Hudson (Ed) (From Bones To Behavior Ethnoarchaeological and Experimental Contributions To The Interpretation of Faunal Remains 1993) PDF

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
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

edited by Susan A. Gregg
Processual and Postprocessual Archaeologies:

Multiple Ways of Knowing the Past
edited by Robert W. Preucel
Quandaries and Quests:

Visions of Archaeology/ s Future
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
ISBN: 0-88104-076-2 Ackn
Library of Congress Catalog Card Number: 91-78336
Edited by Ruth N .- Kissell .

Designed by Linda Jorgensen-Buhman 1. Introc
Production supervised by Donna-Whitfield Butler- Jean I
Word processing by Brenda Blythe Wells
Graphics layout by Thomas Gatlin
Frontispiece by J. Massey
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
rsity
Contents
Figures ix
Tables XV
Acknowledgments 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 9
3. The Place of Hominids among Predators: Interspecific

Comparisons of Food Procurement and Transport
Mary C. Stiner 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 18. Densi

Two Strategies of Carcass Recovery Insigl
Alice M. Emerson 138 R. Lee
9. Bone Assemblages at Base Camps: A Further Consideration 19. Discu

of Carcass Transport and Bone Destruction by the Hadza Anne
Henry T. Bunn 156
20. Concl
10. Discussion: Subsistence and Settlement Interpretations
D.K.'
S.Ken
fames F. O'Connell · 169
III. Social Interaction: Food Sharing, Preparation Contr

and Consumption, and Questions of Gender
and Ethnicity
11. Gaps in the Zooarchaeological Analyses of Butchery: Is

Gender an Issue?
Diane Gifford-Gonzalez 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 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 342
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 349
169
Contributors 353
181
200
228
247
263
273
301
Figures
2-1. Predatory strategies concerning number of predators

participating in hunt and n~mber of prey taken 11
3-1. Geographic distribution of four Middle Paleolithic

(Mousterian) cave sites in west-central Italy 40
3-2. Examples of the two anatomical representation patterns

for medium ungulates in Italian Middle Paleolithic
cave sites 41
3-3. /Anatomical patterns for medium ungulate remains in

carnivore- and hominid-generated shelter faunas 47
3-4. The proportion of head and horn parts to limbs, (H+H) /L,
for medium ungulate remains by predator 48
3-5. The relative frequenCies of horn parts (HORN /L) to head

parts (HEAD /L) for medium ungulate remains in predator-
collected faunas 49
3-6. Comparison of (H+H)/L medians and ranges for predator

series using a linear ratio axis 51
3-7. Comparison of CH+H)/L medians and ranges for predator

series using a logged ratio axis 51
3-8. 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
ix
xI Figures
4-5. Seasons.:.of-death of red deer as determined by ~ementum 8-8. E

annuli analysis 74 n
4-6. Mixed age profile (an overlay of attritional and prime age 8-9. F
patterns?) of red deer 75 p
5-1. Relative representation of all elements 88 8-10. F
p
5-2. Relative representation of selected element$ 90
9-1.' F
5-3. Average fragment length by element 92 e
6-1. Ache camp, 9 January 1985 105 9-2. R
e
6-2. The Buchu site 110
9-3. R
7-1. The Kua research area in the Republic of B'otswana 119
9-4. R
7-2. Four bar graphs illustrating complementary kill/butchery
site and campsite bone assemblages 124
9-5. R
7-3. Average element abundances plotted against Standardized e.
Food Utility index for Kua gemsbok kill sites and base camp
assemblages 129 11-1. A
v
7-4. Relationship of skeletal completeness index to kill site
processing time 130 11-2. 1v
a·
8-1. Comparative general utility model values for bison with
proximal and distal limb units evaluated separately 141 11-3. 1v
0·
8-2. Comparative general utility model values for bison with

limb units evaluated as whole elements 142 12-1. A
8-3. Total fat utility models for bison 143 12-2. c

li
8-4. Skeletal fat models for bison 143
13-1. A
8-5. Marrow fat models for bison 144 a
8-6. Frequency of unit transport as indicated by scalogram 13-2. SJ

analyses for (a) alcelaphine antelope and (b) buffalo 146 m
8-7. Frequency of unit transport as indicated by scalogram 14-1. L1

analysis for giraffe 147
14-2. P:
cc
Figures I xi
8-8. Relationship between bison skeletal fat yield and a proxy
74 measure of relative transport cost 148
8-9. Relationship between bison nonskeletal fat yield and a

75 proxy measure of relative transpprt cost 149
88 8-10. Relationship between bison bone grease fat yield and the
proportion of grease to skeletal fat yield 151
90
9-1. Representation in percent (%MNU) of different skeletal
92 elements of all taxa transported to base camps by the Badza 159
105 9-2. Representation in percent (%MNU) of the different skeletal
elements of zebra transported to base camps by the Badza 160
110
9-3. Representation in percent (%MNU) of zebra bones at SN-A 163
119
9-4. Representation in percent (%MNU) of impala bones at SN-A 164
y
124 1 Representation in percent (%MNU) of different skeletal
9-5.
elements of all taxa at SN-A 165
?d
mp
129 11-1. A visual model of the movement of faunal remains through
various sites 191
130 11-2. Model showing animal processing activities likely to occur

at each locality type 192
141 11-3. Model showing bone modifications and discards likely to

occur at each locality type 193
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
143
13-1. Animal body parts accumulated by Okiek households over
144 a six-month period 236
13-2. Spatial distribution of the faunal remains of a widely shared

146 animal and a little-shared animal from a !Kung camp 240
14-1. Location of the study area in the Ituri Forest, Zaire 252
147
14-2. Plan map of an Efe campsite showing the primary structural
components 253
I
xii Figures
14-3. Plan map of a Lese village showing the primaty ~tructural 17-5.
components 254
14-4. Plan map of an elephant processing site showing the primary 17-6.
structural components and discarded bones 257
16-1. Comparison of the rank order of 29 skeletal parts based on 17-7.

Binford's modified general utility index and standardized
grease index 277
16-2. Proportion of long bone fragments that bear at least one 17-8.
tooth mark 281
16-3. Size-specific destruction/ deletion of long bone epiphyses 284

17-9.
16-4. Relationship between two-measures of epiphyseal fragment
deletion and destruction 285
17-10.
16-5. Relationship between the extent of epiphyseal removal
and the proportion of shaft £ragmen ts that bear at least one 17-11.
tooth mark 287
16-6. Qualitative model of the degree of competition among 18-1.

scavengers of assemblages of hammerstone-broken long
bones 288 19-1. ...c
~
16-7. Relationship between bulk density and the sequence by
which skeletal parts of sheep were observed to be removed/ 19-2. E
from 17 Berkeley simulated sites by spotted hyenas 289 l
16-8. Pre- and post-ravaging MNE estimates for 33 experimental 19-3. ~
assemblages at Berkeley 290 c
c
17-1. Plot of the relationship between MNI and the actual number
. of individuals observed ethnographically 306
17-2. Comparison of percentage contribution of each taxon based

on MNI and actual number of individuals observed
ethnographically 308
17-3. Plot of the relationship between MNI survival and carcass

size as measured by live ~eight 308
17-4. 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
2-1. Examples of Hunting Strategies for Selected Social Mammalian

Predators 14
2-2. 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
3-3. , Results of Rank-Ordered Comparisons of Predators 52

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. Variation in the Degree to Which Meat Is Shared among

Contemporary Hunter-Gatherers 229
13-2. Distribution of Animal Carcasses Through Transport and ·

Sharing 234
13-3. Food Sharing and Contrasts Between Subsistence and Social

Organization in Contemporary Hunter-Gatherers 242
16-1. Composition of the Experimental Sample 278

sponsore1
17-1. Comparison of :MNI and Actual Number of Individuals 307 Investiga~
have con·
17-2. Comparison of NISP and Actual Number of Individuals 311 tion with
are Georl
17-3. Actual Number of Elements Versus MNE for Four Taxa 315 Center; B
director c
17-4. Survival per Element and Part for Four Taxa 316 tion; Tom
Joan Con:
18-1. Correlation Coefficients Between Skeletal Part Frequencies The staff1
and Bone Density and the MGUI 329 the logist

other cor
18-2. Correlation Coefficients Between Percent Weight Loss of willingne
Carnivore-gnawed Long Bone Ends and Bone Density 332
18-3. Summary Distribution of Correlation Coefficients for 20 New

Assemblages and 67 Previously Analyzed Assemblages 333
18-4. Correlation Coefficients Between Bone Frequencies for Bqvid

Size Classes and Bone Density 335
18-5. Classification of 20 Correlation Coefficients Between Bone

Density and Small and Large African Bovids 336
19-1. Variables Known to Affect Skeletal Part Frequencies 345

229
234
cial
Acknowledgments
242
278 This volume, and the conference from which it developed, is

sponsored by the Visiting Scholar Program of the Center for Archaeological
307 Investigations, Southern Illinois University at Carbondale. A great many people
have contributed to the smooth operation of the program during my associa-
311 tion with it, and I gratefully acknowledge their support. Included among them
are George Gumerman and Don Rice, the past and present directors of the
315 Center; Brian Butler, associate director; Kim Smiley, curator; Patrice Teltser,
director of publications; Donna Butler and Brenda Wells, publications produc-
316 tion; Tom Gatlin, graphics layout; Carolyn Taylor, administrative assistant; and
Joan Corse, Center secretary. Ruth Kissell did an exceptional job as copy editor.
es The staff.bf the Division of Continuing Education helped in essential ways with
329 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 experimen-
tally or ethnoarchaeologically, to better understand the links between particu-
lar 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 anal-
ysis they use to address behavioral questions, incorporating studies of denti-
tion, 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 mea-
sures 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 ~-
I
~
This volume, like the conference from which it _came, is structured around lil Part Ill
four broad themes. The first three concern aspects-of cultural behavior of
interest to archaeologists: subsistence strategies, settlement patterns, and ~
I
between r
field, the s
patterns of social interaction. The fourth theme concerns the impacts of certain ! action of·
nonhuman behaviors and processes that can affect our ability to see the
cultural behaviors that interest us as archaeologists. ·
I ·patterns o
_,~ has been a
Part I focuses on several aspects of the interpretation of subsistence strate- less attent
gies, including methods of meat acquisition, such as scavenging and a variety I ·conzal~z.
of hunting techniques, differences between foraging and collecting strategies, Gifford-
the role of cooperation in certain types of hunting, considerations of the skills preparatio
represented by different hunting strategies operating at various stages of likely to h
hominid evolution, adaptations to temperate climates and seasonal variations that a foct
in prey resources, and issues of anatomical utility. earlier stal
Steele and Baker raise the question of what aspects of human hunting are port, wouJ
uniquely human and of how they might relate to the evolution of a successful impacts o:
adaptive strategy for hominids. Stiner looks at models for the meat acquisition terns, sugl
strategies practiced by Neandertals and anatomically modern humans during influence i
the Middle and Upper Paleolithic in Italy, suggesting seasonal variation in the breakage I
role of scavenging in a temperate climate. Pike-Tay and Knecht combine two and cookii
different types of experimental data, one concerning bone tool manufacture sharing or
and the other concerning dental evidence for season of death and age struc- the distanc
ture of the prey population, to model hunting strategies in the Upper tion with 1
Paleolithic of France.~ with the aim of locating these strategies on the forager- duce patt
collector continuum. Enloe takes a different approach to a similar problem, between k
using controlled data on long bone fragmentation to differentiate between data on an
forager-like processing for immediate return and collector-like bulk process- them fora~
ing of accumulated reserves. · between st
Part IT deals with the interpretation of settlement patterns from skeletal part in the patt
profiles, seasonality measures based on tooth eruption and wear/ and charac- the faunal
teristics of site size and spatial distributions ·of bone. These archaeological Part IV
clues are evaluated as evidence for site function and duration of occupation. influence 1
Jones uses ethnoarchaeological data from Ache foraging camps to examine both wild
patterns associated with very short-term occupations, including small site taphonom:
area, clustering of discarded bone near hearth areas where almost all activities Blum en:
took place, and comprehensive representation of skeletal elements reflecting · captive hy
the lack of logistical transport. Bartram reviews Kua transport patterns for carnivores
gemsbok kills, noting that utility models alone do not account for variability importanc
in carcass treatment and that processing strategies, particularly those that epiphyses
remove meat from bone prior to transport, must be considered when using and midsh
skeletal element profiles to interpret site function. Emerson examines the use related pr1
of various types of anatomical utility indices to interpret transport strategies camps, sul
from skeletal part profiles, focusing on single kills of large game and the role :rvt:l\TEs que
of fat in processing and transport decisions. Bunn reviews the varying trans- tically relic
port patterns documented for the Badza, particularly as they relate to the can be dra
relative abundance of axial and appendicular elements in base camps, sug- tic dogs S<
gesting that quantity and quality of marrow may play an important role in ethnoarch'
decision making. frequency
~~
]
l
:Ji,
'il·
Jj
,_,il:
jjl
Introduction I 3
tured around Part III deals with aspects of social interaction, including the relationship
Ll behavior of between meat processing in residential camps and meat processing in the
patterns, and field, the sharing of meat between members of different camps, and the inter-
1acts of certain action of two ethnic groups at a kill site. These papers focus attention on
ity to see the patterns of behavior that act to modify a~d redistribute bone after the game
has been acquired and transported. Such patterns have traditionally received
istence strate- less attention in interpretations of faunal remains, as pointed out by Gifford-
; and a variety ·Gonzalez.
ing strategies, Gifford-Gonzalez questions our apparent neglect of the final stages of meal
1s of the skills preparation and consumption, suggesting that they are the behaviors that are
~ous stages of likely to have the greatest effect on patterns recovered archaeologically and
mal variations 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 trans-
n hunting are port, would be productive. Oliver addresses this topic in his discussion of the
)f a successful impacts of Badza butchering and cooking practices on bone breakage pat-
~at acquisition terns, suggesting that carcass size and the cooking techniques available will
umans during influence the degree to which bone is fragmented and that greater attention to
ariation in the breakage patterns may allow us to trace the development of both butchering
t combine two and cook,ing techniques through time. Marshall considers the impacts of meat
I manufacture sharing 16n skeletal part distribution between Okiek households, noting that
:md age struc- the distances between households, typically a kilometer or more, in combina-
in the Upper tion with differences in hunting success and patterns of sharing, act to pro-
m the forager- duce patterns similar to those traditionally associated with differences
1ilar problem, between kill sites and residential camps. Fisher. presents ethnoarchaeological
~tiate between data on an elephant butchery camp used jointly by two ethnic groups, one of
bulk process- them foragers, the other hortkulturalists; material evidence of the interaction
between societies with different patterns of mobility and food storage is seen
n skeletal part in the patterning of hearths and structural remains as well as the structure of
tr, and charac- the faunal assemblage.
uchaeological Part IV focuses on two particular types of nonhuman phenomena that
occupation. influence the composition of faunal assemblages: destruction by carnivores,
ps to examine both wild and domesti~, and density-mediated attrition. Both represent
.ing small site taphonomic processes of major importance in many archaeological sites.
st all activities Blumenschine and Marean use experimental observations on wild and
ents reflecting captive hyenas to improve our understanding of how ·scavenging by wild
rt patterns for carnivores impacts bone that has been processed by hominids; they stress the
for variability importance of collecting data on the placement of tooth marks, the ratio of
.rly those that epiphyses to shafts, and the distinction, in skeletal part profiles, of epiphyseal
d when using and midshaft fragments. Hudson uses ethnoarchaeological data to address a
Lmines the use related problem, the impacts of resident domestic dogs in modern forager
port strategies can1ps, suggesting that destruction of bone by Aka dogs renders the use of
Le and the role MJ\JEs questionable, that taxonomic ranking by MJ\JI and NISP remains statis-
varying trans- tically reliable, and that prey species with live weights of less than a kilogram
If relate to the can be dramatically underrepresented archaeologically in sites where domes-
e camps, sug- tic dogs scavenge bone. Lyman uses a collection of 87 archaeological and
Jortan t role in ethnoarchaeological assemblages to assess the extent to which skeletal part
frequency is predicted by bone density alone, finding density a reasonable
41 J. Hudson-
predictor in almost half the cases. Theoret
Several interpretive concerns and analytic-approaches crosscut the thematic best differ
groupings of the papers. Those concerns operate at two levels. One is ent types c
methodological-how to confidently read the behavior of interest from the particular
archaeological remains. The other is theoretical-how did early hominids differenti.:
successfully evolve into modern humans? sometime:
Analytically, three aspects receive attention by several authors: the identifi- although:
cation of limb bone shaft fragments; the interpretation of skeletal part profiles; both anci1
and the choice of appropriate scales of analysis. The importanc~ of identifying yielded m
shaft fragments is brought out in the papers by Bartram, Blumenschine and Throug:
Marean, Bunn, Hudson, Lyman, and Oliver. Multiple causes, both human and motivate(
nonhuman, act to preferentially destroy articular ends. Faunal analyses that served to
ignore shaft fragments are likely to underrepresent the importance of appen- concerns,
dicular elements, biasing skeletal part profiles and potentially influencing points ou1
measures of taxonomic abundance as well. optimal f(
In the papers presented here, skeletal part profiles are used to interpret a being opt]
wide range of cultural behaviors, including meat transport, site function, and bene1
sharing, trade, and hunting success, and to document noncultural agents of conferenc
bone destruction such as carnivores and density-dependent attrition. The collector '
current diversity of opinions about associations between particular profiles opment oj
and particular behaviors and destructive agents suggests that we are still ing, as we:
exploring and defining the critical variables and that problems of equifinality Graysm
have yet to be resolved. whether o
The importance of choosing a scale of analysis appropriate to the research highly pa:
question is discussed by several participants, including Blumenschine and made exr
Marean, Bunn, Hudson, Lyman, O'Connell, Oliver, and Stiner, although there hypothesi:
is no consensus about which scale might be the most appropriate. tales that~
Blumenschine, Marean, and Lyrrtan advocate more detailed approaches to part of sci
skeletal part profiles, with attention to fine-scale variations in bone' density, as often fails
they may represent the real cause of differences in preservation ahd identifia- very simi]
bility. Stiner and Hudson, on the other hand, suggest that more robust mea- case of as:
sures, produced by lumping elements into broader body part categories, Behrensm,
might be more useful archaeologically. Comments by Bunn, O'Connell, and science is
Oliver raise questions about when body part analyses should focus on hope for u
particular species and when the analyses should group species into size logically.
classes. The Hadza data suggest that the use of size classes rather than taxa as One of 1
the analytic ~ategory is advantageous because it increases sample size but can insights be
obscure potentially significant differences in the way particular species are The pape1
treated. ideas abot
;j became cl
Sample size, in terms of the number of controlled assemblages available for
any given study, is a concern for many authors, particularly those working which to:
"I experimen
with ethnoarchaeological data. Case studies are presented as opportunities for
insight, and tests of the hypotheses generated against additional data sets are better sen
recommended. Stiner points out that a logical next step in the application of I
frameworl
small-scale, fine-grained ethnoarchaeological studies to archaeological prob- :,I extra step
lems will be the combination of multiple assemblages to simulate the coarser :! can be tes1
':
scale of archaeological signatures. ;[ situations.
;I
:t
H
:J
I
m
m
m
II:!:
Introduction I 5
Theoretically, these papers are characterized by repeated interest in how to

.t the thematic best differentiate the roles of scavenging and hunting, how to identify differ-
~vels. One is ent types of hunting and the kinds of cultural behaviors associated with them,
~rest from the particularly as they relate to cooperation and planning skills, and how to
trly hominids differentiate foragers from collectors. There is also repeated reference, albeit
sometimes oblique, to theoretical models 'of optimization. Common to many,
s: the identifi- although not all, papers is an underlying assumption that hunter-gatherers,
1 part profiles; both ancient and modern, used acquisition and processing strategies that
of identifying yielded maximum meat or fat while minimizing labor effort or risk.
tens chine and Throughout the conference's discussion sessions, comments were often
thhuman and motivated by ethnoarchaeological experience, and such comments often
analyses that served to push theoretical frameworks beyond existing assumptions and
mce of appen- concerns, advocating more critical evaluations of popular models. O'Connell
.y influencing 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
to interpret a being optimized and we typically lack the data to measure the relevant costs
site function, and benefits. Both O'Connell and Jones question, as did Yellen during the
:ural agents of conference's final roundtable discussion, the usefulness of the forager-
attrition. The collector distinction when applied to evolutionary models about the devel-
icular profiles opment pf human abilities to plan and cooperate, noting that modern forag-
at we are still ing, as well as modern collecting, regularly incorporates those skills.
of equifinali ty 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
o the research highly particularistic and that frequently their theoretical importance is not
1enschine and made explicit. Watson and others respond that the scientific method of
:tlthough there hypothesis testing requires particularistic studies and that even the cautionary
appropriate. tales that seem to characterize much ethnoarchaeological work are an essential
approaches to part of scientific progress. Styles. points out that ethnoarchaeological research
Jne density, as often fails to take into account the fact that different processes can produce
L and identifia-
very similar results and that ethnoarchaeology, when it stops with the first
~e robust mea- case of association of behavior with material remains, is stopping too soon.
ut categories, Behrensmeyer also notes this habit of oversimplification and observes that the
>'Connell, and science is still young. She charts the trends we can expect as we balance our
>uld focus on hope for understanding against the reality of the data available to us archaeo-
~cies into size logically.
er than taxa as One of the ai:rns of the conference and of the resulting volume was to share
1le size but can insights based on the richness of ethnoarchaeological and experimental data.
lar species are 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
~s available for became clear in the periods of open discussion that we see many ways in
those working which to improve our use of the faunal record and the· types of data that
portunities for experimental and ethnoarchaeological research methods can produce. We can
l1 data sets are better serve the archaeological community by making explicit the theoretical
application of frameworks and the behavioral questions that concern us and by taking the
~ological prob- extra step beyond particularistic descriptions to build analytic methods that
ate the coarser 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 con-
cerned 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 archaeo-
logical bones to cultural behavior.
I.
1oth those con-
other 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 hunt-
ing. However, there has not been a clear distinction in the anthropologi-
cal 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 per-
taining to human hunting and predatory strategies quickly reveals two direc-
tions 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 incorpo-
ration 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

9
10 I D. G. Steele cmd B. W. Baker
live game (Bartholemew and Birdsell 1953; Dart 1953; Isaac 1976, 1978, 1980, Numt
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 relation-
ships of predation and scavenging. To date, analysts have emphasized that
human predation involves tool using, social hunting, differential labor, coop-
eration, 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, improved t:
researchers have emphasized that humans,·unlike many predators, have an and theyw
omnivorous diet consisting of both plant and animal matter that makes sub- As an ex,
sistence under marginal conditions possible and of value (Wolpoff 1980). ing from 0
Of the behavioral features that have been taken as hallmarks of human had succes;
predation, the social aspect of human hunting has been one of the first recog- behavior o
nized by scholars, the most intriguing, but at the same time one of the most 1970, 1972J
difficult features to clearly recognize in the archaeological record or character- contention ·
ize as distinctly human. The r~ason for the difficulties in distinguishing Curio, fa:
human social hunting patterns from the social hunting patterns of other and attack€
predators is that two distinct hunting strategies are commonly being consid- ally. In thj
ered simultaneously and referred to under a variety of names Stich as social gazelles we:
hunting, communal hunting, communal drives, and cooperative hunting. zebra was:
The two hunting strategies we consider separately are social hunting and groups rest
multiple predation. Of these two strategies, multiple predation is the less made it po:
understood and is potentially one of the most distinctive hallmarks of human supported
predation. Figure 2-1 illustrates the distinctiveness of the two strategies con- most likely
cerning how many predators are involved versus how many prey specimens by itself." ·
are taken. The figure also indicates the relationship of the two phenomena. In reviev
human sod
the primar·
Early Views on Social Hunting prey could.
sees the evi
One of the most complete reviews of predatory behavior, which horses, bisc
unfortunately does not include an analysis of human predatory behavior, is larger than
that of Curio (1976). While most of the nonhuman predators reviewed by major facto
Curio hunt as single predators, 'he also reviewed an array of social predators Since few o
as diverse as spiders and lions. Among the social hunters, hunting of single of larger p
prey specimens during one hunting episode was the rule. The hunting of considered,
single prey provided social predators with certain advantages: social hunters among hun
Multiple Predation 111
76, 1978, 1980, Number of Predators Hunting Number of Prey Taken

)rd 1981, 1984;
182; Bunn and
Single Hunter ~ Single Specimens
.ressed in this
about human .
, of the human Social Hunting Group Multiple Spec1mens
man foraging
o the issue of
y function and
Sequential Predation Mass Predation
)gical relation- of Single Specimens in Singl·e Episode
lphasized that in Separate but
Associated Episodes
.allabor, coop-
!ms, utilization
Figure 2-1. Predatory strategies concerning number ofpredators partici-
nption of prey pating in hunt and number of prey taken.
·e 1968; Marks
, 1984a, 1984b;
Additionally, improved their success ratio when hunting in groups rather than individually,
lators, have an and they ,were able to hunt bigger prey.
1at makes sub- As an1example, single hyenas hunting wildebeests had success rates vary-
,off 1980). ing from 0 to 18% of the time, while two or more hyenas hunting wildebeest
1r ks of human had success rates varying from 32% to 100% of the time, depending on the
the first recog- behavior of the wildebeest upon confrontation with the predators (Kruuk
'ne of the most 1970, 1972). Bertram (1979) provided similar evidence that supported the
~d or character- contention that social predators tend to have higher success rates.
distinguishing Curio, following Kruuk (1970), also pointed out that spotted hyenas hunted
tterns of other and attacked larger prey when hunting socially than when hunting individu-
{ being consid- ally. In this example, the average hyena group size hunting Thomson's
; such as social gazelles was 1.2 individuals, while the average.hyena hunting group size for
e hunting. zebra was 10.8 individuals. Thus, for nonhuman social predators, hunting in
al hunting and groups resulted in more efficient capture of individual prey specimens and/ or
tion is the less made it possible to captu,re specimens of larger prey. This view was recently
1arks of human supported by Packer and Ruttan (1988:189) who stated that "cooperation is
strategies con- most likely when an individual has a low probability of capturing a large prey
Jrey specimens by itself."
)henomena. 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,
ehavior, which horses, bison, bear and a wide variety of ungulates (all animals substantially
)ry behavior, is larger than humans), it is impossible to deny that social hunting has been a
rs reviewed by major factor in making it possible for humans to prey upon those animals.
;ocial predators Since few other advantages to social hunting in humans other than the taking
1nting of single of larger prey species or the enhancement of hunting success rates were
fhe hunting of considered, the traditional conclusion was that the evolution of social hunting
;: social hunters 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 not defined
that may have been·a scavenger and/or an opportunistic predator analogous prey and pr
to the modern common chimpanze~ (Wolpoff 1980). With the adaptation to a game sped
more open savanna grassland environment, a segment of this lineage evolved where a nu
into the Hominidae and developed a greater reliance on ~eat, acquired either number of'
through predation (Isaac 1978, 1983; Wolpoff 1980) or scavenging (Binford haps with o
1981, 1984; Blumenschine 1986a, 1986b, 1987, 1988a, 1988b, 1989; Potts 1982, To faciliti
1984a,.1984b; Shipman 1983, 1986). With the evolution of Homo erectus, the cooperative l
adaptive shift was assumed to be toward a more.structured social hunting of act of two o
single individuals of larger prey (Binford 1981; Shipman et al. 1981; Shipman emphasis is
and Walker 1989; Thompson 1979:135-179; Washburn and Lancaster 1968:297; no refereno
Wolpoff 1980). (1976:199) c
To recapitulate the traditional sequence of hypotheses and correlated predators.
assumptions concerning human predatory patterns: (1) social hunting has In additic
.involved predominantly the hunting of solitary prey, and (2) as social hunting added emp
evolved, or the social hominids undertook social hunting, it primarily pro- communal]
vided the advantage of preying on larger species or enhancing the success rate hunting," iJ
of the hunt. · hunting grc
In a large measure, because both of the hypotheses and assumptions either cooperative
were based on at least some eviden·ce or followed a logical sequence of two ormon
assumptions, it was, until recently, difficult to consider alternative or com- Research
plementary hypotheses. Within recent years, researchers began to consider relatively 1
ecological models and optimal foraging theory as an explanation why human apparently
and nonhuman predators hunt the way they do and how such practices may Examples c
have affected human evolution (cf. Driver 1990:21-28; Earle and Christenson fishes (Cur:
1980; Hawkes 1990, 1991; Hawkes et al. 1982; K. Hill 1982, 1988; Hill et al. Murie 1944
1987; Jochim 1976:19-45; Jones 1990:70-71; Kamil et al. 1987; Kamil and (Teleki 1972
Sargent 1981; Packer and Ruttan 1988; Shipman and Walker 1989; Lee and De
Slobodchikoff 1988; Smith 1980, 1983; Stephens and Krebs 1986; 'furner 1989; human sod
Webster and Webster 1984; Winterhalder 1987; Winterhalder andiSmith 1981). In Curio'
Such models emphasize the maximization of benefits versus costs in the pro- except one,
curement of food. single prey
predators h
social hunti:
Human Predatory Strategies hunting its'
killed. The '
By separating the question of how many predators participated in single hunt
the hunt from the question of how many prey were taken during a hunting practice, thE
episode, it permits us to focus on the issue of how adequate animal resources instances. 1
are acquired rather than on the social behavior of the hunt. Anthropologists predation.
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 interpre-
tive framework.
Saunders (1977) for example, noted that entire families of mammoth may
have been hunted prehistorically, in addition to the taking of single individ- social organ
uals by communal efforts. Driver (1990:12) stated that communal hunting is as a separa1
qr··
Multiple Predation j13
~arly hominoid not defined by the number of prey taken but may involve varying numbers of
ator analogous prey and predators. In addition, Kooyman (1990:330) noted: "Hunting any big
:tdaptation to a game species can be accomplished using two basic techniques: communal,
lneage evolved where a number of hunters coordinate their efforts (and usually procure a
:tcquired either number of animals), and individual, where hunters work in isolation (or per-
1ging (Binford . haps with one or two other hunters), usually to capture a single prey animal."
'89; Potts 1982, To facilitate the separation of these phenomena, the phrases social hunting,
mo erectus, the cooperative hunting, and communal hunting are used here strictly to refer to the
,cial hunting of act of two or more predators actively pursuing or killing prey. In this vein, the
1981; Shipman emphasis is on the number of predators involved in the act of hunting, with
:aster 1968:297; 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
:md correlated predators.
al hunting has In addition to the number of hunting participants, Frison (1987:182), placed
; social hunting added emphasis on aspects of planning and temporary authority in defining
primarily pro- communal hunting. Hayden (1981:421) distinguished "small-scale cooperative
the success rate hunting," involving two to four individuals, from a "large-scale communal
hunting group," involving five or more individuals hunting in a coordinated,
1mptions either cooperative fashion. Driver (1990:12) viewed communal hunting involving
al sequence of two or mbre cooperating hunters with a preconceived plan.
'native or com- Reseafch on nonhuman predators has revealed that there are in fact a
~an to consider relatively large number of social hunters and that social predation has
on why human apparently evolved independently in a wide variety of animals (Table 2-1).
1 practices may Examples of social hunting can be seen among spiders (Wilson 1972) and
nd Christenson fishes (Curio 1976:210) and among mammals such as canids (Mech 1970;
L988; Hill et al. Murie 1944), hyenaids (Kruuk 1970, 1972)~, felids (Schaller 1972), pongids
'87; Kamil and (Teleki 1973), and hominids (Coon 1971; Davis and Reeves 1990; Forbis 1978;
Walker 1989; Lee and Devore 1968; and Nitecki and Nitecki 1987 are general sources on
:6; Turner 1989; human social hunting with extensive bibliographies).
nd Smith 1981). In Curio's (1976) discussion of social hunting, all examples presented,
~osts in the pro- 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
: participated in single hunting episode either may be a single animal or many animals. In
1ring a hunting practice, the majority of social hunters do kill single animals, but not in all
nimal resources instances. These exceptions are highly significant to understanding human
~nthropologists predation.
r of prey versus
mgh none have
sin an interpre- Multiple Predation
mammoth•.may By restricting the definition of social hunting to the question of the

f single indi vid- social organization of the predator, it i~ possible to more thoroughly examine
mnal hunting is 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
~
("I:)
Sequential Mass ("I:)
Single Prey Predationa Predationa Prey Sizeb Author Labelc Reference ~

Taxa ~
;::::
:;.::;...
Order Carnivora S,M,L Cooperative Macdonald (1984:62) tJj
*
Family Canidae hunting ;E.
Canids tJj
Canis latrans ~
Coyote ~
-....:
Canis lupus S,M,L Pack hunting Anderson (1982);
Gray wolf * * Macdonald (1984:58-61);
Mech (1970); Sharp (1978).
Canis rufus S,M,L Pack hunting Anderson (1982);
Red wolf * * Macdonald (1984:5&-61);
Mech (1970); Sharp (1978)
Canis dingo S,M,L Pack or Anderson (1982);
Dingo * cooperative Macdonald (1984:84)
hunting ·
Canis aureus S,M,L Cooperative Macdonald (1984:64)
Golden jackal * hunting
Canis mesomelas S,M,L Cooperative · Macdonald (1984:64)
Silverbacked jackal * hunting
Cuon alpinus -...___
S,M,L Cooperative Anderson (1982);
Dhole, red dog * hunting Macdonald (1984:80)
Lycaon pictus · S,M,L Cooperative or Anderson (1982);

African, or cape hunting * ·organized Macdonald (1984:76-79)
dog hunting
Table 2-l.-continued
~~. ~~ ~~~~-~~ )~-~-·
Cuon alpinus S,M,L Cooperative Anderson (1982);

Dhokred dog hunting Macdonald (1984:80)
Lycaon pictus S,M,L Cooperative or Anderson (1982);

African, or cape hunting organized Macdonald (1984:76--79)
dog hunting
Table 2-1.-Continued
Multiple Prey
Sequential Mass
Taxa Single Freya Predationa Predationa Prey Sizeb Author Labelc Reference
Speothos venaticus s;M Pack or Anderson (1982);

Bush dog * cooperative Macdonald (1984:85)
hunting
Family Hyaenidae S,M,L Group hunting Kruuk (1972:89, 204);
Crocuta crocuta * * * and mass killing Macdonald (1984:154-
Spotted, or laughing 158)
hyena
Hyaena hyaena * S,M,L Group hunting Macdonald (1984: 154-
Striped hyena 158)
Family Felidae S,M None Anderson (1982)
Felis pardalis
Ocelot
Panthera leo S,M,L Communal, Macdonald (1984:29-35); a:::
Lion pride or Schaller (1972:251, 437) c
1--'
multiple ::r.
~
hunting ro-
Order Cetacea S,M Cooperative Macdonald (1984:169) 1-d
'"1
Family Delphinidae hunting ro
* * * 0..
Delphinus delphis Pl
::r.
Common dolphin 0
::s
Orcinus orca '>f-
S,M Cooperative Macdonald (1984: 190- -.........
Killer whale * * hunting 191) (.11
Table 2-1.-Continued ~
0\
>:J
Multiple Prey
0
Sequential Mass t:.l'l
.......
Taxa Single Prey1 Predationa Predationa Prey Sizeh Author Labezc Reference ('I)
~
('I)
- l':l
;::s
Family Monodontidae S,M Cooperative Macdonald (1984:201) ~
Delphinaptems leucas hunting, group lJ:!
Beluga, belukha, or white herding
whale
;E
tci
l':l
Order Primates S,M Simple Macdonald (1984:377);
~
Family Cercopithecidae cooperative Strum (1981 :262, 274-275) """'!
Papio cynocephalus hunting,

Savanna baboon coordinated
hunting
Family Pongidae S,M Simple Lawick-Goodall (1968);
Pan troglodytes cooperation, Macdonald (1984:424);
Chimpanzee cooperative Teleki (1973, 1981:323,
hunting 331-333); Wrangham
(1975)
Family Hominidae S,M,L Communal, Coon (1971); Davis et al.
Homo sapiens cooperative, 1986; Davis and Reeves
Humans social, and (1990); Forbis (1978); Lee
group hunting and Devore (1968);
Nitecki and Nitecki
(1987), and references
therein
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
t:L s=- ~ t;;· '"d rt re as s sr s
8 ~- i:'r ~ ~ s=- 0' B fti ~ ~ s=-
<~!TOro~oo~~H~ooo~~o~ororo~~'"d::roro~o..~~~ro
< OQ s· s s !:I0 &. s o.. ~- ~ s=- --o o s· o.. ~
0 ~~~o~oo@~OQ 0 n
~ ~ m ~ ~ ~ ro" ::r ~ ~ ~ Pl (ti ~ PJ 5" ~ •s ...
~ ~ 00 0 ~ ~ - .,..- ~· ~ (5" ~ !:Is· h3 2. 2· p ::t' ~ ~ 0..
1-'4 (i) PJ 8 p;·
::;::' OO
....,_. ~
o i:'r a...... '"d::r ro
~ 0 '"'
s
0.. oo
s· < ~ ...ro~ ~ ro • ~ ~ u '"d 0.. :=; n- ~ ~ 11> n- "" ~ 0..
Q...
;::l.
~
~ ~ ~
?\ o
Pl 11>
H ~ s ~ !T o.. ~ s:- ::r s 0..
...... !4. 1-j rt o
11> .......
o.. Ft ~ o.. Q. .r> ft g ro
>-+.
.-J
e ~ s::r. ::::· ::r. o ;r ~ '"""'"',
~ ~ OQ
H .......
...... .-J
~ ........ 11>
..... 00
8 OQ~ ·~ 11>
~ ....
,.J<:;
0~>-~->-->-+,·
0 '" 11>
~
~ ~
..... ~
1-'•0Q
00
0
11><>-t-,0•
.......
.............. ::::M ~ ~
..
~
0
11>
..... ('"\ -
>-1->00 ......
~ ?\
'" ..... 0
!T.-::::
~ - ~ ........
(t Pl ::=:
~-!:Is~~!:::!
1-f
0
0
~ ~· Pl
;::::>
~- ~
>-!->
- 11> .
~
,.J
11>
vuoo>-1"-\~
~ ~ ::s:
"'"
~ ~ s
11>
-::tro.-J
,.J
P>'
0 ~ .... ,.J
~ ~ ~ ~ 2.. --g § ~- s:- g- % § .g. e ~ 0' ffi ~ §..~ st- ~ ~ ~ ~- ~ ~ : ~ ~ -. ~ '"M ft ~ ~ E. g :! §- ~ _; ~ ri o.. ffi"
,.J
o..
social hunters during a single hunting episode. Curio's (1976) monograph

documented that the majority of nonhuman predators kill single animals dur-
ing a hunting episode. There are, however, instances when nonhuman preda-
tors do kill more than one animal (Table 2-1). We use the phrase "multiple
predation," following Steele (1982) and Packer and Ruttan (1988), to refer to
those instances when more than one prey specimen is taken in a relatively
short time during a single hunting episode.
There are two ways multiple predation can occur (Figure 2-1). Mass preda-
tion, one form of multiple predation, occurs when several prey are killed
.s
QJ
during a single hunting episode. Sequential predation, the second form of
1-<
QJ multiple predation, occurs when a sequential series of prey are taken during
-:8
different but associated episodes of hunting.
One of the best documented examples of mass predation practiced by a
nonhuman predator concerns the African lion. Schaller (1972:437) docu-
mented that lion prides have killed, in mass, wildebeest, zebra, Thomson's
gazelle, and buffalo. The number of animals killed during single episodes
varied from two to five animals (Schaller 1972:251, 254). In these instances, the
mass killing resulted from separate lions within the pride successfully killing
individual specimens during the final rush.
A patt¢rn of predation related to mass predation, though not the same, is
the pheiomenon of sequential predation. Schaller (1972) reported that lions
would kill additional single specimens during separate hunting episodes
while on a single foray ifopportunities arose. Mech (1970) and Sharp (1978:68)
reported that wolves under certain conditions took more specimens than they
consumed during a sequence of separate hunting episodes while on a single
foray. This occurred when lead wolves chasing a herd of caribou outdistanced
the rest of the pack, left freshly killed carcasses, and continued the pursuit.
Mech interpreted this as a mechanism for the lead wolves to supply the
slower, nonhunting wqlves of the pack with fodd. Thi3 pattern of multiple
killing differs from mass predation as defined here in that sequential preda-
tion involves killing single specimens during different hunting episodes while
on a single foray.
By this definition, human hunters have practiced sequential predation if
more than one animal is taken during a single hunting foray. Since such a
broad definition of sequential predation would seem to include many, if not
most, human hunting forays, it raises the question of the heuristic value of the
term. We feel this broad definition emphasizes the ubiquity of taking more
game on a hunting foray than the human hunter can consume, and this is the
purpose of the strategy. On the other hand, we find in practice mass predation
is the more definable an¢, identifiable form of multiple predation and is an
easier phenomenon to examine. To provide a more operational definition of
sequential predation, we reserve the use of the term for those instances when
the taking of more than one prey is a clear-cut strategy and the kills are taken
close to one ano~her in time. Examples of such sequential predation would be
when one hunter or a few hunters intercept a caribou herd and from a single
stand kill several animals; a similar stand by a bison hunter who would kill
several animals in a short time would be an example of sequential predation.
18 I D. G. Steeleand B. W. Baker
While multiple predation, including mass and .sequential predation, has However
been documented in lions and wolves, mul.tiple predation, particularly mass tic ally take~
predation, of larger prey is in fact a. rare predatory strategy among nonhuman occur when
mammalian predators. Of the terrestrial nonhuman mammalian social preda- surround, c
tors, multiple predation of large prey has been identified only in three taxa as predation i~
far as we are aware: the wolf, the lion, and the hyena. Further, it is undocu- is undoubtE
mented among solitary predators. tunistic eve
Among human hunters, multiple predation is commonly documented, par- hunts is th:
ticularly mass predation, the more recognized a~d elaborate form of multiple animal reso
predation. Several researchers (Anel11960, 1969; Coon 1971; Forbis 1978) pro- Whenmt
vided comprehensive surveys df human hunting methods involving _mass is considen
kills. Waselkov (1978), in addition, examined mass predation of deer in detail. predation i~
All of those researchers, however, considered social hunting and mass killing on an incre
as inseparable events and focused their interests on the social aspect of the conditions (
hunters. · · bison, Asiar
Forbis (1978), for instance, stated that the focus of his study was largely elephants, c
restricted to hunting methods ·involving large. numbers of humans joined taken in epi
together to secure herds of land mammals. In particular, he was seeking his reviews
parallels to the comparatively recent bison drives of the North American (1960, 1969:
plains (Forbis 1978:3). Coon (1971) also considered social hunting and mass Melanesia, r.
killing as inseparable phenomena. Using the Efe pygmies of the Ituri Forest of as being tak
Africa as an example, he noted that during most of the year stalking individ- Coon (19:
ual animals by two or more hunters was the common form of hunting. Once a monkeys, sc:
year, however, band members would unite to participate in communal hunts ple predatic
with the express purpose of killing large numbers of game during single hunt- whales, and
ing episodes (Coon 1971:88). Coon then proceeded to review similar hunting Additionall-
episodes involving mass killing in other human groups. mass, while
Structurally, such communal hunts were considered by Forbis (1978:3) to be turkey. Furi
like military formations requiring tight community organization, ~inbued with Scott (198C
magico-religious sanctions governing both the hunters' and the preys' behav- rhinocerose
ior. The massive communal hunts were also commonly associated with com- North Arne
plex hunting devices, such as nets, traps, fences, weirs, pits, or the use of fire. perhaps m;
Most recently, Nitecki (1987:5) stated: 'We assume that man was first a episodes.
solitary hunter and gatherer, and that human evolution progressed toward In additic
the later-developed high degree of social cohesion and cooperation that was a · killed in m:
necessary precondition of massive purposeful hunting." Thus, Nitecki numerous i
assumed that social cooperation is a necessary requisite to the taking of vertebrate lc:
animals in mass. accoutremeJ
The consideration of mass predation in humans only in the context of However, in
communal hunts analogous to the bison drives of North America and the net ing the larg
hunts of the Ituri Forest pygmies has obscured the broad use of multiple Consequent:
predation by human hunters irrespective of the social organization of the for all kinds
hunters. We prefer to consider multiple predation the more common preda- hunting, fis
tory strategy used by humans, and the social hunt, or communal hunting obscures thE
involving mass predation, as but one specialized form of mass predation, pattern of p:
specifically, the form of mass predation by humans that is best documented animal resm
and recognized. efficient a m
predation, has However, mass predation also occurs when a single hunter opportunis-
~ticularly mass tically takes two animals, a female and her offspring, for instance. Or it can
mgnonhuman occur when a group of hunters during a day-to-day hunting episode is able to
n social preda- surround, drive, or corner more than one specimen. In those instances mass
n three taxa as predation is a predatory strategy practiced as opportunities arise and one that
~, it is undocu- . is undoubtedly eagerly sought. The common feature of these simple, oppor-
tunistic events of mass predation and the large, well-organized communal
:umented, par- hunts is that both are designed for, and result in, the greatest recovery of
rm of multiple animal resources possible given the circumstances.
~rbis 1978) pro- When multiple predation-both mass predation and sequential predation-
lvolving mass is considered irrespective of the social behavior of the hunters, multiple
: deer in detail. predation is seen as a virtually ubiquitous phenomenon practiced by humans
td mass killing on an incredibly diverse number of species and under a wide variety of
1. aspect of the conditions (Table 2-2). Forbis (1978) identified grasshoppers, rabbits, caribou,
bison, Asian gazelles, guanacos, rheas, pumas, rats, African antelope, buffalo,
ly was largely elephants, deer, horses, mountain sheep, and Asian elephants as having been
t umans joined taken in episodes of multiple predation (usually mass predation). Similarly, in
2 was seeking his reviews of communal hunting in North America and Australia, Anell
)rth American (1960, 1969) identified a wide variety of marsupials in Australia, pigs in
.ting and mass Melanesi~, and rabbits, deer, bison, caribou, and pronghorn in North America
~ Ituri Forest of as being taken in episodes of multiple predation.
:1lking individ- Coon (1971) further reported records of mule deer, sika deer, Old World
unting. Once a monkeys, sambur deer, axis deer and African elephants being taken in multi-
mmunal hunts ple predation episodes. Nelson (1969, 1973) reported the taking of ducks,
:1g single hunt- whales, and seals in mass in Alaska by Eskimo and North American Indians.
imilar hunting Additionally, Hickey and Steele (1978:168) reported musk-ox being taken in
mass, while Scherger (1966) cited the multiple predation of North American
.s (1978:3) to be turkey. Further, Klein (1989:325) reported the tak~ng of eland in mass, while
1, imbued with Scott (1980, 1986) suggested that Old World mammoth and woolly
~ preys' behav- rhinoceroses may have been taken by multiple predation. For prehistoric
:tted with com- North American fauna, bison (Frison 1987; Reeves 1983; Wheat 1972) and
he use of fire. perhaps mammoth (Saunders 1977) were taken in multiple predation
mn was first a episodes. · ·
;ressed toward In addition to those animals, there is documentation of humans' having
tion that was a killed in mass a wide variety of marine and freshwater fishes, as well as
Thus, Nitecki numerous inver.-tebrates. Since the harvesting of invertebrates, fishes, and
the taking of vertebrate land animals requires such different behavior patterns and cultural
accoutrements, it is difficult to imagine any theme unifying the patterns.
the con text of However, in the harvesting of all three kinds of food, the advantage of gather-
ica and the net ing the largest amount that can be used with the least effort is apparent.
se of multiple Consequently, techniques to acquire organisms in mass have been developed
1ization of the for all kinds of animals. From our perspective, the separation of discussions of
)mmon preda- hunting, fishing, and slow game harvesting (Coon 1971; Lee 1968:41-42)
nunal hunting obscures the fact .that multiple predation of land mammals is part of a larger
ass predation, pattern of predation where the strategy is to acquire an adequate amount of
;t documented animal resources for a group larger than those participating in the hunt in as
efficient a manner as possible.
N
Table 2-2. Selected Examples of Taxa Taken by Humans in Multiple Predation Episodes as Related to Environment 0
and Geography
9
0
(/")
Taxa Environment Geographic Location Reference ,....._
(':)
(':)
'R)
$::)
Class Insecta Desert, deciduous forest, Worldwide Coon (1971:166-168); Sutton (1988) ~
Insects tropical forest, grassland, $::l..
savanna ~
Class Mollusca Aquatic (marine and Worldwide Coon (1971:72) ;s
Molluscs freshwater); terrestrial IJ:j
$::)
(land snails) 8""
Class Osteichthyes Aquatic, marine, riverine Worldwide Coon (1971:71-72, 141-146);Nelson ""''
Bony fish (1973:55-70)
Class Aves Coastal, inland Worldwide Coon (1971:71)
Birds
Class Reptilia Riverine Southwest New Guinea Anell (1960:8)
Crocodiles
Class Mammalia Desert Australia Anell (1960:2)
Order Marsupialia
Family Macropodidae
Kangaroos
Order Lagomorpha Desert American Southwest Driver (1940:186); DuBois (1940:13-14); W.
Family Leporidae W. Hill (1982:52); Spier (1928:112-113)
Rabbits and hares
Order Rodentia Grassland Africa Usher-Wilson (1947:31-32)
Rats
Order Carnivora Forest South America Cooper (1946:143)
Family Felidae
Puma
Taxa Environment Geographic Location Reference

Africa Usher-Wilson (1947:31-3:2.)
Order Rodentia Grassland
Rats Cooper (1946:143)
Forest South America
Order Carnivora
Fa.ntily Felidae
Puma
Taxa Environment Geographic Location Reference

-~
Order Pinnipedia Aquatic, coastal Alaska Nelson (1969:226)

Family Phocidae
Phoca vitulina
Harbor seal
Family Odobenidae Aquatic, coastal Alaska Nelson (1969:360-361)
Odobenus rosmarus
Walrus
Order Artiodactyla Tropical forest South America; Melanesia Anell (1960:8); Holmberg (1950:25);
Family Suidae Metraux (1946:43)
Peccaries
Family Cervidae Arctic, forest edge Alaska, northeastern Asia Birket-Smith (1959:86); Blehr (1990);
Rangifer Gordon (1990); Jochelson (1926:378);
Reindeer or caribou Mathiassen (1928:58); Porsild (1920:304-
307); Spiess (1979:104-130)
Odocoileus Dedduous forest, North America, southern Heidenreich (1971:206); Hill (1938:185);
White-tailed and mule deer woodland, forest edge United States, American Morgan (1904:336); Swanton (1946:317);
Southwest, eastern United Waselkov (1978)
States
Cervus nippon Forest Asia Coon (1971:71) ~
Sika or Japanese deer E..
p-.
Family Antilocapridae Grassland North America, Canada, Arkush (1986); Brumley (1984); Davis and '"d
.......
Antilocapra americana Great Basin Fisher (1990); Egan (1917); Hill (1938:149); ro
Pronghorn Regan1934; Stephen (1936:276-277) ~
'"1
ro
Family Bovidae Plains grassland North America Davis and Wilson (1978); Dibble and 0..
Pl
Bison spp. Lorrain (1968); Forbis (1962:69); Verbicky- p-.
Bison Todd (1984); Wheat (1972) 0
~
Ovibos moschatus Arctic and tundra Canada Hickey and Steele (1978) -
Musk-ox N
~
Table 2-2.-Continued N
N
Taxa Environment Geographic Location Reference ~

0
Ovis canadensis Mountainous, plains North America, American Driver (1982); Frison (1985, 1987:201-205); C/)
..;..
~
Bighorn or mountain sheep Southwest, Plains Frison et al. (1990); Spier (1928:100) ~
·~
Synceros caffer Grassland Africa Usher-Wilson (1947:31-32) :>::l
~
African buffalo ~
Cephalophus monticola Forest Africa Hudson (1991) 0::1
Blue duiker ~-
Family Camelidae Grassland, savanna South America Cooper (1946:143)
0::1
L£zma guanicbe ~
Guanaco ~
Order Perissodactyla Open and mountainous France Clark (1969:66); Olsen (1989)
Family Equidae temperate grasslands
Equus caballas
Modem horse
Equus sp. Grassland Africa Roscoe (1924:66)
Zebra
Equus sp. Central Asia Atkinson (1860:260)
Order Proboscidea Grassland Africa Coon (1971:111); Usher-Wilson (1947:31-
Family Elephantidae 32) '
Loxodonta africana
African elephant
Elephas maximus Forest Thailand Graham (1924:53-56)
Asian elephant
? Mammuthus Grassland, arctic steppe North America Saunders (1977)
Mammoth
Order Cetacea Aquatic Northern Europe Clark (1947:89)
Family Delphinidae
Globicephala sp~
Pilot whale
Note: Emphasis of the table is on manunals. References for insects, molluscs, fishes, and birds are exemplary only.
~~
::n o~
s· o
PJ
g?. s· ~ e: e: iS g?.
~ ~ 'T.I ;:r ::+: ~ ""d qs
OQ
8 &]
~ l"\ ....,
8
~ ,.,e.
st 8~ !?' ~Pl m
o
8 st ~ P3 > ~ s- § ~ s- ~ ~
Pl < . . , g?. 8 ;;r p.. ro r.. ~ ro ro
~- g g st
Er
~ ~ ~ ~ ~- ~
PJ ~ ~ 8 ~ ro ;g
PJ
~ <
~
< PJ ~ ~
rl- 1""'1 !:j
() \ol"' 0 () ;:J (!) ....... PJ rl- Vl t.:::.. .......rl- (!) ....... p.. ;:J p.....
""""' rl- ....... ..... ;:J ........ ;:J
~· p:~ 5· Q ~ ~] ~ m~ g PJ~ ~ ~ ~: s: 1 ; ~ ""d§:: ~~ [ :::..~ ~~ ; 0~ '-<:;. ft 0~-~...., ~ ~~ ~ ·Os s[ ~X ::r'~ § ~ g~ ~

,.J ........ ,.J ,.J 11.1 \.j J () () 11.1
() PJ ....,
OQ <; v .....(!) """"' PJ v -.....
ffi" Vl Vl
& ff
Vl...., ::r' H-.
~ ~ roC§.~ s· ~ ~ : § § ~ ""d ~ g. 2: s;.. ~ n tn" ci ~ E· ~ s;.. ~ :::r ~ ~ ~ ~- ;:r ~ e.. ~ ~ ~ -~ ~ ~ s ~ ~ s.
,.J .............. rl- rl- 11.1 (D ... • 11.1
Q PJ Vl
Multiple Predation I 23
While other animals such as whales feeding on krill or anteaters feeding on

a nest of termites can be viewed as multiple predators in this broader devel-
opment of the concept, the human use of multiple predation remains defini-
tive, although not unique, because humans are one of the very few terrestrial
predators to utilize multiple predation to prey upon larger, swifter game as
well as smaller species from an incredibly' wide range of classes. In particular,
humans utilize the more specialized form of mass predation to a greater
extent than all other predators.
An examination of the behavior of the prey species being killed documents
that humans have taken in mass animals of the land, water, and air; sessile
and mobile animals; asocial and social animals; nocturnal and diurnal
animals; arboreal and terrestrial animals; animals that flee upon confrontation
~ :>.. with predators and those that stand and defend themselves and their con-
OC)
~
-a0 specifics. In essence, no matter what the structure of the behavior of the prey
0"1
c ~
<13 species, humans have attempted to utilize multiple predation as a technique
~
1-o
0
ro t~ to harvest it.
Geographically, the authors discussed above have documented multiple
(1)
(1)
1-<
<13
predation in North America, South America, Europe, Asia, Africa, Australia,
<I)
'E and the Melanesian Islands. The range of habitats where animals have been
;.a taken b;'f''humans in mass includes arctic tundras, boreal forests, mixed decid-
"0
Q)
~
<13 uous woodlands, tropical forests, temperate grasslands, tropical grasslands,
0...
0
V'l
(1)
..c
Asian steppes, temperate deserts, tropical deserts, woodland savannas, and a
~ <I)
variety of aquatic environments. Again, it appears that wherever humans
r..Ll u::
<I)~
~ u
<I)
:;:3
have lived, and in whatever ecosystems they have hunted, there is evidence
~
~ ;:g that multiple predation has been used as a hunting strategy. It also is docu-
z0
s mented that multiple predation has been practiced by hunters and gathers,
.l!l agriculturalists, and members of city-state societies.
u
(1)
u
).., .s Multiple predation has been reported as part of the day-to-day activities of
)..,
lJ
1-<
..8 hunters (Coon 1971:83) as well as large, seasonal, communal hunts, such as
ii <I)
(1) those typified by bison drives. Here, too, irrespective of the social pattern or
~ u
~
j
~ makeup of the hunters, the teChnique of multiple predation has been used.
~
, With this distinction having been made between the number of predators
j
r:;
~ -~
'*
p:::
ui participating in the hunt' and the number of prey taken, it is now possible to
ll) ~ Cil
ll)
& ask under what conditions we would expect to find social hunting and/or
~
j ..:X:
§ <13 multiple predation practiced and what is their ecological and evolutionary
s significance. Packer and Ruttan (1988) and Steele (1982) have made a similar
~
0
-~
separation between prey number and predator number. Packer and Ruttan's
::0
(1)
distinction of the two strategies was based on data from 60 species and 28
.i'3 different studies of group hunting. Though their study did· not include
~ (1)
'"0 -:5 primate data, their conclusions have important ramifications for understand-
<Jl
~,.!:;
!IS
~.g_5}
:s . .......
0,
-~
<I)
ing cooperative hunting in humans.
-t:: .....
..... 0 !IS-..S QJ ro
..c Following Kruuk (1970), Packer and Ruttan noted that individuals will hunt
~ 8 "QJC)j!"Ca 0..
2~ u
l-4~u~
&,.!:;
~
s single prey in groups rather than solitarily only when individual hunters have
low foraging efficiency. They also stated, "Species that capture multiple prey
~~
QJ '"""" ..... ~
"-!
'"Cl ~ ~ 0
('.·
1-4
O~OP...
..8c;~
~ in a single hunt are not faced with similar disadvantages from grouping and
should not be similarly constrained to hunt alone by high individual foraging
efficiency" (Packer and Ruttan 1988:189). Finally, they viewed cooperative
241 D. G. Steele tJnd B. W. Baker
hunting in many species as resulting from gregariousness, rather than causing is required
it. Packer and Ruttan (1988:189) identified one exception to this last conclu- the effectiv
sion. In species that take multiple prey, cooperative hunting often appears to organizati1
be an important factor in the formation of groups. using; (4) ~
We argue that social hunting in humans coevolved with the predatory servation c
strategy of multiple predation, as well as with -the hunting of large game, and Considerir
that multiple predation is one of the most definitive types of human predatory kill and he:
strategies. Multiple predation is far more definitive than social hunting or big and food 1
game hunting because it is the rarest predatory strategy among nonhuman system, na
predators but is common among human predators. Finally, we see the three the consun
strategies of multiple predation, hunting larger game, and an increased effi- transportii
ciency in hunting as a complex of strategies designed to procure the greatest ized the F
amount of animal resources·in the most efficient manner possible. participate
Once multiple _predation is recognized as a separate practice from social While it
hunting, additional questions arise: How is it adaptive, and what behavioral conjunctioJ
features of the predator make it possible? Hyenas, wolves, lions, and humans the transp
are all large social predators. Consequently, the ·food intake to maintain the exploring.
body mass of the social groups of these species must be large. Multiple preda- until they
tion provides the needed food as an alternative to taking larger prey or arrival, ht:
increasing single prey hunting success rates. Accepting this line of reasoning consumpti
alone, one would anticipate, then, that any large social predator, living or need may
extinct, may have incorporated multiple predation as a hunting strategy. behaviors-
Of the four documented practitioners of multiple predation (hyenas, lions, est elabora
wolves, and humans), multiple predation is practiced on a wider range of tion, occu:
prey by humans, in more contexts, and more frequently. The strategy prob- commonly
ably developed among humans because humans as predators have the small- be stored t
est percentage of the group participating in the hunt. In virtually all social the resoun
predators, most of the adults and all but the youngest juveniles participate in drying.
the hunt. In humans, typically only the physically active adult and adolescent If multif
males participate in hunting terrestrial vertebrates. These huniers usually tion of foe
represent less than one-half of the group membership. Consequently, for food, then
hunting to be a successful part of the overall subsistence strategy, the hunters either coe,
must return with a greater portion of meat to share with the nonhunting human bel
members of the group. This is not to belittle or ignore the harvesting of plant human pn
foods nor the foraging of small and slow game that is more commonly done multiple "F
by those who do not traditionally hunt: Rather, it is to emphasize that if hunt- sequential
ing is a successful part of a human omnivorous subsistence strategy, surplus the more e
animal resources must be returned and shared, and that places a responsibil- Similarly,
ity on the hunter to return with the most resources possible. This can be strategy as
accomplished by being a more efficient killer of single prey, killing larger maintenan
prey, or killing more prey per hunting episode through multiple predation. It resources v
is informative that in more arid and temperate environments where plant logical recc
resources are not plentiful, or in seasons when they are not plentiful, hunting predation
becomes more important, and multiple predation becomes a more prevalent multiple p
form of hunting. 1989; Strau
If multiple predation is to be an effective predatory strategy for humans, a While m
series of related behavioral patterns associated with hunting and subsistence envision fi·
~r than causing is required. We feel that six associated behavioral patterns are most central to
1is last conclu- the effective use of multiple predation: (1) sharing of food; (2) a complex social
ften appears to organization supporting the hunting and the sharing of the food; (3) tool
using; (4) delayed consumption of food; (5) transportation, storing, and pre-
the predatory servation of the surplus of the kill; and .<6) effective group communication.
1rge game, and . Considering the six behavioral patterns, hyenas, wolves, and lions share the
.man predatory kill and have a complex behavioral repertoire that facilitates social hunting
hunting or big and food sharing. However, only humans have a complex communication
mg nonhuman system, namely, language; use tools extensively in the hunting of game; delay
re see the three the consumption of the food consistently; have elaborated the mechanisms of
. increased effi- transporting, preserving, and storing the surplus; and have virtually canon-
1re the greatest ized the practice of redistributing· the resources among individuals who
>le. participated in the hunt as well as those who did not.
ice from social While it is not necessary to elaborate on human tool use and sharing in
vhat behavioral conjunction with human predation, the delayed consumption of resources and
lS, and humans the transportation, storage, and preservation of the surpluses are worth
to maintain the exploring. While some nonhuman predators delay the consumption of food
\1:ultiple preda- until they return to den sites or regurgitate their stomach contents upon
larger prey or arrival, f1umans practice a delayed consumption of food that permits the
ne of reasoning consu117Ption of the surpluses during future times of need. That future time of
:lator, living or need may be a few days or months away. It is associated with this suite of
~strategy. behaviors-delayed consumption, storage, and preservation-that the great-
t (hyenas, lions, est elaboration and reliance on multiple predation, particularly mass preda-
wider range of tion, occurs. In northern climes during winter months, surplus caribou,
~ strategy prob- commonly procured in mass or in planned sequential predation forays, may
have the small- be stored by freezing (Binford 1981). In warmer climes or in warmer seasons,
tually all social the resources of multiple prey may be preserved, or cured, by smoking and
~s participate in drying. .
· and adolescent If multiple predation is dependent at least in part upon delayed consump-
1unters usually tion of food, tool using, and ·transporting, storing, and preserving surplus
nsequently, for food, then multiple predation as a commonly practiced hunting strategy
~gy, the hunters either coevolved with t_hem or evolved after they were established in the
the nonhunting human behavioral repertoire. Given the present level of our understanding of
vesting of plant human prehistory, it is difficult to establish when humans became effective
:ommonly done multiple predators. On the basis of their simplicity; we anticipate that
:ize that if hunt- sequential predation and opportunistic mass predation evolved first and that
trategy, surplus the more elaborate and stylized forms of mass predation evolved from them.
~s a responsibil- Similarly, we anticipate that multiple predation became a more adaptive
>le. This can be strategy as body mass of humans increased, thus requiring more energy for
{, killing larger maintenance, and as humans expanded into harsher climates where plant
ple predation. It · resources were scarcer during some or all seasons. UnfortUnately, the archaeo-
1ts where plant logical record is unclear on the evolutionary pathway through which multiple
entiful, hunting predation evolved. By Upper Paleolithic times, however, the evidence for
more prevalent multiple predation, particularly mass predation is well documented (Olsen
1989; Straus 1987).
y for humans, a While multiple predation would be practiced as opportunities arose, we
and subsistence envision five conditions ideal for the practice of mul~iple predation, particu-
I
26 D. G. Steele and B. W. Baker
larly mass predation: (1) preying upon social species that herd or school and For n
can be surrounded or driven; (2) preying upon small and/ or solitary species species,
multipl~
of prey that occur in high enough densities to warrant their being driven in a
particular direction or surrounded; (3) living in temperate or cooler regions arthrop1
that make possible the preservation and storage of food surpluses for longer the gam
periods of time as an adaptation against periods of limited access to animal Usually,
resources; (4) existing in societies where at least at select times people congre- gatherir
gate in large enough groups to conduct large social hunts; and (5) existing in predatic
societies where specific game provides a marketable resource exploited by the the larg(
society. to provi
A reexamination of the list of prey upon which humans have practiced Then
multiple predation documents that many of them are herding mammals such and unc
as bison, caribou, musk-ox, sheep, horses, and elephants. They have been tary fisr
recognized as the preferred prey of human soci.al hunters, and their remains ing of p
form some of the most spectacular evidence of the hunting expertise of prehis- or poise
toric humans. Those species, and the multiple predation of them, become all frequen'
the more essential in environments where they' form the major resource of the Multi'
peoples (e.g., the caribou in the northern latitudes of the Americas and the quently,
bison in the North American Great Plains). In those environments, subsistence norther:
depends on the efficient taking of these species, and multiple predation equivalE
becomes a major hunting strategy. Most of the herd species flee as a response in Nortl
to predator stress (e.g., bison, caribou, horse, sheep), but others (musk-ox and Americc;
elephant) may respond by standing and forming defense groups. Multiple may no:
predation practices have been developed to counter both defense responses. may be
Most of the research on social hunting examines the social hunting of the impossi1
herd species such as bison, caribou, and sheep. In those societies that were em latit1
dependent upon these animals, social hunting took on all of the well- ceous fo
developed cultural trappings typical of the communal hunt. They include the strategy
tightly structured social organization of the hunters, the magico-religious African
sanctions governing the behavior of the hunters and the huntE:fd, and the use to plant
of complex tools including nets, traps, fire, and fences. Certainly, it is the most ability o
complex and distinctive form of mass predation, but it is not the only form. made it
Opportunistic cases of multiple predation frequently occur in those societies scale.
as well.· Hunti
From an ecological perspective, schooling fish also should be considered as tion, als,
a social prey subject to multiple predation. Many coastal societies emphasize ticed me
the seining, trapping, and use of throw nets for the capture of schooling fish. North A
Predation on schooling fish or trapping fish in an impoundment may be relative!
undertaken by a single fisherman-hunter, or it may be an organized effort effective
involving many individuals. In this regard it is noteworthy that in northern tions, p<
latitudes of Eurasia and North America many societies are successful because hunts cc
they practice the multiple predation of both terrestrial mammals and fish, Among-
each in its proper season of harvest. When one notes the range of species coincide
taken that are social, it is apparent that size and behavior of the prey, and the While
habitat in which they occur, dictate whether single predators can be effective it obviot
or whether a social hunt is required. It is also apparent that the taking of more predatic
than one specimen is a viable and practiced option for increasing the harvest. societies
or school and For multiple predation to be particularly effective on small or isolated

species, the prey must occur in high densities within a given habitat. Also,
>litary species
1g driven in a multiple predation on species such as rabbits, monkeys, deer, and terrestrial
arthropods is most effective with large groups of hunters who can surround
:ooler regions
the game or drive the prey into nets or natural traps such as water or canyons.
ses for longer
Usually, the hunts take place at times of large social gatherings of hunting and
:ess to animal
gathering groups, and fewer solitary species seem amenable to this form of
>eople congre-
(5) existing in predation. We suspect that in many instances these drives provide food for
ploi ted by the the large social groupings gathered temporarily and may not be used as much
to provide surplus for preservation and storage.
ave practiced The multiple predation of solitary fish, however, can occur more frequently
1ammals such and under a wider variety of circumstances. For instance, the seining of soli-
tary fish can be accomplished by a few hunter-fishers. Alternatively, the dry-
ey have been
ing of pools can congregate the solitary fish so they can be effectively seined
their remains
or poisoned. The mass predation of drought-concentrated fish may occur
rtise of prehis-
frequently enough to be a predictable resource at certain times.
m, become all
Multiple predation, particularly mass predation, seems to occur more fre-
·esource of the
quently, or has been reported more frequently, among societies occupying
~ricas and the
northern latitudes than those inhabiting more equatorial latitudes. The
ts, subsistence
equival~nt reliance on mass predation of caribou and bison by some societies
ple predation
in North America does not seem to have ·equivalents in Central and South
! as a response
(musk-ox and America, and possibly the reliance on mass predation of herd species in Africa
1ups. Multiple may not have been as common as in some prehistoric Eurasian societies. It
e responses. may be that the higher mean seasonal temperatures complicate, or make
impossible, the effective preservation of the surplus. Certainly, among north-
1unting of the
em latitude hunters and gatherers, the practice of preserving animal proteina-
ties that were
ceous foods is common and widespread and is typically an essential survival
l of the well-
strategy for winter months. Among South American, Southeast Asian, and
ey include the
African societies, preservation of foods is less common and largely restricted
tgico-religious
to plant foods. In those more equatorial environments the year-round avail-
d, and the use
ability of animal resources and the greater reliability of plant resources have
r, it is the most
made it unnecessary to practice mass predation as frequently or on as large a
the only form.
scale.
those societies
Hunting societies practicing multiple predation, particularly mass preda-
considered as tion, also must have enough hunters to permit multiple predation to be prac-
:ies emphasize ticed more than opportunistically. For example, along the northwest coast of
North America, the taxonomically abundant and rich environment permitted
schooling fish.
relatively large population densities of hunters and gatherers, and there they
lment may be
effectively harvested the anadromous salmon. Among Great Plains popula-
ganized effort
tions, particularly with _the advent of horse-based hunting societies, bison
.at in northern
hunts could occur year-round, as well as when several bands congregated.
:essful because
Among the caribou hunters of the American northlands, spring and·fall hunts
mals and fish,
coincided with the largest social gatherings.
nge of species
While we haye concentrated on mass predation as a subsistence adaptation,
~ prey, and the
it obviously fits as a strategy in a market economy as well. Particularly, mass
an be effective
taking of more predation of game in the nineteenth-:-century market economy of Western
g the harvest. societies is one of the clearest examples of it. In North America, examples of
I>
;~\, ;, , ,<.
I
28 D. G. Steele ~nd B. W. Baker
mass predation include the mass killing of bison in stands, and the mass
killing of waterfowl ·along the East Coast with the use of small boat-mounted
cannons loaded with shot and scrap metal. But the use of game in a market 1. Throu.
fashion may well have occurred prehistorically for bison hides (Creel 1991). In to mean th
Africa, Wilkie and Curran (1991) have convincingly argued for the spread of that pursui
net hunting as a form of mass predation among the Mbuti because of its and Crade1
greater efficiency for harvesting marketable quantities of game. (1981) vie¥
In summary, under the above-listed conditions we expect to see mass pre- they are tw
the encoun
dation occurring in its most complex form, but we feel it is important to see
ual spontm
those instances as examples of the maximum expression of mass predation,
point -becOl
not the only examples. Rather, we see multiple predation and mass predation that it has·
in a broader context, one that is universally practiced but reached its flores- on the basi:
cence only under a combination of specific circumstances. of the resot
Conclusions
The human hunting repertoire (Figure 2-1) can be seen to include a Anderson,:
1982 G
series of resource exploitation strategies: (1) a single hunter may take a single
Anell, B.
specimen; (2) a single hunter may take multiple specimens through sequential 1960 .fj
predation; (3) a single hunter may take multiple specimens through mass pre- grapru
dation; (4) a social hunting group may take single specimens; (5) a social 1969 R
hunting group may take multiple specimens through sequential predation; Upsali
and finally(6) a social hunting group may take multiplespecimens through Arkush, B.:
mass predation. Multiple predation is practiced by only a few nonhuman 1986 A
mammalian predators, and only in limited instances. Humans are unique biology
among mammalian social predators who practice multiple predation, particu- Atkinson, 1
larly mass predation, and who seem to use it on a wide variety of game of a 1860 T
wide range of sizes and under a wide variety of circumstances. Wf! view mul- Bartholeme
tiple predation as an alternative strategy to increase the animal resources pro- 1953 E
cured in the most efficient manner. It has developed to its greatest extent in Bertram, B.
1979 S1
humans for a variety of reasons. Multiple predation is most common in those
Ecosysi
societies where hunting is a major subsistence strategy for at least a part of the Unive1
year and where environmental conditions permit the storage of the surplus. Binford, L.:
1981 B,
1984 Ft
Acknowledgments Birket-Smit
1959 T1
Thanks are extended to Cristi Assad, David L. Carlson, Lee Cronk, Blehr, 0.
Gary B. DeMarcay, Jean Hudson, Virginia Massey, James O'Connell, Jeffrey 1990 c
Saunders, Harry J. Shafer, Brian S. Shaffer, and two anonymous reviewers for Huma:
Reeve~
constructive conversations pertaining to the subject or critical readings of
Blumenschj
drafts of the manuscript. Brian S. Shaffer produced Figure 2-1. We thank the
1986a C
secretarial staff of the Department of Anthropology, Texas A&M University,
Seaver
for originally typing the manuscript. 1986b Et
and the mass Note

boat-mounted
ne in a market 1. Throughout this paper we use the terms hunting and predation interchangeably
(Creel1991). In to mean the pursuit of animals in order to capture or kill them, regardless of when
r the spread of that pursuit begins or whether the prey is large and swift or small and sessile. Isaac
because of its and Crader (1981) restricted hunting to the taking of large mobile prey, but Harding
(1981) viewed predation and hunting as similar. Ingold (1987:87-100) argued that
they are two separate concepts: hunting reflects the intention to procure game before
) see mass pre-
the encounter of prey while predation is spontaneous. We feel that once an individ-
nportant to see ual spontaneously encounters prey and actively begins seeking it out, it has at that
1ass predation, point become a hunter. Thus, the end product, if the animal is captured or killed, is
nass predation that it has been hunted. We further feel that distinguishing predation and hunting
ched its flares- on the basis of prey characteristics maskS basic similarities practiced in the harvesting
of the resources.
References
Anderson, S. (editor)
~en to include a 1982 Guide to Mammals. Simon and Schuster, New York.
3.Y take a single Anell, B.,
mgh sequential 1960 /Hunting and Trapping Methods in Australia and Oceania. Studia Ethno-
ough mass pre- graphica Upsaliensia 18.
ns; (5) a social 1969 Running Down and Driving of Game in North America. Studia Ethnographica
1tial predation; Upsaliensia 30.
:imens through Arkush, B. S.
:ew nonhuman 1986 Aboriginal Exploitation of Pronghorn in the Great Basin. Journal of Ethno-
3.ns are unique biology 6:239-255.
dation, particu- Atkinson, T. W.
ty of game of a 1860 Travels in the Region of the Upper and Lower Amoor. Harper's, New York.
. We view mul- Bartholemew, G. A., and J. B. Birdsell
l resources pro- 1953 Ecology and the Protohominids: American Anthropologist 55:481-498.
Bertram, B. C. R.
eatest extent in
1979 Serengeti Predators and Their Social Systems. In Serengeti: Dynamics of an
,mmon in those Ecosystem, edited by .A. R. E. Sinclair and M. Norton-Griffiths, pp. 221-248.
ast a part of the University of Chicago Press, Chicago.
f the surplus. Binford, L. R.
1981 Bones: Ancient Men and Modern Myths. Academic Press, New York
1984 Fauna( Remains from the Klasies River Mouth. Academic Press, New York.
Birket-Smith, K.
1959 The Eskimos. Methuen, London.
son, Lee Cronk, Blehr, 0.
Connell, Jeffrey 1990 Communal Hu~ting as a Prerequisite for Caribou (Wild-Reindeer) as a
1s ~reviewers for Human Resource. In Hunters of the Recent Past, edited by L. B. Davis and B. 0. K.
Reeves, pp. 304-326. Unwin Hyman, London.
cal readings of
Blumenschine, R. J.
1. We thank the 1986a Carcass Consull}ption Sequences and the Archaeological Distinction of
&M University, Scavenging and Hunting. Journal of Human Evolution 15:639-659.
1986b Early Hominid Scavenging Opportunities: Implications of Carcass Availability in
I
the Serengeti and Ngorongoro Ecosystems. BAR Interna~ional Series 283. Oxford.· Memo:
1987 Characteristics of an Early Hominid Scavenging,Niche. Current Anthro- Driver, H. I
pology 28:383-407. · 1940 V\
1988a An Experimental Model of the Timing of Hominid and Carnivor_e Influence Archae1
on Archaeological Bone Assemblages. Journal of Archaeological Science 15:483- Driver, J. C.
502. 1982 E;
1988b Reinstating the Early Hominid Scavenging Niche: A Reply to Potts. Current Radde
Anthropology 29:483-486. 1990 M
1989 A Landscape Taphonomic Model of the Scale of Prehistoric Scavenging Recent
Opportunities. Journal of Human Evolution 18:345-371. Londo:
Brumley, J. H. DuBois, C.
1984 The Laidlaw Site: An Aboriginal Antelope Trap from Southeastern Alberta. 1940 vv
In Archaeology in Alberta, edited by D. Burley, pp. 96-127. Occasional Paper No. ologya;
23. Archaeological Survey of Alberta, Edmonton. Earle, T. K.,
Bunn, Henry 1980 M
1982 Meat-Eating and Human Evolution. Unpublished Ph.D. dissertation. Univer- York.
sity of California, Berkeley. Egan, H.
Bunn, H. T., and E. M. Kroll _ 1917 Pi
1986 Systematic Butchery by Plio/Pleistocene Hominids at Olduvai Gorge, Forbis, R. G
Tanzania. Current Anthropology 5:431-452. 1962 Tl
Clark, J. G. D. Bulletit
1947 Whales as an Economic Factor in Prehistoric Europe. Antiquity 21:84-104. 1978 Sc
1969 World Prehistory: A New Outline. Cambridge University Press, London. 8.
Coon, C. Frison, G. C
1971 The Hunting Peoples. Little, Brown, Boston. 1985 PI
Cooper, J. M. Rocky
1946 The Patagonian and Pampean Hunters. Bureau of American Ethnology Sympoc
Bulletin 143(1):127-168. Edmor
Creel, D. 1987 PI
1991 Bison Hides in Late Prehistoric Exchange in the Southern Plains. American egies. :
Antiquity 56:36-40. / NiteckJ
Curio, E. Frison, G. C
1976 The Ethology of Predation. Springer Herlag, Berlin. 1990 Pt
Dart, R. A. North.
1953 The Predatory Transition from Ape to Man. International Anthropological and Reeves
Linguistic Review 1:201-209. Gordon, B. (
Davis, L. B., and J. W. Fisher, Jr. 1990 w
1990 A Late Prehistoric Model for Communal Utilization of Pronghorn Antelope L. B. D;
in the Northwestern Plains Region, North America. In Hunters of the Recent Past, Graham, W.
edited by L. B. Davis and B. 0. K. Reeves, pp. 241-276. Unwin Hyman, London. 1924 Sit
Davis, L. B., N. Noe-Nygaard, and B. 0. K. Reeves Hall, R. L., a
1986 Communal Land Mammal Hunting and Butchering. World Archaeological 1978 w
Congress. Allen and ·unwin, London. Harding, R.
Davis, L. B., and B. 0. K. Reeves (editors) 1981 A1
1990 Hunters of the Recent Past. Unwin Hyman, London. orous F
Davis, L. B., and M. Wilson (editors) Univer:
1978 Bison Procurement and Vtilization: A Symposium. Plains Anthropologist Hawkes,K.
Memoir14. 1990 w
Dibble, D. S., and D. Lorrain Uncertli
1968 Bonfire Shelter: A Stratified Bison Kill Site, Val Verde County, Texas. Texas Westvi1
1991 Sh
283. Oxford. Memorial Museum Miscellaneous Papers No. 1. Austin, Texas.

