on a portion of the medullary surface or other morphology indicative of their derivation from a proximal or distal shaft. All near-epiphyseal fragments can be grouped into eitJ'ter proximal shafts and distal shafts, as used by Marean and Spencer (1991). Long bone midshafts, at least on bovids, have no cancellous bone. The term shaft encompasses both near-epiphyseal and midshaft'fragments and is comparable to the term limb shafts used by Bunn (e.g., 1986). Tooth ·marks were identified only after bones had been thoroughly mascerated by boiling to produce clean, grease-free surfac~s. Identification followed procedures and criteria of Blumenschine and Selvaggio (1988, 1991). A 16-power hand lens aided all identifications. Tooth marks on cortical and medullary surfaces occur as roughly circular to equilateral depressions (pits), or more linear marks (scores; see also Binford 1981; Bunn .1981; Potts and Shipman 1981; Shipman and Rose 1983). The internal cross-sectional topography is U-shaped for scores (following Bunn 1981) and usually bowl-shaped but occasionally more angular for pits. Unlike other types of marks, internal surfaces show conspicuous crushing regardless of whether the pressure was downward only (pits) or horizontal, producing a drag effect (scores). Pun'ctures .and furrows (Binford 1981) are, respectively, subcircular and linear depressions through the thin cortical bone that covers trabecular bone of epiphyses. The vast :majority of marks identified as tooth marks are much less conspicuous than the punctures, furrows, or dense patches of deep pits and scores that occur along severely gnawed edges of the bone. Instead, they are often isolated, faint, and shallow (but with a still appreciable depth), which, along with the distinctive internal crushing, can be detected and distinguished from other marks only upon close examination under strong, low-incident light with the aid of a hand lens. Scanning electron microscopy is unnecessary, and optical microscopy reveals little of diagnostic value that the hand lens does not. Our experience with control assemblages shows that thesff obscure marks are as accurately diagnostic of carnivore action as marks associated with heavy gnawing.
Tooth-Marking Figure 16-2a shows the incidence of tooth-marked long bone fragments for three sets of simulated sites and the Serengeti carnivore-only sample. Blumenschine (1988) has reported the data upon which the carnivore-only and fat-rich Serengeti simulated sites are based. Here the sample is expanded, and values for tooth-marking are reported more informatively as assemblage means rather than as total sample averages so that variability can be expressed. Mean incidence of tooth-marked fragments is shown for all bones, and for different bone portions (epiphyseal fragments, near-epiphyseal fragments and midshaft fragments). A bone fragment was counted as being tooth-marked regardless of the number or conspicuousness of the marks it bore; many tooth-marked specimens preserve only a single, inconspicuous mark.
l
a. By
--- 1Q( "0 '
rn
+
8(
'-"' +-'
c
Q) (J L Cl)
0...
c
6(
4(
0 Cl)
::::;E
-,~
l
b. By
I J I
2(
1Ql ,......._ '"0 (/)
81
T'""'
+ ...__, +-'
c (l) u L (l)
0...
61
4
c 0
(l)
~
2
1 I
I
A Carnivore's View 281 a portion of ivation from ~rouped into md Spencer 1s bone. The nents and is
a. By Bone Portion
· - carnivore only t:s:sl Serengeti simulated site (fat-rich) IZZl Serengeti simulated site (fat-depleted) ~ Berkeley simulated site
,...... 100 "'0 .· (f)
~hl y
mascer.on followed m, 1991). A cortical and ssions (pits), 1; Potts and nal topogra)Owl-shaped rks, internal >ressure was ·ores). Puncr and linear 1lar bone of re much less eep pits and ~ad, they are -pth), which, .istinguished .ow-incident 1nnecessary, te hand lens lese obscure s associated
+
'-.J
80
+'
c
0)
u
60
L 0)
Q._
c
40
0
0)
L
20 0
NEF
EPI
MSH
TOTAL
Bone Portion
b. By Size Group
100
-Size 1 & 2 t:s:sl Size 3 & 4
~
"0 (/)
80
or-
+
"-J
+-'
c
60
(!)
0
L. (!)
(l_
g bone frag~ni vore-only :h the carnihe sample is rmatively as Lriability can wwn for all r-epi physeal ted as being the marks it :onspicuous
40
c 0
(!)
L
20 0 CARNIVORE ONLY
SERENGETI SIMULATED SITES
BERKELEY SIMUlATED SITES
Figure 16-2. Proportion of long bone fragments that bear at least one tooth mark, by (a) bone portion: EPI =epiphyseal fragment, NEF =nearepiphyseal fragment, MSH = midshaft fragment, and (b) carcass size group (see text for definitions).·
282 I R. ]. Blumenschine and C. W. Marean Carnivores extracting marrow and grease from whole long bones typically produce a high (mean = 84%, min. = 67%, max. = 100%) proportion of fragments with at least one tooth-mark (Blumenschine 1988). Epiphyseal fragments that are not consumed completely are invariably tooth-marked, while a slightly lower proportion of near-epiphyseal fragments (85%) and midshaft · fragments (83%) are tooth-marked. The segmental distribution and overall incidence of tooth-marking in the carnivore-only sample contrasts strongly with those produced by carnivores scavenging hammerstone-broken long bones. The overall incidence of toothmarking in all three simulated site samples is significantly lower~ with sample means not exceeding 20%. The highest incidence, obtained for one fat-rich Serengeti site, is 45%. The substantially lower values for the simulated sites are primarily due to the very low incidence of tooth-marked mi_dshafts. Here, carnivores typically ignore hammerstone-generated midshaft fragments because they have been deprived of virtually all nutrient value (Blumenschine 1988). Our direct observations of site disturbance at the Berkeley Colony confirm this pattern: only rarely, a hyena may briefly "mouth" a midshaft but will invariably reject it without biting. Those simulated site midshafts that are tooth-marked are therefore derived exclusively from hammerstone-generated epiphyseal fragments through consumption of cancellous ends by hyenas. Hyenas focus on epiphyseal fragments because they are the only bone portions in hammerstone-generated assemblages of long bones that contain nutrients (grease; Blumenschine 1988). In destroying epiphyses, hyenas will produce tooth-marked near-epiphyseal fragments in high frequencies. An interesting exception to this rule is provided by the two simulated sites composed of fat-depleted bones. The absence of tooth-marked near-epiphyseal and midshaft fragments in the two assemblages (Figure 16-2a) is a result of an absence of on-site epiphyseal destruction. Remaining epiphyses are commonly tooth-marked in this sample, but only with tooth punctures inlo cancellous bone, and not pits or scores ori cortical surfaces. The hyenas disturbing these sites apparently needed to use their sense of taste to sample the nutrient value of grease in the epiphyses' cancellous bone before rejecting it as a whole unit. Variability in the incidence of tooth-marked fragments among the simulated sites is probably also based on competition. Competition among scavengers of simulated sites will influence whether destruction of hammerstonegenerated epiphyseal fragments occurs on-site, or only after epiphyses are transported away from the site to minimize competition while feeding. Onsite epiphyseal destruction will generate maximal levels of tooth-marked fragments, while complete deletion prior to destruction will leave the simulated site with no tooth-marked fragments but nevertheless depleted numbers of epiphyses. This premise can only be supported at this time qualitatively. Nonetheless, in experiments at Berkeley using a single hyena, we have commonly observed it to destroy many epiphyses at or near the site, unless a hot sun forced it to seek shade or an infestation of wasps forced a retreat from the main concentration of bones. However, when two or more hyenas are present, subordinate individuals will commonly remove any
epiphyses tr consuming 1 tion would E sampled by Capaldo (19~ et al. 1988). Interestin~
marking at marking inc Serengeti (f contrasts wi1 larger anima (Figure 16-21 In summa ultimately d Carnivores ' 84% of bon epiphysis t< broken long all bone spe1 to midshaft, marks. The~ secure distin access to ass~
I - J
I
l
l.