:urrent Anthro- Driver, H. E.
1940 Wappo Ethnography. University of California Publications in American
:Uvore Influence Archaeology and Ethnology 36:179-220.
Science 15:483- Driver, J. C.
1982 Early Prehistoric Killing of Bighorn Sheep in the Southeastern Canadian
to Potts. Current · Rockies. Plains Anthropologist 27:265-271.'
1990 Meat in Due Season: The Timing of Communal Hunts. In Hunters of the
)ric Scavenging Recent Past, edited by L. B. Davis and B. 0. K Reeves, pp. 11-33. Unwin Hyman,
London.
DuBois, C.
leastem Alberta. 1940 Wintu Ethnography. University of California Publications in American Archae-
donal Paper No. ology and Ethnology 36:1-148.
Earle, T. K., and A. L. Christenson
1980 Modeling Change in Prehistoric Subsistence Economies. Academic Press, New
~rtation. Univer- York.
Egan, H.
1917 Pioneering the West, 1846-1848. Salt Lake City.
Olduvai Gorge, Forbis, R. G.
1962 The Old Women's Buffalo Jump, Alberta. National Museum of Canada
Bulletin 180:56-123.
dty 21:84-104. 1978 /Some Facets of Communal Hunting. Plains Anthropologist Memoir 14(23):3-
>s, Loridon. 8. _;
Frison, G. C.
1985 Prehistoric and Early Historic Mountain Sheep Procurement in the Central
Rocky Mountains. In Contributions to Plains Prehistory: The 1984 Victoria
~rican Ethnology Symposia, eidted by D. V. Burley, pp. 267-276. Archaeological Survey of Alberta,
Edmonton.
1987 Prehistoric, Plains-Mountain, Large-Mammal, Communal Hunting Strat-
Plains. American egies. In The Evolution of Human Hunting, edited by M. H. Nitecki and D. V.
Nitecki, pp. 177-222. Plenum Press, New York
Frison, G. C., C. A. Reher, and D. N. Walker
1990 Prehistoric Mountain Sheep Hunting in the Central Rocky Mountains of
North America. In Hunters of the Recent Past, edited by L. B. Davis and B. 0. K
:-zthropological and Reeves, pp. 208-240. Unwin Hyman, London.
Gordon, B. C. · -
1990 World Rangifer Communal Hunting. In Hunters of the Recent Past, edited by
:1ghorn Antelope L. B. Davis and B. 0. K. Reeves, pp. 277-303. Unwin Hyman, London.
'Jf the Recent Past, Graham, W. A.
ryrnan, London. 1924 Siam. De La More Press, London.
Hall, R. L., and H. S. Sharp
l Archaeological 1978 Wolf and Man: Evolution in Parallel. Academic Press, New York.
Harding, R.
1981 An Order of Omnivores: Nonhuman Primate Diets in the Wild. In Omniv-
orous Primates, edited by R. Harding and G. Teleki, pp. 191-214 .. Columbia
University Press, New York.
ns Anthropologist Hawkes,K.
1990 Why Do Men Hunt? Some Benefits for Risky Strategies. In Risk and
Uncertainty in Tribal and Peasant Economies, edited by E. Cashdan, pp. 145-166.
11.ty, Texas. Texas Westview, Boulder, Colo.
1991 Showing Off: Tests of an Hypothesis about Men's Foraging Goals. Ethology
32j D. G. Steele a.nd B. W. Baker
and Sociobiology 12:29-54. Perspec

Hawkes, K., K. Hill, and J. F. O'Connell 103. Co:
1982 Why Hunters Gather: Optimal Foraging and the Ache of Eastern Paraguay. Jochelson, V\i
American Ethnologist 9:379-398. 1926 Th
Hayden, B. History
1981 Subsistence and Ecological Adaptations of Modem Hunter/Gatherers. In Jochim, M. P
Omnivorous Primates: Gathering and Hunting in Human Evolution, edited by R. S. 1976 Ht.
0._ Harding and G. Teleki, pp. 344-421. Columbia University Press, New York Press, r-.
Heidenreich, C. Jones, B. A.
1971 Huronia. McClelland and Stewart, Toronto. . 1990 Pa
Hickey, C. G., and D. G. Steele the Sou
1978 Predator /Prey Relations Between Man arid Muskox. American Quaternary Davis a
Association Abstracts of the Fifth Biennial Meeting. Karnil, A. C.,
Hill, K. 1987 Fo1
1982 Hunting and Human Evolution. Journal of Human Evolution 11:521-544. Karnil, A. C.,
1988 Macronutrient Modifications of Optimal Foraging Theory: An Approach 1981 Fa;
Using Indifference Curves Applied to Some Modem Foragers. Human Ecology Garlanc
16:157-197. Klein, R. G.
Hill, K., H. Kaplinf K. Hawkes, and A. M. Hurtado 1989 Th
1987 Foraging Decisions among Ache Hunter-Gatherers: New Data and Implica- Chicagc
tions for Optimal Foraging Models. Ethology and Sociobiology 8:1-36. Kooyman, B.
Hill, W. W. 1990 M<
1938 The Agricultural and Hunting Methods of the Navaho Indians. Yale University Recent j
Publications in Anthropology No. 18. Hyman
1982 An Ethnography of Santa Clara Pueblo New Mexico. University of New Mexico Kruuk,H.
Press, Albuquerque. 1970 Int
Holmberg, A. R. and The1
1950 Nomads of the Long Bow. Smithsonian Institution, Institute of Social Anthro- 1972 Th1
pology, Publication No. 10. Lawick-Goo'
Hudson,J. 1968 Th
1991 Nonselective Small Game Hunting Strategies: An Ethnoar9haeological Animal.
Study of Aka Pygmy Sites. In Human Predators and Prey Mortality, /edited by M. Leakey,M. I
C. Stiner, pp. 105-120. Westview Press, Boulder, Colo. 1971 Olt
Ingold, T. Univers
1987 The Appropriation of Nature: Essays on Human Ecology and Social Relations. LeeR. B.
University of Iowa Press, Iowa City. 1968 Wl
Isaac, G. L. · InMan
1976 The Activities of Early African Hominids: A Review of Archaeological Lee, R. B., an
Evidence from the Time Span Two-and-a-Half to One Million Years Ago. In 1968 MCi
Human Origins: Louis Leakey and the East African Evidence, edited by G. L. Isaac Macdonald, ·
and E. McCown, pp. 483-514. W. A. -Benjamin, Menlo Park, Calif. 1984 Tht
1978 The Food Sharing Behavior of Protohominids. Scientific American 38:90-108. Marks, S.
1980 Casting the Net Wide: A Review of Archaeological Evidence for Early
1976 Lm
Hominid Land-Use and Ecological Relations. In Current Argument on Early Man,
Washin,
edited by L. Konigsson, pp. 226-251. Pergamon Press, New York.
1983 Bones in Contention: Competing Explanations for the Juxtaposition of Mathiassen, '
Early Pleistocene Artifacts and Faunal Remains. In Animals and Archaeology: 1928 Me:
Hunters and Their Prey, edited by J. Clutton-Brock and C. Grigson, pp. 3-18. BAR 6(1).
International Series 163. Oxford. Mech, L. D.
Isaac, G., and D. Crader 1970 Tht
1981 To What Extent Were Early Hominids Carnivorous? An Archaeological Museur
Schaller, G. B.
445-475. 1972 The Serengeti Lion. University of Chicago Press, Chicago.
Schorger, A. W.
NYork. 1966 The Wild Turkey. University of Ok~ahoma Press, Norman.
Scott, K.
ks of the United 1980 Two Hunting Episodes of Middle Palaeolithic Age at La Cotte de Saint-
Brelade, Jersey (Channel Islands). World Archaeology 12:137-152.
1986 The Large Mammal Fauna. In La Cotte de Saint-Brelade 1961-1978: Excava-
icago. . tions by C. B. M. McBurney, edited by P. Callow and J. M. Cornford, pp. 109-138.
Chicago. Norwich, U.K.
Sharp, H. S.
unting, edited by 1978 Comparative Ethology of the yYolf and the Chipewyan. In Wolf and Man:
)rk. Evolution in Parallel, edited by R. L. Hall and H. S. Sharp, pp. 55-79. Academic
Press, New York.
Shipman, P.
1983 Early Hominid Lifestyle: Hunting and Gathering or Foraging and Scaveng-
)f Upper Palaeo- ing? In Animals and Archaeology: Hunters and Their Prey, edited by J. Clutton-
Brock and C. Grigson, pp. 31-49. BAR International Series 163. Oxford.
1986 Scavenging or Hunting in Early Hominids: Theoretical Framework and
l t, Rinehart and Tests. American Anthropologist 88:27-43.
Shipman,,P., W. Bosler, and K. L. Davis
1981 ;Butchering of Giant Geladas at an Acheulian Site. Current Anthropology
uralist 132(2):159- 22:257-264.
Shipman, P., and A. Walker
1989 The Costs of Becoming a Predator. Journal of Human Evolution 18:373-392.
tmalian Predators Slobodchikoff, C. N. (editor)
and Psychology of 1988 The Ecology of Social Behavior. Academic Press, New York.
Smith, E. A. .
1980 Evolutionary Ecology and the Analysis of Human Foraging Behavior: An Inuit
~d to Be of Norse Example from the East Coast of Hudson Bay. Ph.D. dissertation, Cornell University,
Ithaca, N.Y. University Microfilms Int~rnational, Ann Arbor.
1983 Anthropological Applications of Optimal Foraging Theory: A Critical
at Olduvai Gorge, Review. Current Anthropology 24:625-651.
·sity, Cambridge, Spier, L.
1928 Havasupal Ethnography. Anthropological Papers of the American Museum of
8-347. Natural History 29:83-392.
Iunter I Gatherers. Spiess, A. E.
nan Adaptation in 1979 Reindeer and Caribou Hunters: An Archaeological Study. Academic Press, New
ernie Press, New York.
Steele, D. G.
1982 Patterns of Human Predation. Paper presented at the 47th annual meeting
1:120-135. of the Society for American Archaeology. Program and Abstracts, p. 88.
Minneapolis, Minn.
~dings of the Utah · Stephen, A. M.
1936 Hopi Journal. Columbia University Contributions to Anthropology No. 23.
Columbia University Press, New York.
Stephens, D. W., a~dJ. R. Krebs
1986 Foraging Theory. Princeton University Press, New Jersey.
by E. Johnson, pp. Straus, L. G. ·
Jbock. 1987 Hunting in Late Upper Paleolithic Western Europe. In The Evolution of
Metraux, A. , . Schaller, G. E
1946 The Caingang. Bureau of American Ethnology Bulletin 143(1):445-475. 1972 The
Morgan, L. H. Scherger, A.
1904 League of the Ho-De-No-Sau-Ne or Iroquois. Dodd, Mead, New York. 1966 The
Murie,A. Scott, K.
1944 The Wolves of Mount McKinley. Fauna of the National· Parks of the United 1980 Tw
States, Fauna Series 5, Washington, D.C. Brelade,
Nelson, R. K. 1986 Th1
1969. Hunters of the Northern Ice. University of Chicago Press, Chicago. tions by'
1973 Hunters of the Northern Forest. University of Chicago Press, Chicago. Norwicl
Nitecki, M. H. Sharp, H. S.
1987 The Idea of Human Huntmg. In The Evolution of Human Hunting, edited by 1978 COJ
M. H. Nitecki and D. V. Nitecki, pp. 1-9. Plenum Press, New York. Evolutio;
Nitecki, M. H., and D. V. Nitecki (editors) Press, N
1987 The Evolution of Human Hunting. Plenum Press, New York. Shipman, P.
Olsen, S. L. 1983 Ear
1989 Solutre: A Theoretical Approach to the Reconstruction of Upper Palaeo- ing? In
lithic Hunting Strategies. Journal of Human Evolutipn 18:295-327. Brock ar
Oswalt, W. H. 1986 Sea
1973 Habitat and Technology: The Evolution of Hunting. Holt, Rinehart and Tests. A1
Winston, Dallas. Shipman, P., ·
Packer, C., and L. Ruttan 1981 But
1988 The Evolution of Cooperative Hunting. The American Naturalist 132(2):159- 22:257-2
198. Shipman, P., ,
Peters, R., and L. D. Mech 1989 ThE
1975 Behavioral and Intellectual Adaptations of Selected Mammalian Predators Slobodchikof
to the Problem of Hunting Large Animals. In Socioecology and Psychology of 1988 The
Primates, edited by R. H. Tuttle, pp. 279-300. Aldine, Chicago. Smith, E. A.
Porsild, M. P. 1980 Eva
1920 On Eskimo Stone Rows in Greenland Formerly Supposed to Be of Norse Example
Origin. Geographical Review 10(5):297-309. Ithaca, r\
Potts, R. , 1983 An1
1982 Lower Pleistocene Site Formation and Hominid Activities at Olduvai Gorge, Review.
Tanzania. Unpublished Ph.D. dissertation, Harvard University, Cambridge, Spier, L.
Mass. 1928 Ha,
1984a Home Bases and Early Hominids. American Scientist 72:338-347. Naturall
1984b Hominid Hunters? Problems of Identifying the Earliest Hunter/Gatherers. Spiess, A. E.
In Hominid Evolution and Community Ecology: Prehistoric Human Adaptation in 1979 Reh
Biological Perspective, edited by R. Foley, pp. 129-166. Academic Press, New 'York.
York. Steele, D. G.
Reeves, B. 0. K. 1982 Patt
1983 Six Millenniums of Buffalo Kills. Scientific American 249(4):120-135. of the S
Regan, A. B. Minneap
1934 Some Notes on the History of the Utah Basin. Proceedings of the Utah Stephen, A. N.
Academy of Science, Arts, and Letters 11:55-64. 1936 . Hop
Roscoe, J. Columbi
1924 The Bagesu. Cambridge University Press, Cambridge. Stephens, D. '
Saunders, J. J. 1986 Fora
1977 Lehner Ranch Revisited. In Paleoindian Lifeways, edited by E. Johnson, pp. Straus, L. G.
48-64. Museum Joumal, West Texas Museum Association, Lubbock. 1987 Hur
Schaller, G. B.
:45-475. 1972 The Serengeti Lion. University of Chicago Press, Chicago.
Schorger, A. W.
rYork. 1966 The Wild Turkey. University of Ol4ahoma Press, Norman.
Scott, K.
:s of the United 1980 Two Hunting Episodes of Middle Palaeolithic Age at La Cotte de Saint-
Brelade, Jersey (Channel Islands). World Archaeology 12:137-152.
1986 The Large Mammal Fauna. In La Cotte de Saint-Brelade 1961-1978: Excava-
:ago. tions by C. B. M. McBurney, edited by P. Callow and J. M. Comford, pp. 109-138.
:hicago. Norwich, U.K.
Sharp, H. S.
nting, edited by 1978 Comparative Ethology of the yYolf and the Chipewyan. In Wolf and Man:
·k Evolution in Parallel, edited by R. L. Hall and H. S. Sharp, pp. 55-79. Academic
Press, New York.
Shipman, P.
1983 Early Hominid Lifestyle: Hunting and Gathering or Foraging and Scaveng-
f Upper Palaeo- ing? In Animals and Archaeology: Hunters and Their Prey, edited by J. Clutton-
Brock and C. Grigson, pp. 31-49. BAR International Series 163. Oxford.
1986 Scavenging or Hunting in Early Hominids: Theoretical Framework and
, Rinehart and Tests. American Anthropologist 88:27-43.
Shipman, f., W. Bosler, and K. L. Davis
1981 ;Butchering of Giant Geladas at an Acheulian Site. Current Anthropology
ealist 132(2):159- 22:257-264.
Shipman, P., and A. Walker
1989 The Costs of Becoming a Predator. Journal of Human Evolution 18:373-392.
1alian Predators Slobodchikoff, C. N. (editor)
1d Psychology of 1988 The Ecology of Social Behavior. Academic Press, New York.
Smith, E. A.
1980 Evolutionary Ecology and the Analysis of Human Foraging Behavior: An Inuit
l to Be of Norse Example from the East Coast of Hudson Bay. Ph.D. dissertation, Cornell University,
Ithaca, N.Y. University Microfilms International; Ann Arbor.
1983 Anthropological Applications of Optimal Foraging Theory: A Critical
' Olduvai Gorge, Review. Current Anthropology 24:625-651.
.ty, Cambridge, Spier, L.
1928 Havasupai Ethnography. Anthropological Papers of the American Museum of
·347. Natural History 29:83-392.
mter/ Gatherers. Spiess, A. E.
m Adaptation in 1979 Reindeer and Caribou Hunters: An Archaeological Study. Academic Press, New
nic Press, New York.
Steele, D. G.
1982 Patterns of Human Predation. Paper presented at the 47th annual meeting
20-135. of the Society for American Archaeofogy. Program and Abstracts, p. 88.
Minneapolis, Minn.
ings of the Utah Stephen, A. M.
1936 Hopi Journal. Columbia University Contributions to Anthropology No. 23.
Columbia University Press, New York
Stephens, D. W., al).d J. R. Krebs
1986 Foraging Theory. Princeton University Press, New Jersey.
E. Johnson, pp. Straus, L. G. ·
ock. 1987 Hunting in Late Upper Paleolithic Western Europe. In The Evolution of
Human Hz/nting, edited by M. H. Nitecki and D. V:Nitecki, pp. 147-176. Plenum Wilson, E. C
Press, New York. ' 1972 TJ
Strum,S. C. Winterhaldt
1981 Processes and Products of Change: Baboon Predatory Behavior at Gilgil, 1987 T1
Kenya. In Omnivorous Primates: Gathering and Hunting in H_uman Evolution, edited Model.
by R S. 0. Harding and G. Teleki, pp. 255-302. Columbia University Press, New Harris
York. Winterhald1
Sutton, M. Q. 1981 H
1988 Insects ·as Food: Aboriginal Entomophagy in the Great Basin. Ballena Press Univer
Anthropological Papers No. 33. Ballena Press, Novato, Calif. Wolpoff, M.
Suzuki, A. · 1980 Pc
1975 The Origin of Hominid Hunting: A Primatological Perspective. In Soda- Wrangham,
ecology and Psychology of Primates, edited by R. H. Tuttle, pp. 259-278. Aldine, 1975 TJ
Chicago. Unpub
Swanton, J. R.
1946 The Indians of the Southeastern United States. Bureau of American Ethnology
Bulletin No. 137.
Teleki, G.
1973 The Predatory Behavior of Wild Chimpanzees. Bucknell University Press,
Lewisburg, Pa.
1981 The Omnivorous Diet and Eclectic Feeding Habits of Chimpanzees in
Gombe National Parl<, Tanzania. In Omnivorous Primates: Gathering and Hunting
in Human Evolution, edited by R. S. 0. Harding and G. Teleki, ·pp. 303-343.
Columbia University Press, New York.
Thompson, P.R.
1979 Rethinking Human Evolution. Ph.D. dissertation, Rutgers University, New
Brunswick, N.J. University Microfilms International, Ann Arbor.
Turner, A.
1989 The Concept of the Ecological Niche and Its Application to Studies of
Hominid Evolution. Archaeozoologia 3(1,2):111-119.
Usher-Wilson, L. /
1947 An Acholi Hunt. The Uganda Journal11:30-37.
Verbicky-Todd, E.
1984 Communal Buffalo Hunting among the Plains Indians: An Ethnographic and
Historic Review. Occasional Paper No. 24. Archaeological Survey of Alberta,
Edmonton.
Waselkov, G. A.
1978 Evolution of Deer Hunting in the Eastern Woodlands. Midcontinental
Journal of Archaeology 3(1):15-34.
Washburn, S. L., and C. S. Lancaster ,
1968 The Evolution of Hunting. In Man the Hunter, edited by R. B. Lee and I.
DeVore, pp. 293-303. Aldine, Chicago.
Webster, D., and G. Webster
1984 Optimal Hunting and Pleistocene Extinction. Human Ecology 12(3):275-289.
Wheat, J. B.
1972 The Olsen-Chubbuck Site: A Paleo-Indian Bison Kill. Memoirs of the Society for
American Archaeology No. 26.
Wilkie, D. S., and B. Curran
1991 Why Do Mbuti Hunters Use Nets? Ungulate Hunting Efficiency of Archers
and Net-Hunters in the Ituri Rain Forest. American Anthropologist 93:680-689.
147-176. Plenum Wilson, E. 0.

1972 The Insect Societies. Harvard University Press, Cambridge, Mass.
Winterhalder, B.
havior at Gilgil, 1987 The Analysis of Hunter-Gatherer Diets: Stalking an Optimal Foraging
Evolution, edited Model. In Food and Evolution: Toward a Theory of Human Food Habits, edited by M.
~rsity Press, New. Harris and E. B. Ross, pp. 311-339. Temple University Press, Philadelphia.
Winterhalder, B., and E. A. Smith
1981 Hunter-Gatherer Foraging Strategies: Ethnographic and Archeological Analyses .
n. Ballena Press . University of Chicago Press, Chicago.
Wolpoff, M. H.
1980 Paleoanthropology. Alfred A. Knopf, New York.
>ective. In Socio- Wrangham, R.
2.59-278. Aldine, 1975 The Behavioural Ecology of Chimpanzees in Gombe National Park, Tanzania.
Unpublished Ph.D. dissertation, University of Cambridge.
~rican Ethnology
niversity Press,
:::himpanzees in
ring and Hunting
~ki, pp. 303-343.
University, New
)n to Studies of
Ethnographic and
rvey of Alberta,
Mid continental
R. B. Lee and I.
':Y 12(3):275-289.
of the Society for
:iency of Archers
;t 93:680-689.
standards:
fore adapt:
populatior
about the 1
'to thehabi"
This res1
3. The Place of Hominids among sapiens bet
point of vi
Predators: Interspecific taken fro:
MacArthru
Comparisons of Food Procurement what rend
change in
and Transport tors. The f,
for mappi1
Mary C. Stiner By compa
framewod
two closel-
Abstract: This study investigates ungulate procurement strategies and cally simil:
skeletal part transport habits of a variety of social predators in order to rela tionshi
develop a comparative baseline for testing hypotheses about ecological late prey a
differ-=nces between Middle and Upper Paleolithic hominids. The analy-
sis focuses on the anatomical composition of faunal assemblages in shel-
ter and den contexts, representing the final destinations of transport
trajectories. Comparisons of modern predators document strong links
between skeletal representation in transported faunas and the principal
modes of procurement (hunting and scavenging).- Differences in trans-
port choices associated with each class of foraging strategy stem from the was prom
combined influences of differential persistence of skeletal parts at pro- Middle Pa
curement sites, fat-protein contents of certain prey body parts (particu- central It:
larly the cranium), and the spatial and seasonal charactdristics of food Moscerini1
supply. The scales at which the strategies are combined in time and record are
space distinguish predator adaptations to a large extent. The compara- Upper Pal
tive framework supplied by nonhuman predators reveals significant aspects an
niche separation between Middle and Upper Paleolithic hominids in FourteeJ
Mediterranean Europe. reveal twc
species, pr
known as
Introduction ungulate a
Numbero
Controversies about the European Middle Paleolithic generally each regie
center on whether N eandertals were like modern people or somehow graph wm
evolutionarily distinct in their adaptations. An interesting but difficult One anata
challenge underlies debates on this topic: no one would be surprised to learn sists almo
that hominid adaptations varied over time, but independent measures or tently assc
(Stiner 195
From Bones to Behavior: Ethnoarcbaeological and Experimental Contributions to the Interpretation of from Gro
Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights resembles
reserved. ISBN 0-88104-Q76-2. mandible~
38
Hominids among Predators I 39
standards are needed to systematically assess ecological"distance" and there-

fore adaptive change between Neandertal and anatomically modern human
populations. The overreaching problem concerns how to address hypotheses
about the behavioral ecology of any archaic hominid without sim:ply turning
to the habits of the sole modern representative for an explanation.
This research investigates possible shifts in the predatory niche of Homo
ng sapiens between roughly 120,000 and 10,000 years ago from an interspecific
point of view (Stiner 1990a, 1990b, 1991b, 1991c, 1991d). The study is under-
taken from the perspective of niche theory (sensu MacArthur 1968;
MacArthur and Levins 1967), andrather than working from the standpoint of
rement what rendered humans unique relative to other mammals, it traces adaptive
change in terms of some very basic rules affecting all large terrestrial preda-
tors. The foraging niches of social carnivores serve as independent measures
for mapping possible changes in human niche during the Upper Pleistocene.
By comparing Middle and Upper Paleolithic data sets within this larger
framework, it is possible to address questions of whether and to what extent
two closely related, quasi-contemporary hominid populations were ecologi-
t strategies and cally similar. This presentation focuses on only one dimension of foraging, the
:~.tors in order to relationship between predators' strategies for obtaining medium-sized ungu-
::~.bout ecological late prer' and the skeletal parts predators carry to shelters.
nids. The analy-
mblages in shel-
ms of transport Background
ent strong links
nd the principal
The comparative analysis of the food transport habits of predators
~rences in trans-
;y stem from the
was prompted by interpretive difficulties I encountered while studying
tal parts at pra- Middle Paleolithic faunal collections from four coastal caves in Latium, west-
y parts (particu- central Italy (Stiner 1990a): Grotta Breuil, Grotta Guattari, Grotta dei
teristics of food Moscerini, and Grotta di Sant' Agostino. Some aspects of this archaeological
ted in time and record are anomalous when compared to what we are accustomed to seeing in
t. The compara- Upper Paleolithic and later human sites in the same region, whereas other
reals significant q.spects are easily recognized as the products of hunting.
hie hominids in Fourteen Mousterian faunal assemblages from the four caves (Figure 3-1)
reveal two distinct patterns of anatomical composition for the same prey
species, principally red .deer (Cervus elaphus) and ari extinct form of wild cattle
known as aurochs (Bos primigenius). By way of example, Figure 3-2 divides the
ungulate anatomy into nine anatomical regions using standardized Minimum
Number of skeletal Element (MNE) counts (Table 3-1); the observed count for
ithic generally each region is divided by an expected value, meaning that the bars in the
~ or somehow graph would be of equaJ height if complete bodies of prey were represented.
; but difficult One anatomical profile (lower graph, a case from Grotta dei Moscerini) con-
·prised to learn sists almost exclusively of head parts and, in one cave sequence, is consis-
tt measures or tently associated with exploitation of marine shellfish and aquatic tortoises
(Stiner 1990a, 1993). The other kind of anatomical profile (upper graph, a case
'he Interpretation of from Grotta Breuil) verges toward anatomical completeness and closely
1tions, Occasional
rersity. All rights
resembles later Upper Paleolithic fa.unas of Italy and Spain. Crania and
mandibles occur in roughly equal proportions regardless of which general
I
40 M. C. Stiner
0 40
I I
.km
MNE
COBS)
EXP
Figure 3-1. Geographic distribution of four Middle Paleolithic (Mouste-

rian) caves sites in west-central Italy: (B) Grotta Breuil, (G) Grotta
Guattari, (M) Grotta dei Moscerini, (S) Grotta di Sant'Agostino. ' MNE
(OBS)
\_EXP
anatomical profile is represented, a fact whose significance will become
apparent in later discussion.
One kind of anatomical profile or the other dominates a stratigraphic
sequence of any given site, representing thousands of years of shelter reuse by
Neandertals. Available geological, faunal, and technological eyidence sug-
gests that this spatial variation in resource use represents adjustments by the
hominids to prevailing foraging opportunities in the area over time; coastal
habitats were certainly affected by any change in sea level (Stiner and Kuhn
1992). However, because the head-dominant anatomical pattern occurs only at
the older sites in the study sample, a linear shift in foraging adaptations with-
in the Middle Paleolithic cannot be firmly excluded from further consideration
in future research.
I began by trying to explain the dtfferences between the faunas in terms of
in situ bone destruction, archaeological recovery biases, and sample size.
While the influence of these factors cannot be ruled out entirely, they fall far
short, singly a,nd together, of accounting for the variation in the Middle Paleo- archaeolo1
lithic data set (Stiner 1990a, 1991b, 1991d). This leaves the strong possibility cavefaunc:
that the biases at least partly reflect hominid transport decisions. end of trm
Yet I also found that the varia,tion in the archaeological record could not be of foraginl
explained by ethnoarchaeological observations of modern human beings. The obtained.
level atwhich the archaeological differences are expressed is much coarser in CollapsJ
time and space than scales of observation commonly used in ethnoarchaeolog- solution f<
ical research; the variation of interest is apparent only by comparing many be necessa
) 40
I I
km
MNE
COBS)
EXP
olithic (Mouste-
~uil, (G) Grotta
gostino. MNE
(OBS)
\_EXP
e will become
L stratigraphic
hel ter reuse by
evidence sug-
stments by the
~r time; coastal
iner and Kuhn HORN/ NECK UPPER UPPER FEET
ANTLER FRONT HIND
. occurs only at
1ptations with- HEAD AXIAL LOWER LOWER
FRONT HIND
r consideration
Figure 3-2. Examples of the two anatomical representation patterns for
1as in terms of medium ungulates in Italian Middle Paleolithic cave sites.
j sample size.
ly, they fall far
Middle Paleo- archaeological assemblages from several sites. Moreover, each Mousterian
Jng possibility cave fauna represents debris from many foraging events mixed together at the
,, end of transport trajectories. Here I equate the subject shelters with some kind
·d could not be of foraging hub to which food was brought rather than the place where it was
an beings. ;fhe obtained.
mch coarser in Collapsing various ethnoarchaeological data sets might seem the obvious
moarchaeolog- solution for bringing them in line with the archaeological problem. It would
nparing many be necessary forth~ purposes of this study, however, t.o compile all Nunamiut
421 M. C. Stiner
Table· 3-1. Expected MNE Values (a) by Skeletal Element and by T

Anatomical· Region for Artiodactyls and (b) Expected Ratio Values
for Complete Artiodactyl Prey
b. Expected Ri
a. Expected MNE Values by Skeletal Element and by Anato_mical Region for
Artiodactyl Prey
Ratio Name
By Anatomical Eleme!lt By Anatomical Regiona
(H+H)/L
Shed antler 2 *HORN 2
Hom/unshed antler
HORN/L
1/2cranium 2
1/2 mandible 2 *HEAD 4
HEAD/L
Atlas 1
Axis 1 NECK 7
Cervical vertebra 5
Note: All ratios
Thoracic vertebra 13 aExpected valt
Lumbar vertebra 7 combination of
Sacral vertebra 1 AXIAL 49
Innominate 2
Rib 26 (e.g., Binforc
0' Connell e1
Scapula 2
Humerus 2 * UPPER FRONT 4 handful of Cc
cal contexts '
Radius 2 providing, a·
Ulna 2 * LOWER FROfiT 6 create. IndeE
Metacarpal 2 through stuc
adjustments
Femur 2 * UPPERHIND 2 how human
with environ
Tibia 2 tive archaeol
Calcaneum 2 as much as
Astragalus 2 * LOWERHIND 8 prehistoric ·s
Metatarsal 2 understand
nonhumanp
1st phalanx 8
2dphalanx 8 FEET 24
3d phalanx 8
1
Note: Adjustments have been made for cases involving equids as prey. T
aAsterisk= used to calculate the expected value for (H+H)/L. imen tal surv
some genera
for a varietJ
variation ob:
rt and by Table 3-l.-Continued

:io Values
b. Expected Ratio Values for Complete Artiodactyl Prey
cal Region for
Ratio Name Description Expected Value
cal Regiona
(H+H)/L Hom MNE + Head MNE 0.30
2 divided by Limbs MNE
HORN/L Hom or antler MNE < 0.10a

divided by Limbs MNE
4
HEAD/L Head MNE divided 0.20
by Limbs MNE
7
Note: All ratios employ raw (unstandardized) :MNE counts.

a Expected ;value will be less than 0.10 due to individual age in horned species and to a
combinatipn of age, sex, and season in cervids.
49
(e.g., Binford 1978), or Aka (Hudson 1990), or Badza (e.g., Bunn et al. 1988;
O'Connell et al. 1988a, 1988b), or !Kung data (e.g., Yellen 1977) into just a
handful of cases. Such a procedure would obscure a diverse array of ecologi-
:.'RONT 4
cal contexts and habitat-specific foraging strategies across the world without
providing, at this stage of knowledge, any meaning to the madness it would
FRONT 6 create. Indeed, attempts to collapse reference data can only be legitimized
through studies that explicitly integrate "pattern recognition," experimental
adjustments in scales of observation, and sound theoretical concepts about
-IIND 2 how human adaptations (and their archaeological consequences) should vary
with environment. .This has not yet happened to the extent required for effec-
tive archaeological application; it is a future problem for ethnoarchaeologists
as much as for other r_esearchers hoping to use the comparative data for
HIND 8 prehistoric studies. More of the appropriate kinds of information needed to
understand the Italian Middle Paleolithic cases presently exist for modern.
nonhuman predators, hence the analytical approach used here.
24
The Sample and the Methods
The archaeological problem dictates the scale for designing exper-
imental surveys of what other kinds of predators do. The purpose is to isolate
some general ru1e(s) about variation in patterns of anatomical part transport
for a variety of social carnivore species that might help account for the
variation observed in the Mousterian 'faunas. Here, I compare assemblages
441 M. C. Stiner
collected by six social predators: striped hyenas (Hyaena hyaena), brown the southe;
hyenas (Hyaena brunnea), spotted hyenas (Crocuta ·crocuta), wolves (Canis fallow deer
lupus), and Middle and Upper ~aleolithic humans. All cases come from caribou (a.l
shelters or excavated dens at which the transported products of many pro- Eastern cas
curement events were jumbled together. Brain (1981
In keeping with the archaeological problem, multiple cases are considered Some of the
for each bone-collecting predator. The carnivore den faunas are from all over smaller spe
the world; most are modern in origin, but the sample also includes several they are for
from Upper Pleistocene Italy. Unlike the curre~t situation for ethnoarchaeo- Measure:
logical data bases, information is available for understandi11g the causes of Bones.in ar
variation in the foraging habits of nonhuman predators by environment and, for reasons
in fact, proves quite useful for later interpretations (reviewed in Stiner 1990a, a wide arra
1991 b). The data sources are described and quantified elsewhere (Stiner reconstruct
1991b) and will not be reiterated here. The Middle and Upper Paleolithic sample on i
hominid cases-, forming the center of the investigation, are from features ("I
Mediterranean Europe only (Italy [Stiner 1990a, 1991b] and Spain_[Altuna :MNE repn
1986]). The geographical restriction on the hominid;..generated cases is in- occur in pa:
tended to control for variation in environment as much as possible, in 'effect The cou:
narrowing the potential array of causes that might account for differences in articular en
hominid foraging responses. and mandi'
All Pleistocene faunas have been attributed to a specific predator (human or or bony arc
nonhuman) in advance of this study, based on taphonomic evidence (Stiner relatively o
1990a, 1991d). The taphonomic analyses were accomplished in three succes- determinat
sive steps: (1) species representation served as a first indication of the poten- wherever 1
tial array of bone-collecting species that might have contributed to the forma- important f
tion of each assemblage; (2) patterns of bone damage (gnawing, salivary and ity in the f1
digestive erosion, burning, toolmarks, distinctive fracture patterns, weather- bone-based
ing, and abrasion) were used to identify the main bone collectors-modifiers often domi
and, to the extent possible, the sequence of modifiers if more t)lan one was and tarsals
indicated; and (3) the age structure of the common carnivore· species, and considerati1
presence of carnivore latrines, were used for evaluating assemblages sus- on relative
pected to represent den occupations. transportal
Because the taphonomic analyses, assays of bone preservation, and archae- their decisi1
ologist-imposed recovery biases are major undertakings for each site, readers Finally, i1
specifically interested in these issues should consult other publications on the- determini:n
data set (Stiner 1990a, 1991d, 1993). It is worth summarizing here that 1988; TurnE
recovery biases had a more serious effect on the anatomical contents of the for any of
faunas excavated between 1939 and about 1955 than did preservation condi- paten tial in
tions. Axial elements suffered most, one of the reasons for concentrating on comparisor
horn/ antler, head, and limb parts of prey in much of the presentation (see shaft·for an
also Stiner 1991b:459-463). In praise of older collections, the limitations set by faunas exc<
recovery practices of earlier decades in the study area were imposed in a complete n
systematic fashion and thus have not erased all possibility for gaining insights excavatiom
about Middle and Upper Paleolithic subsistence practices in coastal Italy. Paleolithic)
Each "case" considered in the anatomical comparisons represents one (or occupation
two similar) prey species from an assemblage. Medium-sized ungulate prey- features in ·
the most common taxa in all faunas-are considered almost exclusively. For all cases, tl
yaena), brown the southern European assemblages, this size category includes red deer,
wolves (Canis fallow deer (Dama dama), and ibex (Capra ibex). North American cases involve
:es come from caribou (a.k.a. reindeer, Rangifer tarandus), and in the African and Middle
s of many pro- Eastern cases, this size category includes Class II and III bovids as defined by
Brain (1981:275) and occasionally other taxa, such as domesticated donkey.
are considered Some of the variation outlined below may· not apply to significantly larger or
e from all over · smaller species (Stiner 1990a); certainly none of the distinctions are so clear as
Lcludes several they are for medium ungulates (see Stiner 1991b).
ethnoarcha~o Measures of anatomical representation are based on MNE counts for bone.
~ the causes of Bones in archaeological assemblages are typically smashed into many pieces
rironment and, for reasons beginning with the processing habits of people, followed by any of
.n Stiner 1990a, a wide array of other destructive phenomena. MNE represents an attempt to
~where (Stiner reconstruct the full, minimum array of whole animal bones represented in a
per Paleolithic sample on the basis of the fragments, using the most common unique skeletal
o"n, are from features ("portions" or 11landmarks") for each type of element. In this study,
Spain [Altuna MNE represents the sum of right and left sides for elements that naturally
ed cases is in- occur in pairs.
ssible, in effect The counts for limb bones employ unique landmarks on or near the
r differences in articular ends and any of a variety of distinctive bony features for the cranium
and mandible (e.g., petrous, occipital condyle, incisive,mandibular condyle,
ator (human or or bony itrchitecture of the maxilla or mandible behind the cheek tooth row if
vidence (Stiner relatively complete; see Stiner 1991b:460-462). Teeth are not used for the MNE
n three succes- determinations because they can distort head counts relat!ve to postcrania
·n of the poten- wherever preservation conditions are less than ideal. This point is very
d to the forma- important for understanding the anatomical comparisons: where comparabil-
g, salivary and ity in the face of differential preservation is a critical concern. Interestingly,
terns, weather- bone-based MNE counts for head parts are high in most of the assemblages,
ctors-modifiers often dominating the graphed profiles of anatomical representation. Carpals
~ than one was and tarsals other than the astragalus and calcaneum arc also excluded from
re species, and consideration (Table 3-1) because of their small size. The study instead focuses
;emblages sus- on relatively large bones that, in principle, might represent the minimum
tr:ansportable parcels genuinely attractive to foragers, thereby influencing
Jn, and archae- their decisions·to transport those items.
ch site, readers Finally, it should be acknowledged that an analyst's choice of landmarks for
lications on the determining limb MNE could affect the numbers perceived (Bunn and Kroll
z:ing here that 1988; Turner 1989), although it is not necessarily true as a matter of definition
contents of the for any of the bone-collecting predators considered.l In recognition of the .
ervation condi- potential impact of how MNE was determined in the study sample, controlled
~ncentrating on comparisons of the relative frequencies of epiphyseal landmarks versus mid-
esen ta tion (see shaft forami-nae (and cet:tain muscle attachment scars) were conducted for
nitations set by faunas excavated from three cave sites in west-central Italy, each involving
? imposed in a complete recovery of bone (Stiner 1991b:460-462). The faunas are from recent
~aining insights excavations at Grotta Breuil (Middle Paleolithic), Riparo Salvini (late Upper
:tstal Italy. Paleolithic), and.Buca della lena (a series of Upper Pleistocene carnivore deli
resents one (or occupations, primarily spotted hyena). The sites occur in shallow solution
.ngulate prey- features· in limestone bedrock, and they share similar sedimentary settings. In
~xclusively. For all cases, the epiphysis-based method yielded equivalent or slightly higher
46j M. C. Stiner
MNE counts for limbs relative to counts based on midshaft landmarks STR
HYE
(primarily foraminae). Sediments were screened at the, early excavations of
Moscerini, Guattari, and Sant' Agostir,to, but bones were retained or discarded
based on systematic but selective criteria (Stiner 1990a). While. such an
H
0
A
N [
analysis was not possible for collections excavated in earlier decades, the
results of the control study defend the use of the epiphysis-based counting
H
E
A
D [123'
method for the older collections from caves within the study area. ..
w
a... H
>- E
A
1- D
Variation in Anatomical Representation z L .
0::: E
w G
s
Figure 3-3 summarizes variation in anatomical representation for 1-
1-
medium-sized ungulates in faunas generated by the six pret?-ators at shelters. <(
a...
Predators are listed on the x-axis, and anatomical profile types appear on the NC
E 0
AM
y-axis. The profiles range from being head- and/ or horn-dominated to more A p
L L
y E
or less anatomically complete, and then on to slightly "leggy." The __chart T
E
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 scav-
engers (brown and striped hyenas), and cases generated by known obligate L
E
hunters (wolves) appear at the opposite corner in the bottom right. The G
s
generally diagonal arrangement of cases across the chart suggests that some
larger relationship exists between procurement strategies and part transport ·~
123•
choices.
Experimental situations in which obligate scavengers are provisioned with Fi~
whole medium-sized ungulate carcasses near their dens (outlier case in lower car:
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 sapiens. Thera
more often than other items due, at least in part, to what is usually l~ft over at foraging strat
death sites. Predators that primarily hunt move relatively more 1parts, and distinguish prt
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 He
and scavenging, such as spotted hyenas, show good representation through-
out the possible range. The same is true for the Middle Paleolithic series. Al
A surprisingly simple anatomical ratio, (H+H)/L, describes much of the horn/ antler ilJ
variation among cases (Figure 3-4), not only in the kinds of patterns typically interesting kir
produced by each predator but also in-the modes and ranges of variation. The sion before col
(H+H)/L index is obtained by combining unstandardized horn or antler plus ing processin~
head counts and then dividing the sum by the total unstandardized number relative freque
of limb elements (except the phalanges; see Table 3-1). It should be noted that 'shelter faunas.
the relative amounts of leg parts vary most, whereas heads and/ or horns are While both l
well represented in nearly all cases. scavenging, al
The ranges of variation in the {H+H)/L ratio are especially important for when their ~
seeing niche differences among the predators because the general tactics- Pleistocene), ~
hunting and scavenging-are not unique to any of the species considered and back only he a<
certainly will not, in and of themselves, distinguish subspecies of Homo consume and ,
aft landmarks STRIPED
HYENA
BROWN
HYENA
SPOTTED
HYENA
MP
HOMO
UP
HOMO WOLF
excavations of
~d or discarded
Nhile such an
~r decades, the
H
0
A
N
H
DOD[];
fT---i 1.----i 123456789; ~
~riD~ rr=J
E
A
)ased counting .• D
ea. ..
tjBd~
w
D.. H
>- E
A
r D
&
z L
0:: E
w G
s
1resentation for f-
f-
tors at shelters. <(
D..
[JJ[ffi][]J
NC
; appear on the
.
E 0
AM
.inated to more
123456789~: iiliD~: l.:lii1~
A p
l L
y E
gy." The chart T
E ~~
:otal number of
! lJWJlliJ llullllJ
B
. obligate scav- i
cnown obligate L
""'""'i """'"I
E I horn/antler
2 head
tom right. The G
s 3 neck
~
4 axial
5 upper front
sests that some 6
7
lower front
upper hind
I
l part transport 123456789:

8
9
lower hind
feet
: : : 123456789:
·ovisioned with Figure 3-3. Anatomical patterns for medium ungulate remains in
~r case in lower carnivore- and hominid-generated shelter faunas (from Stiner 1991b).
rybod y likes to
tuations subject
ls and/ or horns sapiens. The ranges of variation, on the other hand, can be tied to predominant
rally left over at foraging strategies in control cases for carnivores and thus in some way
1ore parts, and distinguish predatory adaptations more generally.
series is packed
Jmparison and
)f both hunting Hom- and Head-Dominated Faunas
tation through-
hie series. A high (H+H)/L ratio value can arise from elevated proportions of
es much of the horn/ antler and/or head parts, as apparent from Figure 3-3. Two potentially
Ltterns typically interesting kinds of information are obscured by this ratio and merit discus·
,f variation. The sian before considering variation among species. The first concerns the differ-
n or antler plus ing processing capabilities of hominids and hyenas. The second concerns the
trdized number relative frequencies of cranial versus mandibular elements in head-dominated
ld be noted that shelter faunas. ·
ld I or horns are While both horn- and head-dominated faunas suggest a heavy emphasis on
scavenging, all species of hyena tend to focus on horn (and apparently antler
y important for when their geographical ranges overlapped with cervids during the
eneral tacMcs- Pleistocene), whereas Middle Paleolithic hominids in Italy tended to bring
considered and back only heads (see also Stiner 1991d and associated discussion). Hyenas can
)ecies of Homo consume and completely digest virtualiy any bone tissue (Guthrie and Smith
481 M. C. Stiner .
!:::,.
(M4:29)
8~----------~----------~------------------~
6
H+H HEAD
L
L
4
EX
EXP
STRH BH SH MP UP w
Ncases: 2 3 11 16 45 6
Figure 3-4. The proportion of head and horn parts to limbs, (H+H)JL, for Fi
medium ungulate remains by predator (order on axis as in Figure 3-3; pa
from Stiner 1991b). fm
cal
1990; Kruuk 1972; Sutcliffe 1970), apparently making horns and antlers worth
collecting regardless of what else might be attached to them. Hominids, on the (but not the n
other hand, are very good at opening bone cavities but generally avoid ravaged carni
ingesting dense bone. It is also significant in the case of the Mo~~terian that combined to
elaborate processing technologies appropriate for grinding trabecular bone, only because
and containers for boiling food, simply did not exist (see also Speth 1989; Paleolithic fa·
Stiner and Kuhn 1992). that scavengi
Figure 3-5 illustrates the horn (/antler) to head contrast between hyena- and articulated sk
hominid-collected shelter faunas. When either horn or head parts are overly
abundant relative to limbs in shelter assemblages, the hyena-generated cases
tend to be strongly biased toward horn elements (high HORN /L; see Table 3- Fe
1), whereas hominid-generated cases. tend to be strongly biased toward head
elements (high HEAD /L). The tight cluster of points near the graph intercept Ta
consists primarily of Upper Paleolithic cases. (H+H)/L rati(
The second point, having to do with the relative abundance of cranial because few c
versus mandibular elements within the head region, may have implications occupy similE
for the sources of scavenging opportunities exploited by hominids. Actualistic for example,
studies by Haynes (1980, 1982) and Blumenschine (1986:35-38) suggest that if linear. ratio a
heads are obtained from predator kills the skull would often have been safely markedly aml
separated away fron1 the mandible, and the latter destroyed by primary provided on t
feeders. An expectation drawn from these actualistic cases is that the cranium by sample siz1
....
(M4:29)
*HYENA
.A. MP HOMO
• VARIOUS
. ' 4
HEAD *
L 3
*
2 ....
....
*
*
* *
*
oj) 2
EXP
w HORN
6 L
bs, (H+H)/L, for Figure 3-5. The relative frequencies of horn parts (HORN/L) to head
s in Figure 3-3; parts (HEAD/L) for medium ungulate remains in predator-collected
faunas. Outlier cases are marked according to predator; undesignated
cases conform to expected (EXP) values (from Stiner 1991b).
d. an tiers worth
)minids, on the (but not the mandible) is the most common prize available to scavengers from
enerally avoid ravaged carnivore kills. In this archaeological study, crania and mandibles are
tfousterian that combined to form the "head unit"-the numerator in the (H +H) /L ratio-
~abecular bone, only because they occur in r.oughly equivalent frequencies in the Middle
.so Speth 1989; Paleolithic faunas. Equivalent frequencies of crania and mandibles suggests
tHat scavenging opportuDities arose largely from nonviolent deaths, where
een hyena- and articulated skulls remained intact, rather than from ravaged predator kills.
~artsare overly
senerated cases
/L; see Table 3- Foraging Specialization and Predator Niches
~d toward head
graph intercept Table 3-2 presents lowest, highest, and· median values for the
(H +H) /L ratio by predator type. Striped and brown hyena cases are combined
ance of cranial because few cases are available for comparison and these predators appear to
ve implications occupy similar foraging niches in geographically separate regions (compare,
lids. Actualistic for example, Mills 1984a, 1984b; Skinner et al. 1980). Working first with a
) suggest that if linear ratio axis (Figure 3-6), it is clear that the ranges of variation differ
ave been s~fely markedly among predators and that the relationships are nonlinear. N values
ed by primary provide9 on the right side of the graph show that the ranges are not explained
II
:1at the cranium by sample size.
50 IM. C. Stine,r
Table 3-2. Summary Statistics for (H+H)/L by Predator Type
(H+H)/L striped/br
Predator Type N Low High Median
spo
Striped/brown hyena sa 0.82 28.00 4.64

MF
Spotted hyena 11 0.23 4.31 0.81
MPhominid 16 0.19 29.00 0.57 UP
UP human 45 0.12 1.25 0.43
Wolf 6 0.05 0.18 0.13
Note: Expected (H+H)/L value for complete carcasses= 0.30.

aArtificial provisioning case removed.
Obligate scavengers and species that both hunt and scavenge have much
wider ranges of variation. The scavenging element is largely responsible for
this because it emphasizes opportunism. The Mousterian series is the most
variable of all; some cases are head-biased whereas othei,"s are meaty or verge
on anatomical completeness. In contrast; Upper Paleolithic cases dating to
between 20,500 and 10,000 years ago display only the second aspect of the two
potential extremes, and the ranges are very narrow for both Upper Paleolithic
humans and wolves. The latter two predators almost certainly scavenged striped/br
from time to time, but if so, they usually had access to whqle carcasses,
suggesting they were better prepared to locate and monopolize windfall spo·
opportunities. i
Figure 3-7 presents the same data, using a logged ratio axis. The two
extremes of the (H+H)/L ratio correspond to habitual scavenging and habit- MP
ual hunting, as observed in modern nonhuman predator species. The vertical
dotted line on the graph represents the expected value for the natural prey UP
skeleton (see Table 3-1) and provides a baseline for comparison. This graph
deemphasizes the higher end of the (H+H)/L range, focusing instead on
variation below and just above the expected value.
The fact that some predator series display constricted ranges of variation
while others do not suggests that the differences relate to foraging niches in
some way. The next problem is to determine just how different is "different."
This question is directly relevant to the degree of foraging specialization
exhibited by each predator. The (H+H)/L ratio values are rank-ordered in
Table 3-3 and subjected to pair-wise comparisons using a Kruskal-Wallis one-
way ANOVA test for ranked data. This test makes fewer assumptions about
the data than parametric tests, circumventing the problem of nonlinearity in
the distributions. The rank-order comparisons reveal significant differences
between all predator pairs save one: Middle Paleolithic hominids and spotted
·Type exp
striped/brown
• I
N
hyena 5
Median
spotted
hyena 11
4.64 I
MPHomo 16
-
0.81 I
0.57 UP Homo 45
0.43 6
wolf '
0.13
0 10 20 30
H+H
-L-
Figure 3-6. Comparison of (H+H)/L medians and ranges for predator

nge have much
series using a linear ratio axis. N refers to number of cases; "exp" refers
/ to the expected value based on the living prey anatomy; and vertical line
responsible for
through each bar represents the median for that series.
ries is the most
meaty or verge
exp ~median
cases dating to
spect of the two
pper Paleolithic N
inly scavenged striped/brown
hyena 5
hole carcasses,
Jolize windfall spotted
*
hyena
· axis. The two

* 11
ging and habit-

ies. The vertical
MPHomo
* 16
he natural prey UP Homo 45

5on. This graph *
;ing instead on wolf.
*• 6
~es of variation
·aging niches in .03 0.1 10 30
tt is different."
11
H+H
-L-
~ specialization
ank-ordered in LIMBS COMPLETE HORNS/HEADS
;kal-Wallis one- HUNTING SCAVENGING
lmptions about
nonlineari'ty in Figure 3-7. Comparison of (H+H)/L medians and ranges for predator
~ant differences series using a logged ratio axis. N refers to number of cases; "exp" refers
.ids and spotted to the exp~cted value based on the living prey anatomy .
521 M. C. Stiner
Table 3-3. Results of Rank-Ordered Comparisons of Predators Based on there is a nutrit

(H+H)/L Ratio becomes an im
and where larg
carnivore kills c
Spotted MP UP (reviewed in S
Predator Type Hyena Hominid Human Wolf
crania and mar
consistent witr
Striped/brown )(2 =5.94 )(2 = 4.97 )(2 = 12.67 )(2 = 17.50 strong link to
hyena p = .015 p = .026 p < .001 p = .006 particularly latE
cooler parts of
Spotted )(2 =1.07 )(2 = 14.02 )(2 =11.01 condition.
hyena p = .3oa p < .001 p < .001 Soft cranial '
MPhominid )(2 =5.89 )(2 =12.53 tongue, have t
p = .015 p < .001 carnivorous fo
because, amon
UP human )(2 = 14.57 (proteinaceous
p < .001 et al. 1961); and
organs is quite
relatively unaff
aNonsignificant value (df = 1). for sustaining 1:
The food val
hyenas are essentially indistinguishable. It should be noted, however, that the terms of the rat
spotted hyena series is compiled across a large geographical cline extending for adult domE
from South Africa to Mediterranean (Upper Pleistocene) Europe, whereas the model only be
Middle Paleolithic series represents only one small region of Italy. However, the·
The chances are very slim of mistaking other predators with one another at buildup and los.
this scale of comparison. It is impossible, for example, to confuse the Middle stable fats in th
Paleolithic series with the Upper Paleolithic one; the range of variation is great sequence for p<
for the former and quite restricted for the latter, even though both ~~ries come although the s'
from generally analogous environments. 2 The data imply that the Middle domestic cattle
be leaner as a r
Paleolithic series is like that of the spotted hyenas in that these hominids
other species fc
relied on either scavenging or hunting, depending on local circumstances. In
contrast, striped and brown hyenas were dedicated to scavenging, whereas Figure 3-8 si
Upper Paleolithic humans and wolves were largely dedicated to hunting com paring the
existing fat (ri<
wherever they took up residence.
1986; NLSMB
range from a l
Nutritional Implications of Head-Collecting estimated level
fed and under
Why should scavenging hominids transport heads-and just about fat/protein val
only heads-to shelters? The outstanding proportion of head parts relative to and tongue) all
lower limbs and feet (see Figure 3-3) in the Italian cases is not especially Of the torso or
faithful to Speth's (1987) part choice model based on ungulate marrow deple- . organs, but thi~
tion sequences as a function of season. The persistence of crania at death sites Current moe
has been demonstrated in previous experimental research (e.g., Blumenschine by hominids fc
1986; Brain 1981; Haynes 1980, 1982) and could partly explain the attraction of Speth 1989:33
head parts for bone-collecting predators, especially scavengers. However, ungulates mete
~dators Based on there is a nutritional element as well, and that is where environmental setting
becomes an important consideration. In biomes ruled by freeze-thaw cy~les
and where large predator guilds remain intact, nonviolent deaths ratherr'fuan
carnivore kills appear to be the principal sources of scavenging opportunities
(reviewed in Stiner 199Gb, 1991c). The generally equivalent frequencies of
Wolf crania and mandibles in the head-dominated Mousterian faunas of Italy are
consistent with this interpretation. Nonviolent deaths in turn show a very
,)(2 = 17.50 strong link to starvation and tend to be concentrated in certain seasons,
p = .006 particularly late winter through late spring. Thus, scavenging opportunities in
cooler parts of the world may involve primarily animals in poor nutritional
X2 = 11.01 condition.
p < .001 Soft cranial tissues, such as the brain, thymus, and, to some extent, the
X2 = 12.53 tongue, have two unusual nutritional properties from the perspective of a
p < .001
carnivorous forager: (1) soft head tissues contain relatively high fat levels
because, among other things, they are the loci of myelinated nerve fibers
X2 = 14.57 (proteinaceous cells sheathed in fat, often called "white matter," e.g., Kimber
p < .001 et al. 1961); and (2) unlike muscle fat and yellow marrow, most fat in the head
organs is quite stable relative to changes in prey health. Fat in head parts is
relatively :unaffected by starvation, because neurological function is essential
for sustaiping life in all circumstances.
The food value of head organs can be compared to other soft body parts in
terms of the ratio of total fat to total protein (sensu Speth and Spielmann 1983)
:>wever, that the
for adult domestic cattle (Table 3-4). Domestic cattle are used as a general
cline extending
•pe, whereas the model only because extensive nutritional data are available for this taxon.
aly. However, the example approximates the relationship between the potential
h. one another at
buildup and loss differential for metabolizable fats in postcranial tissues and
fuse the Middle
stable fats in the head tissues of ungulates more generally. The fat depletion
sequence for postcranial soft tissues generally corresponds to marrow fat loss,
rariation is great
although the schedules of depletion vary somewhat (Speth 1990:152). While
Joth series come
domestic cattle are selectively bred to produce more fat and wild prey tend to
:hat the Middle
be leaner as a rule, the basic relationships among the tissues are analogous in
these hominids
other species for reasons of basic organ function .
.rcumstances. In
Figure 3-8 simulates the opposing nutritional states-fat and starved-by
~nging, whereas
comparing the fat/protein ratio in cattle muscle cuts, trimmed to 1/2 inch of
1ted to hunting
existing fat (rich) versus completely trimmed of all visible fat (lean; USDA
1986; NLSMB 1988). The fat to protein values in postcranial muscle mass
range from a high of 134% to a low of 30%. To keep things in perspective,
tg estimated levels in captive white-tailed deer (DelGiudice et al. 1990) in well-
fed and undernourished states change from 63% to 10%, respectively. The
-and just about fat/protein values for three kinds of soft head tissues of cattle (brain, thymus,
parts relative to · and tongue) all are on the high side of the differential for muscle fat in cattle.
s not especially Of the torso organs, only the pancreas displays fat levels equivalent to head
~ marrow deple- organs, but this fat would only be present in well-fed animals.
tia at death sites Current mode1s relating fat metabolism in prey to food transport decisions
;., Blumenschine by homi.nids focus simply on sequences of fat depletion in marrow (but see
the attraction of Speth 1989:337). The models are built primarily on observations that
tgers. However, ungulates metabolize both yellow marrow and postcr~nial soft tissue fats in
54! M. c. Stiner
Table 3-4. Summary Data on Fat and Protein
Content of Soft Cranial and Postcranial Tissues of
Adult Domestic Cattle per 100 g Portion
Total Fat/ CATTLE:

Soft Tissue Total Protein
fat
protein
Head organs
Brain 0.95
Thymus 1.67
Tongue 1.08
Muscle cuts, fully and partly

trimmed of metabolizable fat
Lean plus fat 1.34 DEER:

Lean 0.30
fat
protein
Torso organs
Liver 0.19
Pancreas 1.18
Spleen 0.16
Fi~
CrCi
Sources: USDA (1986); NLSMB (1988). tai
response to nutritional stress. If we look only at fat reserves that can be are resistant I
metabolized as a way of building transport choice models, we would expect much incenti·
starved prey to contain no appreciable nutritional value, or at most we would This is clear i
expect whatever fat remained in starved prey to be found mainly in the lower . reported hen
limb bones (e.g., Speth 1987). seasonal chat
Human beings have exceptionally large brains that continue to grow after [Binford 1978:
birth, and there is considerable neeq for fats during neurological development !Kung [Yellen
in utero and in children under five years of age (e.g., Crouch 1972). Those age
groups, along with pregnant and lactating mothers, constitute a large fraction
of human populations living in mobile and other traditional societies. Many of
the soft tissues of the prey cranium are always good food; equally important,
they are partly or wholly encased in bone just like marrow. The contents of H<
prey head parts therefore are r.elatively easy to extract without the benefit of framework p
tools for grinding and boiling food, in contrast to the situation for processing emphasis on
fat-rich trabeculae (e.g., Speth 1989; Yellen 1977). Head parts would be of by the cumuL
greatly increased value if the ungulate was in poor condition when it died or as variation i:
in any situation where fat is seasonally or chronically scarce. Moreover, heads since the lattE
MUSCLE HEAD ORGANS

5 of (unstable) (stable)
1.SO
'tal Fat/ CATTLE:

rl Protein 1.00
fat
protein
0.95 .so .so

1.67
1.08
rich lean brain thymus tongue
1.34 DEER:
0.30
I .so
tilt
protein .
?
0.19 rich lean

1.18
0.16 Figure 3-8. Comparative data on total fat to total protein content in soft
cranial organs and postcranial muscle mass of adult cattle and white-
tailed deer.
~rves that can be are resistant parts, and because they require considerable processing, there is
we would expect much incentive to move them to protected locations, like dens and shelters.
Lt most we would This is clear from the food transport habits of carnivores and, although not
:linly in the lower reported here, in ethnoarchaeological cases as well (compare, for example,
seasonal changes in the treatment of head parts by the Nunamiut Eskimo
nue to grow after [Binford 1978:77-84, 149-151] and nearly constant interest in head parts by the
;ical development !Kung [Yellen 1977]).
. 1972). Those age
te a large fraction
3ocieties. Many o~ Conclusion ·
qually important,
'· The contents of Holding preservation and recovery practices constant, the- general
out the benefit of framework provided by the (H+H)/L variable may diagnose the relative
on for prosessing emphasis on hunting and scavenging in predatory adaptations, as expressed
arts would be of by the cu.mulative products of food transport. The contrast is reflected largely
n when it died or as variation in the amount of limb bones relative to head and/ or horn parts,
Moreover, heads since the latter are a nearly universal features of the sh~lter faunas examined.
561 M. C. Stiner
On the other hand, treatment of head parts and the extent to which they stand Placing the co:
out in transport trajectories may also provide valuable 'clues about foragers' of predators pro·
responses to seasonal or chronic scarc.ity of fats, regardless of whether they are in animal comm
hunters, scavengers, or both. The two contingencies, procurement- strategy simply cannot g
and nutritional considerations, are difficult to separate, in this analysis, dichdtomy oftex
although it should be noted that ungulate age structures in the Mousterian one. At best, it c
head-dominated faunas point independently to scavenging (Stiner 1990b, by omnivorous J
199lc). The association of the two. contingencies in the control data suggests improving our t
. that foraging strategy and fat scarcity could be inextricably linked in natural · we are no long,
foraging situations, since hunters have more choices than scavengers both at were hunters or
the level of transport decisions and in the selection of live prey (nutritional separate large p
condition varies among individuals at any given time). of Homo sapiens.
The relationships between procurement and transport habits documented Two addi tio:
here emerge only at relatively broad scales of comparison, and it is at this hominid transp
level that they are relevant to the problem of distinguishing predator niches. tendencies in ar
The major differences described among predators, and between the Middle and the relatio
and Upper Paleolithic specifically,·are really about the scale at which variation marrow indexe~
in foraging strategies is expressed in time and space. We can expect that every or predicting p2
one of these predators scavenged, and most of them hunted too. Certainly, and/ or resourcE
Mousterian hominids were capable hunters, even if they did not always trials in the rea
choose to hunt. The Mousterian scavenging pattern stands out because the choices by livin
strategy was somehow spatially and temporally separated from hunting, a rather different
striking contrast to modern human situations where the products of both These data in t
strategies are pooled at the same residential places (e.g., Bunn et al. 1988; easier to compr1
O'Connell et al. 1988a, 1988b). Limited evidence indicates that Upper The problem
Paleolithic foragers in Italy also scavenged from time to time (e.g., involving a more far-rea
rarer species, such as horse; Stiner 1990a), but the visible fallout of that behav- regardless of w
ior was normally swamped by the more consistent and pervasive products of ethnoarchaeolo.
hunting at each shelter they used. 1 vation possible
In niche theory, the trophic relationships among species in communities are Unfortunately,
defined not only by the resources that the animals depend upon but also by assumption the:
how and when they use them (e.g., MacArthur 1968; MacArthur and Levins "seeing" beha\
1967; Root 1974; Wiens 1977). Ungulates are one set of resources many preda- archaeological ·
tors share an interest in but use differently. We can expect that head parts will are jumbled, rr
represent something special to carnivorous foragers whenever and wherever stricti y in the~
fat is scarce. We can also expect that ancient hominids needed a good bit of fat between specie
in their diets just as we do now, because fat is critical for neurological devel- and their prey,
opment in children, as well as for meeting caloric needs of human individuals relationships a
of all ages. The need has not changed, but the archaeological data indicate repeated. Henc
significant differences in how hominids went about getting fat during various viable match v
periods of the Upper Pleistocene and probably differences in how ungulate evolution. Dete
exploitation offset the use of other (perhaps now invisible) resources. This addressed. is p2
contrast is greatly reinforced by other lines of archaeological evidence (Stiner Ethnoarchae1
and Kuhn 1992), including prey age selection patterns (Stiner 1990b, 1991c), research on thE
small animal exploitation (Stiner 1990a, 1993), seasons of shelter use (Stiner be taken as cri t
1990a), as well as aspects of tool manufacture, reuse, and transport pertaining have already rr
to mobility (Kuhn 1990, 1991). needs to take,
Nhich they stand Placing the comparisons of hominid foraging habits within the larger world
; about foragers' of predators provides independent insights on the changing place of hominids
whether they are in animal communities, something that an exclusive focus on human beings
trement strategy simply cannot give. Such a perspective also shows that the simple heuristic
ln this analysis, dichotomy often made between hunting and scavenging is probably a false
L the Mousterian one. At best, it only narrowly describes one aspect of the resource base used
g (Stiner 1990b, by omnivorous hominids. This dichotomy has served its intended purpose by
·ol data suggests improving our understanding of what hominids are and were as foragers, but
linked in natural we are no longer dealing with questions of whether Mousterian hominids
avengers both at were hunters or scavengers or both. It is the rules of strategic combination that
?rey (nutritional separate large predator adaptations and perhaps also those of two subspecies
of Homo sapiens. ·
Jits documented Two additional points are worth summarizing, one about modeling
and it is at this hominid transport choices and another about ways of "seeing" behavioral
predator niches. tendencies in archaeological records. In modeling food transport by hominids
veen the Middle and the relationships of those choices to foraging tactics, it is clear that
t which variation marrow indexes alone are insufficient for ranking nutritional returns of parts
~xpect that every or predicting part transport choices, especially in situations where scavenging
d too. Certain! y, and/ or resource stress are suspected. Models are developed and improved by
did not always trials in the real world. In this study, interspecific comparisons of transport
out because the choices ~y living carnivores alongside nutritional analyses of head tissues set
from hunting, a rather different values on skeletal body parts of medium-sized ungulate prey.
nod ucts of both These data in turn make enigmatic faunas of the Italian Middle Paleolithic
3unn et al. 1988; easier to comprehend.
ttes that Upper The problem of isolating behavioral tendencies in archaeological records is
e (e.g., involving a more far-reaching concern.· When we collect primary data in the field,
tut of that behav- regardless of whether it involves archaeological or modern phenomena (as in
lsive products of ethnoarchaeology), we naturally and justifiably seek the finest scales of obser-
vation possible in order to allow for a multitude of futu:re uses of those data.
communities are Unfortunately, this very practical issue is sometime confused with an
1pon but also by assumption that finer scales of observation automatically bring us closer to
rthur and Levins "seeing" behavior as an adaptation or phenotype in the past. Very ancient
~ces many preda- arch~aeological records are usually disappointing for their lack of detail; they
lt head parts will are jumbled, mixed, or deflated, and we often rate their scientific potential
·er and wherever strictly in these terms (Stiner 1991a). Yet, some ecological· relationships
l a good bit of fat between specie~, such as between coexisting predators and between predators
urological devel- and their prey, are only perceptible at broad scales of comparison because the
tman individuals relationships are about the rates at which tactics, actions, and decisions are
cal data indicate repeated. Hence, the grain most typical of archaeological records presents a
1t during various viable match with many. of the problems of behavioral ecology and human
in how ungulate· evolution. Determining scales of observation appropriate to the problem being
) resources. This addressed is part of the analytical challenge.
evidence (Stiner Ethnoarchaeology presents exciting but largely unexploited possibilities for
.er 1990b, 1991c), research on the Paleolithic, its most commonly stated aim. The point need not
1elter use (Stiner be taken as criticism of ethnoarchaeological studies conducted to date-many
1sport pertaining have already made significant contributions. However, ethnoarchaeology now
needs to take, or at least pursue more vigorously, issues concerning spatial
. .
581 M. C. Stiner
and temporal scales at which observed behavioral patterns are meaningful Reft
with regard to human adaptations if it is to serve archaeological concerns
effectively. The questions of potentially meaningful scales of observation, and Altuna, J.
methods and contingencies for collapsing. data on foraging, for exanlple,can 1986 The Me
and should be ethnoarchaeologists' problems, not simply t9 be relegated to Cave: Stone ,
archaeologists hoping to use ethnoarchaeological data as bases of comparison. Straus & G.
University P
Binford, L. R
Acknowledgments 1978 Nunamiut
Blumenschine, R.
I thank Jean Hudson for her invitation to participate in the 1991 1986. Early R
From Bones to Behavior conference, hosted by Southern Illinois University, the Serengeti
Brain, C. K.
Carbondale, and for her efforts to see this paper through the publication pro-
1981 The Hu
cess. I am grateful to Lew Binford, Steve Kuhn, and Diane Gifford-Gonzalez
Bunn, H. T., and
for inspiration and critical evaluations during all of the stages of this work. 1988 Reply t
Thanks are also due to my European colleagues, A. G. Segre, E. Segre-Naldini, and Interpn
A. Bietti, P. Cassoli, D. Cocchi, M. Piperno, A. Radmilli, C. Tozzi, and G. Bunn, H. T., L. E.
Manzi. The final version of this paper has benefited from comments by 1988 Variah
Donald Grayson, Robert Leonard, R. Lee Lyman, James O'Connell, John ing, and Ca1
Speth, John Yellen, and two anonymous reviewers. Various stages of the Crouch, J. E.
research were supported by the American Association of University Women, 1972 Fundic
the L. S. B. Leakey Foundation, the Institute for International Education DelGiudice, G. [
(Fulbright Program), and a dissertation improvement grant (BNS-8618410) 1990 Effects
from theN ational Science Foundation. Profiles of v
Diamond,J. M.
1975 The A:
munities, ed
Notes Cambridge,
Guthrie, R. D., a1
1. Cats are not considered in this study, yet it is interesting to note that, in the
1990 A Con
author's unpublished research on modem mountain lion kill and scat a?~emblages
and the Pr
from New Mexico, limb shaft fragments from mule deer with intact foraminae
Council for
occurred in roughly equivalent proportions to limb epiphyses. While observations
Haynes, G.
that MNE counts can differ because of epiphysis-based versus central shaft-based
1980 Prey B
counting methods are important and useful means for evaluating potential biases, it
Bone Sites.
may be premature to assume that epiphysis-based counts are, as a matter of course,
1982 Utilize
inferior to midshaft counts in all carnivore-ravaged assemblages.
2. This analysis contrasts Middle and Upper Paleolithic series as whole temporal
Arctic 35:26
units. It is not designed to address the possibility of trends within the Middle Hudson,J.
Paleolithic and thus does not test hypotheses about the rates of change before, across, 1990 Advan
or after the ''Middle-Upper Paleolithic transition" as presently defined. Other kinds Aka Pygmi,
University
of analyses do in fact reveal trends within the Italian Middle Paleolithic, both in
Kimber, D. C., C
animal exploitation and lithic technology (see Stiner and Kuhn 1992). These trends
1961 Anat01
appear to shift in the direction of classically Upper Paleolithic exploitation patterns by
Kruuk,H. ·
the late Mousterian but in themselves do not constitute evidence of "vectored"
1972 The Sp
adaptational change in the study area. Moreover, it is not yet clear whether differ-
ences between Middle and Upper Paleolithic lifeways in coastal Italy really center on
Kuhn, S. L.
the "transition" period as defined elsewhere in Europe (Stiner 1990b, 1991c). 1990 Divers
Mousterian
Albuquerq
U'e meaningful References

)gical concerns
Jservation, and Altuna, J. ·
'r example, can 1986 The Mammalian Faunas from the Prehistoric Site of La Riera. In La Riera
be relegated to Cave: Stone Age Hunter-Gatherer Adaptations in Northern Spain, edited by L. G.
of comparison .. Straus & G. A. Clark, pp. 237-274 and (Appendix B) 421-479. Arizona State
University Anthropological Research Paper No. 36.
Binford, L. R
1978 Nunamiut Ethnoarchaeology. Academic Press, New York
Blumenschine, R. J.
ate in the 1991 1986 Early Hominid Scavenging Opportunities: Implications of Carcass Availability in
ois University, the Serengeti and Ngorongoro Ecosystems. BAR International Series 283. Oxford.
Brain, C. K.
ublication pro-
1981 The Hunters or the Hunted? University of Chicago Press, Chicago.
fford -Gonzalez Bunn, H. T., and E. M. Kroll
~s of this work.
1988 Reply to "Fact and Fiction about the Zinjanthropus Floor: Data, Arguments,
Segre-N aldini, and Interpretations;' by Lewis Binford. Current Anthropology 29(1):135-149.
Tozzi, and G. Bunn, H. T., L. E. Bartram, and E. M. Kroll
comments by 1988 Variability in Bone Assemblage Formation from Hadza Hunting, Scaveng-
'Connell, John ing, and Carcass Processing. Journal of Anthropological Archaeology 7:412-457.
s stages of the Crouch, J. E.
rersi ty Women, 1972 /Functional Human Anatomy. Lea and Febiger, Philadelphia.
mal Education DelGiudice, G. D., L. D. Mech, and U. S. Seal
(BNS-861841 0) 1990 Effects of Winter Undernutrition on Body Composition and Physiological
Profiles of White-Tailed Deer. Journal of Wildlife Management 54(4):539-550.
Diamond, J. M.
1975 The Assembly of Species Communities. In Ecology and Evolution of Com-
munities, edited by M. L. Cody and J. M. Diamond, pp. 342-444. Belknap Press,
Cambridge, Mass.
note that, in the Guthrie, R. D., and R. L. Smith
1990 A Comparison of Wolf and Spotted Hyaena Dietary Strategies in Bone Use
;cat assemblages
and the Production of Pseudoartifacts. Paper presented at the International
ntact foraminae
Council for Archaeozoology, Washington, D.C.
1ile observations
Haynes,G. .
:1tral shaft-based
1980 Prey Bones. and Predators: Potential Ecologic Information from Analysis of
otential biases, it
Bone Sites. Ossa 7:75-97.
natter of course,
1982 Utilization and Skeletal Disturbances of North American Prey Carcasses.
Arctic 35:266-281. ·
whole temporal
thin the Middle Hudson,J.
ge before, across, 1990 Advancing Methods in Zooarchaeology: An Ethnoarchaeological Study among the
ned. Other kinds Aka Pygmies. Ph.D. dissertation, University of California at Santa Barbara.
University Microfilms, Ann Arbor.
.eolithic, both in
12). These trends Kimber, D. C.~ C. E. Gray, C. E. Stackpole, and L. C. Leavell
1961 Anatomy and Physiology. MacMillan, New York.
atlon patterns by
Kruuk,H.
:e of "vectored"
1972 The Spotted Hyaena. University of Chicago Press, Chicago.
~ whether differ-
Kuhn, S. L.
{ really cent~er on
1990 Diversity Within Uniformity: Tool Manufacture and Use in the Pontinian
. 1991c).
Mousterian of Latium (Italy). Ph.D. qissertation, University of New Mexico,
Albuquerque. University Microfilms International, Ann Arbor.
I
L
I
60 M. C. Stiner
1991 "Unpaaking" Reduction:. Lithic Raw Material Economy in the Mousterian Predatory
of West-Central Italy. Journal of Anthropological Archaeology 10:76-106. pp. 149-1:
MacArthur, R. H. · 1991d A Ta
1968 The Theory of Niche. In Population Biology and Evolution, edit~d by R. C. Guattari (
Lewontin, pp. 159-176. Syracuse University Press, Syracuse. J993 SmaJ
MacArthur, R. H., and R. Levins Hunting,
1967 The Limiting Similarity, Convergence, and Divergence of Coexisting Anthropol
Species. The American Naturalist 101:377-385. Mellars, i
Mills, M.G. L. Stiner, M. C., a
1984a The Comparative Behavioural Ecology of the Brown Hyaena (Hyaena brun- 1992 SubE
nea) and the Spotted Hyaena (Crocuta crocuta) in the Southern Kalahari. Koedoe, Italy. Am1
Supplement:237-247. Sutcliffe, A.
1984b Prey Selection and Feeding Habits of the Large Carnivores in the Southern 1970 Spot
Kalahari. Koedoe, Supplement:281-294. 227:1110-
National Live Stock and Meat Board (NLSMB) Turner, A.
1988 Nutrient Values of Muscle Foods, Edition 1. Washington, D.C. 1989 Sam
O'Connell, J. F., K. Hawkes, and N. Blurton Jones . Partial RE
1988a Hadza Hunting, Butchering, and Bone Transport and Their Archaeological Klasies R
Implications. Journal of Anthropological Research 44(2):113-161. United States :
· 1988b Hadza Scavenging: Implications for Plio /Pleistocene Hominid Subsistence. 1986 Com
Current Anthropology 29(2):356--363. Human
Root, R. Washing
1974 Some Consequences of Ecosystem Texture. In Proceedings of the SIAM-SIMS Wiens, J. A.
Co11ference on Ecosystem Analysis and Prediction, edited by S. Levin, pp. 83-97. 1977 Onl
Philadelphia. Yellen, J. E.
Skinner, J. D., S. Davis, and G. Llani 1977 Cul1
1980 Bone Collecting by Striped Hyaenas, Hyaena hyaena, in Israel. Paleontology In Experi
in Africa 23:99-104. · pp. 271-~
Speth,J. D.
1987 Early Hominid Subsistence Strategies in Seasonal Habitats. Journal of
Archaeological Science 14:13-29. /
1989 Early Hominid Hunting and Scavenging: The Role of Meat flS an Energy
Source. Journal of Human Evolution 18:329-343.
·1990 Seasonality, Resource Stress, and Food Sharing in So-Called "Egalitarian"
Foraging Societies. Journal of Anthropological Archaeology 9:148-188.
Speth, J. D., and K. A. Spielmann
1983 Energy Source, Protein Metabolism, and Hunter-Gatherer Subsistence
Strategies. Journal of Anthropological Archaeology 2:1-31.
Stiner, M. C.
1990a The Ecology of Choice: Procurement and Transport of Animal Resources by Upper
Pleistocene Hominids in West-Centratitaly. Ph.D. dissertation, University of New
Mexico, Albuquerque. University Microfilms International, Ann Arbor.
1990b The Use of Mortality Patterns in Archaeological Studies of Hominid Pred-
atory Adaptations. Journal of Anthropological Archaeology 9:305-351.
1991a The Community Ecology Perspective and the Redemption of "Contami-
nated" Faunal Records. In Perspectives on the Past: Theoretical Biases in Mediter-
ranean Hunter-Gatherer Research, edited by G. A. Clark, pp. 229-242. University
of Pennsylvania Press, Philadelphia.
1991b Food Procurement and Transport by Human and Non-Human Predators.
Journal of Archaeological Science 18:455-482.
1991c An Interspecific Perspective on the Emergence of the Modern Human
Hominids among Predators I 6i
the Mousterian Predatory Niche. In Human Predators and Prey Mortality, edited by M. C. Stiner,
-106. pp. 149-185. Westview Press, Boulder, Colo.
1991d A Taphonomic Perspective on the Origins of the Faunal Remains of Grotta
?dited by R. C. Guattari (Latium, Italy). Current Ant~ropology 32(2):103-117.
1993 Small Animal Exploitation in the Italian Mousterian, and Its Relation to
Hunting, Scavenging, and Gathering. Archaeological Papers of the American
! of Coexisting Anthropological Association, edited by G. Larsen-Peterkin, H. Bricker, and P.
Mellars, in preparation.
Stiner, M. C., and S. L. Kuhn
ta (Hyaena brun- 1992 Subsistence, Technology, and Adaptive Variation in Middle Paleolithic
:a.lahari. Koedoe, Italy. American Anthropologist 94(2):12-46.
Sutcliffe, A.
in the Southern 1970 Spotted Hyaena: Crusher, Gnawer, Digester, and Collector of Bones. Nature
227:1110-1113.
Turner, A.
1989 Sample Selection, Schlepp Effects and Scavenging: The Implication of
Partial Recovery for Interpretations of the Terrestrial Mammal Assemblage from
· Archaeological Klasies River Mouth. Journal of Archaeological Science 16:1-11.
United States Department of Agriculture (USDA)
nid Subsi~tence. 1986 Composition of Foods: Beef Products. United States Department of Agriculture,
Human Nutrition Information Service, Agriculture Handbook No. 8-13.
Washington, D.C.
:the SIAM-SIMS Wiens,J. A.
evin, pp. 83-97. 1977 Qn Competition and .Variable Environments. American Scientist 65:590-597.
Yellen, J.1t. .
1977 ·Cultural Patterning in Faunal Remains: Evidence from the !Kung Bushmen.
ael. Paleontology In Experimental Archaeology, edited by D. Ingersoll, J. E. Yellen, W. Macdonald,
pp. 271-331. Columbia University Press, New York.
itats. Journal of
at as an Energy
~d "Egalitarian"
:8.
·rer Subsistence
!Sources by Upper
tiversity of New
Arbor.
:Hominid Pred-
51.
m of "Contami-
:iases in Mediter-
-242. University
1man Predators.
1-iodern Human
Aurignacian
the nature o1
understandi:
both fauna'
T~y 1989, 1~
Perigordian :
4. Uncovering Technological, Our apprc
of and proje
Organizational, and Seasonal dental anal,
elaphus) anc
Strategies of Paleolithic Paleolithic p
Hunting: Experimental strategies dt
(ca. 27,000-2
Contributions the hunting
projectile po
Anne Pike-Tay and Heidi Knecht
E
1
Abstract: This paper presents experimental aspects of ongoing research
concerning the nature of hunting during the Early Upper Paleolithic of 1
western Europe. Information gained from (1) replicative manufacture of
and projectile experimentation with organic projectile technology and (2) 1
analyses of teeth of modem control samples of red deer/wapiti and within the D
reindeer I caribou is examined in light of its potential contribution to the
is, for them
recognition of technological, organizational, and seasonal components of
Early Upper Paleolithic subsistence strategies. Aurignacian
B.P. and is fc
tation of the
Peri gordian
Introduction and VII, as
1978; Borde:
Some researchers (cf. Enloe 1992; Gamble 1986; Jacobs 1990; Lindly
1968; David
and Clark 1990:254; Smirnov 1990; Straus 1990) place "a significant adaptive
Two sorts
shift'' including a major change from "forager" to "logistical collector" (sensu
of Upper Pe
Binford 1980) subsistence strategies at around 20,000 B.P., that is, within the
tion with 1
Late Upper Paleolithic (LUP). The proponents of this post-20,000 B.P. adaptive
shift generally view subsistence systems of the Early Upper Paleolithic (EUP) Protomagcia
and (2) age-
much the same as the opportunistic forager pattern that has come, via Binford
(1984), to characterize the Middle Paleolithic (ca. 150,000 to 35,000 B.P.), the of the sites <
(levell), anc
cultural period of the Neandertals. For the LUP, the faunal evidence from a
number of Magdalenian sites (e.g., Bratlund 1990; Enloe and David 1989) 1
appears to be consistent with specialized reindeer hunting and a logistical
collector strategy. However, we do not yet have enough information from F
elements of
From Bones to Behavior: Ethnoarchaeological and Experimental Contributions to the Interpretation of Paleolithic h
tile points o
reserved. ISBN 0-88104-Q76-2. hooks. Obvi
traps or snm
62
Paleolithic Hunting Strategies I 63
Aurignacian and Gravettian sites to make a definitive statement concerning

the nature of EUP hunting tactics. It is our intent to contribute to an increased
understanding of hunting strategies during the EUP through assessment of
both fauna and technology. The present analysis focuses on red deer (Pike-
Tay 1989, 1991a) and organic projectile points (Knecht 1991) from Upper
Perigordian levels of sites in the Dordogne region of southwest France.
Our approach is an actualistic one based upon (1) replicative manufacture
of and projectile experimentation with EUP organic projectile points and (2)
al dental analyses of modern control samples of red deer /wapiti (Cervus
elaphus) and reindeer/ caribou (Rangifer tarandus), species that were important
Paleolithic prey. In this paper we limit our inquiry to the nature of hunting
strategies during one cultural period within the EUP, the Upper Perigordian
(ca. 27,000-23,000 B.P.), and focus primarily on red deer as prey. The nature of
the hunting technology is addressed through exploration of just one type of
projectile point, that is, bone single-be~eled points.
Background on Early Upper Paleolithic Hunting

Technology and Prey
ngoing research
'er Paleolithic of The Perigordian Study Sample
: manufacture of
:hnology and (2) The discussion that follows is _based on data from several sites
leer/wapiti and within the Dordogne region of France (Figure 4-1). Within the region, the EUP
ntribution to the
is, for the most part, represented by a succession of two cultural periods, the
J components of
Aurignacian and the Gravettian. The Aurignacian begins as early as 35,000
B.P. and is followed by the Gravettian at around 29,000 B.P. The local manifes-
tation of the Gravettian is usually referred to as the Upper Perigordian. The
Perigordian includes industries that have been labeled 'Perigordian IV, V, VI,
and VII, as well as Protomagdalenian and Aurignacian V (cf. Bordes 1958,
1978; Bordes and de Sonneville-Bordes 1966; Bricker and David 1984; Clay
bs 1990; Lindly
1968; David 1985; Movius 1975, 1977).
~icant adaptive
Two sorts of data have been combined to address the question of the nature
>llector" (sensu
of Upper Perigordian hunting strategies: (1) results of projectile experimenta-
t is, within the
tion with bone single-beveled points like those from Perigordian VI,
10 B.P. adaptive
Protomagdalenian, and Aurignacian V levels of the site of Laugerie-Haute
leoli thic (EUP)
and (2) age-at-death and season-of-death of red deer from Perigordian levels
ne, via Binford
s,ono B.P.), the of the sites of Le Flageolet I (level 7), La Ferrassie (level D2), Roc de Combe
ridence from a (level1), and Les Battuts (levelS).
j David 1989)
Projectile Technology
nd a logistical
)rmation from Primary to assessment of hunting techniques is recognition of
elements of hunting technology in the material record. Material evidence of
~e Interpretation of Paleolithic hunfing technology is usually found in the form of durable projec-
ttions, Occasional
tile points or other armatures of hunting weapons, such as harpoons or fish-
ersity. All rights
hooks. Obviously, devices constructed of perishable materials, for example,
traps or snares, wot;1ld be missing from the material record.
64j A. Pike-Tay and. H. Knecht
In EUP sites,
ivory), are usua
jectile points h
assemblages (C
Chafelperroniar
ing Chatelperr·
changes in hun
earliest uneq;ui·
antler projectilE
words, it seems
dent with the bE
Within the
Perigordian (ca
several kinds o:
stemmed lithic
employed d urir
2) were used c
Perigordian VI,
While only the '
have been reco'
may be postulai
tile points and F
Faut
In g•
western Europe
1983) indicate t
region, there is
The Dordogne
second point, C
Pataud and Roc
Dordogne and <
more abundant
of what Chase c
1989) argues fo:
nant prey sped•
evidence of intE
egy. In responsE
climatic phases
I, the frequencie
km Dordogne) do r
0 20 reflect more anc
Two very diff
seasonal moveJ
Figure 4-1. Map· of southwest France: 1 = Le Flageolet 1; 2 = La climatic change
Ferrassie; 3 = Roc de Combe; 4 = Les Battuts; 5 = Abri Pataud; 6 = ing or confusin~
Laugerie-Haute (after Delpech 1983:10).
In EUP sites, projectile points, eit!ter lithic or organic (i.e., bone, antler, or
ivory), are usually the only material evidence of hunting technology. No pro-
jectile points have been consistently identified among Middle Paleolithic
assemblages (Chase 1989), nor have they been definitively recognized in
Chatelperronian levels (ca. 35,000 B.P.; cf. Plisson and Schmider [1990] regard-
ing Chatelperronian points). This fact alone raises questions concerning
;:;:,) changes in hunting strategies from the Middle Paleolithic to the EUP. The
earliest unequivocal evidence of hunting technology in western Europe is
antler projectile points recovered from early Aurignacian contexts. In other
words, it seems that projectile technology makes its initial appearance coinci-
dent with the beginning of the EUP.
-( Within the Aquitaine Basin in, southwest France, throughout the
·~· ......
Perigordian (ca. 27,000-'24,000 B.P.; i.e., the local variant of the Gravettian),
several kinds of lithic and organic projectile points were used. For example,
stemmed lithic projectile points in the form of Font-Robert points were
employed during the Perigordian V and bone single-beveled points (Figure 4-
2) were used during the subsequent phases of the Perigordian (i.e., the
Perigordian VI, Protomagdalenian, and Aurignacian V; Knecht 1991, 1992).
While only the armatures of the hunting weapons, that is, the projectile points,
have been recovered from EUP sites, the nature of the projectile technology
may be postulated on the basis of results of replicative manufacture of projec-
tile points and projectile experimentation.
Fauna
In general, the faunal assemblages from sites throughout south-
western Europe (see Chase 1989:323-325, Fig. 19.1a, b, c and Fig. 19.2; Delpech
1983) indicate that during the EUP (1) with the exception of the Dordogne
region, there is no evidence of heavy reliance upon a single species, and (2)
The Dordogne evidence suggests an emphasis on reindeer. Regarding the
second point, Chase (1989) suggests that the reindeer-dominated sites of Abri
Pp.taud and Roc de Combe may be masking the true diversity of prey in the
Dordogne and could simply have been locations where reindeer were always
more abundant than other herbivores. These sites, then, would have been part
of what Chase calls an "eclectic" subsistence system. In contrast, White (1985,
1989) argues for relationships between site size, topographic location, domi-
nant prey species, and site seasonality. He views sites such as Abri Pataud as
evidence of intensified migration hunting of reindeer as a subsistence strat-
egy. In response to White's ideas, Delpech (1983:172) argues that even for the
climatic phases for which_ we have the fullest faunal documentation, the Dryas
I, the frequencies of reindeer bones at sites in southwest France (including the
Dordogne) do not correspond to topographic location or site size but rather
reflect more and less frequented habitats within the animals' range.
Two very difff?rent things are being addressed by the above researchers: (1)
seasonal movements of prey and (2) movements triggered by long-term
~eolet 1; 2 = La climatic change (i.e., shifts in species ra:J;lge). It is critical that we avoid conflat-
1bri Pataud; 6 = ing or confusing the two types of movement. There is an enormous difference
66 IA. Pike-Tay and H. Knecht
Met
Proje
Disc1
~I
typological grou
tics, is no lange:
importance of e:
· (cf. Arndt and N
1988; Cattelain
context of the p
use of Perigordi.
the nature of th
points, and (2) tl
Replicative eX]
a performed in ar
tion. Raw mater:
ous stages of co
b stigmata on ard
J L
) 1.-
production. ThE
through experirr
Projectile exper
of the hafting, u
preliminary exp
points was con
fonctionelle des p(
imental prograr
projectile technc
gists, faunal ana
to Paleolithic s·
launched with a
position.
For experimen
of force necessc:
initially determi
1990). The forcE
c d
constant. N otabl
Figure 4-2. Single-beveled poirzts from the site of Laugerie-Haute (Musee hand-thrown sp
National de la Prenistoire,,Les Eyzies). Lepetz 1988 for c
both mass and S]
it could have bee
between (1) specialization of hunting strategy (even seasonally) on one of lar spear, the ere
several available species and (2) ~ingle species emphasis as a by-product of always the same
ecological restrictions on availability. Our goal is to distinguish archaeologi- depth of penetra
cally the first approach, which is consistent with a collector strategy, from the the experiments 1
second, which is more consistent with opportunism. We hope to achieve this and hard tissues
by the methods outlined below.
Paleolithic Hunting Strategies 167
Methods
Projectile Experimentation
Discussion of the function of Paleolithic tools and weapons on
typological grounds, that is, on the basis of purely morphological characteris-
tics, is no longer viable. Archaeologists have become intensely aware of the
importance of experimentation in the technological study of material culture
(cf.. Arndt and Newcomer 1986; Barton and Bergman 1982; Carrere and Lepetz
1988; Cattelain 1986; Plisson and Geneste 1989; Stodiek 1990). Within the
context of the present discussion, experimentation in both manufacture and
use of Perigordian single-beveled bori.e points was carried out to ascertain (1)
the nature of the complete hunting weapons, that is, not just the projectile
points, and (2) the manner in which the weapons functioned.
Replicative experiments in the manufacture of single-beveled points were
performed in an attempt to determine the Perigordian sequence of produc-
tion. Raw material selected for production, manufacture waste, pieces at vari-
ous stages of completion, and microscopic and macroscopic manufacturing
stigmata on archaeological projectile points provided Clues to the sequence of
productjbn. The hypothesized manufacturing strategies were then tested
through experimental replication of the projectile points.
Projectile experiments were carried out to begin to come to an understanding
of the hafting, use, and performance of the single-beveled points. A series of
preliminary experiments in the use of several types of EUP organic projectile
points was conducted in conjunction with the program of Technologie
fonctionelle des pointes de projectiles pre1listoriques (TFPPP). This ongoing exper-
imental program has been designed to allow for the study of prehistoric
projectile technologies through the collaboration of archaeologists, technolo-
gists, faunal analysts, and others. Projectile points, identical in size and form
to Paleolithic specimens, are hafted to wood handles. These spears are
launched with a crossbow into fresh goat cadavers suspended in anatomical
position.
Fo,r experimentation with EUP bone and antler projectile points, the amount
of force necessary for the point to penetrate the hide of the animal was
initially determined (cf; Carrere and Lepetz 1988; Cotterell and Kamminga
1990). The for<;e with which the projectiles hit the animal was then kept
constant. Notably, the force used was equivalent to forces achieved by either
erie-Haute (Musee hand-thrown spears or spears projected with spearthrowers (see Carrere and
Lepetz 1988 for details concerning these variables). Since force is a function of
both mass and speed, a spear's mass would greatly affect the means by which
it could have been projected. For the experiments, given the mass of a particu-
nally) on one of lar spear, the crossbow was set at a tension such that the force of impact was
a by-product of always the same. For each shot, velocity of the projectile, location of contact,
lish archaeplogi- depth of penetration, and any damage to the projectile were noted. Following
trategy, from the the experiments, the animal was butchered and internal damage to both soft
>e to achieve this and hard tissues was observed.
68j A. Pike-Tay and H. Knecht
Reliabi.lity and Maintainability in Hunting Technologies Sease