I T epiphyses t< shaft fragmE (near-epiph) tify carnivor sity of their 1 Loss of ha on-site destr to levels pro Such severe exceeds by a sample. Los: composed o: (Figure .16-3) Variabilit) sample is h: tencies in bo one assembli a (size 1) T
A Carnivore's View I 283 nes typically ·tion of frag'hyseal frag·ked, while a nd midshaft
epiphyses they are able to acquire to a remote corner of the enclosure before consuming them (Marean etal. 1992). In free-ranging conditions, such deletion would effectively remove any resulting fragments from the area typically sampled by excavation, a premise that is being investigated systematically by Capaldo (1990) with free-ranging carnivores in the Serengeti (see also Binford et al. 1988). Interestingly, bone size does not seem to influence the incidence of toothmarking at the simulated sites. There is virtually no difference in toothmar king incidences between size 1 I 2 and size 3 I 4 animals for both the Serengeti (fat-rich) and the Berkeley simulated sites (Figure 16-2b). This contrasts with the carnivore only sample, where the more robust diaphyses of larger animals seem to promote a higher incidence of tooth-marked fragments (Figure 16-2b). In summary, the frequency and distribution of carnivore tooth-marking is ultimately determined by the presence or absence of within-bone nutrients. (::arnivores consuming whole bones produce tooth marks on an average of 84% of bone fragments, and the frequency decreases only slightly from epiphysis to midshaft. In contrast, carnivores scavenging hammerstonebroken long bones generate tooth marks on only an average of about 20% of all bone specimens, and the frequency dwindles significantly from epiphysis to midsWaft, with less than approximately 10% of the midshafts bearing tooth marks. These overall and segmental patterns of tooth-marking provide a secure distinction between primary carnivore access and secondary carnivore access to assemblages of long bones.
Lrking in the y carnivores nee of toothwith sample one fat-rich nulated sites shafts. Here, ·agments belumenschine eley Colony midshaft but 1afts that are le-generated ·hyenas. e only bone that contain hyenas will' uencies. An ~d sites comr-epiphyseal 1 result of an ;es are com; into cancelLS disturbing : the nutrient it as a whole
Long Bone Epiphysis: Shaft Fragment Ratios The above shows that carnivores typically :onsume or destroy epiphyses to access grease but ignore nutritionally unattractive, ungreasy shaft fragments. This principle suggests that the ratio of epiphyseal to shaft (near-epiphyseal plus midshaft) fragments (NISP values) can be used to identify carnivore involvement with a bone assemblage and to measure the intensity of their ravaging. Loss of hammerstone~generated epiphyseal fragments through deletion or on-site destruction by carnivores can reduce epiphyseal:shaft fragment ratios to levels produced by carnivores destroying whole long bones (Figure 16-3). Such severe epiphyseal loss in the fat-rich Serengeti simulated site sample exceeds by an order of magnitude the mean value for the hammerstone-only sample. Loss of epiphys~al fragments was less severe in the simulated sites composed of fat-depleted bones and in those created at the Berkeley Colony (Figure 16-3). · Variability in epiphyseal:shaft fragment ratios for the hammerstone-only sample is high ..Part of this variability results from unavoidable inconsistencies in bone fracturing methods. However, much of the variability is due to one assemblage (NAA39G) compos.ed ~f forelimb and hindlimb long bones of a (size 1) Thomson's gazelle (Figure 16-3). Bones of such small animals
lg the simuamong scavtrnmerstoneJiphyses are feeding. On>oth-marked 11 leave the ~ss depleted at this time ingle hyena, ~ear the site, sps forced·a :wo or more remove any
l
"
1 !
2841 R. ]. Blumenschine and C. W. Marean
1.6 ~Size 1 & 2 lZZl Size 3 & 4
(f)
+' 4--
0
...c
1.2
(/)
D
0.
NM39G (size 1) (f)
" ;;-
+'
4--
~
({)
~ ~ 0...
0
...c (/)
(/)
(/}
0
(/}
w ~ L
E
CJ)
0
>. >
...c
..._;
Q)
...0
,.......
"~ ·- 0 . Q)
0.8
0.4
::J
z
0
0...
1
L
0
··a_ w Q)
L
(f)
0...
0.
...0 ......_,_
E
fat-rich CARNIVORE ONLY
::J
z
fat-depleted · Berkeley SIMULATED SITES
HAMMERSTONE ONLY
0.
Figure 16-3. Size-specific destruction/deletion of long bone epiphyses as measured by epiphyseal:shaft fragment ratios for all experimental samples. · Fi
typically require only one blow to access all marrow with a probe. This minimizes the production of shaft fragments in comparison to bones of larger animals where addi tiona! blows toward the proximal and distal ends are needed to remove all marrow. If the resulting very high epiphyseal:shaft fragment ratio (1.56) of this assemblage is excluded, the hammerstone-only mean drops to .36, a value still greater than the me;;tns for any of the simulated site samples. Aside from the above example, carcass size does not seem tci have a consistent effect on epiphyseal: shaft fragment ratios. In fact, epiphyseal loss is similar for assemblages using size 1/2 and size 3/4 bones across the spectrum of site formation scenarios examined (Figure 16-3). A tight (r = -.94) negative relationship exists between the post-ravaging epiphyseal:shaft fragment ratio and the observed percentage change in the original number of epiphyseal fragments (Figure 16-4).1 This shows that the epiphyseal:shaft fragment ratio accurately measures depletion of epiphyseal fragments by carnivores. Percentage change in the number of epiphyseal fragments was calculated using the following formula:
b-a %change=-- x 100 b where b is the number of epiphyses before disturbance, and a is the number remaining at or near the site after disturbance. Values forb were recorded for all simulated sites upon breakage, while those for the carnivore-only sample were obtained by simply multiplying the estimated number of whole bones
dt
of ra destroyed b) , more severe stone-on!y as fragments ari nally thoughi The two rr independent always resuli relationship ~ be described fairly immun ber and seqw Variability probably a p (1991) has qt analyses sug duced is in( decreases~ Gi seal:shaft fra: the Serengeti than that exp the Berkeley
J:'•
'1.00·'····'····:.:·
I
A Carnivore's View j285 0 Serengeti carnivore only D Serengeti simulated site (fat-rich)
0.6
'V Serengeti simulated site (fat-depleted) 6. Berkeley simulated site
(/)
+-' '+-
0 Hammerstone only
0
..c
(f) ,.-....
'-...~ (/)Q) ·-(Jl 0.4 (/)
0
>. >
..c 0 o_ L
·o_
l
w (/) L
Q) _Q
Q o_
0.2
E
y = .32- .0031x r = .94
:J
z MERSTONE ONLY
e epiphyses as experimental
1e. This minines of larger , ;tal ends are !al:shaft fragLe-only mean [mulated site n to have a hysealloss is the spectrum ost-ravaging hange in the .OWS that the >f epiphyseal f epiphyseal
; the number recorded fbr -only sample whole bones
......,
0.0+-~~~~-,--~.-~--.-~-.--~-.~~w
-40
100 60 20 40 80 0 -20 % Change in Number of Epiphyseal Fragments [ (No. before - No. after) / No. before x 1 00]
Figure 16-4. Relationship between two measures of epiphyseal fragment deletion and destruction, including known percentage change in number of epiphyseal fragments, and post-ravaging epiphyseal:shaft fragment ratios.
destroyed by two. Increasingly positive percentage change values indicate more severe epiphyseal loss. Negative values, obtained for most hammerstone-only assemblages (Figure 16-4), reflect the post-boiling addition of shaft fragments arising from the removal of periosteum that bound fragments originally thought to represent a single hammerstone-broken fragment. The two measures of epiphyseal deletion/ destruction are not completely independent. A 100% change in the number of epiphyseal fragments will always result in an epiphyseal:shaft fragment value of zero. Nonetheless, the relationship shows that the intensity of epiphyseal removal by carnivores can be described by a single measure-epiphyseal:shaft fragment ratios-that is fairly immune to differences in carcass size, fat content of bones, and the number and sequence of consumer species breaking bones. Variability in the epiphyseal:shaft fragment ratio between simulated sites is probably a partial function of competition among site scavengers. Marean (1991) has quantified competition· for 33 Berkeley Colony sites. Preliminary analyses suggest that as the number of hyenas per number of limbs introduced is increased (i.e., increased competition), survival of epiphyses decreases. Given this relationship, one can infer that the lower mean epiphyseal:shaft fragment ratios obtained for the fat-rich simulated sites created in the Serengeti results from a higher level of competition among site scavengers than that experienced by the Berkeley sites. This is not surprising, given that the Berkeley hyenas were fed .at regular 24-hour intervals and that no more
------~
l
j
j 286j R. ]. Blumenschine and C. W. Marean
j 1
than four hyenas were ever given access to a single. site. Such conditions are likely to promote less competition than those which can characterize the site disturbance of free-ranging hyenas. In summary, carcass size does not ·severely affect the proportionate survival of limb epiphyses and shafts. Epiphyseal:shaft fragment ratios correlate well with the depletion of epiphyses by carnivores across the spectrum of site formation scenarios examined here. Hence, this simple ratio value of fragment counts (NISPs) can provide an accurate estimate of epiphysis loss through carnivore agencies. Variation in epiphysis:shaft ra~ios reflects the intensity of carnivore destruction, and in a way that appears to be sensitive to levels of competition among carnivores.
j
10
i
l
I
I l ~
~ ......., (/)
8
t
res entation >res entation $agents of .ndMunson ed to dogs. ium duiker, Table 17-1). al exhibited somewhat (1 kg) and L duiker (18 lon by dogs lt, and MNI e in simple >w extreme
variability in archaeological survival as measured by MNI. Sample size may be influencing the cluster of cases with small body size and high survival rates. Examination of Figure 17-4 suggests that sample size does correlate with :MNI survival, such that species represented in the original discard assemblage by only one or two individuals are more likely to be fully represented by MNI, regardless of body size. This is not surprising, given the fact that :MNI is calculated from the most abundant surviving element. For an MNI of one, any identifiable element will serve. To register higher values of MNI, the same element must survive from each. of the original carcasses, and the probability that this will occur can be expected to decrease as the number of carcasses increases .
NISP Figure 17-5 and Table 17-2 illustrate the relationship between NISP and actual numbers of individuals. Although NISP is not theoretically linked to numbers of individuals in the same way that MNI is, it is often used as a measure of relative taxonomic abundance. Of the 149 animals originally contributing to the assemblage, 898 identifiable fragments were recovered, representing an average of 6 fragments per carcass. In spite of the degree of loss of identifiable bone, the correlation between NISP and actual individuals per taxon is statistically significant (r =·.921, probability of error< .001, df = 14, N = 16), as is rank order (rs = .786, probability of error oe:: .01).
308
I J. Hudson
30
25 20
1(
!B.
15
ca
::::J
10
"'C
:~
"'C
Forest Rat
Squirrel Porcupine
Blue Oui ker
MediLJll Oui k.er
Tortoise
Civet
Monkey
Yellowbacked Dui ker
Mongoose
Lizard
-coc:· ::::J ..... 0
rrelated , 9 of the density. 1 assem;es (60%) density~ responcrushing or some with the rtionally wgorovgnificant 2nted by re repre,tatistical :Grayson 'esults of 7 assem-
)rrelated ~s 1, 2, 3) nediated quencies reside in
I
%of 87 Total Assemblages
Difference Between 67 and 87 Assemblage Samples
10.3 9.2 1.2 32.1 34.5 1.2 0.0 10.3 1.2
4.3 -1.3 -0.3 -3.7 3.2 -0.3 0.0 -1.6 -0.3
I,
It
I I I !
t
l
I l
Il
I
Coefficient Set Class Class 1 (destruction) Class 2 Class 3 Class 4 (destruction) Class 5 Class 6 Class 7 (destruction) Class 8 Class 9
%of 20 Additional Assemblages 25.0 5.0 0.0 20.0 45.0 0.0 0.0 5.0 0.0
%of 67 Original Assemblages 6.0 10.5 1.5 35.8 31.3 1.5 0.0 11.9 1.5
I
I I
Note: See Tjiible 18-1 for details and Figure 18-1 for definitions of coefficient classes.
I
I
I l
I i
I'
I
I I I ~ !-
J
I Il I J
l
j
i I ~
!
!i
the fact that density-mediated attrition subsumes many taphonomic processes such as carnivore gnawing, fluvial transport, fragmentation, and crushing from overburden weight, whereas the MGUI is meant to monitor only differential transport and use of carcass parts. Thus, there simply are more ways for a bone assemblage to become correlated with bone density than with the MGUI. Considering the total sample of 87 assemblages, 9 of the 30 "ethnoarchaeological" assemblages (30%) are positively correlated with bone density, whereas 28 of the 57 "archaeological" or prehistoric assemblages (49%) are positive.., ly 'correlated with bone qensity. That difference in proportions is weakly statistically significant (arcsine transformation ts = 1.75, 0.1 > p > 0.05). These results indicate that there is some potential for postdepositional destruction to more frequently result in: significant positive correlations ·of prehistoric bone assemblages with bone density than for ethnoarchaeological bone assemblages to correlate positively with density. A final comment seems warranted regarding the absence of any Class 7 assemblages (Figure 18-1). That such assemblages should be extremely rare, if not nonexistent, can be explained by the fact that bone frequencies in Class 7 assemblages would have to not only indicate gourmet or bulk utility strategies but also correlate positively with density. Given that the correlation of the MGUI with density indicates bones that rank high in utility tend to rank low in density and that bones that rank low in utility also tend to rank high in density, it is not surprising that Class 7 assemblages are not to be found in the 87 assemblages I have examined.