Bleed (1986) has outlined a model that stresses reliability and ReceJ
maintainability as design alternatives Jor optimizing the availability 9f hunt- particularly cerr
ing weapons. He describes the characteristics of reliable and maintainable lated" teeth for c
systems as follows (Bleed 1986:Table 1): animals from ar
Reliable systems
outer root ceme
observed and in·
1. Overdesigned components. Recent advan
2. Understressed during use. (e.g., Burke 199(
3. Parallel subsystems and components, that is, redundant and standby and Albright 1S
parts are built into the system. factors, coi:npar1
4. Carefully fitted parts and generally good craftmanship.
the most impor
5. Generalized repair kit including basic raw materials. ·
6. Maintained and used at different times.
extend cementt
7. Maintained and constructed by a specialist. appropriate moe
issues surround
Maintainable systems composed of su
1. Generally light and portable. studied for a v'
2. Subsystems are arranged in series, that is, each part has one unique [Pike-Tay 1989,
function. general goal of t:
3. Specialized repair kit that irlcludes ready-to-use extra components. variables that af
4. Modular design. ing the regularit
5. Design for partial function. specimens (from
6. Repair and maintenance occur during use. to the Perigordia
7. Maintained by user.
8. Overall easily repaired, that is, "serviceable." 1."RE
both;
Reliable systems are generally more specialized than maintainable systems.
_latitu
They are advantageous when the period of use is scheduled: that is, for hunt- Paleo
ing weapons, when it is possible to determine the time at which the hunt will differ
occur or when game will be encountered. Construction and· maintenance dayle:
activities can be performed prior to the hunting "season" when the weapons 2.For
are not needed; reliable weapons will be ready to function for the scheduled annul
capture of predictable resources. May.
Maintainable hunting weapons are more generalized and are designed for 3. For
easy reparation and modification. Maintainability is desirable in systems that adjacE
are needed continuously and on unpredictable schedules. They are ready to of the
use at all times and in a variety of situations. Modular design and inter- 4.For
changeability of parts allow for e'asy repair and maintenance. The mal a
cons~
"serviceable" characteristics of maintainable weapons systems allow for main-
obsefl
tenance by the user, that is, the hunter, even when reparation during use is
necessitated. #>
It is Bleed's (1986) contention that maintainable weapons systems are Agel

appropriate for hunters using a forager approach for the capture of game,
whereas logistical collectors are beSt equipped with reliable weapons systems. Thea
In the present context, aspects of morphology, hafting techniques, and main- assemblages wer
tenance of Perigordian single-beveled points will be assessed in terms of the with eruption an
characteristics of reliability and maintainability outlined above. whenever those
ogies Seasonality: Dental Annuli Analysis

reliability and Recent analyses of the incremental structure of mammalian teeth,
lability of hunt- particularly cementum annuli analysis, underline the potential of even iso-
d maintainable lated teeth for determining age-at-death and season-of-death of individual
animals from archaeological contexts. Seasonal variation in the deposition of
outer root cementum and internal dentine in teeth has been systematically
observed and investigated by biologists for over sixty years (cf. Gordon 1991).
Recent advances in method and technique for archaeological application
(e.g., Burke 1990; Lieberman et al. 1990; Pike-Tay 1989, 1991a, 1991b; Savelle
lant and standby and Albright 1989) include quantification, controlled exploration of causal
factors, comparative microscopy, and computerized image analysis. One of
p. the most important advances is the recognition that before attempting to
extend cementum annuli analysis to archaeofaunas it is imperative that
appropriate modern control samples be compiled to clarify species dependent
issues surrounding annuli formation (Gordon 1982). Such control samples
composed of sufficient numbers of individual animals are presently being
studied for a variety of species (e.g., horse [Burke 1990], red deer /wapiti
t has one unique [Pike-Tay 1989, 1991a, 1992]; and reindeer/caribou [Pike-Tay 1992]). The
general goal of the study of control samples is twofold: (1) to assess the many
components. variables that affect annuli formation and (2) to test our assumptions regard-
ing the regularity of annuli formation. Data furnished thus far by the control
specimens (from Pike-Tay 1989, 1991a, 1991b) include information applicable
to the Perigordian sample discussed here:
1. "Rest" or "winter" band (annuli) formation and cessation times vary

both among species, and for cervids at least, according to geographic
:tinable systems.
latitude. This is cause for optimism in extending the technique to
that is, for hunt- PaleolithiC faunas because, while "seasonality" may have indeed been
ch the hunt will different during the Pleistocene, causal primacy may be related more to
1d maintenance daylength-triggered hormonal responses than to climate .
.en the weapons 2. For red deer/wapiti inhabiting regions from 40° to 60° north latitude,
)r the scheduled annuli (winter pand) formation begins in January and ceases by early
May. ·
ue designed for 3. For red deer/wapiti, the first wide growth band (zone) is deposited
! in systems that adjacent to the dentine boundary and forms pre-eruption. Determination
1ey are ready to of the time of the first increment's appearance is critical for proper aging.
~sign and inter- 4. For species of the Cervidae and Bovidae thus far examined, the opti-
intenance. The mal area for observing an even, undistorted deposit of cementum is
; allow for main- consistently that .nearest the tooth cervix. The most unreliable area for
observation is ~etween the roots and at the root apex.
)n during use is
ms systems are Age Profiles

apture of game,
eapons systems. The ~ge (mortality) profiles of the red deer from the Perigordian
ques, and main- assemblages were constructed from cementum annuli counts. They concurred
l in terms of the with eruption and wear estimates of th~ partial maxillae and hemi-mandibles
whenever those dental series were present. With the exception of one site,
I
70 A. Pike-Tay and H. Knecht
sample sizes have been too small for effective application of Klein's (Klein and org•
Cruz-Uribe 1984) crown height program (Pike~Tay 1991b~146). the
Hunting strategies may sometimes. be inferred from the reconstructed age age
profiles of Paleolithic prey populations. Following Klein (1982), Stiner (1991), sho·
and others, the interpretive framework adopted here defines three character- c. T
istic age profiles for ungulates; age profiles are compared to· attritional, catas- rela
trophic, and prime-age-dominated patterns. The attritional pattern has the bye
highest ·number of individuals in the youngest age class, then drops off or s
. steeply with a second rise in the oldest age class . Attritional patterns may coo·
reflect the relatively easy catch of youngest and oldest age cohorts or scaveng- 2. If, 0
by the
ing from natural deaths (as these gtoups exhibit the highest rate of mortality).
respm
The catastrophic pattern is characterized by a large number of individuals in shoulc
the youngest age class and gradually decreases in numbers down to the oldest a. T
age cohort, reflecting an ideal living population. Catastrophic patterns may be app
the result of a drive style hunt or a natural death assemblage generated by a by~
catastrophic event. The prime-age-dominated pattern may reflect sufficient logi
hunting skill to pick off the "biggest and best/ selection after a drive for the lirni
prime game, or selection when scavenging following a catastrophic event. b. .A
ally
Discussion of Methods dri'
unc
How, then, do the analytical tools of projectile experimentation, a nil
age structure of prey population, and season of exploitation of prey popula- at tl
tion enable us to distinguish the archaeological traces of specialization of c. T
hunting strategy on one of several available species from single species age:
emphasis as a by-product of availability? As noted previously, the behavioral The
sue:
correlates of the former case are consistent with a· collector strategy. In this
smc:
situation, specialization infers decision making based on the costs and benefits indi
of particular taxa. In the second situation, which is more consistept with a
versatile and opportunistic approach to hunting, single species/ emphasis
suggests a more general response to local (including seasonal) variability in
resources.
To operationalize our approach archaeologically, we pose the following Techn
hypotheses:
Exp~r
1. If the faunal sample being considered is the result of a specialized has demonstrate~
hunting strategy whereby one of several available species has been and E. Peyrony C
selected, then we should expec;:t: to a bevel with a
a. The technology and hunting techniques to be "reliable" (sensu Bleed
During experime
1986), that is, highly specialized and essentially failure free. Such tech-
adhesion. The irr
nologies would be recognizable in the archaeological record by the
presence of weapons such as barbed harpoons or points, multi- the cancellous b<
pronged spears, or a multiplicity of types of projectile points or other intentionally mac
devices, such as fishhooks, for the specialized capture of particular tl:le end of the h
subsistence resources. · points were som1
b. Since the sending out of task forces characterizes the collector strat- distal end of the
egy, the season-of-death should be consistent with times of aggregation slippage of the 1
for migratory species (e.g., fall and spring for reindeer/caribou), or if superior surfacE
ein' s (Klein and organized game drives have occurred, the"8eason-of-death should be
the same for several animals who together constitute a catastrophic
:onstructed age age profile (indicative of mass kills). Regardless of time of year, we
~),Stiner (1991), should expect seasonality to cluster in sites of collectors.
three character- c. The age profile to be either prime-dominated or catastrophic (with
ttritional, catas-. relatively high numbers of individual animals). Reflected is selection
pattern has the by considerably skilled hunter(s) for optimal yield (prime-dominated)
then drops off or sufficient skill to take all ages (catastrophic) in successful, probably
Ll patter_ns may cooperative drives.
::>rts or scaveng- 2. If, on the other hand, a versatile and opportunistic approach is taken
by the hunters, with single species emphasis suggesting a more general
te of mortality).
response to local (including se.asonal) variability in resources, then· we
f individuals in should expect:
wn to the oldest a. Technologies to be maintainable" and versatile, with generalized
11
patterns may be application to a variety of circumstances. They would be characterized

generated by a by spare or interchangeable parts that would be recognized archaeo-
~fleet sufficient logically by standardization of both size and shape. Moreover, only a
~ a drive for the limited variety of hunting devices/weapons would be present.
>phic event. b. As with logistical collectors, the season-of-death may cluster season-
ally since foragers also use sites in this manner. However, large-scale
drives or intensive high-yield hunting episodes would probably not be
undertaken. Therefore, season-of-death will reflect one or a few
xperimen ta tion, animals spread across a given season rather than many animals taken
of prey popula- at the same time (the latter being indicative of mass kills).
pecialization of c. The age profile to be mixed (i.e., a combination of attrition and prime-
t single species
age) and composed of relatively small numbers of individual animals.
·,the behavioral These are the ages that would be taken with devices and techniques
such as snares, traps, deadfalls, encounter ambush, stalking, and
strategy. In this small-scale surrounds, all of which involve a single animal or a few
>sts and benefits individual animals.
msistent with a
1ecies emphasis
11) variability in The Upper Perigordian as a Case Study
e the following Technology: Single-Beveled Points
Experimental manufacture and use of single-beveled bone points
: of a specialized has demonstrated the efficacy of the hafting mechanism first proposed by D.
species has been 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.
ble" (sensu Bleed
During experimentation, resin was applied to the beveled surfaces to increase
re free. Such tech-
:.:al record by the adhesion. The irregular sm::face of the bevel created by the natural openings of .
:>r points, multi- the cancellous bone from which it was manufactured (Figure 4-2a), or from
le points or other intentionally made striae, allowed for increased tenacity between the base and
ture of particular 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
:he collector'strat- distal end of the bevel (Figure 4-2d). It is likely that these incisions reduced
leS of aggregation slippage ·of the ligature with which th~ point was bound to the shaft. The
~er I caribou), or if superior surface opposite the distal end of the bevel of a number of
72IA. Pike-Tay and H. Knecht
(n=64)
25
4-!
Figure 4-3. Proposed technique for hafting single-beveled points.
Perigordian single-beveled points is flattened and marked with oblique or Figure

perpendicular incisions (Figure 4-2c). This combination of features would Laugeri.
allow for additional cohesion of lashing and/or adhesive. With the produc-
tion of textured bevels, incised shafts, and flattened surfaces, there appears to
have been an emphasis on maintaining the attachment of the haft in the Season1
design of Perigordian single-beveled points. This may reflect that the spear
was highly stressed at the level of the haft during use. The ba
Results of the projectile experiments suggest that armed with spears tipped surrounding the n
with single-beveled bone points Perigordian hunters would h~ve been because all past 1
capable of dispatching both large and small animals. With forces well within Dordogne has beeJ
the range of spears projected with spearthrowers, the spears were found to noted differences
easily enter and even pass through the animal. The projectile trajectory was from the region ar
unimpeded by the streamlined contours of the single-beveled points attached nomenon. Howe\
to single-beveled handles. important Upper
Given the range of single-beveled point lengths (Knecht 1991), it is probable First, since our ain
that the points were resharpened following breakage during manufacture, that additional·pn
handling, or use. Among the assemblage from Laugerie-Haute are several small portion of o:
single-beveled points that appear to have been resharpened (Figure 4-2b). It that of reindeer.
would be possible to resharpen a broken point in the field, that is, during a Season-of.:.death
hunting episode, without detaching it from the spear shaft. The relatively and 4-6) are from I
unrefined appearance of the retooled points may indicate ad hoc resharpening for study in part b
of the projectile points by a hunter. The tight range of variation of bevel angles a fair amount of J
(Figure 4-4; Knecht 1991) suggests the possibility of reusing the same spear sample was camp
shaft for the hafting of different single-beveled points. Spare projectile points teeth recovered w
could therefore have been carried by the hunter to serve as ready-to-use extra accounted for. Cor
components. ber of individuals
Appendix C) proh:
(n=64)
25
201---····'"---·---------------------··------······-··--···--···----·...............
15 .......................................
0
4-5 6-7 8-9 10-11 12-13 14-15
~led points.
Bevel Angle (0 )
~ with oblique or Figure 4-4. Distribution of bevel angles of single-beveled points from
f features would Laugerie-Haute.
With the produc-
' there appears to
1f the haft in the Seasonality and Age Profiles of Upper Perigord ian Prey: Red Deer
~ct that the spear
The background discussion above pertained primarily to issues
rith spears tipped surrounding the nature of reindeer hunting during the EUP in the Dordogne
rould have been because all past research concerning Upper Paleolithic subsistence in the
·orces well within Dordogne has been concerned with a single species emphasis on reindeer. We
.rs were found to noted differences of opinion as to why reindeer dominates the faunal lists
ile trajectory was frbm the region and even.derived our main research questions from this phe-
d points attached nomenon. However, most of the data presented here pertain to another
important Upper Perigordian cervid prey-the red deer-for two reasons.
)91), it is probable First, since our aim is to illuminate the nature of EUP hunting, it is important
ing manufacture, that additional prey species be considered. Second, the data represent just a
-Iaute are several small portion of ongoing research in which red deer analyses have preceded
:i (Figure 4-2b). It that of reindeer.
, that is, during a Season-of-death determinations q.nd age profiles of the red deer (Figures 4-5
ft. The relatively and 4-6) are from Pike-Tay (1989, 1991b). The faunal assemblages were chosen
hoc resharpening for study in part because of their excellent preservation. Although there were
on of bevel angles a fair amount of hemi-mandibles and partial maxillae, the majority of the
.g the same spear sample was composed of isolated teeth. All DP4, P4, Ml, and M2 red deer
e projectile. points teeth recovered were analyzed to insure that every individual red deer was
·eady-to-use extra accounted for. Conservative linking of teeth to determine the minimum num-
ber 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
LE FLAGEOLET I, level 7 LA FERRASSIE, level D2

late fall through late spring spring, summer
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 Summer late Sull}mer
4 0 4
Number of teeth sectioned =18 Number of teeth sectioned =17
Dental MNI = 12 Dental MNI = 12
RED DEER HUNT RED DEER HUNT
LES BATIUTS, levelS ROC DE COMBE, 1 (a,b,c)

winter 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
le
B.
late W/early Sp late W/early Sp
early Spring early Spring In sum, se<

Spring Spring
1.
late Spring late Spring (l,
late Sp/early S
le
late Sp/early S
b<
Summer
Summer ca
late Summer late Summer nc
0.2 0.4 0.6 0.8 0.2 0.4 0.6 0 ..8 1.2
2.
St
Dental MNI = 2
Number of teeth sectioned = 5 8(
Dental MNI = 3
RED DEER HUNT 3.
RED DEER HUNT
p<
d
4.
Figure 4-5. Seasons-of-death of red deer as determined by cementum 1)
annuli analysis by Pike-Tay (1989). d~
PERIGORDIAN SAMPLE
1RASSIE, level D2 age profile of RED DEER based on
,ring, summer
individual teeth
16 number of teeth sectioned
14 !-························
12 !--·······················
10
of teeth sectioned =17

)ental MNI = 12
2
D DEER HUNT
0
.5-2 3-4 5-6 7-9 10+
age classes
>E COMBE, 1(a,b,c) MIXED-AGE-PATTERN (n=29)

1er through early winter
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:
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 pyre-
naica) 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
'r of teeth sectioned = 5
Dental MNI = 3
80.1% (Delpech 1983:348).
3. Les Battuts (level 5) shows winter hunting of red deer. Red deer com-
::D DEER HUNT
pose !).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
ined by cementum 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 reasons previot
at Le Flageolet I, La Ferrassie, Les Battuts, and Roc de Combe show a broad (1) the age pro
mix of all age classes (an overlay of prime age and attritional patterns? [Figure (reflecting high
4-6]) derived from relatively small numbers of individuals in each level. Age tive fuscavengin~
profiles were constructed for each individual site, but as they are detailed aggregation for
elsewhere (Pike-Tay 1989, 1991b) and differences among them are slight, they for a substantia
are combined here. profile (indicat
forces would 1
Seasonality and Age Profiles of Upper Perigordian Prey: Reindeer hunting techni
redundant par1
Analysis of the reindeer from level4 of Abri Pataud (Spiess 1979), nized by the pr
which is roughly contemporaneous with the assemblages from which red deer ing special typE
were sampled, offers data that complement the findings outlined above. H a generaliz
Spiess (1979:185-208) claims that the reindeer-dominated fauna from all (1) the age prof
Upper Perigordian levels of Abri Pataud suggest fall, winter, and early spring and composed '
(cold season) hunting. He estimates an MNI of 26 reindeer for level4 (Spiess numbers that v
1979). . and deviCes sw
As regards the demography of hunted reindeer during the Upper small-scale sm
Perigordian at Abri Pataud, Spiess (1979:185) notes the consistently small (however, as la
MNis, sees the same demographic pa,ttern throughout all Upper Paleolithic not expected, tl
levels, and concludes: given season, 1
being suggesti'
1. The total demography of all bands of caribou wintering within range versatile, with
of the Abri Pataud was not significantly different in composition from
would be char
the total "natural" population. .
2. The hunters from the Abri Pataud randomly selected animals to be recognized arcl
killed, or used a hunting technique that did not allow selectivity (and We recognizE
was not a large-scale-drive, either). egy other relev
ence and absen
Whilereindeer was the dominant taxon at Abri Pataud, Spiess's a/~alysis of remains, estim2
faunal data from other species present led him to generalize for all levels: levels concerne
"Animals were killed singly or in small numbers, including everything from Our purpose he
chamois to 1500-kg cattle. There is no evidence of large-scale drives or slaugh- hand and to su~
ters" (Spiess 1979:185). In sum, our F
are:
Conclusions 1. A~
and
Which approach, then, was taken by the Upper Perigordian attril
hunters at the sites considered here? Are the data presented above, limited as all a
seco:
they are at present, consistent with specialization of hunting strategy or with a 2. Ba
more general response to local availability of resources? As noted previously, .dian
the behavioral correlates of the former, the collector type of approach, would no tic
include decision making based 011 the costs and benefits of particular taxa. sing:
The second situation is more consistent with a generalized and opportunistic evid,
foraging approach. Ther
In light of the Upper Perigordian data, we now return to the set of hypothe- spec:
ses pertaining to specialized collectors and generalized foragers. For the tile t
n levels sampled reasons previously elaborated, we would expect that for specialized collectors
Je show a broad (1) the age profiles of the prey would be prime-dominated or catastrophic
?atterns? [Figure (reflecting highly skilled individual hunters, successful game drives, or selec-
L each level. Age tive scavenging); (2) the season-of-death would be consistent with times of
hey are detailed aggregation for migratory species or, if game drives were employed, the same
n are slight, they 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
Prey: Reindeer hunting techniques would be "reliable;" that is, overly specialized with
redundant parts, essentially failure-free. Such technologies would be recog-
ud (Spiess 1979), nized by the presence of special purpose projectile points designed for hunt-
n which red deer ing special types of game or for use under special hunting conditions.
outlined above. If a generalized, opportunistic approach were taken, then we would expect
fauna from all (1) the age profile to be mixed (i.e., a combination of attrition and prime-age)
and early spring and composed of relatively small numbers of individual animals (the ages and
Jr level 4 (Spiess numbers that would be taken by sufficiently skilled hunters with techniques
and deviCes such as snares, traps, deadfalls, encounter ambush, stalking, and
~ing the Upper small-scale surrounds); (2) the season-of-death to also cluster seasonally
1nsistently small (however, as large-scale drives or intensive high-yield hunting episodes are
Tpper Paleolithic 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
~ring
within range versatile, with generalized application to a variety of circumstances. They
composition from would be characterized by spare or interchangeable parts that would be
:ted animals to be
recognized archaeologically by morphological standardization.
w selectivity (and We recognize that to ultimately distinguish foraging from a collector strat-
egy other relevant site data must be considered as well. Of interest are pres-
ence and absence of full residential residues in sites, body part bias in prey
less's analysis of remains, estimates of group size per site, degree of temporal resolution of the
ze for all levels: levels concerned, and season~of-death and age profiles for all species present.
everything from Our purpose here, however, is to offer a preliminary assessment of the data in
drives or slaugh- 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
per Perigordian attrition. Therefore, either the hunters were skilled enough to pick from
all ages or they cooperated to procure mixed age herds. However, the
above, limited as
second approa5=h is not supported by the low numbers of individuals.
;trategy or with a. 2. Based on the analysis of bone single-beveled points, Upper Perigor-
toted previously, dian projectile technology appears quite adequate to accommodate the
approach, would notion of fairly skilled hunters. Several characteristics of the Perigordian
f particular taxa. single-beveled points, including standardization of bevel angle and
nd opportvnistic evidence of retooling, suggest a portable, "serviceable" weapons system.
There does not seem to be evidence of special-purpose projectiles for
Le set of hypothe- special game or hunting conditions. In sum, Perigordian organic projec-
oragers. For the tile techl).ology appears to be, in Bleed's terqlinology, "maintainable"
78j A. Pike-Tay and H. Knecht
Roe, pp. 165-
and thereby indicative of a foraging hunting strategy rather than logis- Barton, R. N. E., a
tical collecting. -
1982 Hunten
3. Although seasons-of-death o.f the red deer from four sites and of rein-
World Archae
deer from the Abri Pataud during the Upper Perigordian appear· to
Binford, L.
cluster seasonally (with the exception of Roc de Combe), detailed break-
198h Willow
downs (Pike-Tay 1989, 1991b) show small numbers of animals taken
Archaeologi<
across a given cold or warm season during the Upper Perigordian rather
1984 Faunal1
. than many animals taken simultaneously. The spread of kills at La
Bleed, P.
Ferrassie (Figure 4-5b) across the warm SE?ason illustrates this well. Thus,
1986 The Op
the seasonal distribution of red deer kills is consistent with Spiess's
American Anj
(1979:185) interpretation of cold-season reindeer hunting at Abri Pataud: Bordes, F.
"Animals were killed singly or in small numbers, ... There is no evi-
1958 NouveU
dence of large-scale drives or slaughters." 62:205-244.
1978 Le Prot<
In conclusion, our preliminary data regarding age structure and season-of- la Societe Pre!
death of prey do not appear to be consistent with a collector strategy at least Bordes, F., and D.
as concerns the procurement of red deer during the Upper Perigordian at Le 1966 Protomc:
Flageolet I, La Ferrassie, Les Battuts, and Roc de Combe and of reindeer at Bratlund, B.
Abri Pataud. Upper Perigordian organic projectile technology seems to be 1990 The Hu
lacking several optimal design features characteristic of the weapons systems International
of logistical collectors. Much future work remains to be conducted to elicit the Bricker, H., and N
exact nature of EUP hunting techniques and strategies. However, the limited 1984 Excavati
evidence available at present suggests that a generalized and opportunistic 3) Assemblage
foraging approach was operationalized for the capture of game resources Burke, A.
during the Upper Perigordian. · 1990 Une Etu
dans le Sud·
Colloquelntt
Carrere, P., and S.
Acknowledgments 1988 Etude dt
Memoiredel
We wish to thank Jean Hudson, Jean-Michel Genestek James Cattelain, P.
O'Connell, Randy White, and an anonymous reviewer for valuable c~mp1ents 1986 Traces l
on an earlier version of this paper. We thank Fran<;oise Delpech and Jean- Helinium 26:1
Philippe Rigaud for generous access to faunal materials from southwest Chase, P.
France and Jean Guichard for kind permission to study the collection from 1989 How Di
Laugerie-Haute at the Musee National de la Prehistoire, Les Eyzies. The Perspective ,
Flageolet I data should be considered preliminary since the materials await Revolution, ec
further analysis and publication. Funding for the archaeological analyses was sity Press, Pri
provided by the W enner-Gren Foundation for Anthropological Research, the Clay, R. B. .
L. S. B. Leakey Foundation, the National Science Foundation (Grant #BNS- 1968 The Prot1
8711320), and New York University Dean's Dissertation Fellowships. Current of Anthropol,
funding for dental control sample analysis is from the National Science Cotterell, B., and J
Foundation (Grant #BNS-9023662). 1990 Mechani1
bridge.
David,N. ·
· 1985 Excavatil
References Assemblages t
Arndt, S., and M. Newcomer Prehistoric R1

1986 Breakage Patterns on Prehistoric Bone Points: An Experimental Study. In Delpech, F.
Studies in the Upper Palaeolithic of Britain and Northwest Europe, edited by D. A. 1983 Les Faun
Paleolithic Hunting Strategies j79
Roe, pp. 165-173. BAR International Series 296. Oxford.
tther than logis- Barton, R. N. E., and C. A. Bergman
1982 Hunters at Hengistbury: Some Evidence from Experimental Archaeology.
;ites and of rein-
I
World Archaeology 14:237-248. ·
~dian appear to
Binford, L.
, detailed break- 1980 Willow Smoke and Dogs' Tails: H~nter-Gatherer Settlement Systems and
£ animals taken Archaeological Site Formation. American Antiquity 44:4-20.
·rigordian rather ·
1984 Faunal Remains from Klasies River Mouth. Academic Press, Orlando.
d of ·kills at La Bleed, P.
; this well. Thus, 1986 The Optimal Design of Hunting Weapons: Maintainability or Reliability.
1t with Spiess's American Antiquity 51:737-747.
; at Abri Pataud: Bordes, F.
There is no evi- 1958 Nouvelles Fouilles a Laugerie-Haute Est: Premiers resultats. L'Anthropologie
62:205-244.
1978 Le ProtomagdalEmien de Laugerie-Haute Est (fouilles F. Bordes). Bulletin de
and season-of- la SoCiete Prehistorique Franqaise 75:501-521.
trategy at least Bordes, F., and D. de Sonneville-Bordes
rigordian at Le 1966 Protomagdalenien ou Perigodie VII? L'Anthropologie 70:113-122.
. of reindeer at Bratlund, B.
~y seems to be 1990 The Hunting Economy of Stellmoor. Paper presented at the Colloque
~apons systems Interpational: La Chasse dans la Prehistoire, Treignes, Belgium.
ted to elicit the Bricker, If·, and N. David ·
rer, the limited 1984 / Excavation of the Abri Pataud, Les Eyzies (Dordogne): The Perigordian VI (Level
i opportunistic 3) Assemblage. American School of Prehistoric Research Bulletin 34.
Burke, A.
;arne resources
1990 Une Etude Squelettochronologique des Equides au Paleolithique Superieur
dans le Sud-Ouest de la France: Etude Preliminaire. Paper presented at the
Colloque International: La Chasse dans la Prehistoire, Treignes, Belgium.
Carrere, P., and S. Lepetz
1988 Etude de la Dynamique des Pointes de Projectiles. Elaboration d'une Methode.
Memoire de Maitrise de l'Universite de Paris !-Pantheon Sorbonne.
;eneste, James Cattelain, P.
~able comments 1986 Traces Macroscopiques d'Utilisation sur les Propulseurs Paleolithique.
pech and Jean- Helinium 26:193-205.
·om southwest Chase, P. .
::ollection from 1989 How Different Was Middle Paleolithic Subsistence? A Zooarchaeological
es Eyzies. The Perspective on the Middle to Upper Palaeolithic Transition. In The Human
naterials await Revolution, edited by P. Mellars and C. Stringer, pp. 321-337. Princeton Univer-
al analyses was sity Press, Princeton.
11 Research, the Clay, R. B. .
t (Grant #BNS-:-
1968 The Proto-Magdalenian Culture. Unpublished Ph.D. dissertation, Department
V'ships. Current of Anthropology, Southern illinois University, Carbondale.
Cotterell, B., and J. Kamminga
:ttional Science
1990 Mechanics of Pre-Industrial Technology. Cambridge University Press, Cam-
bridge.
David, N.
1985 Excavation of the Abri Pataud, Les Eyzies (Dordogne): The Noaillian (Level 4)
Assemblages ·and the Noaillian Culture in Western Europe. American School of
Prehistoric Research Bulletin 37. ·
Delpech, F.
11ental Study. In 1983 Les Faunes du Paleolithique Superieur dans le Sud-~uest de la France. Cahiers
edited by D. A.
I
80 A. Pike-Tay and H. Knecht
du Quatemair~ No. 6. CNRS, Paris. 1991a L'Anal

Enloe, J. Paleo, 3:149-
1992 Subsistence Organization in the Early Upper Paleolithic: Reindeer Hunters 1991b Red Dl
of the Abri du Flageolet, Couche V. In Before Lascaux: The Complex Record of the Seasonality. ·
Early Upper Paleolithic, edited by H. Knecht, A. Pike-Tay, and R. White. CRC 1992 Develc
Press, Boca Raton, Fla., in press. 'CrownHei~
Enloe, J., and F. David of the Sociel
1989 Le Remontage des Os par Individus: le Partage du Renne chez les Plisson, H., and ]
Magdaleniens de Pincevent. Bulletin de la Societe Prehistorique Franqaise 86(9):275- 1989 AnalyE
281. rente), du F
Gamble, C. (Dordogne).
1986 The Palaeolithic Settlement ofEurope. Cambridge University Press, Cam- Plisson, H., and 1
bridge. 1990 Etude j
Gordon, B. Renne a Ar
1982 Tooth Sectioning as an Archaeological Tool. National Museum of Man, ologique. In
Archaeological Survey of Canadq, Canadian Studies Report No. 14. pp. 309--312.
1991 Archaeological Seasonality Using Incremental Structures in Teeth: An Annotated Savelle, J., and D
Bibliography. Zooarchaeological Research News, New York. 1989 Campa
Jacobs, K. · Analysis. Pc;
1990 Comment on "Symbolism and Modem Human Origins" by J. Lindly and canArchaec
G. Clark. Current Anthropology 31(3):233-261. Smirnov, Y.
Klein,R. 1990 Comm1
1982 Age (Mortality) Profiles as a Means of Distinguishing Hunted Species from G. Clark. Cu
Scavenged Ones in Stone Age Archaeological Sites. Paleobiology 8:151-158. Spiess, A.
Klein, R., and K. Cruz-Uribe 1979 Reindee
1984 The Analysis of Animal Bones from Archeological Sites. University of Chicago York.
Press, Chicago. Stiner, M.
Knecht, H. 1~91 The Us
1991 Technological Innovation and Design During the Early Upper Paleolithic: A atory Adapt
Study of Organic Projectile Technologies. Unpublished Ph.D. dissertation, Depart- Stodiek, U.
ment of Anthropology, New York University. · 1990 Jungpa
1992 Early Upper Paleolithic Approaches to Bone and Antler Proje¢tile Tech- such. Experi
nology. Archaeological Papers of the American Anthropological Association; in press. Mitteilunge
Lieberman, D., T. Deacon, and R. Meadow Naturkunde
1990 Computer Image Enhancement and Analysis of Cementum Increments as Straus, L.
Applied to Teeth of Gazella gazella. Journal of Archaeological Science 17. 1990 Comm1
Lindly, J., and G. Clark G. Clark. Cu
1990 Symbolism and Modern Human Origins. Current Anthropology 31(3):233- White, R.
261. 1985 Upper I
Movius,H. tence and Set;
1975 Excavation of the Abri Pataud, Les Eyzies (Dordogne). American School of 1989 Husbar
Prehistoric Research Bulletin 30. the EvidencE
1977 Excavation of the Abri Pataud, Les Eyzies (Dordogne). Stratigraphy. American
School of Prehistoric Research Bulletin 31.
Peyrony, D., and E. Peyrony
1938 Laugerie-haute pres des Eyzies (Dordogne). Archives de l'Institut de Paleon-
tologie Hurnaine Memoire 19.
Pike-Tay, A.
1989 Red Deer Hunting in the Upper Paleolithic of Southwest France: A Seasonality
Study. Unpublished Ph.D. dissertation, Department of Anthropology, New York
University.
1991a L'Analyse du Cement Dentaire chez les Cerfs: L' Application en Prehistoire.
Paleo, 3:149-166. ,
Reindeer Hunters 1991b Red Deer Hunting in the Upper Paleolithic of Southwest France: A Study in
'11plex Record of the Seasonality. BAR International Series 569, Tempus Reparatum. Oxford.
ld R. White. CRC 1992 Developing a Control Sample for Archaeology of Dental Annuli and
Crown Heights in Rangifer tarandus. Paper presented at the 57th Annual Meeting
of the Society for American Archaeology, Pittsburgh.
. Renne chez les Plisson, H., and J.-M. Geneste
'ranqaise 86(9):275- 1989 Analyse Technologique des Pointes a Cran Solutreenes du Placard (Cha-
rente), du Foumeau du Diable, du Pech de la Boissiere et de Combe Sauniere
(Dordogne). Paleo 1:65-106.
rsity Press, Cam- Plisson, H., and B. Schmider
1990 Etude Preliminaire d'une Serie de Pointes de Chatelperron de la Grotte du
Renne a Arcy-sur-Cure: Approche Morphometrique, Technologique et Trace-
vfuseum of Man, ologique. In Paleolithique Moyen Recent et Paleolithique Superieur Ancien en Europe,
J;14. pp. 309-312. Memoires du Musee de Prehistoire d'lle de France No. 3.
'eeth: An Annotated Savelle, J., and D. Albright
1989 Comparison of Preparation and Examination Techniques in Dental Annuli
Analysis. Paper presented at the 54th Annual Meeting of the Society for Ameri-
' by J. Lindly and can Archaeology, Atlanta.
S:mimov, Y.
1990 /Comment on "Symbolism and Modem Human Origins" by J. Lindly and
~nted Species from G. Clark. Current Anthropology 31(3):233-261.
lj 8:151-158. Spiess, A.
1979 Reindeer and Caribou Hunters: An Archaeological Study. Academic Press, New
rersity of Chicago York.
Stiner, M.
.1991 The Use of Mortality Patterns in Archaeological Studies of Hominid Pred-
pper Paleolithic: A atory Adaptations. Journal of Anthropological Archaeology 9:305-351.
:;sertation, Depart- Stodiek, U.
1990 Jungpalaolithische Speerschleudern und Speere-ein Rekonstruktionsver-
~r Projectile Tech- such. Experimentelle Archtiologie in Deutschland. Sonderdruck aus Archaologische
·ociation, in press. Mitteilungen aus Nordwestdeutschland Beiheft 4. Staatlichen Museum fiir
Naturkunde und Vorgeschichte.
urn Increments as Straus, L.
nee 17. 1990 Comment on "Symbolism and Modem Human Origins" by J. Lindly and
G. Clark. Current Anthropology 31(3):233-261.
·apology 31 (3):233- White, R.
1985 Upper Paleolithic Land-Use in the Perigord: A Topographic Approach to Subsis-
tence and Settlement. BAR International Series 253. Oxford.
nerican School of 1989 Husbandry and Herd Control in the Upper Paleolithic: A Critical Review of
the Evidence. Current Anthropology 30(5):609-632.
igraphy. American
nstitut de Paleon-
:mce: A Seasonality
pology, New York
information for
learn from the1
this paper is to
nation of these
Faunal analy
questions, app:
5. Ethnoarchaeology of Marrow and Cruz-Urih
of species (e.g
Cracking: Implications for the Olsen 1964; Scl
paleoenvironrr
Recognition of Prehistoric issues of archc:
Subsistence Organization 1987; Noe-Nyg
the contexts of
another impor1
James G. Enloe ogy, that is, int
Other studies :
Abstract: HunterI gatherer subsistence varies. in organizational strategy, 1954, 1955; Yel:
both intraculturally and interculturally, which should be manifest in sentation. Fe~
patterning of faunal remains. Certain resources may be entirely con- patterns on be
sumed upon acquisition, before other resources are sought. Other 1957; Frison ~
resources may be acquired in sufficient quantity for mass processing and researchers ha~
deferred consumption through storage. This contrast in consumption examination oJ
practices can be exhibited in the treatment of bones from which marrow
and Kroll (1981
is extracted, contrasting (1) marrow cracking that is incidental to eating
the meat off the bones with (2) that which is systematic mass processing fragments and
outside of the context of cooking and consuming meat. problem furthE
Three samples of faunal materials from two ethnohistorically known Our ability 1
Nunamiut Eskimo sites are used to develop recognition criteria. One strategies in th
sample is drawn from a bone dump associated with mass marrow pro- inferences are
cessing; the other two are drawn from kitchen middens that we7e accum- agreements ar
ulated as results of direct consumption activities. The sam,ples were faunal assemb]
analyzed and contrasted for attributes that should be rnonitorable in ers. Recent eth
archaeological assemblages. Significant differences between mass pro- often jump to <
cessing and consumption were noted in element representation, frag- to behavior or
ment size, and variation in fragment size. These criteria can be employed provide a brid
to distinguish between residues resulting from different processing tech- in archaeologi
niques and can lead to the identification of prehistoric subsistence
behavior resp<
strategies in archaeological contexts.
ology and the
What are the
Introduction inform on the
What, after unretouched lithic debitage, could be more ubiquitous have relevancE
and underutilized in the archaeological record of prehistoric hunter I gatherers One of the r
than long bone shaft splinters? They represent a largely untapped source of has been Binf
(1980), which:
From Bones to Behavior: Ethnoarchaeological i:md Experimental Contributions to the Interpretation of tence pursuit~
Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional mobility (Kell
Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights (Enloe 1993; E1
reserved. ISBN 0-88104-076-2. intended to fir
82
Subsistence Organization I 83
information for the zooarchaeologist. What do we do with them? What can we

learn from them about the behavior of prehistoric humans? The purpose of
this paper is to suggest one potential analytical goal derived from the exami-
nation of these fragments from the archaeological record.
Faunal analysis is a growing field of investigation, covering a wide range of
questions, approaches, techniques, and analytical goals (Chaplin 1971; Klein
w and Cruz-Uribe 1984). Issues range from the identification and differentiation
of species (e.g., Brown and Gustafson 1979; Gilbert 1980; Lawrence 1951;
:he Olsen 1964; Schmidt 1972) to animal population dynamics, paleoecological or
paleoenvironmental issues (e.g. Grayson 1984), or to the critical taphonomic
issues of archaeological site formation processes (e.g. Burgett 1990; Lyman
1987; Noe-Nygaard 1977). While they are very important for understanding
the contexts of human adaptation and interpretation of bone accumulations,
another important focus concerns a more anthropological aspect of archaeol-
ogy, that is, interpretation of faunal assemblages in terms of human behavior.
Other studies have examined butchering techniques (e.g. White 1952, 1953,
ational strategy, 1954, 1955; Yellen 1977), often focusing on body part segmentation and repre-
l be manifest in sentation. Few have dealt specifically with bone splinters per se. Fracture
be entirely con- patterns on bones have been used to search for or identify tools (e.g., Dart
~ sought. Other 1957; Frison 1970, 1974,· 1978; Kehoe 1967; Sadek-Kooros, 1972). Other
s processing and researchers have more directly addressed the problems of butchering through
in consumption examination of bone splinters and shaft fragments. Binford (1978) and Bunn
n which marrow and Kroll (1986) have struggled with the proportional representation of shaft
idental to eating
mass processing fragments and articular ends, while Todd and Rapson (1988) have carried this
problem further to investigate spatial patterning within archaeological sites.
:;torically known Our ability to make interpretations about the organization of subsistence
.on criteria. One strategies in the archaeological record depends of necessity on inference. The
ass marrow pro- inferences are all too frequently based on conventions for interpretation, tacit
:hat were accum- agreements among archaeologists as to how to read the patterning in the
te samples were faunal assemblages to understand lifeways of prehistoric hunters and gather-
~ monitorable in ers. Recent ethnoarchaeological work (e.g., Lupo 1990) has shown us that we
:ween mass pro- often jump to conclusions without sound bridges linking bones in the ground
·esentation, £rag- to behavior on the ground. The research described here was undertaken to
can be employed provide a bridging argument between material remains that might be found
· processing tech- in archaeological sites and an understanding of the organization of human
oric subsistence
behavior responsible for their creation and deposition. This is the method-
ology and the purpose of ethnoarchaeology: linking statics with dynamics.
What are the characteristics observable in static material remains that can
inform on the nature of dynamic behaviors in the past, behaviors that may
1ore ubiquitous have relevance to some theoretical model of human organization?
1nter I gatherers One of the most useful heuristic devices for the study of hunter-gatherers
tpped source of has been Binforq' s distinction between foraging and logistical .collecting
(1980), which suggests a basic characterization of the organization of subsis-
the Interpretation of tence pursuits .. It has been approached by means of studies of residential
~ations,Occasional mobility (Kelly 1983), site structure (Simek 1987), and faunal assemblages
versity. All rights (Enloe 1993; Enloe and David 1989). ~is study of long bone shaft fragments is
intended to find ways to discriminate from the archaeological record between
841 J. G. Enloe
behaviors resultil}g from subsistence organization of foragers and behaviors often overloo1
resulting from that of logistical collectors. In the former,·a signal characteristic rewards are fn
is that of immediate consumption of resources upon their acquisition. If food cable to such c
packages are too large to be immediately consumed by their acquirers, food tions from spe
sharing is likely to take place (Kaplan and Hill 1985). In contrast, the subsis- uals; who kno
tence organization of logistical collectors is directed toward the acquisition of ously depends
storable food resources. That activity most often occurs in highly seasonal goals here are ·
environments, where there is both a restriction or limitation in acquisition of cal samples ar
food during certain times of the year (e.g., winter) and an aggregated food through ethno;
resource with limited time for acquisition (e.g., migrating caribou). The logis-
tical collection of such a resource yields a considerable quantity of meat and
marrow at one time. Economies of scale dictate an efficient processing of such Da
resources, that is, some sort of systematic butchering and protessing for
storage. The contrast between foraging and logistical collecting has been Ob:
suggested as characterizing the transition from the Middle and Upper Paleo- tarandus) bonE
lithic (Binford 1980, 1982a, 1982b; Mellars 1989; Simek 1987). The ability to known situati<
recognize that. difference in subsistence organizati.on would aid us tremen- interior of Ale
dously in study of the evolution of human adaptive patterns and culture. samples came
When the dichotomy is applied to the kinds of processes that are respon- House. This "'
sible for the creation of bone splinters, it can be reduced to a question of the Stone tools fig
context of marrow processing. Marrow is a particularly important food re- site. Sample PJ
source for people whose diets are largely derived from faunal resources processing fqr
(Speth 1983; Speth and Speilmann 1983). Other than postdepositional or non- cessed, except
human agents (Binford 1981, 1984; Bonnichsen and Will 1980; Brain 1981; any). A large:
Haynes 1983; Meyers et al. 1980), the activity of marrow processing is the best marrow was 1
explanation for the fact that bones found in archaeological sites are usually domestic cook
broken (Binford and Bertram 1977; David 1972; Delpech and Rigaud 1974; Usually, joints
Martin 1910; Street 1990; Zierhut 1967). How does the extraction of marrow the bone were
reflect on the difference between foraging and logistical collectipg? One was the prodt
possibility is to derive expectations about the characteristics of remains that third sample o
would pertain to immediate consumption of sequentially acquired food ver- occupied in 19·
sus those that would result from mass processing of a single input and subse- 528) consists c
quent withdrawal from storage for consumption. In the former case, meat samples were 1
would be minimally processed, that is, carcasses would be disarticulated just the contents w
enough for transport and distribution to consumers. Further processing might of activities the
go no further than reduction of package sizes for cooking. Meat would be highlight the 1
cooked on the bones and eaten, and the bones would be cracked by each cracking incidE
consumer to obtain the marrow immediately thereafter. In the logistical case,
meat would be filleted, removed from the bones in order to be dried and
stored, and the bones would be mass processed at one time to obtain the A11
marrow (Binford 1978). As far as the marrow is concerned, it is a matter of
individual breakage of the bones immediately after meat consumption or AI
collective processing of bones entirely aside from meat consumption. cessing of raw
Therefore, if the goal is to distinguish the difference between intentional · posed. In the :
mass processing for marrow and extraction of marrow incidental to meat marrow utility
consumption, the necessary data to be examined are the durable residues processed at t
from both procedures, that is, bone splinters. This archaeological resource is breaking them
Subsistence Organization j85
; and behaviors often overlooked because it is difficult to deal with and the conventional
al characteristic rewards are frequently meager. Bone splinters rarely yield information appli-
uisition. If food cable to such conventional goals as determining past environmental condi-
acquirers, food tions from species representation or counting minimum numbers of individ-
~ast, the subsis- uals; who knows how many splinters come from a single bone? That obvi-
e acquisition of ously depends on how many times the bone is broken, and to what ends. The
1ighly seasonal· goals here are to examine some attributes that can be observed in archaeologi-
n acquisition of cal samples and to relate them to subsistence organization behavior known
ggregated food through ethnoarchaeological research.
bou). The logis-
ity of meat and
>cessing of such Data Base
processing for
~cting has been Observations were made on assemblages of caribou (Rangifer
d Upper Paleo- tarandus) bone splinters, which had been recovered from ethnographically
. The ability to known situations among the Nunamiut Eskimos of the Brooks Range in the
aid us tremen- interior of Alaska by Lewis R. Binford (1978:428-447, 1983:176-184). Two
ld culture. samples came from distinct activity areas at a site known as Palangana's
hat are respon- House. This winter residence was occupied in the late nineteenth century.
question of the Stone tools figure significantly in the processing of faunal materials from the
10rtant food re- site. Sample P1 (N = 181) consists of material from debris associated with mass
.unal resources processing fqr marrow. The bones had not been previously cooked or pro-
1sitional or non- cessed, except for dismemberment and removal of meat (if there had been
'80; Brain 1981; any). A large number of bones were broken open at one time, and the raw
ssing is the best marrow was removed. Sample P2 (N = 624) consists of material from a
ites are usually domestic cooking dump, that is, debris resulting from consumption of meals.
d Rigaud 1974; Usually, joints ofmeat were cooked in a stew pot. After the meat was eaten,
tion of marrow the bone were cracked open and the marrow was eaten. The dump, therefore,
:ollecting? One was the product of an accumulation of repea.ted consumption events. The
of remains that third sample comes from a site known as the Bear site. This winter house was
uired food ver- occupied in 1948. Steel tools were used during the occupation. Sample B3 (N =
lput and subse- 528) consists of materials from a kitchen midden similar to that of P2. These
mer case, meat samples were derived from restricted areas (1m2), in order to avoid mixing of
;articulated just the contents with debris from other activities. They do not represent the range
rocessing might of activities that might have been carried out at either site but were selected to
Meat would be highlight the differences between mass processing of ni.arrow and marrow
racked by each cracking incidental to meat consumption at meals.
~ logistical case,
o be dried and
e to obtain the Analysis
lt is a matter of
onsumption or A model for differentiating between breakage during mass pro-
tption. cessing of raw m.arrow and breakage for immediate consumption can be pro-
·een intentional posed. In the first case, elements present should primarily reflect their high
~dental to meat marrow utility (Binford 1978). When a large quantity of the same elements are
1rable residues processed at the same time, there should be a standardized procedure for
;ical resource is 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 Tab]
relatively few impact cones per element and relatively long fragments, reflect-
ing 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 Element
would be more variably located and more numerous, and fragments would be
relatively short. Humerus
In order to test those expectations, observations were made on each frag- Radius.
ment for element, length, and number of impact cones. Table 5-1 presents the Metacarpal
element representation of the three samples. Femur
It must be noted that not all fragments could be identified as to skeletal Tibia
element. That is a very common problem in dealing with bone splinters. The Metatarsal
unidentified fragments most likely come from the shafts of the major upper Phalanges
long bones, that is, humerus, radius, femur, and tibia. Recognizability is a Unidentified
critical criterion here. The major landmarks on the bones are mostly located at metapodial
Unspecified
the articular ends, and there are long spans of what are essentially tubes Total
between them. Various fragmentation processes result in the formation of metapodialsa
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 aIncludes all metac;
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 assemblages de
easily recognized, but the anterior portion of the metacarpal shaft is much less dominated by rr
distinctive. Therefore, anterior metacarpal fragments lacking any portions of To clarify the;
the posterior surface could have been placed in the unspecified category. percentages, di1
Posterior portions of either of those elements lacking proximal, cJistal,_ or were omitted ft
cranial portions had to be placed into the unidentified metapodial category .. threeassemblag
The final category (total metapodials) combines all of the fragments of assemblage, wE
metacarpals, metatarsals, and unidentified metapodials. Figure 5-2, whE
indices (Binforc
Element Representation (Binford 1978:81
residue assemh
The first test of the expectations deals with element representation element represe
in each assemblage (see Table 5-1). Strong preference for marrow-bearing assemblage, lie
elements in assemblage composition i~ argued to reflect mass-processing elements of the
events, whereas no preference or preference for meat-bearing elements is crosses the lines
argued for consumption incidental to meat consumption. Figure 5-1 displays both meat and:
the relative percentages of each element for all three cases. It is clear from this ences in elemen1
figure that assemblages P2 and B3 are almost identical to one another, while bearing element
assemblage P1 appears radically different. All three track very closely through both of the cons
the upper limb elements. Where .they diverge is in the unspecified and lines. Elements 1
metapodial categories, with relative differences of at least 36% and 43%, the line for the
respectively. If, as I suspect, the unspecified fragments really do come from has relatively 1·
the upper limb bones, then what we are seeing is a divergence between important marrc
similarity of th1
vould result in Table 5-1. Bone Splinters from Three Nunamiut Samples
~ents, reflect- -
>le 4.6). In the
tility. Since the Pl P2 B3
hey would be Mass Meal Meal
not expect that· Processing Midden Midden
tpact fractures Element NISP % NISP % NISP %
aen ts would be
Humerus 8 4 267 11 50 9
~ on each frag- Radius 18 10 81 13 52 10
;-1 presents the Metacarpal 2 1 5 1 7 1
Femur 11 6 63 10 55 10
i as to skeletal Tibia 24 13 52 8 70 13
~splinters. The Metatarsal 37 20 43 7 23 4
te major upper Phalanges 0 0 1 0 0 0
snizability is a Unidentified
metapodial 62 34 34 5 25 5
ostly located at Unspecified 19 10 278 246
46 47
sentially tubes Total
e formation of metapodialsa 101 56 82 13 55 10
nents were put I
than the upper alncludes all metacarpals, metatarsals, and unidentified metapodials.
·ior surfaces, so
mterior surface
distinctive and assemblages dominated by the meat-bearing elements and the assemblage
aft is much less dominated by marrow-bearing elements.
a.ny portions of To clarify that difference, a second calculation was made of the relative
:ified category. percentages, disregarding the unspecified fragments _(Table 5-2). Phalanges
imal, distal, or were omitted from consideration .since there was only one identified in all
odial category. three assemblages. Similarly, metacarpals, which constituted only 1% of each
assemblage, were omitted. The percentages are graphically displayed in
~ fragments of
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
; representation element representation. The line for assemblage P1, the marrow processing
tarrow-bearing assemblage, lies substantially below the other two for the meat-bearing
Lass-processing elements of the upper limbs but climbs sharply to the right of the graph. It
ng elements is crosses the lines for the other two at the tibia, which has significant values for
1re 5-1 displays both meat ahd marrow. In this figure the contrast demonstrates the differ-
; clear from this ences in element representation between the assemblages biased toward meat-
another, while bearing elements and the one biased toward marrow-bearing elements. Here,
closely through both of the consumption debris assemblages, P2 and B3, exhibit relatively flat
nspecified , and lines. Elements vary between 9% and 25% of the total assemblage. In contrast,
36% and 43%, the line for the marrow-processing assemblage exhibits a significant rise. It
· do come from has relatively low values for meat-bearing elements but high values for
gence between 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
Humerus
Radiocubitus
Q) Femur
Ol
co Tibia
§ 30+-----------------------~--------~--4-~------4 Metatarsal
()
05 Unidentified
a.. metapodial
Total
metapodials
Note: Adjusted rela

rnents.
HM RD MC FM TA MT PH US MP MPs
Element
Frag
, _ P1 -1- P2-*- 83
The:
The three samp:
Figure 5-1. Relative representation of all elements. other taxa with
elements examil
ing was not prE
happening in the two situations and that we will be able to see the difference
samples were dJ
in the material residues.
have operated t
We can further test the differences to see if they are statistically d,ifferen t.
not allowed to r
First, the P2 and B3 assemblages were subjected to a chi-square an~lysis for
tethered in dog ·
the relative proportions of element representations, using the raw counts. Low
Table 5-3 sun
frequency cells for metacarpal and phalanx were eliminated, so only the total
samples. This r1
metapodials were used to represent metacarpals, metatarsals, and unidenti-
the average len~
fied metapodials. The resultant X2 = 9.99 is not significant at the .05 level, so
while those of
we cannot reject the null hypothesis that the samples were drawn from the
bearing elemen·
same population. This finding supports the similarity of their makeup, and we
samples P2 and
can therefore infer that they are likely to be derived from the same formation
the differences t
processes.
and metatarsah
The same test was performed between Pl and P2, and between Pl and B3.
lowest below tr
In each of the tests the results are quite different from those of the first test,
with the highes·
and they are very similar to one another. The X2 values are very high (257.01
is the very fact t:
and 244.29), particularly in contrast to that of the first test (9.99), and are
Overall, the r
highly significant at less than the 0.001level. This strongly confirms that both
pared with 72.7,
times the samples being compared<had been drawn from radically different
Figure 5-3 that
populations and that P2 and B3 are similar in the order of magnitude of their
bones in sampl
differences from Pl. These results support the idea that there is a significant
length? The moe
difference in element representation between the sample drawn from inten-
of breakage shO'
tional marrow processing and those drawn from meal residues.
Table 5-2. Adjusted Relative Percentages of Identifiable Elements
Element Pl P2 B3
Humerus 5 19 18
Radiocubitus 11 23 18
Femur 7 18 20
Tibia 15 15 25
Metatarsal 23 12 8
Unidentified
metapodial 38 10 9
Total
metapodials 62 24 20
Note: Adjusted relative percentages disregard metacarpals, phalanges, and unspecified frag-
ments.
MP MPs
Fragment Length
The second aspect to be explored deals with the size of fragments.
The three samples included only specimens from caribou so that problems of
other taxa with different body sizes would not affect size relationships for the
elements examined in this analysis. Additionally, evidence of carnivore gnaw-
ing was not present on bones from these samples. As previously stated, the
~e the difference samples were drawn from restricted locations to limit the processes that might
have operated to modify the assemblages. On the Nunamiut sites, dogs were
tically different.
not allowed to roam freely and scavenge bones all about the camp. They were
tare analysis for tethered in dog yards (Binford 1978:429, 431).
·aw counts. Low
Table 5-3 summarizes average length and varianc_e .for each element in the
so only the total
samples. This relationsh~p is displayed graphically in Figure 5-3. In general,
5, and unidenti-
the average lengths of element fragments -are fairly similar between P2 and B3,
the .05 level, so
while those of Pl are consistently greater. Average lengths of the meat-
jrawn from the
bearing elements of the upper limbs were fairly consistent, implying that in
nakeup, and we
samples P2 and B3 bones were being subjected to the same treatment. Most of
same formation
the differences that are apparent between P2 and B3 occur in the metapodials
and metatarsals. It is the length of the unspecified fragments that plunges
veen P1 and B3.
lowest below the mean fot each case. For P2 and B3, those are the elements
of the first test,
with the highest frequencies, which in itself pulls the mean down. Perhaps it
ery high (257.01 .
is the very fact that these fragments are so short that they are unidentified.
(9.99), and are
Overall, the mean length of fragments for sample P1 is 143.26 inm, com-
nfirms that both
pared with 72.73 mm and 57.92 mm for samples P2 and B3. It can be seen in
dically different
Figure 5-3 that ·the difference in length applies to all of the elements. The
tgnitude of their
bones in sample P1 are clearly longer, but do they show less variation in
~ is a significant
length? The model for mass processing dictates that a standardized procedure
1wn from inten-
of breakage should result in less overall variability in ~ength than a fortuitous
5.
90 I ]. G. Enloe
Element representation Ta
70
60
Element
50
Q)
Overall
OJ
cu 40
c
Q)
(.)
(D 30 Humerus
0...
20
Radius
10
Metacarpal
0
FM HM TA RD MT MP MPs
Element Femur
I-ll- p1 -t- P2 ---*-- 83

Tibia
Figure 5-2. Relative representation of elements: meat utility to left, mar-
row utility to right.
Metatarsal
or incidental breakage during meal consumption. Table 5-3 presents the stan- Metapodial
dard deviations for the mean lengths for each element, but those are difficult
to interpret. Since there are rather large differences in mean length/ between
samples, raw standard deviation figures are not so informative. Rather, the Unspecified
coefficient of variation, which indicates the proportion of the mean repre-
sented by the standard deviation, can better inform us about the variation
within and between samples. Overall, the coefficient of variation for sample
Pl (0.430) is lower than that for either sample P2 or B3 (0.483 and 0.482). It is Meat Bias
interesting that the latter two are so similar to one another. Let us further look
at the variation between the lengths of meat-bearing elements as opposed to Marrow Bias
marrow-bearing elements. Again, in each case the coefficient of variation for
sample Pl is lower (0.400 and 0.356) than that for sample P2 (0.460 and 0.536)
or for sample B3 (0.484 and 0.417). The value for the high marrow utility
elements of sample Pl is the lowest of these six. In summary, fragment lengths
are both greater and less variable for the marrow-processing assemblage,
confirming the expectations of the model.
· men ts will sh
Impact Cones define the lirr
caused by a si:
The third aspect of the model deals with impact fractures, which £ragmen ts of t
are manifested in visible impact cones on the bone fragments. Not all frag- Table 5-4.
Subsistence Organization 91 I
Table 5-3. Fragment Lengths
Mean Standard
Length Deviation Coefficient
Element Sample (mm) (mm) of Variation
Overall P1 143.26 61.55 0.430

P2 72.73 40.15 0.483
B3 57.92 27.97 0.482
Humerus P1 10;3.39 25.04 0.242
P2 50.20 27.67 0.551
B3 66.08 19.82 0.230
Radius P1 123.45 39.18 0.317
P2 85.43 34.82 0.408
B3 70.69 34.12 0.483
Metacarpal P1 199.50 64.35 0.323
P2 60.82 39.16 0.644
MPs B3 58.38 17.89 0.306
Femur P1 108.62 23.16 0.213
P2 71.85 28.79 0.401
B3 70.26 27.25 0.388
Tibia P1 132.32 44.87 0.339
P2 86.98 30.32 0.349
tility to left, mar- B3 76.62 30.49 0.398
Metatarsal P1 151.07 63.99 0.424
P2 119.45 65.46 0.548
B3 71.65 25.29 0.353
·esents the stan- Metapodial P1 180.30 56.30 0.312
ose are difficult P2 122.37 59.41 0.485
engths between B3 73.30 35.39 0.483
ive. Rather, the Unspecified P1 70.67 37.38 0.529
he mean repre- P2 59.08 24.48 0.414
1t the variation B3 42.62. 17.01 0.399
tion for sample
Meat Bias P1 109.53 43.90 0.400
and 0.482). It is
P2 66.02 30.37. 0.460
: us further look B3 56.45 27.35 0.484
s as opposed to Marrow Bias P1 169.97 60.54 0.356
of variation for P2 117.09 62.82 0.536
0.460 and 0.536) B3 71.16 29.68 0.417
marrow· utility
~agment lengths
ng assemblage,
ments will show signs of impact cones. Spiral fracture intersections may
define the limits of any particular fragment, even if they were originally
caused by a single impact or multiple impacts that left impact cones on other
ractures, which fragments of the same bone. Data from the three samples are summarized in
ts. Not all frag- Table 5-4.
921 ]. G. Enloe
Fragment length Ta
200
180
160 NISP
Total Cones (N,
140 Average cones
E Standard de'
E
..c
+-' 120 Variation cot
OJ
c Total length
Cll
_j
100 Average length
Standard de'
80 Length/ cone
60
40 these samples
HM RD MC FM TA MT MP US MPs distinguishin~
Element deriving from
1-- P1 -+- P2--*-- 83

Cc
Figure 5-3. Average fragment length by element.
Th
known deriva
for recognizir
In general, there appear to be more cones per fragment in sample P1 (1.18) processes res1
than in either P2(0.82) or B3 (0.43). This is true for all elements. Again, P2 and criteria for di~
B3 are quite similar to one another, while Pl appears sli?htly diver:ge~t. The and incidenta
standard deviations, however, are quite large, approaching or exceeding the tiona!, special:
mean in each case. The coefficients of variation emphasize the slightly smaller be some signj
variability in sample P1 (0.991) and the similarity of the slightly greater and length vc:
variability of samples P2 and B3 (1.72 and 1.72). These differences are not
dump locatio
statistically significant, and they are not diagnostic for our purposes, but they upper limb be
do hint at the differences in the formation processes of each of the samples. sumption mid
Examination of the average length per impact cone of fragments of the various be longer and
elements does not reveal any pattern of distinction between the three samples. middens.
This index is quite variable in all three ,samples, ranging from 39.3 to 203.2 in Given the s
P1, from 57.0 to 138.8 in P2, and from 51.1 to 233.0 in B3.
identified ele1
The presence of impact. cones, by frequency or fragment length, does not truly diagnos1
appear to be a satisfactory diagnostic criterion for distinguishing marrow- examining lor
processing remains from meal residues. Perhaps because the fragments result- problem of ec
ing from the former activity are longer, they are likely to retain a greater num- other cultural
ber of impact points. It may also be that cooking has reduced the structural
tions of metaF
integrity of cortical tissue of the diaphyseal shafts, making them easier to
of this researc
break or easier to break into smaller pieces with fewer blows. No matter samples were
which situation pertains, it renders at least the kinds of observations made on
each activity
Table S-4. Summary of Impact Cones and Average Length
Pl P2 B3
NISP 181 624 528

Total Cones (N) 214 511 229
Average cones 1.18 0.82 0.43
Standard deviation 1.17 0.96 0.75
Variation coefficient .991 1.72 1.72
Total length 25929.38 45385.06 30582.73
Average length 143.26 72.73 57.92
Standard deviation 61.55 40.15 27.97
Length/ cone 121.2 88.8 133.5
these samples unusable as diagnostic criteria or signature characteristics for

us MPs distinguishing between assemblages deriving from meal residues and those
deriving~ from specialized marrow processing.
Conclusions
This research has examined a few attributes of several samples of
known derivation from ethnographic sources, in search of diagnostic criteria
for recognizing characteristics of material remains that might indicate the
;ample Pl (1.18) processes responsible for their creation. In this case, it has been a search for
'·Again,; P2 and criteria for discriminating between bones splinters that are the result of meat
r divergent. The and incidental marrow consumption and those that are the result of inten-
,r exceeding the tional, specialized processing of raw bones for their marrow. There appear to
slightly smaller be some significant differences in element representation, fragment length,
slightly greater and length variation between the two processes. Mass marrow-processing
erences are not dump locations appear to contain more metapodial fragments relative to
rposes, but they upper limb bone and unidentified element fragments than do immediate con-
of the samples. sumption middens. The fragments from mass-processing locations appear to
ts of the various be longer and to be of a more uniform length than those from consumption
e three samples. middens.
l 39.3 to 203.2 in
Given the small sample size of three assemblages of only a few hundred
identified elements in each, the results cannot be considered conclusive or
ength, does not truly diagnostic. They may be more profitably viewed as a starting point for
ishing marrow- examining long bone shaft splinters under other controlled conditions. The
·agments result- problem of equifinality has not been sufficiently addressed. There may be
t a greater num-
other cultural or taphonomic factors that might yield similarly high propor-
d the structural tions of metapodials of uniform length, for example. The primary importance
them easier to of this research has been its contrastive nature on a single site. Two of the
1ws. No matter samples were drawn from different a"ctivity areas of the same site. Clearly,
rations made on each activity does not define the overall subsistence strategy of the site's
94j ]. G. Enloe
occupants. The very fact, however, that mass marrow processing was present David, F.
on the site indicates its importance in overall subsistence. The application of 1972 Temoim
this type of analysis only to small samples of archaeological sites runs the very habitat magdc.
great risk of not discovering the appropriate activity areas. It should be pp. 295-320.
Delpech, F., and J
applied to the analysis of different concentrations of debris or activities on
1974 Etude d
sites that have already been excavated over large surfaces, as suggested by niveau d'hal
O'Connell (1987). When properly and cautiously applied, it is hoped that this H. Camps-Fc;
ethnoarchaeological research will add to the means available to zooarchae- . Enloe, J. G.
ologists for learning· about past organizations from the sparse documents of 1993 Subsiste
the archaeological record. of the Abri c
Paleolithic, ec
Press, Boca F
References Enloe, J. G., and F
1989 Le remo
Binford, L. R. leniens de P
1978 Nunamiut Ethnoarchaeology. Academic Press, New York. Prehistorique
1980 Willow Smoke and Dogs' Tails: Hunter-Gather~r Settlement Systems and Frison, G.
Archaeological Site Formation. American Antiquity 45:4-20. 1970 The Gler
1981 Bones: Ancient Men and Modern Myths. Academic Press, New York. 1974 The Casp
1982a The Archaeology of Place. Journal of Anthropological Archaeology 11:5-31. York.
1982b Comment on White. Current Anthropology 23:177-181. 1978 Prehistor
1983 In Pursuit of the Past. Thames and Hudson, New York. Gilbert, G. M.
1984 Faunal Remains from Klasies River Mouth. Academic Press, New York. 1980 Mammal;
Binford, L. R., and J. B. Bertram Grayson, D. K.
1977 Bone Frequencies and Attritional Processes: In For Theory Building in Archae- 1984 Quantita
ology: Essays on Faunal Remains, Aquatic Resources, Spatial Analysis, and Systemic Academic Pr
Modeling, edited by L. R. Binford, pp. 77-153. Academic Press, New York. Haynes, G.
Bonnichsen, R., and R. T. Will 1983 Frequen1
1980 Cultural Modification of Bone: The Experimental Approach in Faunal Anal- Modem Surf
ysis. In Mammalian Osteology, edited by B. M. Gilbert, pp. 7-30. B. Miles Gilbert, Kaplan, H., and K
Laramie, Wyo. . / 1985 Food Sh
Brain, C. K. · · Anthropology
1981 The Hunters or the Hunted? An Introduction to African Cave Taphonomy. Kehoe, T. F.
University of Chicago Press, Chicago. 1967 The Bam
Brown, C. L., and C. E. Gustafson Kelly, R. L.
1979 A Key to Postcranial Skeletal Remains of Cattle/Bison, Elk, and Horse. Reports of 1983 Hunter-(
Investigations 57. Laboratory of Anthropology, Washington State University, 277-306.
Pullman. Klein, R. G., and I<
Bunn, H., and E. Kroll 1984 The Anal
1986 Systematic Butchery by Plio-Pleistocene Hominids at Olduvai Gorge, Press, Chicag
Tanzania. Current Anthropology 27:431-452. Lawrence, B.
Burgett, G. R. · 1951 Post-cra1
1990 The Bones of the Beast: Resolving Questions of Faunal Assemblage Formation with Notes o
Processes Through Adualistic Research. Unpublished Ph.D dissertation, University ogy and Ethn1
of New Mexico, Albuquerque. Mass.
Chaplin, R. E. , Lupo,K.
1971 The Study of Animal Bones from Archaeological Sites. Seminar Press, London. 1990 Cutmar1
Dart, R. A. Tanzania. Pa
1957 The Osteodontokeratic Culture of Australopithecus Prometheus. Transvaal Council for A
Museum Memoirs 10.
Subsistence Organization 95 I
;ing was present David, F.
te application of 1972 Temoins osseux. In Fouilles de Pincevent: Essai d'analyse ethnographique d'un
habitat magdalenien (la Section 36), edited by A. Leroi-Gourhan and M. Bn§zillon,
tes runs the very
1s. It should be pp. 295-320. Vile Supplement aGallia_Prehistoire.
Delpech, F., and J-P. Rigaud
or activities on
1974 Etude de la fragmentation et de la. repartition des restes osseux dans un
as suggested by niveau d'habitat paleolithique. In L'Industrie de L'Os dans la Prehistoire, edited by
; hoped that this H. Camps-Fabrer, pp. 47-55. Editions de l'Universite de Provence.
le to zooarchae- Enloe, J. G.
:;e documents of 1993 Subsistence Organization in the Early Upper Paleolithic: Reindeer Hunters
of the Abri du Flageolet. In Before Lascaux: The Complex Record of the Early Upper
Paleolithic, edited by H. Knecht, A. Pike-Tay, and R. White, pp. 101-116. CRC
Press, Boca Raton. ·
Enloe, J. G., and F. David
1989 Le remontage des os par individus: le partage .du renne chez les Magda-
leniens de Pincevent La Grande Paroisse, Seine-et-Marne. Bulletin de la Societe
Prehistorique Franr;aise 86:275-281.
nent Systems and Frison, G.
1970 The Glenrock Buffalo Jump, 48C0304. Plains Anthropologist Memoir 10.
wYork. 1974 The Casper Site: A Hell Gap Bison Kill on the High Plains. Academic Press, New
Jlogy 11:5-31. York.
1978 ;Prehistoric Hunters of the High Plains. Academic Press, New York.
Gilbert, G. M.
Jew York. 1980 Mammalian Osteology. B. Miles Gilbert, Laramie, Wyo.
Grayson, D. K.
3uilding in Archae- 1984 Quantitative Zooarchaeology: Topics in the Analysis of Archaeological Faunas.
lysis, and Systemic Academic Press, Orlando.
New York. Haynes, G.
1983 Frequencies of Spiral and Green-Bone Fractures on Ungulate Limb Bones in
:h in Faunal Anal- Modem Surface Assemblages. American Antiquity 48:102-114.
'. B. Miles Gilbert, Kaplan, H., and K. Hill
1985 Food Sharing among Ache Hunter-Gatherers of Eastern Paraguay. Current
Anthropology 262:223-246.
Cave Taphonomy. Kehoe, T. F.
1967 The Boarding School Bison Drive. Plrlins Anthropologist Memoir 4.
Kelly, R. L. ·
' Horse. Reports of 1983 Hunter-Gatherer Mobility Strategies. Journal of Anthropological Research '39:
State University, 277-306.
Klein, R. G., and.K. Cruz-Uribe
1984 The Analysis of Animal Bones from Archaeological Sites. University of Chicago
Olduvai Gorge, Press, Chicago.
Lawrence, B.
1951 Post-cranial Skeletal Characteristics of Deer, Pronghorn, and Sheep-goat
~mblage Formation . with Notes on Bos and Bison. Papers of the Peabody Museum of American Archaeol-
tation, University ogy and Ethnology, Harvard University 35, no. 3. Peabody Museum, Cambridge,
Mass.
Lupo, K.
Press, Lonqon. 1990 Cutmarks and Carcass Acquisition Strategies: The Hadza of Eastern
Tanzania. Paper presented at the Sixth International Conference, International
1etheus. Transvaal Council for Archaeozoology, Washington, D.C.
96j ]. G. Enloe
Lyrnan,R.L , Nos. 3, 4, 5, a1
1987 Archaeofaunas ,and Butchery Studies: A Taphonomic Perspective. In 1955 Observat
Advances in Archaeological Method and Theory, vol. 10, edited by M. B. Schiffer, pp. Nos. 7, 8, and
249-337. Academic Press, New York · Yellen, J. E.
Martin, H. 1977 Cultural·
1910 La percussion osseuse et les esquilles qui en derivant: Experimentation. In Experimer,
Bulletin de la Societe Prehistorique Franqaise 7:299-304. Macdonald, F
Mellars, P. A. Zierhut, N. W.
1989 Major Issues in the Emergence.of Modem Humans. Current Anthropology 30: 1967 Bone Ure
349-385. 1:33-36.
Meyers, T. P., M. R. Voorhies, and R. G. Comer
1980 Spiral Fractures and Bone Pseudotools at Paleontological Sites. American
Antiquity 45:483-489.
Noe-Nygaard, N. .
1977 Butchering and Marrow Fracturing as a Taphonomic Factor in Archaeolog-
ical Deposits. Paleobiology 3:218-237.
O'Connell, J. F.
1987 Alywara Site Structure and Its Archaeological Implications. American Antiq-
uity 52:74-108. ·
Olsen, S. J.
1964 Mammal Remains from Archaeological Sites. Papers of the Peabody Museum
of Archaeology and Ethnology, Harvard University 56, no. 1. Peabody Museum,
Cambridge, Mass.
Sadek-Kooros, H.
1972 Primitive Bone Fracturing: A Method of Research. American Antiquity 37:
369-382.
Schmidt, E.
1972 Atlas of Animal Bones for Prehistorians, Archaeologists and Quaternary Geolo-
gists. Elsevier Publishing, Amsterdam.
Simek,J. F.
1987 Spatial Order and Behavioral Change in the French Palaeolithic. Antiquity
61:25-40. I
Speth,J. D.
1983 Bone Counts and Bison Kills: Decision Making by Ancient Hunters. University of
Chicago Press, Chicago.
Speth, J. D., and K. A. Spielmann
1983 Energy Source, Protein Metabolism, and Hunter-Gatherer Subsistence Strat-
egies. Journal of Anthropological Archaeology 2:1-31.
Street, M.
1990 Butchering Activities at the Early Mesolithic Site Bedburg-Konigshoven,
Rhineland, F.R.G. Cranium 7(1):25-43. ,
Todd, L. C., and D. J. Rapson
1988 Long Bone Fragmentation and Interpretation of Faunal Assemblages:
Approaches to Comparative Analysis. Journal of Archaeological Science 15:307-
325.
White, T. E.
1952 Observations on the Butchering Technique of Some Aboriginal Peoples: I.
American Antiquity 17:337-338. ·
1953 Observations on the Butchering Technique of Some Aboriginal Peoples, No.
2. American Antiquity 19:160-164.
1954 Observations on the Butchering Technique of Some Aboriginal Peoples,
Subsistence Organization j97
Nos. 3, 4, 5, and 6. American Antiquity 19:254-264.
Perspective. In 1955 Observations on the Butchering Technique of Some Aboriginal Peoples,
.1. B. Schiffer, pp. Nos. 7, 8, and 9. American Antiquity21:170-178 .
Yellen, J. E.
1977 Cultural Patterning in Faunal Remains: Evidence from the !Kung Bushmen.
~xperimentation. In Experimental Archeology, edited by p. Ingersoll, J. E. Yellen, and W.
Macdonald, pp. 271-331. Columbia University Press, New York
Zierhut, N. W.
·Anthropology 30: 1967 Bone Breaking Activities of the Calling Lake Cree. Alberta Anthropologist
·1:33-36. \
1 Sites. American
1r in Archaeolog-
. American Antiq-
Peabody Museum
abody Museum,
can Antiquity 37:
2uaternary Geolo-
olithic. Antiquity
ers. University of
)ubsistence Strat-
rg-Konigshoven,
:~.1 Assemblages:
~z Science 15:307-
iginal Peoples: I.
inal Peoples~ No.
>riginal Peoples,
II. Settlement Patterns:
Site Function and Duration
of Occupation
6. The Archaeological Structure of a
Short-Term Camp
Kevin T. Jones
Abstract: Behavioral observations and archaeological study among the

Ache, Paraguayan hunter-gatherers, indicate that in short-term camps
(occupied for one night) most activities take place within 1.5 m of camp
fires. After abandonment, the debris (mostly faunal remains) clearly
reflects this fire-focused activity pattern. As short-term camps are not
cleaned, all debris is in primary context. Foraging for immediate con-
sumption by the Ache results in the introduction of complete animal
carcasses into camps and a characteristic pattern of body part represen-
tation. The patterns produced by short-term occupations are distinctive
and can be used to test inferences about prehistoric provisioning and
positioning systems.
Introduction
Short-term camps are probably the most com1non type of archaeo-
logical site. Foragers, collectors, hunters, gatherers, farmers, herders, and even
archaeologists make them. I make several of them myself each year. We see
them each time we conduct an archaeological survey, and despite often hav-
ing spent the previous night in one, we generally do an extremely poor job of
investigating them.
Occupied for a few days or less, the archaeological content of a short-term
camp is usually unremarkable. Such sites are commonly neglected by
researchers in favor of more spectacular, complex localities, yet a significant
portion of the archaeological record of hunter-gatherers is likely represented
by such scant remains. Most hunter-gatherer subsistence strategies are flexible
and variable-and may involve within a relatively short time span foraging for
immediate consumption, foraging for delayed consumption and transporta-
tion, and a range in between. The better we become at recognizing different

reserved. ISBN 0-88104-076-2. ·
101
1021 K. T. ]ones
strategies in the· archaeological. record, the better .we will be at testing Tab!'
hypotheses about the nature and organization of prehistoric settlement and
subsistence systems. Identifying short-.term camps and evaluating their role in
larger systems of positioning and provisioning is crucial. Site Area (1
In this paper I describe some aspects of short-term ca:rnp archaeology,
focusing on the hunting strategies, butchering, consumption, use of camp 11/22 22.'
space, and refuse disposal of the Ache. Distinctive patterns noted in Ache
camps may be useful in identifying similar strategies represented archaeologi- 11/24 24.,
cally. Most noteworthy is that the camps are small and exhibit no secondary
disposal of refuse or size-sorting of artifacts. Aspects of animal butchery and 11/26 29.1
disposal of faunal remains may also be diagnostic of a short-term encamp-
1/09 18.1
. ment and can be used in conjunction with observations on the spatial
arrangement of refuse to clarify interpretations of the strategies represented in 1/10 21.1
short-term sites.
1/11 19.1
Spatial Organization of Behavior among the Ache Mean 22.1
The Ache are traditional hunter-gatherers who until recently for- ± 3.~
aged in the thick neotropical forests of eastern Paraguay (Borrero and
Yacobaccio 1989; Clastres 1972; Hawkes et aL 1982; Hill and Hawkes 1983).
Coming into peaceful contact with Paraguayans only within the past 20 years,
the Northern Ache now live on a Catholic mission and forage part-time, farm of people (249-2!
part-time. Foraging trips are taken regularly by most Ache and generally last also larger than J
from one week to one month. Band-sized groups (20-35 people) hunt and (Fisher and Strid
gather in looping treks through the forests surrounding the mission. in area, were ocCJ
Ache foraging bands are usually composed of four to seven nuclear-family- ing from 2 to 30 i
sized groups. They are very mobile, moving camp each day except during of camp space pe
heavy rain. Each morning when camp is broken the men leave to hul)t and the than Ache camp
women and children walk through the forest, usually stopping several times tures (1.6 m2/per
to gather plant resources. They meet with the men late in the day, construct (1988) and others
camp, and cook and eat the foods they have obtained. Camp areas are cleared size and length o:
of underbrush, and each family builds a fire on the perimeter of the cleared Among desert
area. Camps are usually roughly circular or oval, ringed by hearths, and vary constrained, cam
between 3m and 10m in diameter. ber of occupants
Six Ache camp sites (Table 6-1) occupied in 1984 and 1985 for a single night short-term Kua S
each averaged 22.6 m2 in total area. This amounted to 0.90 ± 0.13 m2 per al. 1991:Table 2) c
person (including children), and 1.71 ± 8.20 m2 per adult. Each site was ringed the camps avera1
by five or six fires. The mean number of individuals per fire was 4.43 ± 0.31, or term camps fron
2.30 ± 0.15 adults per fire. The amount of camp space per fire was 4.00 ± 0.61 m 2 /person (Nic:
m2. The relatively small camp size is likely related to the necessity of clearing reported by 0' C
a space in the dense undergrowth, a space that will only be utilized for a person. Other re
single night. i11 ter household a
In comparison, a relatively short-term camp of the Aka Pygmy (occupied for camp size in
for seven days) covered 110 m2, an average of 6.88 m 2 per person (Hudson illustrating the c
1988, personal communication 1991). Longer-term Aka sites (occupied for 75 between househ
and 71 days) are considerably larger, over twice the size for the same number considered to be i
A Short-Term Camp !103
·ill be at testing Table 6-1. Site Area and Population for Six Ache Foraging Sites
:c settlement and
ating their role in
Site Area (m2) N Fires N Indiv's. N Adults M2/person M2Jfire
np archaeology,
lon, use of camp 11/22 22.7 6 26 14 0.9 3.8
lS noted in Ache
nted ·archaeologi- 11/24 24.8 6 26 14 1.0 4.1
[bit no secondary
11/26 29.0 6 26 14 1.1 4.8
nal butchery and
)rt-term encamp- 18.0 6 24 12 0.8 3.0
1/09
s on the spatial
.es represented in 1/10 21.6 5 24 12 0.9 4.3
1/11 19.6 5 24 12 0.8 3.9
the Ache Mean 22.6 0.9 4.0
mtil recently for- ± 3.9 0.1 0.6

ay (Borrero and
d Hawkes 1983).
the past 20 years,
;e part-time, farm of people (249-254 m2, or 15.56 and.15.88 m2 per person. Efe Pygmy sites are
Lnd generally last also larger than Ache sites, although they are occupied for considerably longer
)eople) hunt and (Fisher and Strickland 1989, 1991). Nineteen Efe Campsites averaged 242.7 m2
1ission. in area, were occupied for an average of six weeks, and had populations rang-
n nuclear-family- ing from 2 to 30 individuals (Fisher and Strickland 1989:475-477). The amount
:~.y except during of camp space per person ranged from 4.39 m 2 to 22.67 m 2, considerably larger
·e to hunt and the than Ache camps, although the amount of space per occupant in Efe struc-
ing several times tures (1.6 m 2/personrls comparable to the 0.9 m2 in Ache Camps. As Hudson
he day, construct (1988) and others have noted, there appears to be a relationship between camp
areas are cleared size and length of occupation.
ter of the cleared Among desert hunter-"gatherers, where potential camp size may be less
1earths, and vary constrained, camp area appears to be highly variable but related to the num-
ber of occupants and the length of time the site is occupied. A sample of 25
for a single night short-term Kua San camps for which applicable data are available (Bartram et
90 ± 0.13 m2 per al. 1991:Table 2) average 10.5 m 2/occupant. When length of stay is factored in,
~h site was ringed the camps averaged 4.22 m2/person-day. A sample of protohistoric shorter-
vas 4.43 ± 0.31, or term camps from the Western Desert of Australia averaged 30 m2, or 6.8
·e was 4.00 ± 0.61 m 2/person (Nicholson and Cane 1991). Longer-term Hadza base camps
:essity of clearing reported by O'Connell and others (1991) averaged 796 m2, and 19.3 m2/
be utilized for a person. Other researchers, investigating longer-term camps, and including
interhousehold areas in site definitions, have published much higher figures
=>ygmy (occupied for camp size in Australia (e.g., Gould and Yellen 1987; O'Connell 1987),
person (Hudson illustrating the difficulties in estimating camp size: whether the distance
(occupied for 75 between household areas is great or. small and whether households are
the same number considered to be individual sites or part of a larger site ~ffect comparisons.
1041 K. T. Jones
Susan Kent (1991), investigating factors that influence site size, number and
size of structures, and site area per person among San arid Bantu speakers, has
found that the strongest predictor of site area is "anticipated mobility/' an
estimate of the amount of time people expect to stay in a particular location.
The population of a site was also significantly correlated with site size, but
actual mobility, the amount of time that was actually spent at the site, was not
significantly correlated with site size.
Nearly all activities in an Ache camp take place within the ring of fires.
From the time camp is made until it is broken the next morning, people spend
most of their time in the camp, very near to their family hearth. But for occa-
sional trips out of camp to retrieve water or nearby resources, other than
children playing during daylight hours, all of the members of the band stay in *: ad1
camp. After dark, nobody leaves camp unless urgently called by nature be-
cause the forest is dangerous. 0
Within the confines of the small circle of fires, animals are butchered, plants
processed, baskets woven, children nursed, arrows sharpened, food con-
Figu
sumed, and jokes told. The fires form a protective ting about the camp, create
smoke that deters insects, cook food, heat warped arrow shafts, soften stiff
palm fronds, and warm chilled bodies. Each fire is the focus, the magnet, for
all activities. Nearly all activities in camp take place within less than one For;
meter of a fire.
The
Behavioral observations on space use were taken in three Ache camps
obtained by th
occupied on successive nights in January 1985. The method consisted of first
small (< 25 kg;
constructing a camp map and then taking a space-use census every 15 min-
animals too lar
utes, indicating on the map the location of each adult and noting what he or
location of the 1
she was doing. From the maps, the person's distance from the fire was com-
cooked, and us1
puted (Figure 6-1). These data are summarized in Table 6-2. The mean dis-
meat is left ove
tance from a fire for all adult activities covered in the sample was .60.7 em±
ing. The leftove
22.4 em, N = 175. There was little difference between males (60.5 f 21.4, N =
may be eaten a
88), and females (60.9 ± 23.5, N = 87). Interestingly, debris-produCing activi-
duction of near
ties, such as eating and implement repair, were conducted in closer proximity
When leftovers
to the hearth (57.0 ± 24.0, N = 46) than activities that produced no debris, such
of tidbits, notrE
as resting or childcare (62.1 ± 21.2, N = 129). That holds true for both males
The pattern <
and females. It may be accounted for by use of the fire for light, use of the fire
well documen
as an aid in a task such as straightening arrow shafts, or, when eating, use of
Harako 1981:54
the fire fo~ cooking. admittedly arh
The distribution of debris left in the camp when abandoned closely mirrors
can generally l:
this pattern of space use. Most debris is discarded within one meter of hearths
substantial part
(Jones 1983, 1984). The debris is all in primary context, as short-term Ache
sumed within a
camps are not swept or cleaned. The distribution of larger items (> 5 em) coin-
cides with the distribution of all debris, being only slightly skewed toward the
outside of camp. Were secondary disposal practiced, the distribution of larger
items would be expected to fall to the outside of the primary area of smaller
debris. (A comparison of distributions of micro-refuse to macro-refuse should Eth1
indicate covariance between micro- and macro-refuse in a short-term camp curement, proc
and divergent distributions in longer-term camps). In the Ache case, the dis- value to prehisi
tribution of all debris is a direct reflection of the pattern of space use. to the behavio:
A Short-Term Camp j105
;ize, number and

ntu speakers, has
ed mobility," an
Lrticular location.
rith site size, but
; the site, was not
the ring of fires.