3341 R. L. Lyman
Other Issues Large Versus Small Bovids a.t Klasies River Mouth The potential meaning of variation in skeletal part frequencies at the South Mrican site of Klasies River Mouth (Binford 1984, !'989; Klein 1989; Turner 1989) may be revealed when considered in the light of Marshall and Pilgram's (1991) conclusions. Small bovids at Klasies and at the pastoral site of Ngamuriak described by Marshall (1986; Marshall_ and Pilgram 1991) are represented by proportionally more proximal limb elements, whereas large bovids are represented by proportionally more distal limb elements. This pattern is explained by Marshall and Pilgram (1991) as resulting from more consistent and/ or intensive exploitation of the within-bone nt,ttrients (grease and marrow) of cattle bones and less consistent and/ or intensive exploitation of those nutrients in the smaller bones of caprines. Klein (1989) presented correlation coefficients between the frequencies of skeletal parts of five size classes of bovids from Klasies and bone density; they are summarized in Table 18-4. Binford (1981) presented bone frequency data for six taxa of bovids whose remains were recovered from a hyena den. Those remains represent at least three size classes of bovids. I (Lyman 1991a) correlated those bone frequencies with density; the coefficients are presented in Table 18-4. The pattern that emerges is clear. Skeletal part frequencies of small bovids (< 84 kg live weight) do not correlate with bone density, whereas skeletal part frequencies of large bovids (> 84 kg live weight) do correlate with bone density (chF = 10.07, p = 0.005). Granting that density is inversely correlated with the amount of grease in a bone, the coefficients in Table 18-4 suggest more frequent and/ or more intensive exploitation of that grease in large bovids than in small bovids, regardless of the taphonomic agent. Is this pattern sustained when other assemblages are included? To show t)1at it is, the Swartkrans Member 2 size class III bovid remains, the Swartkrans Member 2 size class II bovid remains, and the Kromdraai A size class II bovid remains as listed in Table 18-1 can be added. As well, I recalculated MAU frequencies for large and small bovids for Makapansgat from Dart (1957) and for large and .small bovids for Richardson's (1980) hyena den data. Addition of those seven assemblages does not significantly alter the result obtained by using only those coefficients listed in Table 18-4 (Table 18-5; chF = 8.00, p = 0.005). There seem to be two ways to pursue this issue. First, if the densitymediated attritional agents that produced the pattern of coefficients in Tables 18-4 and 18-5 were in fact extracting within-bone nutriel'\ts, I would expect more percussion marks (Blumenschine and Selvaggio 1988), flake scars (Lyman 1987), or gnawing marks on the bones of larger taxa than on the bones of small taxa. However, such data are not available for these assemblages. The second way to evaluate the more intensive exploitation of within-bone nutrients in larger bovids is to examine other assemblages. Relevant data are, however, unavailable. On one hand, bone frequency data for four bovid taxa collected by Andrew Hill and described by Binford (1981:214-215) were
Hyena dE
Klasies R
Ngamurii
avalues ar bcoefficieJ to 84 kg, II from Klein ccoefficier dcoefficie: (1978) 90-n
consider' remains· a passivE mula ted remains, On the ·c fromNoJ all but or ungulate lope (Ani considen two in T; that·fall; eluded w those thr positive!~
can data sizes of 1
rI i i
~
i l
I I at ~in 1989; hall and al site of 991) are ~as large tts. This 1m more ; (grease loitation ~ncies
Density-Mediated Attrition I 335 Table 18-4. Correlation Coefficients (r 5) Between Bone Frequencies for Bovid Size Classes and Bone Density
II !
I
Hyena denb (Binford 1981)
!
1' a
1
Klasies River mouthc
~ncies
of ity; they ncy data n. Those a) correented in of small whereas late with ly correble 18-4 ;rease in tt. Is this hat it is, 1.s MemII bovid ~dMAU
957) and '\ddition :tined by 8.00, p =
densityn Tables d expect ke scars he bones ges. 1in-bone data are, 1vid taxa 5) were
Ngamuriakd
Bovid Size Cla$S
rs
Grysbok (I) Springbok (II) Reedbuk (ll) Topi (Ill) Wildebeest (Ill) Kudu/ eland (Ill/IV) Small (I) Small-medium (ll) Large-medium (ill) Large (IV) Very large (V) Caprine (ll) Cattle (lli-IV)
0.39 0.25 0.25 0.65 0.58 0.64 0.14 0.25 0.53 0.73 0.55 0.09 0.43
p 0.15 > 0.15 < 0.005 < 0.005 < 0.005 0.46 0.18 0.005 0.0001 0.003 0.72 0.07
Summed MNia 143 117 141 85 232 112 258 245 313 544 140 299 140
avalues are sum of MNI for each skeletal part included. bcoefficients from Lyman (1991b); size classes I-IV after Brain (1981): I = 0 to 23 kg, II= 23 to 84 kg, III= 84 to 296 kg, IV=> 296 kg. Size class V after Klein (1975, 1989). Original data from Klein (1975). ccoefficients from Klein (1989). dcoefficients recalculated from corrected data in Marshall and Pilgram (1991). Binford's (1978) 90-month-old sheep weighed 44.9 kg.
considered in my earlier analysis (Lyman 1991a). However, because those remains were from surface contexts, represent modern animals, and represent a· passive accumulation a~ross a large landscape rather than an actively accumulated assemblage like the other assemblages considered in Table 18-4, the remains were deemed inappropriate for consideration in Tables 18-4 and 18-5. On the other hand, while there are many assemblages- of bones recovered from North American sites considered in my earlier analysis (Lyman 1991a), all but one of those assemblages involves taxa> 84 kg in live weight. The only ungulate in North America that is< 84 kg live weight is the pronghorn antelope (Antilocapra americana) whoseremains make up two of the assemblages considered ih Table 18-1.-Because the three assemblages (one in Lyman 1991a~ two ih Table 18-1 here) are the only North American assemblages with taxa that fall in the small size class, comparisons like that in Table 18-S are precluded with those assemblages. It is perhaps important to note, however, that those three North American small ungulate assemblages are all correlated positivgly with density and thus suggest that the pattern shown by the African data (Table 18-5) may be a function ofthe wide range of available body sizes of ungulates ~n that continent. That is, African organisms (whether
3361 R. L. Lyman Table 18-5. Classification.of 20 Correlation Coefficients Between Bone Density and Small (< 84 kg live weight) and Large(> 84 kg
live weight) African Bovids Live Weight Size
84kg
Number correlated with bone density
3
9
Number not correlated with bo.ne densitya
7
1
aChi2 = 8.00, p = 0.005.
hominid or hyaenid or whatever) may have selected bones of larger taxa instead of bones of smaller taxa for extraction of w~thin-bone nutrients because they had several size classes to choose from, whereas North American · organisms had fewer size classes from which to choose.