** *
* **
* **
**
l
ng, people spend ****** ** *****
rth. But for occa-
uces, other than
~*******
f the band stay in * : adult time point
ed by nature be-
0 1m
Jutchered, plants
>ened, food con- Figure 6-1. Ache camp, 9 January 1985.
the camp, create
hafts, soften stiff
;, the magnet, for
~in less than one
Foraging for Immediate Consumption
I
The Ache forage each day and generally consume everything
ree Ache camps obtained by the following day. Most of the animals killed by the Ache are
consisted of first small (< 25 kg) and are usually carried whole to the evening camp. When
us every 15 min- animals too large to easily carry are taken, camp will be made close to the
oting what he or location of the kill. Animals are not butchered in the field. They are butchered,
;he fire was com-
cooked, and usually entirely consumed in the same camp. Occasionally, some
2. The mean dis- meat is left over and will be carried along when camp is broken in the morn-
,le was 60.7 em ± ing. The leftovers are consumed along the trail by the women and children or
(60.5 ± 21.4, N = may be eaten at the following day's camp. This strategy results in the intro-
?roducing activi- duction of nearly complete skeletons of all animals taken in that day's camp .
. closer proximity When leftovers are taken from the camp, they tend to be a random assortment
d no debris, such of tidbits, not related to body part utility..
le for both males The pattern of carrying complete carcasses of smaller animals to camps is
;ht, use of the fire well documented among hunter-gatherers (e.g., Bartram et al. 1991:101;
1en eating, use of Harako 1981:540; summarized in Jones 1984; Lee 1979:219). (I use 25 kg as an
admittedly arbitrary definition of small, since an animal that size or smaller
d closely mirrors can generally be carried by a single person.) Small animals often make up a
~meter of hearths substantial part of hunter-gatherer meat diet (Jones 1984) and are usually con-
short-term Ache sumed within a relatively short time following acquisition.
ms (> 5 em) coin-
:ewed toward the·
ribution of larger The Big Bone Bias
y area of smaller
:ro-refuse should Ethnoarchaeological descriptions of hunter-gatherer animal pro-
short-term camp curement, processing, and consurnption have proven to be of considerable
che case, the dis- value to prehistoric archaeology by providing models relating faunal remains
ace use. to the behaviors that produced them (e.g., Binford 1978; Bunn, et al. 1988;
1061 K. T. Jones
Table 6-2. Distance from Fire (em) for Activities in Three Ache Camps 1:43 I
bitsc
right
Debris Producing Nondebris Producing Both cling
men1
Men 58 ±25, N =30 62± 19,N= 58 61 ±21, N = 88
1:47
Women 54 ±22, N = 16 62±24, N= 71 61 ±23,N= 87 sharJ
steel
Both 57 ±24,N= 46 62±22,N=22 61 ± 22, N = 175
1:50.
Note: Observations made at 15-minute intervals; total of 3.75 hours observation time for 5 1:51
adult women and 5 adult men. way
fire.
O'Connell et al. 1988; Yellen 1977). Most of the work to date has involved 1:52
medium- to large-sized game, the most comprehensive work focusing on
caribou (Binford 1978). In addition, much of the analytical effort has focused 1:53
on relating butchering and discard patterns to logistical strategies involving
transport and delayed consumption of animal resources. 1:54
Smaller game anhnals have received considerably less attention (cf. Jones
1:55
1983; Stahl .1982; Yellen 1990) as have systems of animal procurement for
and
immediate consumption. There are a number of reasons for this. Small-animal ''bla:
remains are more difficult in many ways to deal with than are the remains of asp<
larger animals. They are more difficult to recover archaeologically, more likely
to suffer postdepositional damage and destruction, and more difficult to iden- 2:01
tify; they are often introduced to sites by nonhuman agents and, because of verb
their small body size, may be more likely to be thought of as insignificant into
dietary items. In the following section, I describe some aspects; of Ache
butchering and then discuss some ways data from the Ache can be .applicable 2:10
to archaeology.
The following is excerpted from my notes taken in 1984 on Ache butchering 2:14
a white-lipped peccary (Tayassu pecari):
2:27
Francisco shot @ 11 am 11/23 40 kg white-lipped peccary. skin
and
1:00 pm Meet women, build fire.
Although pe
1:15 Fire is ready. the way they 4
animals are bt
1:20 Jito throws carcass on fire, turning over and over, scrapes vigor-
ously with a flat stick, hot, heavy work. Throws it off fire at 1:35. consumption,
animals are bn
1:35 Begins cutting transverse cut through skin anterior of penis, then can be either rc
longitudinally left of ce!}ter up through muscles of abdominal cavity. ·butchered and
Loosens guts with hands, hands to a woman. Scoops pooled blood out of be either roash
abdominal cavity with hands, calls dogs (2), they drink it from leaves. available for h
Pulls out liver, hands it to a woman who puts it directly in fire. butchered for l
Continues cleaning abdominal cavity. monkeys are oj
A Short-Term Camp 1107
tree Ache Camps 1:43 Grandma works on gvts and other organs. Saves heart, lungs, little
bits of meat, begins cooking them right on the ashes and coals. Removes
right forelimb with scapula by taking a long cut along the chest, encir-
Both cling near the head of the humerus, (1:46) pulls back skin, severs attach-
ments,lays on leaves (1:47).
61±21,N=88.
1:47 Starts on other fore quarter. Pauses to have Enrique Tykwanangi
61 .± 23, N = 87 sharpen my knife (folding Gerber) on another knife blade (used like a
steel).
61 ± 22, N = 175
1:50 Left front quarter removed. Organs are roasting.
~rvation time for 5 1:51 Cuts with axe through 'ribs on both sides of sternum, split all the
way up to clavicles, cut off with knife. Hands sternum piece to women at
fire.
1te has involved 1:52 With axe, cuts ribs loose from vertebrae on both sides.
·ark focusing on
ffort has focused 1:53 Chop through pubis.
1tegies involving
1:54 Uses knife to remove penis and testicles.
ten tion (cf. Jones
procurement for 1:55 Cut through skin up along vertebral neural spines removing loin
and meat along vertebral column with skin and ribs as a pair of
1is. Small-animal
''blankets." Work carefully to get as much skin and meat off in one piece
re the remains of as possible.
cally, more likely
difficult to iden- 2:01 Chop through axis to remove head. Chop through third lumbar
; and, because of vertel:>ra, removing hind portion as single unit. Chop vertebral column
f as insignificant into two shorter lengths.
aspects of Ache
:an be applicable 2:10 Remove and separate hind quarters by chopping through sacrum.
Ache butchering 2:14 All major pieces are cooking.
2:27 Grandfather returns to camp, begins working on head. Removes

try. skin and flesh off mandible, to cook as a smaller "blanket," mandible
and rest of cranium cooked separately.
Although peccaries are the largest animals frequently taken by the Ache,
the way they are butchered is consistent with the ways most other game
animals are butchered. Animals are butchered for cooking and immediate
·er, scrapes vigor-
.re at 1:35. consumptiOn, which structure the ways they are dismembered. Initially,
animals are broken into segments suitable for the method of cooking, which
:ior of penis, then can be either roasting or boiling. In the case of peccaries, parts are minimally
::1bdominal cavity. butchered and cooked on large wooden roasting racks. Smaller animals may
ooled blood out of be either roasted or boiled or both, and segmented accordingly. As the pots
ink it from leaves. available for boiling are generally small (volume of 6 liters or less), animals
t directly in fire. butchered for boiling will be cut into smaller segments. Some animals such as
monkeys are often boiled and are butchered into pot-sjzed pieces. Armadillos
108j K. T. Jones
may be boiled or·roasted and are butchered differently depending on the Brea
method to be used. For ·roasting, the head and tail are removed, and the rest of
the body roasted in the carapace directly on the coals. For boiling, the head, 10:5(
tail, and limbs are removed and boiled, while the carapace with ribs is teun
roasted.
In general, Ache never skin animals. They singe off the hair and cook the 10:5~
animal in its skin. The "blanket of flesh" technique is used whenever feasible, him
especially on animals with substantial subcutaneous fat, such as peccaries and
10:5L
pacas. Initial butchering reduces animals to a manageable size for cooking and hisn
is done without regard to subsequent distribution. Consumption begins as
soon as the meat is cooked. Organs are often consumed before some other 10:5!
parts are even on the fire. Small pieces are cut from the cooked edges-of larger
roasting segments for immediate consumption. As cooked -pieces become 10:5~
available, they are passed from the cooking fire to others in the camp. When smal
larger animals such as peccaries are taken, the process of cooking and eating mad
can last well into the night and continue in the morning. The secondary
butchering is thus for consumption and is done on the cooked animal. Parts 11:01
are segmented as they are ready and are not cut according to a predetermined eats
package size. Secondary butchering generally is done with a knife, although a
machete or axe may be used to cut through bone when necessary. · 11:0:
This pattern of butchering produces relatively fewer cut marks on bone betV\
than would one in which uncooked animals ·are more completely dismem- 11:m
bered. Final butchering is done when the animal is fully cooked and cutting is and
less difficult. A study of faunal remains from 10 Ache camps showed cut
marks on only about 8% of the bones (306/3,939); however, a single larger 11;QL
animal in the collection showed cut marks on 24% of the bones (50/209; Jones back
1984). Cut marks noted on assemblages of larger animal remains among the toM
Hadza range from 24% to 28% (K. Lupo, personal communication 1991).
Other large animal assemblages showed cut marks on 8% to 69% of the bones 11:0(
(e.g., Binford 1981; Bunn 1983), so clearly there are variables other than just Sud
animal size at work.
Marrow-bearing bones from smaller animals (e.g., <15 kg) are generally
opened at the articular ends and the marrow sucked or pushed out. The ends Acl
may be bitten off or chopped off with a machete or knife. This leaves a dis-
tinctive bone cylinder. Of 124 small-animal long bones from 10 Ache camps Am
that were opened for marrow extraction, 41% were broken in mid-diaphysis of determining
and 59% had the articular ends removed (Jones 1984). The production of through time. 1
small-animal bone diaphysis cylinders may be useful in some cases in distin- the Ache make
guishing human- from raptor-produced assemblages (see Hockett 1991). rectangular she
Ache break marrow-bearing bones of larger animals by creating a spiral total site area "'
fracture down the shaft of the bone or by breaking the shaft into numerous ably smaller th
fragments. Following are descriptions from my notes of two incidents of 8.75 m 2•
marrow extraction from peccary long bones. Note that marrow extraction is The faunal n
done for immediate consumption as is all Ache butchering and processing. If within 1m of tl
no one feels like eating marrow, unopened bones will lie where dropped on skull and arma1
the ground. When someone feels like eating marrow, then the bone will be 3 m from the sl
opened and the marrow consumed. located about 1
pending on the Breaking a marrow bone. White-lipped peccary humerus, fully cooked.
l, and the rest of
)iling, the head, 10:50 Julio Kuaregi uses knife to carefully clean bone and scrape perios-
ce with ribs is teum from it.
ir and cook the 10:53 Jamo Tatugi (Julio's father) calls for Bepegi (Julio's son) to bring
Lenever feasible, him the bone to eat.
l.S peccaries and
10:54 He sets it by fire for about 1 minute, then breaks it with the back of
for cooking and his machete: first takes proximal end off, eats and sucks marrow out of it.
ption begins as
:ore some other 10:55 Splinters shaft, sucks e9-ch piece, drops each when done.
l edges of larger
pieces become 10:59 Splits distal end down center with sharp edge of machete, using
he camp. When small log as an anvil. Sucks one piece, then breaks it in two with back of
king and eating machete.
The secondary
~d animal. Parts 11:01 Uses tip of machete to scoop spongy bone out of the distal end and
, predetermined eats it off the machete. Done by 11:02.
nife, although a
Lry. 11:02 Starts on tib-fib. Splits the two apart to get the little meat left
/ between.
marks on bone
lletely dismem- 11:03 Chops prox end on diagonal with sharp edge of machete and sucks
d and cutting is and picks at the piece he cut off.
Lps showed cut
, a single larger 11:04 Works way down diaphysis with a series of about 20 blows from
s (50/209; Jones back of machete blade: Passes a big section of lateral shaft with marrow
ains among the to Marsano and his brother.
mication 1991).
9% of the bones 11:06 They all take twigs and poke them into the cavity to get juices.
other than just Suck and pick, all done at 11:08.
~) are generally
d out. The ends Ache Archaeology: The Buchu Site
lis leaves a dis-
10 Ache camps A month-old Ache site was collected in January 1985 with the goal
t mid-diaphysis of determining how well observations made during site occupation hold
~ production of
through time. The site had been roofed with a temporary ramada-like shelter
! cases in dis tin-
the Ache make when it rains. Four fires were placed at the drip lines of the
<ett 1991). rectangular shelter, one on each side. The sheltered area was 3.0 m by 2.5 m, a
reating a spiral total site area within the shelter of 7.5 m2. Sheltered sites tend to be consider-
into numerous ably smaller than open sites: the area covered on another sheltered site was
ATO incidents of
8.75 m 2 • ·
)W extraction is\
The faunal remains were almost entirely confined to the area under and
.d processing. If within 1m of tl}e sheltered area (Figure 6-2). A few larger pieces (the monkey
ere dropped on skull and armadillo pelvis) were found outside the area, up to approximately
he bone will be 3m from the shelter. Interestingly, se~eral bones were found in a tiny shelter
located about 1m from the main shelter. This was a small shelter constructed
110 IK. T. Jones
Animal
pet
Monkey (Cebu
Armadillo
(Dasypus nover
'~
.
..I
\
,'
(?) Boa c6nsttil
(?)
·~ NoiD
1
TOTAL
I
Note: Collected
shelter
leftovers carr
extraction an1
by pets. The
armadillo fen
0
taining little.
reasonably cc
Figure 6-2. The Buchu site. tion and proc
for a pet or pets, likely a young coati or peccary, and the bones found there
were the only pieces showing extensive gnawing damage. .1
A total of 312 bones were collected from the site (Table 6-3; armadillo
carapace fragments were not collected but would have numbered in the In
hundreds). Most of the elements (220) were from a single boa constrictor. Two the archaeolo.
l\.1NI armadillos we~e represented by 82 NISP, a single monkey by a skull and for a very sha
mandible, and 8 pieces were unidentifiable fragments. Approximately 5% (17) habitation. T1
of the bones exhibited cut marks, in line with the 8% figure from occupied concentratior
sites, and 11% (34) had been noticeably burned. secondary di
The two bones from the monkey cle~rly represent 11leftovers" carried from campsites, wl
another camp. This was confirmed by informant testimony. The snake case of the A
appears to be completely or nearly completely represented. The body part locales, or she
representation for the two armadillos shows 100% representation for the head, blage campo:
forelimb (humerus and ulna), and pelvis. Other larger parts are represented immediate cc
by 50% to 75% of those expected, while smaller parts (vertebrae, ribs, foot expected in '\
parts) are less than 50% present. A notable exception is the femora, repre- occur.
sented by a single proximal portion. The patten
In general, the pattern of body part representation is consistent with the described her
expectation that complete animals were introduced to the site. The single sufficient to s
monkey represents leftovers, and the missing armadillo femora may represent duce similar 1
Table 6-3. Buchu Site Faunal Remains
Animal NISP MNI Cut Marked (%) Burned(%)
Monkey (Ce.bus apella) 2 2 (100) 0
Armadillo
(Dasypus novemcinctus) 82 2 4 (5) 7 (9)
(?) Boa constrictor

(?) 220 1 11 (5) 21 (10)
NoiD 8 0 6 (75)
TOTAL 312 4 (11) 17 (5) 3
0
I Note: Collected (100%) one month after abandonment.
=o_ I
I leftovefs carried away from the site, they may have been broken for marrow
~
/ extraction and reduced to small fragments, or they may have been consumed
/ by pets. The most likely explanation is that they were carried away, as an
armadillo femur is a very robust bone, difficult to break for marrow, and con-
taining little. From the distribution of parts and their representation, we could
reasonably conclude that this assemblage is the result of a short-term occupa-
tion and processing for immediate consumption.
1es found there

Conclusions
6-3; armadillo
In this paper I have described some aspects of Ache camp use and
tmbered in the
the archaeologically rele,vant patterning that results. Ache camps are occupied
:onstrictor. Two
for a very short tilne and provide an example of the extreme end of short-term
y by a skull and habitation. The patterning evident in Ache camp structure, especially in the
imately 5% (17)
concentration of activities and refuse around fire hearths and little or no
from occupied
secondary disposal of refuse, is expected to be characteristic of short-term
campsites, whether they are the result of generalized foraging, such as in the
·s" carried from
case of the Ache, transient camps, camps at specialized resource extraction
1ny. The snake
locales, or short-term base camps. The patterning in butchering, faunal assem-
The body part
blage composition, and body part representation is related to foraging for
on for the head,
immediate consumption and use of small animals, and the patterns may be
are represented
expected in varying degrees to accompany those strategies wherever they
~brae, ribs, foot
occur.
femora, repre-
The patterning noted in archaeological sites, whether like the patterns
described here or in other reported ethnoarchaeological observations, is not
;istent with the
sufficient to stand alone in identifying its causes. Divergent causes can pro-
site. The single
duce similar patter_ns, and similar patterns can have 9ivergent causes. When
1 may represent
1121 K. T. ]ones
linked to hypotheses about the kincl of occupation expected in a given circum-· Hunter-Gal
stance, the patterns identified ethnoarchaeologkally, along with the variabil- edited byE
ity noted among and between differing kinds of sites, can be powerful tools Binford, L. R
for use in illuminating the archaeological re.cord. If a short-term occupation of 1978 Nunart
a particular site is expected or suspected, the patterns descriJ?ed here can be 1Q81 Bones:
used in conducting a test of the expectation. The strategy for data collection Borrero,L. A., a1
1989 Etnoa:
can be designed to obtain relevant information (e.g., exposure of large areas Americanist,
around hearths), and the data can l:>e contrasted with patterns described for Bunn,H. T.
short-term and longer-term sites, resulting in a relativ_ely more robust solution 1983 Comp:
than would simple pattern-matching. Gatherer C
One of the most important contributions ethno·archaeology can make to Den near I\
archaeology is to provide examples of how strategies are reflected in the edited byJ.
archaeological record. It does the archaeologist no good to know what an 163. Oxford
Ache campsite looks like. Few archaeologists will ever excavate an Ache site. Bunn, H. T., L. E
It is incumbent on the ethnoarchaeologist to provide ways for the data to be- 1988 Variab
come useful as a tool for real archaeologists. The patterns I have described are ing, and Ca
the result of short-term campsites, processing for immediate consumption, Clastres, P.
and small-animal use. Those strategies need not always co-occur; they happen 1972 Chroni,
to in the case of the Ache. Expectations about how the strategies may be Fisher, J. W., Jr.,
1989 Ethnoc
represented in an archaeological site qm be posed in opposition to expecta- Campsites.
tions about how longer-term sites, processing for delayed consumption, and 1991 Dwelll
large-animal use will be represented. We need not siinply use ethnoarchaeo- Ethnoarchae,
logkal information to make ethnographic analogy but can incorporate it into A. Boismie1
investigation via hypothesis testing. The strength of such tests, and the inter- ological Ser
pretations and explanations that derive from them, is enhanced by consider- Gould, R. A., anc
ing the body part representation, damage morphology, and distribution as 1987 Man t
important aspects of the information relevant to the behavioral strategies that Tropical Fo
produced them. Harako, R.
1981 The C1
Ituri Forest
Acknowledgments Teleki, pp. •
Hawkes, K., K. t
The research reported here was supported by the NSF (Grant 1982 Whyt
#BNS-8309834), the NIH (Grant #1 R01 HD16221~01A2), the Evert American Et
Foundation, and the Utah Division of State History. I thank Jean Hudson for Hill, K., and K. t
1983 Neotrc
inviting me to participate in the conference, for sharing unpublished data with
Responses oj
me, and for fruitful discussions of tropical forest fo),'agers. I also thank Kim 139-188. Ac
Hill, Hillard Kaplan, Ana Hurtado, and Kristen Hawkes for the large parts Hockett, B.S.
they played in all aspects of my participation in the Ache research. David B. 1991 Towar
Madsen, Steven R. Simms, James F. O'Connell, and Duncan W. Metcalfe American A1
contributed ideas and useful criticism. Jean Hudson and one anonymous Hudson,J.
reviewer were also most helpful in improving the manuscript. 1988 Spatia:
possession 1
·Jones, Kevin T.
References 1983 Forage
Scavengers,t
Bartram, L. E., E. M. Kroll, and H. T. Bunn and A. S.1
1991 Variability in Camp Structure and Bone Food Refuse Patterning at Kua San Calgary.
A Short-Term Camp jl13
a given circum- Hunter-Gatherer Camps. In The Interpretation of Archaeological Spatial Patterning,

ith the variabil- editedby E. M. Kroll andT. D. Piice, pp. 77-148. Plenum Press, New York
powerful tools Binford, L. R
1978 Nunamiut Ethnoarchaeology. Academic Press, New York.
n occupation of
1981 Bones: Ancient Men and Modern Myths. Academic Press, New York
Jed here can be
Borrero, L.A., and H. D. Yacobaccio
data collection 1989 Etnoarqueologia de Asentamientos Ache. Journal de la Societe des
e of large areas Americanistes LXXV: 7-33. Buenos Aires.
ts described for Bunn,H. T.
robust solution ·1983 Comparative Analysis of Modern Bone Assemblages from a San Hunter-
Gatherer Camp in the Kalahari Desert, Botswana, and from a Spotted Hyena
~y can make to Den near Nairobi, Kenya. In Animals and Archaeology 1: Hunters and Their Prey,
·eflected in the edited by J. Clutton-Brock and C. Grigson, pp. 143-148. BAR International Series
know what an 163. Oxford.
.te an Ache site. Bunn, H. T., L. E. Bartram, and E. M. Kroll
~ the data to be- 1988 Variability in Bone. Assemblage Formation from Hadza Hunting, Scaveng-
re described are ing, and Carcass Processing. Journal of Anthropological Archaeology 4:412-457.
~ consumption,
Clastres, P.
1972 Chronique des Indiens Guayaki. Plon, Paris.
ur; they happen Fisher, J. W., Jr., and H. C. Strickland
ategies may be 1989 , Ethnoarchaeology among the Efe Pygmies, Zaire: Spatial Organization of
tion to expecta- Ca,npsites. American Journal of Physical Anthropology 78:473-484.
1sumption, and 1991' Dwellings and Fireplaces: Keys to Efe Pygmy Campsite Structure. In
~ ethnoarchaeo- Ethnoarchaeological Approaches to Mobile Campsites, edited by C. S. Gamble and W.
:orporate it into A. Boismier, pp. 215-236. International Monographs in Prehistory, Ethnoarchae-
;, and the inter- ological Series 1. Ann Arbor.
ed by consider- Gould, R A., and J. E. Yellen
distribution as 1987 Man the Hunted: Determinants of Household Spacing in Desert and
tl strategies that Tropical Foraging Societies. I ournal of Anthropological Archaeology 6:77-103.
Harako, R.
1981 The Cultural Ecology and Hunting Behavior among Mbuti Pygmies in the
Ituri Forest, Zaire. In Omnivorous Primates, edited by R. S. 0. Harding and G.
Teleki, pp. 499-555. Columbia University Press, New York.
Hawkes, K., K. Hill, and J. F. O'Connell
he NSF (Grant 1982 Why Hunters Gq.ther: Optimal Foraging and the Ache ofEastern Paraguay.
~2), the Evert American Ethnologist 9:379-398.
Hill, K., and K. Hawkes
~an Hudson for
1983 Neotropical Hunting among the Ache of Eastern Paraguay. In Adaptive
.ished data with Responses Qf Native Amazonians, edited by R. B. Hames and W. T. Vickers, pp.
also thank Kim 139-188. Academic Press, New York.
the large parts Hockett, B. S.
earch~ David B. 1991 Toward Distinguishing Human and Raptor Patterning on Leporid Bones.
1n W. Metcalfe American Antiquity 5~:667-679.
'ne anonymous Hudson,J.
1988 Spatial Analysis of Faunal Rema1ns in Hunter-Gatherer Camps. Ms. in
possession of author.
Jones, Kevin T.
1983 Forager Archaeology: The Ache of Eastern Paraguay. In Carnivores, Human
Scavengers, and Predators: A Question of Bone Technology, edited by G. M. LeMoine
and A. S. MacEachern, pp. 171-191; Archaeological Association, University of
·ming at Kua San Calgary.
1141 K. T. Jones
1984 Hunting and Scavenging by Early Hominids: A Study in Archeological Method
and Theory. Unpublished Ph.D. dissertation, Deparb:nent of Anthropology,
University of Utah, Salt Lake City.
Kent, S.
1991 The Relationship Between Mobility Strategies and Site Structure. In The
Interpretation of Archaeological Spatial Patterning, edited by E: M. Kroll and T. D.
Price, pp. 33-59. Plenum Press, New York
Lee, Richard B. 7. p
1979 The !Kung San: Men, Women and Work in a Foraging Society. Cambridge
University Press, Cambridge.
p
Nicholson, A., and S. Cane
1991 Desert Camps: Analysis of Australian Aboriginal Proto~historic Campsites.
E
In Ethnoarchaeological Approaches to Mobile Campsites, edited by C. S. Gamble and
W. A. Boismier, pp. 263-354. International Monographs in Prehistory, Ethno- Lt
archaeological Series 1. Ann Arbor.
O'Connell, J. F. ·
· 1987 Alyawara Site Structure and Its Archaeological Implications. American At
Antiquity 52:74-108. ha
O'Connell, J. F., K. Hawkes, and N. Blurton Jones De:
1988 Hadza Hunting, Butchering, and Bone Transport and Their Archaeological ca:
Implications. Journal of Anthropological Research 44:113-162. ole
1991 Distribution of Refuse-producing Activities at Hadza Residential Base
Camps: Implications for Analyses of Archaeological Site Structure. In The Inter- an
pretation of Archaeological Spatial Patterning, edited by E. M. Kroll and T. D. Price, (0
pp. 61-76. Plenum Press, New York ast
Stahl, Peter W. etc:
CaJ
1982 On Small Mammal Remains in Archaeological Context. American Antiquity
47:822-829. rrr
Yellen, J. E. as:
1977 Archaeological Approaches to the Present. Academic Press, New York. ole
1990 Small Mammals: !Kung San Utilization and the Production of Fa~nal Assem- sal
blages. Ms. in possession of author. ;'
co:
be:
cal
pr1
po
inf
In
Til
tify site type:
nutritional di:
bone assemb
From Bones to Be.

Faunal Remains, 1
Paper No. 21. Q

reserved. ISBN C
:heological Method
f Anthropology,
Structure. In The
L Kroll and T. D.
7. Perspectives on Skeletal Part
ciety. Cambridge
Profiles and Utility Curves from
storic Campsites.
Eastern Kalahari Ethnoarchaeology
C. S. Gamble and
rehistory, Ethno- Laurence E. Bartram, Jr.
ations. American Abstract: Recent ethnoarchaeological research in the Botswana Kalahari

has documented how bone assemblages accumulate at Kua (San) camps.
Data from over 30 camps reflect variability in Kua hunting,· butchery,
ir Archaeological carcass transport, and bone discard patterns of interest to zooarchae-
ologists.
Residential Base Bone transport patterns are the focus of this paper. Analytical results
ture. In The Inter- are reported for bone assemblages from kill/butchery camps of gemsbok
(Oryx gazella) and their complementary base camp assemblages. The
n and T. D. Price, assemblages are described in terms of their formation histories and skel-
etal part profiles. Carcass size, season, group size, and distance to base
Lmerican Antiquity camp influenced butchery and transport strategies. These, in turn, dete-
rmined the skeletal part composition of the primary butchery camp bone
assemblages, sometimes in unexpected ways. Of interest to archae-
:wYork. ologists is the considerable variability evident among assemblages of the
of Faunal Assem- same "type."
Because zooarchaeologists partly assess site function through the
construction of utility curves, identifying the kinds of butchery I discard
behavior that result in specific assemblage patterns has important impli-
cations for the construction of utility curves and for archaeological inter-
pretations of site type/function in general. Meat drying is identified as a
potent, and relatively unstudied, distorting force of great antiquity
influencing skeletal part profiles.
Introduction
This paper is concerned in general with how archaeologists iden-.
tify site types from faunal remains. More specifically, I ask, How well do
nutritional differences in carcass parts explain skeletal element frequencies in
bone assemblages from Kalahari forager camps with known formation

115
116j L. E. Bartram, J.r.
histories? I shall limit the discussion to a single species, the southern African tionships with 1
oryx (Oryx gazella), or gemsbok. With two exceptions, the carcasses considered which nutritio-c
here were all hunted and butchered by the same group of Kua foragers during lines have comE
a 2 1/2-month period. I present skeletal part data from nine gemsbok site types. Othe
kill/butchery sites and data on their representation in seven complementary have·also been c
base camp assemblages. I conclude that factors other than nutritional utility Binford appli
strongly influence skeletal part composition of the assemblages. formation histo1
cally derived u
Skeletal Part Frequencies and Nutritional Variability expectations, kil
rare high-utility
For more than a century archaeologists have tried to make behav- generally more
ioral sense of variability in skeletal element abundances. Inferences about because of their
prehistoric human behavior and site function have been based on the recogni- potential of util
tion that, unlike corn-dogs-on-a-stick at the carnival, the quantities of meat context.
and fat on the "sticks" of animal carcasses are variable. The butchery and Binford's infl
transport decisions of prehistoric people are also thought to have been medi- front of zooarcll
ated by this recognition (e.g., Binford 1978, 1981, 19_84; Chase 1985; Emerson tions of Binfon
1990a, 1990b; Frison 1974; Grayson 1989; Jones and Metcalfe 1988; Lartet and Grayson 1988, ~
Christy 1865-1875; Lyman 1985; Metcalfe and Jones 1988; Perkins and Daly Mayer 1983), an
1968; Speth 1983; Thomas and Mayer 1983; Wheat 1972; White 1954). focused on ref
Evidence from even the earliest archaeological sites indicates that hominids economic anatc
processed carcasses that were larger than they could possibly carry intact Caro 1986; Bor:
(Bunn 1982; Bunn and Kroll 1986). Intuitively, one would expect successful reduced the con
hominids to make rational economic choices and to make butchery and trans- more easily calc
port decisions with food value in mind. As a result of butchery and transport, been attemptin~
the bones of a single carcass end up in one or more clusters on the landscape quency (Bartrar
and define the places used during various stages of carcass processing. The Bunn et al. 1988
decisions of hominids would be archaeologically detectable as distinctively
patterned skeletal element frequencies according to this view. . Tro~
With other taphonomic biases removed, one would expect analysis/to reveal
groups of assemblages related by similarities in element frequencies. The Deli,
groups would be composed of assemblages with similar processing and suggested by th
transport histories, reflecting aspects of subsistence organization and an embraced utilir
appreciation by the hominids for the differences in food value between descriptive stat
carcass parts. been so readily
Working among the caribou-hunting Nunamuit, Lewis Binford was the first thirst for analyt
to juxtapose relative bone frequencies with quantitative measures of eco_nomic ior. Thirsty, we
anatomy. He collected data from the carcasses of three of the most famous For utility inc
animals ever to die for science: a caribou and two sheep. From these animals, to know how a<
Binford developed a number of quantitative indices of nutritional utility. The of a bone asse
most often cited of these is what he called the MGUI, the Modified General revealing the in
Utility Index, a combined measure of the meat, marrow, and grease associated factors, both pr
with each bone. warning by ex
Binford illustrated relationships with scatterplots of standardized skeletal '(Grayson 1988:
part abundances arrayed against their respective MGUI values. He reasoned faced by zooarc
that regression lines with negative slopes would be expected to characterize sidering other
kill/butchery sites from which nutritious elements were removed, while rela- 1984), in evalua
Skeletal Part Profiles 1117
outhern African tionships with positive slopes would characterize assemblages from sites to
:tsses considered which nutritious elements had been transported. The resulting regression
foragers during lines have come to be known as "utility curves," thought to be indicative of
nine gemsbok site types. Other theoretical variants such as ''bulk" and "gourmet" curves
complementary have also been discussed as they might reflect different transport strategies.
utritional utility Binford applied his method to Nunamuit bone assemblages with known
~s. formation histories. The resulting scatterplots were compared to the theoreti-
cally derived utility curves and assessed for goodness-of-fit. According to
ty expectations, kill/butchery sites generally had abundant low-utility parts and
rare high-utility parts. By contrast, at residential sites high-utility parts were
to make behav- generally more abundant, although there was greater variability in this group
nferences about because of their complex formation histories. Overall, though, the explanatory
I on the recogni- potential of utility curves was demonstrated convincingly in the Nunamuit
.antities of meat context.
Le butchery and Binford's influential research propelled utility curve analysis to the fore-
1ave been medi- front of zooarchaeological methodology. Several purely archaeological applica-
~ 1985; Emerson tions of Binford's methodology have been published (Binford 1981, 1984;
1988; Lartet and Grayson 1988, 1989; Landals 1990; Speth and Speilmann 1983; Thomas and
~rkins and Daly Mayer 1983), and more are appearing. Since his original research, others have
1954). focused/ on refining the technique. Some have worked on enlarging the
~s that hominids econoniic anatomy data base for different species (e.g., Blumenschine and
.bly carry intact Caro 1986; Borrero 1990; Emerson 1990a, 1990b). Others have successfully
xpect successful reduced the complexity of deriving utility indices for other taxa by developing
chery and trans- more easily calculated alternatives (Metcalfe and Jones 1988). Still others have
y and transport, been attempting to refine the accuracy of our methods for estimating part fre-
m the landscape quency (Bartram and Bunn n;d.; Bartram et al. 1991; Bunn and Bartram n.d.;
processing. The Bunn et al. 1988; Marean and Spencer 1991).
as distinctively
Trouble for Utility Curves
.nalysis to reveal
requencies. The Delighted with the prospects for archaeological interpretation
processing and suggested by the relatively clear-cut Nunamuit examples, archaeologists have
liz a tion and an embraced utility curves. In fact, utility curves have nearly become a standard
value between descriptive statistic. Donald Grayson has written, J/That these curves have
been so readily adopted by archaeologists is in large part a function of our
ord was the first thirst for analytical techniques that will validly inform on past human behav-
1res of ecqpomic ior. Thirsty, we drink" (Grayson 1988:123).
:1e most famous For utility indices to have real explanatory meaning, it is obviously essential
n these animals, to know how accurately the curves one generates reflect the original contents
onal utility. The of a bone assemblage. Taphonomic work conducted arou!),d the globe is
:odified General . revealing the importance of understanding the effects of a host of nonhuman
;rease associated factors, both predepositionally and postdepositionally. Grayson echoed this
warning by extending his metaphor-"the waters are not always pure"
ardized skeletal (Grayson 1988:123). He was referring to a potentially widespread problem
es. He reasoned faced by zooarchaeologists attempting to use utility curves without first con-
L to characterize sidering other factors, especially bon_e density (e.g., Grayson 1989; Lyman
>ved, while rela- 1984), in evaluating part frequencies. His overall message was one of caution
118j L. E. Bartram, Jr.
in the use of utility curves for arclmeological inference._

It must be stressed,· however, that even in the complete absence of tapho-
nomic and sampling distortions, interpreting skeletal part frequencies remains
a tricky business (Bartram et al. 1991:78--:79). We have yet to sort out what
kinds of human behavior leave interpretable element diffe~ences before we
can employ utility curve analyses as widely as we might like, even if we be-
come able to conclusively eliminate taphonomic biases from our data. It is
here that-ethnoarchaeological studies are important.
The remainder of this paper is devoted to explori;ng the factors that condi-
tion skeletal part frequencies at camps of the Kua of southern Africa. I shall
consider (1) the effect of transport costs on Kua bone assemblage composition
and (2) one transport cost reduction strategy that may be relevant to even the
oldest archaeological sites.
The data to illustrate those points come from my analyses of gemsbok
kill/butchery assemblages and their complementary base ca1np assemblages. I
first turn to a description of the context of the research.
The Eastern Kalahari and the Kua

The Environment
The study area is located in the Republic of Botswana near the
eastern border of the Central Kalahari Game Reserve (Figure 7-1). The lack of
ink on government maps says much about this area of the Kalahari.
Unconsolidated aeolian sands form its surface, and except for rare dunes and
pans, the region is as flat as a pancake.
The climate of the region is subtropical and semiarid, and precipitation falls
almost exclusively between November and April. The vegetation in/this part
of the Kalahari is open Acacia savanna, punctuated with rare "i&1ands" of
trees. The presence of significant sources of fluids in the tubers and fruits of
several plant species enables animals and people to live in a region otherwise
characterized by a total lack of surface water. In many of its features, the
region is similar to the area inhabited by the G /wi described by Silberbauer
(Silberbauer 1981) and Tanaka (Tanaka 1980).
More than 50 mammalian species, plus abundant birds and reptiles, inhabit
this part of the Kalahari. Among them, at least a dozen are hunted regularly
by the Kua (Table 7-1), and about a dozen more are taken less frequently. The
gemsbok considered in this paper are at the larger end of the size range, Figt
weighing between about 200 and 250 kilograms (Smithers 1983).
The Kua residential spe

ranges from se<
The Kua belong to the poorly defined category of Bushmen, San, Xu a rna y be fm
or Basarwa (Elphick 1977:xxi-xxii, 4-7, 23-42; Lee 1979:29-35; Silberbauer different times
1981:3-6; Tobias 1978:1-3) but refer to themselves only as Kua. Kua who were cattle-post econ
full or part-time foragers were the subject of this study. An economic and Kua residenti
the period of ob
:bsence of tapho-
ruencies remains
:o sort out what
ences before we
~, even if we b~
n our data. It is
~:tors
that condi-
~nAfrica. I shall
age composition
vant to even the
-ses of gemsbok
p assemblages. I Central Kalahari
Game Reserve
G/wi Area
I
:swana near the
7-1). The lack of
f the Kalahari.
: rare dunes and
recipitation falls
ttion in this part 0 50 100 200 KM
are "islands" of
ers and fruits of
·egion otherwise 0 Capital
its features, the District Boundary
i by Silberbauer Reserve Boundary
f"...../ River
reptiles, inhabit
mnted regularly
; frequently. The
the size range, Figure 7-1. The Kua research area in the Republic of Botswana.
3).
residential spectrum is evident among the inhabitants of the region that
ranges from sedentary cattle-post living to mobile foraging (Hitchcock 1978).
f Bushmen, San, Kua may be found from one end of this economic spectrum to the other. At
-35; Silberl{auer different times of the year, some Kua shift from complete dependence on the
. Kua who were cattle-post economy to full-time foraging .
1 economic and Kua residential mobility produced four distinct types of occupations during
the period of observation: (1) base camps (sensu Binford 1978:488, 491; Binford
I
120 L. E. Bartram, Jr.
Table 7-1. Some Mammals Hunted by the Kua · est shaded area
the liver and a
Metapodials wE
Taxon Cominon Name Weight Range (kg)
Dec1
Giraffa camelopardalis Giraffe 703.0-1395.0
Taurotragus oryx Eland 400.0-700.0 Att1
Oryx gazella Gemsbok 210.0-240.0 transport at thE
Connochaetes taurinus · Blue wildebeest 180.0-250.0 effect on detern
Tragelaphus strepsiceros Greater kudu 119.6-258.2 extent. of furth
Alcelaphus buselaphus Red hartebeest 105.0-156.0 made by evalu
Orycteropus afer Antbear 40.4-64.5 time of day the
Antidorcas marsupialis Springbok 30.4-47.6 distance to the 1
Hyaena brunnea Brown hyena 28.0-47.5
One of four t
Sylvicapra grimmia Duiker 15.3-25.4
Hystrix africaeaustralis Porcupine 13.6-24.1
in minimally d
Raphicerus campestris Steenbok 8.9-13.2 parts, mainly m
Proteles cristatus Aardwolf 7.7-10.0 For gemsbok
Manis temmincki Pangolin 4.5-14.5 The third (mea
Otocyon megalotis Bat-eared fox 3.2-5.4 gems bok and k
Pedetes capensis Springhare 2.9-3.9 for giraffe only.
The time of c
desire to return
Note: Average adult minimum and maximum weights for each taxon are from Smithers and a large cart
(1971, 1983).
would be no fu
body parts bac]
and Binford 1966) of hot dry, cool dry, and rainy season varieties, (2) transient people were a\
camps, (3) special-purpose locations, and (4) special-purpose camps where an the kill so that
overnight stay was involved and a more substantial structure was con- lions in the are
structed. I lions or other
hunter did not
his prints to thE
Formation of Kua Bone Assemblages In any case,
extensive carcc:
Primary Butchery of Gemsbok biltong, or sun-
sun, hanging .i:
At the rainy season and cool dry season camps hunters were
During the rain
occupied with the bow-hunting of larger game (principally gemsbok [Oryx
it frmn becomi
gazella], eland [Taurotragus oryx], and greater kudu [Tragelaphus strepsiceros]).
Larger game was more easily taken during this season for several reasons but weight of 4.65 :
primarily because of the fresh availability of arrow poison. dramatically n
Primary butchery of a large carcass commenced immediately upon its dis- which there are
covery by a tracking party. The usual butchery program for gemsbok began at reduce significc:
the shade at a r
or very near the death site, by chopping off the horns and skinning the
carcass, which was then eviscerated and quartered. Rib slabs were removed .in bil tong weig
mum valuefor
by chopping the ventral ends of the ribs away from the sternum and chopping
weights are a vc;
the rib heads and vertebral processes along each side of the vertebral centra.
The head and vertebral column were segmented into three or four units by Marrow bon
chopping and cutting. Quartered units were then usually moved to the near- biltong produc
est shaded area for more extensive processing. A fire was built nearby to roast
the liver and a portion of a rib slab removed early in the butchery process.
Metapodials were frequently removed, skinned, roasted, and cracked.
'eight Range (kg)
Decisions Influencing Carcass !ransport
703.0-1395.0
400.0-700.0 At this stage the factors that conditioned additional processing and
210.0-240.0 transport at the primary butchery site are of interest because of their critical
180.0-250.0 effect on determining the composition of the resulting bone assemblages. The
119.6-258.2 extent of further butchery depended upon consensus decisions that were
105.0-156.0 made by evaluating four key variables: (1) size of the carcass obtained, (2)
40.4-64.5 time of day the carcass was recovered, (3) number of available carriers, and (4)
30.4-47.6 distance to the base camp.
28.0-47.5 One of four transport strategies was chosen: (1) carrying the entire carcass
15.3-25.4
in minimally disarticulated units, (2) carrying extensively processed carcass
13.6-24.1
8.9-13.2
parts, mainly meat, (3) caching the meat, and (4) moving camp to the kill.
7.7-10.0 For gemsbok, only the first two of the transport strategies were observed.
4.5-14.5 The third (meat caching) was observed only for eland and only reported for
3.2-5.4 gemsbok and kudu, and the last (moving camp to the kill site) was reported
2.9-3.9 for giraffe only.
The dme of day a carcass was recovered was occasionally at odds with the
desire to return to camp before nightfall. If a carcass was found late in the day,
re from Smithers and a large carrying party was available at or near the kill site, there normally
would be no further butchery, just the transport of the large, bone-containing
body parts back home. On the other hand, if it was late and only one or two
ies, (2) transient people were available to carry meat, a windbreak or hut was constructed at
:amps where an the kill so that butchery could continue the following day. The presence of
cture was con- lions in the area made this an unpopular decision, but sharing the kill with
lions or other scavengers was not looked upon favorably, either. When a
hunter did not return to camp, the next morning others would usually track
his prints to the kill site, hoping to assist in carrying the meat.
. In any case, if butchery continued at the kill site, the next step involved
extensive carcass defleshing. The goal of this stage was the production of
biltong, or sun-dried meat. The meat was cut into strips and dried in the direct
sun, hanging in trees, bushes, or on quickly constructed branch platforms.
s hunters were
During the rainy season this step also helped to preserve meat and prevented
gemsbok. [Oryx
it from becoming fly blown. Biltong was tied into bundles with an average
:us strepsiceros]).
weight of 4.65 kg (N = 31). Despite the extra time it took, biltong production
era! reasons but
dramatically reduced the weight of transported meat. On a sunny day, of
which there are many in .the Kalahari, the process of biltong production could
~ly upon its dis-
reduce significantly the weight of the load to be carried back to camp. Even in
·msbok began at
the shade at a rainy season camp eight hours of drying resulted in a 29% loss
.d skinning the
in biltong weight. The weight loss was far more pronounced than this mini-
: were removed
mum value for biltong dried in the direct sun at kill sites, although no precise
n and cho:Rping
weights are available.
rertebral centra.
Marrow bones exposed during defleshing were cracked and eaten during
)r four units by
biltong production at the kill site instead of being tran~ported unbroken back
ved to the near-
L
to the campsite. The resulting fragments were discarded at the breakage loca- pieces, the spe1
tion, leaving a pile of epiphyseal and limb shaft fragments discarded next to anatomical poE
the anvil on which they were cracked. Biltong production and subsequent imens, like lim~
marrow cracking had profound effects on the composition of bone- assem- conjoined or p
blages related to these animals, as large amounts of debo,ned meat were cise,paid off in
transported to the camp sites. ably less time-<
total number o:
sites to which 1
Analysis blages, a certa:
eluded assignr
Sample Characteristics and specimen
abundance esti
Ten gemsbok are included in the analyses, but miss.ing values pre- All- bone spe
vented some carcasses from figuring in certain analyses. Eight of the gemsbok and Crader's P
were killed by Kua hunters during the late rainy and early dry seasons of (Gifford and C:
1986; the other two were killed about one and two years earlier, respectively. taxonomic ider
For these carcasses, complete data were unavailable about the distance to base
camps and details of the transported assemblage. All were adult or nearly Que
adult animals. Cor
The gemsbok were all killed with poisoned arrows. Tracking/butchery skeletal eleme1
parties ranged in siz2 from one to seven people. The butchery camps were ments into laq
occupied from a few hours to as long as 36 hours, and ranged from 3 to 7 raised through
kilometers from base camps. Two camps with missing data on this variable and previous!:
were considerably more distant, as much as 15 or 20 kilometers from base This was espec:
camp. been scavengec
Bone Collection seal portions ((
Bunn 1982; Bu
Complete kill/butchery and campsite bone assemblages from the Spencer 1991).
nine gemsbok were collected for subsequent analysis. Every effort w,as made Once MNEs
to ensure 100% recovery, but there is no doubt that some of the smallest bones each carcass. 1
were missed. Recovered bones were cataloged, plotted on maps of the camps, dance and th(
and shipped for more detailed study at the University of Wisconsin, Madison, (Binford 1978,
as part of a larger project on Kua ethnoarchaeology. dividing the el4
The bone collections were made at varying intervals after discard. Two of a Minimum A1
the nine kill/butchery sites considered here had been abandoned for at least each element _i
one year at the time ofcollection, four others were visited within a week of blage. The que
abandonment, one was mapped about a month after abandonment, and three abundance as a
other assemblages were collected immediately upon abandonment. At the This statistic
last-named camps, details of gemsbok butcheries that took place there were blages each inc:
also recorded. The base camp assemblages were collected immediately after 7-2 are expres~
occupation or, in several cases, during the occupation. The varying collection element are pn
intervals allowed the effects of postdepositional taphonomic factors to be Another exp
expressed in the assemblages. . the sum of the
Bone Coding in a complete
single value so
Prior to recording, the entire collection of bone specimens from The completen
each camp was laid out on lab tables. Beginning with the most diagnostic for the magnitt
breakage loca- pieces, the specimens were laid out in a pattern approximating their original
:carded next to anatomical position to help with element identification. Less diagnostic spec-
nd subsequent imens, like limb shafts, vertebra fragments, and rib shaft fragments, were then
>f bone assem- conjoined or positioned for detailed comparison to the larger sets. This exer-
.ed meat were cise paid off in low frequencies of nonidentifiable specimens. It was consider-
ably less time-consuming for the single carcass kill/butchery sites, where the
total number of specimens was usually less that 400, than it was for the camp-
sites to which more than one gemsbok had been transported. In those assem-
blages, a certain number of fragments, mainly vertebral processes and ribs
eluded assignment to a specific carcass despite my best efforts at conjoining
and specimen comparisons. Nonetheless, I have confidence in the resulting
abundance estimates. ·
ing values pre- All bone specimens were then coded, using a modified version of Gifford
)£ the gemsbok and Crader's Part/Portion/Segment system for recording element frequencies
dry seasons of (Gifford and Crader 1977), along with other details about butchery marks and
·r. . respectively. taxonomic identity (Bartram 1993; Bartram et al. 1991).
listance to base
Ldult or nearly Quantification of Part Abundances
Comprehensive MNEs (e.g., Bunn 1986) were computed for each
king/butchery skeletal.element in the assemblages by attempting to conjoin all bone frag-
~y camps were ments i~to larger, more identifiable sets. In many cases, MNE values were
ed from 3 to 7 raised through the conjoining exercise, where the identity of several smaller
n this variable and previously nondescript bone fragments was established conclusively.
ters from base This was especially evident among limb bones from collections known to have
been scavenged by brown hyenas and jackals, who made off with the epiphy-
seal portions (e.g., Bartram 1993; Bartram et al. 1991:134; Blumenschine 1988;
Bunn 1982; Bunn et al. 1980; Bunn and Kroll 1988; Klein 1975; Marean and
1lages from the Spencer 1991).
Jort was made Once MNEs had been determined, %MAU values were also calculated for
smallest bones each carcass. This is a standardized measure of relative skeletal part abun-
s of the camps, dance and the most commonly used measure in utility curve analyses
nsin. . Madison, (Binford 1978, 1984; Thomas and Mayer 1983). The %MAU is derived by (1)
dividing the element's MNE by its frequency in a complete skeleton, yielding
iscard. Two of a Minimum Animal Units value (MAU) and (2) then dividing the MAU for
1ed for at least each element in the assemblage by the highest MAU value for that assem-
thin a week of blage. The quotients are then multiplied by 100, to express each element's
1ent. . and three abundance as a percentage of the most frequent element.
nment. At the This statistic is particularly appropriate here because the original assem-
ace there were blages each include the bones of a single carcass. The horizontal axes in Figure
nediately after 7-2 are expressed in %MAU units, ·and %MAU statistics for major skeletal
ying collection element are presented for each camp in Table 7-2.
: factors to be Another expression of carcass completeness was also derived by dividing
the sum of the MNEs estimated for each carcass by the sum of those elements
in a complete a.nimal. This fraction summarizes carcass completeness on a
single value so that whole assemblages can be arrayed against other variables.
Jecimens from The completeness statistic obscures precisely which elements are responsible
.ost diagnostic for the magnitude of its value, but these data are visible in other graphs.
A B ~
0
c:i
CRA CRA .:·. ..... ·.: . 0.... lrl
MAN MAN ,
·...... .. ,: ...., ··:: . .. ,....:, cc= 0
CER CER :c:" ·.. 7 2§ c:i
~
THOR ~ THOR ...........,,.,. . . ,...>:· :·: .·,.
~
:.'/ ~
Cl)
LUM ~ LUM • ·' ·~····· ·.,·.~···C.,io;1··<';0~!"'41;:<'''""'''~··''''-- """•'?<···,-:10.!:,...,
e!.
Cl)
~
Cl::: 0
c:i
SAC SAC ........... ·. ·.:... :..,::·;.,: .:·: .·.::::: .::::;..;:,,:,::;:,·..; .. .: •.:· c..,) ~
~
PEL
RIB
SCAP
.<.
.,
PEL
RIB
SCAP
·::
.,
a ~
~
0
c:i
HUM 0 HUM 0 ~ 0
RAD !. RAD !. ~
tl) aJ'
!....::;
c:i
CARP 3'
cr
CARP · .. :·:::.:'''' 3'
cr
s::\U
MCM MCM ,. . . . . .,. . . . .,, . . :,:·:·:? ,,,....:,:,:.:.:.,,:: 0
........ ..
~ 2§ c:i
FEM :J: FEM ::1: \U
........
~
TIB s· TIB 5'
~
TAR 9: TAR !:':':·
e: ........ &:
0
c:i
MTM
3' MTM , .....................
3' ~ \5
cr cr ~
\U
........
0/100% 50 100/0% 0/100% . 50 100/0% Q
0
K1 KX8E·ORYXG2 TS KH-ORYXG3 \U c:i
~ ~
Cf)
Butchery Site Camp Site Butchery Site Camp Site ~ 0
Q ~
:::r::
c:i
~
c D ~
Cl)
~
0
c:i
CRA ::.< >·.............. CRA ......> ......... ,. :·,.::,:..·: ...... . .... :,: \U V)
MAN MAN C.)

~ 0
CER 1 ......... ,.. :::::>:,:'. CER Q ~
Cl::: c:i
THOR THOR ·:--.
~ ~
LUM
SAC
~ ..
:::; ·:· ·:· : ,;· •.· ..;. '•..... ··,..,c.·r.<;'<>:·,<'l"<>i~>..ft:.
~ LUM
SAC ..· ., .. '';fo1:···~~-1;o;:;;·~·~·'<:O..!
~ ~ ~
Cij
0....
0
0
PEL , ·:.,· ,..::: ....:;

.... ......
PEL
~
.. .............
0
RIB RIB / tl)
1......1
c:i
SCAP
HUM
··:·:':,;:.·:.....,, ,.., ....
.,
0
SCAP
HUM
I .,
0
\U
;:s
........
~
0
RAD !. RAD ..,. :··-=: :.,.;,: '··'· .... ·: ,:.: ·::,.... ~
~ ~ c:i
0
..--
CARP 3' CARP 3' -..J
cr cr
MCM ....... ··.·.: .... MCM ~ Cl::: 0
c:i
~
~
FEM :I: FEM :X:
TIB .. , ...:::.. .. :;.:;..: s· TIB s·
•
·
TAR 9: !2:: 0 .
3' TAR 3' ~

Cl:::
l::j c:i
MTM ,,
...., ......................... ;-::;
....
cr MTM cr
I
0/100% 50 100/0% 0/100% 50 100/0% r--i <: 0

c:i
KM X2·0RYXG2 HZ X2·0RYXG3 •
t--.. ~
Butchery Site Camp Site Butchery Site Camp Site C1J 0
1'"'"'4
~
~
e5 c:i
Figure 7-2. Four bar graphs illustrating complementary kill/butchery ~

site and campsite bone assemblages. Each graph (A to D) represents a ~
~ rtl
~ "fj
single gemsbok. Dark bars represent bones recovered at butchery site; -.'3
-ir
rtl
3:
light bars represent bones transported to campsite. Plaid areas of bars } 0
represent skeletal elements represented in both locations because of -.'3
fracturing. Gaps reflect missing bones (see text).
c;· ~;::;. tJ..,
m~···~ I '·'
II\ I
: I
;::; c:r' ' - « 3 .·. . 6 -:::
3 ... 1··•. ·'· ~·
g
I
en :. Q) 1·. 5: •;. ·'
i>::l
::;::: ._, ?\"
-...; ...... ('~) ..... "0 ·• ::0 ...
..
1•·.· "0
I
I
1 .. I ·;· I
~·?·.·g
•
~ ·.•· . g
g" ~
(") "'1::3 :::::: ... I u I ji
r:,en~~~ It I
,
<D •. ..a. <D .•• I I· i•: I ·:·~ I I.

B
.... o
~«('~)
._, en ::;:::
...... :···
:..
(o)
0
0 !
....
(o) ..a.
0
0
I
I'
I•
I I
I !".'
I\ I I·'
g;&-cn~g. . C; (:)
1:'••
en n. qwnaJo:l
~ QW!IPU!H QWUPU!H QW!IaJO:I
i>::l ....... IBPCV ~ IB!XV
~~~i>::l~
0
Table 7-2. %MAU Values for Major Gemsbok Skeletal Elements by Carcass
Camp/Carcass CRA MAN CER THOR LUM SAC PEL RIB SCP HUM RAD CARP MCM FEM TIB TAR MTM PHA
:;too/watsia 0.0 0.0 0.0 0.0 100.0 0.0
0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 50.0
Hitzonu 0.0 0.0 . 85.7 100.0 100.0 100.0
50.0 69.2 100.0 100.0 50.0 41.7 100.0 50.0 50.0 50.0 50.0 12.5
KeArn 0.0 0.0 0.0 46;2 100.0 0.0
0.0 38.5 0.0 0.0 0.0 0.0 50.0 0.0 0.0 10.0 50.0 0.0
Koka//a/u I 100.0 100.0 100.0 92.3 100.0 100.0 0.0 30.8 100.0 100.0 50.0 0.0 100.0 100.0 100.0 40.0 100.0 75.0
Koka//a/u II 100.0 100.0 85.7 92.3 100.0 o.o o.o 30.8 100.0 o.o 5o.o ·o.o 100.0 o.o 5o.o o.o 1oo.o 91.7
Minipara 100.0 100.0 28.6 0.0 33.3 0.0 50.0 26.9 100.0 100.0 50.0 0.0 50.0 100.0 100.0 0.0 100.0 8.3
Senyabe 100.0 100.0 28.6 46.2 100.0 100.0 o.o· 26.9 100.0 o.o 100.0 0.0 0.0 0.0 50.0 50.0 100.0 41.7
Doatara 100.0 100.0 85.7 100.0 100.0 100.0 100.0 53.8 100.0 100.0 100.0 50.0 100.0 100.0 100.0 0.0 100.0 8.3
Taela 100.0 100.0 0.0 0.0 0.0 0.0 50.0 19.2 50.0 100.0 50.0 0.0 50.0 . 50.0 50.0 10.0 50.0 0.0
X2-0RYXG2 100.0 100.0 28.6 38.5 83.3 0.0 50.0 30.8 100.0 100.0 100.0 0.0 50.0 100.0 100.0 0.0 50.0 0.0
X2-0RYXG3 100.0 100.0 0.0 0.0 0.0 100.0 50.0 0.0 0.0 0.0 100.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
KH-ORYXG2 100.0 0.0 71.4 46.2 66.7 100.0 100.0 65.4 0.0 100.0 100.0 91.7 0.0 100.0 100.0 80.0 0.0 0.0
KH-ORYXG3 100.0 100.0 28.6 92.3 100.0 100.0 100.0 84.6 100.0 100.0 100.0 91.7 100.0 100.0 100.0 50.0 100.0 33.3
EB-ORYXG1 0.0 0.0 0.0 0.0 16.7 100.0 50.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 en
?\
KX8E-ORYXG1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 100.0 0.0 41.7 100.0 100.0 0.0 0.0 0.0 0.0 m..
ro
.......
KX8E-ORYXG2 0.0 0.0 0.0 0.0 0.0 100.0 100.0 0.0 0.0 50.0 0.0 100.0 0.0 0.0 0.0 80.0 0.0 0.0 ~
""0
Kill Average 66.7 66.7 46.0 53.0 81.5 44.4 27.8 32.9 72.2 55.6 50.0 10.2 61.1 44.4 55.6 17.8 72.2 31.9
~
Camp Average 57.1 42.9 18.4 25.3 38.1 71.4 64.3 25.8 28.6 64.3 57.1 46.4 35.7 57.1 42.9 30.0 21.4 4.8 ""0
'"i
~
.........
Note: CRA =cranium/maxillae, MAN= hemimandibles, CER =cervical vertebrae, THOR= thoracic vertebrae, LUM =lumbar vertebrae, SAC= sacra, PEL= innominates, ro
fJ)
RIB= ribs, SCP =scapulae, HUM= humeri, RAD = radioulnae, CARP= carpals, MCM =main metacarpals, FEM =femora, TIB =tibiae, MTM =main metatarsals, PHA = .......
phalanges. ~
126j L. E. Bartram, Jr.
SFUI Values for Gemsbpk

To measure part utility, I used Metcalfe and Jones's Standardized Table 7-3. G
Whole Bone Food Utility Index (SFUI)'values (Metc;alfe and Jones 1988). This
was frustrating because I was forced to use Binford's caribou once again as a DeathfPrimary Seaver,
Butchery Site
proxy for the values I would much rather have had for gemsbok. A combina-
tion of factors conspired to keep me from obtaining gemsbok food utility Koka//a/u II
indices, not the least of which was the justifiable desire of the Kua to let the KeArn
feasting begin. It may be possible to construct a composite gemsbok utility Senyabe :;eoo
index from existing data, but for the time being, I must defend my usage by Hitzonu N
resorting to familiar arguments about the common anatomical features of Senyabe
artiodactyls, be they caribou or gemsbok! It would be surprising indeed if
Koka//a/u I
there were not highly significant rank order correlations between the food
values for gemsbok and caribou parts. , :;eoo/watsia
I now turn to a discussion of the results of the analyses. Mini para Hye
Taela Hye
Results Doatara
ay =evidence for sea•
Intercamp and Intracamp Comparisons of Part Abundance bproduct of previous
cEntire suppurating h
As a first step in the analysis, scatterplot and correlation matrices revealed no evidenc
of part abundances for behaviorally similar assemblages were inspected for
patterning. I expected to see a number of positive linear relationships, illus-
trating indirectly their similar formation histories. Surprisingly, only a few
weak relationships, expressed as low rank-order correlation coefficients were
sized carcass, t:
apparent between these assemblages. to a specific ca
The lack of correlation between examples of the same type of site can also ments by scave:
be visualized with another kind of graph. Because the destination of each
Three additic
carcass was known, it was possible to watch the effects of butchery and trans- 7-2 (B to D). Ea
port on the part composition of the assemblages by summing the 1\1{NEs for
recovered. One
each pair of assemblages. Figure 7-2 illustrates this with the bones from four
functionally id~
carcasses recovered at both the kill/primary butchery site and the residential 2B, C, and D re]
base camp to which part of the carcass was transported. the same elem
The four charts each represent the bones of a single animal. It may be best
assemblages as
understood as two bar graphs facing one another; this is reflected in the followed by pa:
reversible scale at the bottom. The vertical axis lists major skeletal elements, of cross mends
grouped by skeletal region (i.e., axial, forelimb, and hindlimb), as indicated by
for intersite cor
the labels to the right of the chart. The length of a dark bar originating from
ed. Axial elemE
the left of the chart is proportional to the abundances of those elements recov-
because they a
ered at the kill/butchery site. The bars coming from the right side of the chart
conduct limb b1
represent the percentage of bones from the same animal transported and
until cracked f<
discarded at the camp site. In the example from Koka/ I a/u I, complete
epiphyses, bro1
representation of crania, mandibles, cervical, thoracic, and lumbar vertebrae is
shown by a dark bar stretching all the way across the graph. On the other Part
hand, the light bars for sacrum, half-pelvis, and carpals indicate that the ToE
complete inventory for these parts was recovered at the campsite. Gaps in the was explained
bar represent loss in transport (see note "c," Table 7-3), or in the case of camp each assemblag
assemblages where some axial elements came from more than one similarly Utility Index.
s Standardized Table 7-3. Gemsbok Kill/Butchery Sites-Behavioral and Completeness Data

mes 1988). This
once again as a Death/Primary Scavenged?a Processing No. of Person- %Carcass Distance Carriers Base Camp
Butchery Site Time (hrs) Butchers Hoursb Completeness (km) (Carcass No.)
Jk. A combina-
ok food utility· Koka//a/u II N 6:30 3 19.50 59.72% 3.0 KH (ORYXG2)
~ Kua to let the KeArn N 1:22 2 2.75 33.33% 4.2 X2 (ORYXG2)
;emsbok utility Senyabe~oo N 2:00 4.00 5.00% 5.0 4 KH (ORYXG4)
d my usage by Hitzonu NC 4:49 24.10 75.00% 5.5 7 X2 (ORYXG3)
.cal features of Senyabe N 5:00 20.00 43.06% 5.9 4 KX8E (ORYXG1)
ising indeed if
Koka//a/u I N 9:00 27.00 68.06% 6.0 3 KX8E (ORYXG2)
:ween the food
~oo/watsia y 4:00 2 8.00 1.39% 8.5 4 KH (ORYXG3)
Mini para Hyenas 5:00 4 20.00 37.50% 10.0 MP/EB/KX-8E
Taela Hyenas 12:00 12.00 23.61% ?1
Ooatara 80.56% ?1
ay =evidence for scavenging, most likely by brown hyenas, N = unscavenged,? =scavenging uncertain.
ndance bproduct of previous two columns.
'Entire suppurating hindlimb intentionally abandoned at kill site (see Figure 7-20). Later visit to site
~lation matrices revealed 70
evidence of the bones; removed by scavengers.
e inspected for
ionships, illus-
~ly, only a few
>efficients were
sized carcass, they reflect my inability to relate some specimens conclusively
to a specific carcass. In four cases, complete destruction or removal of ele-
of site can also
ments by scavenging carnivores may be the cause of the gap .
.nation of each
Three additional complementary distributions of bones are shown in Figure
1ery and trans-
7-2 (B to D). Each graph represents a single carcass and where its bones were
; the NINEs for
recovered. Once again, the charts illustrate the extreme variability between
ones from four
functionally identical kinds of site assemblages. The plaid pattern in Figure 7-
the residential
2B, 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
It may be best
assemblages as a result of the fragmentation of these parts during butchery
eflected in the
followed by partial transport of the fragments. In fact, I have found a number
~letal elements,
of cross mends between kill sites and campsites. The archaeological potential
as indicated by
for intersite conjoining, although daunting, is thus convincingly demonstrat-
·iginating from
ed. Axial elements apparently present the likeliest candidates for conjoining
~lements recov-
because they are generally fragmented during primary butchery. The Kua
ide of the chart
conduct limb butchery at the joints, and limb elements normally remain whole
ansported and
until cracked for marrow. Occasional exceptions include incompletely fused
lu I, complete
epiphyses, broken off in joint-directed chopping.
Jar vertebrae is
. On the other Part Abundance and Food Utility
:licate that the To examine the extent to which variability in element frequencies
ite. Gaps in the was explained by nutritional utility, I next plotted the abundance values for
.e case of camp each assemblage against Metcalfe and Jones's Whole Bone Standardized Food
1 one similarly
Utility Index.
1281 L. E. Bartram,]~.
It is here that the influence of food utility considerations would be expected B

to emerge. Kill sites would be expected to appear as reverse utility curves, 90
while transported camp site assemblages would have positive slopes. How- 80
ever, no compelling relationships could be discerned in these scatterplots. 70 CRA RAI
Based on the reasoning thatsample size problems at a si~gle carcass site • •
60
may be misrepresenting central tendencies that would appear in a larger sam- CARP
•
ple, plots were constructed for the average element abundances from both kill ~50 eMAN
~ MCM
sites and their complementary base camp assemblages (Figure 7-3). Once ~ 40 • TA
c
again, no convincing relationships were apparent; ev~n the "average" assem- 30
THOR
•
blages of each type seem equally amorphous. 20
c
Behavioral Data 10 PHA

•
Data of a more behavioral nature (Table 7-3) help. to explain the 0
0 0 0
,....
observed variability in the gemsbok bone assemblages. The carcass complete- C\1
ness statistic described earlier was plotted against the number of butchers, Standa1
number of carriers, distance to camp, and processing time of each carcass. No
significant relationships were evident between any of these variables and Figu
carcass completeness. Food
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 Why do the a:
highly significant correlation (p = .005) was observed. As one would expect, that the camps t
the longer that processing activities were conducted at a kill site, the more an assumption
bones were abandoned there. The fact that this strong correlation exists in the explain the obs<:
absence of compelling nutritional relationships strongly suggests that factors principal assum
other than nutritional utility determine the composition of Kua bone assem- toric faunal asse
blages. ported or aband·
at a residential ~
ported with tha
/ 1988:487).
Discussion
This assumpt1
What Explains Kua Assemblage Composition? Because in man·
and transported
If nutritional utility is so effective in explaining the composition of upon which the
Nunamuit assemblages and apparently so ineffective in the Kua case, what anticipated utili
accounts for the difference? What we see ope
It cannot be merely the use of inappropriate utility indices. Despite the one without the
reasonable expectation that functionally identical sites would exhibit positive at the primary b1
linear relationships and high correlation coefficients, inspection of the scatter- The lack of as:
plot matrix generated for all pairs of kill/butchery sites revealed virtually no stands in sharp ,
systematic correlations in part frequencies. It is extremely unlikely that this carcass completE
lack of similarity among assemblages of the same type would be suddenly would be discar'
rescued by the refinements conferred by the use of utility data derived from to available ·proc
gemsbok. Both the overall lack of correlation between the very same types of turns to, What c
camps and the widely variable abundance rankings between assemblages interpretation of
implies that factors other than nutritional utility are conditioning the content variable situatim
of the assemblages. available carrier:
.tld be expected Base Camps Kill Sites

90 90~------------------~
utility curves, LUM
e slopes. How- 80 HUM 80 •

:atterplots. 70 CRA RAD
• SAC FEM 70 CRA
MTM
• SCAP
•
gle carcass site • • •
PEL
MMAN
60 ?O • MCM HUM TIB

[n a larger sam-. CARP
• TIB THOR e RAD • •
~50 eMAN
• ~50 •
s from both kill ~ MCM LUM ~ • SAC
•
FEM
ure 7-3). Once ~ 40 • TAR • ~ 40

CER
SCAP
" PHA RIB
verage" assem- 30
THOR e
• CERMTM • •RIB 30 • •
e •
PEL
20 • 20 TAR
•
CARP
10 PHA 10 •
•
to explain the O+-~~~~~~~~~~
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
~ N M V ~ ~ ~ ~ m ~ ~ N M V ~ ~ ~ ~ m ~
·cass complete-
Standardized Food Utility Index Standardized Food Utility Index
er of butchers,
::t.ch carcass. No
variables and Figure 7-3. Average element abundances plotted against Standardized
Food Utility index for Kua gemsbok kill sites and base camp assemblages.
nt processing a
Jerson-hours of
Why c,lo the assemblages not exhibit the predicted relationships if we know
sure 7-4) that a
that the camps are functionally identical? Metcalfe and Jones have identified
would expect,
an assumption built into the use of utility curves that I feel does much to
site, the more
explain the observed lack of patterning in the Kua case. They note that "the
on exists in the
principal assumption underlying the use of utility indices to interpret prehis-
sts that factors
1a bone assem- toric faunal assemblages is that bone and associated meat always are trans-
ported or abandoned as a unit, and therefore, the presence of a particular bone
at a residential site typically is interpreted to mean that the body part trans-
ported with that element was transported to the site" (Metcalfe and Jones
1988:487).
This assumption is clearly not warranted for the examples described here.
B.ecause in many instances meat was stripped from the bones at the kill site
and transported as biltong, the presumed linkage between bone and meat
composition of upon which the meaningful use of these curves depends is broken. Thus, the
~uacase, what anticipated utility-based relationships are not reflected in the bone refuse.
What we see operating in the Kua examples is a '1>ulk" transport strategy, but
~s. Despite the one without the bones to evidence its operation because of biltong production
~xhibit positive at the primary butchery location.
1 of the scatter- The lack of assemblage similarity between functionally homologous camps
ed virtually no stands in sharp contrast to the highly significant correlation between kill site
likely that this carcass completeness and processing intensity. The probability a given bone
d be suddenly would be discarded at the kill site or transported seems to be strongly related
a derived from to available processing time. With this in mind, the question for archaeologists
· same types of turns to, What determines available processing time? Unfortunately for the
n assemblages interpretation of individual archaeofaunal assemblages, the answer is highly
lng the content 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
Kua Gernsbok Kill/Primary Butchery Sites Kua examples i
90
80
70
2
r :0.695
r = 0.812
5
@ Scavenged
e Unscavenged
J 8
HZ
@
/
KO
We have see:
importance, bu
fact that the Kt:
seems that the
of the Kalahari
the only real f
value of each c;
, K2 whole carcass
60 8
played an.impc
50 ranking of spe
SY lowest-ranked
8
40 ©l MP
threshold."
KM For the Kua,
8
30 carcass someti1
©l TL filleting and su
20 the limb bones,
elements; they
10 breaking them
SG the meat is dry
0~----~--~--~----~------~------~----~ carrying the ca
0 5 10 15 20 25 30 enticement to 1
gemsbok).
Thus, kill sit1
Processing Time (Person-Hours)
the extra proce
limitations of tl
Figure 7-4. Relationship of skeletal completeness index to kill site proces-
occasionally im
sing time (in person-hours). in order not to
now add the be
ering transpor1
motivation born of hunger and the requirements of social obligations, influ-
geographically
ence the decisions made at specific sites as well.
skeletal part fre
Transport Costs Arc)
In a recent paper on kangaroo bone transport, O'Connell and WhE
Marshall (1989) listed three factors that determine the selection of body parts them) expresse1
for transport: (1) the weight or nutriti9nal value of edible tissue attached to should we expE
various body parts; (2) the kill site processing costs; and (3) the relative bene- shown that the
fits of consumption at the kill site versus transport and consumption at the reflect transpor
residential base. but reflect inst
There is support for the suggestion that these same factors operate in other influence that n
areas as well. Indeed, Binford (1978:90) identified "transport difticulties, time The fact tha1
limitations on labor activities, and location of the kill site relative to residential .hunters pursue
or consumer locations" as factors that may influence which parts of a carcass kinds of surplu
are chosen for transport, even though the desire to maximize nutritional be obtained wi
benefits in transport was the overarching concern among the Nunamuit. The transport strate
conditioning as.
r Sites 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
HZ 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
25 30 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
to kill site proces- 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 low-
ering transport costs by reducing weight. This option is obviously variable
Mgations, influ- geographically, but it has important implications for the interpretation of
skeletal part frequencies from archaeological sites.
Archaeological Implications
0' Connell and Where else should we expect to see the kinds of patterns (or lack of
'n of body parts them) expressed in the Kua data? In which archaeological subsistence contexts
;sue attached to should we expect utility analyses to come up short for similar reasons? I have
Le relative bene- shown that the interassemblage differences in skeletal part frequencies do not
;umption at the reflect transport strategies based upon some ranked scale of carcass part value
but reflect instead a variety of situational constraints that overwhelm any
operate in other influence that nutritional ranking may have otherwise exerted.
iifticulties, time The fact that utility indices are at their best as explanatory tools where
ve to residential hunters pursue large herds of gregarious prey is probably no accident. The
arts of a ca,rcass kinds of surpluses available at mass kills, or in contexts where carcasses may
1ize nutritional be obtained with almost predictable regularity, lend themselves to selective
Nunamuit. The transport strategies, and thus, the appearance of part utility as a key factor
conditioning assemblage composition.
l
1321 L. E. Bartram~Jr.
By contrast, these conditions qearly do. not obtain .in the modern Kalahari logical inferenc
environment, where transport costs are reduced by increasing processing ethnoarchaeolo:
costs at the kill site. The Kua example~ discussed here all involve low popula- of kill site meat
tion densities of both hunters and prey; infrequent, solitary, medium-to-large well as a searcl
animals; limited transport capabilities; and semiarid enviro~ents with high clearly warrant
insolation. meat drying or
It is precisely in these kinds of contexts that we should expect to find similar think in genera:
strategies pursued prehistorically. Interestingly, it is from broadly similar to see utility in'
paleoenvironmental contexts that much of the best ~vidence for hominid sub- the con texts in
sistence behavior is derived. This issue demands additional ethnoarchaeologi- applied researc
cal consideration and much wider attention froni. archaeologists working in recognizi11g me
areas where meat drying is a potentially effective way of reducing transport record.
costs.
Importantly, there are no technological restrictions of this effective way of
reducing transport costs, even with Oldowan assemblages. All that is required Cor
for biltong production with lithic-based technology is sharp-edged flakes to SevE
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 1.lr1
enormous antiquity; the benefits of meat drying are simply too great to be tran
ignored. sets
The implications of this simple and dramatically effective cost-saving strat- byr
egy for current debates on the meaning of skeletal part frequencies are far- logi·
reaching. Until we are able to demonstrate conclusively in a given archaeolog- rom
ical context that defleshing and meat drying were not taking place and that ofn
2.0
meat and bone were indeed traveling together, we are on increasingly thin ice
botr
regarding statements about the meaning of assemblage composition and car- tran
cass part transport. inso
What is the archaeological"signature" of kill site defleshing and meat dry- 3.C
ing? As I have shown, the Kua provide quantitative support for t;he notion not'
that processing time and kill site carcass completeness will exhibit a signif- 4. T
icant positive relationship. This is not, however, what most archaeologists Old,
wish for when they think of an archaeological signature! The answer must 5. Y\
almost certainly include other lines of evidence. For example, if large ratic
assemblage size and high artifact or feature density suggest a residential site, rate:
relatively complete skeletal inventories may be reflecting low processing 6.A
times at kill sites. This, in turn, may be a function of relatively large pre- tal. 1
historic group sizes to provide transport, high prey densities, or other factors othe
and
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
Ac1
carcasses at the death sites of animals or may be suggesting the lack of a need ·I am
to do so. The latter case would imply relatively effective carcass acquisition .the From Bone~
strategies. In any case, further consideration is called for. anonymous re\
It is clear from the Kua ex~mple that skeletal part frequencies may be a poor Wisconsin Kala
reflection· of the whole nutritional story. Skeletal part frequencies must be National Scienc
considered an important, but likely incomplete, line of evidence in archaeo- Fulbright Progr
Skeletal Part Profiles j133
10dern Kalahari logical inference about subsistenc~ behavior and site function. Additional
5ing processing ethnoarchaeological consideration of the factors that determine the likelihood
lve low popula- of kill site meat drying in particular kinds of paleoenvironmental contexts, as
:ledi urn -to-large well as a search for unambiguous, "diagnostic" evidence of meat drying, is
nents with high 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
:t to find simila~ think in general terms about the types of contexts in which we should expect
broad! y similar to see utility indices influencing transport decisions. If we better understand
or hominid sub- the contexts in which food utility plays a role, we face smaller risks of mis-
hnoarchaeologi- applied research effort and erroneous conclusions, and a better chance of
~ists working in recognizing meat drying as a viable processing strategy in the archaeological
.ucing transport record .
effective way of
. that is required
Conclusions
·edged flakes to Several conclusions may be reached from the above discussion:
·ative climate. It
d practice with 1. In tropical and subtropical Africa, the sun is a potent tool in reducing
too great to be transport costs. Here, and in similar regions, drying meat more than off-
sets the additional processing costs associated with biltong producing
ost-saving strat- by reducing transport costs. This issue invites additional ethnoarchaeo-
uencies are far- logical and archaeological consideration, especially where paleoenvi-
ven archaeolog- ronmental data suggest that meat drying was a potentially effective way
; place and that of reducing transport costs.
2. Conditions that favor this strategy include low population densities of
~asingl y thin ice
both hunters and prey, infrequent and single carcass availability, limited
•osition and car- transport capabilities, and arid or semiarid environments with high
insolation.
g and meat dry- 3. Current interpretations of variability in skeletal part frequency have
t for the notion not considered the effects of defleshing and drying.
exhibit a signif- 4. The technology required for. efficient meat drying is present in
t archaeologists Oldowan assemblages.
lle answer must 5. Where bone assemblages are likely to have been scavenged, incorpo-
ample, if large ration of limb ·shaft fragments into MNE estimates is critical to accu-
residential site, rately represent the contents of the assemblage.
low processing 6. Archaeologists should not attempt to identify site function from skele-
i vel y large pre- tal part frequencies alone but should identify it in. conjunction with
or other factors other evidence, such as number of hearths, site size, and assemblage size
and diversity.
~s judged to be
l part patterning
1 inventories for
Acknowledgments
mce to process
Le lack of a need I am grateful to Jean Hudson for the opportunity to participate in
cass acquisition 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
'Smay be a'poor Wisconsin Kalahari Project was made possible by the financial support of the
tencies must be National Science Foundation and the Institute for International Education's
~nee in archaeo- 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 Olduvai Gm
Robert K. Hitchcock f0r introducing us to thefield area. In Gaborone, Alec 15:673-690.
Campbell, James Denbow, Angier Pe~vey, and Ruby Apsler provided much Bunn, H. T., and I
needed advice and assistance. Special gratitude is extended to Keith aRd Irene n.d. Dutara:
Whitelock and Debswana Mining Pty., Ltd. (Orapa), for e;ndless logistical authors.
support and extraordinary kindness. I thank Iron Gabatswane and Bunn, H. T., L. E.
1988 Variabi
Keikantsemang "Shakes" Tshikhinya for translations between English and
Scavenging,,
Kua and project members Lisanne Bartram, Henry Bunn, Christian Gurney, 457.
Marty Jakobs, and ·Ellen Kroll for their conscientio~s and enthusiastic assis- Bunn, H. T., J. W.
tance in the field. I especially thank Henry Bunn for countless discussions and K. Behrensm
for his patient supervision of my graduate research. Finally, I thank the Kua 1980 FxJj 50:
for their tolerance, good humor, and friendship. 12:109-136.
Bunn, H. T., and I
1986 System;:
References Tanzania. Cu
1988 Reply tc
Bartram, L. E., Jr, . and Interpret
1993 An Ethnoarchaeological Analysis of Kua San (Botswana) Bone Food Chase, P. G.
Refuse. Unpublished Ph.D. dissertation, Department of Anthropology, Univer- 1985 On the·
sity of Wisconsin-Madison, in preparation. Sites. P. A. C.
Bartram, L. E., and H. T. Bunn Elphick, R.
n.d. An Ethnoarchaeological Perspective on the Klasies Pattern, Limb Bone 1977 Khoikhoi
Fragmentation, and Zooarchaeological Analyses. Ms. in possession of authors. Emerson, A. M.
Bartram, L. E., E. M. Kroll, and H. T. Bunn 1990a Archaeol
1991 Variability in Camp Structure and Bone Food Refuse Patterning at Kua San bison. Unpul
Camps. In The Interpretation of Archaeological Spatial Patterning, edited by E. M. ton State Uni
Kroll and T. D. Price, pp. 77-148. Plenum Press, New York. 199Gb Carcass
Binford, L. R. ed at the Sixt
1978 Nunamiut Ethnoarchaeology. Academic Press, New York. Institution, V\
1981 Bones: Ancient Men and Modern Myths. Academic Press, New Yorky Frison, G. C.
1984 Faunal Remains from Klasies River Mouth. Academic Press, New Yqrk. 1974 The Cas1
Binford, L. R., and S. R. Binford · New York.
1966 A Preliminary Analysis of Functional Variability in the Mousterian of Gifford, D.P., and
Levallois Facies. American Anthropologist 68:238-295. 1977 A Camp
Blumenschine, R. J. Antiquity 42:2
1988 An Experimental Model of the Timing of Hominid and Carnivore Influence Grayson, D. K.
on Archaeological Bone Assemblages. Journal of Archaeological Science 15:483- 1988 Danger '
502. Anthropolog~
Blumenschine, R., and T. Caro York
1986 Unit Flesh Weights of Some East African Bovids. Journal of African Ecology 1989 Bone Tn
24:273-286. Archaeological
Borrero, L.A. Hitchcock, R. K.
1990 Fuego-Patagonian Bone Assemblages and the Problem of Communal 1978 Kalahari
Guanaco Hunting. In Hunters of the Recent Past, edited by L. B. Davis and B. 0. Agriculturalist
K. Reeves, pp. 373-399. Unwin Hyman, London. of Local Gove
Bunn,H.T. . Jones, K. T., and D
1982 Meat-Eating and Human Evolution: Studies on the Diet and Subsistence Practices 1988 Bare Bor
of Plio-Pleistocene Hominids in, East Africa. Unpublished Ph.D. dissertation, Archaeological
Department of Anthropology, University of California, Berkeley. Klein, R. G.
1986 Patterns of Skeletal Representation and Hominid Subsistence Activities at 1975 Paleoant:
~al District, and Olduvai Gorge, Tanzania, and Koobi Fora, Kenya. Journal of Human Evolution
::;aborone, Alec 15:673-690.
provided much Bunn, H. T., and L. E. Bartram
Keith and Irene n.d. Dutara: An Eland Kill Site in the Kalahari, Botswana. Ms. in possession of
authors .
.dless logistical
Bunn, H. T., L. E. Bartram, and E. M. Kroll
batswane and 1988 Variability in Bone Assemblage Formation from Hadza Hunting,,
~n English and
Scavenging, and Carcass Processing. Journal of Anthropological Archaeology 7:412-
ristian Gurney, 457.
:husiastic assis- Bunn, H. T., J. W. K Harris, G. Isaac, Z. Kaufulu, E. Kroll, K. Schick, N. Toth, and A.
:iiscussions and K. Behrensmeyer
: thank the Kua 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.
:ma) Bone Food Chase, P. G.
)pology, Univer- 1985 On the Use of Binford's Utility Indices in the Analysis of Archaeological
Sites. I?. A. C. T. 11:287-302.
Elphick, R/
:ern, Limb Bone 1977 Khoikhoi and the Founding of White South Africa. Ravan Press, Johannesburg.
ion of authors. Emerson, A. M.
1990a Archaeological Implications of Variability in the Economic Anatomy of Bison
rning at Kua San bison. Unpublished Ph.D. dissertation, Department of Anthropology, Washing-
edited by E. M. ton State University, Pullman.
1990b Carcass Product Yields ii1 Bison Bison and Hunter Selection. Paper present-
ed at the Sixth International Council for Archaeozoology Meeting. Smithsonian
Institution, Washington, D. C.
N York. Frison, G. C.
few York. 1974 The Casper Site: A Hell Gap Bison Kill on the High Plains. Academic Press,
New York.
e Mousterian of Gifford, D. P., and D. C. Crader
1977 A Computer Coding System for Archaeological Faunal Remains. American
Antiquity 42:225-238.
nivore Influence Grayson, D. K.
I Science 15:483- 1988 Danger Cave, Last Supper Cave, and Hanging Rock Shelter: The Faunas.
Anthropological Papers 66, pt. 1. American Museum of Natural History, New
York.
i African Ecology 1989 Bone Transport, Bone Destruction, and Reverse Utility Curves. Journal of
Archaeological Science 16:643-652.·
Hitchcock, R. K.
1of Communal 1978 Kalahari Cattle Posts: A Regional Study of Hunter-Gathers, Pastoralists, and
Davis and B. 0. 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
bsistence Practices 1988 Bare Bones Archaeology: Bone Marrow Indices and Efficiency. Journal of
D. dissertation, Archaeological Science 15:415-423.
Klein, R. G.
~nee Activities at 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 Tobias, P. V. (edit
Research 5:275-288. · 1978 The Bu:
Landals, A. Rousseau, C:
1990 The Maple Leaf Site: Implications of the Analysis of Small-scale Bi$on Kills. Wheat, J. B.
In Hunters of the Recent Past, edited by L. B. Davis and B. 0. K. Reeves, pp. 122- 1972 The Ols
151. Unwin Hyman, London. Society for A
Lartet, E., and H. Christy (editors) White, T. E.
1865-1875 Reliquae Acquitanicae: Being Contributions to the Archaeology and 1954 Observ;
Paleontology of Perigord and Adjoining Provinces of Southern France. Williams and Nos. 3, 4, 5, c
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 Archaeo-
logical 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. Anthro-
pological Papers 59. American Museum of Natural History, New York.
Skeletal Part Profiles j137
frica. Quaternary Tobias, P. V. (editor)

1978 The Bushmen: San Hunters and Herders of Southern Africa. Human and
Rousseau, Cape Town.
scale Bison Kills. Wheat, J. B.
Reeves, pp. 122- 1972 The Olsen Chubbuck Site: A Paleo Indian Bison Kill. SAA Memoirs No. 26.
Society for American Archaeology, Washington, D.C.
White, T. E.
Archaeology and 1954 Observations on the Butchering Technique of Some Aboriginal Peoples,
ce. Williams and Nos. 3, 4, 5, and 6. American Antiquity 19:254-264.
:iety. Cambridge
Jurnal of Archaeo-
lasses. 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. Anthro-
wYork.
Index (FUJ
scalogram
Hadza hu
"schlepp e
rather thar
the hunte:
8. The Role of Body Part Utility in between SJ
of axial un
Small-scale Hunti~g under Two cited by OJ
Strategies of Carcass Recovery interpretat
abundance
also leads
Alice M. Emerson greatly infl
transport a
Analyse~
Abstract: Data provided by O'Connell and colleagues (1990) on the although <
· transport of body parts to base camps. are examined for Hadza hunting appendicu
of large game animals under conditions of small-scale kills. Axial and
the "schleJ
appendicular unit transport is considered with respect to different mea-
caloric yiel
sures of bison body part utility and discussed in light of the expectations
of th~ "schlepp effect" (Perkins and Daly 1968; White 1952). The greater and marrO'
appendicular utility suggested by that model is found·to apply only ity does pt
when skeletal fat or marrow is the focus of utility assessments. The vari- they are m·
ability in carcass unit transport to base camps by Hadza hunters is exam- sit food ne1
ined with respect to two strategies of carcass transport-a Maximum based upo1
Carcass Recovery strategy, which over time yields assemblages domi- and the otl
nated by axial elements, and a Limited Carcass Recovery strategy, which Expectatio:
yields assemblages dominated by appendicular and some axial units. ported assE
Transport and processing costs are evaluated with respect to body part dicular pal
utility and suggested to be important factors influencing elimination of · small-scale
elements from the transport assemblage to meet in-transit food needs. The rem
Versus Ap]
the expectc
Introduction body part
Recent observations by Bunn and colleagues (1988) and O'Connell section, Ha
and associates (1988, 1990) of modern Hadza hunter-gatherers suggest that ed by O'C
small-scale hunting of large game (the taking of one or two individual ani- Strategies
mals) permits greater flexibility in carcass transport practices than is possible ences in thE
with large-scale kills. This flexibility is reflected in what first appears to be high utili!)
inconsistent relationships between the transport of body parts and models of Modificati<
general body part utility, such as the Modified General Utility Index (MGUI) nated from
as developed by Binford (1978) and simplified by Metcalfe and Jones (1988),
yielding various standardized or unstandardized versions of the Food Utility
Paper No. 21. © 1993 by the Board of Trustees, Southern illinois University. All rights
reserved. ISBN 0-88104-076-2. earlier by 'Y
138
Strategies of Carcass Recovery 1139
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
ltyin 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
·Two cited by O'Connell and colleagues (1990) as reason to question archaeological
rery 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),
es (1990) on the although clearly not an African species, suggest that the expectation that
)r Hadza hunting appendicular elements are of greater value than axial units, as suggested by
e kills. Axial and the "schlepp effect," is unfounded unless the evaluation is based upon the
to different mea-
caloric yield of skeletal fat (determined from the fat content of bone grease
f the expectations
l952). The greater
and marrow) or marrow alone. The analyses also suggest that body part util-
nd to apply only ity does~play an important role in transport decisions at small-scale kills, but
:sments. The vari- they are mediated by evaluations of transport costs and immediate or in-tran-
l hunters is exam- sit food needs. Two possible strategies of carcass recovery are recognized: one
art-a Maximum based upon an initial expectation of complete carcass recovery and transport
:semblages domi- and the other based upon limited recovery and transport of high utility parts.
':'1 strategy, which Expectations of how carcass units are subsequently eliminated from trans-
some axial units. ported assemblages through use or abandonment clarify why axial or appen-
pect to body part dicular parts might dominateassemblages transported to base camps from
ng elimination of small-scale kills.
sit food needs. 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
) and O'Connell section, Hadza Transport Practices, is a brief review of transport data present-
ers suggest that ed by O'Connell and colleagues (1990). It is followed by Carcass Transport
individual ani- Strategies and Assemblage Compositions, which focuses upon how differ-
than is possible ences in the initial decision of whether to transport the complete carcass or the
st appears to be high utility parts result in assemblages that are distinctive in composition.
:sand models of Modifications to Transport Assemblages examines how body parts are elimi-
y Index (MGUI) . nated from the initially selected transport assemblage. Conclusions follow.
nd Jones (1988),
the Food Utility
Axial Versus Appendicular Utility and the
' the Interpretqtion of ~"'Schlepp Effect"
gations, Occasional
iversity. All rights 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
CERVICA
modes of acquisition (scavenging versus hunting). White's (1952) interpreta- ..J
.o;
x.o; THO
tion of bison use reflects this expectation. He suggests that axial units of ll
hunted bison were often stripped of meat and the associated skeletal elements Cl
discarded at· the kill site, while the appendicular skeleton was transported
back to camp. He attributes this practice to differential transport costs and PR
marrow yields. White's work suggests that assessments of value may be made
PRO XI
on the utility of component products of a carcass unit rather than on general DIS
utility. a:
:5:::> PRO XII
Analyses presented here demonstrate that the relative utility of axial and u DIS
0 ANT
appendicular units varies depending upon which carcass products are in- ~.o;
cluded in the assessment. Four types of bison body part utility models
(Emerson 1990b) are evaluated with respect to their relative axial anq appen-
dicular yields: (1) two general utility models-one (Emerson 1990b) calculated PRO XI
Dl~
from the caloric yields of skeletal fat, muscle protein, intramuscular and other POST
dissectible fat (the Total Products model [S]MAVGTP) and the other calcu-
lated 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, in-
tramuscular, 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 (Emerso
grease and marrow; and (4) two Marrow Fat models ([S]MAVGMAR and of axial
[S]AVGMAR) based upon the caloric yield of marrow alone.l Carcass compo- differen1
sition data (Emerson 1990b) from four range-fed bison (yearHhg and 4-year- Like I
old males and 7-year-old and 16-1/2-year-old females) provide the data for units rel
the models. It is expected that generation of similar data for African species rider a-v
would reflect similar relationships between major carcass divisions. nouncec
Figure 8-1 compares two models of general utility for bison in which prox- limb un
imal and distal limb values are determined. Before discussing relative axial- lated. Tl
appendicular division utility differences, one must first account for the differ- whole u
ences evident in the models that are based upon data from the same animals. complet
The Total Products nlodel is based upon caloric yields of carcass products that units) ar
are standardized so that each body part value is expressed as a percentage of the valu
the highest yielding unit. The standardized values of the limb units are then for each
modified to account for the effects of riders. In essence, the rider averaging 8-1a, wl
routine (developed by Binford 1978) raises the value of most limb units to Product
account for the fact that lower-valued parts are sometimes transported be- units. Tl
cause of their location next to, or between, higher-valued units. femur,tr
The bison Food Utility Index model (also modified to account for riders) is Ifwhc
based upon Metcalfe and Jones's (1988) simplified model construction. the majc
Although body part values are based upon weight rather than caloric yield, in (S)AVG:
bison this has little effect on the comparative rankings of the elements have hi~
tances the least
tblage returned a. TOTAL PRODUCTS MODEL ((SIMAVGTPI b. FOOD UTILITY INDEX MODEL ((SIAVGFUI)
units are more 0 20 40 60 80 100 0 20 40 60 80 100

.I
ldance of these
IT_J
SKULL I I 14.2 SKULL I 25.3
=fi I I I I
ATLAS I I 6.4 ATLAS
I I I I 4.0
tggest different ·AXIS I I
I
I
I
7,8 AXIS
I
I I
I
5.1
_, CERVICAL VERTEBRAE (3-7) I' 58.8 , CERVICAL VERTEBRAE 13·71 I I 38.6
952) interpreta -. x<< THORACIC VERTEBRAE l 84,7
..J
x<< THORACIC VERTEBRAE I

I
I
I
I
47.4
t axial units of LUMBAR VERTEBRAE I 82.9 LUMBAR VERTEBRAE I I

I
I
I
45.1
SACRUM-PELVIS II I 54.7 SACRUM-PELVIS
~: I I 34.7
celetal elements CAUDAL VERTEBRAE 'I I I
I 1.5 CA.UDAL VERTEBRAE I I I I
I
0.9
RIBS 100.0 RIBS I I 62.3
ras transported STERNUM I: I 52.9 STERNUM 1: : I 32.4
~
:;port costs and
~
SCAPULA I I 31.6 SCAPULA I I 25.5
I I I I
PROXIMAL HUMERUS I 31.8 PROXIMAL HUMERUS I 27.5
I I
te may be made DISTAL HUMERUS 1
I
I
I.
25.1 DISTAL HUMERUS I
I
I
I
27.5
PROXIMAL RADIUS-ULNA I 16,5 PROXIMAL RADIUS-ULNA 1· I 19.1
:han on general DISTAL RADIUS-ULNA
I
I
I
I
I
12.1 DISTAL RADIUS-ULNA I
I
I
I 14.9
CARPALS I 6.6 CARPALS I I 8.6
a: I I I
:5 PROXIMAL METACARPAL I I
I
3.9 j PROXIMAL METACARPAL I
I I 5.4
ity of axial and ::>
u DISTAL METACARPAL I
I
I 2.6 ::>
u DISTAL METACARPAL I I
I
3.8
roducts are in-

0 ANTERIOR PHALANGES I
I
I
1.9 0 ANTERIOR PHALANGES I I 2.7
~< PROXIMAL FEMUR I I 69.4

~ PROXIMAL FEMUR 100.0
m
I
utility models DISTAL FEMUR I I 69.4 <( DISTAL FEMUR 100.0
I
PROXIMAL TIBIA I I
40.8 PROXIMAL TIBIA II I 57.7
xial and appen- I I
~i
DISTAL TIBIA I 25.5 DISTAL TIBIA I 36.6
I I
)90b) calculated
TARSALS I
I
I
I
I
13.6
7.5
TARSALS
PROXIMAL METATARSAL
I
I
II 19.4
PROXIMAL METATARSAL 10.8
I I I I
DISTAL METATARSAL I DISTAL METATARSAL I I 6.5
cular and other POSTE,RIOR PHALANGES
I
I
I
4.5
2.9 POSTERIOR PHALANGES
I I
3.9
he other calcu-
Jne weight (the
es (1988); (2) a Figure 8-1. Comparative general utility model values for bison with
of skeletal, in- proximal and distal units evaluated separately.
.testinal depots,
viA VGSKF and
Jresent in bone (Emerson 1990b). The differences evident in Figure 8-1 between the rankings
:AVGMAR and of axial and appendicular parts in the two models are attributed to other
Carcass compo- differences in model construction.
ing and 4-year- Like Binford's MGUI, FUI models exaggerate the values of appendicular
de the data for units relative to axial units beyond that which may be accounted for by the
African species rider averaging routine (Emerson 1990a). That exaggeration is most pro-
ions. nounced in MGUI and FUI models that generate separate proximal and distal
in which prox- limb unit values and is less extreme when whole element values are calcu-
~ relative axial- lated. The exaggeration occurs because the MGUI and FUI models use (1) the
lt for the differ- whole unit value to calculate the values of proximal and distal ends (e.g., the
~ same animals. complete humerus value is used for both the proximal and distal humerus
;s products that units) and (2) some grouped unit totals (e.g., the tibia plus tarsals) to calculate
a percentage of the values of smaller component units (e.g., the complete tibia value is used
J units are then for each of the proximal and distal tibia, and tarsal units). As evident in Figure
rider averaging 8-la, when appendicul~r unit values are not exaggerated, as in the Total
:;t limb units to Products model, most axial units have higher values than most appendicular
transported be- units. The (S)AVGFUI model, however, falsely exaggerates the value of the
femur,tibia, and other appendicular units relative to the axial units.
.nt for riders) is H whole unit .values are used instead of values for proxirnal and distal ends,
I construction. the major source of appendicular unit value exaggeration is eliminated in the
caloric yield, in (S)AVGFUI construction. In Figure 8:-2, both general utility models tend to
f the elements have higher axial than appendicular utility values. The major exceptions are
1421 A.M. Emerson
a. TOTAL PRODUCTS MODEL U&IMAVGTPI b. FOOD UTILITY INDEX.MODEL (ISJAVGFUII
a. PRO
0 20 40 60 80 100 0 20 40 60 80 100
~~
lI 25:3
~
SKULL I I 10.4 SKULL I
I I I · ATLAS I
I 9.1
ATLAS I 4.7 I
I I I I
AXIS I I I 5.7 AXIS
I '
I I
9.1
_, CERVICAL VERTEBRAE {3-71

' 'I I 41.4
.....
CERVICAL VERTEBRAE {3-7)
THORACIC VERTEBRAE
I
J
I
I
I
I
38.6
47.4
CERVICAL VERTEBR~
<( THORACIC VERTEBRAE J 61.7 <(
x I I THORACIC VERT
x LUMBAR VERTEBRAE J
I
60.4 LUMBAR VERTEBRAE I I I 45.1 -'
<(
LUMBAR VERT
<( I <(
I I x
~ :I
I
~~
SACRUM-PELVIS 39.8 SACRUM-PELVIS I 34.7 SACRUM-
I I <(
I ·cAUDAL VERTEBRAE I I 1.0

CAUDAL VERTEBRAE L1 I CAUDAL VERT
RIBS J : 73.3 RIBS I I 62.3
I
STERNUM II :I I 38.4 STE[INUM II i
I
32.4
STI
~~
SCAPULA I 27.5
~
SCAPULA I I 28.4 sc
I I I I
I 28.4 HUMERUS I I 27.5
HUMERUS I I I PROXIMAL HUI
I I I 19.2
RADIUS-ULNA I I 19.7 RADiUS-ULNA I DISTAL HUI
I I
a: I I a: CARPALS I I I 10.7
CARPALS I I I 10.6 PROXIMAL RADIUS
:3 METACARPAL I I I 6.0
:3::> METACARPAL
I
I
I
I
I
I
6.5 DISTAL RADIUS
::> I I I u
u ANTERIOR PHALANGES I I 3.4 ANTERIOR PHALANGES I I I 4.0
0 I 0 CA
~
FEMUR 100.0
~
FEMUR 100.0 PROXIMAL METAC
TIBIA ft 58.1 TIBIA JiI I 57.7
=?IIl
I <( DISTAL METAC,
<(
F?rl
TARSALS I I 30.0 TARSALS I 30.0 ANTERIOR PHAL,
I I I
I METATARSAL I I 16.1
METATARSAL I I 15.9 PROXIMAL I
I I I 6.8
POSTERIOR PHALANGES I 6.4 POSTERIOR PHALANGES DISTAL!
PROXIMAl
DISTAl
Figure 8-2. Comparative general utility model values for bison with limb TA
PROXIMAL MET AT,
units evaluated as whole elements. DISTAL MET AT,
POSTERIOR PHAL
the femur and tibia, which have values either higher than or comparable to Fi~
the high value axial units. It is evident, therefore, that the assumed higher unit
value of the limb elements, as expressed in the general expectations of the
"schlepp effect," is applicable only to the femur and sometimes the tibia with
respect to general utility. Furthermore, this limited greater utility assessment
is not evident in models that evaluate proximal and distal ends separately. a. MODIFI!
Figure 8-3 presents two Total Fat models where (a) proximal and distal ends
are evaluated separately and (b) whole limb unit values are used. Spetp (1983)
suggests that fat yield may be the focus of utility assessments when ~arbohy CERVICAL VERTEBR~
drates are unavailable or in low supply. If total fat utility is calculated, most _,
<(
THORACIC VERT
axial units have higher value than most appendicular units, regardless of x
<(
LUMBAR VERT
SACRUM-I
whether whole or proximal and distal unit values are determined. CAUDAL VERT
Two other models, focusing on skeletal fat or marrow alone, do provide STE
sc
assessments of body part utility in which appendicular units are more often of PROXIMAL HUI
DISTAL HUI
higher value than axial units. Two skeletal fat models based on caloric yields PROXIMAL RADIUS
of marrow and bone grease illustrate relative body part values where w.odifi- DISTAL.RADIUS
CA
cations to account for riders have (Figure 8-4a) and have not (Figure 8-4b) PROXIMAL METAC
DISTAL METAC.
been applied. The unmodified model may provide the best assessment of ANTERIOR PHALt
comparative utility when bones are evaluated individually, and the modified PROXIMAL I
DISTAL I
model is pertinent when utility is assessed with respect to intact limb trans- ~
PROXIMAl
DISTAl
port. Figure 8-5 illustrates relative marrow utility. Such an evaluation might TAl
be important where methods of processing bone grease are undeveloped or PROXIMAL MET AT,
DISTAL MET AT,
unused. POSTERIOR PHALI
Appendicular units have higher utility than most axial tinits only in models
focusing on skeletal and marrow fat. Assemblages that show the transport of
Fig~
appendicular -rather than axial skeletal elements may indicate that low utility
to ac
TOTAL FAT MODELS (IS]MAVGTF)

WGFUII
a. PROXIMAL AND DISTAL ENDS b. WHOLE BONE UNIT?
60 80 100
25.3 0 20 40 60 80 100 0 20 40 60 80 100
I
I
I
I
9.1
9.1
ATLAS
AXIS ~
I
I
I
I :
I
II 5.9
7.1
ATLAS
AXIS ~
I
I
I
I :
I
:I 5.9
7.1
I I
I
38.6 CERVICAL VERTEBRAE 13-7) I I I 50.7 .CERVICAL VERTEBRAE (3-71 I I 50.7
I 47.4 THORACIC VERTEBRAE 100.0 THORACIC VERTEBRAE 100.0
..J ..J
I <
45.1 < LUMBAR VERTEBRAE 99.5 LUMBAR VERTEBRAE 99.5
I
x x
I
I
I
34.7
1.0
< SACRUM-PEL VIS
CAUDAL VERTEBRAE :
jl
I
I
I
I
I
54.0
1.8
< SACRUM-PEL VIS
CAUDAL VERTEBRAE I I
I l
I
:I
54.0
1.8
~ I J
62.3 RIBS 93.0 RIBS 93.0
r
32.4 STERNUM 1: 58.6 STERNUM 1: 58.6
2.7.5 SCAPULA I 16.7 SCAPULA I I 30.4
I
r
I I
27.5 PROXIMAL HUMERUS 17.1 HUMERUS I I I
30.4
I I
19.2 DISTAL HUMERUS 15.3 RADIUS-ULNA I I 22.0
I I
10.7 PROXIMAL RADIUS-ULNA 11.~ CARPALS I I 13.0
I I
6.5 DISTAL RADIUS-ULNA 8.9 METACARPAL I I I 8.4
I I I
4.0 CARPALS 5.6 ANTERIOR PHALANGES I I 5.4
100.0 ~ PROXIMAL METACARPAL 3.9 ~
::>
FEMUR llI 76.7
57.7
::::J
u DISTAL METACARPAL 3.1 u TIBIA J I 48.6
30.0 i5 ANTERIOR PHALANGES 2.7 i5 TARSALS
: I I
I 26.9
~
=7:
1'---'-----1 I
~
16.1 PROXIMAL FEMUR I I 16.1
38.7 METATARSAL I I
~~
8.8 38.7 POSTERIOR PHALANGES I I
< DISTAL FEMUR < I I
I 9.2
PROXIMAL TIBIA 25.4 I I I
I I I
I I I
DISTAL TIBIA 17.0 I I
I I I
TARSALS 10.1 I I I I
1ison with limb PROXIMAL METATARSAL
I
I
I
6.4 I
I
I
I
I
I
I
I
DISTAL METATARSAL I I I
I 4.6 I
POSTERIOR PHALANGES I 3.5 : I
I
I
I I
omparable to Figure 8-3. Total fat utility models for bison with (a) proximal and distal
umed higher unit values and (b) whole limb unit values.
tations of the
the tibia with SKELETAL FAT MODELS
ty assessment
eparately. a. MODIFIED FOR RIDERS (IS]MAVGSKFI b. NOT MODIFIED FOR RIDERS IISIAVGSKFI
:1d distal ends 0 20 40 eo 80 100 0 20 40 60 so 100

.. Speth (1983) ATLAS
l I I I 1.6 ATLAS
I
I
I i I
I
1.6
AXIS I I I 1.1 AXIS I I 1.1
•hen carbohy- I
~I hi
CERVICAL VERTEBRAE (3-7) I I I 3.3 CERVICAL VERTEBRAE (3-7) I I I 3.3
I I I I I I
culated, most THORACIC VERTEBRAE
LUMBAR VERTEBRAE
I
I
I I
16.8 ..J
<
THORACIC VERTEBRAE
LUMBAR VERTEBRAE I
I
I I
16.8
I I 18.3
x I 18.3
regardless of SACRUM-PEL VIS I I
I
70.6 < SACRUM-PELVIS j I 70.6
I
~~-I ~~
I
CAUDAL VERTEBRAE I 2.9 CAUDAL VERTEBRAE 2.9
i. RIBS I
I
I 38.7 RIBS
I
I 38.7
~, do provide STERNUM I I 3.1 STERNUM l I I 3.1

SCAPULA lJ L 53.7 SCAPULA u I_ ! I 11.8
more often of PROXIMAL HUMERUS I 95.6 PROXIMAL HUMERUS ·I 95.6
DISTAL HUMERUS II 77.2 DISTAL HUMERUS I : 58.7
caloric yields PROXIMAL RADIUS-ULNA I I 67.4 PROXIMAL RADIUS-ULNA li I 57.7
I
~
DISTAL RADIUS-U'LNA 59.1 DISTAL RADIUS-ULNA I I I 50.8
Nhere modifi- I I I
~ ~
CARPALS 39.2 a: CARPALS 8,3
(Figure 8-4b) ~ PROXIMAL METACARPAL
I
I
I
I 29.2 :s:::> PROXIMAL METACARPAL
I I
I
I
I 16.9
:::>
u I I
i5
DISTAL METACARPAL I 24.2 0
i5
DISTAL METACARPAL I I I 19.2
tssessment of I ! I I I
~<
21.6
~<
ANTERIOR PHALANGES ANTERIOR PHALANGES 19.1
. the modified PROXIMAL FEMUR 100.0 PROXIMAL FEMUR I 94.7

DISTAL FEMUR Hio.o DISTAL FEMUR 100.0
ct limb trans- ·'
PROXIMAL TIBIA I 97.1 PROXIMAL TIBIA I 94.2
DISTAL TIBIA 78.0 DISTAL TIBIA ~ iI 58.9
luation might I
~
TARSALS I I 51.6 TARSALS 25.1
I I
I
=i!
PROXIMAL METATARSAL I 37.5 PROXIMAL METATARSAL I 18.9
Ldeveloped or DISTAL METATARSAL I 30.5 DISTAL METATARSAL
I
I I I 23.5
POSTERIOR PHALANGES I 25.4 POSTERIOR PHALANGES I I I 20.3
I
1nl y in models
e transport of Figure 8-4. Skeletal fat models for bison (a) modified and (b) not modified
1at low utility to account for the effect of riders.
1441 A. M. Emerson
MARROW FAT MODELS ses work only

a. MODIFIED FOR RIDERS IISIMAVGMARI b." NOT MODIFIED FOR RIDERS IISIAVGMARI
order of selecti<
Figures 8-6 a:
40 0 60
ATLAS
0 20 60 80 100
0.0 ATLAS
20
I
40 80
I
100
0.0
ses reported b~
I
AXIS I 0.0 AXIS I
I.
0.0 terns of carcasE
CERVICAL VERTEBRAE (3·71 I CERVICAL VERTEBRAE (3·71 0.0
.J THORACIC VERTEBRAE
I
I
0.0
0.0 .J THORAC:c VERTEBRAE
I
I 0.0 Alcelaphus buse
<( <(
I
x
<(
LUMBAR VERTEBRAE ~
SACRUM-PELVIS
I
I
0.0
3.9
R
<(
LUMBAR VERTEBRAE
SACRUM-PELVIS ~
I
0.0
3.9
typical of Had~
I·
I
CAUDAL VERTEBRAE
I 0.0 CAUDAL VERTEBRAE I
I
0.0 movement of '
RIBS I o.o RIBS 0.0
STERNUM I 0.0 STERNUM I i 0.0 buffalo ( Synceri
SCAPULA : 40.6 SCAPULA I I 1.3
of the pattern a
HUMERUS I 79.8 HUMERUS t 79.8
RADIUS-ULNA II 69.4 RADIUS-ULNA ! I 58.9 Hadza general
0:
I ir· I
0:
: I 0.0
l=Ji
::5 CARPALS 43.4 CARPALS .
I ::5 I
I dicular than ax
hri
~ ~ 17.3
u METACARPAL 30.3
u METACARPAL I I
Ci ANTERIOR PHALANGES I I .I 15.9 Ci ANTERIOR PHALANGES I I I l 1.4
in Figure 8-7, sl
~
<(
FEMUR
TIBIA
93.5
100.0 ~
<(
FEMUR
TIBIA
I 87.0
100.0 transported an
I
II
R
TARSALS 60.6 - TARSALS I 0.0
I I I identified as e1
~
METATARSAL I 40.8 METATARSAL I I 21.1
POSTERIOR PHALANGES I I 22.2 POSTERIOR PHALANGES I I ! 3.5
species carcass
species, differe
Figure 8-5. Marrow fat models for bison (a) modified and (b) not modi- analysis is not
fied to account for the effect of riders. selection exceF
selected for tra
selected for re1
axial elements (based on skeletal or marrow fat yields) were abandoned at the alternative stra
kill, but most likely not before the high utility fat and muscle was removed. when a kill is rr
This is comparable to the traditional expectations of the "schlepp effect." In
contrast, a pattern of skeletal element selection in which most axial units are
transported and few appendicular units arrive at base camp is not suggested Ca1
by any of the utility models. It does not appear that one can account for such
assemblages on the basis of utility differences alone. Rather, it is more likely Co1
that transport decisions, processing costs, and enroute food consumption or AtE
snacking may explain the variability in transport practices. Both p~tterns of ments of trans1
carcass transport to base camps (i.e., one showing greater frequency of axial the carcass and
unit transport and the other showing more frequent appendicular unit trans- lowest utility).
port) are evident in the Badza data presented by Bunn and colleagues (1988) Carcass RecovE
and O'Connell and associates (1988,1990). chosen, the cm
considerably..v
entire carcass l
Hadza Transport Practices frmn the assem
use of this stra
Through scalogram analyses of Badza hunting and carcass trans-
port data, O'Connell and colleagues (1990) found that more axial than appen- returned to ba
selected for foe
dicular units are typically returned to base camps, but that it varies both
The alcelaphinE
between and within species. As they note, scalogram analyses rank carcass
part transport across individual cases of hunting and transport. The stated ity possible in a
The alternati
assumptions of the analysis are that (1) if parts are moved, then high ranked
recovery is no1
parts are among those moved; and (2) all higher ranked parts will be moved,
ported or disc
if low or intermediate ranked elements are moved, and this pattern continues
assessments of
hierarchically (O'Connell et al. 1990:304). They also note that such scale analy-
frequent returr
ses work only if units can be ranked by utility and if rank determines the
{ISIAVGMAR)
order of selection. ·
Figures 8-6 and 8-7 represent reproductions of some of the scalogram analy-
60
40 80 100
ses reported by O'Connell and colleagues (1990:303-304), showing two pat-
0.0
o.o terns of carcass recovery and transport. The alcelaphine antelope (hartebeest,
0.0
0.0 . Alcelaphus buselaphus, and wildebeest, Connochaetes taurinus) data are most
0.0
3.9
typical of Hadza transport practices. This pattern is characterized by greater
0.0 movement of axial units but shows some limb unit transport as well. The
0.0
0.0 buffalo (Syncerus caffer) data include one case (Figure 8-6b, case HB12) typical
I I 1.3
of the pattern attributed to the "schlepp effect." Reportedly, this is atypical of
II 79.8
Badza general practices. This case shows more frequent recovery of appen-
~
58.9
I 0.0
I I I 17.3 dicular than axial elements. The giraffe (Giraffa caJ?Jelopardalis) data presented
I ! i 1.4
in Figure 8-7, show a mixture of axial and appendicular units most frequently
l 87.0
100.0 transported and differ from Figure 8-6 in the greater frequency of values
I 0.0
I
I
I
I
I 21.1 identified as errors (circled symbols) along the gradient. Although variable
I I ! 3.5
species carcass composition might contribute to the variability seen between
species, differences within species may occur if the initial assumption of the
:nd (b) not modi- analysis is not met. That is, utility and rank sometimes may not determine
selection except in how elements are eliminated from transport rather than
selecteqi for transport. That might happen if the complete carcass is initially
selected for return. To understand this distinction, one must recognize two
mndoned at the alternative strategies of carcass recovery that hunters must decide between
! was removed. when a kill is made .
.lepp effect." In
t axial units are
s not suggested Carcass Recovery Strategies and Assemblage
ccount for such
t is more likely Compositions
~onsumption or
At small-scale kills, an initial decision is m.ade, based on assess-
:oth patterns of ments of transport constraints, to transport the entire carcass or to subdivide
quency of axial the carcass and abandon sorrie parts or skeletal elements (presumably those of
:ular unit trans- lowest utility). The decisions may be characterized as Maximum and Limited
>lleagues (1988) Carcass Recovery strategies, respectively. Depending upon which strategy is
chosen, the composition of the assemblage transported to base camp varies
considerably. When a Maximum Carcass Recovery strategy is employed, the
entire carcass is selected for return, and parts are subsequently eliminated
from the assemblage to meet food needs enroute to base camp. With repeated
i carcass trans-
use of this strategy, axial units should dominate the assemblage of all parts
ial than appen- returned to base camps because, generally, they are hot the parts that are
: it varies both selected for food use away from camp. Why they are not is discussed below.
es rank carcass The alcelaphine data (Figure 8-6a) appear to be representative of the variabil-
ort. The stated ity possible in a Maximum Carcass Recovery assemblage composition.
en high ranked The alternative, Limited Carcass Recovery strategy, is one in which total
recovery is not possible and subsets of the carcass are selected to be trans-
will be moved,
.ttern continues ported or discarded. The extent of discard is variable, depending upon
1ch scale analy- assessments of transport constraint. O.ver time, it generally results in the more
frequent return of appendicular than axial skeletal elements to base camps.
1461 A.M. Emerson
a. ALCELAPHINE ANTELOPE b. BUFFALO GIRAFFE
CASE NUMBER
CASE NUMBER 2 6 11 HB17 CASE NUMBER HB6 HB12
. ELEMENTS TRA
ELEMENTS TRANSPORTED ELEMENTS TRANSPORTED
RIBS
CERVICAL VERTEBRAE T T T T HUMERI T T THORACIC VER'
THORACIC VERTEBRAE T T T T RADIOCUBITI CD T LUMBAR VERTE
LUMBAR VERTEBRAE T T T T CARPALS CD T PELVIS
· PELVIS T T T T METACARPALS T T SCAPULAE
RIBS T <D T T F/PHALANGES T T FEMORA
SKULL T T T SCAPULAE T t HUMERI
MANDIBLE T T T FEMORA T t RADIOCUBITI
SCAPULAE T T T TIBIAE T t TIBIAE
A/PHALANGES T T T TARSALS T t A/PHALANGES
FEMORA T T METATAR.SALS T t CARPALS
TIBIAE T T A/PHALANGES T t METACARPALS
TARSALS T T SKULL T F/PHALANGES
METATARSALS T T MANDIB.LE T TARSALS
HUMERI T T CERVICAL VERTEBRAE T MET A TARSALS
RADlOCUBITI T THORACIC VERTEBRAE T CERVICAL VERT
CARPALS T LUMBAR VERTEBRAE T SKULL
METACARPALS T PELVIS T MANDIBLE
F/PHALANGES T RIBS t CD
Fig·
Figure 8-6. Frequency of unit transport as indicated by scalogram analy- anal
ses for (a) alcelaphine antelope and (b) buffalo (Source: O'Connell et al. fore
1990:303-304). T =more than 66% transported; t =34%-66% trans-
ported; - =less than 34% transported; circled symbols identify errors
where values are out of place along the gradient. dicular units s
highest yields
/ accounting for
Presumably, skeletal elements are eliminated from the initial transport assem- Another fac-
blage if they have (1) low yields of skeletal fat and (2) low skeletal fat yields Limited Carca:
per unit of bone weight, indicating high transport costs. from incomplE
Figure 8-8 illustrates the relationship, for bison, between skeletal fat yield muscle remove:
and yield per unit .of fresh bone weight (which is used as a proxy measure of than that resu
relative transport cost).2 The low yields of skeletal fat and relatively low skele- under circums
tal fat yields per unit of bone weight are evident for most axial elements in the losses from det
figure. The sacrum-pelvis and ribs, however, have higher yields of fat and fat yields (othe:
low-to-moderate transport costs; thus, they might be transported more fre- of processing e
quently under a Limited Carcass Recovery strategy than other axial units. In associated higr
Figure 8-6b, case HB12, where a Limited Carcass Recovery strategy is evident, low yields of s1
the ribs (but not the pelvis) show more frequent transport than other a.Xial In Figure 8-5
units. That is not true in case HB6, but it appears to represent a Maximum measure of rei
Carcass Recovery event. .potential losse~
Figure 8-8 also illustrates the relatively low value of the scapula and distal high. The thora
limb elements (i.e., the phalanges, metapodials, carpals, and tarsals). The dis- to the greatest
tal limb elements are more often eliminated from the transport assemblage of losses are largE
giraffe carcasses, as evident in Figure 8-7, cases 72, HB27, and 25. Appen- However, they
GIRAFFE
CASE; NUMBER HB22 72 HB27 25 HB19 HB21 HB29 39 41

HB6 HB12
ELEMENTS TRANSPORTED
rED
RIBS T CD CD T
T T THORACIC VERTEBRAE T 0 T T
CD T LUMBAR VERTEBRAE T 0 CD T
CD T PELVIS T T T
T T SCAPULAE T CD T T
T T FEMORA T T T T
T HUMERI T T 0 T
T RADIOCUBITI T T .T
T TIBIAE T T T
T R/PHALANGES T T t
T CARPALS T
T METACARPALS T
T F/PHALANGES T CD
T TARSALS T
T METATARSALS T
T CERVICAL VERTEBRAE
T SKULL
T MAN biBLE
CD I
Figure 8-7. Frequency of unit transport as indicated by scalogram
;calogram analy- analysis for giraffe (Source: O'Connell et al. 1990:304). See Figure 8-6
O'Connell et al. for explanation of symbols.
4%-66% trans-
' identify errors
dicular units such as the femur, humerus, tibia, and radius-ulna have the
highest yields of skeletal fat and yields per unit of bone weight, thus
accounting for their higher transport priority.
ansport assem- Another factor that may affect which units are included in the initial
~letal
fat yields Limited Carcass Recovery transport assemblage is potential calorie losses
from incomplete fleshing. O'Connell and colleagues (1988:128) note that
:eletal fat yield muscle removal associated with the axial skeleton tends to be less complete
JXY measure of than that resulting from deboning appendicular units. This suggests that
vely low skele- under circumstances where fat is a highly desired carcass product calorie
elemen ts in the losses from deponing would be higher in axial than appendicular units, where
elds of fat and fat yields (other than skeletal fat) are higher. Thus, if potential losses and costs
>rted more fre- of processing exceed the expected costs of transport, then axial elements with
r axial units. In associated high yields of total fat may be transported despite their relatively
tegy is evident, low yields of skeletal fat.
1an other axial In Figure 8-9, the relationship between nonskeletal fat yield and a proxy
1t a Maximum measure of relative transport cost is shown. This figure illustrates where
potential losses of fat calories from deboning of the bison carcass would be
pula and distal high. The thoracic and lumbar vertebrae and the ribs are the axial units subject
.rsals). The"dis- to the greatest losses of fat calories from incomplete fleshing. Whether the
: assemblage of losses are large enough to offset processing and transport costs is unknown.
nd 25. Appen- However, they are the units more often transport~d when giraffe carcass
1481 A.M. Emerson
18000
3200
t:, AXIAL UNITS 0 FEM

0 APPENDICULAR UNITS
16000
2800
14000.
2400
12000.
0 TIS
c;;
2000 0 HUM (.)
:6
~ a- l
0 lU
==-
0
....J
w >=
....J :;::: 1- 10000.
lJ.:J
>=
1-
<( Lt
1- u.. -l
<( 1600 0 RUL ....J <(
1-
u..
....J ~ lU
....J
<( lU lU
....J
1- w ~
w ::.::: (/)
....J 8000.
w C::, rJ'J
SPEL z
~
::.:::
Cl) 0
1200
z .....J
~
0
!:::.
6000·
800
6 RIBS
OMT 4000-
0 MC
400 THOR
0 TARS
C::,
t:, LUMB 0 APHA
0 0 PPHA
SCAP 2000.
0 CARP
0
0.00 0.20 0.40 0.60 0.80 1.00
SKELETAL FAT (kcal) I FRESH BONE WEIGHT (g) 0.

0
PROXY MEASURE OF RELATIVE TRANSPORT COST
Figure 8-8. Relationship between bison skeletal fat yield and a proxy
measure of relative transport cost. Based upon yields from a Spring-killed Figu
Adult Female (SAF) bison (Source: Emerson 1990b). See endnote for meas
abbreviations. Adul
18000
6 11-lOR
0 FEM
~RIBS 6 AXIAL UNITS
16000
LUMB 0 APPENDICULAR UNITS
14000
12000
0 TIS (ij
(J
6
0
_J
0 w
_J
w >=
>= ~ 10000
~
_J
LL.
_J
,1:(
+-- STER 0 FEM
~ '_J /W
w
_J w
w (/)
~
~
(/) 8000 6 SPEL
z ~
0 LL.
z . ...J
g~
6000
4000
0 HUM
2000
0 SCAP
6
1.00 6 AX
AT
CARP
PPHA TARS
MC MT o ns
0.00 0.20 0.40 0.60 0.80 1.00

T
SKELETAL FAT (kcal) I FRESH BONE WEIGHT (g)
PROXY MEASURE OF RELATIVE TRANSPORT COST

·ield and a proxy
m a Spring~killed Figure 8-9. Relationship between bison nonskeletal fat yield and a proxy
. See endnote for measure of relative transport ·cost. Based upon yields from a Spring-killed
Adult Female (SAF) bison (Source: Emerson 1990b).
I
150 A.M. Emerson
recovery follows a. Limited Carcass Recovery strategy (Figure 8-7, cases 72,
HB27, and 25). ·
2000
Excepting elimination of carcass units that are tainted or consumed, an un-
modified Maximum Carcass Recovery· transport assemblage should include
all skeletal elements. A Limited Carcass Recovery assemblage, on the other
hand, should include the limb elements and those axial units' that have high 1800
skeletal fat yield and large potential fat losses from field fleshing. One might
expect a Limited Carcass Recovery strategy to be evident where distance or
carcass size becomes too great or the labor pool too S!-Uall to opt for complete
recovery. 1600
Modifications to Transported Assemblages. 1400
Regardless of which carcass recovery strategy is used, the assem-

blage of bones that arrives at a base camp frequently differs from that initially
selected. At the kill site, body parts maybe left unused because they are 1200
tainted. Others may be selected for consumption at the kill site or at interme..:
diate snacking locations (Bunn et al. 1988) on the return to base camp. Those ~
units selected for use and discard (opposed to those abandoned because of 0
.....J
w
unacceptable quality) should be controlled by assessments of processing and >= 1000
w
transportation costs. It has already been shown that processing costs due to (/)
potential fat losses are high for axial elements. Furthermore, most of these us0:
units have little marrow and only limited amounts of bone grease; thus, they 0
800
rarely would be the focus of snacking. The exception to this might be the ribs
that may be subdivided and partially consumed in the field, without increas-
ing package numbers. It seems likely, however,that consumption in the field
should focus on appendicular elements. 600
Some appendicular elements might be eliminated from the assep1blage
transported to base camp as a function of their relative yields of marrow and
bone grease because of differences in the processing costs of the two carcass
products. Marrow, which may be obtained by cracking bones and removing 400
the contents of the medullary cavity, is relatively easily obtained. Bone grease,
however, is dispersed in the cancellous tissue and requires more effort to
collect. Differences in processing costs, therefore, suggest that marrow should
200
be the focus of snacking.
Bone grease and marrow, however, are not evenly distributed throughout
the skeleton. Selecting a unit with high marrow fat yield may also result in the
selection of one that also has high grease fat yield. The objective of selection 0
should be one in which processing costs are kept low. Thus, elements with 0
high bone grease fat yields should be transported and processed at base
camps, while elements with lower yields may be used in the field.
The relationship between grease fat yield and the proportion of skeletal fat
yield that is bone grease is illustrated for bison in Figure 8-10. In this graph, as
the proportion of skeletal fat that is derived from bone grease increases, the Figur
proportion that is obtained from marrow decreases. Whole element and prox- propor
imal/ distal element values are both shown. killed,
re 8-7, cases 72,

2000
msumed, an un- 0 WHOLE UNITS
should include e HALFUNITS 0 FEM
ge, on the other
; that have hig~ 1800
1ing. One might
here distance or
)pt for complete
1600
s 1400
Lsed, the assem- e DFEM

0 SPEL
om that initially
~cause they are 1200
:e or at interme-
Lse camp. Those ~01 OHUM
>ned because of ....J'
processing and ~ 1000

UJ
~ng costs due to (/)
~, most of these u:i

a: 118 0
(.9
·ease; thus, they RUL 0
BOO
1ight be the ribs e PHUM
rVithout increas- PTIB e 0 RIBS
tion in the field
600 PFEM e
:he assemblage
of marrow and DRUL e
the two carcass
' and removing 400
/THOR
~d. Bone grease, DHUM 0 - TARS
PRUL M OMT
more effort to APHA
0
OMC C»-.1Te
marrow should 0 LUMB
200 DMC e 0
DTIB e 0
PPHA SCAP
ted throughout 0 CARP
.lso result in the CAUD 0,.../ C3C7
-AX.
:ive of selection 0 --.:::-AT
, elements with 0 ~0 40 60 80 100 STER
>cessed at base MOSTLY ( GREASE YIELD (kcaij ) • 100 MOSTLY
~ld. MARROW SKELETAL FAT YIELD (kcal) GREASE
n of skeletal fat
:n this graph, as
e increases, the Figure 8-10. Relationship between bison bone grease fat yield and the
ment and prox- proportion of grease to skeletal fat yield. Based upon yields from a Spring-
killed Adult Female (SAF) bison (Source: Emerson 1990b).
152j A.M. Emerson
Based upon Figure 8-10, one might expect that units·in which skeletal fat is Removals of
dominated by marrow and in which bone grease yields are also low (e.g., the initial assessme1
metapodials) are likely to be processed and consumed in transit and the The use of the 1
exhausted bones discarded. Units that have high marrow fat yields, but also sequent removi
have high grease fat yields (e.g., the tibia, humerus, and radius-ulna), may be unless they incr
more frequently transported to base camps. Units that have low grease fat the skeletal ele1
yields and little or no marrow (e.g., the phalanges, scapula, carpals, and selection of eros
tarsals) may be discarded or less efficiently processed in the field after remov- that increases. in
ing usable tissue. Assessments of cost must include evaluations of loss, extra is less acceptabl
processing, or transport costs necessitated by breaking whole bones into ing episodes.
smaller units. Thus, loss from not processing bone grease in the metapodials With respect
would be low, but similar treatment of the humerus or femur would result in model is strong
major relative loss. discussions of
Equally important to the selection of units for snacking is the extent of instrumental to
carcass subdivision. Based upon reports by O'Connell and colleagues (1990), it under condition
appears that the integrity of limbs is often maintained in transport. Under transport constr
these circumstances, in-transit use and discard may-progress within a single of hunting and
limb unit rather than between limb units. In essence, the rider concept is of assemblage 1
evident here. The rationale for treating units· in this manner rather than before assumir
according to relative utility appears to be related to assessments of transport O'Connell and
cost. Presumably, limb subdivision would increase package numbers and thus hunting-deri vee
increase the difficulty of transport. It should be a
port costs are cr
scale kills. Furth
Conclusions ity of carcass c
It appears that four factors-initial assessments of transportation human and non
constraints, processing cost, skeletal fat yield, and snacking prior to the return in which attenti
to base camp-are responsible for some of the variability in bone assemblage range of variabi
composition seen in the Hadza data presented by O'Connell and coileagues standing of sma
(1990). Specifically, it appears that an initial decision of whether to transport occur after their
the entire carcass (excluding tainted pieces) or only the most useful parts are critical to ou
(which may be determined by something other than general utility), when
followed by subsequent eliminations of parts to meet consumption needs, can Ack
account for (l) the high proportion of axial elements returned to base camps,
(2) the variability in appendicular element discard, and (3) the variability in lWOl
carcass treatment between and within taxa. Some differences may be due to for their helpful
carcass variability between species, to poor animal condition, or age-related to acknowledgE
yield differences, but the initial assessment of transport costs appears to be the Anthropology :
main factor influencing assemblage composition. figures. The bis<
Metcalfe (sensu O'Connell et al. 1990:312) has suggested that processing as part of a dis~
and transport costs, together with the ratio of the yield of edible to inedible Foundation thrc
tissue, should be a good indicator of decisions to process and transport body graciously provj
parts. It is probable that initial assessments of transport constraints determine
whether a Maximum or Limited Carcass Recovery strategy is employed. The Not~
expectations of the Maximum Carcass Recovery strategy are not in conflict
with Metcalfe's suggestion. It merely represents the case where initial trans- 1. Acronyms fc
portation costs are considered so low as to be unimportant. When they are
:h skeletal fat is Removals of elements from the transport assemblage subsequent to the
.o low (e.g., the initial assessment is probably dependent upon processing and transport costs .
:ransit and the The use of the Maximum Carcass Recovery strategy should dictate that sub-
yields, but also sequent removals are not subject to evaluations relative to transport costs
;-ulna), may be unless they increase costs (e.g., by increasing package numbers). Removal of
low grease fat. the skeletal elements from most or all of a limb rather than the piecemeal
a, carpals, and selection of cross-limb subdivisions based upon utility might be an indication
ld after remov- that increases in transport costs brought about by increasing package numbers
lS of loss, extra is less acceptable than some loss due to less efficient processing during snack-
ole bones into ing episodes.
he metapodials With respect to the expectations of the "schlepp effect," it is clear that the
would result in model is strongly tied to assessments of transport constraints. The original
discussions of this factor clearly state that distance and carcass size are
s the extent of instrumental to its application. To fairly question the validity of the model
~agues (1990), it under conditions of axial unit transport dominance, it must first be shown that
msport. Under transport constraints do exist. On the other hand, it seems that the complexity
within a single . of hunting and transport decisions demands more sophisticated assessments
lder concept is of assemblage composition than simple axial or appendicular dominance
er rather than before a~suming a hunting or scavenging mode of acquisition. Clearly
1ts of transport O'Conn7ll and colleagues (1990) have shown that both may dominate in
mbers and thus hunting..:derived assemblages.
It should be apparent that more studies of processing, deboning, and trans-
port costs are critical to our understanding of carcass transportation at small-
scale kills. Further evaluation of this problem also depends upon the availabil-
ity of carcass composition data for African species. Additional studies of
transportation human and nonhuman predators and scavengers such as that by Stiner (1991),
or to the return in which attention is given to the details of assemblage compositions and the
me assemblage range of variability in assemblage types, are important to our ultimate under-
and colleagues standing of small-scale hunting. Clearly, major changes to assemblages often
ter to transport occur after their arrival at base camps, and additional studies of those factors
st useful parts are critical to our understanding of the archaeological record.
. utility), when
~tion needs, can Acknowledgments
to base camps,
.e variability in I would like to thank Jean Hudson and two anonymous reviewers
may be due to for their helpful comments on an earlier draft of this paper. I also would like
. or age-related to acknowledge Timothy A. Kohler and the Washington State University
>pears to be the Anthropology Department for their assistance in the preparation of the
figures. The bison carcass composition data used in this paper were gathered
hat processing as part of a dissertation research project supported by the National Science
lble to inedible Foundation through grant BNS-8406036. Keri Woodall and Nannette Kistler
transport body graciously provided me with a home while I completed the final manuscript.
1ints determine
employed .. The Notes
not in conflict
re initial trans- 1. Acronyms for utility models include. indicators of their specific construction.
When they are preceded by an (S), the model data have been standardized.
1541 A. M. Emerson_
Standardization is accomplished by 4ividing the value for. each part by that of the Perkins, D., and
highest valued part and ·multiplying by 100. This permits utility values to range 1968 AHun
between zero and 100 but does not alter their relative ranking. Acronyms may also Speth,J. D.
have a leading M. This character indicates that the appendicular data have been 1983 Bison F.
modified to account for riders. By definition of the origi:11al authors (Metcalfe and Stin~r, M. C.
fanes 1988), all FUI models have been modified and do not necessitate the leading M. 1991 Food I
The presence of AVG as part of the acronym indicates that the model is based upon I ournal of A1
data averaged from four bison of various age and sex classes. Models for bison of White, T. E.
specific age and sex classes are presented elsewhere (Emerson 1990b). 1952 Obsen
. 2. The following abbreviations are used in Figures 8-8 through 8-10: American A1
APHA = anterior phalanges MC = metacarpal

AT= atlas MT = metatarsal
AX= axis PFEM = proximal femur
CARP = carpals PHUM = proximal humerus
CAUD = caudal vertebrae PMC = prmdmal metacarpal
C3C7 = third through seventh PMT = proximal metatarsal
cervical vertebrae PPHA = p9sterior phalanges
CERV = cervical vertebrae PRUL = proximal radius-ulna
DFEM = distal femur PTIB = proximal tibia
DHUM = distal humerus RIBS~ ribs
DMC = distal metacarpal RUL = radius ulna
DMT = distal metatarsal SCAP = scapula
DRUL = distal radius-ulna SPEL =sacrum-pelvis
DTIB = distal tibia STER =sternum
FEM= femur TARS = tarsals
HUM= humerus THOR = thoracic vertebrae
LUMB = lumbar vertebrae and flank TIB =tibia
References I
Binford, L. R.
1988 Variability in Bone Assemblage Formation from Hadza Hunting, Scaveng-
ing, and Carcass Processing. Journal of Anthropological Archaeology 7:412-457.
Emerson, A. M.
1990a Carcass Product Yields in Bison bison and Hunter Selection: Utility Model
Construction and Interpretations of Use. Paper presented at the Sixth Inter-
national Conference of the Council for Archeozoology, Washington, D.C.
1990b Implications of Variability in the Economic Anatomy of Bison bison. Ph.D. dis-
sertation, Washington State University. University Microfilms, Ann Arbor.
Metcalfe, D., and K. T. Jones
1988 A Reconsideration of Animal Body Part Utility Indices. American Antiquity
53:486-504.
O'Connell, J. F., K. Hawkes, and N. Blurton Jones
1988 Hadza Hunting, Butchering, and Bone Transport and Their Archaeological
Implications. Journal of Anthropological Research 44:131-161.
1990 Reanalysis of Large Mammal Body Part Transport among the Hadza.
Journal of Archaeological Science 17:301-316.
t by that of the Perkins, D., and P. Daly

alues to range 1968 A Hunters' Village in Neolithic Turkey. Scientific American 219(5):96-106.
nyms may also Speth,J. D.
lata have been 1983 Bison Kills and Bone Counts. University of Chicago Press, Chicago.
; (Metcalfe and Stiner, M. C.
the leading M. 1991 Food Procurement and Transport by Human and Non-human Predators.
l is based upon Journal of Archaeological Science 18:445-482.
~ls for bison of White, T. E.
1952 Observations on the Butchering Techniques of Some Abonginal Peoples: I.
·American Antiquity 17:337-338.
lr
nerus
:arpal
:arsal
langes
us-ulna
·brae
:tting, Scaveng-
7:412-457.
: Utility Model
he Sixth Inter-
m, D.C.
ison. Ph.D. dis-
:m Arbor.
erican Antiquity
Archaeological
ng the Badza.
people by suggestil
costs of transportin
at kill sites after de
from inside them,
axial and girdle ele1
Both groups doc1
9. Bone Assemblages at Base tities of edible mea
bone assemblages 1
Camps: A Further Consideration dance of limb elem
cator of a transpor
of Carcass Transport and Bone many ·axial parts, E
those elements toE
Destruction by the Hadza abundance of liml::
that the most com1
Henry T. Bunn brae, scapulae, and
Because archaec
skeletal parts in be
Abstract: Ethnoarchaeological observations of carcass transport and pro- specific topics of E
cessing by Hadza hunter-gatherers in northern Tanzania provide a use- consider how and ·
ful but complex basis for evaluating archaeological bone assemblages. they do. In, this p
The Hadza typlcally transport essentially entire field-butchered carcasses developed by 0'0
of most animals to base camps, where vertebrae and some other ele- transport in relatio
ments are boiled for fat. Limb representation is high in the Hadza trans- tal composition of
port data and in bone assemblages from Hadza base camps. Axial and I consider the prir
girdle element representation is high in the transport data because of the Hadza camps· and
availability of boiling technology, but reduced, especially for vertebrae,
tions in the bone a.
in ~ase camp assemblages because processing by the Hadza destroys less
durable bones. With appropriate methodology, high limb representation
remains a sound indicator of transported assemblages.
The :H
andC
Introduction Using
Recent ethnoarchaeological studies of Hadza hunter-gatherers in 0' Connell and co:
northern Tanzania, by two independent groups of researchers, present carcass transport
archaeologists with a dilemma. One group (Bunn et al. 1988, 1991) concluded impact on archa~c
that in transporting carcasses of all but the largest animals (i.e., giraffe, and the behavior
elephant) from kill sites to base camps th~ Hadza typically (1) try to carry sented a scalogra
entire carcasses in field-butchered units, (2) abandon mainly axial units at kill ported to base cm
sites when circumstances warrant, and (3) consequently accumulate the entire transport costs is
range of skeletal elements in approximately natural anatomical frequency but bone assemblage~
with a notably high representation of limb bones. The other group (O'Connell blades, backbone
et al. 1988, 1990) concluded essentially the opposite pattern for the same groups 4-6) are ra
of axial parts; 01
From Bones to Behavior: Ethnoarchaeological and Experimental Contributions to the Interpretation of elements). More r
Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional ued to claim that
Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights des as the basis f
reserved. ISBN 0-88104-076-2. logical sites.
156
Carcass Transport 1157
people by suggesting that the Badza typically (1) try to reduce the energetic
costs of transporting skeletal elements, (2) consequently abandon limb bones
at kill sites after defleshing them and extracting and consuming the marrow
from inside them, and (3) preferentially transport and accumul'ate mainly
axial and girdle elements at base camps.
Both groups documented the transport to Badza base camps of large quan-
tities of edible meat and other tissue, but the archaeological signatures of the
bone assemblages would differ. Bunn and colleagues suggested that an abun-
ttion dance of limb elements, as documented at Badza base camps, is a good indi-
cator of a transported assemblage and that a major incentive for carrying as
ne many axial parts, especially vertebrae, as the Hadza do is the ability to boil
those elements to extract grease. O'Connell and colleagues suggested that an
abundance of limb elements is a poor indicator of significant transport and
that the most common elements transported to Badza base camps are verte-
brae, scapulae, and pelves.
Because archaeologists commonly use the proportional representation of
skeletal parts in bone assemblages to examine overall site function and more
1sport and pro- specific topics of subsistence and diet, there is a pressing need to carefully
provide a use- consider how and why the Hadza select for transport the carcass portions that
.e assemblages. they do:/In ,this paper (1) I examine aspects of the interpretive framework
hered carcasses develop~d by O'Connell and colleagues for their evaluation of Hadza carcass
orne other ele- transport in relation to my observations of the Hadza, (2) I report on the skele-
te Hadza trans- tal composition of the bone assemblages from two Badza base camps, and (3)
nps. Axial and I consider the principal processes involved in bone assemblage formation at
t because of the
'{ for vertebrae, Badza camps and the likely causes of differences between the skeletal propor-
z:a destroys less tions in the bone assemblages·and observations of all of the researchers.
· representation
The Hadza Pattern Reported by O'Connell
and Colleagues
Using data on the transport of skeletal elements of 39 carcasses,
r-gatherers in O'Connell and colleagues (1988) described a cause-and-effect relationship of
hers, present carcass transport by the Hadza that would, if correct, have a significant
91) concluded impact on archaeologists' understanding of the processes of carcass transport
(i.e., giraffe, and the behavi.or of ancient human foragers. O'Connell and colleagues pre-
~) try to carry sented a scalogram analysis showing that the typical skeletal parts trans-
lal units at kill ported to base camps are axial rather than limb elements. A desire to reduce
tlate the entire transport costs is emphasized as the causal factor, and it is claimed that the
frequency but bone assemblages at Hadza base camps should be predominantly shoulder
up (O'Connell blades, backbones, and pelves. The bones of very large carcasses (i.e., size
for the same groups 4-6) are rarely transported (i.e., one of one eland with transport mostly
of axial parts; one of five giraffes with transport of both axial and limb
e Interpretatiqn of elements). More recently, O'Connell and colleagues (1988, 1990) have contin-
tions, Occasional ued to claim that their results call into question the use of high limb frequen-
~rsity. All rights cies as the basis for identifying carcass transport at Plio-Pleistocene archaeo-
logical sites.
1581 H. T. Bunn
The Hadza Pattern Reported by Bunn ~nd Colleagues SAC

Using data on the transport of skeletal elements of a smaller num- LU~
ber of 25 carcasses, Bunn and colleagues (1988, 1991) described a different
pattern of carcass transport by the same Badza as were studied by O'Connell PEL 1/~
and colleagues and different causal factors. Bunn and colleagues (1988) char- THOF
acterized the typical pattern of carcass transport by the Badza as mainly
involving· the transport of entire smaller carcasses (size groups 1-2) and even CER'
larger ones (size group 3), withthe transport of very large carcasses (size
SCAI
group 4, N = 2 buffalo) focused mainly on limbs and meat (for examples of
size group classifications based on live animal weight, see Brain 1981; Bunn et HUt
. al. 1980, 1988; Klein 1976). Variations from this pattern (i.e., cases where axial
skeletal parts were transported to base camps instead of limb .elements) were RA[
mainly attributed to the reluctance of a small number of Badza individuals to FEt
share high-quality marrow fat with the people back at the base camp. The
largest giraffe carcasses (size group 5) were variably transported in direct ME.
response to differences in transport distance. Bone transport was significant Til
for five of seven carcasses of giraffe. The abundance of limb relative to axial
elements in the bone assemblages from a Badza base camp was cited in MAN 1/:
support of the observations.
Discussion
That two independent studies of ·the same Hadza individuals
could reach seemingly opposite conclusions emphasizes. the need for larger
Fi;
sample sizes and a consistent analytical approach. To detect patterns in a
ele
small carcass sample, Bunn and colleagues (1988) examined the data from a
by
broad perspective, using animal size groups rather than individual species. In
contrast, O'Connell and colleagues (1988) divided their data by species, which
probably is justifiable, given that the Hadza handle some species of equivalent casses of all
size differently (see below), but to do so reduced the analytical units to elements at ki
unconvincingly small sizes (e.g., the most abundant species, impala and zebra, The data f1
were represented by only nine and eight carcasses, respectively). from O'Conn
Larger sample sizes would permit more conclusive analysis at both general carcasses (49'
and specific levels. Accordingly, Bunn and colleagues (1991, 1993) returned in from Bunn et
1988 for more field observations, and in 75 days of uninterrupted observation, sectioned car•
the Hadza obtained and consurned 81 carcasses. That period of research more for axial part
than doubled the previously available· sample size to 110 carcasses, and it deviated fror
decisively confirmed our initial conclusions. ments in the
Figures 9-1 and 9-2 summarize observations by Bunn and colleagues (1991, the typical He:
1993) of the transport of carcasses of all taxa and of the most abundant taxon comparable tc
(zebra), respectively) All skeletal parts except skulls and ribs were trans- elements awa
ported to base camps more than 90 percent of the time. Vertebrae, scapulae, and impala fr
and pelves are typically transported to base camps by the Hadza, but the Bunn and co:
point, contra O'Connell and colleagues, is that the limb bones are also trans- Hadza, as des
ported to base camps by the Badza in comparably high frequency. In other Some uncal
words, the Hadza typically transport essentially entire field-butchered car- Hadza. Basec
~lleagues
SAC 100
:maller num- LUM ~~;:;:;:;:;:;~~~~;:;:;:;:;:;~:::;:;:;~;:;:;~ 9 9' 5
i a different
)y O'Connell PEL 1/2
; (1988) char- THOR
~a as mainly
-2) and even CERV
.rcasses (size
SCAP
examples of
l981; Bunn et HUM
:; where axial
~ments) were
RAD
1dividuals to FEM
e camp. The
ted in direct MET
3.S significant TIB
ative to axial
was cited in MAN 1/2
1 CRA
RIB
0 20 40 60 80 100
L individuals
~ed for larger
Figure 9-1. Representation in percent (%MNU) of different skeletal
patterns in a
elements of all taxa transported to base camps by the Hadza, as observed
~data from a
by Bunn and colleagues.
wl species. In
pedes, which
of equivalent casses of all but the largest taxa, while abandoning mainly some axial
tical units to elements at kill sites (Bunn et al. 1988).
tla and zebra, . The data from Bunn and colleagues can also be combined with the data
from O'Connell et al. to further increase sample size. Nearly half of all zebra
both general carcasses (49%, N = 40: 8 carcasses from O'Connell et al. 1988; 32 carcasses
3) returned in from Bunn et al. 1988, 1991, 1993) were transported to base camps as entire
l observation, sectioned carcasses, and an additional 37% were entirely transported except
esearch more for axial parts, such as skulls and ribs. Only a minority of carcasses (13%)
:asses, and it deviated from this pattern through the variable abandonment of limb ele-
ments in the way that O'Connell and colleagues (1988, 1990) concluded was
eagues (1991, the typical Badza pattern_. Patterning in the larger sample of impala (N = 37) is
undant taxon comparable to zebra with only about 20% involving significant discard of limb
; were trans- elements away from base camps. Thus, after combining the sample of zebra
~ae, scapulae, and impala from O'Connell and colleagues with the much larger sample from
:tdza, but the Bunn and colleagues, the predominant pattern of carcass transport by the
.re also trans- Badza, as described by Bunn and colleagues (1988), does not change .
ncy. In other Some uncommon taxa in the same size ranges are treated differently by the
utchered car- Badza. Based on a small sample, the differential transport of alcelaphine
160 jH. T. Bunn
CERV three times as r

1990).
LUM O'Connell ar
SAC costs motivate~
medi urn-to-lar~
THOR In making that
Daly (1968). 0'
SCAP and Daly mod1
HUM Daly were eqt:
1952, 1953, 195·
RAD likely to be trar
FEM and Daly [1968
who are familic
TIB (1968), howevE
MET tween their vi:
assemblages sl
PEL 1/2 · predominance
RIB essentially reit
Lartet and Ch:
CRA argued for the
limb bones of l;
MAN 1/2
and colleagues
0 20 40 60 80 100 and Daly modE
What is new
Figure 9-2. Representation in percent (%MNU) of the different skeletal size of carcass
elements of zebra transported to base camps by the Hadza, as observed by that the limb h
Bunn and colleagues. difficult to tra:
bones of the la:
low because of
antelopes (hartebeest and wildebeest) often does involve the introduction of principle appl
axial skeletons and limb meat into base camps, with the consumption of limb deflesh limbs,·
marrow and discard of limb bones occurring near kill sites. Ignoring observa- sites, and carry
tions and informant information about Badza motives reported by Bunn and But would tl
colleagues (1988), O'Connell and colleagues (1988, 1990) repeatedly ascribe of size groups
this alcelaphine pattern, including impala and zebra with it, to a Badza desire large giraffes
to reduce transport costs by typically abandoning limb elements at kill sites. attached to a 1
Bunn and colle~gues (1988) suggested that a reluctance to share .high-quality this practice re,
portions led to the pattern. The Badza explain it in yet another way (personal 2 and 3 carcas
observation 1988): hartebeest and wildebeest marrow is more abundant but of and bones) of;
lower quality than zebra marrow and is, therefore, eaten at kill sites instead of the hindquartE
brought back to base camps to feed children. In contrast, zebra marrow is about 3 kg. By
sweet and highly valued, and zebra limbs are transported to base camps in hold the hindq
order to provide children with high-quality marrow. Further research aimed . the defleshed 1
at defining the relationship between marrow quantity and quality for different is about 1.5 kg
taxa is clearly warranted. Preliminary experimental results confirm the Badza meat.) From tr
distinction regarding marrow abundance, with wildebeest humeri yielding cost from trar
wooden pole.··
three times as much marrow by weight as zebra humeri (personal observation

~ 100 1990).
~ 100 O'Connell and colleagues have argued that the desire to reduce transport
costs motivates the Hadza to typically discard at kill sites the limb bones of
~ 100 medium-to-large species, including impala, zebra, hartebeest, and wildebeest.
~ 99.5 In making that argument, they have reasserted the reasoning of Perkins and
Daly (1968). O'Connell and colleagues (1990:302) discuss the 'White/Perkins
j 98.4 and Daly model" as though the transport models of White and Perkins and
898.4 Daly were equivalent: they summarize the transport model of White (e.g.,
1952, 1953, 1954), who argued that limb elements of larger animals were more
j 98.4 likely to be transported than axial elements, and then they assert that "Perkins
~ 98.4 and Daly [1968] made a similar argument'' (O'Connell et al. 1988:142). Readers
who are familiar with White (e.g., 1952, 1953, 1954) and with Perkins and Daly
898.4 (1968), however, will recognize that there are fundamental differences be-
~ 96.8 tween their views. In arguing that the skeletal composition of transported
assemblages showed more complete representation of smaller animals and a
93.5 predominance of limb bones over axial elements of larger animals, White
essentially reiterated the reasoning of the nineteenth-century prehistorians
Lartet and Christy (1865-1875). In direct contrast, Perkins and Daly (1968)
argued fot the abandonment at kill sites and for the nontransport of all of the
limb bon~s of larger animals to reduce transport costs. The views of O'Connell
and colleagues (1988, 1990) are consistent only with this aspect of the Perkins
and Daly model. .
What is new in the argument of O'Connell and colleagues (1988, 1990) is the
size of carcass to which the reasoning is applied. Perkins and Daly suggested
fferent skeletal that the limb bones of very large wild cattle (size group 4) were too heavy and
as observed by difficult to transport from kill sites; the likelihood of extensive transport of
bones of the largest animals, such as elephants, is widely acknowledged to be
low because of their weight. O'Connell and colleagues contend that the same
principle applies to substantially smaller carcasses, leading the Hadza to
trod uction of
deflesh limbs, break limb bones for marrow and discard the fragments at kill
.ption of limb sites, and carry the axial skeleton and limb meat back to the base ca~p.
~ring observa-
But would this reduce transport costs in a meaningful way? For carcasses
by Bunn and of size groups 2 and 3, when limb meat is defleshed at kill sites (as with very
Ltedly ascribe large giraffes and long. transport distances), the meat is bundled up and
Hadza desire attached to a w~oden pole, which is then carried over the shoulder. Would
-s at kill sites.
this practice reduce transport costs in a meaningful way for smaller size group
~high-quality
2 and 3 carcasses, such as impala and zebra? The intact hindquarter (meat
vay (personal and bones) of an impala and zebra weigh about 4 kg and 20 kg, respectively;
undant but of the hindquarter limb bol).es (femur, tibia, and metatarsal) of a zebra weigh
~tes instead of
about 3 kg. By comparison a wooden pole stout enough and long enough to
ra marrow is hold the hindquarter meat of a zebra would probably weigh about as much as
~ase camps in
the defleshed limb bones. (The actual weight of a typical Hadza digging stick
~search aimed
is about 1.5 kg and is perhaps large enough for carrying a weighty bundle of
y for different meat.) From this comparison, it is hard to see much of a savings in transport
rm the Hadza cost from transferring the meat from the pole-like, articulated limb to a
meri yielding wooden pole. When one consiqers that an intact articulated limb is, without
1621 H. T. Bunn
any modification,,an especially convenient package for transporting over the

shoulder and that a Hadza man has no noticeable probl~m carrying even the F
larger hindquarter of a buffalo (35-40 ~g) for a distance of 5 km, the merits of c
wooden poles and defleshing limbs to reduce transport costs seem question-