Greater Density Values When I originally chose the bone density values that were to be correlated with the MGUI and MAU values, my choices were somewhat "subjective" and geared toward "averages" for the anatomical region being quantified (Lyman 1984:287). Recent discussions of those choices (Rapson 1990; Gifford-Gonzalez, personal comm~nication) have suggested that the choices do not represent realistic ones and that perhaps the greatest density value for an anatomical region would be a better choice (see also LymaJil-1992). _ Following that suggestion, I recalculated correlation coefficients betw'een the 87 assemblages and the greatest density value for particular anatomical portions. That resulted in only six (7%) of the assemblages having to be reclassified: two prehistoric assemblages changed from Class 4 to Class 5, three prehistoric assemblages changed from Class 5 to Class 4, and one ethnoarchaeological assemblage (Klippel et al.'s wolf-gnawed deer) changed from Class 5 to Class 4. These minimal changes are no doubt due to the fact that the density values I have consistently used are quite tightly correlated with the greater density values others have suggested sho,uld be used (r5 = 0.88, p < 0.0001, N = 31). The net result of using the greater-density values is that two assemblages are removed from Class 5 and added to Class 4. Thus, while the results of my original analysis do not change significantly when the greater density values are used, the lessons here should be clear. First, density-mediated destruction is a very real phenomenon that must be dealt with during the analysis of skeletal part frequencies. That is so because with the maximum density values, 39 of the 87 assemblages (45%) are positively correlated with density, which, while only a 3% increase over the results when the traditionally utilized density values are employed, indicates that as many as half of the assemblages that we might study may well have been
subjectec the 95 sc skeletal I extent of dictated scan site ysis of tl 150% of much gn proxima] ful analy tion and
and Sorg on tapho 1991; Pu: data two ical and c not been this anal· Europe,. Only 6 o: original, relates tc (1988:12~
among t: prehistor for exa1 (1986:641 rion [for data prm hominidE Regard what iss various u and Ree\ reported, sented or are left v. alternativ are identj are now c l'v1NE val
Density-Mediated Attrition I 337
84kg 9 1
r taxa Lts beerican
:to be ewhat being apson ,at the ensity 1992). en the 1l porclassi~e pre:haeo]ass 5 .ensity ;reater 11,N= blages cantly clear. ust be ~cause
~ posiresults :hat as
~been
subjected to density-mediated attrition. Second, quantitative units should be the 95 scan sites I originally defined (Lyman 1984) rather than some arbitrary skeletal part. The 95 scan sites provide 95 data points with which to assess the extent of density-mediated destruction, rather than the more usual 25 to 30 dictated by the MGUI. Density-mediated destruction of the most proximal scan site on the radius, for instance, may not mean that the proximal diaphysis of the radius was also destroyed, as the latter has a bulk density that is 150% of the former (scan sites RA1 and RA2 in Lyman 1984) and thus has a much greater chance of surviving density-mediated attrition. Noting that the proximal end is missing but the proximal shaft is not, then, could be a powerful analytic tool that allows us to disentangle the effects of differential destruction and differential transport.
Discussion With the publication of at least two edited volumes (Bonnichsen and Sorg 1989; Davis and Reeves 1990) and various articles and dissertations on taphonomy (e.g." Blumenschine 1989; Hudson 1990; Marshall and Pilgram 1991; Purd,ue et al. 1989; Rapson 1990; Stiner 1990) since my compilation of data two years ago, I expected a wealth of new assemblages, both archaeological and ethnoarchaeological, that I could add to my sample. That simply has not been the case in the literature I have examined, although I note that since this analysis I have found a half dozen assemblages from the Near East and Europe, all published prior to 1990, which could be added to the analysis. Only 6 of the 20 new data sets !.examine here appeared since I compiled the original data set (Marshall and Pilgram's 1991 data excluded). Perhaps this relates to a stronger belief in the mid to late 1980s that, to paraphrase Potts (1988:122), skeletal element frequency data are not adequate to distinguish among the diverse attritional and accumulational agents that acted upon prehistoric bone assemblages. That, however, does not seem likely because, for example, we find statements such as this by Blumenschine (19.86:641): "Body part [frequencies] seem to represent the most telling criterion [for distinguishing hunted from scavenged faunal assemblages. Such data provide] unambiguous evidence for the timing of access to carcasses by hominids versus carnivores." Regardless of whether one occupies the Potts or the Blumenschine position, what is surprising is that analysts still interpret bone frequencies, producing various utility graphs of their zooarchaeological data (various papers in Davis and Reeves 1990), yet the data upon which those graphs are based are not reported, save in a graphed or other less than explicit form. The values represented on a bivariate scatterplot are often extremely difficult to derive, so we are left with a graph that represents an interpretation but no "raw" data. An alternative is to find the %:MNI per skeletal element listed, but whether those are identical to what Binford (1978) originally labeled %MNI values and what are now called (%)MAU values is seldom clear. Another alternative is to read MNE values per complete skeletal element (e.g., Blumenschine 1986, 1989),
3381 R. L. Lyman
which masks the important data 40 R. L. Lyman blage from Swartklip I, South-Western Cape Province, South Africa. Quaternary Research 5:275-288. . · 1989 Why Does Skeletal Part Representation Differ Between Smaller and Larger Bovids at Klasies River Mouth and Other Archaeological Sites? Journal of Archaeological Science 16:363-382. · !Clippel, W. E., L. M. Snyder, and P. W. Parmalee 1987 Taphonomy and Archaeologically Recovered Mammal Bone from Southeast Missouri. Journal of Ethnobiology 7:155-169. Lyman,R. t.. 1984 Bone Density and Differential Survivorship of Fossil Classes. Journal of Anthropological Archaeology 3:259-299. 1985 Bone Frequencies: Differential Transport, In Situ Destruction, and the MGUI. Journal ofArchaeological Science 12:221-236. 1987 Archaeofaunas and Butchery Studies: A Taphonomic P.erspective. In Advances in Archaeological Method and Theory, vol. 10, edited by M. B. Schiffer, pp. 249-337. Academic Press, San Diego. 1990a Zooarchaeology. In Archaeological Data Recovery at Hatiuhpuh, 45WT134, Whitman County, Washington, edited by D. R. Brauner, pp. 98-138. Oregon State University Department of Anthropology report to the Walla Walla District Army Corps of Engineers, Corvallis. 1990b Zooarchaeology of 45GR445. Unpublished report to Archaeological and Historical Services, Eastern Washington University, Cheney. 1991a Taphonomic Problems with Archaeological Analyses of Animal Carcass Utilization and Transport. In Reamers, Bobwhites, and Blue-Points: Tributes to the Career of Paul W. Parmalee, edited by J. R. Purdue, W. E. Klippel, and B. W. Styles, pp. 125-138. Illinois State Museum Scientific Papers 23. Springfield. 1991b Zooarchaeology of 10IH1017. In Prehistory and Paleoenvironments at Pittsburgh Landing: Data Recovery and Test Excavations at Six Sites in Hells Canyon National Recreation Area, West Central Idaho, edited by K. C. Reid, pp. 373-432. Washington State University Center for Northwest Anthropology Project Report No. 15. Pullman. 1992 Anatomical Considerations of Utility Curves in Zooarchaeology,Journal of Archaeological Science 19:7-22. r Lyman, R. L., and M. J. O'Brien 1987 Plow-zone Zooarchaeology: Fragmentation and Identifiability. Journal of Field Archaeology 14:493-498. Marean, C. W.,and L. M. Spencer 1991 Impact of Carnivore Ravaging on Zooarchaeological Measures of Element Abundance. American Antiquity 56:645:_658. Marshall, F. 1986 Implications of Bone Modification,in a Neolithic Faunal Assemblage for the Study of Early Hominid Butchery and Subsistence Practices. Journal of Human Evolution 15:661-672. Marshall, F., and T. Pilgram 1991 Meat Versus Within-Bone Nutrients: Another Look at the Meaning of Body Part Representation in Archaeological Sites. Journal of Archaeological Science 18:149-163. Oliver, J. S. , 1986 The Taphonomy and Paleoecology of Shield Trap Cave (24CB91), Carbon County, Montana. Unpublished Master's thesis, University of Maine at Orono. Potts, R. 1988 Early Hominid Activities at Olduvai. Aldine de Gruyter, New York
Purdue,J. 1989 Upla Miss< Rapson, C 1990
Resea New Richardso 1980
Paleo1 Stiner, M. 1990
Pleist Univ~
Turner, A. 1989 Partie: Klasi1
Density-Mediated Attrition 1341
ternary Larger
ArchaeSouth-
Irnal of md the
:ive. In ffer, pp. WT134, m State District cal and
Carcass
es to the dB. W.