able. The Hadza individuals who occasionally behave in thi~ way do achieve PEL 1
an energetic benefit, but not the one envisioned by O'Connell and colleagues.
Rather, the individuals who consume limb marrow by themselves guarantee H'
that this ·nutritious food item winds up in their stomachs instead of being R
consumed by someone else in the base camp.
1
Bone Assemblages at Hadza Base Camps CE
Leaving aside the causal factors of differential skeletal transport, I M

have shown that an .abundance of limb bones at base camps does characterize MAN 1
the transported component of at least the commonest animals (impala and
zebra) utilized by the Hadza. Using the patterns of. skeletal transport TH
described by O'Connell and colleagues (1988, 1990; i.e., paucity of limb bones;
abundance of scapulae, vertebrae, and pelves), it would not be possible to
sc.
recognize Hadza base camps from the carcass portions most often transported ~
to them, nor would it be possible to recognize Hadza base camps by studying
the bone assemblages found at them (Bunn et al. 1988, 1991)! s
The main problem is that the skeletal representation in the only two bone
assemblages analyzed so far from Hadza base camps is complex and not
completely in agreement with any data set on carcass transport to base camps
by the Hadza (Bunn et ru. 1988, 1991, 1993). Both base camps were occupied
for several months by the Hadza and then abandoned for two to three years Fig
before analysis. Although some Hadza do infrequently keep pet dogs, no dogs aH
were present at the base camps reported here. From the observations of
carcass transport by Bunn and colleagues, very high proportions of all skeletal contribute to ;
elements would be predicted, with some underrepresentation of axial parts involved, the l
(especially skulls and ribs). As noted above, for example, approximately 87% the appropria1
of the zebra carcasses in the combined sample reported by Bunn and col- structions fron
leagues and O'Connell and colleagues were transported in their entirety or the ethnoarchc:
were lacking only the skull and some ribs. As reported previously for one the two bone
Hadza base camp assemblage (BA-1; Bunn et al. 1988, 1991) and below for following.
another base camp assemblage (SN-A), limb representation is reasonably First, Hadza
high, especially when the bones of large animals of size group 3 are viewed ent, which was
separately as a group (Bunn et al. 1991) or by species (Figure 9-3 for zebra at base campsites
SN-A). Representation of smaller animals is more variable (Figure 9-4 for with this mills
impala at SN-A), and for all sizes of animals, significant underrepresentation bones and butc
of many elements is evident (Figure 9-5 for SN-A). This is especially evident duct butcherir
for vertebrae and ribs. might prompt
. showing no in·
Taphonomic Processes at Hadza Base Camps interviews for
observation.
Many factors may contribute to the complex skeletal patterns of Second, the J
bone assemblages at Hadza base camps. Briefly reviewing them here may brae, ribs, and
1rting over the

rying even the FEM 100
L, the merits of CRA
eem question-
ray do achieve PEL 1/2
nd colleagues. HUM
.ves guarantee
;tead of being RAD
TIB
CERV
tal transport, I MET
~s characterize MAN 1/2
s (impala and
~tal transport THOR
of limb bones; SCAP
be possible to
~n transported RIB
JS by studying
SAC
)nly two bone LUM
1plex and not
to base camps o· 20 40 60 80 100
,vere occupied
Figure 9-3. Representation in percent (%MNU) of zebra bones at SN-A,
to three years
a Hadza base camp.
dogs, no dogs
Jservations of
' of all skeletal contribute to an understanding of the sequences of taphonomic processes
of axial parts involved, the biases they introduce in the surviving bone assemblages, and
)Ximately 87% the appropriate analytical techniques for reaching sound behavioral recon-
3unn and col- structions from ancient archaeological bone assemblages. Contrasts between
eir entirety or the ethnoarchaeological observations of carcass transport by the Hadza and
iousl y for one the two bone assemblages from Hadza base camps may result from the
1nd below for following.
is reasonably First, Hadza "behavior may be different when no anthropologists are pres-
1 3 are viewed ent, which was the situation during the formation of both the BA-1 and SN-A
-3 for zebra at base campsites (cultural anthropologists and ethnoarchaeologists are familiar
Figure 9-4 for with this millstone). For example, an anthropologist's perceived interest in
~epresentation bones and butchery might unintentionally prompt Hadza informants to con-
~ciall y evident duct butchering activities mainly in the presence of anthropologists, or it
might prompt the opposite reaction. For anthropologists, the alternative of
showing no interest in carcasses and bones leads to a reliance on informant
interviews for skeletal data, which may often be less precise than direct
s observation.
:al patterns of Second, the Hadza damage and destroy less durable bones (such as verte-
Lern here may brae, ribs, and some limb epiphyses) during butchery and consumption. It
1641 H. T. Bunn
PEL 1/2 100 PEL 1/:

SCAP CRI
CRA FEI
HUM HUI
FEM Tl
MAN 1/2 MAN 1I
LUM RAI
CERV SCA
RAD CER
TIB THO I
SAC ME
THOR LU~
MET SA(·
RIB Rll
0 20 40 60 80 100
Figure 9-4. Representation in percent (%MNU) of impala bones at SN- Fig

A, a Hadza bas,;: camp. eler,
certainly is a significant factor, given that the Hadza typically use small metal I have also
axes to chop spongier elements, especially vertebrae and pelves; into frag- useful measur
ments that will fit into cooking pots for boiling and grease extraction. That that the unde
makes it difficult to incorporate most fragments into estimates of abundance. base" sites inc
The Hadza also gnaw and consume bones (see Oliver, this volume). for prolonged
Third, hyenas, jackals, and other scavengers consume or remove bones after an observed c
the Hadza abandon a site, if the bones are still fresh enough to be palatable. hunter-gatherE
The attraction that recently abandoned bone food refuse at human sites pro- the hunters ab
vides for scavenging carnivores is widely recognized, and measures to over- bro-vvn hyena~
come this biasing effect on proportions of limb and other elements have been limb epiphysE
defined and implemented in the analysis of bone assemblages from the Plio- weeks or mo1
Pleistocene sites at Olduvai Gorge, Tanzania, and Koobi Fora, Kenya (Bunn bones from sc
1982, 1986; Bunn et al. 1980; Bunn and Kroll 1986, 1988). With limb bone scavengers. Li:
representation, these measures entail (1) a thorough effort to incorporate limb because, with
shaft specimens, which are much less susceptible than epiphyses to carnivore based on abun
scavenging after processing by humans for meat and marrow, into measures If the poten1
of original limb proportions and (2) the use of general taxonomic categories, the underrepn
such as families or mammal size groups. Comparable methods that emphasize least with bra
the denser, more durable portions of other skeletal elements can also be used. epiphyses ratl
Carcass Transport j165
~ 100 PEL 1/2

.6 CRA
FEM
HUM
TIB
MAN 1/2
RAD
SCAP
CERV
THOR
MET
LUM
I
/SAC
RIB
----., 20 40 60 80 100
0
100
zla bones at SN- Figure 9-5. Representation in percent (%MNU) of different skeletal
elements of all taxa at SN-A, a-Hadza base camp.
1se small metal I have also argued that the proportions of limb epiphyses can provide a
lves, into frag- useful measure of the intensity and duration of occupation by humans and
xtraction. That that the underrepresentation of limb epiphyses at Plio-Pleistocene home 11
of abundance. base" sites indicates that they were not continuously occupied by hominids
me). for prolonged periods (Blinn 1987, 1991; Bunn et al. 1988). This follows from
ove bones after an observed contrast in limb representation between short-term Kua San
o be palatable. hunter-gatherer sites in the Kalahari Desert, where bones- are still fresh when
Lman sites pro- the hunters abandon the site and limb epiphyses are extensively scavenged by
asures to over- brown hyenas, and long-term Hadza base camp sites, where scavenging of
ents have bee:h limb epiphyses. is much less pronounced. By their sustained presence for
from the Plio- weeks or months at base camps, the Hadza unintentionally protect most
' Kenya (Bunn bones from scavengers u.ntil the bones have lost their nutritional appeal to
Tith limb bone scavengers. Limb representation in the Hadza bone assemblages bears this out
.corporate limb because, with rare exception, minimum number of element (MNE) values are
es to carnivore based on abundant epiphysis specimens rather than on limb shafts.
into measures If the potential for scavenger bias at the Hadza base camps is reduced-, then
mic categories, the underrepresentation of vertebrae and ribs is unlikely to result from it. At
that emphasize least with brown hyenas in the Kalahari, scavenging was focused on limb
1 also be used. epiphyses rather than on axial parts, although some experimental observa-
166IH. T. Bunn
tions of spotted hyenas indicate alternative scavenging sequences are possible Tanzania Com1
(Bunn 1986; Bunn and Kroll 1988; Capaldo 1990; Mareah, personal commu- their valuable
nication 1991). coprincipal inv
Fourth, trampling, subaerial weathering, and postdepositional sediment O'Brien, and Ji
profile compaction may eliminate less durable skeletal ele~ents. Subaerial my Anthropol<
weathering may have eliminated bones that the Badza accumulated decades from discussio
earlier, but the bones at the studied base camps had only been subjected to Jean Hudson,-
two to three years of exposure before analysis. Sediment profile compaction earlier versions
has recently been emphasized as a major determinant of skeletal proportions
at archaeological sites, with destruction of less durable elements, such as
vertebrae and some epiphyses (Grayson 1989; Klein 1989). The biasing effects
-No
that profile compaction may produce at archaeological sites can be overcome 1. Minimum r
by the same methods cited above in the discussion of carnivore scavenging of labeled MNI (m
bones. At the Badza sites, moreover, the bones were not subjected to sediment Klein 1976) or 1
profile compaction, -and yet they display some of the same characteristics, for more consistent
example, the underrepresentation of vertebrae. That suggests that the other minimum numl
processes outlined above may have a more primary role than profile com- skeletal part a b1:
paction in modifying skeletal proportions at archaeological sites. Further = lumbar; PEL 1
study of these primary taphonomic processes- in modern contexts will cer- HUM=humeru
tainly enhance reconstructions of ancient behavior from skeletal proportions 1/2 = halfmand
at archaeological sites.
ReJ
Conclusion
Binford, L. R.
Carcass transport by the Badza is complex, and some of the 1981 Bones:
determining factors woul_d be invisible in archaeological contexts. Based on 1984 Faunal
Brain, C.K.
the largest available data set, the typical Badza practice is to try to transport
1981 The H
essentially entire field-butchered carcasses to base camps for further,process- University
ing and consumption; exceptions involve the very largest carcasses,/~arcasses Bunn,H. T.
with poor-quality marrow, and attitudes of particular Badza about sharing 1982 Meat-e
fat-rich marrow. The data on carcass transport indicate that bone assemblages of Plio-Plei
at Badza base camps should, at least initially, contain high frequencies of Departmen
almost the entire range of skeletal elements. Yet bone destruction by the 1986 Patter:
Badza at their base camps significantly modifies skeletal proportions. The Olduvai Gc
contrast between the high incidence of vertebrae in the Badza data on carcass (8):673-690
transport and the low incidence of vertebrae in the bone assemblages from 1987 Plio-P
Badza base camps reveals this process particularly well. Defining what these Congress
taphonomic processes and effects are should enable archaeologists to identify Protohistor
carcass transport from high limb proportions at archaeological sites. 1991 A Tap
Revie-w of A
Bunn, H. T., L. I
Acknowledgments 1988 Varial
Scavengin~
I thank Jean Hudson for· inviting me to present a paper at the From 457.
Bones To Behavior conference and in this volume. Field research was funded Bunn, H. T., J. V
by a National Science Foundation grant in 1986 and by a University of K. Behrensmeye
1980 FxJj 51
Wisconsin Graduate School Faculty Research grant in 1988. I thank the
12:109-136.
es are possible Tanzania Commission for Science and Technology for research clearance. For
sonal commu- their valuable participation in the field research, I thank Ellen Kroll, the
coprincipal investigator in 1986, and Larry Bartram, Christian Gurney, Chris
)nal sediment O'Brien, and Jim Oliver. For dedicated laboratory work, I thank students in
:nts. Subaerial my Anthropology 391 course, Bones for the Archaeologist. I have benefited
1lated decades from discussions with the participants in' this research project, and I thank
n subjected to Jean Hudson, two anonymous reviewers, and Ellen Kroll for comments on
le compaction earlier versions of this paper.
al proportions
1ents, such as
biasing effects Note
1 be overcome
1. Minimum number of units (MNU) is the same value that other researchers have
scavenging of labeled MNI (minimum number of individuals by skeletal part; e.g., Binford 1981;
ed to sediment Klein 1976) or MAU (minimum animal units; e.g., Binford 1984). The MNU label is
~acteristics, for more consistent with conventional usage of other minimum number labels (e.g.,
that the other minimum number of individuals, minimum number of elements). The following
1 profile com- skeletal part abbreviations are used in Figures 9-1 through 9-5: SAC= sacrum; LUM
sites. Further = lumbar; PEL 1/2 = half pelvis; THOR = thoracic; CERV = cervical; SCAP = scapula;
.texts will cer- HUM = humerus; RAD = radius; FEM = femur; MET = metapodial; TIB = tibia; MAN
al proportions 1/2 =half mandible; CRA =cranium; RIB= rib.
/
References
Binford, L. R.
l some of the 1981 Bones: Ancient Men and Modem Myths. Academic Press, New York.
exts. Based on 1984 Faunal Remains from Klaszes River Mouth. Academic Press, Orlando~ Fl.
Brain, C.K.
'Y to transport
1981 The Hunters or the Hunted? An Introduction to African Cave Taphonomy.
trther process- University of Chicago Press, Chicago.
sses, carcasses Bunn,H. T.
about sharing 1982 Meat-eating and Human Evolution: Studies on the Diet and Subsistence Patterns
.e assemblages of Plio-Pleistocene Hominids in East Africa. Unpublished Ph.D. dissertation,
frequencies of Department of Anthropology, University. of California, Berkeley.
uction by the 1986 Patterns of Skeletal Representation and Hominid Subsistence Activities at
)portions. The Olduvai Gorge, Tanzania, and Koobi Fora, Kenya. Journal of Human Evolution 15
lata on carcass (8):673-690. . .
·mblages from 1987 Plio-Pleistocene Hominid Land Use Patterns. Paper presented at 11th
.ng what these Congress of the Union Internationale des Sciences Pn§historiques et
ists to identify Protohistoriques, Mainz, Germany.
ites. · 1991 A Taphonomic Perspective on the Archaeology of Human Origins. Annual
Review of Anthropology 20:433-467.
1988 Variability in Bone Assemblage Formation from Hadza Hunting,
Scavenging, and Carcass Processing. Journal of Anthropological Archaeology 7:412-
1er at the Fron1 457.
:h was funded Bunn, H. T., J. W. K. Harris, G. Isaac, Z. Kaufulu, E. Krolt K. Schick, N. Toth, and A.
University of K. Behrensmeyer
L I thank the 1980 FxJj 50: An Early Pleistocene Site in Northern Kenya. World Archaeology
12:109-136.
168jH. T. Bunn
Bunn, H. T., and E. M. Kroll

1986 Systematic Butchery by Plio/Pleistocene Hominids at Olduvai Gorge,
Tanzania. Current Anthropology 27:43 J-452.
1988 Reply to Fact and Fiction about the Zinjanthropus Floor: Data, Arguments,
and Interpretations by Lewis Binford. Current Anthropology 29 (1):135-149.
Bunn, H. T., E. M. Kroll, and L. E. Bartram ·
1991 Bone Distribution on a Modem East African Landscape and Its Archaeolog-
ical Implications. In Cultural Beginnings: Approaches to Understanding Early 10. D
Hominid Life-Ways in the African Savanna, edited by J. D. Clark, pp. 33-54.
Romisch-Germanisches Zentralmuseum, Mainz, Germany.
s,
Bunn, H. T., E. M. Kroll, C. J. O'Brien, J. S. Oliver, C. F. Gurney, and L. E. Bartram
1993 New Data on Carcass Transport and Bone Assemblage Formation by Ja]
Hadza Hunter-Gatherers. Ms. in possession of authors.
Capaldo, S.
1990 Paper presented at Society of Africanist Archaeologists meeting, ln1
Gainesville, Fl.
Grayson, D. K.
1989 Bone Transport, Bone Destruction, and Reverse Utility Curves. Journal of
Th
interpretation
. Archaeological Science 16:643-652. with particul;
Klein, R. G.
human behav
1976 The Mammalian Fauna of the Klasies River Mouth Sites, Southern Cape
sions: (1) the i~
Province, South .Africa. South African Archaeological Bulletin 31:75-98.
logical record
1989 Why Does Skeletal Part Representation Differ Between Smaller and Larger
Despite the g
Bovids at Klasies River Mouth and Other Archaeological Sites? Journal of Archae-
here see the s
ological Science 16:363-381. ongoing disp·
Lartet, E., and H. Christy
uniformi tariar
1865-75 Reliquiae Acquitanicae: Being Contributions to the Archaeology and
living world, ·
Paleontology of Perigord and Adjoining Provinces of Southern France, edited by T. R.
tion can be in'
Jones. Williams and Norgate, London.
Given that c
0' Connell, J. F., K. Hawkes, and N. Blurton Jones
1988 Hadza Hunting, Butchering, and Bone Transport and Their Archaeological 1.
Implications. Journal of Anthropological Research 44:113-161. pr
1990 Reanalysis of Large Mammal Body Part Transport among the Hadza. rec
Journal of Archaeological Science 17:301-316. lar
Perkins, D., and P. Daly we
1968. A Hunters' Village in Neolithic Turkey. Scientific American 219(11):97-106. ciz
White, T. E. in!
1952 Observations on the Butchering Technique of Some Aboriginal Peoples: I. 2.
American Antiquity 17:337-338. a£1
1953 Observations on the Butchering, Technique of Some Aboriginal Peoples, ap
No.2. American Antiquity 19:160-164.
1954 Observations on the Butchering Technique of Some Aboriginal Peoples, The papers
Nos. 3, 4, 5, and 6. American Antiquity 19:254-264. grouped for d
ies of prey se:
and applicatic
tive archaeoll
From Bones to BeJ
Faunal Remains, E
Paper No. 21. ~
reserved. ISBN 0
iuvai Gorge,
, Arguments,
35-149.
; Archaeolog-
:anding Early 10. Discussion: Subsistence and
c, pp. 33-54.
Settlement Interpretations
. Bartram
'ormation by James F. O'Connell
sts meeting,
Introduction
res. Journal of
These papers are all concerned with the same general problem, the
interpretation of variation in archaeological bone assemblage composition,
with particular attention to its implications for the reconstruction of past
mthern Cape human behayior. In principle, the problem has two interdependent dimen-
L sions: (1) thr identification of behaviorally significant variation in the archaeo-
:r and Larger logical record itself and (2) the development of frameworks for interpreting it.
vnal of Archae- Despite the generally inductive nature of the discipline, most contributors
here see the second as more immediately critical. They also agree (despite
ongoing dispute elsewhere over the utility of analogy and the validity of
haeology and uniformitarian approaches) that interpretive frameworks are best built i~ the
iited by T. R. living world, where behavior, its determinants, and its archaeological reflec-
tion can be investigated simultaneously.
Given that consensus, two broad sets of questions emerge:
rchaeological 1. How do modem humans, especially those who ~ely on wild animal
prey for a substantial part of their subsistence, create an archaeological
; the Hadza. record? How do they· select their prey, process it, move it across the
landscape, consume it, damage the bones in the process and discard the
waste? What are the archaeological consequences of theiractions, espe-
(11):97-106. cially with respect to site distribution, bone assemblage composition,
intrasite spatial organization, and bone damage morphology?
tal Peoples: I. 2. What factors determine this behavior? How are they likely to have
affected it in the past? How is an understanding of their effects best
inal Peoples, applied to the resolution of specific archaeological problems?
inal Peoples, The papers under review speak to several of these is~:mes. They can be
grouped for discussion under four headings: ethnographic or actualistic stud-
ies of prey selection, processing and transport, intrasite spatial distribution,
and applications of current knowledge of these (and other) topics to substan-
tive archaeological problems. Most are more concerned with identifying
169
170 [ f. F. O'Connell
patterns and possible analogues in the living world than with explaining them 1988; Lyman J
in nomothetic terms. Despite their descriptive merit, this- emphasis ultimately base camps, <
limits their utility with respect to substantive problems in prehistory. It is a transport reliE
criticism not only of these particular works but also of the field at large (see portion of ele:
also Gifford-Gonzalez 1991). and (2) for an:
are axial part:
Prey Selection part utility an
Binford's (197
Anthropologists have long been interested in the origin and evo- ous parts of c;
lution of large mammal hunting among humans . Many have argued that it ing aspects oJ
conditions other aspects of behavior (including the sexual division of labor, element selec
group size and composition, patterns of seasonal movement and territoriality) equal), stron~
and their variability through time. Certainly it determines the initial structure archaeologica:
of archaeological bone assemblages. Subsequent
Steele and Baker present the only paper in the volume directly concerned and Jones (19t
with explaining prey selection among humans. They focus on "multiple pre- are unnecessc
dation," defined as "the near instantaneous killing·of more than one prey 9r transport patt
the serial killing of individual prey." This is characterized as a distinctive 1990) report t
feature of human hunting, one of a set of strategies "designed to procure the species, contn
greatest amount of animal resources in the most efficient manner possible." appendicular
The paper's principal strengths are its comparative, cross-taxonomic perspec- elements are s
tive and its appeal to theory and models from evolutionary ecology; its main notion (as m<
weakness, the unsupported assumption about efficiency. Though widely tissue alone d1
' i
shared, this idea deserves critical evaluation. · size and distc
A recent paper by Hawkes and colleagues (1991) on a related topic, special- transported, j
ized large mammal hunting, shows the problematic quality of "efficiency" in argue that ele
this context. Hawkes and her colleagues compared caloric returns from large inedible tissw
mammal hunting by Hadza foragers with those potentially available from the implicatio
small mammal hunting and trapping in the same habitat. Though lOng-term manner consi
averages for the first are significantly higher, variation in individu~l success Barlow 1992).
rates makes it a risky strategy, not the best one to pursue if the goal were to Papers by I
provide one's dependents with animal tissue on a daily basis, at least in that this general p
habitat. From the perspective of individual hunters, small mammal hunting tices, all con tJ
and trapping, combined with opportunistic large animal hunting and the whole carcas:
appropriation of meat from other hunters, meet this goal more efficiently. representatior
Nevertheless, Hadza men remain specialized big game hunters. As Hawkes are left at the J
(1990, 1991) has argued elsewhere, their objective may be to attract favorable 200 kg) and th
attention to themselves as successful hunters, thereby increasing their oppor- where appen1
tunities for access to women and ensuring better treatment for their offspring. hunter's or cal
From this perspective, the key question about multiple predation may be his or her a tteJ
about goals, not efficiency. Bunn does r
patterns he dE
identified· or
Processing and Transport
important det
Archaeologists have long recognized variation in skeletal element made similar
representation in faunas recovered from prehistoric sites and interpreted it by combined dab
appeal to differences in element density and resistance to attrition (Grayson The key iss1
Subsistence and Settlement 1171
laining them 1988; Lyman 1984), patterns in eleme.nt selection for transport from kill sites to
,is ultimately base camps, or a combination of the two. Until recently, arguments about
.story. It is a transport relied heavily on T. E. White's (1952) propositions that (1) the pro-
at large (see portion of elements removed from a kill varies inversely with prey body size
and (2) for any given taxon, limb elements are more likely to be removed than
are axial parts, both patterns operating as a function of differences in body
part utility and transport costs. Neither proposition was actually tested until
Binford's (1978) research among the Nunamiut. Binford showed that (1) vari-
sin and evo- ous parts of caribou and sheep could be scaled on a series of indices measur-
~gued that it ing aspects of nutritional utility and (2) the scales predict the rank order of
ion of labor, element selection for transport and, by extension (all other things being
:erri tori ality) equal), strongly condition the relative frequency of skeletal parts found in
tial structure archaeological assemblages.
Subsequent work has modified those results in two ways. First, Metcalfe
y concerned and Jones (1988; see also Jones and Metcalfe 1988) show that Binford's indices
nultiple pre- are unnecessarily complex; simpler measures of utility predict Nunamiut
one prey or transport patterns equally well. Meanwhile, O'Connell and colleagues (1988,
a distinctive 1990) report that among the Badza (1) transport patterns differ across prey
1procure the species, contradicting White's generalization that hunters preferentially move
2r possible." appendi)illar parts from kill to· base across all large-bodied taxa; (2) some
mic perspec- elements' are stripped of edible tissue prior to transport, challenging Binford's
)gy; its main notion (as modified by Metcalfe and Jones) that weight of attached edible
'ugh widely tissue alone determines transport rank; and (3) transport costs (including prey
size and distance from kill to base) determine the proportion of elements
)pic, special- transported, just as White had predicted. O'Connell and colleagues further
fficiency" in argue that element rank for transport is determined by the ratio of edible to
.s from large inedible tissue per element and by the costs of removing one from the other,
ailable from the implication being that they should vary by el~ment across prey taxa in a
;h long-term manner consistent with observed· transport patterns (see also Metcalfe and
dual success Barlow 1992).
goal were to Papers by Bartram, Bunn, Emerson, and Jones address various aspects of
least in that this general problem. Bunn makes three points about Badza transport prac-
mal hunting tices, all contrary to the work of O'Connell and colleagues (1988, 1990): (1)
ing and the whole carcasses are typically transported intact, implying no differential
~ efficiently. representation of body parts as a function of transport costs; (2) where parts
As Hawkes are left at the kill, the animal involved is generally large (mean adult weight>
tct favorable 200 kg) and the parts left are more likely to be axial than appendicular; and (3)
their oppor- where appendicular parts are left in the field, the explanation involves a
~ir offspring. hunter's or carrier's monopolization of marrow for personal consumption, not
:ion may be 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
etal element made similar claims before (Bunn et al. 1988), but subsequent analysis of the
rpreted it by combined. data sets (O'Connell et al. 1990) showed them to be unsupported.
)n (Grayson 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 interesting (
part transport. O'Connell and his colleagues' analysis indtcates that transport among other
patterns vary from time to time, place to place, sample to sample, depending makes it diff
on such factors as the number and type(s) of prey taken, the number· of has observec
carriers available, and the distance(s) from kill(s) to base c~p(s). In short, assemblages
there is no "typical" pattern. This is an asset, not an obstacle, to analysis. It role in huma
enables one to explore alternate explanations for observed variation in carcass predict from
processing and transport within a single ethnographic case, in this instance, mals should
among the Hadza. Attempting to identify the normal_ or modal Hadza pattern resulting wa
in one or another sample of cases effectively precludes this. how they n
Bartram's paper on gemsbok butchery and transport by the Kua makes (including tl
three main points: (1), varying, sometimes high, numbers of bones are stripped possible. Fe·
of meat and other edible tissue and left at kill sites, probably as a function of among modE
weight reduction for transport (cf. Bunn, this volume); (2) as a result, simple Hudson.1991
utility indices based on the amount of meat, marrow, and grease originally achievement
associated with various elements do not accurately predict their probability of tant in that t
transport; and (3) element transport patterns do not-reflect a body part rank- driven work
ing of any kind. Though the first two observations are not unprecedented, the 1991; Hill et'
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
s
least strongly conditions transport (e.g., Binford 1978; Emerson 1990, this vol- T
ume; Metcalfe and Jones 1988; O'Connell et aL 1990). If Bartram were right, research on
the prospects for interpreting archaeologically observed differences in body pioneering v
part representation in terms of transport by hunters might well be quite refuse in an
limited. The situation he reports is complicated in that some assemblages in Gonzalez 191
his sample were ravaged by carnivores prior to collection. Until this effect is the investig1
controlled by analyzing whatwas actually transported rather than rEj'covered archaeologic.
ethnoarchaeologically, one cannot be certain whether the absence oH~~vidence to transport i
for a transport scale is real or a function of secondary attrition. a well-founc
Emerson's report on bison body part utility is an important contribution. aspects of si
Her observation that parts may be scaled quite differently depending on the analogy, an E
currency used to measure nutritional benefit is particularly instructive. It sug- Enloe and
gests that very different patterns in body part assemblage composition might do not addn
result from differential transport, depending on the hunter's goals and the because of j
condition of the carcass on encounter. It would be interesting to know how determinanb
processing costs might affect those decisions, but Emerson's experiments were age morpho
undertaken before it became apparent such costs might be critical. Her criteria that c
attempt to apply her bison data to the empirical problem of variability in (Binford 198
carcass transport by the Badza, specifically in cases involving giraffe, is in- set and pro'
triguing; but one may be skeptical about its results, partly because parts of archaeologic
animals as morphologically different as bison and giraffe are unlikely to be which forag:
ranked in sufficiently similar ways on any scale and partly because of the those strateg
absence of data on processing costs. . associated w
Jones describes small marnmal transport and its implications for bone representatic
assemblage composition at a base camp created by Ache foragers. Though I examples pn
!rn" of body interesting (not least because of the contrast with well-documented cases
.at transport among other groups involving larger prey), the anecdotal quality of the work
, depending makes it difficult to apply beyond the Ache context. As Jones (1984) himself
number of has observed, small mammals are an important part of many archaeological
:s). In short, assemblages from the late Pliocene onward, yet uncertainty persists over their
, analysis. It role in human diets (e.g., Grayson 1991). It would be useful if one were able to
min carcass predict from general principles the circumstances under which small mam-
1is instance, mals should be taken, how they should be processed for consumption and the
tdza pattern resulting waste discarded, what archaeological patterns would result, and
how they might be distinguished from the effects of other processes
Kua makes (including the actions of nonhuman bone accumulators); but this is not yet
are stripped possible. Few comprehensive descr"iptions of small mammal exploitation
L function of
among modern hunters have been undertaken with these goals in mind (cf.
!Sult, simple Hudson 1990, this volume; Yellen 1991a, 1991b), and none yet enable their
;e originally achievement. A detailed analysis of the Ache data could be especially impor-
robability of tant in that they can be placed in the context of a large body of theoretically
y part rank- driven work on foraging, prey selection and food sharing (e.g., Hawkes 1990,
edented, the 1991; Hill et al. 1987; Kaplan and Hill1985) .
. processing
tram's third
settings that Site Structure
rmines or at
~90, this vel- The past 15. years have witnessed many reports of actualistic
were right, research on hunter-gatherer site structure, beginning with Yellen's (1977)
1ces in body pioneering work on the !Kung. Few, however, treat the distribution of bone
ell be quite refuse in any detail (cf. Bartram et al. 1991; Binford 1983, 1987; Gifford-
emblages in Gonzalez 1989). Even they involve little analysis of the kind needed to guide
this effect is the investigation of patterns in bone distributi~n that might be observed
n recovered archaeologically, other than in general terms. Unlike the situation with respect
of evidence to transport and processing, archaeologists currently lack even the promise of
a well-founded theoretical basis for predicting form and variation in most
ontribution. aspects of site structure. Their only interpretive tool is direct ethnographic
tding on the analogy, an essentially nonexplanatory basis for argument. ·
ctive. It sug- Enloe and Jones make important contributions to the growing literature but
sition might do not address the central problem. Again, Jones's single case is interesting
)als and the because of its. simplicity but difficult to employ analytically because its
) know how determinants are unclear. Enloe analyzes body part representation and dam-.
iments were age morphology in three Nunamiut middens with the goal of identifying
~ritical. Her criteria that distinguish storage-dependent collectors from nonstoring foragers
ariability in (Binford 1980). He deserves credit for tackling an initially unpromising data
;iraffe, is in- set and producing some intriguing results. In order to make use of them
use parts of archaeologically, one needs to know more about the circumstances under
1likely to be which foraging or collecting might be expected, how sites associated with
cause of the those strategies are likely to be organized internally, and why faunal remains
associated with various areas within sites display the patterns in body part
ns for bone representation and damage morphology they do, at least in the Nunamiut
ers. Though examples presented.
174 I J. F. O'Connell
Archaeological Applications another que

circumstano
Ultimately, the payoff for. actualistic research comes with the define the lil
application of its results to substantive problems in prehistory. Two papers in and Knecht':
the sections under review tackle this difficult challenge. . implication.
The first, by Pike-Tay and Knecht, deals with aspects of subsistence and determinant
technology in the Upper Paleolithic of southwestern France, the goal being to ologists have
determine· whether they represent the activities of "opportunistic" foragers or In the last
"specialized" collectors. The exercise is ambitious ~d potentially important body part're
but open to criticism on two grounds. First, as Pike-Tay and Knecht see it, tals at sites
foragers choose prey opportunistically as a function of their local abundance, control data
while collectors specialize in taking particular species from among an avail- and nonhurr
able array, essentially as a function of cost-benefit considerations. The distinc- ing ambush
tion differs from that originally proposed by Binford (1980)~ who defined prey. Thee
reliance upon stored foods as the key criterion. It is also inconsistent with important.
standard ecological models of prey selection, which emphasize the impor- Impressiv
tance of cost/benefit considerations for all predators and downplay the role of first involvi
abundance alone as a predictor of selection under most circumstances (e.g., ethnoarchae
Bettinger 1991; Stephens and Krebs 1986). an importan
A second problem involves the archaeological correlates of foraging and for any one
collecting. Among collectors, Pike-Tay and Knecht expect "reliable" technol- more than c
ogy, defined by "the presence of such weapons as barbed harpoons or points, habitat-sped
or a multiplicity of points"; for foragers; "maintainable" technology and the meaning to 1
absence of evidence for "large [scale] game drives or intensive high-yield transport ta'
hunting episodes." The validity of those and other criteria mentioned are nonhuman I
open to question. For example, both Hadza and Alyawarra foragers use barb- modern hm
ed projectile weapons; Hadza hunters routinely carry as many as three differ- Nunamiut i
ent types of arrows. Similarly, Great Basin ethnography and archaeology are sample relat
rich in evidence of large-scale game drives, all by people acting as foragers, at though the I
least in the seasons when the drives took place (e.g., Egan 1917:238-239; ical contexts
Steward 1938). I data here an
Though Pike-Tay and Knecht do not develop the argument, thete are at one might a~
least two reasons to be interested in the forager-collector contrast in the poorly unde
European Paleolithic, one good, the other not. The latter, mentioned at the (at least as F
conference and elsewhere, entails the view that collecting might be a critical patterns, eitl
indicator of what Binford (1989) calls "planning depth," the ability to concep- The secon'
tualize and meet subsistence needs over long periods of time through food Despite the
storage. Binford himself has appealed to this notion in distinguishing Middle exercise. Var
from Upper Paleolithic subsistence patterns and, by extension, archaic sapi- tral tendenc:
ens' cognitive abilities from those of modern humans. A moment's reflection paints for in
undercuts the argument. The fact that modern human foragers do not store grounds of s
food does not imply they lack the ability to think far ahead: witness the time
depth in marriage arrangements among traditional Australian Aborigines, likely to be
most of whom were foragers. Conversely, the habit of storing food, found in reasonable a
many nonhuman organisms (Vander Wal11990), does not necessarily mean tal foraging
they think like we do. ance, both v
The other reason turns on the observation that Europe was unoccupied Upper Paleo
during periods of full glacial climate prior to the most recent cycle (Gamble this issue m
1986). Following Binford's argument, the difference might well involve re- central tend~
liance on storage. Whether it also involves some new conceptual ability is predator-sea
another question. In any case, one should be able to predict the ecological
circumstances under which storage becomes the optimal subsistence strategy,
es with the define the likely archaeological indicators, and run the test. In short, Pike-Tay
vo papers in and Knecht's question has an interesting and extremely important substantive
implication. Addressing it successfully will require more knowledge of the
;istence and determinants of storage and their archa~ological implications than archae-
~oal being to ologists have at present.
I foragers or In the last piece under consideration, Stiner reports two distinct patterns in
y important body part representation for red deer and aurochs accumulated by Neander-
necht see it, tals at sites in west-central Italy. She interprets them by comparison with
abundance, control data from bone assemblages accumulated by other predators (human
ng an avail- and nonhuman) as the product of two distinct foraging strategies, one involv-
The distinc- ing ambush hunting, the other scavenging of selected parts from winter-killed
rho defined prey. The exercise is original; the results intuitively appealing and very
sistent with important.
! the impor- Impressive as it is, Stiner's analysis can be challenged on two points, the
ly the role of first involving internal consistency. Though one might have expected the
:tances (e.g., ethnoarchaeologically observed bone accumulations of modern hunters to be
an important reference dimension, Stiner rejects them on grounds that data
:>raging and for any one case are too few to be informative. Lumping information from
>le" technol- more t~an one J/would obscure a diverse array of ecological contexts and
rrs or points, habitat..specific foraging strategies across the world without providing ... any
ogy and the meaning to the madness it would create." Hunter-gatherer prey selection and
~ high-yield transport tactics are also said to be less well understood than those of the
!n tioned are nonhuman predators she employs for comparison. While many data sets on
~rs use barb- modern hunter-gatherers are indeed quite small, Binford's (1978) on the
three differ- Nunamiut includes information on more than 275 prey animals, a large
taeology are sample relative to most in the nonhuman suite used in the analysis. Further,
; foragers, at though the point about the problems of lumping data from different ecolog-
117:238-239; ical 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
there are at one might agree that variati01,1in human bone accumulation practices remains
.trast in the poorly understood, it seems unlikely to make much difference to the analysis
loned at the (at least as presented here) since explanations of variation in the nonhuman
be a critical patterns, either between or within taxa, play little role in the argument.
y to concep- The second objection is essentially the one raised above about Bunn's paper.
lrough food Despite the occasional reference to explanation, this is a pattern recognition
hing Middle exercise. Variance observed in the modem world is reduced to a series of cen-
trchaic sapi- tral tendencies· "typical" of each bone accumulator; these serve as reference
t' s reflection points for interpreting the Neandertal samples. The approach is justified on
do not store grounds of scale: the fundamental niche is the target of interest, and all that is
,ess the time likely to be reflected archaeologically, at least in this time range. This is a
Aborigines,
:>d, found in reasonable argument as far as it goes. But if Stiner's inferences about Neander-
:saril y mean tal foraging are correct, the problem then becomes accounting for the vari-
ance, both within the Middle Paleolithic sample, and between that and the
unoccupied Upper Paleolithic. Again we confront the problem of explanation. Addressing
de (Gamble this issue must attend to the variance subsumed in the characterization of
. involve re- central tendencies in prey selection and bone accumulation among modern
tal ability is predator-scavengers, human and nonhuman.
176 I ]. F. O'Connell
Summary 1983 In F
1987 Res
Recent research on the composition of faunal assemblages repre- and Theo;
sents some of the best work now being done in archaeological method and Kent, pp
theory. The papers reviewed here reflect some of the strength _of this research, 1989 Isol
especially in their reliance on the living world as a source of information preach.:
about the past. The key problem entails their use of that information. Analysts Pleistoce1
here and elsewhere have generally been concerned with identifying "typical" Cambrid
patterns in faunal assemblage composition, as observed ethnographically or Bunn, H., L. I
/l.ethologically," and using them as the basis for interpreting variation in 1988 Var
assemblage composition in the past. Stiner's work shows some of the gains to · ing and 1
be made from this approach; the ongoing argument about bone transport Egan, H.
among the Hadza suggests its potential pitfalls. However successful it may be, 1917 Pi01
it has a limit: some patterns represented in ~he past are not matched in the Estate, R
present, and even the "matches" (as in the Italian Mousterian case) create Emerson, A.
problems of explanation. Either way, one confronts the problem of accounting 1990 Arc
for archaeologically observed variability. bison. Pl
As most contributors to this volume would agree, the living world is the versity,:
place to develop the necessary frameworks for explanation, primarily because Gamble, C.
potentially pertinent variables can be observed there directly. Questions of 1986 The
prey selection, processing and transport, and their implications for variation Cambric
in assemblage composition s,eem especially tractable, partly because they lend Gifford-Goru
themselvE~s to quantitative cost/benefit analysis and partly because they can 1989 Mo
be tackled yvith the use of theoretical models developed in other fields, cation, e•
notably behavioral ecology. Site structure presents a more challenging prob- the First
lem, primarily because of the current absence of any such models. It is impor- 1991 Bm
tant that analysts pay more attention to those issues and that they undertake a in Zooa1
more theoretically driven approach to ethnography and ethnoarchaeology. Grayson, D.
Without that, archaeologists will be in no position to address the problems in 1988 Dal
prehistory that attracted their interest in the first place. pologicc=
/ 1991 Th1
Reamers.
Acknowledgments edited 1
Museun
Jennifer Graves, Donald Grayson, Kristen Hawkes, and Duncan Hawkes,K.
Metcalfe provided useful comments on earlier versions of this paper. 1990 Wl
Tribal ar.
Boulder
References 1991 She
and Sod
Bartram, L., H. Bunn, and E. Kroll
Hawkes, K.,:
1991 Variability in Camp Structure and Bone Food Refuse Patterning at Kua San
1991 Hu
Hunter-Gatherer Camps. In The Interpretation of Archaeological Spatial Patterning, Fora gin
edited by E. Kroll and T. D. Price, pp. 77-148. Plenum Press, New York of the Ro
Bettinger, R. Hill, K., H. K
1991 Hunter-Gatherers: Archaeological and Evolutionary Theory. Plenum Press, New 1987 Fm
York cations.
Binford, L. Hudson,J.
1990 Ad
1980 Willow Smoke and Dogs' Tails: Hunter-Gatherer Settlement Systems and
Aka Py~
Archaeological Site Formation. American Antiquity 45:4-20. Califorr
1983 In Pursuit of the Past. Thames and Hudson, London.

1987 Researching Ambiguity: Frames of Reference and Site Structure. In Method
ages repre- and Theory for Activity Area Research: An Ethnoarchaeological Approach, edited by S.
lethod and Kent, pp. 449-512. Columbia University Press, New York.
.is research, 1989 Isolating the Transition to Cultural Adaptations: An Organizational Ap-
nformation proach. In The Emergence of Modern Humans: Biocultural Adaptations in the Later
n. Analysts Pleistocene, edited by E. Trinkaus, pp. 18-41. Cambridge University Press,
.g "typical" Cambridge.
tphically or Bunn, H., L. Bartram, and E. Kroll
·ariation in 1988 Variability in Bone Assemblage Formation from Badza Hunting, Scaveng-
the gains to ing and Carcass Processing. Journal of Anthropological Archaeology 7:412-457.
e transport Egan, H.
1 it may be, 1917 Pioneering the West, 1846-1878. Major Howard Egan's Diary. Howard R. Egan
:hed in the Estate, Richmond, Utah.
~ase) create Emerson, A.
accounting 1990 Archaeological Implications of Variability in the Economic Anatomy of Bison
bison. Ph.D. dissertation, Department of Anthropology, Washington State Uni-
rorld is the versity, Pullman. University Microfilms, Ann Arbor.
·ily because Gamble, C.
uestions of 1986 The Palaeolithic Settlement of Europe. Cambridge University Press,
>r variation Cambridge.
:e they lend Gifford-Gonzalez, D.
se they can 1989 Modem Analogues: Developing an Interpretive Framework. In Bone Modifi-
ther fields, cation, edited by R. Bonnichsen and M. Sorg, pp. 43-52. Center for the Study of
tging prob- the First Americans, Orono, Maine.
It is impor- 1991 Bones Are Not Enough: Analogues, Knowledge, and Interpretive Strategies
lndertake a in Zooarchaeology. Journal of Anthropological Archaeology 10:215-254.
·chaeology. Grayson, D. ·
>roblems in 1988 Danger Cave, Last Supper Cave, and Hanging Rock Shelter: The Faunas. Anthro-
pological 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,
edited by R. Purdue, W. Klippel, and B. W. Styles, pp. 99-109. Illinois State
Museum Science Papers No. 23. Springfield.
1d Duncan 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
:; at Kua San
1991 Hunting Income Patterns among the Badza: Big Game, Common Goods,
rl Patterning,
Foraging Goals and the Evolution of the Human Diet. Philosophical Transactions
>rk.
of the Royal Society, London B 334:243-251 ~
Hill, K., H. Kaplan, K. Hawkes, and A. M. Hurtado
t Press, New
1987 Foraging Decisions among Ache Hunter-Gatherers: New Data and Impli-
cations for Qptimal Foraging Models. Ethology and Sociobiology 8:1-36.
Hudson,J.
1990 Advancing Methods in Zooarchaeology, An Ethnoarchaeological Study among the
)ystems and
Aka Pygmies. Ph.D. dissertation, Department of Anthropology, University of
California, Santa. Barbara. University Microfilms, Ann Arbor.
178 I f. F. O'Connell
T I
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<aplan, H.,. and K. Hill
1985 Food Shari~g among Ache Foragers: Tests of Explanatory Hypotheses.
III. s
Current Anthropology 26:223-245. I
Lyman,R. L. .
1984 Bone Density and Differential Survivorship of Fossil Classes. Journal of a
Metcalfe, D., and K. R. Barlow <
1992 A Model for Exploring the Optimal Tradeoff Between Field Processing and
Transport. American Anthropologist 94:340-356. ·
a
Metcalfe, D., and K. Jones
1988 A Reconsideration of Animal Body Part Utility Indices~ American Antiquity
53:486-504.
O'Connell, J., K. Hawkes, and N. Blurton Jones
1988 Hadza Hunting, Butchering and Bone Transport and Their Archaeological
Implications. Journal of Anthropological Research 44:113-162.
1990 Reanalysis of Large Mammal Body Part Transport among the Hadza. Jour-
nal of Archaeological Science 17:301-316.
Stephens, D., and J. Krebs
1986 Foraging Theory. Princeton University Press, Princeton.
Steward,J.
1938 Basin-Plateau Aboriginal Sociopolitical Groups. Bureau of American Ethnology
Bulletin 120.
Stiner,M.
1991 An Interspecific Perspective on the Emergence of the Modern;Human
Predatory Niche. In Human Predators and Prey Mortality, edited by M. Stiner, pp.
149-186. Westview Press, Boulder, Colo.
Vander Wall, S.
1990 Food Hoarding in Animals. University of Chicago Press, Chicago.
White, T. E.
1952 Observations on the Butchering Techniques of Some Aboriginal Peoples: I.
American Antiquity 17:337-338.
Yellen, J.
1977 Archaeological Approaches to the Present: Models for Reconstructing the Past.
Academic Press, New York.
1991a Small Mammals: !Kung San Utilization and the Production of Faunal
Assemblages. Journal of Anthropological Archaeology 10:1-27.
1991b Small Mammals: Post Discard Patterning of !Kung San Faunal Remains.
Journal of Anthropological Archaeology 10:152-192.
I
I
,.,!.
ogical Method
rsity of Utah,
cy. Journal of
III. Social Interaction:
Hypotheses.
Food Sharing, Preparation
es. Journal of
and Consumption, and
Questions of Gender
ocessing and and Ethnicity
can Antiquity
rchaeological
Hadza. Jour-
m Etlmology
iern Human
VL Stiner, pp.
I,
Lal Peoples: I.
ting the Past.
m of Faunal
:1al Remains.
11. Gaps in Zooarchaeological
Analyses of Butchery: Is
Gender an Issue?
Diane Gifford-Gonzalez
Abstract: Ethnoarchaeological research on bones has given considerable

attention to field processing and transport ded$ions and little attention
to subsequent subdivision, processing, and discard. Likewise, experi-
mental research has concentrated on primary butchery activities and un-
/ cooked bones, seldom addressing culinary processing. Various reasons
may exist for this research bias: the atypical nature of faunal acquisition
and processing decisions in several widely read studies, assumptions
about carcass processing by early hominids, and narrowly construed
applications of foraging theory models to modern humans. Another
cause is a cultural bias toward seeing immediate prey acquisition as the
aspect of hunting behavior most worthy of study, with social subdivision
of prey and household processing commensurately undervalued as
determinants of patterning in faunal deposits. This correlates with an
androcentric view of human foraging. Notable exceptions to those biases
indicate the strengths of an end-product-focused approach as a means of
balancing actualistic research with bones.
Introduction
Over ten years ago, I had a conversation with Lewis Binford about
the Nunamiut hunters he knew, and I was struck by a story he told me. He
had asked someNunamiut men if they usually ate the marrow from metapo-
dials of caribou at the kill site. They asked him whether it was spring (the
season of lean caribou) or fall (the season of fat caribou), how far they were
from home, and whether their wives were pregnant (Lewis Binford, personal
communication 1980). ·

181
1821 D. Gifford-Gonzalez
This story has ·stayed with me since then, saying something about zooar-
chaeological research that I could not figure out for a long time. In a way, this
v
essay springs from it. It addresses 'Y"hat I perceive has been missing from A
recent ethnoarchaeological analyses of butchery and carcass processing. The blages, I enc<
gaps alluded to in the title are unstudied topics, and be~ind them stand irtg. I needec
unheeded persons and seldom-mentioned tasks. What is missing from recent from a Dass
ethnoarchaeological analyses is serious treatment of the culinary end of the weeks and t
processi-ng spectrum and its major influence on the structure of faunal bones. Abou
assemblages. _ and jagged 1
My discussion takes two perspectives on this gap. First, I argue here that the other pastorz
full range of activities that incorporate animal resources into human nutri- Such.brea1
tion-and the material effects of such tasks on bone-are not being suffi- logical sites l
ciently studied. Specifically, culinary strategies and tactics are intimately tied other ethnoa
to the nutritional benefits derived from faunal resources, they can drive field also convinc
butchery decisions, and they determine the patterning of faunal refuse, even literature. Cc
in field situations. Second, I contend that lack of attention to the influence of samples que;
cooking and cuisine stems from a general devaluation of the work of women, bone since th
and of that of anyone who is not a hunter, in processing the bodies of animals. assemblages.
This results not in "politically incorrect" research, but methodologically exposed to 1
flawed research, in which a narrow segment of the provisioning spectrum is hundreds of
privileged for study, while the overall movement of animal bodies through a same region
subsistence system-and the impact on processing decisions at every stage of weathering.
it-is slighted. The experi
I recently (Gifford-Gonzalez 1989a) made a detailed analysis of bone modi- cooked bone
fications of some modern mammal bones assemblages created by Dassanetch al. 1988; Dav
people on the northeastern shore of Lake Turkana, in Kenya. Dassanetch 1988). Uncoo
people speak a Cushitic language and, when I worked with them, lived in for understaJ
several territorial sections around the north end of Lake Turkana, practicing a loading. But
mix of pastoralism, hunting and fishing, and cultivation. The local fi:mdscape cooked long -
contained a variety of abandoned sites, testifying to several aspeets of their a wider meth
mobile and flexible subsistence strategy. My earlier analyses focused on the My immec
activities that gave the, sites their form and produced the artifactual and (boiling) or t
dietary debris left behind (Gifford 1977, 1978, 1980; Gifford and Behrensmeyer dynamic loac
1977). My reanalysis compared Dassanetch carcass processing practices with was that at 1
those documented by other archaeologists and ethnoarchaeologists including been done in
Frison (e.g., 1970, 1974; Frison and Todd 1987), Binford (1978, 1981, 1984), bone to norn
Bunn and Kroll (1986), Bunn and colleagu.es (1988), O'Connell and colleagues detectable al
(1988), and Guilday and colleagues (1962). I was especially interested in the work has beE
effects of exposure to fire on bone breakage patterns, the impacts of cooking collagen stru
on placement and intensity of cut marks, and the effects of situationally varied argued that :
processing equipment on patterns of bone modification among one cultural "horizontal h
group. On all counts, I found virtually no comparative ethnoarchaeological moisture thr'
literature, even in subsequent publications (e.g., O'Connell et al. 1990; response of ~
O'Connell and Marshall 1989). An exception is Oliver (this volume). In the dynamic loac
process I identified what seem to me to be some serious gaps in the extant stress whole
literature on dismemberment, selective transport, and bone breakage. 1972; Bonfielc
Gaps in Zooarchaeological Analysis 1183
; about zooar- What Is the Effect o£ Cooking on Long Bone Fracture?
In a way, this
missing from Analyzing long bone breakage patterns in my Dassanetch assem-
rocessing. The blages, I encountered a gap in the literature on the effects of culinary process-
d them stand ing. I needed to explain high rates of jagged, transverse fractures on bones
1g from recent . from a Dassanetch pastoral camp. The camp was inhabited for about six
:uy end of the weeks and then abandoned. Four weeks later, I mapped and collected the
ure of faunal bones. About one-third of the long bones from the camp displayed transverse
and jagged break surfaces, as did the same proportion of long bones from
e here that the other pastoral camps sampled in the region (Gifford-Gonzalez 1989a).
human nutri- Such breaks are rare in long bone s.amples from African Stone Age archaeo-
)t being suffi- logical sites I have analyzed. Illustrations and descriptions of bone fracture in
.ntimately tied other ethnoarchaeological studies (Binford 1978, 1981; Yellen 1977a, 1977b)
:an drive field also convinced me that the breakage pattern was as yet unreported in the
tl refuse, even literature. Colleagues who worked with paleontological and Stone Age bone
te influence of samples questioned whether the long bones actually were fractures of fresh
xk of women, bone since they had never seen such breaks except on weathered bone in their
ies of animals. assemblages. I discounted weathering as a factor $ince the bones had been
hodologicall y exposed to the elements for a maximum of 10 weeks and since none of
1
lg spectrum is hundre'f's of wild animal long bones I have longitudinally monitored in the
jies through a same region (Gifford-Gonzalez 1984) displayed such fractures at any stage of
every stage of weathering.
The experimental and ethnographic literature on fracture focused on up-
of bone modi- cooked bones (e.g., Binford 1978,1981; Bonnichsen 1973, 1978, 1979; Bunn et
JY D assanetch al. 1988; Davis 1985; Johnson 1985; Mengoni-Gofialons 1980; O'Connell et al.
a. D as san etch 1988). Uncooked bones are a logical place to start since they provide a baseline
:hem, lived in for understanding properties of bone as a material when subject to dynamic
a, practicing a loading. But the literature's nearly exclusive emphasis on breakage of un-
)Cal landscape cooked long bones presented me with an hnmediate interpretive problem and
;pects of their a wider methodological question.
::>cused on the My immediate question was whether sustained heating, in either moist
rtifactual and (boiling) or dry (roasting) media, affects bones' ability to respond to later
Behrensmeyer dynamic loading when the bone is cracked for marrow. The answer, I found,
practices with was that at present we do not know because little systematic research has
;ists including been done in this area. Shipman and colleagues (1984) established that heating
~' 1981, 1984), bone to normal cooking temperature ranges does not create microscopically
.nd colleagues detectable alterations in the inorganic component of bone. However, little
·erested in the work has been done on the effects of cooking on bones' moisture content or
cts of cooking collagen structure, which might affect their tensile strength. Johnson (1985)
ion ail y varied argued that spiral fractures characterize very fresh bone, and more jagged
~ one cultural "horizontal tension failure" is typical of bone that has lost' some of its original
trchaeological moisture through exposure to the air. A voluminous literature exists on the
l et al. 1990; response of small segments of bone as a material to various stresses (static,
)1 ume). In the dynamic loading, etc.), but these studies neither aim to mimic cooking nor to
in the extant stress whole bones or bone segments (e.g., Amprino 1958; Ascenzi and Bell
kage. '1972; Bonfield 1981, 1987; Bonfield and'Li 1966; Simkin and Robin 1974).
A few articles exist on alterations to bone collagen fibrils at temperature Bt

ranges that replicate cooking (e.g., Bonar and Glimcher 1970; Richter 1986; A:
Sedlin 1965). Since collagen gives bone its resiliency under stress, the alter-
ations are especially critical to the question of whether cooking can make Li·
bones break differently after cooking. The studies cited indicate that collagen (Binford 1978,
fibers liquify with heating and then reaggregate structurally with cooling. The al. 1988, 1990;
reaggregated fibers may be shorter than the original, and full bone resiliency factors as
is not recovered. The results suggest that cooked bones may respond differ- 1.
ently to dynamic loading than uncooked ones. Experimental breakage studies 2.
by Horwitz (1987) and Black (1989) suggest this may be the case, but more 3.
controlled experimentation is needed. In fact, Oliver's chapter in this volume 4.
reports jagged, transverse diaphysis fractures on long bones that were ex- 5.
posed directly to fire by Badza processors prior to breakage for marrow 6.
extraction. They are virtually identical to those in the Dassanetch assemblages, I have argu
supporting the sequence of processing events originally proposed in my anal. . an animal's s]
ysis (Gifford-Gonzalez 1989a) . be constrainil
With the exception of the work of Richter, Horwitz, and Black, none of the dismemberm
investigations were presented by archaeologists. In fact, the situation could be employed, ac(
seen as even bleaker since Black and Horwitz ·were in part motivated to do that factors a
their studies after hearing me expatiate in a zooarchaeology course on the lack tactics used:
of experimental data on the handling of cooked bones. One must comb the
materials science and biological literature to find these citations. I 7.
8.
9.
Raw, Cooked, and Cut: Omitted Distinctions Those factors
Recent discussions of dismemberment versus filleting marks rest two especiall)
on the presumption that flesh is normally cut off when body segments are Culinary te
fresh and uncooked. Average frequencies of occurrence of cut marks pn bones transport. He
processed by modern humans have been the basis of arguments aqout early butcher aims
hominid subsistence (e.g., Bunn and Kroll 1986; Shipman 1986a). But ethno- and flesh to 1
graphic cases attest to many instances of cooking meat with the bones left in, manageable ;;
as roasts, in stews, or baked or steamed in subterranean "ovens," with flesh logical acco.u:
removal after cooking (e.g., Catlin 1959; Wissler 1910). Because meat is easier (1988) and 0'
to remove from bones after it is cooked, one could hypothesize that filleting of Dobe !Kung (
uncooked bones imposes time and energy costs that must be balanced against influenced b)
potential benefits of reducing the weight of tissue transported awav from the parts of the c
acquisition site. For groups killing ocly one large ani~al or a few"' large ani- example, bot:
mals at a time, transport costs may not be so crucial a consideration, and limb handling of cc
segments would have been transported to sites of consumption with bones (1988) in fact
included. Were marrow bones of meat-rich body segments transported and mals' axial se;
cooked in joints of meat, marrow extraction might be deferred until "dessert." technology.
More knowledgeable reconstruction of prehistoric processing patterns re- Likewise, tl
q:uires better ethnographic documentation of contrasts between cut mark camp sites is
patterns on bones cooked with the meat on and raw meat removal marks. Nunamiut, ar
There is a need as well for extensive evaluation of the nutritional benefits and pots as part o
energetic costs of various cooking and other culinary processing techniques with this tech
and technology (see Emerson, this volume). cracking of b<
temperature Butchery and Transport Decisions: Underreported

Richter 1986; Aspects
~ss, the alter-
ng can make Literature on determinants of field butchery and transport
that collagen (Binford 1978, 1981, 1984; Bunn and Kroll1986; Bunn et al. 1988; O'Connell et
tcooling.The al. 1988, 1990; O'Connell and Marshall1989; Yellen 1977a) has focused on such
)ne resiliency factors as
spond differ- 1. size of the animal, relative to that of the human processors;
akage studies 2. number of animals requiring immediate processing;
tse, but more 3; distance of the animal from the destination of its products;
1 this volume 4. number of persons in carrying party;
that were ex- 5. condition of the carcass at the time encountered; and
~ for marrow 6. time of day.
assemblages, I have argued (Gifford-Gonzalez 1989a), as have others in this volume, that
din my anal- an animal's size, anatomical structure, and nutritional composition appear to
be constraining conditions or outer parameters within which a variety of
<, none of the dismemberment and defleshing tactics and strategies may successfully be
ttion could be employed, accordingto such considerations. Ethnographic literature indicates
tivated to do that fact~J'S additional to those noted above may be critical in determining
se on the lack tactics u~ed:
ust comb the
7. gear at hand to effect field processing;
8. processing technology available at destination site; and
9. ultimate form or forms the animal products will take.
Those factors have been given less attention than they deserve, and the last
tg marks rest two especially merit discussion.
segments are Culinary techniques and technology are intimately related to rates of bone
:uks on bones transport. How an animal is disjointed and filleted depends on whether a
s about early butcher aims to produce joints of meat to roast on a fire, segments of bones
1). But ethno- and flesh to boil in a pot, boneless cuts to be sliced and dried as jerky, or
bones left in, manag~able and quickly frozen segments for winter storage. Ethnoarchaeo-
s," with flesh logical accounts by Binfqrd (1978) for the Nunamiut, Bunn and colleagues
neat is easier (1988) and O'Connell and colleagues (1988) for the Hadza, and Yellen for the
1at filleting of Dobe !Kung (1977b) indicate that kill-site butchery decisions and tactics are
anced against influenced by expectations about how the meat, marrow, and other useful
way from the parts of the dead animals will be stored or processed for consumption. For
ew large ani- example, both Bunn and colleagues (1988) and Yellen (1977b) note special
ion, and limb handling of carcass segments headed for boiling in pots. Bunn and colleagues
n with bones (1988) in fact contend that the fundamental decision to transport larger ani-
1sported and mals' axial segments is predicated on the existence of a pot-boiling extractive
Ltil"dessert." technology.
~ patterns re-
Likewise, the ultimate form of the bone debris at residential or short-term
~en cut mark camp sites is determined by cooking technology. Hadza, !Kung, Dassanetch,
noval marks. Nunamiut, and other documented modern groups all have metal or ceramic
l benefits and pots as part of their culinary equipment, and meat is often prepared in stews
tg techniques with this technology. Investigators report a range of hearthside chopping and
cracking of bone an~ bone units prior to cooking, both to facilitate fitting the
186j D. Gifford-Gonzalez
bones into cook pots and to liberate fat, marrow, and blood (e.g., Binford 1978; research and
Yellen 1977b). Among such groups, pots may be absent -from hunting stands Sim ul taneow
or other sites created by hunting parti~s, leading to different processing and atmosphere, i
damage patterns on bone assemblages produced by one group of people, in act to drive b1
functionally disparate sites. I have outlined the differing imp;:tcts of situation- (Blinn et al.
ally variable processing gear on Dassanetch mammal bone assemblages earlier work,
(Gifford-Gonzalez 1989a) and will not elaborate on this important consid- starkly repetl
eration here. mass kill situc
Among people who store animal products, the for~ of the products drives will be remo·
field butchery activities. Ethnographically documented muscle-stripping they, are frac·
tactics at mass bison kills by historic Plains Indians (Frison 1970, 1974; Frison cooking. The:
et al. 1976; Frison and Todd 1987; Johnson 1978; Wheat 1972) were aimed at stances and s1
quickly removing the most readily transported meat units .from as many The early h
animals as possible. The overriding consideration was to dry flesh for trans- concentratior
port and storage, sometimes augmented by extraction and processing of fat logical resear
(Densmore 1918; Speth and Spielmann 1983). These culinary goals produce blages in Plic
distinctive patterns of bone modification and discard at kill/butchery sites 1~81, 1983, 19
(e.g., Frison 1970; Frison and Todd 1987). Binford (1978) reports another 1990). Those·
pattern of mass-kill butchery among Nunamiut caribou hunters. Instead of goals (e.g., S
extensive muscle lifting, selective flesh removal, situationally tuned dismem- Whether of tl
berment, and transport of higher-utility, bone-bearing body segments to the assumptic
secondary caches or residential camps are the norm. in their writir
As discussed by others in this volume, patterns of wholesale extraction of were extracte
edible tissue and/ or discard of bone at mass kill sites differ from those or products a
reported for hunting peoples who kill fewer large animals at a time and at a and Kroll 19
steadier rate over the year. The ethnographic literature suggests that bones are differen tl y m1
more rarely discarded in the field when animals are encountered on a one-by- (the cower in~
one basis than when encountered in substantial numbers (e.g., Bunn et al. hunter strolli
1988; Marshall, this volume; O'ConneUet al. 1988, 1990; Yellen 1977b)jTo take knowledge, c
an archaeological example, processing marks on large ungulate bones at the employed to:
Eschelman site, a residential settlement into which single deer and elk entered gulp" scenaril
over a span of decades, reflect substantially different patterns of limb segmen- However, I
tation and muscle removal than do bones of like-sized animals at mass pro- away from co
cessing sites (Guilday et al. 1962). ery. Here I sh
Despite those indications of the role of culinary goals and technology in examining th
structuring faunal assemblages, recent ethnoarchaeological literature has zooarchaeolo
largely noted them as asides to discussions of field processing. To use a faunal differentially
metaphor, this approach may be putting the cart before the horse. When I
considered this fact in combination with the scanty research on the influence
of cooking on bone breakage and cut mark patterns, I was led to ask why the
effects of culinary processing on bone had not been deemed a suitable
research topic by archaeologists. Tl:
zooarchaeolo
within the fie
Why Has Culinary Processing Been Ignored? ment since th1
Reasons for the research emphasis on uncooked bone are several. have been ch
First, centrally influential works on butchery, such as Frison's Great Plains see as importe
Gaps in Zooarchaeological Analysis j187
Binford 1978; research and Binford's Nunamiut project, focus on mass kill assemblages.
mting stands Simultaneous kills of many animals create a kind of "mass production"
ocessing and atmosphere, in which considerations of spoilage, transport, and storage inter-
of people, in act to drive butchery in an efficiency-focused direction. Recent Hadza research
: of situation- (Bunn et al. 1988; O'Connell et al. 1988, 1990), as well as Yellen's (1977b)
assemblages earlier work, indicate that single-animal processing may not parallel the
rtant consid- starkly repetitive, cost-benefit, discard and modification patterns typical of
mass kill situations. Mass kill situations also heighten the likelihood that meat
)ducts drives will be removed prior to cooking, that long bones will be fractured raw, if
cle-stri pping they are fractured at all, and that bones will be discarded on-site, without
. 1974; Frison cooking. These cases may not adequately direct our attention to the circum-
·ere aimed at stances and strategies forming most archaeological bone deposits.
om as many The early hominid focus of much actualistic research also contributes to the
~sh for trans- concentration on uncooked body segment processing. Many ethnoarchaeo-
:essing of fat logical researchers want to apply their findings to interpreting bone assem-
oals produce blages in Plio-Pleistocene archaeological sites (e.g., Binford 1981, 1986; Bunn
utchery sites 1981, 1983, 1986; Bunn and Kroll1986; Bunn et al. 1988;.0'Connell et al. 1988,
orts another 1990). Those who use experimental research on butchery marks have similar
·s. Instead of goals (e.g., Shipman 1981, 1986a, 1986b, 1989; Shipman and Rose 1983).
ned dismem- Whether/of the pro-hunting or pro-scavenging camp, those researchers share
segments to the assufnption that their target populations lacked fire and-it seems implicit
in their writings-subjected animal products to minimal processing once they
extraction of were extracted from carcasses. Early hominids' transporting carcass segments
r from those or products away from a carcass is considered likely (e.g., Binford 1984; Bunn
:ime and at a and Kroll 1986), although different researchers say early hominids were
hat bones are differently motivated in distancing themselves from a large animal death site
on a one-by-r (the cowering scavenger hotfooting it for a safe haven versus the efficient
, Bunn et al. hunter strolling home with the choicest cuts of a kill). No one has, to my
'77b). To take knowledge, explicitly discussed what. tactics early hominids might have
bones at the employed to process animal foods, past an implicit "smash and/ or slash and
d elk entered gulp" scenario.
imb segmem- ' However, I believe they are not the only reasons why attention is directed
at mass pro- away from cooked bone and from the culinary end products of animal butch-
ery. Here I shift focus from the internal logic of zooarchaeological practice to
~chnology in examining the structure of assumptions and the standards of research in
.terature has zooarchaeology. Specifically, I want to examine the role of gender bias in
, use a faunal differentially valuing certain types of research questions.
)rse. When I
the influence
ask why the Zooarchaeo1ogical Research as Gendered Practice
d a suitable
The lack of attention to cooking and culinary end products in
zooarchaeology is, I believe, attributable to unconscious androcentric bias
within the field . Despite having undergone trerrlendous rrlethodological fer-
ment since the early 1970s, the subjects of study and debate in zooarchaeology
~are several. have been chosen according to an underlying valuation of what researchers
Great Plains see as important activities. This view favors hunting-especially male pursuit,
1881 D. Gifford-Gon_zalez
dispatch, and butchery of prey-{)ver just about any other activity involving ings bee a usE
animals, even the su·pposedly patriarchal practice o'f pastoralism. The patterning in
paradigmatic emphasis at least implicitly characterizes large animal hunting Specificall
as the central part of past human adaptations, usually ignoring the role of studies of a
food-processing technology and of the social allocation of fqod resources as p~rhaps rnm
pivotal adaptive issues. It characterizes field butchery and transport as male hunting pen
activity and central to understanding "adaptation," while secondary process- affect the ul
ing-storage or culinary-is less interesting, less central to adaptation, and ethnographi'
often "left to" marginally noted females, if discussed.at all. helping acqt
But it is even worse than that. Not even all men are deemed worthy of study for storage
in this context. The economic roles of nonhunting males and children of either Mackenzie:
. sex are downplayed in ethnographic accounts of hunting, except insofar as 1983; Wheat
they constitute part of the carrying party hunters mobilize to get their prey to grease, and
residential camps. Nonhunter processing inputs and impacts on bone are not managing s
included in discussions of either energetic costs incurred or nutritional bene- accounts, cal
fits reaped by such secondary processing of animal products, or treatments of and women,
the form of debris that hits the ground. I contend that they are seriously use of animz
incomplete accounts of both hominid tool-mediated subsistence and the activ- monitored.
ities that form archaeological sites. Second, a:
Concentration of carcass processing research on the pre-culinary, field Kroll 1986; E
phase reflects, I believe, the "male equals public, female equals private" field butcheJ
dichotomy of our own culture, in which the really important decisions in life ~of prey acqu
are seen as going on outside the sphere of women and home (Collier and decisions ar
Yanagisako 1987; Conkey and Spector 1984; Rosaldo 1980; Strathern 1988). In and social
fact, the dichotomy may actually be reproduced in the "Man the Hunter/ approach th
Woman the Gatherer" stereotypes now common in the general anthropo- ations in mi
logical literature. While it has served a useful role in documenting human on selectior
females as active and productive foragers (e.g., Tanner 1981; Tanner and approach to
Zihlman 1976), the Woman the Gatherer concept can reify our own/Culture's in the optim
ideas about the division of labor between men and women. Specifically, it Blurton Jon
blinds us to the involvement of women and nonhunting personnel in tasks theoretical u
that seek, acquire, and process animal foods and other animal products. I the entire c1
contend that they are seriously incomplete accounts both of hominid tool- studies by C
mediated subsistence and of processes that form archaeological sites. To the from their c<
extent that we underplay certain processing tasks because we think of them as culinary tas:
routine, simple, passive, or fundamentally unproductive (in our culture's impacts of s
gender paradigm, female), we undermine our understandings of processes assessed.
forming the archaeological record on a· regional scale and so undermine the A behavic
effective practice of zooarchaeology. ing uniforrr
I want to make clear that I do not dispute that among modern peoples, most cases than J
of the time, it is men rather than women who usually stalk and kill large scope can 1
animals and that it is reasonable on selective grounds to assume a strong Modern huJ
tendency for males to have done so in the past. Rather, the purpose of my individual !
inquiry is to delineate the unrecognized shape of our discipline, as a conver- inputs as sir
sation in which we all-men and women-participate and take to be our that is, in a
common knowledge. Let us explore some implications for zooarchaeological that structu
methodology. I am concerned that the implicit biases produce less useful find- one's goal j
:y involving ings because our attention is directed away from an important source of
·alism. The patterning in archaeological materials.
nal hunting Specifically, how does this bias negatively impact ethnoarchaeological
; the role of studies of animal processors? Gaps are created in two areas. First, and
·esources as perhaps most superficially, is the a priori assumption that women and non-
port as male hunting personnel are not themselves active field butchers whose decisions
uy process- affect the ultimate form of the animal food base and its bony refuse. The
Jtation, and ethnographic literature abounds with examples of women as major actors in
helping acquire large animals (Mason 1907), in field processing large animals
thy of study for storage or consumption (Demallie 1983; Denig 1930; Dodge 1959;
ren of either Mackenzie 1960; Medicine 1983; Weist 1983), in transporting them (Weist
,t insofar as 1983; Wheat 1972), in extracting bulk ·carcass products such as marrow, bone
:heir prey to grease, and bone juice (Binford 1978; Fletcher and La Flesche 1911), and in
Jane are not managing stored products (Binford 1978). Women, from ethnographic
tional bene- accounts, can be up to their elbows in large animal carcasses. Given that men
·eatments of and women, even of the same household, may have divergent agendas in the
re seriously use of animal products, women's decisions and their consequences need to be
1d the activ- monitored.
Second, and more seriously, recent discussions (Binford 1986; Bunn and
.inary, field Kroll19~6; Bunn et al. 1988; O'Connell et al. 1988, 1990) appear to assume that
1ls private" field butchery decisions are primarily driven by the particular circumstances
lsions in life of prey acquisition in the field (the first six factors listed earlier) rather than by
:Collier and decisions and strategies grounded in the overall nutritional, technological,
~rn 1988). In and social context in which hunters live. The assumption that hunters
he Hunter/ approach their field butchery tasks without such end-product-based consider-
.1 an thropo- ations in mind is, I believe, problematic. Ethnoarchaeological research based
ting humanr on selectionist theory might be expected to take a more comprehensive
Tanner and approach to meat provisioning but in fact has not. Some recent ethnography
tVn culture's in the optimal foraging framework has focused on women and children (e.g.,
ecifically, it Blurton Jones et al. 1989; Hawkes et al. 1989), as is reasonable., given the
.nel in tasks theoretical underpinnings of this approach. Although the costs and benefits of
products. I the entire chain of animal processing would seem to require study as well,
m1inid tool- studies by O'Connell and colleagues (e.g·., 1988, 1990) have thus far omitted
sites. To the from their cost-benefit calculations the time and energy invested in secondary,
k of them as culinary tasks-and the· nutritional benefits of such processing. Nor have the
ur culture's impacts of such processing on actual on-the-ground element frequencies been
Jf processes assessed.
:lermine the A behavioral ecological approach may indeed be a better means of discover-
ing uniformitarian relationships with which to understand archaeological
=oples., most cases than less theory-informed, ad hoc investigations. But I believe their
Ld kill large scope can be expanded beyond what is apparent in published studies.
ne a strong Modern humans are unique in the degree that social cooperation supports
·pose of my individual survival and reproduction. To evaluate human prey processing
as a conver- inputs as simply analogous to those of non-tool-using, less social predators-
:e to be our that is, in a primarily kill-site-based perspective-is to ignore major factors
:haeological that structure bone assemblages among modern hunting hominids. Even if
useful find- one's goal is to discern uniformitarian relationships with which to study
I
earlier hominid forms, whose carcass processing strategies may have been
closer to those of nonhuman carnivores, it is essential to aefine the "modern"
impacts that should be lacking in earlier by-products. .
The question, which needs to be addressed ethnographically, is how much
do nonhunters' agendas, either home"-economic or reproductive, structure
hunter's actions in the field? Binford's story about the hunters and their wives
led me to reflect on the possibility that, beyond the many factors enumerated
in the literature on carcass processing, we need to consider the imperatives of
the household in driving field processing decisions. Men and women in
households engage in common and basic projects of feeding themselves,
rearing and feeding children, supplying dependent elders with shelter and
sustenance, and developing their social lives according to their respective
goals. Given that hunting is largely aimed at satisfying such needs, we might,
as a radical alternative, envision it as an extension of home economics, in
which verbal and nonverbal information from all members of households
heavily influences hunting and field butchery decisions. These are empirically
investigatable issues, accessible to various established forms of systema~ic
analysis.
The problem with recent studies' focus-whatever their theoretical ground-
ing-may really be that they have been overly determined by simplistic
archaeological visions and versions of the past, including a nearly fetishistic
obsession with the search and pursuit phases of hunting. It would be a waste
of human effort if we only used our ethnographic research to flesh out a priori
(and often profoundly ethnocentric) ideas about the sexual division of labor-
and the process of food-getting itself-rather than seeing what is really going
on in our study cases. I am not advocating a purely inductive approach, but
rather one in which research is a continual confrontation of analytical models
and concepts with empirical evidence, with modification of the former in light
of the latter when necessary. /
What Are Productive Research Approaches? F:

Vl
The ethnoarchaeology of human subsistence profits when it takes a gt

task- and product-focused approach. This approach envisions each element in m
a faunal assemblage as an end product of the chain of events through which it
has passed (Figures 11-1 through 11-3). From this perspective, it isirrtportant
to monitor processing impacts and their benefits and costs at all phases of the approach, as
chain. All phases of animal processing, from acquisition through final refuse terns but as
disposal, are studied. Relationships between processing decisions and tactics prod uct-foct:
and their material consequences are isolated at each phase. A task-focused faunal remai
and product-focused approach to faunal materials in ethnographic settings and the pracj
thus asks how sites and their contents form. I advocated this approach from a field.
taphonomic perspective ten years ago, as have Binford (1981) and Lyman Somewhat
(1987), and Behrensmeyer and Kidwell (1985) in paleontology. A task-focused attention to ,
approach was also advocated by Conkey and Spector (1984) as a means of nonhunters'-
highlighting the social roles involved in forming archaeological sites. In this tial archaeol
.y have been
:1e "modern" ACQUISITION SITE
[HUNTING STAND,
is how much AD HOC ENCOUNTER,
ve, structure MASS KILL, SCAVENGE]
d their wives
enumerated
nperatives of
d women in
themselves,
SNACK SITE
1 shelter and
-ir respective
is, we might,
[SPECIAL PURPOSE
WITH ANIMAL INPUTS] [ CACHE SITE
conomics, in
f households ....
I
....
e empirically ....
....
....
)f systematic ....
tical ground- ~, 1-"

JY simplistic
rly fetishistic ~~P
ld be a waste RESIDENTIAL SITE
h out a priori
on of labor- [SHORT- OR LONG-TERM]
:; really go,ing
.pproach, but
ytical models
)!mer in light
Figure 11-1. A visual model of the movement of faunal remains through

various sites in ·a subsistence system that provisions meat through large
·hen it takes a game hunting, specifying several site types at which animal carcasses
ch element in may be processed.
)ugh which it
:is important
phases of the approach, assemblages are viewed not as "biased" remnants of extinct sys-
h final refuse tems but as aggregate evidence about the past states of those systems. A
:lS and ta.ctics product-focused research strategy necessarily attends to the tasks affecting
task-focused faunal remains-including storage, cooking, and refuse disposal practices_:
.phic settings and the practitioners, whoever they are, whether they act "at home". or in the
'roach from a field.
) and Lyman Some·what pa!adoxica:lly, the strategy does not require paying special
. task-focused attention to women. It just involves not ignoring them and their-and other
sa means of nonhunters'-participation in processing chains and tasks that create poten-
l sites. In this tial archaeological sites. Ethnographic cases suggest that participation of
192j D. Gifford-Gon;;alez
. ACQUISITION SITE
DISPATCH OF .PREY
PRE-TRANSPORT PROCESSING
BULK PRODUCT EXTRACTION
[CULINARY PROCESSING]
SNACK SITE
DISPATCH OF PREY
CACHE SITE
PRE-STORAGE PROCESSING
[
CULINARY PROCESSING [CULINARY PROCESSING]
Dl~
VI!
....
.... AN
......
......
Jl~
RESIDENTIAL SITE
PRE-STORAGE PROCESSING
BULK PRODUCT EXTRACTION
CULINARY PROCESSING
REFUSE PROCESSING/DISPOSAL
Fi
Figure 11-2. The same model, showing animal processing activities likely lik
to occur at each locality type.
different age and sex classes in animal processing activities varies tremen-
dously, both situationally within a culture and from one culture to another. M
Our task is really to open our eyes to the irripacts of such variability. to develop le
One researcher who has taken a product-focused approach is Binford, in his faunal remair
Nunamiut research (e.g., 1978, 1981). It meant that he actually spent consider- sarily possib
able time observing women's work. I doubt that Binford did it out of a femi- humans' soc:
nist agenda-in fact, he too, emphasizes male actions and the decisions under- excavate· rna)
lying them. Binford's orientation is, rather, a logical outcome of his persistent expect to acct
concern with the archaeological products of human actions. In other words, in all women tl
order to follow bones to their final resting places, as well as to understand hominid decis
butchery decisions, Binford needed to pay attention to women's-and other combine faur
nonhunters' -work. the activities
ACQUISITION SITE
TRACES· CUTS, CHOPS, SCRAPES,
BLoDol;~~:c~:RA,
....
HORN, ANTLER, BONE
~ p
FRACTURES, BURNING, GNAWING
SNACK SITE CACHE SITE

TRACES: CUTS, CHOPS, TRACES· CUTS, CHOPS,
SCRAPES, FRACTURES, SCRAPES, FRACTURES,
BURNING, GNAWING [BURNING, GNAWING]
,,
I ', \
DISCARD: BLOOD, ',,
VISCERA, HORN, ',, DISCARD· BONES
AN~, BONE >,~
\
~r9 RESIDENTIAL· SITE
~r9
I
TRACES: CUTS, CHOPS, SCRAPES,
FRACTURES, BURNING, GNAWING
I
DISCARD· BLOOD,
.VISCERA, HORN,
ANTLER, BONE
~·p
Figure 11-3. The same model, showing bone modifications and discards
ctivities likely likely to occur at each locality type.
Conclusion
Lries tremen-
e to another. My thesis is that our own culture's gender paradigms have led us
ty. to develop less than comprehensive approaches to human actions affecting
inford, in his faunal remains. I have not argued here that it is easy--or even that it is neces-
ent consider- sarily possible-to "dig ·up gender" or other social aspects of prehistoric
mt of a femi- humans' social lives, not, at least, relying on bones alone. The bones we
isions under- excavate may be products of gendered actions, but what we can reasonably
his persistent expect to access-short of uniformitarian flights of fancy making all men and
her words, in all women the same in all places for all times-is the indirect evidence of
) understand hominid _decisions and actions affecting animals' remains. We may be able to
s-and other combine faunal data with other, independent lines of evidence to ask about
the activities of diff~rent ages and genders in that pa~t system (Conkey and
194j D. Gifford-G()nzalez .
Spector 1984; Gifford-Gonzalez 1991). But to do this we must first expand our ology of Bo;
understanding of the linkages between human decisions and actions and their Press, New
products. That is what should be studied ethnoarchaeologically. Behrensmeyer, j
1985 Tapho
We zooarchaeologists have had great success pursuing a research program
Binford, L. R.
based on a uniformitarian approach to faunal materials as. an entry point to 1977 ForTh
the unknown terrain of ancient hominid life. I do not dispute this method- 1978 Nunan
ological stance since I fully support it (Gifford 1981; Gifford-Gonzalez 1989b, 1981 Bones:
1991). But I think we are not using ethnographic cases to their full potential. 1984 Faunal
Ethnoarchaeology is important not only because .it lets us see the dynamic 1986 Comn
processes that produce archaeological sites and assemblages. It is also crucial Gorge, Tan
because it and other actualistic stUdies in archaeology often lead us to ques- Black, A. T.
tion the realism of our research questions and analytical categories (viz., 1989 The E!J
Binford 1977; Wylie 1989). We use actualistic research to our .fullest advantage Study. Unp
when it informs and re-forms our ideas about what might have happened in Blurton Jones, :r\
the past, rather than simply using it to flesh out our preexisting ideas about it. 1989 Study
To study the use of animal resources in a more realistic way requires a Schedules
edited by \I
wider scope and attention to the influence and impacts of nonhunting house-
Bonar, L. C., anc
hold members. To attend to these aspects of animal-based subsistence and its 1970 Them
archaeological consequences is not to force gender on our analytical Journal ofU
approaches; rather, it is to ungender categories that up to now have been Bonfield, W.
falsely gendered. 1981 Meche;
S.C. Cowjl
Note 1987 AdvaJ
chanics 20:1
1. Analysts dealing with later prehistoric Mediterranean and European faunal Bonfield, W., an
assemblages have emphasized cooking techniques more than have most North 1966 Defon
American researchers. It probably stems from their concentration on food-producing Bonnichsen, R:
economies in which houses and food-processing and distribution methods are central 1973 Some
objects of study. However, as I will discuss later in this paper, I suspect tpat hunting Mammalian
and herding are studied according to androcentric models within Europeanist Missouri A
research paradigms as well. 1978 Critic;
Yukon: AI
pp. 102-11
Acknowledgments Anthropolc
1979 Pleistc
A preliminary version of this paper was drafted during tenure of a
of Man Mer
National Science Foundation Research Opportunities for Women Award, Bunn,H. T.
BNS-8711024. I am grateful to Jean Hudson for inviting me to participate in 1981 Archa
the From Bones to Behavior conference. I thank Carolyn Clark, Shelly fromKooh
Errington, Joan Gero, Donald Grayson, Lee Lyman, and Anna Tsing for their 1983 Evide
valuable comments on drafts of this manuscript. Hominids
Archaeolog}
pp. 21-30. ~
References 1986 Patter
Amprino,R. Olduvai G
1958 Investigations of Some Physical Properties of Bone Tissues. Acta Anatomica 15:673-690
34:161-186 Bunn,·H. T., L. I
Ascenzi, A., and G. H. Bell 1988 Varial
1972 Bone as a Mechanical Engineering Problem. In The Biochemistry and Physi- ing, and Cc
rst expand our ology of Bone, val. 1, 2d ed., edited by G. H. Bourne, pp. 311-351. Academic
~tionsand their Press, New York. ·
Behrensmeyer, A. K., and S. Kidwell
earch program 1985 Taphonomy's Contributions to Paleobiology. Paleobiology 11:109-119.
Binford, L. R.
entry point to
1977 For Theory Building in Archaeology. Aca~emic Press, New York.
e this method- . 1978 Nunamiut Ethnoarchaeology. Academic Press, New York.
onzalez 1989b, 1981 Bones: Ancient Men and Modern Myths. Academic Press, New York.
~ full paten tial. 1984 Faunal Remains from Klasies River Mouth. Academic Press, New York.
e the dynamic 1986 Comment on Systematic Butchery by Plio-Pleistocene Hominids at Olduvai
t is also crucial Gorge, Tanzania, by H. T. Bunn and E. Kroll. Current Anthropology 27:431-452.
~ad us to ques- Black, A. T.
ttegories (viz., 1989 The Effects of Cooking on the Fracture Pattern of Bovine Bones: A Microstructural
llest advantage Study. Unpublished Master's thesis, State University of New York, Binghamton.
·e happened in Blurton Jones, N., K. Hawkes, and J. O'Connell
ideas about it. 1989 Studying Costs of Children in Two Foraging Societies: Implications for
va y requires a Schedules of Reproduction. In Comparative Socioecology of Mammals and Man,
1unting house- edited by V. Standon and R. Fley, pp. 365-390. Blackwell, London.
Bonar, L. C., and M. J. Glimcher
;istence and its 1970 Thermal Denaturation of Mineralized and Demineralized Bone Collagens.
Jur analytical Journal, of Ultrastructural Reseach 38:545..;.557.
.ow have been Bonfield, vf.
1981 Mechanisms of Fracture in Bone. In Mechanical Properties of Bone, edited by
S.C. Cowin, pp. 163-170. ASME, New York.
1987 Advances in the Fracture Mechanics of Cortical Bone. Journal of Biome-
chanics 20:1071-1081.
.uropean faunal Bonfield, W., and C. H. Li
tve most North 1966 Deformation and Fracture of Bone. Journal of Applied Physics 37:869-875.
food-producing Bonnichsen, R.
:hods are central 1973 Some Operational Aspects of Human and Animal Bone Alteration. In
1ect that hunting Mammalian Osteoarchaeology: North America, edited by B. Miles Gilbert, pp. 9-24.
tin Europeanist Missouri Archaeological Society, Columbia.
1978 Critical Arguments for. Pleistocene Artifacts from the Old Crow Basin,
Yukon: A Preliminary Statement. In Early Man in America, edited by A. L. Bryan,
pp. 102-118. Occasional Papers of the University of Alberta Department of
Anthropology Nci. 1. ·
ing tenure of a 1979 Pleistocene Bone Technology in the Beringian Refugium. National Museum
of Man Mercury Series, Archaeological Survey of Canada Paper 89:1-280.
·omen Award, Bunn,H. T.
1 participate in
1981 Archaeological Evidence for Meat-eating by Plio-Pleistocene Hominids
Clark, Shelly from Koobi Fora and Olduvai Gorge. Nature 291:574-577.
Tsing for their 1983 Evidence on the Diet and Subsistence Patterns of Plio-Pleistocene
Hominids at 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
. Acta Anatomica 15:673-690. ~. ~ .
1988 Variability in Bone Assemblage Formation from Hadza Hunting, Scaveng-
1istry and Physi-
1 ing, and Carcass Processing. Journal of Anthropological Ar~haeology 7:412-457.
196j D. Gifford-Gonzalez·
Bunn, H. T., and E. M. Kroll Fossils in J

1986 Systematic Butchery by Plio-Pleistocene HomiRids at Olduvai Gorge, Universit)
Tanzania. Current Anthropology 27:412-452. 1981 Tapb
Catlin, G. DisciplinE
1959 Episodes of Life among the Indians, and Last Rambles. University of Oklahoma, B. Schiffer
Norman. Gifford, D. P., ;
Collier, J. F., and S. J. Yanagisako 1977 Obse
1987 Introduction. In Gender and Kinship: Essays Toward a Unified Analysis, edited Kenya. Qz
by J. F. Collier and S. J. Yanagisako. Stanford University Press, Stanford, Calif. Gifford-Gonza:
Conkey, M., and J. Spector . 1984 Tapr
1984 Archaeology and the Study of Gender. In Advances in Archaeological Method Reports, 1.
and Theory, vol. 7, edited by M. B. Schiffer, pp. 1-38. Academic Press, New York. D.C.
Davis, K. L. 1988 Ship1
1985 A Taphonomic Approach to Experimental Bone Fracturing. and Applications to Anthropol£
Several South African Pleistocene Sites. Unpublished Ph.D. dissertation, State 1989a Ethn
University of New York, Binghamton. from East
Demallie, R. J. 179-246. '
1983 Male and Female in Traditional Lakota Cultur~. In The Hidden Half: Studies of Orono.
Plains Indian Women, edited by P. Albers and B. Medicine, pp. 237-262. 1989b Mod
University Press of America, Lanham, Md. Modijicati1
Denig, E. T. Study oft
1930 Indian Tribes of the Upper Missouri, edited by J. N. B. Hewitt, pp. 375-654. 1991 Bone
Bureau of American Ethnology, Annual Report 46. Washington, D.C. in ZooarcJ
Densmore, F. Guilday, J. E., ]
1918 Teton Sioux Music. Bureau of American Ethnology, Bulletin 61. 1962 Ab01
Dodge, R.I. Lancaster
1959 Our Wild Indians: Thirty-Three Years' Personal Experience among the Red Men · Hawkes, K., J. :
of the Great West. Reprinted. Archer House, New York. Originally published 1989 Hare
1882, A. D. Worthington, Hartford, Conn. · Behavioun
Fletcher, A. C., and F. La Flesche Foley, pp.
1911 The Omaha Tribe. Bureau of American Ethnology, Annua~, Report 27. Horwitz, V. D.
Washington, D.C. 1987 Expe
Frison, G. C. Meeting c
1970 The Glenrock Buffalo Jump, 48 C0304. Plains Anthropologist Memoir No.7. Johnson, E.
1974 The Casper Site: A Hell Gap Bison Kill on the High Plains. Academic Press, 1978 Pale<
New York. Estacada.
Frison, G. C., and L. C. Todd (editors) Davis and
1987 The Horner Site: The Type Site of the Cody Cultural Complex. Academic Press, 1985 Curr
Orlando, Fla. Method an
Frison, G., tvf. Vvilson, and D. J. vVilson NewYor1
1976 Fossil Bison and Artifacts from' an Early Altithermal Period Arroyo Trap in Leakey, R., anc
Wyoming. American Antiquity 41:28-57. 1978 Peopl
Gifford, D. P. N.Y.
1977 Observations of Modern Human Settlements as an Aid to Archaeological Lyman, R. L.
Interpretation. Ph.D. dissertation, University of California, Berkeley. University 1987 Arch
Microfilms, Ann Arbor. Advances 1
1978 Ethnoarchaeological Observations on Natural Processes Affecting Cultural 249-337. j
Materials. In Explorations in Ethnoarchaeology, edited by R. A. Gould, pp. 77-101. Mackenzie, A.
University of New Mexico Press, Albuquerque. 1960 The:
1980 Ethnoarchaeological Contributions to the Taphonomy of Human Sites. In Mississou
Fossils in the Making, edited by A. K. Behrensmeyer and A. P. Hill, pp. 93-106.

Olduvai Gorge, University of Chicago Press, Chicago.
1981 Taphonomy and Paleoecology: A Critical Review of Archaeology's Sister
Disciplines. In Advances in Archaeological Method and Theory, val. 4, edited by M.
lty of Oklahoma, B. Schiffer, pp. 365-438. Academic Press, New York.
Gifford, D.P., and A. K. Behrensmeyer .
1977 Observed Formation and Burial of a Recent Human Occupation Site in
l Analysis, edited Kenya. Quaternary Research 8:245-266.
tanford, Calif. Gifford-Gonzalez, D.
1984 Taphonomic Specimens, Lake Turkana. In National Geographic Research
aeological Method Reports, 1976 Projects, pp. 419-428. National Geographic Society, Washington,
ress, New York. D.C.
1988 Shipman's Shakey Foundations: The Prolonged Drift Cut Marks. American
d Applications to Anthropologist 91:180-186.
ssertation, State 1989a Ethnographic Analogues for Interpreting Modified 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, University of Maine,
~n Half: Studies of Orono.
e, pp. 237-2p2. 1989b Modern Analogues: Developing an Interpretive Framework. In Bane
Modification, edited by R. Bonnichsen and M. Sorg, pp. 43-52. Center for the
Study of the First Americans, University of Maine, Orono.
vitt, pp. 375-654. 1991 ;Bones Are Not Enough: Analogues, Knowledge, and Interpretive Strategies
in Zooarchaeology. Journal of Anthropological Archaeology 10:215-254.
Guilday, J. E., P. Parmalee, and D. Tanner
[. 1962 Aboriginal Butchering Techniques at the Eschelman Site (36LA12),
Lancaster County, Pennsylvania. Pennsylvania Archaeologist 32:59-83.
wng the Red Men Hawkes, K., J. F. O'Connell, and N. Blurton Jones
inally published 1989 Hardworking Hadza Grandmothers. In Comparative Saciaecolagy: The
Behavioural Ecology of Humans and Other Mammals, edited by V. Standen and R.
Foley, pp. 341-366. Blackwell, Oxford.
nual Report 27. Horwitz, V. D.
1987 Experiments on Cooked Bone Breakage. Paper presented at the 52d Annual
Meeting of the Society for American Archaeology, Toronto.
t Memoir No. 7. Johnson, E.
!\cademic Press, · 1978 Paleo~Indian Bison Procurement and Butchering Pattern on the Llano
Estacada. In Bison Procurement and Utiiization: A Symposium, edited by L. B.
Davis and M. Wilson. Plains Anthropologist Memoir 14:98-105.
Academic Press, 1985 Current Developments in Bone Technology. In Advances in Archaeological
Method and Theory, vol. 8, edited by M. B. Schiffer, pp. 157-235. Academic Press,
New York.
i Arroyo Trap in Leakey, R., and R. Lewin
1978 People of the Lake: Mankind and Its Beginnings. Anchor Press, Garden City,
N.Y.
o Archaeological Lyman, R. L.
:eley. University 1987 Archaeofaunas and Butchery Studies: A Taphonomic Perspective. In
Advances in Archaeological Method and Theory, vol. 10, edited by M. B. Schiffer, pp.
ffecting Cultural 249-337. Academic Press, New York.
)U}d, pp. 77-;;:-101. Mackenzie, A.
1960 The Mississouri Indians: A Narrative of Four Trading Expeditions to the
Human Sites. In Mississouri, 1804-1806, for the North-West Company. In Les bourgeois de la
198j D. Gifford-Gonzalez
Compagnie du Nord-ouest: recits de voyages, lettres et rapports inedits relatifs au nord- Shipman, P., an
ouest canadien, edit,ed by L. R. Masson, pp. 362-370. Antiquarian Press, New 1983 Early
Ymk · Approache
Mason,O. T. Simkin, A., and
1907 Hunting. In Handbook of American Indians North of Mexico, edited by F. W. 1974 Fractt
Hodge, pp. 580-581. Bureau of American Ethnology, Washington, D.C. Journal of B
Medicine, B. Speth, J., and K
1983 _Warrior Women: Sex Role Alternatives for Plains Indian Women. In The 1983 Energ
Hidden Half: Studies of Plains Indian Women, edited by P. Albers and B. Medicine, Strategies.
pp. 267-280. University Press of America, Lanham, Md. Strathern, M.
Mengoni-Gofialons, G. _ 1988 The G
1980 Notas zooarqueol6gicas: fracturas en huesos. VII Congreso Nacional de Melanesia.·
Arquelogia, Colonia de Sacramento, Uruguay, 1980. Centro de Estudios de Tanner,N.
Arqueologia, Montevideo. 1981 On Bt:
O'Connell, J. F., K Hawkes, and N. Blurton Jones Tanner, N., and
1988 Hadza Hunting, Butchering, and Bone Transport Practices and Their 1976 Worn
Archaeological Implications. Journal of Anthropological Research 44:113-161. Signs: Jour;
1990 Reanalysis of Large Mammal Body Part Transport among the Hadza. Weist, K.
Journal of Archaeological Science 17:301-316. 1983 Beast
O'Connell, J. F., and B. Marshall Northern
1989 Analysis of Kangaroo Body Part Transport among the Alyawara of Central Women, ec
Australia. Journal of Archaeological Science 16:393-405 America, I
Oliver, J. S. Wheat, J. B.
1989 Analogues and Site Context: Bone Damages from Shield Trap Cave 1972 The C
(24CB91), Carbon County, Montana, U.S.A. In Bone Modification, edited by R. Archaeolog
Bonnichsen and M. Sorg, pp. 73-98. Center for the Study of the First Americans, 1979 TheJ1
University of Maine, Orono. Wissler, C.
Richter, J. 1910 The~
1986 Experimental Study of Heat Induced Morphological Changes in Fish Bone History, A1
Collagen. Journal of Archaeological Science 13:477-481. Wylie, A.
Rosaldo, M. 1 1985 The l
1980 The Use and Abuse of Anthropology: Reflections on Feminis1'l"1-' and Cross- Theory, vo·
cultural Understanding. Signs: Journal of Women in Culture and Society 5:389-417. 1988' Simp
Sedlin, E. Archaeolo
1965 A Rheological Model for Cortical Bone. Acta Orthopaedica Scandinavia 37:29- 1989 Arch
48. Bernstein'
Shipman, P. Science 19:
1981 Applications of Scanning Electron Microscopy to Taphonomic Problems. Yellen, J. E.
New York Academy of Sciences, Annals 376:357-386. 1977a Archc
1986a Scavenging or Hunting in Early Hominids: Theoretical Framework and Academic
Tests. American Anthropologist 88 27-43. 1977b Cultt
1986b Studies of Hominid-Faunal Interactions at Olduvai Gorge. Journal of Human In Experi
Evolution 15:691-706. MacDona]
1989 Altered Bones from Olduvai Gorge, Tanzania: Techniques, Problems, and
Implications of Their Recognition. In Bone Modification, edited by R. Bonnichsen
and M. Sorg, pp. 317-334. Center for the Study of the First Americans,
University of Maine, Orono
Shipman, P., G. S. Foster, and M. Schoeninger
1984 Burnt Bones and Teeth: An Experimental Study of Colour, Morphology,
Crystal Structure and Shrinkage. Journal of Archaeological Science 11:307-325.
' relatifs au nord- Shipman, P., and J. J. Rose
ian Press, New 1983 Early Hominid Hunting, Butchering and Carcass-processing Behaviors:
Approaches to the Fossil Record. Journal of Anthropological Archaeology 2:57-98.
Simkin, A., and G. Robin
edited by F. W. 1974 Fracture Formation in Differing Collagen Fiber Patterns of Compact Bone.
D.C.
:-t, Journal of Biomechanics 7:183-188.
Speth, J., and K. A. Spielmann
Women. In The 1983 Energy Source, Protein Metabolism, and Hunter-Gatherer Subsistence
md B. Medicine, .strategies. Journal of Anthropological Archaeology 2:1-31.
Strathern, M.
1988 The Gender of the Gift: Problems with Women and Problems with Society in
-eso Nacional de Melanesia. University of California Press, Berkeley.
le Estudios de Tanner, N. ·
1981 On Becoming Human. Cambridge University Press, Cambridge.
Tanner, N., and A. Zihlman
tices and Their 1976 Women in Evolution. Part I. Innovation and Selection in Human Origins.
1:113-161. Signs: Journal of Women in Culture and Society 1:104-119.
mg the Ha}iza. Weist, K.
1983 Beast of Burden and Menial Slaves: Nineteenth Century Observations of
Northern Plains Indian Women. In The Hidden Half: Studies of Plains Indian
1wara of Central Womep., edited by P. Albers and B. Medicine, pp. 29-52. University Press of
America, Lanham, Md.
Wheat,J. B.
leld Trap Cave 1972 The Olsen-Chubbuck Site: A Paleo-Indian Biso_;n Kill. Society for American
m, edited by R. Archaeology Memoir No; 26.
First Americans, 1979 The Jurgens Site. Plains Anthropologist Memoir No. 15.
Wissler, C.
1910 The Material Culture of lhe Blackfoot Indians. American Museum of Natural
ges in Fish Bone History, Anthropological Papers, val. 5, pt. 1.
Wylie, A.
1985 The Reaction Against Analogy. In Advances in Ardzaeological Method and
nism and Cross- Theory, val. 8, edited by M. B. Schiffer, pp. 63-111. Academic Press, New York.
'ciety 5:389-417. 1988 Simple Analogy and the Role of Relevance Assumptions: Implications from
. Archaeological Practice. International Studies in Philosophy 2:134-150.
;andinavia 37:29- '1989 Archaeological Cables and Tacking: The Implications for Practice for
Bernstein's "Options Beyond Objectivism and Relativism." Philosophy of Social
Science 19:1-18.
,omic Problems. Yellen, J. E.
1977a Archaeological Approaches to the Present: Models for Reconstructing the Past.
Framework and Academic Press, New York.
1977b Cultural Patterning in Faunal Remains: Evidence from the !Kung Bushmen.
rournal of Human In Experimental Archaeology, edited by D. Ingersolt J. E. Yellen, and W.
MacDonald, pp. 271-331. Columbia University Press, New York.
;, Problems, and
yR. Bonnichsen
irst Americans,
1r, Morphology,
11:307-325.
1986, 1989, 19'
1983). Few da1
reflected in bo
bone. To a larg
ior '(Binford 1~
tioned, resulti1
12. Carcass Processing by the Hadza: behaviors may
differentiatirg
Bone Breakage from Butchery actualistic wor
That is, ethno<
to Consumption focus to selec'
carcass proces~
James S. Oliver 1990; O'ConnE
1981, 1983; Pot
extraction fror
Abstrad: Owing in part to the episodic treatment of human behavior, Johnson 1985).
archaeologists have limited ethnoarchaeological data on how cultu,ral field butchery
behavior may be reflected in bone frequency and modification patterns, explicit treatm·
particularly fractured bone. Detailed observations of butchery-, cooking- The episodi<
preparation-, and consumption-related bone breakage by the Hadza, culties. First,
hunter-gatherers living near Lake Eyasi, Tanzania, are used to examine necessary to c
the organization of carcass processing behavior. Bones broken by the behaviors beh
Hadza reflect different processing behaviors and tool use (heavy vs. light logical asseml::
duty) and occur in different contexts (kill sites and residence camps). and consum p1
These observations indicate that carcass processing by the Hadza is
organized around considerations of (1) animal size and (2) later process-
rnents are cool
ing options, for example, nutrient extraction through bone smashing and or boiling of 1:
bone boiling. The goal of Hadza processing decisions appears to be to related bone
maximize nutrient returns while minimizing field processing,1transport, Should we, fo:
and social costs. Finally, the role of cooking options in structunng Hadza processing taF
carcass transport and breakage decisions is used to construC:t a prelimi- lates because t
nary evolutionary model of carcass transport and processing for groups A related p1
with pre-fire, post-fire and post-bone-boiling technologies. This model how one set<
suggests that the invention of nutrient extraction technologies such as immediate cor
roasting and bone boiling forced changes in other aspects of hominid ing decisions.
carcass processing strategies. butchery and
and particula1
structure earIii
Introduction size and cook
Numerous actualistic studies are available to help differentiate fragmentation
hominid-induced bone frequency and modification patterns from those cre- organization c
ated by "natural," nonhominid agents (e.g., Behrensmeyer et al. 1989; Binford gists are poor]
1981; Blumenschine 1986, 1989; Bonnichsen 1979; Bunn 1981, 1989; Oliver that explain h
cessing beha '\
From Bones to Behavior: Ethnoarchaeological and Experimental Contributions to the Interpretation of archaeologica:
Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional define how th
Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights ous carcass pr
reserved. ISBN 0-88104-076-2. breakage anc
200
From Butchery to Consumption I 201
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 behav-
ior (Binford 1987), where otherwise related activities are artificially parti-
tioned, resulting in a patchwork of data but no appreciation of how several
adz a: behaviors may be integrated into a decision-making strategy. With a focus on
differentiating between hominid- and nonhominid-induced patterns, most
ry 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;
man behavior, Johnson 1985). Carcass processing as a whole, a set of related behaviors from
n how cultural field butchery and transport to cooking and consumption, has not been given
::ation patterns, explicit treatment (but see Binford 1978).
hery-, cooking- The episodic treatment of carcass processing has created a number of diffi-
by the Badza, culties. Ffrst, it has not provided archaeologists with the referential tools
sed to examine necessary to develop expectations about, much less to identify, the various
broken by the behaviors behind bone frequency and damage patterns in most zooarchaeo-
(heavy vs.light logical assemblages where post-transport processing associated with cooking
idence camps). and consumption rri.ay have occurred. We do not know how different ele-
r the Badza is ments are cooked, which bones are broken prior to cooking, whether roasting
nlater process- or boiling of bone creates· visible damages, how cooking- and consumption-
e smashing and
ppears to be to related bone breakage varies by element, taxon, and cooking technique.
sing, transport, Should we, for example, expect bones of smaller ungulates to enter the post-
ucturing Badza processing taphonomic system more fragmented than those of larger ungu-
truct a prelimi- lates because they are less durable?
sing for groups A related problem this treatment engenders is difficulty in understanding
ies. This model how one set of processing options (e.g., dry meat for storage or trade vs.
ologies such as immediate consumption) inay influence earlier or later transport and process-
·cts of hominid ing 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
> differentiate
fragmentation for large or small ungulate limb bones? Finally, because the
·om those ere- organization-of carcass processing behavior has not been defined, archaeolo-
. 1989; Binford gists are poorly prepared to develop evolutionary carcass processing rp.odels
, 1989; Oliver that explain how cultural and technological innovations might influence pro-
cessing behaviors and how those changes in processing behavior maybe
1e Interpretation of archaeologically -defined. vVe need to develop models, for example, that will
tions, Occasional define how the invention and use of fire for cooking may have altered previ-
~rsity. All rights ous 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 The third obj
have altered previous ·carcass processing strategies where the only cooking Hadza carcass
technique was roasting. . tions of some 1
Experimental approaches may provide some solutions to the first problem ative of partie
(e.g., definition of fire-related damages and consumption-rel~ted bone break- carcass proces
age). Clearly, however, the most fruitful way to assess carcass processing as a cooking (boilir
nutrient extraction strategy is with ethnoarchaeological observations. Bone sions, serves a:
modification behaviors have been discussed with varying degrees of detail for cessing for prE
a few groups including the Australian Aborigines (~inford 1983; Gould 1967; nologies. Expe'
O'Connell and Marshall 1989), San (Bartram 1993; Bunn 1989; Yellen 1977), extraction tech
Nunamuit (Binford 1978, 1981), Cree (Bonnichsen 1973; Zierhut 1967), and bone breakagE
Dassanetch (Gifford-Gonzalez 1989). With few exceptions, however, those Pleistocene hm
studies have not approached carcass processing from an organizational per- elements becm
spective, and nm1e provide frequencies of variability in bone breakage behav- smashing and
ior, tools employed, or the resultant damages so that archaeological expecta- parts. Increase'
tions can be developed. fire-based innc
Here I present observations on butchery-, cooking-, and consumptio~ underscore thE
related bone breakage behavior of the Hadza, a small group of hunter- help understm
gatherers living near Lake Eyasi, Tanzania. There are three objectives to organization o
presenting these observations. First, the bone breakage data for each behav- face value, the
ioral context are presented by the animal size class, the frequency of element Pleistocene hor
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 Me
variable in field processing and transport decisions (Bunn et al. 1988;
O'Connell et al. 1990). Data presented here suggest that two other variables The
guide processing and transport decisions, carcass size and later processing rugged semim
options, particularly cooking (see also Bunn et al. 1988). south and east
The second objective is to show that the ultimate goal of many Hadza pological stud
carcass processing decisions (guided as they are by carcass size and later pro- and Bartram 1
cessing options) may be to maximize nutrient returns while minimizing field Woodburn 196
processing and transport costs. Note, however, that I am not explicitly testing bone breakagE
an optimization model; alone, data presented here are insufficient to test such through Octob
a model. Rather, I merely observe that nutrient maximization and reduction of archaeology Pr
field processing costs are reasonable (though perhaps ad hoc) explanations of Direct obser
some carcass ~rocessi~? behavi.ors. These observation.s on the. full range of (Table 12-1).3 1
carcass processing deciSions are 1mportant, not because 1n sorrw 1nstances they more easily tr•
are at odds with previous optimization explanations of Hadza carcass camp, or (3) re:
processing that dealt only with transport (i.e., O'Connell et al. 1990) but of the distal lin
because they establish a broader behavioral context for transport decisions. that provided
Nor do I assume all Hadza carcass processing behavior to be rational and eco- consumption a
nomically based. Several observations suggest that Hadza treatment of dik- behavior wher
dik (Rhynchotragus kirki) and impala (Aepyceros melampus) carcasses may be . to facilitate pic
mitigated by social costs associated with the inability to adequately share such (hearth-sizing;
small animals. Other observations and informant data indicate that taste may sumption if th
also be an important variable in explaining treatment of some animals, partic- breakage.
ularly zebra (Equus burchelli) and wildebeest (Connochaetes taurinus). Parts of mas
~ boiling might The third objective is to point out the obvious archaeological implications of
! only cooking Hadza carcass processing behavior. In part, this objective is met with descrip-
tions of some bone modifications and frequency patterns that may be indic-
~firstproblem ative of particular processing behaviors. Additionally, the organization of
~d bone break- carcass processing by the Hadza, particularly the apparent importance of
Jrocessing as a . cooking (boiling and roasting) in structuring earlier carcass processing deci-
rvations. Bone sions, serves as a basis for a preliminary evolutionary model of carcass pro-
!es of detail for cessing for pre-fire, post-fire, pre-bone-boiling, and post-bone-boiling tech-
3; Gould 1967; nologies. Expectations derived from this model suggest that as new nutrient
; Yellen 1977), extraction techniques developed, changes occurred in carcass transport and
1ut 1967), and bone breakage patterns. In particular, the expectations suggest that Plio-
owever, those Pleistocene hominids lacking fire would have preferentially transported limb
1izational per- elements because the only available nutrient extraction techniques of bone
·eakage behav- smashing and gnawing would have precluded efficient processing of axial
)gical expecta- parts. Increased transport of axial parts is expected for hominid groups after
fire-based innovations in nutrient extraction technology. These expectations
consumption- underscore the difficulties in using carcass processing by modern groups to
up of hunter- help understand the past without defining the variables that structure the
! objectives to organization of a particular behavior. Thus, the data suggest that, if taken at
)r each behav- face value, the Hadza carcass transport decisions are a poor analogue for Plio-
ncy of element Pleistocdne hominids (see also Bunn et al. 1988; O'Connell et al. 1988a, 1990).
i, and the type
,tanding of. the
1 to be a major Methods
n et al. 1988;
)ther variables The Hadza are a small group of hunter-gatherers occupying the
ter processing 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 anthro-
many Hadza pological study for many years (Bunn, this volume; Bennet al. 1988; Bunn
and later pro- and Bartram 1990; O'Connell et al. 1988a, 1988b, 1989, 1990; Vincent 1985;
lnimizing field Woodburn 1964, 1968a,1968b,1970, 1972). Observations on carcass processing
plicitly testing bpne breakage reported here were made during the dry season (August
~n t to test such through October) as part of the 1988 University of Wisconsin-Madison Ethno-
Ld reduction of archaeology Project directed by H. T. Bunn.2
xplanations of Direct observations of bone breakage were made for parts of 24 animals
~ full range of (Table 12-1).3 Butchery is defined as the sectioning of a carcass to (1) create
instances they more easily transportable units, or (2) facilitate carcass part distribution in
-Iadza carcass camp, or (3) remove less desirable units from a larger body part (e.g., severing
· al. 1990) but of the distal limb from the upper limb). This definition is more restricted than
Jort decisions. that provided by Lyman. (1987:252) in that I regard cooking preparation and
:ional and eco- consumption as separate behaviors. Cooking preparation breakage is a limited
ltmentof dik- behavior where the bone is broken immediately prior to cooking specifically
:asses may be to facilitate placement in a cooking pot (pot-sizing) or on a fire for roasting
:ely share such (hearth-sizing) ..Finally, a bone is defined as having been broken for con-
that taste rhay sumption if the marrow and/ or cancellous tissue is eaten immediately after
nilnals, partic- breakage. .
zus). Parts of most carcasses (N = 15) were followed from primary field butchery,
204 I J. S. Oliver
Table 12-1. General Carcass Procurement Data Eyasi) carca
HadzahelpE
The residenc
Taxa Size Classa Sex Age Procurement Method
the bones aJ
those preset
Dik-dik la pb Adult Evening hunt Is anzu procE
Dik-dik la F Juv Day scavenge Bone brea
Impala 2 F Juv Day scavenge ahammerst<
Zebra 3b M Adult Night hunt
Zebra Night hunt
panga or axE
3b M Juv
Wildebeast 3b M Adult Night hunt knife, (5) WJ
Zebra 3b M Adult Night hunt before. and ;
Zebra 3b Trade w I Isanzu technique, a
Zebra 3b pb Adult Isanzu night hunt quent fractu
Zebra 3b Trade w I Isanzu initial break
Impala 2 F Adult Day hunt and site ofb
Zebra 3b M Adult Night hunt in addition
Buffalo 4 M Adult' Night hunt struck to bn
Zepra 3b F Adult Night hunt of cooking
Zebra 3b M Juv Day scavenge removal fro:
Buffalo 4 Adult Day scavenge times were I
Zebra 3b M Adult Night hunt
Warthog 2 Adult Day hunt
Dik-dik la M Day scavenge
Impala 2 M Juv Evening hunt
Impala 2 Adult Day hunt
Cow 4 Trade w I Tatoga First, the ge
Baboon 2C F Adult Evening hunt carcass pro'
Impala 2 F Adult Day hunt sumption, a
.· I during butc
asize class definitions follow those provided by Bunn et aL (1988: Table 1). la = 7-9 kg; lb = dence camp
9-23 kg; 2 = 23-113 kg; 3a = 113-204 kg; 3b = 204-340 kg.; 4 = 340-907 kg; 5 = 907-2722 kg;
description~
6 = > 2722 kg.
ous breakag
bWith fetus.
cThough technically baboons belong in size class lb, this baboon was placed in size class 2
Hadza m1
because baboon bones are significantly stronger than those of other size 1 animals. hunting frc
encounter l
scribed by I
1990), Addi
to secondary butchery and defleshing,in camp, to preparation for cooking, to services for
cooking, and to final consumption. Breakage of bones from parts of other sharing wit
carcasses were observed at various stages from butchery to consumption (N = Hadza area
8). All parts of only one carcass, an adult female dik-dik, were "tracked". com- quently pro
pletely from initial butchery through complete consumption. N = 10, 41
Breakage events occurred at residence camps and kill sites. Snack sites like (Phacochoen
those reported by Bunn and colle9-gues (1988) were not observed. However, a and baboon
few bone breakage events (N = 3) were observed after successful begging for only other
food at a neighboring Tatoga (a pastoralist group) camp and at an Isanzu (Rhynchotra
(agriculturalists, most of whom live in camps at the southern end of Lake 8.3%; Table
Eyasi) carcass processing station4 where nine limb bones from a zebra the
Hadza helped butcher and transport were broken and the marrow consumed.
rement Method
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
lng hunt Isanzu processing camp were transported to a Hadza camp.
:cavenge Bone breakage techniques include (1) striking a bone lying on an anvil with
:cavenge
a hammerstone, (2) clubbing bone onto an anvil, (3) chopping at a bone with a
thunt
thunt pariga or axe, (4) hacking at a bone (with or without the use of an anvil) with a
t hunt knife, (5) wrenching a bone, and (6) gnawing. Bone breakage occurred both
thunt before and after cooking. Many limb bones were initially broken, using one
~ w/ Isanzu technique, and later the exposure of medullary bone was increased by subse-
unighthunt quent fracture, sometimes using a different technique. Direct observations of
~ w/ Isanzu initial breakage techniques for 879 axial and limb bones by animal size class
mnt and site of breakage are reported here (Table 12-2; Figure 12-1). Data recorded,
thunt in addition to breakage technique, are (1) the number of times a bone was
thunt struck to break it for marrow I cancellous tissue consumption, (2) the method
thunt of cooking (Table 12-3; Figure 12-2), and (3) the cooking time until first
scavenge removal from fire or boiling pot for initial consumption (subsequent cooking
;cavenge times w~te not recorded).
thunt
mnt
:cavenge Results
lng hunt
mnt Observations of bone breakage are divided into three sections.
~ w/ Tatoga First, the general features of the sample are presented (Table 12-1). The three
lng hunt carcass processing behaviors, butchery, preparation for cooking, and con-
mnt
sumption, are then discussed. The frequencies of axial and limb bones broken
during butchery, cooking preparation, and consumption at kill sites and resi-
l = 2-9 kg; lb = dence camps are provided in Table 12-2 and Figure 12-1. Finally, general field
= 907-2722 kg;
descriptions of breakage and cooking-induced damages created by the vari-
ous breakage techniques are presented.
:i in size class 2 Hadza meat procurement methods include hunting (day and night ambush
mals. hunting from hunting blinds near game trails and water holes, and day
encounter hunting) and scavenging. These hunting methods have been de-
scribed by Bunn and colleagues (1988) and O'Connell and colleagues (1988a,
1990). Additional procurement methods observed in 1988 include performing
or cooking, to services for the Tatoga (a pastoralist group; one observation) and trading and
:Jarts of other sharing with the Isanzu (agriculturalists living to the south who enter the
mmption (N = Hadza area in the dry season to hunt; three observations). The most fre-
tracked" com- quently procured taxa for the breakage observations are zebra (Equus burchelll;
N = 10, 41.7%) and impala (Aepyceros melampus; N = 5, 20.8%) .. Warthog
nack sites like (Phacochoerus aethiopicus), wildebeest (Connochaetes taurinus), cow (Bas sp.),
d. However, a and baboon (Papio cynocephalus) are each represented by one individual. The
Jl begging ,for only other taxa represe?ted by more than one individual are dik-dik
. at an Isanzu (Rhynchotragus kirki; N = 3, 12.5%) an.d cape buffalo (Syncerus caffer; N = 2,
1 end of Lake
8.3%; Table 12-1).
KILL SITE RESIDENCE CAMP

Panga Hstn& Club& No Site Panga Hstri& Club& No Site
Wrnch Knife I Ax Anvil Anvil Gnaw Data Total Wrnch Knife I Ax Anvil Anvil Gnaw Data Total TOTAL
LlffiD u u u u u u u u u u u ;; u u u ::J J
Size Class IV
Axial 0 0 0 0 0 0 0 0 0 0 0 14 0 0 0 14 14
:Limb 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
r- --
KILL SITE RESIDENCE CAMP

Panga Hstn& Club& ------
No Site Panga Hstn& Club& No Site
Wrnch Knife /Ax Anvil Anvil Gnaw Data Total Wrnch Knife /Ax Anvil Anvil Gnaw Data Total TOTAL
CONSUMPTION
Size Class I
Axial 0 0 0 0 0 0 0 0 2 0 0 2 0 54 0 58 58
Limb 0 0 0 0 0 0 0 0 0 7 0 2 0 2 2 13 13
Size Class II
Axial 0 0 0 0 -0 0 0 0 0 0 0 11 0 0 0 11 11
Limb 0 1 - 0 2 0 0 0 3 0 0 0 17 1 0 0 18 21
Size Class III
Axial 0 0 0 6 2 0 0 8 0 0 1 1 0 0 0 2 10
Limb 0 0 0 0 0 0 0 0 0 0 1 37 1 0 0 39 39
Size Class IV
Axial 0 0 0 0 0 0 0 0 0 0 0 0 o· 0 0 0 0
Limb 0 0 0 0 0 0 0 0 0 0 0 19 0 0 0 19 19
BEHAVIOR SUM ~
0
Size Class I
Axial 0 2 0 0 0 0 0 2 28 93 0 2 0 54 18 195 197
so:l
limb 0 0 0 0 0 0 0 0 0 7 0 2 0 2 _2 13 13 c
.-t-
('J
Size Class II
!:J'
Axial 0 28 0 0 0 0 0 28 0 85 42 11 0 0 0 138 166 ro
>;
Limb 1 1 0 2 0 0 0 4 0 0 9 20 3 0 0 32 36 '--<
r-1-
Size Class III 0 0
Axial 30 0 185 6 2 0 75 298 0 0 67 11 0 0 0 78 376 n
Limb 0 0 0 0 0 0 0 0 0 0 1 42 1 0 0 44 44 0
:::::s
Size Class IV fJJ
Axial 0 0 14 0 0 0 0 14 0 0 0 14 0 0 0 14 28
limb 0 0 0 0 0 0 0 0 0 0 0 19 0 0 0 19 19
sc
'ij
;:t
All Classes 0
Axial 30 30 199 6 2 0 75 342 28 178 109 38 0 54 18 425 767
Limb 1 1 0 2 0 0 0 4 0 7 10 83 4 2 2 108 112 -N:::::s
TOTAL 31 31 199 8 2 0 75 346 28 185 119 121 4 56 20 533 879 0
"'
208 I f. S. Oliver
a. BUTCHERY-RELATED BREAKAGE of AXIAL BON.E~ b. BUTCHERY -RELATED BREAKAGE of LIMB BONES
. %
%
40
20
Ill N Ill N Ill N Ill N

KILL SITE RESIDE:\CE CAMP KILL SITE RESIDENCE CAMP
c. COOKING-PREPARATION BREAKAGE of AXIAL BONES d. COOKING-PREPARATION BREAKAGEofLIMB BONES

100
Size Class I
80 80
Axial
60 60 Limb·
% %
40 40 Size Class II
20 20 Axial
Limb
Ill N 111 N •Ill N Ill N
KILL SITE . RESIDEl'iCE CAMP KILL SITE RESIDENCE CAMP Size Class III
e. CONSUMPTION-RELATED BREAKAGE of AXIAL BONES 100
f. CONSUMPTION-RELATED BREAKAGE of LIMB BONES Axial
100
Limb
80 80
Size Class N
60 60
% % Axial
40 40
Limb
20
20
All Classes
Ill N Ill N lil N III N
Axial
KILL SITE RESIDENCE CAMP KILL SITE RESIDENCE CAMP Limb
!§GNAW mJ CLUB & ANVIL IJjlpANGA I AX I HSTN & ANVIL D KNIFE .WRENCH I TOTAL
Figure 12-1. Axial and limb bone breakage frequencies by breakage tech-
nique, size class (I-IV), and site of breakage that occurred during butch-
processing
ery (a, b), cooking preparation (c, d), and consumption (e, f).
camps (but
breakage (1
Butchery-related Breakage Complete ti
one hunter
Though there are some carcass size and transport-dependent occupants c
butchery patterns (Bunn et al. 1988a; O'Connell et al. 1988a; 1990), carcasses carcasses m
are generally butchered into similar units. After skinning, forelimbs are usual- be summon'
ly removed complete by cutting between the medial scapulae and lateral rib In contra~
surfaces. Hindlimbs are also typically removed as units by cutting through the occurred at
acetabuluw.-fe:rnora joint with a knifei axe, or panga. The distal limbs (tarsals/ b). Transpor
carpals-metapodials-phalanges) of smaller carcasses (i.e, impala) are some- ing the ver-
times removed by chopping at the wrist and ankle joint with a panga or axe. removal. Ril
Rib slabs are removed by hacking through the neck and costal rib area with a they make c
knife or panga, or wrenching the slab laterally, so that some ribs snap at the Further pro'
neck/head and others dislocate from the vertebrae. The vertebral column is arrival.in Ci
usually sectioned into three to five units. Small carcasses may be sectioned these parts t
into cervical-skull, thoracic vertebrae, and lumbar-pelvis-sacrum units. In the efficient
larger carcasses skulls and the pelvis-sacrum may form additional units. site but tha
Carcass size clearly influences the location and intensity of initial butchery. Similarly, li
Size I and II carcasses are usually (87.5% of the carcasses with transport and nearly com1
fLIMB BONES Table 12-3. Frequency of Roasting, Boiling, and Direct Consumption
of Marrow/Cancellous Tissue
L
CAMP
Ill
E of LIMB BONES
IV
Size Class I
Pot-sized
194
Break
Consume
Roast
0
Roast
Break
Consume
30
Break
Consume
Boil
16
Boil
Break
Consume
38
Break
Consume TOTAL
19 284
Axial 54 0 ~7 0 31 0 112
Limb 0 0 0 0 12 0 12
Size Class II
Axial 39 0 30 0 7 0 76
Limb 10 9 4 6 0 30
Ill IV
CAMP Size Class III
of LIMB BONES Axial 72 0 8 0 0 2 82
Limb 5 0 13 5 0 19 42
Size Class IV
Axial 14 0 0 0 0 0 14
Limb i 0 0 5 7 0 7 19
All Class~s
III IV
Axial 179 0 65 0 38 2 284
CAMP Limb 15 27 16 18 26 103
TOTAL 194 92 16 56 28 387
' breakage tech-

. during butch-
c). processing information reported here) transported complete to residence
camps (but see O'Connell et al. 1990) where 85% of the butchery-related
breakage (175 out of 206 bones; Table 12-2; Figure 12-1a, b) was observed.
Complete transport obviously reduces field processing costs. The ability of
one hunter or small hunting party (two to six men, the typical number of
rt-dependent occupants of two to three hunting blinds) to easily transport size I and II
90), carcasses carcasses may minimize· transport costs because additional carriers need not
nbs are usual- be summoned from residence camps.
md lateral rib In contrast, almost all the butchery-related breakage for larger carcasses
g through the occurred at the kill site (304 of 308 bones, or 98.7%; Table 12-2; Figure 12-la,
lmbs (tarsals/ b). Transport of axial parts of larger carcasses (size ill) is facilitated by section-
la) are some- ing the vertebral column and rib slabs into manageable parts after meat
panga or axe. removal. Rib slabs of four to five ribs may be preferred for transport because
ib area with a they make convenient '1Jowls" in which smaller body parts may be placed.
JS snap at the Further processing of the partially defleshed axial parts is delayed until their
1ral column is arrival in camp where boiling and roasting occur. Moreover, transport of
· be sectioned these parts takes place in spite of their low meat and marrow yields because of
·urr1 units. In the efficient nutrientextraction techniques that would be impractical at the kill
:1l units . site but that can be employed at leisure in camp (Bunn et al. 1988:452).
.tial butchery. Similarly, limb-processing costs are minimized at the kill by transporting
transport and nearly complete limbs to camp. In fact, observation .of size ill and IV limb-
210 I f. S. Oliver
most regular]
BREAKAGE of COOKED & UNCOOKED LIMB BONES
related break
100
correlated wi
80 ery-related b1
60 most pot-sizi:
%
4D All Hadza br
men seemed 1
20
The decisic
0
ll UI IV cooking appE
SIZE CLASS ments (Table~
elements obs,
50% of the ot
whereas cool<
Figure 12-2. Comparison of breakage frequencies for cooked and un-
89.1% and 10
cooked limb bones by animal size class.
carcass axial ·
should also b
cooking time~
carrying techniques suggests that the relatively low-utility distal limbs may be
The frequer
transported simply because they make good handles to balance limbs on the
with carcass E
shoulders of carriers. Again, roasting and boiling permit extraction of nutri-
ing pots with
ents from otherwise low utility parts. Thus, delaying the processing of distal
size II carcass
limbs and axial parts and the transporting of the parts apparently reduces
of size III can
field processing and transport costs as well as increasing nutrient yield.
so they wou:
Carcass size also strongly influences choice of tools in butchery-related
suggest that 1
breakage (Table 12-2; Figure 12-la, b). Lighter-duty tools, that is, knives, are
solely to fit in
used to section size I and II carcasses. Ofthe 201 size class I and II axial
tant variables
elements broken during butchery, 168 were broken with knives (83.6%), and
ease of consm
26 (ribs) were broken without tools by wrenching rib slabs laterally. In con-
As with toe
trast, heavy-duty metal tools were used in butchery-related breakage o£ 87.1%
for pot-sizing
(203 out of 233; Table 12-2) of the size class III and IV axial element~/(Figure
frequently USE
12-1a).
breakage (for
Butchery-related breakage of only five limb bones, all carpals or tarsals
accomplished
from size II carcasses, was observed. In all but one case a panga was used to
hammerstonE
chop through the joint. (One astragalus was broken by a wrenching action.)
carcasses. FUJ
There are probably two reasons why only wrist and ankle joints of size II
sumption the:
animals were treated in this manner. First, these joints of larger-sized animals
cancellous tis
are not as easily severed with machetes or axes. Second, limbs of size II
only as many
carcasses need t~ be broken before they will fit into cooking pots, while size I
halves. This s
limb bones fit without breakage. Regardless of site or tool employed, field
Thus, as sugg,
butchery is usually a male activity. Both males and females may deflesh
by behavioral
transported elements, though males usually deflesh larger limbs.
Breakage as Cooking Preparation: Pot- and Hearth-Sizing Co

M;
Twenty-two percent (N ~ 194) of observed bone breakage events
all Badza (Ta
were directed at reducing bone size so they would fit into a metal boiling pot
or, rarely, in a small hearth (Table 12-2). (The mean dimensions of the two cooked, but fc
tion was obs,
most regularly used pots are estimated to be 26 x 15 em.). As with butchery-

related breakage, the tool used in pot-sizing breakage appears to be directly
correlated with carcass size and element type (Figure 12-1c, d). Unlike butch-
ery-related breakage, women who tended boiled foods were responsible for
l
most pot-sizing breakage; children sometimes assisted women in pot-sizing.
All Hadza broke ribs so that they would 'fit in hearths, though children and
men seemed to hearth-size more often than women.
a
The decision whether or not to break bone as a direct preparation for
cooking appears to be determined by carcass size, particularly for axial ele-
ments (Tables 12-2 and 12-3). As carcass size increases, the frequency of axial
elements observed being pot-sized increases. Pot-sizing accounts for about
50% of the observed class I and II breakage of cooked axial bone (93 of 188),
whereas cooked size III and IV axial bones were first pot- or hearth-sized in
oked and un- 89.1% and 100% of the observations, respectively. Greater breakage of large
carcass axial parts not only facilitates placement in pots or in a hearth, but it
should also have the effect of increasing nutrient extraction and decreasing
cooking times by increasing the bone surface areas.
imbsmaybe The frequency of limb bones broken for pot-sizing is not as neatly correlated
limbs on the with carcass size. Limb bones of size !carcasses easily fit into the metal cook-
:ion of nutri- ing pots without breakage. In contrast, just over 32% of cooked limb bones of
3ing of distal size II cdrcasses were pot-sized, but only about 12% of the cooked limb bones
ntly reduces of size III carcasses were pot-sized. No size IV limb bones were broken solely
yield. so they would fit into cooking pots (Table 12-2; Figure 12-1d). The data
hery-related suggest that the decision on whether or not to boil a limb bone and break it
;, knives, are solely to fit in a pot is not mitigated by pot and limb size alone. Other impor-
and II axial tant variables may include marrow volume and how volume influences the
(83.6%), and ease of consuming uncooked versus cooked marrow (see below).
rally. In con- As with tool selection for butchery.;.related breakage, smaller tools are used
:age of 87.1% for pot-sizing breakage of smaller carcasses while heavy-duty tools are more
len ts (Figure frequently used on larger carcasses (Table 12-2; Figure 12-1c, d). All pot-sizing
breakage (for which the tool was recorded) for size I carcasses was easily
tls or tarsals accomplished with a knife. Heavier-duty tools such as machetes or axes and
was used to hammerstones were more frequently employed to break bones of larger
hing action.) carcasses. Further, in contrast to bones broken for immediate marrow con-
1 ts of size II
sumption that were often struck repeatedly to expose more marrow and
ized animals cancellous tissue (see below), those broken for boiling were usually struck
bs of size II onJy as many times as necessary to break the bone into proximal and distal
, while size I halves. This significantly reduced fragmentation of size I and II limb bones.
?loyed, field Thus, as suggested by Binford (1981), details of breakage techniques may vary
may deflesh by behavioral context.
Consumption-related Breakage of Cooked and Uncooked Bone

'ng
~vfarr,ow and cancellous bone were eaten both raw and cooked by
1kage events all Hadza (Table 12-3; Figure 12-2). Marrow from size I animals was always
tl boiling pot cooked; but for 16.7% of the limbs fro~ size II carcasses for which consump-
s of the two tion was observed, the marrow was eaten uncooked (N = 5). For larger
2121 ]. S. Oliver
animals, howeve·r, the marrow was typically eaten uncooked. Of the 42limb Only two 1i:
bones from size III ·animals broken for marrow, 57:1% were uncooked. broken with
Seventy-four percent (14 out of 19) .of the size IV limb bones were broken
fresh. E
Though some meat was usually consumed at death sites ~f larger animals,
only a little over 6% of the observed consumption-related breakage events 1
took place at death sites. Three limbs of an impala were broken for marrow at based on fie:
the kill site as were eight ribs of two zebras. Informants report that skulls, above and tl
. mandibles, and at' least some ribs are almost always roasted at kill sites. Both damage to'
O'Connell and colleagues (1988a, 1988b, 1990) and Bunn and colleagues (1988) Fracture feaf
note a greater frequency of marrow consumption at kill and snack sites than present on
reported here. Although I was not present at the breakage events, informants described fc
reported and site inspection (Bunn, personal communication) confirmed that technique n
the skulls and mandibles of several zebra and several limb elements and ribs bones proce~
of one kudu (Trageiaphus strepsiceros) were consumed at some kill sites. Most of t
As stated above, variables other than pot and limb size may influence cook- (panga and
ing and cooking preparation decisions. The other·variables include marrow on anvils), 1
cavity volume and how volume influences the ease of consuming uncooked wrenching.
versus cooked marrow. That is, marrow volume of size I limb bones may be tures typica
too small to remove efficiently without cooking. Cooking tends to solidify the Selvaggio 15
marrow, making it easier to consume with less spillage than if the bone were ological (e.g
cracked raw. Limbs of size II animals apparently represent a threshold of scars at the 1
marrow volume where some (16.7%) may be broken and marrow consumed flake scars a
directly without cooking. The frequency of consumption of uncooked limb smooth frac
marrow increases for size class III and IV animals, suggesting that the marrow fracture fror
volume is large enough to risk some spillage (Figure 12-2). Again, a nutrient outline, and
extraction maximizing strategy may be indicated. This differential treatment well known
of marrow and cancellous tissue based on carcass size further underscores the Morlan 198(
importance of within-bone nutrients when analyzing skeletal partfrequency Behrensmey
patterns as recently noted by Marshall and Pilgram (1991). producing s
Marrow condition may also influence a decision to cook a bone as illus- or frequenc
trated by the processing of a female impala late in the dry season (the last 1991). That
carcass processing event observed in 1988). Immediately after gutting the relation to c
carcass, the left distal hindlimb was removed and the metatarsal broken. Once other fractUl
inspected, the bone was tossed to the ground, the translucent, runny marrow Because c
dynamic lcia
uneaten. The Badza teenage boys then pronounced the marrow to be bad.
After the carcass was butchered, however, they proceeded to roast the ferrtur and anvil b
notches (wr
and tibia. They then cracked open the ·bones, stated the now white, solidified
marrow to be fine, and proceeded to eat it. This was the only time I saw a Instead of tl
bone broken specifically to evaluate the marrow. by hammer
Tool selection for consumption-related breakage did not vary markedly for notch or fla
animals size class II and larger (Table 12-2; Figure 12-1e, f). Bammerstones fracture (see
and anvils were used to break over 93% of size II, ill, and IV axial and limb Of the 11:
elements broken for marrow and cancellous tissue. Bones from size I animals one location
were treated in a very different'manner, however. Almost 79% of the size I buffalo lim1
bones broken during consumption (N =56; 54 ribs and 2 metapodials) were positive relc
broken by Badza gnawing, and two ribs were broken by a wrenching action. number ofiJ
Jf the 42 limb Only two limb bones were broken with a hammerstone, and seven were
!re uncooked. broken with knives.
; were broken
Bone Breakage Damage Morphologies
arger animals,
=akage events The following descriptions of bone breakage morphologies are
for marrow at based on field observations of bones from both the breakage events reported
>rt that skulls, above and the inspection of bone in refuse piles of occupied camps. Although
kill sites. Both damage to axial parts is mentioned, damage to limbs is emphasized here.
.leagues (1988) Fracture features described for a given breakage technique are not necessarily
tack sites than present on every bone processed in a similar manner. Rather, damages
tts, informants described for each butchery, cooking, and marrow consumption breakage
:onfirmed that technique represent a composite signature of observations of a number of
nents and ribs bones processed in a similar manner.
11 sites. Most of the observed breakage techniques involved dynamic loading
1fluence cook- (panga and axe chops, knife hacks, hammerstone impacts and clubbing bone
elude marrow on anvils), one involved static loading (gnawing), while a third involved
ling uncooked wrenching. Hammerstone and anvil breakage events produced fracture fea-
bones may be tures typical of those noted in the experimental (e.g., Blumenschine and
to solidify the Selvaggio1988; Bonnichsen 1979; Bunn 1989; Johnson 1985) and ethnoarchae-
the bone were ological/(e.g., Binford 1981; Bonnichsen 1973) literature: (1) concentric flake
1 threshold of
scars at the loading and rebound points, (2) broad impact notches, (3} negative
'OW consumed
flake scars at the loading points, (4) hackle marks or other stress features, (5) a
~ncooked limb
smooth fracture surface interrupted only by (6) stress features indicative of
lat the marrow fracture front direction, (7) a smooth, curvilinear (sometimes "spiral") fracture
ain, a nutrient outline, and (8) impact pits and scratches near the loading points. Though it is
ttial treatment well known that carnivore activity (e.g., Binford 1981; Haynes 1983; Hil11989;
nderscores the Morlan 1980; Potts 1988) and some geological processes and trampling (e.g.,
Jart frequency 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
bone as ill us- or frequency of hammerstone and anvil bone breakage (Oliver 1986, 1989,
~ason (the last 1,991). That is;: it is the association of impact features and their context in
er gutting the relation to oth~F damages that defines hammerstone impact damage or any
1broken. Once other fracture agent (Oliver 19.91).
·unny marrow Because chopping and hacking of bone with metal tools are essentially
ow to be bad. dynamic loading actions, damages similar to those produced by hammerstone
:>ast the femur and anvil breakage are created, including negative flake scars and impact
hite, solidified notches (which, if present, tend to be linear rather than oval in appearance).
r time I saw a Instead of the irregular crushed area with a concentric ring of flakes produced
by hammerstones, metal-tool-induced breakage produces a characteristic
r markedly for notch or flat-sided area where the blade cut through the bone prior to bone
fammerstones fracture (see also Lyman 1987).
:txial and limb Of the 112 limb bone breakage events, 47 (42%) were struck at more than
size I animals one location. The mean number of blows struck to dik-dik, Lrnpala, zebra, and
ro of the size I buffalo limb bones is 1.7, 7.1, 9.9, and 14.6, respectively, clearly indicating a
tpodials) were positive relationship between body (and therefore bone) size and the average
nching action. 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 could not be
density and because several bone"s were pounded on not to obtain marrow, row undoubt,
but simply to occupy time.) Repeated blows to different locations on a bone Cancellous
has a number of results. First, it damages the bone to a greater extent than a consisted of 1
single impact at one location. More impact points, fracture lines, and ulti- hcrmmerstone
mately bone fragments are created by repeated impact events (see also Bunn knife or pang
1989). Repeated blows serve not only to straighten the fracture front but also This action n
create many more impact notches than suggested by the early experimental pieces wi tl1 V
literature (e.g., Bonnichsen 1979). Finally, repeated blows after breakage often nique was to
split limbs across the articular surface. , the cortical t
Although most larger limb bones were struck so that diaphyses fragmented impact dents
into numerous pieces, bone tubes were created by striking both the proximal and scratches
and distal diaphyses of several size I and IT limb bones. Aft~r marrow was This suggesh
sucked out, the exposed ends were chewed; without careful inspection of frac- much tissue a
ture 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. D
Post-boiling processing of these limbs for cancellous tissue may create addi-
tional impact damages. n
Boiling bone does not yield visible damage, nor does breaking previously ior-specific br
boiled bone result in a diagnostic breakage morphology. Observed boiling set, those bel
times are short, however, usually less than 11 minutes. Observation of boiling the Hadza. T.
ended after the bone was first removed from the pot and some tissue con- vs. prehistori1
sumed. Many bones were subsequently returned to the pot, but the boiling in all other ca
times were not recorded. Longer boiling times may result in visible damage. 1981; Gifford
Roasting slightly chars some limb bones, but many show no charring at all. Yellen 1977 <
Further, it should be noted that the Hadza typically do not discard bone into theless, the r
hearths. goals serve a
Breakage of roasted limb bones does, however, yield breakage feayures not processing he
at all like those created when green, uncooked bone is broken. Rather than the tion of carcas
smooth, curvilinear fractures typical of hammerstone breakage of fresh limb from initial fi
bones, breakage of roasted limb bones produces jagged fracture lines that are tion-may h
frequently oriented transverse to the bone's long axis; associated damage in- extraction is r
cludes hammerstone and anvil dents and scratches, and impact flakes. Over- minimized. ~
all, broken roasted bones are somewhat reminiscent of bone fractured after define those
significant desiccation or fossilization. Structural alteration of bone due to structuring c
burning (Shipii'.an 1981; Shiprnan et al. 1984) as v.rell as simple desiccation of B.unn et al. 1
outer compact bone may partially explain this fracture pattern. Microscopic available nut
examination of bones fractured after roasting and boiling would be informa- carcass proce:
tive. Gifford-Gonzalez (1989, this volume) has also noted similar bone frac-
tures in faunal assemblages from abandoned Dassanetch camps, which, she 0;
suggests, were created by cooking.
Finally, marrow and cancellous tissue extraction from broken bones created H;
diagnostic damage. Marrow removal was accomplished in a variety of ways consideration
including (1) sucking, (2) pounding th~ exposed end on a rock, and (3) using a organization
knife, stick, or bone fragment as a scraping tool. Binford (1981) reports the maximizing r
same marrow extraction techniques for the Nunamuit. Though cut marks smaller carca~
·nall y greater could not be identified in the field, _the intense use of knives to remove mar-
1tain marrow, row undoubtedly produced cut marks on the medullary wall.
ms on a bone Cancellous bone removal for consumption (both cooked and uncooked)
extent than a consisted of repeatedly striking the diaphysis near the epiphyseal end with
nes, and ulti- hammerstone on an anvil until cancelldus bone was exposed. Once exposed, a
:;ee also Bunn knife or panga was used to cut off pieces of cancellous tissue to eat or suck on.
front but also This action resulted in an area of consumption littered with cancellous limb
experimental pieces with V-shaped notches and/ or planed cancellous tissue. Another tech-
reakage often nique was to strike the exposed cancellous tissue with a hammerstone (while
the cortical bone rested on an anvil). This technique created hammerstone
~s fragmented impact dents and crushed areas of cancellous bone; the anvil left pits, dents,
the proximal and scratches on the cortical surface. Generally, little edible tissue is ignored.
marrow was This suggests bones may be transported to camps to cook and consume as
ection of frac- much tissue as possible.
llt of chewing
I and II limbs
~agmentation. Discussion
'f create addi-
The above data show that Hadza carcass processing yields behav-
1g previously ior-specific breakage patterns related to carcass size and cooking options. As a
~rved boiling set, thos~ behaviors and the resultant patterns are probably specific only to
:ion of boiling the Hadza. That is, we may not conclude that correspondence in one (Hadza
1e tissue con- vs. prehistoric pattern) breakage or frequency pattern implies correspondence
ut the boiling in all other carcass processing or cultural behaviors (see, for example, Binford
sible damage. 1981; Gifford 1980; Gifford-Gonzalez 1991; Oliver 1989; Saunders 1990; and
1arring at all. Yellen 1977 on how analogues may be used to interpret the past). Never-
ard bone into theless, the manner in which carcass size and ultimate nutrient extraction
goals serve as boundary conditions to structure the organization of carcass
e features not processing has a number of archaeological implications. First, this examina-
::tther than the tion of carcass processing behaviors has elucidated how several behaviors-
of fresh limb from initial field butchery, to transport, to cooking-preparation, to consump-
lines that are tion-may be integrated into a decision-making strategy where nutrient
d damage in- extraction is maximized, while transport, field processing, and social costs are
flakes. Over- minimized. Specific bone modification and frequency patterns may help
~actured after define those behaviors. Furthermore, the important role cooking plays in
:bone due to structuring carcass processing behavior for modern people (Binford 1978;
desiccation of Bunn et al. 1988; Gifford-Gonzalez, this volume; Yellen 1977) suggests that
. Microscopic available nutrient extraction techniques also helped determine prehistoric
d be informa- carcass processing and transport strategies.
lar bone £rae-
IS, which, she Organization of Carcass Processing
bones created Hadza carcass processing decisions are clear! y made with due
riety of ways consideration of .carcass size and later nutrient extraction options. Further, the
nd (3) using a organization of carcass processing appears to be structured by the goal of
1) reports the maximizing nutrient returns while min~mizing transport, processing, and, for
~h cut marks smaller carcasses, social costs.
216 I ]. S. Oliver
Processing and transport CO$tS for smaller carcasses are minimized by low ratios ar1
almost always transportingthe complete carcass to residence camps (five of transport sh
six dik-dik and impala carcasses for .which I have transport data). These data may be relat,
are at odds with those provided by O'Connell et al. (1990, 1991), who observ- O'Connell a
ed frequent processing, consumption-of tissue, and eleme11t abandonment at that they do
kill sites. Nevertheless, situational information for six impala and dik-dik taste good (I
carcasses reported here may help to explain the conflicting observations. commented
Specifically, the Hadza decision whether or not to transport a small carcass to sweet. 6 The i
camp seems to be made by evaluating the ability. to adequately share small transport of
carcasses with camp members. Isanzu procE
Three of the small antelope carcasses were transported complete by hunters bones in the
to relatively small camps where they were easily shared. One of the three was ing options i
found dead by a Tatoga herder and the location reported to a Hadza boy who The argur
clearly could not have consumed all of the carcass. Another scavenged dik-dik while other '
was found after the entire camp population was alerted to vultures circling diate desire:
just outside camp. The remains of this animal (all but the right hindlimb, .nature, a risl
pelvis, sacrum, and four lumbar vertebrae) were returned to camp compl~te, tiona! basis.
but not all camp members partook of this food. Most parts of the- animal were social costs c.
eaten by women and small children. One dik-dik was killed at night very near bers and thE
camp and transported complete to the camp edge where it was processed and Similarly, tb
totally consumed by two hunters without the knowledge of other camp mem- wildebeest
bers. Finally, one impala was largely consumed by several teenage boys at the N evertheles~
kill site; The teenage boys' decision to consume most of a female impala at the satisfying of
kill site was apparently made because (1) the adult hunter had given up the Maximizil
animal as lost, (2) the animal was in poor nutritional condition, and (3) they seen in the c.
would have been unable to adequately share meat from such a small carcass axial and lin
with the large Hadza camp. 5 Whether or not these situational variables also whereas axic:
explain the impala transport pattern reported by 0' Connell and colleagues from size I a
(1990) is unknown at this time, but the observations do suggest thlcit evalua- but is not n1
tion of the ability to share the kill may play a determining role in s1ze I and II limbs that a1
carcass transport patterns. marrow and
With some notable exceptions (wildebeest, giraffe, and apparently eland; subject to co:
O'Connell et al. 1988a, 1990, 1992) most carcass parts of larger animals, partic- O'Connel1
ularly zebra, are transported to camp (Bunn, this volume; Bunn et al. 1988; erentially trc
Bunn and Bartram 1990). Again, the overriding consideration seems to be ing higher-u
maximizing nutrient returns by transporting most of the animal back to camp Perkins and
where nutrient extraction (i.e., through bone smashing and boiling) could pro- Bunn 1986; I
ceed at leisure. Carcasses were butchered only to the extent that it facilitated strategy wa~
transport of most carcass parts back to camp using available carriers. With carcasses w
regards to the single wildebeest killed in 1988, consumption of most limb Bunn, this v
marrow at the kill site and transport of axial parts are consistent with earlier however, it'
observations of the unique treatment of wildebeest (Bunn et al. 1988; strongly infl
O'Connell et al. 1988b, 1990). According to O'Connell and colleagues (1990, choose to t:
citing Metcalfe 1989), the transport pattern is explicable in that the probability defleshing c
that a particular element will be transpbrted is related to the ratio of edible to high field f
inedible tissue and the processing costs. Carcass parts with high edible to scraps, can t
inedible tissue ratios are likely to be transported, while bones from those with bral column:
From Butchery to Consumption 1217
ninimized by low ratios are more likely to be abandoned at the kill site. Processing costs and
::amps (five of transport should be inversely related. Thus, wildebeest transport patterns
:a). These data may be related to taxa-specific inedible to edible tissue ratios as suggested by
), who observ- O'Connell and colleagues (1990). Nevertheless, Hadza informants indicate
andonment at that they do not carry wildebeest limbs to camp because the marrow does not
a and dik-dik taste good (H. T. Bunn, personal communication). Conversely, they frequently
obsE;rvations. commented on their preference for zebra meat and marrow because it is so
nail carcass to sweet. 6 The importance of taste in transport decisions is also suggested by the
ly share small 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
ete by hunters bones in the Hadza camp further underscores the importance of later process-
: the three was ing options in transport decisions.
adza boy who The arguments that some processing decisions are economically based
·enged dik-dik while other decisions may be guided by considerations of social costs, imme-
1tures circling diate desires, and taste may ~eem incongruent, but sharing is, by its very
ght hindlimb, nature, a risk reduction strategy whose solution must be evaluated on a situa-
tmp complete, tional basis. As such, the little meat a dik-dik or impala offers may not offset
.e animal were social costs associated with dividing such a carcass fairly among camp mem-
Light very near bers and the nutritional gain on the part of those who decide not to share.
processed and SimilarlY;~ the "good taste//, of zebra meat and marrow compared to that of
ercampmem- wildebe~st may in fact have an as yet unrecognized nutritional basis.
tge boys at the Nevertheless, I do not believe all human behavior is totally rational, and some
~impala at the satisfying of immediate desires certainly occurs.
i given up the Maximizing nutrient extraction while reducing processing costs may be
1, and (3) they seen in the animal-size-dependent cooking preparation and consumption of
L small carcass
axial and limb parts. Pot- and hearth-sizing is minimal for smaller carcasses,
variables also whereas axial parts of larger carcasses are heavily processed. Cooking of limbs
.nd colleagues from size I and II carcasses apparently increases marrow extraction efficiency
st that evalua- but is not necessary for limbs of larger animals. In cor.trast to size I and II
in 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
arently eland; subject to considerable processing and fragmentation.
nimals, partic- O'Connell and colleagues (1988a; 1990)·argue that because the Hadza pref-
nn et al. 1988; erentially transport axial and other low-meat utility parts to camp (abandon-
rr seems to be ing higher-utility wildebeest and impala limb bones at kill sites) the use of the
l back to camp Perkins and Daly model (1968) to reconstruct early hominid foraging (e.g.,
ng) could pro- Bunn 1986; Bunn and Kroll 1986) behavior is open to question. This transport .
3.t it facilitated strategy was not observed in the 1988 work with the Hadza where almost all
carriers. With carcasses were transported nearly complete to residence camps (see also
of most limb Bunn, this volume; Bunn and Bartram 1990). With either transport pattern,
nt with earlier however, it appears that later nutrient extraction options, in this case cooking,
1 et al. 1988; strongly influence Hadza transport decisions. Thus, it appears that the Hadza
leagues (1990, choose to transport axial parts because (1) meat scraps remaining after
:he probabi.lity det1eshing cannot be readily removed without cooking, that is, they have a
tio of edible to high field processing cost, and (2) the available nutrients, including meat
high edible to scraps, can be readily extracted through boiling (a process to which all verte-
om those with bral 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 bones, (4) che,
field processing time is reduced by foregoing time-consuming defleshing and breakage of b
transporting entire limbs to camps w.here defleshing and limb dismember- si ve breakage
ment can take place at leisure and (2) the nutrient value of limb bones is pots, and (7) u
increased by boiling and by intensive breakage for cancellous tissue. Hadza E>ifferential
transport strategies are thus a poor analogue for pre-fire hominids. of smaller ani
Therefore, O'Connell and colleagues' (1990) criticism of carcass transport marrow. or pc
strategies of early hominids reconstructed by Bunn (Bunn 1986; Bunn and fragmented t]
Kroll1986; Potts 1988) because they do not match Ha~za transport decisions is bones, particu
an example of interpretative risks arising from the episodic treatment of of larger. anin
human behavior. There is no reason to suspect·that early hominids who better techniq
lacked cooking technology would· make the same economic decisions as limb shaft fra
groups with more efficient nutrient extraction techniques (see below). This 1991). The re·
perspective is lost if only one component of an overall carcass processing bones from li
strategy-in this case, transport-is the sole focus of attention. Furthermore, carcasses furt:
the stated Hadza concern with taste (poor tasting wildebeest marrow vs. and boiling rr
sweet zebra meat and marrow) and the apparent concern with ability to cessing of sm
adequately share small carcasses demonstrate the difficulties in applying eco- limited mids]
nomic models of human decision making. Economic concerns are obviously light gnaw da
intertwined with social concerns and the need to satisfy personal tastes. The intensi·
This does not mean that economic models such as optimal foraging seri- ed by thedeg
ously misrepresent human behavior and should not be used. Economic des, and dar
models are useful heuristic devices that serve to structure our observations ological asse1
about human subsistence behavior and may sometimes explain much of it. points, a hig:
Many of the observations reported here, for example, seem to fit predictions damage to ca
recently made by Metcalfe and Barlow (1992) in their model of the optimal marrow cons1
trade-off between field processing and transport. Significantly, however, in by numerous
cases where human subsistence behavior does not seemxational (in this case, pieces, and p:
the Hadza treatment of size I and II carcasses and wildebeest), cult4tal vari- ing intensity
ability or other previously unconsidered variables may be highlight~d. Thus, £ragmen ted 1
Hadza carcass transport behavior appears to be mitigated by cultural attitudes animal axial f
or perceptions of sharing and taste and cooking options, in addition to a The diagno
desire to optimize energy procurement while reducing field processing and the potential1
transport costs. Further, it should be noted that although variable cultural of fire has h
attitudes cannot be effectively incorporated into economic models, processing Swartkrans 0
options like cooking technology could be used (with appropriate experimental Analysis of b
data on nutrient yields of various cooking technologies) to calculate the utility animal food '
of transported loads discussed by Metcalfe and Barlow (1992). some concerr
again that thE
Bone Modification and Frequency Patterns ior and the 1
outward sigr
Generally, data presented above suggest that past carcass process- calcined. bon'
ing behaviors may be more fully understood through examination of bone cooking in tl
fracture damage. Significant breakage observations include those related to (1) about using
animal size and behavior dependent use of heavy- and light-duty tools, (2) 1987).
multiple blows struck on large limb bones to expose areas of cancellous bone, Differentia1
(3) use of knives to scrape marrow and cut cancellous tissue from broken is also of con~
From Butchery to Consumption j219
~cause (1) the bones, (4) chewing of small carcass ribs and articular ends of limbs, (5) limited
efleshing and breakage of bones from small carcasses in preparation for boiling, (6) exten-
~ dismember- sive breakage of large carcass axial parts. to facilitate placement in boiling
imb bones is pots, and (7) unique breakage of roasted bones.
:issue. Badza Differential treatment of large and small carcasses suggests that limb bones
ls. of smaller animals (for which minimal br~akage is required to either extract
ass transport marrow or pot-size) may enter the post-processing taphonomic system less
~6; Blinn and fragmented than those of larger animals. Conversely, small mammal axial
rt decisions is bones, particularly vertebrae, may be more highly fragmented than vertebrae
treatment of of larger animals. This observation underscores the necessity of developing
)minids who better techniques for estimating element frequencies (11NE) such as counting
decisions as limb shaft fragments (Bunn 1982; Bunn and Krolll986; Marean and Spencer
below). This 1991). The repeated blows to limb bones and extensive pot-sizing of axial
:;s processing bones from larger animals compared to the minimal processing of smaller
Furthermore, carcasses further suggests that archaeological visibility of marrow processing
t marrow vs. and boiling may be biased against small animals. Nevertheless, marrow pro-
ith ability to cessing of small ungulate limb bones may be indicated by (1) a pattern of
applying eco- limited midshaft impact breakage and (2) the presence of bone tubes with
ue obviously light gnaw damage near articular ends.
tastes. The intensity of limb marrow I cancellous tissue processing may be indicat-
:oraging seri- ed by th~ degree of diaphyseal fragmentation, complete articular end frequen-
~d. Economic cies, and damage to the medullary wall and cancellous bone. Zooarchae-
observations ological assemblages with a large limb bones with few diaphyseal impact
n much of it. points, a high frequency of intact articular pieces, and a lack of crushing
it predictions damage to cancellous tissue may indicate minimal processing such as only
f the optimal marrow consumption or bone boiling. In contrast, assemblages characterized
, however, in by numerous diaphyseal fragments with impact marks, fragmented articular
l (in this case, pieces, and pitted/ crushed/ sliced cancellous bone would indicate a process-
cultural vari- ing intensity closer to that of the Hadz;a. Similarly, assemblages with highly
ighted. Thus, fragmented large ·animal axial parts, but containing relatively intact small
:ural attitudes animal axial parts, likely indicate bone boiling.
1ddition to a i The diagnostic damages created by breakage of roasted limb bones suggests
·ocessing and the potential to directly identify this behavior from sites where the earliest use
~able cultural of fire has been reported, for example, Chesowanja (Gowlett et al. 1981),
ls, processing Swartkrans (Brain and Sillen 1988), and FxJj20 at Koobi Fora (Bellomo 1992).
experimental Analysis of bone breakage from those sites could determine whether or not
ate the utility 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 behav-
ior 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
rcass process- calcined bone may not be very good indicators of either the use of fire or
ltion of bone cooking in the archaeological record. Others have offered similar cautions
: related to (1) about using burned bone as an indicator of cultural activity (e.g., Lyman
luty tools, {2) 1987).
tcellous bone, Differentiation between heavy- and light-duty tool use in carcass processing
fron1 broken is also of concern. Binford (1981) and others have commented on the peculiar
220 I J. S. Oliver
nature of chopping/hacking butcllery with heavy-duty tools, suggesting that marrow-rich

their presence may alter transport strategies. Others (e.g., Bunn et al. 1980; marrow and
Jones 1980; Toth 1985), however, hav~ argued that there may have been func- hammers ton
tionally equivalent stone tools in the past. Detailed examination of zooar- marks, and J
chaeological assemblages for heavy- versus light-duty-tool:-induced damage Plio-PleistocE
patterns should help (1) refine our understanding of carcass processing, (2) Chewing 2
determine whether or not heavy-duty metal tools significantly affected carcass technique av
segmentation or transport strategies, and (3) compliment Toth's tool use, inid tooth .m
curation, and transport data for Plio-Pleistocene hqminids. That is, definition present) to b
of the tool used to create the cut marks at sites that lack stone tools (e.g., GaJiS; blages create
Bunn 1981, 1982) may help define early hominid tool transport strategies in to be a. comr
areas distant from lithic resources (e.g., the Karari escarpment). (Figure 12-1E
Finally, bonebreakage and consumption observations reported here indi- er carcasses r
cate that the Hadza do not frequently create archaeologically visible kill sites. As fire be:
Thus, for technologically similar human groups, few kill sites would be found axial parts '
in the archaeological record. Sites are created where hominids repeatedly removal of r
gather to process and consume food. When group size permits adequate shar- be evidencec
ing, size I and II carcasses are typically transported complete to residence hearth-sizin~
camps where butchery is undertaken; nothing remains at the kill site. Though promote adc
processing activities are significantly greater at kills of size III and IV animals, the advent o
most parts are transported to camp; only a few roasted ribs and possibly a axial parts, ~
smashed skull and mandibles near an ash scatter remain as evidence of the cessed as is
butchery site. Zebra kills at which informants report roasting of skulls and availability o
numerous ribs lacked all but two or three ribs a few weeks after the kill. Data from
Archaeological visibility of kill sites reported by 0' Connell and colleagues meet the pn
(1988a, 1988b, 1990, 1992) would be greater, particularly for large animals. 1986, 1988) ]
Pleistocene s
Innovations in Nutrient Extraction Technology and the Marean, this
Evolution of Carcass Processing fracture patb
of the identi:
The organization of Hadza carcass processing behavior thus (1) 1991), suggE
provides a baseline with which to compare zooarchaeological assemblages available tee
and (2) suggests new ways to approach prehistoric carcass processing behav- increase in n
ior. Specifically, the role late-stage carcass processing options (i.e., nutrient gies may ha~
extraction techniques like bone smashing, roasting, and boiling) play in time, but the
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 s
cutting meat and smashing bone, fire for roasting and bone-boiling facilities or
B
implements), the relationship between processing and transport costs and
has been do<
nutrient value changed.
and animal~
All things being equal, a number of expectations about the evolution of
guide carcas~
carcass processing may be derived from the Hadza data. Nutrient extraction
to be organi
techniques available to Plio-Pleistocene hominids were limited to bone smash-
returns whilE
ing and gnawing (but see Bellomo 1992). Thus, the difficulty in efficiently
size class I a1
extracting nutrients from defleshed vertebrae and ribs without cooking
transport dis
suggests that, without fire, hominids should have preferentially transported
data suggest
.uggesting that marrow-rich limb bones (see also Bunn et al. 1988). Further, I expect that
nn et al. 1980; marrow and cancellous tissue extraction efforts were maximized via intensive
ave been func- hammerstone breakage. Thus, extreme fragmentation, numerous impact
Jion of zooar- marks, and few complete articular ends may characterize limb bones from
iuced damage Plio-Pleistocene zooarchaeological assemblages.
processing, (2) . Chewing and gnawing of bone remains the only other nutrient extraction
Lffected carcass technique available to Plio-Pleistocene hominids. Thus, I would expect hom-
:)th' s tool use, inid tooth marks (which, to my knowledge, cannot be adequately defined at
tt is, definition present) to be more frequent in Plio-Pleistocene assemblages than in assem-
ols (e.g., GaJiS; blages created after the invention of fire. Furthermore, because chewing seems
rt strategies in 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 small-
rted here indi- er carcasses more frequently than those of larger carcasses.
isible kill sites. As fire began to be utilized for roasting, I expect additional transport of
rould be found axial parts and less nutrient-rich limb bones because roasting facilitates
ids repeatedly removal of remaining meat scraps. More efficient nutrient extraction should
:1dequate shar- be evidenced by increased modification of axial bones resulting from both
:e to residence hearth-sizing and consumption practices. Development of roasting pits might
lll site. Though promote additional body part transport and processing of axial parts. With
nd IV animals, the adv7ht of bone-boiling technology, all elements, and especially, perhaps,
:md possibly a axial parts, should show a tendency to be transported and intensively pro-
vidence of the cessed as is the case with the Hadza. That pattern would be mitigated by
~ of skulls and availability of firewood and other processing goals such as storage and trade.
: after the kill. Data from Olduvai Gorge, Tanzania, and Ko.obi Fora, Kenya, appear to
md colleagues meet the predictions for Plio-Pleistocene sites. Bunn (1986; Bunn and Kroll
;e animals. 1986, 1988) has demonstrated preferential transport of limb bones to Plio-
Pleistocene sites like FLK Zinjanthropus and FxJjSO (but see Blumenschine and
~he 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
tavior thus (1) 1991), suggesting that nutrient extraction was being maximized with the
tl assemblages available technology. Exactly how the introduction of fire and the resultant
cessing behav- increase in nutrient extraction capabilities and subsequent boiling technolo-
; (i.e., nutrient gies may have altered transport and processing strategies is unknown at this
>iling) play in time, but the Hadza example suggests that it may have had a significant effect.
rganization of
on techniques.
tone tools for Summary
ing facilities or
Bone breakage by the Hadza from butchery through consumption
Jort costs and
has been documented according to the site of breakage, breakage technique,
and animal size class. Carcass size and later cooking options were.shown to
e evolution of
guide carcass processing decisions. Further, Hadza carcass processing appears
ient extraction
to be organized around an optimization principle of maximizing nutrient
~o bone srrlash-
returns while minimizing transport and field processing costs. Variciliility (i.e.,
' in efficiently
size class I and II animals and wildebeest) in transport decisions not related to
:hout cooking
transport distances may be related to considerations of sharing and taste. The
1y transported
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 carcasses and dz
from more detailed examination of breakage damages. In general, the data Hadza visits to ·
suggest that because animal size play~ such a major role in Hadza decisions taring" or const:J
5. The Hadza
regarding bone breakage a large part of archaeological variability in breakage
ed as people frOJ
patterns may also be due to differences in body size. Finally, the importance 6. In my limit
of cooking in structuring initial butchery and transport decisions suggests that that of some oth
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 Re:
to be incorporated into models of carcass transport and processing strategies
a.nd explanations of bone frequency and modification patterns in zooarchae- Bartram, L. E., J1
()logical assemblages. 1993 An Et
Refuse. U
University
Acknowledgments Behrensmeyer, 1
1989 Nonh1
I thank J. Hudson and the Center for ~rchaeological Investiga- Modificatior
tions, Southern Illinois University at Carbondale, for organizing a truly enjoy- Study ofth
able; intense conference. I thank H. T. Bunn for inviting me to participate in Bellomo, R. V.
1992 Early
his Hadza ethnoarchaeology project and for granting permission to publish
Koobi Fon
some data before the overview of our 1986 and 1988 work is published. I Society of j
thank the Tanzania Commission for Science and Technology for permission to Binford, L. R.
conduct this research. I thank Dr. R. B. McMillan, Director of the Illinois State 1978 Nunan
Museum, for administrative support and research space at the Research and 1981 Bones:
Collections Center. Finally, I thank the Hadza without whose warmth and 1983 In Pur:
patience none of it would have been possible. Expedition support was pro- 1987 Resem
vided by grants to H. T. Bunn from University of Wisconsin-Madison; data and Theory
analysis was supported by dissertation improvement grants from the Wenner- demic Pres:
Gren Foundation for Anthropological Research and the National,'Science Blurnenschine, F
Foundation. This paper has benefited from discussions with H. T. Bujm, B. W. 1986 Carca:
Styles, M.D. Wiant, D. Gifford-Gonzalez, and J. Hudson and from comments Scavengi!lg
by R. L. Lyman and the anonymous reviewers. Any errors, omissions, or un- 1988 An Ex
on Archaec
founded conclusions that remain are my own. This is contribution No. 89 of
502.
the Illinois State Museum Archaeological and Quaternary Studies Program. 1989 ALan
Opportunit
Notes Blurnenschine, F
1988 Percm
1. The Hadza moved into agricultural settlements in 1989 but have since returned haviour. Nt
to the ''bush" and have resumed hunting and gathering. Bonnichsen, R.
2. Support for the 1988 fieldwork was provided by the Graduate School of the 1973 Millie'
University of Wisconsin-Madison. 291.
3. Other project members (H. T. Bunn, C. Gurney, and C. O'Brien) observed pro- 1979 Pleistc
curement of additional carcasses in 1988, which is the subject of a future report of Survej; of Ct
carcass transport and bone breakage ,by the Hadza. A preliminary report of the · Brain, C. K., and
transport patterns for 110 carcasses observed by project members in 1986 and 1988 1988 Evide1
was presented by Bunn and Bartram (1990). 336:464-46(
4. In 1988 groups of Isanzu men came into the Hadza area to hunt, setting up at Bunn,H. T.
least one carcass processing station. They hunted for several weeks and processed 1981 Archa
tay be derived carcasses and dried meat at a processing station near a hunting blind. The purpose of
Lerat the data Hadza visits to Tatoga camps was to obtain maize or animal food by working e'doc-
.dza decisions toring'' or constructing huts and animal enclosures) or begging.
ty in breakage 5. The Hadza surmised that we were leaving in a few days, and camp size increas-
ed as people from other nearby camps began to arrive.
1e importance
6. In my limited experience, I also found zebra meat sweeter and less tough than
; suggests that that of some other animals including warthog, buffalo, and impala .
. fire, roasting
Lthe evolution
; options need References
;ing str'ategies
in zooarchae- 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
cal Investiga- Modification, edited by R. Bonnichsen and M. Sorg, pp. 99-120. Center for the
a truly enjoy- Study of the First Americans, Orono, Maine.
participate in Bellomo, R. V.
on to publish 1992 ~arly Pleistocene Evidence of Hominid-Controlled Fire at FxJj20 Main,
Koo~i Fora, Kenya. Paper presented at the 11th Biennial Conference of the
; published. I
Society of Africanist Archaeologists, University of California at Los Angeles.
permission to Binford, L. R.
e Illinois State 1978 Nunamuit Ethnoarchaeology. Academic Press, New York.
Research and 1981 Bones: Ancient Men and Modern Myths. Academic Press, New York.
~warmth and 1983 In Pursuit of the Past. Thames & Hudson, London.
port was pro- 1987 Researching Ambiguity:· Frames of Reference and Site Structure. In Method
vfadison; data and Theory for Activity Area Research, edited by Susan Kent, pp. 449-:-512. Aca-
:1 the Wenner- demic Press, New York.
ional Science Blumenschine, R. J.
r. Bunn, B. W. 1986 Carcass Consumption Sequences and the Archaeological Distinction of
)m comments Scavenging and Hunting. Journal of Human Evolution 15:639-659.
.ssions, or un- 1988 An Experimental Model of the Timing of Hominid and Carnivore Influence
tion No. 89 of on Archaeological Bon_e Assemblages. Journal of Archaeological Science 15:483-
502.
s Program.
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 Be-
since. returned haviour. Nature 333:763-765.
Bonnichsen, R.
~ School of the 1973 Millie's Camp: An Experiment in Archaeology~ World Archaeology 4(3):277-
291. .
observed pro- 1979 Pleistocene Bone Technology in the Berengian Refugium. Archaeological
.Iture report of Survey of Canada Paper No. 89. National Museum of Canada, Ottawa.
· report of the Brain, C. K., and A. Sillen
1986 and 1988 1988 Evidence from the Swartkrans Cave for the Earliest Use of Fire. Nature
336:464-466.
t, setting up at Bunn,H. T.
and processed 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. Haynes, G.
1982 Meat..,eating and Human Evolution: Studies on the Diet and Subsistence Patterns 1983 Freq
of Plio-Pleistocene Hominids in East Africa. Unpublished Ph.D. dissertation, Modem~
Department of Anthropology, University of California, Berkeley. Hill, A.
1983 Evidence on the Diet and Subsistence Patterns of Plio-Pleistocene Homi- 1989 Bon
nids and Koobi Fora, Kenya and Olduvai Gorge, Tanzania. In Animals and edite9- b)
Archaeology: Hunters and Their Prey, edited by J. Clutton-Brock and C. Grigson, FirstAmt
pp. 21-30. BAR International Series 163. Oxford. Johnso:q~ E.
1986 · Patterns of Skeletal Representation and Hominid Subsistence Activities at 1985 Cur:
Olduvai Gorge, Tanzania and Koobi Fora, Kenya. Journal of Human Evolution Method a;
15:673-690. ·NewYor
1989 Diagnosing Plio-Pleistocene Hominid Activity with Bone Fracture Evi- Jones, P.
dence. In Bone Modification, edited by R. Bonnichsen and M. Sorg, pp. 299-316. 1980 Exp
Center for the Study of the First Americans, Orono, Maine. Paleo lith'
Bunn, H. T., and L. Bartram Lyman,R. L.
1990 Carcass Processing and Bone Ass·emblage Formation by the Hadza Hunter- 1987 ArcJ
Gatherers in Tanzania. Paper presented at the Sixth International Conference of Advances
the International Council for Archaeozoology, Wa~hington, D.C. 249-337.
Bunn, H. T., L. E. Bartram, and E. M. Kroll Marean, C. W
1988 Variability in Bone Assemblage Formation from Hadza Hunting, Scaveng- 1991 Imp
ing and Carcass Processing. Journal of Anthropological Archaeology 7:412-457. Abundar
Bunn, H. T., J. W. K. Harris, G. Isaac, Z. Kaufulu, E. Kroll, K. Schick, N. Toth, and A. Marshall, F., a
K. Behrensmeyer 1991 Mea
1980 FxJj50: An Early Pleistocene Site in Northern Kenya. World Archaeology 12: Part Rep
109-136. 18:149-H
Bunn, H. T., and E. M. Kroll Metcalfe, D.
1986 Systematic Butchery by Plio/Pleistocene Hominids at Olduvai Gorge, 1989 A G
Tanzania. Current Anthropology 29:431-452. Field Pro
1988 Fact and Fiction about the FLK Zinjanthropus Floor. Current Anthropology Americm
29:135-149. Metcalfe, D., a
Fiorillo, A. R. . 1992 A :M
1989 An Experimental Study of Trampling: Implications for the Foss~{ Record. In Transpor
Bone Modification, edited by R. Bonnichsen and M. Sorg, pp. 61-71. Center for the Morlan, R. E.
Study of the First Americans, Orono, Maine. 1980 Tap:
Gifford, D. P. Yukon T
1980 Taphonomy and Paleoecology: A Critical Review of Archaeology's Sister Survey of
Disciplines. In Advances in Archaeological Method and Theory, val. 4, edited by M. O'Connell, J. I
B. Schiffer, pp. 365-438. Academic Press, New York. 1988a Had
Gifford-Gonzalez, D. Implicati
1989 Ethnographic Analogues for Interpreting 1v1odified Bones: Some Cases 1988b Had
from East Africa. In Bone Modification, edited by R. Bonnichsen and M. Sorg, pp. Current I
179-246. Center for the Study of the First Americans, Orono, Maine. 1989 Dist
1991 Bones Are Not Enough: Analogues, Knowledge, and Interpretative Strat- Camps: I
egies in Zooarchaeology. Journal of Anthropological Archaeology 10:255-254. pretation
Gould, R. 61-76. Pl1
1967 Notes on the Hunting, Butchering, and Sharing Game among the 1990 Rea:
Ngatatjara and Their Neighbors ,in the West Australian Desert. Kroeber Anthropo- Journal of
logical Society Papers 36:41-66. 1992 Pan
Gowlett, J. A.J., J. W. K. Harris, D. Walton, and B. A. Wood HadzaK
1981 Early Archaeological Sites, Hominid Remains and Traces of Fire from O'Connell, J. I
Chesowanja, Kenya. Nature 294:125-129. 1989 Ana
Haynes, G.
;istence Patterns 1983 Frequencies of Spiral and Green-Bone Fractures on Ungulate Limb Bones in
). dissertation, Modern Surface Assemblages. American Antiquity 48:102-114.
Hill, A.
stocene Homi- 1989 Bone Modification by Modern Spotted Hyaenas. In Bone Modification,
rn Animals and edite9- by R. Bonnichsen and M. Sorg, pp, 169-178. Center for the Study of the
nd C. Grigson, First Americans, Orono, Maine.
Johnson, E.
tee Activities at 1985 Current Developments in Bone Technology. In Advances in Archaeological
uman Evolution ·Method and Theory, vol. 5, edited by M. Schiffer, pp. 157-235. Academic Press,
New York
~ Fracture Evi- Jones, P.
·g, pp. 299-316. 1980 Experimental Butchery with Modern Stone Tools and Its Relevance for
Paleolithic Archaeology. World Archaeology 12:153-165.
Lyman, R. L.
Hadza Hunter- 1987 Archaeofaunas and Butchery Studies: A Taphonomic Perspective. In
1 Conference of 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
nting, Scaveng- 1991 Impact of Carnivore Ravaging on Zooarchaeological Measures of Element
7:412-457. Abundance. American Antiquity 56:645-658.
'J. Toth, and A. Marshall, f., and T. Pilgram
1991 Meat Versus Within-Bone Nutrients: Another Look at the Meaning of Body
Archaeology 12: Part Representation in Archaeological Sites. Journal of Archaeological Science
18:149-163.
Metcalfe, D.
)lduvai Gorge, 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
nt Anthropology American Archaeology, Atlanta.
Metcalfe, D., and K. R. Barlow
1992 A Model of Exploring the Optimal Trade-off Betwee:n Field Processing and
ossil Record. In Transport. American Anthropologist 94:340-356.
.. Center for the 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
teology's Sister Survey of Canada Paper No. 94. National Museum of Canada, Ottawa.
4, edited by M. 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.
-s: Some Cases 1988b Hadza Scavenging: Implications for Plio/Pleistocene Hominid Subsistence.
nd M. Sorg, pp. Current Anthropology 29(2):356-363.
:1e. 1989 Distribution of Refuse Producing Activities at Hadza Residence Base
pretative Strat- Camps: Implications fo_r Analyses of Archaeological Site Structure. In The Inter-
255-254. pretation of Archaeological Spatial Patterns, edited by E. Kroll and T. D. Price, pp.
61-76. Plenum Press, New York
:1e among the 1990 Reanalysis of Large Mammal Body Part Transport among the Hadza.
roeber Anthropo- 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.
~s of Fire from 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 .. edited b)
Oliver, J. S. 1970 Hur.
1984 Analogues and Site Context: Bone Damages from Shield Trap Cave Museum
(24CB91), Carbon County, Montana, U.S.A. Paper presented at the First 1972 Eco:
futernational Conference on Bone Modification, Carson City, Nev. among :r.
1986 The Taphonomy and Paleoecology of Shield Trap Cave (24CB91), Carbon County, Man, Set
Montana. Unpublished Master's thesis, Institute for Quaternary Studies, Dimbleb'
University of Maine at Orono. Yellen, J. E. ·
1989 Analogues and Site Context: Bone Damages from Shield Trap Cave 1977 Arc!
(24CB91), Carbon County, Montana, U.S.A. In Bane Modification, edited by R. Academi
Bonnichsen and M. Sorg, pp. 73-98. Center for the Study of the First Americans, Zierhut, ·N. W
Orono, Maine. 1~67' Bon
1991 Damage Context Analysis: A Preliminary Analysis of the Amboseli Hyaena 1:33-36.
Den, FLK NN Level 2 and FLK Zinjanthropus Faunal Asse11.1blages. Paper pre-
sented 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 Mega-
fauna 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:
Toth, N.
1985 The Oldowan Reassessed: A Close Look at Early Stone Artifacts. Journal of
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 by R. B. Lee and I. DeVore, pp. 103-110. Aldine, Chicago.

1970 Hunters and Gatherers: Material Culture of the Nomadic Hadza. British
eld Trap Cave Museum, London.
~d at the First 1972 Ecology, Nomadic Movement and the Composition of the Local Group
{. a111ong Hunters and Gatherers: An East African Example and Its Implications. In
, Carbon County, Man, Settlement, and Urbanism, edited. by P. Ucko, R. Tringham, and C.
rnary Studies, Dimbleby, pp. 193-206. Duckworth, London.
Yellen, J. E.
eld Trap Cave 1977 Archaeological Approaches to the Present: Models for Reconstructing the Past.
n, edited by R. ·Academic Press, New York.
:<irst Americans, Zierhut, N. W.
1967 Bone Breaking Activities of the Calling Lake Cree. Alberta Anthropologist
mboseli Hyaena 1:33-36. .
1ges. Paper pre-
n Archaeology,
19(11):97-106.
3.nzania. Nature
:; Past. In Mega-
Agenbroad, J. I.
: Springs, South
:1ic Problems. In
I by J. B. Griffin
~my of Sciences
,r, Morphology
.1:307-325.
3ing Behaviors:
ology 2:67-98 .
.facts. Journal of
~port of Tubers
l32-148.
Jublished Ph.D.
:ed by R. B. Lee
Man the Hunter,

13. Food Sharing and the Faunal
Record
Fiona Marshall
Abstract: Food sharing has long been an important concept in hominid

evolution. Recent ethnoarchaeological research among Okiek hunter-
gatherers of the western Mau Escarpment, Kenya, demonstrates that
transport of bone through food sharing is the major process affecting the
formation of body part profiles in Okiek sites. Variation in body part
profiles accumulated through sharing results from differences in hunting
success and social relationships between households. Sharing of large
animal meat is nearly universal among contemporary hunter-gatherers
and, as among the Okiek, affects faunal assemblages accumulated by
households. As a result, sharing should be considered, as is the logistics
of carcass transport, in the interpretation of faunas from archaeological 1982b; Winterhz
sites. to environment
and to the sociz
1980). Hunter-g
Introduction environments t:
i diate return sul
Sharing of food between households is common in many small- delayed return
scale societies but is best known among hunter-gatherers. Most contemporary and storage.
hunter-gatherers share food to some degree, especially meat. However, there Washburn ar
is a great deal of variation in the degree to which hunter-gatherers share. focus on the po
Some, such as the !Kung (Lee 1979; Marshall 1976; Yellen 1977a, 1977b), .opment of hun
Badza (Woodburn 1968, 1980), Aka (Hudson 1990) and Efe (Bailey 1988) gatherers and n
Pygmies, Western Desert Aborigines (Gould 1967, 1981) and Ache of Para- Isaac (1977a, 15
guay (Hill and Hurtaldo 1989; Kaplan and Hill 1985), share much more hunter-gathere
widely than others, such as the Nunarniut (Binford 1978, 1984) and hunter- humans from a
gatherers of the Northwest Coast of North America (Gould 1981; Hayden base camps by
1990; Table 13-1). tion for large-br
Sharing is seen by many anthropologists as a risk-spreading mechanism teristic of Homo
(Bahuchet 1990; Cashdan 1985; Gould 1981; Kaplan and Hill 1985; Wiessner Isaac based 1
hominids large]
From Bones to Behavior: Ethnoarchaeologica{ and Experimental Contributions to the Interpretation of from diverse rr
Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional argued that the
Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights by extension, p
reserved. ISBN 0-88104-076-2. sive discussion
228
Food Sharing I 229
Table 13-1. Variation in the Degree to Which Meat Is

Shared among Contemporary Hunter-Gatherers
Contemporary Hunter-Gatherer Groups
1 Western Desert Aborigines (Gould 1967, 1981)

!Kung (Lee 1979; Marshall1976; Yellen 1977ab)
Hadza (Barnard and Woodburn 1988; Woodburn 1968)
Ache (Kaplan and Hill 1985)
Aka (Hudson 1990)
Efe(Bailey 1988)
Okiek (Blackburn 1971; Marshalt personal observation 1990)
Iglulik Eskimo (Damas 1975)

ept in hominid
Okiek hunter-
Copper Eskimo (Damas 1975)
nonstrates that Netsilik Eskimo (Damas 1975)
·ss affecting the Nunamiut (Binford 1978, 1984)
n in body part
North West Coast groups (Gould 1981; Hayden 1990)
nces in hunting Great Basin Shoshone (Steward 1938)
haring of large
1nter-gatherers
ccumulated by
' is the logistics
archaeological 1982b; Winterhalter 1986). The degree to which groups share has been related
to environmental variables (Binford 1980, 1984; Gould 1981; Hayden 1990)
and to the social relations of production (Wiessner 1982a, 1982b; Woodburn
1980). Hunter-gatherers who share widely generally live in less productive
environments than those who do not. And hunter-gatherers with an imme-
diate return subsistence organization share more widely than those with a
1 many small- delayed return subsistence organization, who rely on specialized production
contemporary and storage.
[owever, there . Washburn and DeVore (1961) and Isaac (1977a, 1977b) were the first to
.therers share. focus on the potential im·portance of sharing of food for studying the devel-
1977a, 1977b), opment of human social systems. Drawing on detailed studies of hunter-
(Bailey 1988) gatherers and nonhuman primates that took place during .the 1960s and 1970s,
Ache of Para- Isaac (1977a, 1977b) noted that the food-sharing behavior of contemporary
·e much more hunter-gatherers is one of the major features that differentiates modern
!) and hunter- humans from apes. He linked the sharing of food to meat eating and use of
1981; Hayden base camps by early hominids and suggested that this behavior led to selec-
tion for large-brained ho:qtinids and the development of social systems charac-:
1g mechanism teristic of Homo sapiens.
l985; Wiessner Isaac based his argument for the existence of food sharing among early
hominids largely on the spatial concentration of stone and fragmentary bones
he Interpretation of from diverse mammalian taxa at sites like KBS and FLK Zinjanthropus. He