d. nents at Canyon )73-432. t Report
Jurnal of mrna{ of Element ;e for the
f Human
;of Body
l Science 1
County,
Purdue, J. R., B. W. Styles, and M. C. Masulis 1989 Faunal Remains and White-Tailed Deer Exploitation from a Late Woodland Upland Encampment: The Boschert Site (23SC609), St. Charles County, Missouri. Midcontinental Journal of Archaeology 14:146-163. Rapson, D. J. 1990 Pattern and Process in Intra-Site Spatial Analysis: Site Structural and Faunal Research at the Bugas-Holding Site. Unpublished Ph.D. dissertation, University of New Mexico, Albuquerque. Richardson, P. R. K. 1980 Carnivore Damage to Antelope Bones and Its Archaeological Implications. Paleontologia Africana 23:109-125. Stiner, M. C. 1990 The Ecology of Choice: Procurement and Transport of Animal Resources by Upper Pleistocene Hominids in West-Central Italy. Unpublished Ph.D. dissertation, University of New Mexico, Albuquerque. Turner, A. 1989 Sample Selection, Schlepp Effects and Scavenging: The Implications of Partial Recovery for Interpretations of the Terrestrial Mammal Assemblage from Klasies River Mouth. Journal of Archaeological Science 16:1-11.
\
19.
Discussion: Noncultural Processes
interpret human-g could pre brigade, cerning n anthropo menton 1
Anna K. Behrensmeyer Introduction The theme of chapters 16 through 18 is how to identify and control for noncultural processes in order to untangle them from cultural ones that are of interest to archaeologists and anthropologists: Researchers who pursue evidence of human behavior often regard noncultural taphonomic processes as troublemakers that need to be "controlled." A paleobiologist takes a more positive viewpoint on these processes as intriguing records of any and all biological and physical agencies that affect the vertebrate fossil record. Frustration levels are typically much lower for the paleobiologist, who is happy to be able to relate features of a bone assemblage to any one of many possible causal processes. The anthropological focus has, however, catalyzed a great deal of productive taphonomic research and is helping to define the limits for inferring human behavior (or that of any other single agent) from bones. Identifying specific taphonomic processes that interacted with a given bone assemblage is becoming easier with the growth of actualistic studies that show what various agencies, such as wolves, hyenas, and trampling, do /o bones. The zooarchaeologist has a new and improved set of tools and guidelines for identifying who or what has affected such assemblages. Controlling for the noncultural components, that is, "removing the taphonomic overprint/' is much harder and remains a difficult problem for ethnoarchaeological and fossil assemblages alike because (1) most assemblages reflect multiple processes, (2) many different processes produce similar effects, and (3) the same bones may be affected by different processes in varying sequence. Ideally the archaeologist would like to be able to say exactly which features in an assemblage were caused by humans and whic-h were not, but with increased appreciation for the complexity of this problem, we are left wondering whether even unequivocal qualitative analysis is possible. _ The role of taphonomic research over the past decade has often been characterized as that of a "wet blanket/' continually showing that simplistic From Bones to Behavior: Ethnoarchaeologicalcmd Experimental Contributions to the Interpretation of Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights reserved. ISBN 0-88104-076-2.
342
vide disti cultural, cautioh n Thepa they dea: Althougl grap!llc s tially,gen the value Blume1 scavenge human-b on thee> define be identify i assembla tion beca ing resul· be recorc segments ments co hyena b1 Serengef ecosyster entire Pli multicorr strong si1 icked by1 Hudso small ani Republic units are resolutioJ contribu· Blumens(
Noncultural Processes 1343
;ses
interpretations of such features as. breakage patterns or surface marks as human-generated were potentially wrong because other noncultural processes could produce similar effects. As an invited representative of the wet blanket brigade, I feel duty-bound to present some of the discouraging reality concerning noncultural processes, as well as a theoretical treatment of the issue of anthropologically oriented hope versus this reality. First, however, I will comment on the three papers in Part IV.
Comments on Written Contributions
i control 'nes that ) pursue ~rocesses
sa more ·and all ~d. Fruslappy to possible l a great imits for es. ren bone 1at show :o bones. ~lines for s for the >rint," is ;ical and ~ple pro:he same eally the n assemd apprewhether ·en been implistic >Jretation of )ccasional All rights
The papers in Part IV, and many others in the volume as well, provide distinct glimmers of hope for distinguishing at least some of the effects of cultural versus noncultural processes. They also provide expected words of caution regarding interpretations of archaeological bone assemblages. The papers by Blumenschine and Marean and by Hudson are similar in that they deal with the effects of scavengers on human-generated assemblages. Although they derive information from specific experimental and ethnographic situations, they distill from them some interesting ideas about potentially general patterns that could be tested with further work. This underlines the valud of case studies that relate to a larger theoretical framework. Blumenschine and Marean aim their study at documenting the behavior of scavengers (specifically, the spotted hyena) when these animals encounter human-butchered bones and use their observations to test predictions based on the expected effects of scavengers on bone assemblages. Their goal is to define bone assemblage and bone modification features that could be used to identify the "hyena after hominid" dual-component effects in archaeological assemblages. They point out that these two agents cannot be studied in isolation because the activities of one affect those of the other. One of the interesting results of the study is that competition levels in a hyena population may be recorded in the types of damage and degree of removal of edible bone segments. There are several critical points that apply to this paper: (1) experiments conducted in the artificial hyena colony reflect only a limited view of hyena behavior, some of which may not be entirely "natural"; (2) the Serengeti experiments, .while much less artifical, still represent only one ecosystem at one point in time and should not be taken as models for the entire Plio-Pleistocene of Africa; and (3) most "real" situations will involve multicomponent rather than dual-component taphonomic systems, and even strong signatures imposed by hominds and hyenas may be masked or mimicked by other processes qffecting archaeological bone assemblages. Hudson presents a controlled study ofbone assemblages from relatively small animals that are the prey of the Aka forest people in the Central African Republic and how they are affected by scavenging dogs. General body part units are used to .show the effects and are deemed more informative than finer resolution skeletal part tallies. The dominance of heads and limb fragments contributes to a growing number of case studies, including that of Blumenschine and Marean, that show similar patterns in scavenging situa-
T
344IA. K. Behrensmeyer
tions. In spite of an overall loss of 47% of the bone assemblage after scaveng..: ing by dogs, the rank ordering of taxonomic abundance- is similar to that of the pre-dog assemblage. This offers hope that scavenging need not severely alter species representation at an archaeological site, at least for animals under 25 kg body weight. Hudson raises the issue of a prey body-size threshold in scavenging effects, an interesting possibility that needs further study. As in the paper by Blumenschine and Marean, an important problem involves how the assemblage characters Hudson documents would be further modified by other processes en route to becoming an archaeological site. The paper by Lyman is quite different from the first two in its emphasis on the general empirical observation that bone density often explains much of the patterning in modern and archaeological bone assemblages. Lyman's contri-bution is important because it underlines the fact that differe11t processes can result in similar patterns because of the density effect. His data show that modern carnivores do not always cause density-dependent patterning, which is interesting and somewhat contrary to expectation. He also suggests that there may be a threshold for density effects, and that might help to explain why some of the carnivore assemblages do not show this pattern. The dat'a also show that archaeological assemblages are more often correlated with density than modern ones, and Lyman speculates that it may be partly a consequence of burial and, compaction. There is a need for more rigorous process-based explanations for the observations presented in this paper. Otherwise, there is no theoretical framework for predicting under what circumstances density-dependent effects seriously comprise our ability to distinguish cultural signals in bone assemblages. The empirical studies discussed above are informative and heuristic research efforts. They contribute to the growing body of data that provides a foundation for interpreting noncultural processes in bone assemblages. Such studies are essential building blocks that help to generate as well as test new ideas and hypotheses.
At Death 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
Initii Idim Prey Prey
Heal Seas1 Tim~
Habi Pres1 Timi
Pre-Buria 1. Tran 2. Win! 3. Wea1 4. Buric Post-Buri 1. Plan! 2. Insec 3. Soil c 4. Com 5. Diag; - Post-Diso 1. Exca' 2. Meth 3. Ident 4. Meth
Realities of Noncultural Taphonomy The more research that is done on taphonomic processes, the greater our appreciation for the complexity of what happens between the death of an animal and the final preserved state of its skeletal remains. Unraveling taphonomic history and isolating the record of human activity are challenging and sometimes impossible-tasks. There is often great disparity between what we want to know and what we can know, given the limitations of the data. This is less discouraging, however, if we see increased understanding of complexity as a natural phase in the evolution of the relatively new field of zooarchaeological taphonomy, to be followed by improved interpretative frameworks and questions better suited to the available evidence. The complexity involved in taphonomic analysis of noncultural processes can be illustrated using one of the currently popular variables in zooarchaeology: skeletal part frequency. Table 19-llists processes, agents, and characteristics of the bones themselves that are known or suspected to have differential
effects OI assemble affected 1 the stren relative b The pr cultural1 tap honor through: in singlerelatively not humi question Hope is 1~
Noncultural Processes j345 scavengo that of severely tls under ~shold in iy. As in lves how iified by )hasis on _ch of the 's contri~sses can lOW that g, which ests that ) explain The data ted with partly a rigorous .s paper. Nhat cirto distinh.euristic ·ovides a ~es. Such test new
sses, the veen the remains. tivity are :l.isparity nitations i underelatively ·ed interlence. )rocesses Hchaeollracterisfferential
Table 19-1. Variables Known to Affect Skeletal Part Frequencies At Death or Shortly Thereafter 1. Initial agent of carcass modification (species of predator or scavenger) 2. Idiosyncratic behavior of carnivores or scavengers (within same species) 3. Prey species morphology 4. Prey size 5. Health of prey 6. · Season (affects predatorI scavenger behavior) 7. Time of day (as for #6) 8. Habitat 9. Presence of additional resources (e.g., a shade tree) 10. Timing of access by multiple predator /scavengers Pre-Burial 1. Trampling and kicking 2. Winnowing and transport 3. Weathering 4. Burial potential (a function of size, shape, density) Post-Burjal 1. Plant roots 2. Insects 3. Soil chemistry 4. Compaction 5. Diagenesis (mineralization) Post-Discovery 1. Excavation techniques 2. Methods of counting 3. Identification of fragmentary parts 4. Method of quantitative analysis
effects on skeletal part survival. Not all of these variables would affect every assemblage, of course; but most excavated samples probably would be affected by at least several of them. Isolating one agent or process depends on the strength and distinctiveness of its overprint on skeletal part frequencies relative to the effects of other processes. The present state of affairs regarding hope (i.e., probability) of unraveling cultural components in bone assemblages versusthe reality of noncultural taphonomic overprints can be simply portrayed in graphic form (Figures 19-1 through 19-3 ). The chances for distinguishing a cultural signature are greatest in single-component assemblages (i.e., those formed by a single process), for relatively simple questions (Figure 19-1). Such a question might be whether or not hu:rnan activity affected the assemblage at all, whereas a more complex question would involve the nature of the activity or the number of humans. Hope is lower for th.e complex questions and also for a,ssemblages affected by
346J A. K. Behrensmeyer
t
I
FOR SIMPLE QUESTIONS FOR COMPLEX QUESTIONS
HOPE OF UNRAVELING CULTURAL COMPONENT
UN I
I NFI
(E
NUMBER OF TAPHONOMIC PROCESSES/AGENTS
Figure 19-1. Schematic plot of hope (general probability) of deciphering a specific human-generated component in a bone assemblage versus the number of taphonomic agents or processes that affected that assemblage. SINGLE /COMPONENT
t
DUAL COMPONENT
HOPE OF UNRAVELING CULTURAL COMPONENT
/
INTENSITY OF TAPHONOMIC PROCESS
~
Figure 19-2. Schematic plot showing the possible relationship between hope (as defined in Figure 19-1) and the intensity of a particular taphonomic process, such as human' butchery, for single-component, dualcomponent, and multicomponent bone assemblages (see text for further explanation). two or more processes. Another factor to consider, however, is the relative impact of a process on the bone assemblage (Figure 19-2).1 In singlecomponent systems, hope of identifying the taphonomic agent or process is high at relatively low intensity but may fall off with time because most destructive processes eventually eliminate fragile, low-density parts. This
results ir or multic reach.tliE overwhe There case stw theory-b< behavior growth' growth i: from nea might bt Often ho edge oft bolsterec ecologic' numbers hopes fa] emerge 1 amplituc long run, the who] perhaps' (althoug: would fif
Noncultural Processes 1347
THEORY /{OFTEN BORROWED) .--;
,
HOPE OF UNRAVELING CULTURAL INFORMATION (BEHAVIOR)
f f I f
DATA ON VARIABILITY
I
KNOWLEDGE OF "REAL WORLD" ~ (ACTUALISTIC & ARCHAEOLOGIC DATA)
'-JTS
·iphering a ~ersus the emblage.
'p between ar taphonent, dualcor further
e relative n single::Jrocess is use most arts. This
Figure 19-3. Schematic plot suggesting changes in hope (probability) of unraveling cultural information with increased knowledge of the complexity of taphonomic processes. The dashed lines represent the optimistic versus pessimistic initial outlook on the problem based on minimal knowledge.
results in the density-dependent effects noted by Lyman. In dual-component or multicomponent systems, it takes greater intensity of a particular process to reach the maximum hope of identifying this process because its effects must overwhelm those of other processes. There was considerable discussion at the conference regarding descriptive case studies of bone assemblages, especially ip. ethnoarchaeology, versus theory-based hypothesis testing. Both of those approaches to inferring human behavior from material remains can be represented in the context of the growth of zooarchaeology over the past 15-20 years. In Figure 19-3; this growth is presented as the increase of knowledge of the "real world," starting from near zero when new questions are asked. An example of such a question might be, What were the resource utilization strategies of early humans? · Often hopes of answering new questions are high when there is little knowledge of the complexity of the evidence. This optimistic point of view may be bolstered by theory, which is often borrowed initially from another field. The ecological theory of resource optimization is one example. With increased numbers of descriptive studies providing data on complexity and variability, hopes fall as cautionary tales have their day, but eventually new patterns may erp.erge to raise hopes again within a modified theoretical framework. The amplitude of these cycles of optimism and pessimism may be damped in the long run, although new questions or new general theories can appear to reset the whole system. At present, the field appears to be in the first valley and perhaps beginning to climb up the next hill as new hypotheses are developed (although, of course, individual researchers vary greatly as to where they would fit on the curve).
3481 A. K. Behrensmeyer
Areas of Emphasis for the Future There are many positive st~ps that could be taken to better define what we can and cannot know about cultural processes and human behavior from bone assemblages. They include the following: 1. Refinement of our tools of analysis, as advocated in Lyman's paper, including multivariate analysis of assemblage characters such as skeletal part frequencies, species body size, breakage, and bone surface modification. 2. Comparative analysis of archaeological sites, such as demonstrated in the work by M. Stiner, and also more deliberate multivariate comparisons between ethnoarchaeological and archaeological assemblages. 3. Use of computer-assisted simulations based on theoretical predictions, ethnoarchaeological, or noncultural taphonomic assemblages to provide comparative data for multicomponent and time-averaged archaeological assemblages. 4. Testing of different scales of analysis, such as advocated by Hudson and Stiner. For some questions, general categories of body parts may b'e more informative and also more statistically robust than finer categorizations using individual bones or segments.
Note 1. Different processes may generate increasingly similar features in a bone assemblage the longer or more intensely they act upon it, thereby reducing the hope of unraveling cultural components. There is perhaps more chance in multicomponent situations of having at least one process that continues to leave distinctive traces with increased intensity, which is why the nhope line" remains slightly higher for multiI component than for dual-component assemblages in this diagram.
20. made b, through presentE related· major tll points o tal studi itably tc: ference 1 alogical Jo Wats Behrens: Grays some of most etl differen about he produce tions are Cauti< amount think th: ductive theoretic That i: ing and vent us f can lead wise ha' might bE From Bone. Faunal Rer.
Paper No reserved.]
define 1avior paper, keletal nodifiated in )lnpar::s.
predicIges to eraged Iudson
maybe catego-
assemwpe of lponent es with ~multi-
20.
Concluding Discussion: The Role of Actualistic Studies
The following discussion represents an integration of comments made by various participants during the discussion sessions that occurred throughout the two-day conference in 1991. Much more was said than is presented here. Editorial license has been taken in juxtaposing comments on related topics and in restricting the topics to those most closely tied to the major themes of the conference. The aim has been to capture some of the key points of dialogue concerning the role of ethnoarchaeological and experimental studies and the future directions that such actualistic research might profitably takf. Don Grayson served as an important catalyst early in the conference by questioning the utility of some recent approaches to ethnoarchaeological research. Other participants whose comments are cited here are Patty Jo Watson, Diane Gifford-Gonzalez, John Yellen, Susan Kent, and Kay Behrensmeyer. Grayson: I would like to begin by saying that I am not as enthusiastic about some of this ethnoarchaeological work as I used to be. It seems to me that most ethnoarchaeology that has been published to date has made two very different kinds of contributions. Much of it has produced cautionary tales about how complex the world is. Some of it has produced, or at least tried to produce, general statements about the way the world works. Both contributions are important, but they are important in very different ways. Cautionary tales are important. I have certainly learned a tremendous amount from them. But I think the importance of these tales is short-term. I think this is the case because the cautions that we are issuing are largely inductive reactions to issues of the day, and many of the issues lack general theoretical grounding. That is not to say that inductively derived cautionary tales are not interesting and helpful. They help define what we cannot or do not know; they prevent us from treating complex, multivariate situations in simplistic ways; they can lead us to examine the. archaeological record in ways we might not other- · wise have done; and they can suggest ways in which perplexing phenomena might be explained. Nonetheless, this type of ethnoarchaeology is inductively From Bones to Behavior: Ethnoarchaeological and Experimental Contributions to the Interpretation of Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional Paper No. 21. © 1993 by the Board of Truste~s, Southern Illinois University. All rights reserved. ISBN 0-88104-076-2.
349
I
350 Concluding Discussion·
driven and represents particularistic responses to archa,eological analyses. There is a second kind of ethnoarchaeology that I find both far more intriguing and of longer-term value. I will take O'Connell's work as an example. While he has been busy issuing important cautionary tales, it is also true that most of the ethnoarchaeological work that he has done has been embedded in a. particular theoretical approach, which has as its goal understanding why particular cultural systems work the way they work One may or may not like the optimization models that O'Connell employs, but what is important to me is that his work does have a clear theoretical context. Watson: It seems to me that what is at issue here is the perennial debate in many fields about which should be uppermost-particularizing or general- _ izing goals: cautionary tales are particularizing and optimization theory is generalizing. If we are doing science, then of course generali.zing should be stressed: explanation of groups or classes of data has priority over explanation ()f individual cases. ·But you cannot achieve and maintain good generalizations without good particularistic work Don [Grayson] has suggested that a lot of ethnoarchaeological research is inductive, a reaction to current concerns, and will be of passing interest only. In what sense is it inductive, and why is it of short-lived importance? Don may mean that research of the cautionary tale variety is inductive because it does not flow deductively from a specific theoretical framework. He may feel it is of short-lived importance because the results are often inconclusive, or irrelevant to archaeological types of data, or because researchers may make a pass at a subject and then drop it for another. It seems likely to me that what is happening is basic science at a relatively early developmental stage: constructing an observational method to get at the issues of general concern-past human beings, human societies, and human cultures. To use Kay Behrensmeyer's phrasing, the effort is aimed at bringing hope and reality closer together through taphonomic studies, Bifl.fordian middle-range theory, and site formation processes. Results may be short-lived for a very good reason: because a particular question or set of questions has been answered, or has been shown to be a deadend, and it is time to extend inferential boundaries in new directions. O'Connell's research is predicated upon cost-benefit analysis of specific aspects of a particular human society in one part of the world. Presumably, his research is motivated by a belief that optimization theory can be used with considerable profit by himself and others doing "real archaeology" (as Kevin Jones phrased it) on analogous archaeological data. But to do real archaeology about human optimization you would certainly like to be able to sort out spotted hyena from hominids, and faunal assemblages created by schlepp effects, food-sharing, or culinary processes from those created by noncultural factors such as wind, water, dogs, and earthworms. Both particularistic and generalizing approaches are essential and, in fact, inseparable. By using both, we learn a lot along the way and progress toward the goal that I think everyone does have in mind: figuring out how the world of humankind works, regardless of which world is meant-the human past, the human present, the archaeological record, or all three.
GiffoJ decoupl to addre lar, prec systems tures on Yelle1 useful aJ vance.F essential Andr perspect often sej own exF exercise animall conditio: Simil2 groups; fuller raJ If wear aspects_( cally. Kent: sions is· that prol archaeol alone. Behre multiple use a gn particulc of assoc words tc nation o: Yeller often rer data are blagemc con tinge Behre: time-ave Yelle1 needed. months, understc: the factc theoretic
Concluding Discussion 1351 rses. more in~xample.
true that 2dded in ling why r not like mt to m:e lebate in general:heory is 1ould be )lanation neralizasearch is ·est only. ctive bevork. He nconclu1ers may elatively at the i human bringing nfordian ort-lived :ions has J extend ~et
specific ;umably, sed with :.ts Kevin 1aeology Jut spotJ effects, Ll factors in fact, 'toward le world tan past,
~.,
Gifford-Gonzalez: While there is a danger with middle-range research of decoupling it from general theory, I think that we do need actualistic studies to address uniformitarian processes. We need to investigate and identify regular, predictable relationships between dynamic processes and the states of systems that generate them on the one hand, and their archaeological signatures on the other. Yellen: Particularistic studies in response to archaeological questions are useful and important. They typically do involve research issues of broad relevance. For example, I would say that many of the papers presente