1tions, Occasional argued that they supported the notion of meat eating, use of home bases, and,
~rsity. All rights by extension, pooling of resources and sharing of meat. There has been exten-
sive discussion as to what constitutes archaeological evidence for meat eating,
230 IF. Marshall
hunting, scavenging, and use of base camps (Binford 1981, 1986, 1988; maintained a!
Blumenschine 1986; Bunn 1981, 1982; Bunn and Kroll 1986, 1988; Potts 1984, storage, and h
1988; Potts and Shipman 1981; Ship]Jlan 1983, 1986a, 1986b), but there has tude gradient 1
been very little discussion as to what might constitute archaeological evidence ings are used t
for food sharing (but see Binford 1984). Until the last fe:w years, neither yeat. N everthe
archaeologists working in the Plio-Pleistocene nor those working in later time honey is relied
periods have studied the effect of food sharing on the archaeological record or The animals
pursued Isaac's suggestion of the importance of food sharing to understand- lands, and the)
ing the development of human social systems. _ and meat is no
As a result of a recent study of subsistence and the formation of the faunal and arrows by
record by Okiek hunter-gatherers of Kenya, I became unexpectedly aware of individuals of
the importance of food sharing to the formation of Okiek faunal assemblages. diet. Okiek we
The similar experience of Jean Hudson (1990) among the Aka. Pygmies, recent food is acquire
thinking by John Speth (1990) on the nutritional importance of food sharing, very importanj
and study by Jim Enloe and Francine David (1989) on food sharing at the entitled to sorr
Upper Paleolithic site of Pincevent prompted this discussion of the place of portion of theiJ
food, and specifically meat sharing; in the interpretation of the faunal record. food secretly v.;
The
Food Sharing and Okiek Faunas
My
Our project among the Okiek was set up to study subsistence, and among a previ
specifically hunting and the role of bone transport in patterning in body part ment, the Piik
representation in Okiek sites. However, we·found food sharing to be the not all, still USE
single major factor patterning the early stages of faunal accumulation and the ents. Wild anin
body part representation on Okiek sites. their place. Me
domestic anim
Brief Background to the Okiek: General and Historic Patterns still hunt in a
I techniques as t:
The Okiek are Kalenjin-speaking hunter-gatherers of the high alti- the time of re:
tude rain forests of the Mau Escarpment, Kenya. They are delayed return hunting), the Sc
hunter-gatherers and differ from better-known hunter-gatherer groups in People wed
many aspects of their social organization and subsistence systems. In this However, in a
regard they. provide a useful antidote to academic stereotypes of hunter- secret, and we
gatherers based on groups such as the !Kung and Australian aborigines. observed a hig
Okiek subsistence systems and social structure seem likely to have developed hunted, transi
at least in part in response to resource structure and the predictability of sample fits wit
resources on the Mau Escarpment. There is reason to believe that such de- with informa ti<
layed return hunter-gatherer groups may have been more common in the past 1960 and 1990 1
than they are now (Barnard and Woodburn 1988; Woodburn 1980). believe that th
Okiek social organization is lineage based (Blackburn 1971, 1982; generalizing ac
Huntingford 1942, 1954, 1955; Kratz 1977, 1981), and settlement patterns vary
somewhat with the age and social relationships of individuals but are largely Met
based on single nuclear family units. Houses are located on lineage territories
that encompass strips of land running from low to high altitudes on the Mau Obs
Escarpment. A family often uses a base at the lower altitude edge of the forest ulation, transp
and several forest houses at different altitudes. Until recently the Okiek have subsequent rec
Food Sharing I 231
1 1986, 1988; maintained a specialized subsistence system focused on beekeeping, honey

s8; Potts 1984, storage, and hunting, making no systematic use of plants for food. The alti-
but there has tude gradient of the Mau and the somewhat bimodal nature of plant flower-
gical evidence ings are used to maintain an almost continuous flow of honey throughout the
years, neither year. Nevertheless, there are at least two lean periods in the year when stored
g in later time honey is relied upon. ·
p.calrecord or ·The animals of the forest are abundant in all areas except the Mau grass-
o understand- lands, and they do not migrate. They appear to provide a stable food resource,
and meat is not stored. Animals are both hunted with dogs, spears, and bow
1 of the faunal and arrows by parties of young men from several households and trapped by
:edly aware of individuals of all ages. Together honey and meat provide the Okiek staple
1 assemblages. diet. Okiek women do not play an important role in food getting. Nearly all
ygmies, recent food is acquired by men and brought home to the family. Food sharing is thus
:food sharing, very important among the Okiek, and anyone who sees a person with food is
sharing at the entitled to some of it. However, most people make an effort to limit the pro-
of the place of portion of their food that they share outside the immediate family by caching
aunal record. food secretly whenever possible, and honey is not widely shared.
The Piik ap Oom

My work was conducted with Tom Pilgram and Cory Kratz
1bsistence, and among a previously unstudied Okiek group from the western Mau Escarp-
tg in body part ment, the Piik ap Oom (people of bamboo). We found that some people, but
.ring to be the not all, still use the forest systematically in a way similar to that of their par-
ulation and the ents. Wild animals are no longer relied upon as a staple food; maize has taken
their place. Most families have gardens for greens or maize, and some keep
domestic animal herds. Cattle are kept for milk, and goats for meat. But men
Patterns still hunt in a regular way, hunting the same animals and using the same
techniques as their fathers. For those reasons, as well as the political climate at
)f the high alti- the time of research (elephant conservation drive and discouragement of
dela yed return hunting), the sample of animals studied here is small (N = 17).
erer groups in People we did not know well were reluctant to show us animal carcasses.
ystems. In this However, in a small community it is impossible to keep a successful hunt
ves of hunter- secret, and we obtained at least partial data on most kills. In general, we
ian aborigines. observed a high degree of internal consistency in the way that animals were
1ave developed hunted, transported, butchered, and distributed. This small quantitative
redictability of sample fits with general observations and interview data from the area and
e that such de- with information for Kipchornwonek and Kaplelach Okiek collected between
mon in the past 1960 and 1990 (Blackburn 1971; Kratz, personal communication 1990). Thus, I
180). believe that the information discussed here provides a reasonable basis for
rn 1971, 1982; generalizing about the Okiek.
1t patterns vary
; but are largely Methods
leage territories
des on the Mau Observational and sometimes oral data were collected on disartic-
jge of the forest ulation, transport to houses, and distribution of carcass parts. Most data on
the Okiek have subsequent redistribution of carcass parts were oral. These data do not pro-
2321 F. Marshall
vide a comprehensive record of the extent of redistribution of meat that bone ranging :
occurred but illustrate general patterns. typical, especi
Hunted animals in our sample. include giant forest hog (Hylochoerus ported differE
meinertzhageni) weighing 100-250 kg, bushbuck (Tragelaphus scriptus) weigh- handles are a1
ing 25-80 kg, and blue duiker (Cephalophus monticola) weigh.ing 4-10 kg. Giant meat is wrap]
forest hog are the preferred prey of the Okiek; they provide a lot of meat, and often hung fn
the meat is very fat. Blackburn (1971) reports that they are a staple among the carried by ha
Kaplelach and Kipchornwonek Okiek. Giant forest hog are hunted by groups carry it home
of young men fro1n several families, using a pack of dogs (approximately 5-15 6-8 km, but rn
animals) to bring the hog to bay. It is then killed with a spear. Giant forest hog Despite the
are dangerous, and dogs are frequently injured. Bushbuck are the most carcasses are t
common kill in our area. Bushbuck meat is not as fat as giant forest hog but is an animal are
especially liked for its delicate flavor. Bushbuck are hunted both with and animals were
without dogs, generally by several young men. The animal is usually killed meters (Table
with a bow and arrow; occasionally a sharpened bamboo stake or club may be of meat from
used. By contrast, duiker are usually trapped or snared. The blue duiker in results in furt
our sample were recovered in that-way. the landscape
Butchery and sharing data were also collected for cattle, sheep, and goats, dated bones.
since patterns of butchery and rules for sharing of wild and domestic taxa do Sharing foil
not differ. As among other hunter-gatherer groups, wild animal carcasses are upon the natu
"owned" by the successful hunter (Barnard and Woodburn 1988:17) in the share, includii
same way that domestic stock is owned by individuals. In either case, sharing hunter who 1<
of meat from a carcass is obligatory. In addition, ownership of resources, animal. Large
including individual ownership of trees and hives, is part of Okiek culture. carriers, and
Thus, social relationships and patterns of sharing are not upset by individual depends u po:
ownership of domestic animal resources. made. PeoplE
relatives and .
Data about the kill
Animals were disarticulated at the kill site. Butchery f61lowed a during primar
consistent pattern that .resulted in reducing the carcass to lighter and more Food sharir
easily transported portions, by emptying fluids from the stomach and success and sc
intestines, discarding some internal organs, and disarticulating major joints. affected by th1
Pieces of the internal organs, especially the intestines, were fed to the dogs at context of con
the kill sites, and occasionally a piece of liver was eaten by the hunters. But Ht:
meat was not removed from bones at the kill sites. In only one case, when
carriers were few, were any bones abandoned at the site of the kill. In that Bo·
case, the defleshed cranium and one set of carpals from a giant forest hog much variatic
were discarded at the kill site. · both to skill ru
In all other cases, the initial separation of meat from bone took place at a quantitative s
house, either a permanent house lived in by the whole family or, in some quality of thei
cases, a high altitude forest house belonging to the hunter or to his friends or successful as
family. Transport is not an important factor affecting characteristics of Okiek different kind
faunal assemblages for taxa studied. However, the largest animal in our meat sharing '
sample is giant forest hog, weighing 100-250 kg. Transport could play a role So(
for larger taxa such as buffalo. They no longer exist in the study area, so we
have no data on their transport. So'
In transporting entire carcasses home, individuals carry loads of meat and This is partict
I
Food Sharing 233
n of meat that bone ranging from as much as 31 kg to 2-5 kg. But the smaller loads are more
typical, especially of secondary sharing. Portions of giant forest hog are trans-
1g (Hylochoerus ported differently than portions of the smaller bushbuck. Generally vine
criptus) weigh- handles are attached to giant forest hog limbs to fashion backpacks, and the
; 4-10 kg. Giant meat is wrapped around the carrier's body. By contrast bushbuck meat is
ot of meat, and. often hung from a stick carried over the shoulder or wrapped in leaves and
1ple among the carried by hand. Women may sometimes bundle meat up in the skin and
nted by groups carry it home balanced on their heads. In some cases, meat is carried as far as
·oximately 5-15 6-8 km, but more often 0.5-3 km.
;iant forest hog Despite the fact that all parts of duiker, bushbuck, and giant forest hog
~ are the most carcasses are usually carried home by the Okiek, it is rare that all the bones of
)rest hog but is an animal are discarded at a single settlement. In 12 out of 15 observed cases,
both with and animals were shared between 2 and 7 households separated by several kilo-
; usually killed meters (Table 13-2). It is also common for secondary and tertiary distribution
or club may be of meat from the households to follow primary distribution. This usually
blue duiker in results in further scattering of skeletal parts of animals between sites across
the landscape, depending on whether meat is shared with or without asso-
eep, and goats, ciated bones.
)mestic taxa do Sharing follows set rules, but each distribution will be different depending
al carcasses are upon the/nature and composition of the group .of people present. The largest
1988:17) in the share, in~luding the hindlimbs and often the skin, ribs, and pelvis, goes to the
er case, sharing hunter who killed the animal or owned the dogs who helped to catch the
p of resources, animal. Large shares, always including the forelimbs, go to other hunters,
Okiek culture. carriers, and butchers. The degree to which limb joints are disarticulated
t by individual depends upon the number of people to whom primary distributions are
made. People benefiting from the primary distribution share with their
relatives and neighbors, who do the same. Ultimately everyone who knows
about the kill and lives within a few kilometers of a. house that received meat
.ery followed a during primary distribution is likely to accumulate some meat.
~hter and more Food sharing among the Okiek is patterned by two key factors: hunting
~ stomach and success and social relationships. In addition, meat sharing among the Okiek is
1g major joints. affected by the method of animal capture used, the size of the animal, and the
j to the dogs at context of consumption. ·
1e hunters. But Hunting Success
)ne case, when
the kill. In that Both qualitative and quantitative observations suggest that there is
ian t forest hog much variation among individual Okiek in hunting success. This is related
both to skill and to effort spent on hunting. Fifty percent of all animals in our
took place at a quantitative sample were obtained by one family largely as a result of the
ily or, in some quality of their pack of dogs; the other eight households were markedly less
o his friends or successful as hunters. Since successful hunters obtain both more meat and
ristics of Okiek different kinds of meat than other Okiek, this has a clear effect on patterns of
animal in our meat sharing and faunal accumulation.
•uld play a role Social Relationships
Ld y area, so we
Social relationships also affect the distribution of carcass parts.
ds of meat and This is particularly expressed in secondary and tertiary distribution of meat.
234j F. Marshall
Table 13-2. Distribution of Animal Carcasses Through Transport and great variabili·
Sharing can be expres~
ability in body
In a commu:
Body Parts · Number of Houses imately 10 km
Animal Left at Kill Shared Between
the most SUCCE
13-1). Over as
Bushbuck- 0 6 was made up<
Bushbuck 0 >4 tarsals, hindlir
Bushbuck 0 6 were also- large
Bushbuck 0 7
animal and ac
Bushbuck 0 3
brae, one head
Bushbuck 0 3
Bushbuck 0 3 animal but, bE
Bushbuck 0 4 accumulated n
Bu$hbuck 0 3 any animals ar
Bushbuck 0 3 primarily ribs ,
Giant forest hog · 0 >2 Thus, overt
Giant forest hog 2 >7 than do less su
Blueduiker 0 1 many bones, a
Blue duiker 0 1 elements. Lar
Cow 0 >2 hunters are ty]
Cow 0 1 because they r
Goat 0 4 share of and v
eaten by dogs
received from
variety of skE
Meat is shared with relatives and neighbors. Most people in our study area Household 5, 1
were not closely related so most sharing observed was between neighbors. cally of relath
However, Blackburn (1971) observes that among the Kipchornwone,k Okiek posed mostly <
meat is shared with kin (first with fathers and brothers; then witlf fathers, Up to now,
brothers, and sons; with mothers, brothers, and sisters; with fathers, sisters, archaeologists
and children) and neighbors, usually of the same lineage. that bones as~
Animal Size and Method of Capture carry will be tJ
Klein 1976; Pe:
Among Okiek large animals that are hunted with dogs and spears been taken to :
or bows and arrows are always shared, and there do not seem to be differ- been taken to
ences in the rules of sharing for different taxa. However, small animals (less ethnoarchaeol
than 10 kg), duiker, hyrax, rats, and birds, are not shared between hou~eholds. account, it ha~
In addition, animals that are snared or trapped rather than hunted w1th dogs limb bones an
are often not shared, apparently because few people know about the capture other bones (C
and also because many of the animals are small. - and the numbE
In addition to sharing, and factors affecting sharing, the distribution of accumulated b
bones across the landscape is also affected by Okiek settlement patterns and , those interpret
the fact that people live and eat in nuclear family settlements often made up of by less success
a single house. Thus, sharing takes place between rathe~ than within set~le part profiles tJ
ments or potential archaeological sites. As a result of the Influence of huntmg scavenged fam
success and social relationships on faunal assemblage accumulation, there is Fortunately,
Food Sharing I 235
'ransport and great variability in faunal accumulation between households over time. This
can be expressed in terms of species composition but is chiefly seen in vari-
ability in body part composition.
In a community of nine households living on adjacent ridges in an approx-
mber of Houses imately 10 km2 area of the high forest, one household (Household 1), that of
rred Between
the most successful hunter, accumulated inore bones than any other (Figure
13-1). Over a six-month period the assemblage accumulated by the household
6 was made up of a wider range of body parts and more skins with carpals and
>4 tarsals, hindlimbs, and metapodials than that of any other household. There
6 were also large numbers of forelimbs. By contrast Household 4 killed only one
7
animal and accumulated only one hindlimb, one forelimb, one set of verte-
3
3
brae, one head, and one set of ribs. Ho'wever, Household 9 also killed only one
3 animal but, because the householder is very well liked in the community,
4 accumulated not one but four forelimbs. Households 5, 6, 7, and 8 did not kill
3 any animals and thus accumulated few bones and some restricted body parts,
3 primarily ribs and portions offorelimb elements.
>2 Thus, over time the successful Okiek hunters have different access to meat
>7 than do less successful hunters and, as in the case of Household 1, accumulate
1 many bones, a wide range of body parts, and a high proportion of high utility
1 elementr Large numbers of bones are accumulated because successful
>2 hunters are typically involved in many butchery and sharing events and also
1 because they retain a large share of the carcass. In time, because of the large
4 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
)Ur study area Household 5, 6, 7, or 8, accumulate fewer bones and fewer body parts, typi-
~en neighbors.
cally of relatively low utility. In our examples the assemblages were com-
nwonek Okiek posed mostly of ribs and sometimes forelimb or hindlimb elements.
1 with fathers, Up to now, most variability of this kind has largely been interpreted by
~athers, sisters,
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
Jgs and spears been taken to indicate kill sites, and high proportions of. limb elements have
n to be differ- been taken to indicate camps. Lately those assumptions have been tested in
1 animals (less
ethnoarchaeological contexts and modified. Taking butchery costs into
~n households.
account, it has been suggested that vertebrae, scapulae, pelves, and upper
1ted with dogs limb bones are more likely to be transported from kill sites to camps than
mt the capture 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
iistribution of accumulated by successful hunters, such as those from Household 1, resemble
Lt patterns and those interpreted by archaeologists as base camps, whereas those accumulated
:en made up of by less successful hunters, such as Household 5, 6, 7, or 8,look more like body
t within settle-
part profiles that might be interpreted as kill sites, special purpose sites, or
nee of hunting
scavenged faunal assemblages.
lation, there is Fortunately, other features of site structure could be used to distinguish
I
236 F. Marshall
faunas accun
Household 1 (5 Kills) Hou5ehold 5 (0 Kills)
sharing, fror
Cranium volumes of rr
Vertebrae of tool-manu
Ribs~ 1977a). Bone
Forelimb and carnivor
Hindlimb +---r--...---.....-.....--l markers that
vores, from f<
Number Number
these proces~
Household 2 (2 Kills) Household 6 (0 Kills)
and are distir
Cranium Cranium In summar
t
!!I
a..
Vertebrae t::
<ll
a..
Vertebrae ing of food,
Ribs
~
Ribs between site~
~
IJ) Forelimb },:,lilil"""'i i!i i: : : :Wili: : yli !ilir-:: :ilti=i:m IJ) Forelimb !i:@i!liliiiil relationships.
Hindlimb ::::::::~t:!Mllif:::\tl}l\iliili Hindlimb +---.--.-------.----.--l range of site :
2 3 act to prodw
Number Number
1989; Lyman
Household 3 (1 Kill) Household 7 (0 Kills)
Cranium
t
<ll Vertebrae 0
a..
Ribs :itiiiiiiiiili
Forelimb tmi!i!i!i!ii H
Hindlimb +---.---r----r---.--l groups? Desf
contemporar:
Number Number factors that a1
Household 4 (1 Kill) Household 8 (0 Kills) there are two
social rela tim
Cranium
t
variable affec
<ll Vertebrae
a.. in a hunt.
Ribs
€0 Faunas ace
IJ) Forelimb
Hindlimb.
tionships. M<
~~--.---~-.....--l
3 3
parts are givE
Number Number of the hunt,<
Household 9 (1 Kill)
Yellen 1977a,
friends. Seco:
Cranium
the details of
1ij Vertebrae
a.. and even wi
>-
:§ depending 1.:
(Binford 1984
Difference1
Number
gatherers in
(Woodburn
Kaplan 1988;
Figure 13-1. Animal body parts accumulated by Okiek households over a specifically n
six-month period. Number of cranial, vertebral ~oluJ?m, half rib-c~~e, tion among f
forelimb (scapula, humerus, radius, ulna), and hzndlzmb (femur, tzbza, chooses whc
fibula, patella) units. All bones are whole. !Kung (Yeller
I
Food Sharing 237
faunas accumulated at residential.carnps by unsuccessful hunters through

.ills)
sharing, from kill or special purpose sites. In particular, relatively large
volumes of material culture, traces of hut structures, and residues of a variety
J
of tool-manufacturing activities are useful signatures of base camps (Yellen
1977a). Bone modification by carnivores and superpositioning of cut marks
and carnivore teeth marks (Shipman 1986a, 1986b) seem the most reliable
markers that would help to distinguish scavenging from nonhuman carni-
4 5 vores, from food acquired from successful hunters through sharing. However,
er
these processes will be difficult to differentiate as both are logically similar
(ills)
and are distinguished by late access to carcasses.
In summary, the Okiek data document a new behavioral mechanism, shar-
J
ing of food, that can cause considerable variation in body part profiles
between sites as a result of differential hunting success and individual social
relationships. They thus add to a growing body of literature that documents a
range of site formation processes other than selective bone transport that can
4 5 act to produce similar variation in body part profiles (Grayson 1989; Klein
>er
1989; Lyman 1984, 1985; Marshall and Pilgram 1991).
Kills)
J
Other Hunter-Gatherer Groups
How does this finding fit with data from other hunter-gatherer
groups? Despite the range of variation that exists in patterns of sharing among
3 4 5 contemporary hunter-gatherers, there seems to be remarkable consistency in
:>er factors that affect the way shared meat is distributed. Just as among the Okiek,
Kills) there are two key factors affecting the distribution of meat through sharing-
social relationships and hunting skill. In addition, animal size is an important
J
variable affecting the extent of sharing, as is the number of people taking part
in a hunt.
Faunas accumulated by individual households always reflect social rela-
tionships. Most groups have ·rules for division of shared meat. Certain body
3 5
parts are given to the person who killed the animal and to other participants
ber of the hunt, often on the basis of age or other social categories (Gould 1967;
Yellen 1977a, 1977b). Thereafter food is given first to close kin and then to
friends. Secondary and tertiary divisions follow similar patterns. However,
the details of who is entitled to which body part vary greatly between groups,
and even within a group, the way that an animal is divided up will vary
depending upon the social composition of the group at any given time
(Binford 1984; Yellen 1977a, 1977b).
Differences in hunting skill are documented for a wide range of hunter-
gatherers including the !Kung (Lee 1979; Yellen 1977a, 1977b), Badza
(Woodburn 1968), Aka (Hewlett 1988), Efe (Bailey 1988), Ache (Hill and
Kaplan 1988; Kaplan and Hill 1985), and Nunamiut (Binford 1984). Binford
Juseholds over a specifically notes that hunting success is a factor patterning faunal accumula-
'1,half rib-cage, tion among the Nunamiut, as the hunter receives more meat than others and
tb (femur, tibia, chooses who receives meat during carcass distribution. Even among the
!Kung (Yellen 1977a, 1977b) or Western Desert Aborigines (Gould 1967, 1981),
238IF. Marshall
where the hunter who killed the animal receives no more meat than anyone shared, such
else, it appears that successful hunters may receive consistently different and trated at one
possibly higher-utility body parts than do other members of the group (see However,
Speth 1990). Thus, individual hunting success appears to consistently faunal asserr
influence aspects of the formation of faunal assemblages. . rrt~n t is isola
Other factors identified as affecting sharing and faunal assemblages among whole fauna
the Okiek include animal size and method of capture. In general, as among settlement is
the Okiek~ large animals are shared more widely than small animals simply gatherers, sh
because of the amount of meat they provide. However, there is considerable are not affec
variation at the other end of the spectrum. Some groups such as the !Kung converge, si
(Yellen 1977a, 1977b) do not always share small animals; others such as the excavations 1
Aka (Hudson 1990), Western Desert Aborigines (Gould 1967, 1981), and Ache others. Thus.
(Kaplan and Hil11985) of Paraguay share animals of all sizes; Overall, large hunter-gathe
animal bones will tend to be more widely distributed than small animal bones ingofmeat.
as a result of different patterns of sharing. However,
The way in which animals are captured, and the number of hunters clear indicate
involved, influences patterns of sharing among many groups. Animals killed Okiek, body
using communal labor during net hunts among the Aka (Hewlett 1988) and able. This occ
caribou drives among the Eskimo (Ingold 1980) or speared from ocean-going because port:
canoes by Californian Indian groups (Gould 1981) are widely shared. The fact a single cult
that concealment is not possible in large groups is probably a factor in this resulting fro
widespread pattern. By contrast, animals killed by individuals, directly or carcass trans
using traps, are less uniformly shared. They are shared among groups such as tion processe
the Ache (Kaplan and Hill1985) and Western Desert Aborigines (Gould 1967, Fortuna tel
1981) who have a strong sharing ethic but not by others such as the Okiek or sharing, two
Aka (Hudson 1990). effects of wit
A final factor influencing the extent of sharing in sdme cases is the avail- study of the
ability of food. Binford (1984) notes that need is a consideration in Nunamiut Wiessner 19E
decisions about sharing. And Ingold (1980) points out that sharing i~ partic- hold and sib
ularly widespread among Arctic hunter-gatherers both when much food is occurrences 1
available and when food is very short. But there is little information on those sharing. Secc
points from other hunter-gatherer groups. have been u
Despite the apparent cross-cultural consistency in factors affecting sharing same animal,
among contemporary hunter-gatherers, the effect of food sharing on the 1989) and thr
accumulation of fauna in archaeological sites is hard to gauge as there have 1990). Enloe i
been relatively few field studies of faunal accumulation and food sharing. study of fau
However, some field data exist for the !Kung (Yellen 1977a, 1977b), the Aka Pincevent, aJ
(Hudson 1990), the Western Desert Aborigines (Gould 1967), and the blages sugge~
Nunamiut (Binford 1978, 1984). Those studies suggest that the effect of meat The archaE
sharing on household faunal assemblages is similar to the effect of meat ments is muc
sharing among the Okiek. · assemblage '
Individual households accumulate faunal assemblages with different body profiles ·at di
part profiles and sometimes different species; for example, among the Aka camps or os
and the !Kung, the bones of shared animals are scattered between many However, it
different household units within a camp (Binford 1984; Hudson 1990; formation of
Wiessner 1982a; Yellen 1977a, 1977b). And bones of animals that are not logical sites v
I
Food Sharing 239
than anyone shared, such as those of small animals among the !Kung, are spatially concen-
:lifferent and trated at one household unit (Figure 13-2).
e group (see However, the degree to which sharing influences ethnoarchaeological
consistently faunal assemblages as a whole varies with settlement patterns. Where settle-
ment is isolated, as in the Okiek case, sharing takes place between sites and
lages among whole faunal assemblages are affected 'by patterns of sharing. But where
11, as among settlement is nucleated, a more common pattern among contemporary hunter-
.mals simply gatherers, sharing takes place within sites, and faunal assemblages as a whole
considerable are not affected by sharing. Archaeologically, however,the patterns may
:ts the !Kung converge, since whole sites are rarely excavated and spatially restricted
; such as the excavations of nucleated sites are likely to sample some households and not
1), and Ache others. Thus, body part representation in most excavated samples of recent
>verall, large hunter-gatherer faunal assemblages is likely to have been influenced by shar-
mimal bones ing of meat.
However, there is no single pattern of body part representation that is a
~ of hunters clear indicator of sharing. Among other hunter-gatherer groups, as among the
1imals killed Okiek, body part profiles accumulated through sharing are extremely vari-
~tt 1988) and able. This occurs both because the rules for sharing vary between cultures and
ocean-going because portions received through sharing will vary situationally even within
.red. The fact a singlE?/ culture. In addition, some of the more typical body part profiles
factor in this resulting from sharing are similar to those resulting from the logistics of
:, directly or carcass transport, scavenging, and the destruction of bone during site forma-
oups such as tion processes.
(Gould 1967, Fortunately, however, while body part profiles are a major indicator of
the Okiek or sharing, two additional approaches have been taken to recognition of the
effects of within-camp sharing on faunal assemblages. The first of these is the
is the avail- study of the spatial distribution of animal taxa (Binford 1984; Hudson 1990;
:nNunamiut Wiessner 1982a). Hudson has developed a method of comparison of house-
lng is partic- hold and site MNis for distinguishing between food sharing and multiple
nuch food is occurrences of hunting of the same species by different households without
:ion on those sharing. Second, studies of the spatial distribution of body parts of animals
have been undertaken through identification of matching elements of the
:ting sharing same animal, such as right and left humeri, by osteometries (Enloe and David
uing on the 1989) and through refitting of bone fragments (Enloe and David 1989; Hudson
lS there have 1990). Enloe and David· (1989) have pioneered some of these methods in their
)Od sharing. study of faunal material and food sharing at the Upper Paleolithic site of
'7b), the Aka Pincevent, and Hudson's (1990) study of contemporary Aka faunal assem-
i7), and the blages suggests that the methods seem useful.
ffect of meat The archaeological identification of food sharing between isolated settle-
'feet of meat ments is much more difficult. Sharing might be indicated by the variability in
assemblage composition and by the presence of complementary element
lfferent body profiles at different sites. It is also possible that refitting of bones between
ong the Aka camps or osteometries might be used to identify between-camp sharing.
tween many However, it is clear that since food sharing patterns the early stages of the
udson 199'0; formation of the archaeological record recognition of food sharing in archaeo-
that are not logical sites will depend on exceptional conditions of preservation.
I
240 F. Marshall
These fine
C Hut n. ~ - . among conte
entia! transp'
Bones
.Q'... .: ·~ ~ '.) ered, as are 1
..
l-.\. • ·$· .. -:. . . •• ;.· interpreta tio
...
r ..
~'a·
\.:,
\
,---.,.· •:
.....:J
\., / .• .;_
assemblages .
Unfortuna
... ....
\...,;,
' .!..... ·~ .. logistics of e<
part signa tu
activities rna'
site formatio
Widely Shared Remains and post-dis
Marshall an
sharing can a
.0 J . ual carcasses
•-1ft .0 body part pr
tiona! metho'
~ . .~ that infl uen'
. . (}. . integrate faUJ
. ~ . Where ide
studies of pa
in archaeolo:
Little Shared Remains settlements.
relationships
Yellen's (197
Figure 13-2. Spatial distribution ofthe faunal remains of a widely shared organization
animal and a little-shared animal from a !Kung camp (after Yellen 1977b: supposedly r
307). ties that took
On a large
sharing was 1
Discussion and selectior
suggests thaj
The data presented here on patterns of food sharing among the addressed th:
Okiek and other hunter-gatherers provide both specific information on the It has been
effect of sharing of food on the faunal record and a context for reevaluating tion, namely,
the relevance of food sharing for archaeological studies of the development of ity and degrE
human social systems. rely on sped
The Okiek data show that in some contexts sharing of food may be the tarian, and a
single most important factor affecting early stages of the formation of faunal immediate rE
assemblages. It can create body part profiles similar to those commonly hunter-gathE
attributed by zooarchaeologists to differential transport of carcass parts, to 1980).
destruction of bone during food processing, or to post-discard site formation Because oJ
processes. Camps of less successful hunters have relatively fewer, and lower- different lev,
utility, body parts than those of successful hunters, similar to kill sites, special hunter-gathe
purpose sites, or scavenged assemblages. And large animals are generally less archaeologic
completely represented at sites than small q.nimals because they are more material culn
widely shared. ed return sul
Food Sharing 1241
These findings, together with the ubiquity of sharing of large animals

among contemporary hunter-gatherers, suggest that the possibility of differ-
ential transport of elements through food sharing should be routinely consid-
ered, as are transport based on anatomical utility and transport costs, in the
interpretation of patterns of body part representation in archaeological
assemblages.
Unfortunately, however, bone transport due neither to sharing nor to the
logistics of carcass utility and transport costs has distinctively different body
part" signatures. In addition, the body part profiles associated with those
activities may be similar to those created by later processes in the sequence of
site formation, such as the destruction of bone through carcass processing,
and post-discard site formation processes (Klein 1989; Lyman 1984, 1985;
Marshall and Pilgram 1991). Fortunately, in certain circumstances food
sharing can also be recognized by identification of the distribution of individ-
ual carcasses through refitting and osteometric techniques rather than through
body part profiles. The challenge for future field research is to develop addi-
tional methods of recognition that will help to discriminate between processes
that influence the body part composition of faunal assemblages and to
integrate faunal and nonfaunal signatures of specific site formation processes.
Where 1identification of food sharing is possible on archaeological sites,
studies of patterns of sharing may provide insights into larger research issues
in archaeology, such as the social relations within and between prehistoric
settlements. The fact that Okiek faunas reflect hunting success and social
relationships more than site function, such as kill or butchery site, fits with
Yellen's (1977a:135) observation that "seemingly 'abstract' aspects of social
;idely shared organization may be more easily reconstructed from camp debris than
~'ellen 1977b:
supposedly more simple aspects as delineating the nature and range of activi-
ties that took place."
On a larger scale, Isaac (1977a, 1977b) suggested that identification of food
sharing was relevant to understanding the development of social relationships
and selection processes on Plio-Pleistocene archaeological sites. This study
suggests that there are also issues in later prehistory that could usefully be
among the addressed through studies· of food sharing in the faunal record .
.tion on the It has been argued that a link exists between types of subsistence organiza-
eeval ua ting tion, namely, immediate -return and delayed return, and differences in mobil-
~lopment of ity and degree of egalitarianism. Groups with delayed return subsistence who
rely on specialized production and storage tend to be less mobile, less egali-
may be the tarian, and accun1ulate more specialized material culture than groups with
m of faunal immediate return subsistence systems. The former are often labeled complex
commonly hunter-gatherers (Bernard and Woodburn 1988; Hayden 1990; Woodburn .
.ss parts, to 1980).
e formation Because of the relationships between different patterns of sharing and
and lower- different levels of complexity of social organization among contemporary
dtes, special hunter-gatherers .(Table 13-3), identification of patterns of food sharing in the
~nerally less archaeological record could be used in combination with other aspects of
yare more material culture to address such questions as the relationship between delay-
ed return subsistence systems among hunter-gatherers and the environment
242j F. Marshall
Table 13-3. Food Sharing and Contrasts Between Subsistence and Archaeoi
Social Organization in Contemporary Hunter-Catherers 65. Proc
Archaec
Barnard, A.i
Immediate Return Delayed Return 1988 Pre
tion. In
D. RichE
Widespread sharing Restricted sharing
Binford, L. R
Generalized production Specialized production 1978 Nu
Little storage Much storage 1980 Wi
Highly mobile Less mobile
Archae<
More egalitarian Less egalitarian 1981 801
Less complex social systems More complex social systems 1984 Bu
More intergroup interaction Less intergroup interaction ical Arcl
Less complex material culture More complex material culture 1986 Cu
1988 Fac
Sources: Barnard and Woodburn 1988; Binford 1980; Hayden 1990; Wiessner 1982a, 1982b; pretatio
Woodburn 1980. Blackburn, R
1971 Ho
ment of
(see Enloe and Davis 1989) and between social complexity and the adoption of Arbor.
1982 In
food production (see Hayden 1990; Meillasoux 1973).
Neighb1
In conclusion, the identification of sharing in the faunal record is still in its
andR.I
early stages. Recent ethnoarchaeological studies suggest that it is possible and Blumenschin
that there are both methodological research issues that should be approached 1986 Em
and larger research issues that could usefully be approached through studies the Seret
of food sharing in the faunal record. Bunn,H. T.
1981 An
fromKc
Acknowledgments 1982 Me
I
nia, Ber:
I thank the Office of the President and N ationa! Reseaich Council Bunn, H. T.,.
of Kenya for permission to conduct research and the National Museums of 1988 Va
Kenya for their logistical support. I am grateful to the Leakey Foundation and ing, and
to the National Science Foundation for funding. Jean Hudson and Jim Enloe Bunn, H. T.,'
were generous with ideas and information from their own work on food 1986 Sy1
sharing; Tom Pilgram provided logistical support and intellectual challenges, Tanzan]
Corinne Kratz introduced me to the Okiek, and the Okiek themselves pro- 1988 Fac
vided the bulk of the data and all of the surprises. To these people I am very Interpre
grateful. This paper is dedicated to the'memory of Glynn Isaak. Cashdan, E. ,
1985 Co
Man 20:·
References Damas, D.
1975. CeJ
Bahuchet, S. Enloe, J. E., a
1990 Food Sharing among the I?ygmies of Central Africa. African Study Mono- 1989 Le
graphs 11:27-53. Magda!,
Bailey, R. Societe F
1988 The Significance of Hypergyny for Understanding Subsistence Behaviour Gould, R. A.
among Contemporary Hunters and Gatherers.· In Diet and Subsistence: Current 1967 No
Food Sharing I 243
1ce and Archaeological Perspectives, edited by B. V. Kennedy and G. M. LeMoine, pp. 57-
65. Proceedings of the 19th Annual Chacmool Conference, University of Calgary
Archaeological Association, Calgary.
Barnard, A., and J. Woodburn
1988 Property, Power and Ideology in Hunter-Gathering Societies: An Introduc-
tion. In Hunters and Gatherers 2: Property, Power and Ideology, edited by T. Ingold,
D. Riches, and J. Woodburn, pp. 4-31. Berg, New York.
Binford, L. R.
lon
1978 Nunamiut Ethnoarchaeology, Academic Press, New York.
Archaeological Site Formation. American Antiquity 45:1-17.
1981 Bones: Ancient Men and Modern Myths. Academic Press, New York.
1systems
1984 Butchering, Sharing, and the Archaeological Record. Journal of Anthropolog-
raction
ical Archaeology 3:235-257.
rial culture
1986 Current Anthropology Comment. Current Anthropology 27:444-446.
1988 Fact and Fiction about the Zinjanthropus Floor: Data, Arguments, and Inter-
82a, 1982b; pretations. Current Anthropology 29:123-135.
Blackburn, R.H.
1971 Honey in Okiek Personality, Culture, and Society. Ph.D. dissertation, Depart-
ment of Anthropology, Michigan State University. University Microfilms, Ann
~ adoption of Arbor.
1982 1In the Land of Milk and Honey: Okiek Adaptations to Their Forests and
l is still in its Neighbors. In Politics and History in Band Societies, edited by E. Leacock, E. Lee,
and R. B. Lee, pp. 283-305. Cambridge University Press, New York.
possible and
Blumenschine,R. J.
approached 1986 Early Hominid Scavenging Opportunities: Implications of Carcass Availability in
Jugh studies the Serengeti and Ngorongoro Ecosystems. BAR International Series 283. Oxford.
Bunn,H. T.
1981 Archaeological Evidence for Meat-Eating by Plio-Pleistocene Hominids
from Koobi Fora and Olduvai Gorge~ Nature 291:574-577.
1982 Meat-eating and Human Evolution. Ph.D. dissertation University of Califor-
1
nia, Berkeley. University Microfilms, Ann Arbor.

:1rch Council Bunn, H. T., Bartram, L. E., and E. M. Knoll
Museums of 1988 Variability in Bone Assemblage Formation from Hadza Hunting, Scaveng-
mdation and ing, and Carcass Processing. Journal of AnthropologicalArchaeology 7:412-457.
1d Jim Enloe Bunn, H. T ., and E; M. Kroll
ork on food 1986 Systematic Bftchery by Plio/Pleistocene Hominids at Olduvai Gorge,
1 challenges, Tanzania.Current Anthropology 27:431 -452. ·
nsel ves pro- 1988 Fact and Fiction about the FLK Zinjanthropus Floor: Data, Arguments, and
>le I am very Interpretations. Current Anthropology 29:135-149.
Cashdan, E. A
1985 Coping with Risk: Reciprocity among the Basarwa of Northern Botswana.
Man 20:454-474.
Damas, D.
1975 Central Eskimo Systems of Food Sharing. Ethnology 11:220-239.
Enloe, J. E., and F. David
~ Study Mono- 1989 Le Remontage des Os par Individus: le Partage du Renne Chez les
Magdaleniens de Pincevent (La Grande Paroisse, Seine-et-Marne). Bulletin de la
Sodite Prehistorique Franc;aise 86:275-281.
tee Behaviour Gould, R. A. ·
;tence: Current 1967 Notes on Hunting, Butchering, and Sharing of Ga1,11e among the Ngatatjara
244j F. Marshall
and Their Neighbors in the West Australian Desert. .The Kroeber Anthropological Lee, R.
Society Papers 36:41-66. 1979 The
1981 Comparative Ecology of Food-Sharing in Australia and Northwest Universil
California. In Omnivorous Primates; edited by Robert S. 0. Harding .and Geza Lyman,R. L.
Teleki, pp. 422-454. Columbia University Press, New York. ,1984 Bon
Grayson, D. K. Anthropo
1989 Bone Transport, Bone Destruction, and Reverse Utility Curves. Journal of 1985 Bon
Archaeological Science 16:643-652. MGUI.Jc
Hayden,,B. Marshall, L.
1990 Nimrods, Piscators, Pluckers, and Planters: The Emergence of Food Pro- 1976 Sha:
duction. Journal of Anthropological Archaeology 9:31-69. Kalahari,
Hewlett, B. S. R. B. Lee
1988 Sexual Selection and Paternal Investment among Aka Pygmies. In Human Marshall, F., c:
Reproductive Behavior: A Darwinian Perspective, edited by Laura Betzig, Monique 1991 Mec
Borgerhoff Mulder, and Paul Turke, pp. 263-276. Cambridge University Press, Part Rer
. Cambridge. 18:149-11
Hill, K., and A. Hurtaldo Meillasoux, C
1989 Hunter-Gatherers of the New World. American Scientist 77:5:437-443. 1973 On
Hill, K., and H. Kaplan African E.
1988 Tradeoffs in Male and Female Reproductive Strategies among the Ache: Oxford.
Part 1. In Human Reproductive Behavior: A Dmwinian Perspective, edited by Laura O'Connell, J. :
Betzig, Monique Borgerhoff Mulder, and Paul Turke, pp. 277-289. Cambridge 1990 Rea
University Press, Cambridge. nal of An
Hudson,J. Perkins, D., aJ
1990 Identifying Food Sharing Archaeologically: An Ethnoarchaeological Study 1968 A :f.
among the Aka. Paper presented at the Sixth International Conference on Hunt- Potts, R.
ing and Gathering Societies, Fairbanks, Alaska. 1984 Hm
Huntingford, G. W. B. 1988 Earl
1942 The Social Organization of the Dorobo. Anthropos 46:1-48. Potts, R., and
1954 The Political Organization of the Dorobo. Anthropos 49:123-148. 1981 Cut
1955 The Economic Life of the Dorobo. Anthropos 50:602-634 / Nature 2~
Ingold, T. Shipman, P
1980 Hunters, Pastoralists, and Ranchers. Cambridge University Press, Cambridge. 1983 Ear:
Isaac, G. L. ing? In 1
1977a The Food Sharing Behavior of Protohuman Hominids. Scientific American 31-50. B.
238:90-108. 1986a Sea·
1977b Food Sharing and Human Evolution: Archaeological Evidence from the Tests. Ar
Plio-Pleistocene of East Africa. Journal of Anthropological Research 34:311-325. 1986b Stw
Kaplan, H., and K. Hill Evolutior.
1985 Food Sharing among Ache Foragers: Tests of Explanatory Hypotheses. Speth, J.D.
Current Anthropology 26:223-246. · 1990 Sea:
Klein, R.G. Foragin~
1976 The Mammalian Fauna of the Klasies River Mouth Sites, Southern Cape Steward, J.
Province. South Africa. South African Archaeological Bulletin 31:75-98. 1938 Basi
1989 Why Does Skeletal Part Representation Differ Between Smaller and Larger Ethnolo~
Bovids at Klasies River Mouth and Other Archaeological Sites? Journal of Archae- Washburn, S.
ological Science 16:363-381.
Kratz, C. A. · 1961 Soc
1977 The Liquors of Forest and Garden: Drinking in Okiek Life. Unpublished edited b'
Master's thesis, Wesleyan University, Middletown, Conn. Wiessner, P.
1981 Are the Okiek Really Masai? Or Kipsigis? or Kikuyu? Cahiers d'Etudes 1982a Bey
Africaines. 79:355-368. of Hunte
Food Sharing 1245
Anthropological Lee, R.
1979 The !Kung San: Men, Women, and Work in a Foraging Society. Cambridge
nd Northwest University Press/ Cambridge.
ding and Geza Lyman, R. L.
1984 Bone Density and Differential Survivorship of Fossil Classes. Journal of
.rves. Journal of · 1985 Bone Frequencies: Differential Transport/ In Situ Destruction, and the
MGUI. Journal of Archaeological Science 12:221-236.
Marshall/ L.
:e of Food Pro- 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.
nies. In Human Marshall, F., and T. Pilgram ,
etzig/ Monique 1991 Meat Versus Within-Bone Nutrients: Another Look at the Meaning of Body
niversity Press/ Part Representation in Archaeological Sites. Journal of Archaeological Science
18:149-163.
Meillasoux, C.
437-443. 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,
10ng the Ache: Oxford.
~dited by Laura O'Connel1/ J. F., Hawkes, K./ and N. Blurton Jones
89. Cambridge 1990 ;Reanalysis of Large Mammal Body Part Transport among the Hadza. Jour-
nal of Archaeological Science 17:301-316;
Perkins, D./ and P. Daly
~ologicalStudy 1968 A Hunters Village in Neolithic Turkey. Scientific American 219:97-106.
renee on Hunt- 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
48. 1981 Cutmarks Made by Stone Tools on Bones from Olduvai Gorge, Tanzania.
Nature 291:574-577.
Shipman/ P
ss, Cambridge. 1983 Early Hominid Lifestyle: Hunting and Gathering or Foraging and Scaveng-
ing? In Animals and Archaeology, edited by J. Clutton-Brock and C. Grigsori, pp.
~ntific American 31-50. BAR International Series 163. Oxford.
1986a Scavenging or Hunting in Early Hominids: Theoretical Framework and
lence from the Tests. American Anthropologist 88:27-43.
:4:311-325. 1986b Studies of Hominid-Faunal Interactions at Olduvai Gorge. Journal of Human
Evolution 15:691-706.
y Hypotheses. Speth, J.D.
1990 Seasonality, Resource Stress and Food Sharing in So-Called "Egalitarian"
Foraging Societies. Journal of Anthropological Archaeology 9:148-188.
Southern Cape Steward, J. ·
98. 1938 Basin-Plateau Aboriginal Groups. Smithsonian Institution. Bureau of American
,er and Larger Ethnology Bulletin 120.
urnal of Archae- Washburn, S. L., and I. DeVore
1961 Social ~ehavior of Baboons and Early Man. In Social Life of Early Man,
. Unpublished edited by S. Washburn/ pp. 91-105 Aldine, Chicago .
Wiessner, P.
~ahiers d' Etudes 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
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.
14. F
1980 Hunter-Gatherers Today and Reconstruction of the Past. In Soviet and E
Western Anthropology, edited by Ernest Gellner, pp. 95-117. Columbia University
Press, New York. · E
Yellen, J. E.
1977a Archaeological Approaches to the Present: Models for Reconstructing the Past.
l
1977b Cultural Patterning in Faunal Remains: Evidence from the !Kung Bushmen.
In Experimental Archaeology, edited by D. Ingersoll, J. Yellen, and W. Macdonald, Jc
pp. 271-331. Columbia University Press, New Yorl<.
Al
in
tu
so
te1
Tl
iZI
Fa
in
co
Fa
inJ
tic
sti
Ill
Tl
tion, transpir
where in the
arose for int(
foragers (hun
Prehistoric
many parts 1
beginnings o
From Bones to Be
Faunal Remains,
Paper No. 21. ~
reserved. ISBN C
nies. In Politics
Lee, pp. 61-84.
ment. Journal of
ed by R. B. Lee 14. Foragers and Farmers: Material

:. In Soviet and Expressions of Interaction at
nbia University
Elephant Processing Sites in the
ucting the Past.
lturi Forest, Zaire
(ung Bushmen. John W. Fisher, Jr.
W. Macdonald,
Abstract: Interactions between foragers and farmers probably took place

in many parts of the world prehistorically after the emergence of agricul-
tural economie~. Such interactions likely had important economic and
social consequences. Learning to identify prehistoric forager-farmer in-
teractions is vital if archaeologists are to investigate such occurrences.
This study examines forager-farmer interactions in the context of special-
ized sites associated with elephant butchery and processing in the Ituri
Forest, Zaire, with the goal of establishing criteria for identifying these
interactions, using attributes of site structure and faunal assemblage
composition. Contrasting subsistence adaptations between these Ituri
Forest farmers and foragers, especially regarding food storage, translate
into recognizable material differences at the sites. The short-term dura-
tion of occupation of these sites and various other factors condition their
structure. ·
Introduction
The origin and elaboration of agriculture as a subsistence adapta-
tion, transpiring in southwestern Asia some 10,000 years ago and later else-
where in the world, opened a novel set of human associations. Opportunities
arose for interactions between sedentary horticulturalists and neighboring
foragers (hunter-gatherers) who shared prehistoric landscapes.
PrehistoriC interactions· between foragers and farmers likely took place in
many parts of the world. Eastern North America, for example, saw the
beginnings of indigenous plant domestication between 4000 and 3000 B.P.
From Bones to Behavier: Ethnoarchaeological and Experimental Contributions to the Interpretation of

Paper No~ 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights
247
248 I J. W. Fisher,]r.
Food production emerged as an important economic strategy a millennium Archaeologis

later, and during the following centuries indigenous domesticates com- farmer intera
manded increasing dietary importa~ce. Indigenous domesticates were re- exists, thus, f
placed by maize between A.D. 800 and 1.100, accompanied by the advent of interactions.
greater sociopolitical complexity (Smith 1989). The persistence of foragers in Recognizin
eastern North America during the emergence and development of agriculture and foragers
potentially created opportunities and incentives for interactions with farmers. a relatively u:
The Northern Great Plains of North America, likewise, was the setting during foragers and .
both prehistory and the historic period for interactions between village- ever, present
dwelling horticulturalists and nomadic foragers (Wood 1974). Farmers of the bones in tra~
North American Southwest exchanged foods and material goods during pre- Southern PL
historic and historic times with nomadic foragers who inhabited the adjacent (Spielmann 1 1
Southern Plains (Spielmann 199lb). The prehistoric European landscape Archaeolo!

offered opportunities for forager-farmer interactions during the spread of increasingly 1
agriculture (Milisauskas 1978); indeed, such would seemingly be the case ences from d
everywhere in the world where farming developed in or diffused into an tion of reside
existing forager milieu. · · studied by aJ
Numerous ethnohistoric and ethnographic cases of forager-farmer inter- between for;
actions can be cited from North America, Africa, Southeast Asia, and else- Hitchcock 19
where (e.g., Bailey et al. 1989; Ewers 1954; Headland and Reid 1989). Such the spatial o:
associations take a variety of forms, fulfill many functions and needs., and can specialized c<
yield complex relationships that involve exchange ·(of foods, material goods, processing b~
labor, ideas, and people), participation in each other's ceremonial and ritual Ituri Forest, ~
events, and the like. These examples offer tantalizing insinuations to the remains, the]
significance of forager-farmer interactions prehistorically.
Identifying Forager-Farmer Interactions Archaeologically

Tl
Archaeologists have a variety of strong incentives for d~veloping compared to
method and theory to identify instances of forager-farmer interactions archaeo- acteristics po
logically. Recognizing the existence of interactions is a prerequisite to study- cialized carcc
ing the presence and nature of such relationships when nonarchaeological nature of fo
evidence is absent or inadequate. Forager-farmer interaction as a cultural hypothesized
phenomenon possesses its own intrinsic interest as a topic of study. Moreover, products wit
failure to recognize and investigate the nature of forager-farmer relationships bison huntin~
can result in incomplete or misleading perceptions about economic, social, 1991a). Arch
and political adaptations. The ongoing debate over the time depth and nature nature of su
of forager interactions with agropastoralists in the Kalahari Desert of southern villages. Ana
Africa (Solway and Lee 1990; Wilmsen and Denbow 1990) underscores the least send UI
crucial importance of archaeological input. The tropical rain forest of central farmers had 1
Africa offers another example of the fundamental importance of clarifying the
time depth of forager-farmer interactions. Recent ecologically oriented studies
have challenged the traditional anthropological notion that foragers (Pygmies) T
inhabited the rain forest for untold millennia prior to the advent of horticul-
ture (Bailey et al. 1989; Hart and Hart 1986; Headland and Reid 1989). Bailey T1
and colleagues (1989) hypothesize that tropical forests in Africa and elsewhere ethnographic
might have been uninhabitable until domesticated foods were available. generate insi
Foragers and Farmers I 249
a millennium Archaeologists can test that hypothesis by tracing the antiquity of forager-
sticates com- farmer interactions in the Ituri Forest and other tropical forests. Ample cause
ates were re- exists, thus, for archaeologists to take interest in the study of forager-farmer
the advent of interactions.
of foragers in Recognizing archaeologically the existence of interactions between farmers
of agriculture and foragers is a challenging task. The presence of exotic material items offers
with farmers. a relatively uncomplicated line of evidence for establishing exchange between
;etting during foragers and farmers (Spielmann 1991a; Wood 1974). Such evidence can, how-
¥een village- ever, present ambiguities. It is difficult to resolve, for example, whether bison
,'armers of the bones in trash middens of Southwestern farmers represent exchange with
ls during pre- Southern Plains foragers or hunting forays by the farmers themselves
:i the adjacent (Spielmann 1991a:42). ·
an landscape Archaeological investigation of forager-farmer relationships will become
the spread of increasingly stronger as we develop sound method and theory to make infer-
y be the case ences from diverse classes of evidence. Settlement pattern, spatial organiza-
fused into an tion of residential sites, and the composition of faunal assemblages have been
studied by archaeologists and ethnoarchaeologists to document interactions
·farmer inter- between foragers and farmers (or pastoralists; e.g., Brooks et al. 1984;
sia, and else- Hitchcock 1987). This study presents an ethnoarchaeological examination of
d 1989). Such the spatial organization and faunal assemblage composition of temporary,
eeds, and can specializkd campsite/processing sites.associated with elephant butchery and
aterial goods, processing by Efe Pygmy foragers and sedentary Lese horticulturalists of the
tial and ritual Ituri Forest, Zaire, and attempts to develop criteria to identify, using material
.ations to the remains, the presence of foragers and horticulturalists at the sites .
Why Specialized, Short-term Campsite/Carcass Processing Sites?

ally
. These sites possesses several important qualities. They can be
)f developing compared to ethnoarchaeologically studied residential ~amps to seek char-
tions archaeo- acteristics potentially diagnostic of forager-farmer interaction. Moreover, spe-
lsite to study- cialized carcass processing sites might be more sensitive in some ways to the
rchaeological n~ture of forager-farmer interactions than residential sites. It has been
as a cultural hypothesized, for example, that Southern Plains foragers who exchanged food
:iy. Moreover, products with Southwestern farmers might have actively sought to prevent
· relationships bison hunting by the farmers in order to attain sole access to bison (Spielmann
1omic, social, 1991a). Archaeologists might have difficulty recognizing the competitive
th and nature nature of such ·a relationship in the bison bone-filled middens at farmer
rt of southern villages. Analysis of specialized bison kill/butchery sites will, however, at the
derscores the least send up an archaeological warning flag if it can be demonstrated that
rest of central farmers had not participa~ed.
clarifying the
lented studies
ers (Pygmies) Theoretical Perspectives
1t of horticul-
. 1989). Bailey The Efe-Lese interaction presented here constitutes a particularistic
md elsewhere ethnographic case that can be integrat~d into alarger theoretical context to
~re available. generate insights of general significance. The determinants of subsistence-
250 I ]. W. Fisher, [r.
settlement patterns, of the spatial organization of forager residential camp~ among theN
sites, and of the composition of animal bone assemblages offer baselines for parts from kil
examining these specialized campsite/processing sites. meat, marroV\
Binford's (1980) analysis of ethnographically known hunter-gatherer subsis- hood of tran
tence-settlement strategies distinguishes two fundamental types of organiza- (butchery, trm
tiona! strategies, which he designates as foragers and collectors (the term forager 1988; O'Conn
in this specialized meaning is to be distinguished from my use elsewhere in general goal o
this paper of the word forager to denote hunter-gatherers in general). The costs in trans!
determinants of subsistence-settlement organization lie in environmental The Efe-Le:
characteristics, including the distribution and availability of food resources: subsistence-s~
where resource distribution is relatively undifferentiated, forager tactics are exclusive fore
effective; and where critical resources have differential availability either The Efe acqu:
spatially or seasonally, a logistical strategy is favored (Binford 1980; Kelly than brief P'
1983). The forager_ and collector organizational strategies each embody a intervals.
broad range of variability and form a continuum rather than diametrically The Lese m
opposed types (Binford 1980, 1987). keyed on cult
The forager strategy is characterized by the absence or rarity of food stor- consume wile
age; rather, food is gathered on a daily basis. Foragers generally obtain seed (Bailey a
resources on trips that begin and end at the residential base on the same day. The symbic
The majority of manufacturing, maintenance, and processing activities occur Efe and Lese
at the residential camp, and_the material record at the site is conditioned by the calories E
length of occupation, group size, and seasonal variation in activities. The Lese gardens
collector strategy differs in important ways from the forager strategy and constrained n
yields a different material record. Collectors practice food storage during at day's journey
least some of the year. Residential camp moves are relatively infrequent. They and Stricklan1
obtain food and other resources through logistically organized, special- The special
purpose procurement parties that leave the residential camp and create a processing el1
variety of site types over the landscape (Binford 1980). and subsiste1
The spatial organization of residential campsites, as documented ethno- resulting in u
archaeologically among several hunter-gatherer groups including the !Kung composition 1
and other San (Basarwa) of southern Africa, the Hadza of Tanzania, the
Alyawara of Australia, and the Efe of Zaire, indicate the existence of a general
pattern with variant expressions. The general pattern includes the presence of
"household, communal, and special activity areas" (O'Connell et al. 1991:72). TI
Considerable variability exists in such attributes as total site size, number of Ituri Forest, -_:2
special activity areas, distance between dwellings, and so forth. Circum- Efe by the a1
stances that influence spatial organization in~lude social, economic, environ- Strickland 19
mental, and situational factors, such as 'kinship ties, the degree to which food the Efe and I
is shared between households, the length (actual or anticipated) of occupation, still active (Be
the presence of dangerous predators, seasonality in resource availability, and The Ituri P,
density of vegetation (Binford 1987; Fisher and Strickland 1991; Gargett and above sea le'
Hayden 1991; Hitchcock 1987; O'Connell et al. 1991; Yellen 1976). 2,000 rnm of 1
Animal butchery, carcass transport, processing activities, and consumption three months
practices have been the target of efhnoarchaeological research among various rainfall at the
groups of hunter-gatherers. Although considerable variability in patterns of DeVore 1989)
carcass use has been documented between and within groups, generalization The Efe ar
about underlying determinants might be possible. Binford's (1978) research consisting of
lential camp- among the Nunamiut Eskimo has shown that collectors transport carcass
baselines for parts from kill site to residential camp selectively; the greater the amount of
meat, marrow, and grease available from a body part, the greater the likeli-
therer subsis- hood of transport. Variability in forager behavior regarding carcass use
' of organiza- (butchery, transport, and processing) is considerable (Binford 1987; Bunn et al.
Le term forager 1988; O'Connell et al. 1988), but perhaps underlying this variability is the
elsewhere in general goal of realizing the greatest net returns in nutritional value relative to
general). The costs in transport and processing (0~-Connell et al. 1988).
1vironmental The Efe-Lese case can now be placed in context. The Efe follow a forager
od resources: subsistence-settlement strategy. This is not surprising because "the most
~er tactics are exclusive foragers are best knownfrom equatorial forests" (Binford 1980:7).
ability either The Efe acquire food on a daily basis and do not store provisions for more
d 1980; Kelly than brief periods. They move residential camps at relatively frequent
:h embody a intervals.
diametrically The Lese are sedentary, village-living horticulturalists. Their subsistence is
keyed on cultivated plants, especially cassava and bananas, although they also
of food stor- consume wild plant and animal foods. They store some crops for food and for
erally obtain seed (Bailey and DeVore 1989).
:he same day. The symbiotic, or mutualistic (Spielmann 1991b), relationship between the
:tivities occur Efe and Dese influences the Efe subsistence-settlement pattern. Two-thirds of
mdi tioned by the calolies Efe consume come from cultivated foods obtained largely from
ctivities. The Lese gardens (Bailey and DeVore 1989). The Efe settlement pattern is one of
strategy and constrained mobility; they move relatively frequently_ but never more than a
1ge during at day's journey from Lese villages and gardens (Bailey and DeVore 1989; Fisher
~equent. They and Strickland 1989, 1991).
.zed, special- The specialized campsite/processing site where Efe and Lese cooperate in
and create a processing elephant carcasses constitutes a setting at which forager campsite
and subsistence patterns are juxtaposed with those of sedentary farmers,
ten ted ethno- resulting in unique patterns in site spatial organization and bone assemblage
ng the !Kung composition that are not present at the residential sites of either group.
['anzania, the
:e of a general
te presence of Background of the Efe and Lese
~tal. 1991:72).
The Efe and Lese discussed here liv,e in the northeastern part of the
~e, number of
Ituri Forest, Zaire. Ethnoarchaeological fieldwork was undertaken among the
)rth. Circum- Efe by the author and Helen Strickland for a year in 1984-85 (Fisher and
mic, environ- Strickland 1989). Our research is part of a long-term research project among·
:o which food the Efe and Lese involving numerous investigators that began in 1980 and is
)f occupation, still active (Bailey and DeVore 1989).
:tilability, and The Ituri Forest is a tropical rain forest. The study area lies at about 900 m
; Gargett and above sea level, is situated at about 2°15' north latitude, and receives about
2,000 mm of rainfall annually (Figure 14-1). A pronounced dry season of about
consumption three months' duration occurs between December and February, and monthly
:nong various rainfall at that time is reduced to about 1/30 of the annual total (Bailey and
ln patterns~of DeVore 1989).
;eneralization The Efe are seminomadic foragers. They live in small, temporary camps
978) research consisting of about 2-50 residents. The length of time that they inhabit a camp
252 I f. W. Fisher, Jr.
Figure 14-1. Location of the study area in the Ituri Forest, Zaire.
varies from a few days to three months or more and averages about six weeks
(Bailey and DeVore 1989;Fisher and Strickland 1989). The spatial organization
of Efe residential campsites is of particular interest here, for comparison to the
organization of elephant processing sites. Efe campsites conform to a single,
general pattern of organization (Figure 14-2), although each camp is unique in
details of layout. Efe campsites in our sample (N = 30) range in size from 44
m2 to 532m2. Camps are conta.ined within an oval area cleared of underbrush
and small trees. Dwellings are situated near the perimeter of the clearing, with
doorways usually facing toward the center. Trash heaps occur at ~11 campsites
and are situated immediately beside and behind dwellings; they begin as piles
of cleared vegetation and grow during the life of the camp as food refuse,
broken implements, ashes, and so forth are discarded. The Efe build fires in-
side dwellings and outside dwellings near the door (Fisher and Strickland
1989, 1991). labor, hor
The spatial organization of Lese villages also is of interest for comparative· and impl1
purposes. The Lese live in permanent villages, with populations ranging from Lese alsc
15 to 100 residents (Bailey and DeVore 1989). Lese villages are situated in marriage.
clearings from which all trees and other vegetation have been removed. The
cleared ground around and between village structures is kept bare of vegeta-
tion and debris. Each village contains one or more wattle-and-daub dwellings,
a roofed, wall-less baraza under which men sit to escape the sun or rain or to
socialize, and a mafika (kitchen) that contains one or more fires and is struc-
turally very similar to the baraza, where women cook and socialize (Figure 14- Lese und
3). < usuallyh
The Efe and Lese share long-term, complex, and close social and economic provides
relationships with each other (Bailey and DeVore 1989). Those ties involve the Lese. Dri£
exchange of food, material goods, and labor. The Efe provide the Lese with to an out
. Zaire.
,out six weeks

1 organization
parison to the
m to a single,
p is unique in
1 size from 44
Jf underbrush
clearing, with
t all campsites
begin as piles Figure 14-2. Plan map of an Efe campsite showing the primary struc-
s food refuse, tural components.
build fires in-
nd Strickland
labor, honey, wild game, and other forest products in return for clothing, tools
r comparative and implements, and cultivated foods. The relationship between the Efe and
ranging from Lese also entails sociocultural domains such as ritual service and inter-
re situated in marriage.
removed. The
~are of vegeta-
mb dwellings, Elephant Butchery and Processing
rr or rain or to
; and is struc- Elephant butchery and processing is an enterprise that the Efe and
[ze (Figure 14- Lese undertake cooperatively. Elephants are killed at irregular intervals,
usually by ivory hunters using a high-powered rifle. The death of an elephant
and economic provides a carcass for butchery, processing, and consumption by local Efe and
es involve the Lese. Dried meat can be stored by the Lese for months for consumption or sale
the Lese with to an outside market. Traditional techniques such as pit traps, heavy spears
254 I f. W. Fisher, Jr.
ings, Lese dv
local materi
residential c;
broad leaves
shows less la
the sapling :
impression tl
at residentia
several post~
rear. of the s
Drying rack~
about 1m ab
dry and smo'
to 848m2, wi
Butchery t
five hours. :M
drying on ra'
68 hours (thE
tion while h t
consumed at
disarticulate'
are chopped
the surface o
for their mm
boiled for gr
have comple
erably: in on
other case rr
Figure 14-3. Plan map of a Lese village showing the primaryltructural number of p
components. are importan
elements alrr
suspended over trails, and spear hunts were used to kill elephants in the past. innominates.
A high-powered rifle yields the same end-product as traditional technolo- of marrow a'
gies-an individual elephant carcass. sion of butch
We observed elephant butchery and processing on three occasions and
visited and mapped eight other sites where those activities had occurred from
one week to about six years prior to our visit (Fisher 1992). Elephants usually
are killed deep in the forest, several hours' walk from the nearest village and c
camp. The people taking part in butchery and processing reach the carcass the material
one or two days after the animal's death, and butchery begins, in our experi- dence found
ence, after 30-65 hours have elapsed since the animal's death. The number of structure (sp;
participants ranges from 25 to 35 Lese and Efe adults of both sexes, plus This questi
children. The people set up a short-term camp within 10-70 m of the carcass the specializE
where they live and process the carcass parts. They butcher the carcass using a Efe and the'
relatively simple technology: metal spears, knives, machetes, axes, hatchets, can be distir
pole levers, and vine ropes. ments to est;
The temporary camp has four principal physical components: Efe dwell- material evid
ings, Lese dwellings, drying racks, and fires. All are made on the spot using
local materials. Efe dwellings are virtually identical to those at their
residential camps-a dome-shaped frame of saplings that is covered with
broad leaves. Efe dwelling construction at the specialized sites sometimes
shows less labor investment than at residential camps: the leaf covering over
the sapling frame was incomplete or absent in some cases, and it is my
impression that doorways sometimes were substantially larger than the norm
at residential campsites. The Lese construct a simple lean-to consisting of
several posts supporting a flat or curved roof that slopes downward to the
rear of the structure. On occasion the Lese will use an Efe-style dwelling.
Drying racks are made of poles laid horizontally on a frame and are raised
about 1m above the ground. One or 'more fires are burned under the rack to
dry and smoke meat. The size of the sites ranges, in our sample, from 194m2
to 848m2, with an average of 444m2 (N = 6, s.d. = 235.48 m2).
Butchery typically begins in the morning and is accomplished in three to
five hours. Meat is cut from the carcass and carried to the temporary camp for
drying on racks. The duration of drying ranges, in our experience, from 18 to
68 hours (the maximum time reported here represents an unusual prolonga-
tion while hunters sought another elephant kill). Small quantities of meat are
consumed at the camp. Some bones, after having been stripped of meat and
disarticJiated, are carried to the camp to extract marrow and grease. Bones
are chopped into pieces and boiled in aluminum pots, and the grease rises to
the surface of the water and is scooped off. Long bones are especially desired
for their marrow I grease content. Ribs and vertebrae also are chopped and
boiled for grease and to facilitate meat removal from the vertebrae, which
have complex shapes. The extent to which bones are processed varies consid-
erably: in one case many bones were left unprocessed at the kill site; in an-
other case most bones were processed; and a third was intermediate. The
wry structural number of people present and the nutritional status of the animal probably
are important influences. A general pattern emerges, in which certain skeletal
elements almost never are processed: the cranium, mandible, scapulae, and
ts in the past. innominates. This almost certainly has to do with the relatively low quantity
nal technolo- of marrow available from these bones (Fisher [1992] presents ·a fuller discus-
sion of butchery and processing).
ccasions and
,ccurred from Identifying the Presence of Efe and Lese
han ts usually
st village and Can the presence of both the Efe and the Lese be established from
h the carcass the material record at these specialized sites? Two interrelated lines of evi-
n our experi- dence found at temporary elephant processing sites will be examined: site
:le number of structure (spatial organization) and certain attributes of the bone assemblage.
1 sexes, plus This question will be analyzed in two steps. The first step is to establish that
)f the carcass the specialized sites can be differentiated from the residential campsites of the
1rcass using a Efe and the villages of the Lese. Secondly, having determined that these sites
l(es, hatchets, can be distinguished from residential sites, I will attempt to develop argu-
ments to establish that the presence of Efe and Lese can be discerned from
s: Efe dwell- material evidence.
2561 ]. W. Fisher, h·
Several lines of ·evidence provide strong indications that these specialized

processing sites are not residential sites of either the Efe or the Lese. The
spatial organization of the sites does not conform to that of either group. The
sites clearly are a strong departure from Lese villages with wattle-and-daub
dwellings, barazas, and large, cleared areas. The similarity between elephant
processing sites and Efe campsites is greater, but strong dissimilarities are
evident. One factor is the unplanned, even."incoherent," spatial layout of the
processing sites (Figure 14-4). Dwellings are situated throughout the site
rather than conforming toEfe campsite arrangement along the camp perime-
ter. Moreover, the doors of dwellings do not uniformly face into the center of
the camp. The short duration of occupation (which is anticipated prior to
. camp construction) probably influences the nature of social relations, and
people apparently are more casual in arranging the locations of their
dwellings. In addition, relatively little effort is devoted to clearing away
saplings, underbrush, or other obstructions, and people must accommodate to
their presence when situating their dwelling. The short duration of occupation
is thus expressed by the low investment in structuring the arrangement of
dwellings and in site preparation.
Equally significant are the distribution and· the composition of the bone
assemblage. Discarded bones are distributed both within and outside the
N
"perimeter" of the camp. This departs from Efe practice at residential sites,
where refuse is deposited on trash heaps situated along the camp perimeter.
Efe routinely clean up refuse that accumulates within the circle of huts at
residential camps and redeposit it on trash heaps. The length of site occupa-
r
tion conditions these patterns, with less investment in maintenance at sites
0
briefly occupied. More. careful site maintenance and redeposition of refuse
occurs when the duration of occupation increases (Fisher and Strickland 1989,
1991; Hitchcock 1987; O'Connell1987). Fi~
prij
The bone assemblage at elephant processing sites consists almost exclu-
sively of the bones of one species-elephant. This is a departure ifrom Efe
consumption practices at residential camps, where various species are intro- with unneces~
duced. Moreover, the bone assemblage at elephant processing sites consists of debris generat
a single individual. At Efe residential camps we can expect the accumulation Can we deb
of parts from multiple individuals. It should be noted that animal bones were farmers? The
not abundant in our experience at Efe residential camps. Several factors might from the Lese.
account for this, including low visibility of bones due to generally small size those of the Ef
and reduction during processing and consumption, consumption of bones by 20), whereas E
camp dogs, and scavenging by wild animals after site abandonment (Fisher This discrepa1
and Strickland 1991:223). Long-term foe
Of powerful diagnostic value is the fact that discarded materials at process- large quantitiE
ing sites consist almost exclusively of elephant bone, with few or no other tence ada ptati
types of debris. This is unknown at Efe campsites, where trash heaps contain a large drying 1
variety of discarded items: baskets, potsherds, cloth, bow staves, palm oil butchery I pro
nuts, animal bones, and so forth. The specialized nature of the elephant pro- weighing som
cessing sites and the short occupation are responsible for that pattern. The The larger dr
participants take to these sites only essentials; they do not burden themselves smaller racks<
se specialized
the Lese. The
er group. The
tttle-and-daub
veen elephant
~mil ari ties are
llayout of the
~hout the site
camp perime-
) the center of
Jated prior to
relations, and
:ions of their
Jearing away
commodate to
of occupation
Tangement of
n of the bone
d outside the
:iidential sites,
np perimeter.
·de of huts at
)f site occupa-
~nan ce at sites , bone
o 3m '*f_J!- trash heap
.tion of refuse
rickland 1989,
Figure 14-4. Plan map ofan elephant processing site showing the
almost exclu-
primary structural components and discarded bones.
ture from Efe
des are in tro- with unnecessary paraphernalia. The brief occupation limits the amount of
ltes consists of debris generated. . . .
accumulation Can we determine that these sites had been occupied by both foragers and
.al bones were farmers? The size of drying racks offers a means for distinguishing the Efe
~ factors might
from the Lese. The Lese build drying racks that are more than twice as long as
:tlly small size those of the Efe; The average length of Lese racks is 2.44 m (s.d. = 0.86 m; N =
m of bones by 20), whereas Efe drying racks average 1.12 min length (s.d. = 0.19 m; N = 18).
nment (Fisher This discrepancy is a function of their very different subsistence strategies.
Long-term food storage at ·permanent villages by the Lese can accommodate
als at process- large quantifies of dried meat. Food storage is not a component of the subsis..:
w or no other tence adaptation of the Efe. Consequently, the Efe have no incentive to use
eaps contain a large drying racks. When returning to their residential sites at the end of
Lves, palm oil butchery /processing, the Lese are burdened with large baskets of meat
elephant pro- weighing some 20-30 k, whereas the Efe carry just a few small pieces of meat.
t pattern. The The larger drying racks of the Lese often can be distinguished from the
en themselves smaller racks of the Efe by the presence ·of two closely spaced fires underneath
r·
l '
I
258 f. W. Fisher, Jr.
them or by the presence of one.very long fire that runs practically the entire ence of Efe a
length of the rack. The small 'drying racks of the Efe are ·served by a single fire. of similar p1
Dwelling type also distinguishes. the Efe from the Lese. The ·Efe typically behavior.
construct a dome-shaped dwelling of the kind used in their residential camps, The subsis
whereas the Lese usually construct a lean-to with a ~eaf-covered roof, irfg the arch1
although occasionally they use an Efe-style hut. Dwelling shape, however, the more ge1
would be difficult to ascertain archaeologically. regarding th
A weakly expressed tendency toward physical separation of Efe dwellings tant determil
from those of the Lese is evident. At about 85% o( the sites in our sample, the tities of surF
Efe had situated their dwellings in one or more clusters that were slightly Efehave no:
separated from Lese dwellings. At some sites a portion of Lese dwellings also toward surp
were clustered, but others were situated some distance from their nearest Lese Lese subsistE
neighbor. The distance between adjacent Efe huts within clusters ranges from from domesi
about 0.5 m to 3m, while the distance between adjacent Lese dwellings ranges delayed. foot
from about 1.5 m ·to 12.5 m. The utility of clustering of dwellings rests, how- con sum ptior
ever, on determining the location and shape of dwellings archaeologically, ence to then
which poses operational challenges. Binford 1980
The details of elephant bone-distribution at the temporary processing sites We can e)
is more problematical for identifying the presence of foragers and farmers and storage by ll
for distinguishing the affiliation of a particular dwelling or a subarea within more like th
the site. At one site most discarded bone fragments were located near Lese ward killing
dwellings, while another site showed the opposite pattern. Generally, bones putting ups
are distributed in a manner that precludes associating them firmly, on the conditions, i
basis of spatial proximity, with dwellings_of the Efe or the Lese. The same would diffe
seems to hold true for the distribution of specific skeletal parts. Both the Efe resemble m<
and the Lese process marrow-rich bones (long bones) and marrow-poor bones food in bulk
(ribs and vertebrae). However, the Lese obtain the larger share of bones for The durat:
processing. Indeed, the Efe sometimes chop bones for the Lese to boil. The certain site a
Lese often claim the most productive bones, but it is difficult to "¢ee" this in ery I process:
the subsequent distribution of discarded pieces. · departs fron
tenance (dec
ized, short-if
Conclusions at Efe reside:
The immE
This study has addressed the issue of using material evidence at sped alized. I
specialized elephant butchery I processing sites to establish the presence of segmentsfrc
both foragers and farmers. The underlying determinants that shape the is more easil
evidence at these sites include subsistence-settlement strategy and situational, conditions c
environmental, and social factors. This study extends the theme that archaeol- caribou. Thu
ogists can and should develop methods for identifying forager-farmer inter- The envin
actions. viable mean:
Bone assemblages offer various lines of strong evidence to differentiate an alternativ
elephant processing sites from Efe and Lese residential sites and to. distinguish This stud·
the processing sites as special activity locations. Various attributes of the rna terial pai
spatial organization of the elephant processing sites provide strong indicators produced th
that both foragers and farmers participated in creating these sites. Bone means to op
assemblages do not, however, provide a secure means of identifying the pres- the ground.
ally the entire ence of Efe and Lese. The absence of distinguishing attributes reflects the use
y a single fire. of similar processing techniques by both groups and unstructured discard
~ Efe typically behavior.
iential camps, The subsistence-settlement strategies of the Efe and Lese are key to elevat-
covered roof, ing the archaeological utility of this Efe-Lese case from the particularistic to
:tpe, however, the more general. The contrasting subsistence strategies of the Efe and Lese
regarding the accumulation and storage of surplus food represents an impor-
Efe dwellings tant determinant. The Lese are strongly concerned with acquiring large quan-
ur sample, the tities of surplus meat during elephant butchery and processing, whereas the
were slightly Efe have no interest in putting up large stores of meat. The differing attitudes
dwellings also toward surplus meat are understandable in the broader context of Efe and
ir nearest Lese Lese subsistence adaptations. The Lese, whose subsistence derives primarily
~s ranges from from domestic plants cultivated on a seasonal cycle, incorporate a strategy of
rellings ranges delayed food returns and storage of food. The day-to-day acquisition and
.gs rests, how- consumption of food by the Efe, on the other hand, is explainable with refer-
haeologicall y, ence to the nature of wild food availability in tropical forest environments (see
Binford 1980; Kelly 1983).
rocessing sites We can expect variation as a function of increasing dependence on food
.d farmers and storage by hunter-gatherer groups. Collectors might be expected to behave
ubarea within more like/ the Lese than the Efe. Nunamiut Eskimo subsistence is geared to-
Lted near Lese ward killing large numbers of caribou during brief periods of availability and
nerally, bones putting up stores for consumption at other times (Binford 1978). Under those
firmly, on the conditions, the material record of large mammal butchery and processing
ese. The same would differ from that produc~ by foragers such as the Efe and might
;. Both the Efe resemble more closely that of farmers who are concerned with processing
>w-poor bones food in bulk for storage.
·e of bones for The duration of site occupation stands out as an important determinant of
;e to boil. The certain site attributes. The casual spatial organization of the specialized butch-
0 usee" this in ery /processing sites, particularly the location and orientation of dwellings,
departs from the structure and regularity at Efe residential camps. Site main-
tenance (cleanup, keeping debris out of the way) is desultory at the special-
ized, 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
al evidence at specialized processing sites. Smaller mammals can be transported whole or in
te presence of segments from kill site to residential site, but intensive processing of elephants
:tat shape the is more easily accomplished by moving the people to the carcass. Analogous
nd situational, conditions can arise if hunters slay multiple individuals of, say, bison or
that archaeol- caribou. Thus, the Ituri sites are relevant for other settings.
~-farmer inter- 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
) differentiate an alternative option in suitable environments.
to distinguish This study has been concerned with attempting to link expectations for
ributes of the material patterning with the underlying determinants of the behavior that
ong indicators produced the material record. It has devoted less attention to developing the
3e sites. Bone means to operationalize archaeological.criteria for identifying the patterns in
fying the pres- 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 Bunn, H. T., I
dwellings archaeologically? Reconstructing the location of dwellings at Efe 1988 Var
residential campsites has been ad9.ressed by Fisher and Strickland (1991). ing, and
Further work of this nature is required.. · Ewers,J. C.
The vagaries of preservation of material evidence in, the archaeological '1954 ThE
record presents a related issue. Almost all classes of evidence at the processing pretatim
sites are organic and vulnerable to decay and obliteration. Preservation is Fisher, J. W.,:
contingent on regional, local, and microenvironmental conditions. The red- 1992 Obt
Spring~
dened, oxidized·earth left by past fires probably has the greatest potential for
andJ. D
detection. Elaboration on the kinds of evidence that might or might not be of Color,
preserved archaeologically will not be pursued here. Preservation obviously is Fisher, J. W., .
crucial to archaeology, but I am more concerned in this presentation with 1989 Eth
learning to understand the material record to monitor human behavior. The Campsit
constellation of multiple lines of evidence provides a more powerful indicator 1991 Dv.
. of forager-farmer interactions than does any individual line of evidence in Ethnoarc
isolation. Case Stu
national
Gargett, R., a
Acknowledgments 1991 SitE
for Arcl
Helen Strickland's participation in the fieldwork reported here was byE.M.
vital to its success, and I am very grateful for her contributions. This fieldwork Hart, T. B., m
was funded by a grant from the National Science Foundation to Irven DeVore 1986 Thi
and Peter Ellison. I am very grateful to the Efe and Lese, among whom we Forests:
lived and worked, for their cooperation. I thank Jean Hudson for inviting me Headland, T.
1989 Hu
to participate in the From Bones to Behavior conference and for helpful
Current
comments on an earlier draft of this paper. Hitchcock, R
1987 Sec
Residen
References archaeol1
Press, I\
Bailey, R. C., G. Head, M. Jenike, B. Owen, R. Rechtman, and E. Zechenter
Kelly, R. L.
1989 Hunting and Gathering in Tropical Rain Forest: Is It Possible? American
1983 H1:
Anthropologist 91:59-82.
39:277-~
Bailey, R. C., and I. DeVore
Milisauskas,
1989 Research on the Efe and Lese Populations of the Ituri Forest, Zaire. Ameri-
1978 Eu
can Journal of Physical Anthropology 78:459-471.
O'Connell, J.
Binford, L. R.
1987 Al
Antiquit
O'Connell, J.
Archaeological Site Formation. American Antiquity 45:4-20.
1988 Ha
1987 Researching Ambiguity: Frames of Reference and Site Structure. In.Method
Implica
and Theory for Activity Area Research: An Ethnoarchaeological Approach, edited by S.
1991. Di:
Kent, pp. 449-512. Columbia University Press, New York.
Camps:
Brooks, A. S., D. E. Gelburd, and J. E. Yellen
pretati01
1984 Food Production and Cuiture Change among the !Kung San: Implications
pp. 61-:
for Prehistoric Research. In From Hunters to Farmers: The Causes and Consequences
Smith, B. D.
of Food Production in Africa, edited by J.D. Clark and S. A. Brandt, pp. 293-310.
1989 Or
University of California Press, Berkeley.
Foragers and Farmers 1261
the location of Bunn, H. T., L. E. Bartram, and E. M. Kroll
rellings at Efe 1988 Variability in Bone Assemblage Formation from Badza Hunting, Scaveng-
:kland (1991). ing, and Carcass Processing. Journal of Anthropological Archaeology 7:412-457.
Ewers, J. C.
Lrchaeological 1954 The Indian Trade of the Upper Missouri Before Lewis and Clark: An Inter-
the processing pretation. Missouri Historical Society Bulletin 10:429-446. ·
reservation is Fisher, J. W., Jr.
lons.· The red- 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
:t potential for
and J. Day, pp. 51-81. Denver Museum of Natural History and University Press
might not be
of Colorado, Niwot.
n obviously is Fisher, J. W., Jr., and H. C. Strickland
entation with 1989 Ethnoarchaeology among the Ere Pygmies, Zaire: Spatial Organization of
behavior. The Campsites. American Journal of Physical Anthropology 78:473-484.
erful indicator 1991 Dwellings and Fireplaces: Keys to Efe Pygmy Campsite Structure. In
>f evidence in Ethnoarchaeological Approaches to Mobile Campsites: Hunter-Gatherer and Pastoralist
Case Studies, edited by C. S. Gamble and W. A. Boismier, pp. 215-236. Inter-
national 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
)rted here was by E./M. Kroll and T. D. Price, pp. 11-32. Plenum Press, New York.
rhis fieldwork Hart, T. B., and J. A. Hart
Irven DeVore 1986 The Ecological Basis of Hunter-Gatherer Subsistence in African Rain
mg whom we Forests: The Mbuti of Eastern Zaire. Human Ecology 14:29-55.
Jr inviting me Headland, T. N., and L.A. Reid
1989 Hunter-Gatherers and Their Neighbors from Prehistory to the Present .
.d for helpful
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 Ethno-
archaeological Approach, edited by S. Kent, pp. 374-423. Columbia University
Press, New York.
nter
K~lly, R. L.
sible? American
1983 Hunter-Gatherer Mobility Strategies. Journal of Anthropological Research
39:277-306.
:t, Zaire. Ameri- Milisauskas, S.
1978 European Prehistory. Academic Press, New York.
O'Connell, J. F.
1987 Alyawara Site Structure and Its Archaeological Implications. American
nt Systems and Antiquity 52:74-108.
O'Connell, J. F., K. Hawkes, and N. Blurton Jones
ture. In Method 1988 Badza Hunting, Butchering, and Bone Transport and Their Archaeological
Implications. Journal of Anthropological Research 44:113-161. ·
zch, edited by S.
1991 Distribution of Refuse-Producing Activities at Badza Residential Base
Camps: Implications for Analyses of Archaeological Site Structure. In The Inter-
.n: Implications
pretation of A~chaeological Spatial Patterning, edited by E. M. Kroll and T. D. Price,
pp. 61-76. Plenum Press, New York.
rzd Consequences
it, pp. 293-310. 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. 15. I
Spielmann, K. A. (editor)
1991b Farmers, Hunters, and Colonists: Interaction Between the So.uthwest and the
Southern Plains. University of Arizona Press, Tuc$on. B
Wilmsen, E. N., and J. R. Denbow
1990 Paradigmatic History of San-speaking Peoples and Current Attempts at
Revision. Current Anthropology 31:489-524. It
Wood, W.R.
1974 Northern Plains Village Cultures: Internal Stability and External Relation- Fi
ships. Journal of Anthropological Research 30:1-16. archaeologicc
Yellen, J. E. 20 years. My
1976 Settlement Patterns of the !Kung: An Archaeological Perspective.· In biased by m:
Kalahari Hunter-Gatherers: Studies of the !Kung San and Their Neighbors, edited by Indian sites i1
R. B. Lee and I. DeVore, pp. 47-72. Harvard University Press, Cambridge. 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
15. Discussion: Social Interaction
1west and the
Bonnie W. Styles
Attempts at
Introduction
nal Relation- First, I would like to confess that I have never done any ethno-
archaeological research, but I have been doing zooarchaeological research for
20 years. My interests are in paleoecology and subsistence, and my views are
·spective. In biased by my experiences with faunal remains from prehistoric American
'rs, edited by Indian sites in the Midwestern United States. I have certainly been inspired by
ridge. 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 aban-
don "grab-bag analogy from the ethnographic present" and collect the infor-
mation 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 experimen-
tal 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 archaeo-
logical record" in mind: that is, with a goal of understanding "the determi-
nants 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 assem-
blages. 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 Trustee~, Southern Illinois University. All rights
reserved. ISBN 0-88104-076-2. ·
263
2641 B. W. Styles
followed ethnoarchaeological research, although perhaps not as closely as I from roastin

should have. Howev.er, I can guarantee that I have dohe a lot of reading since Along the
Jean Hudson called me to participate in the conference. missing fror
I will focus my comments on the.papers and major topics for this session, culinary enc
that is, the recognition of ethnic differences, food sharing, and the impacts of ture of faun;
food preparation and consumption on the record. I will attempt to synthesize ern Kenya,
the major points raised in the four papers in Part III and examine their appli- dismemberr
cability to archaeological research. jagged, trar
mately link
Bone 1\tlodification, Cooking Techniques, because th(
uncooked bl
and Consumption that butche:
The papers by Diane Gifford-Gonzalez and J. S. Oliver note the about how t
role that cooking practices play in the patterning of faunal assemblages. be stored m
Oliver analyzes breakage patterns associated with butchering, preparation for available at
cooking, and consumption. Based on his study of the Hadza, he notes that the products· wi
decision whether or not to break a bone for cooking and the tool used for this butchering;:
"pot- or hearth-sizing" appeared to be primarily determined by carcass size Accordin~
and element type. Pot-sizing breakage of axial elements increased with body reflects an E
size. Pot-sizing of limb elements was mitigated by the amount of marrow and meat is like]
cancellous tissue and whether marrow was to beeaten raw or after boiling. tured raw, i1
He argues that ''behaviors-from initial field butchery, to transport, to the use off
cooking-preparation, to consumption-may be integrated into a decision- bias" within
making strategy where nutrient extraction is maximized, while transport, field and butcher
processing, and social costs are minimized." tions. At m<
Oliver argues that the breakage and other bone modifications related to animals of;
butchery and preparation for cooking and consumption, such as for the pot numbers, ar
sizing and marrow extraction that he observed, would have archaeological ed through:
signatures. The short period of boiling used by the Hadza left no visible dam- on the time
age and did not produce a diagnostic breakage pattern. However, 1roasting, as kill sites do
also observed by Gifford-Gonzalez, left breakage patterns (jagged edges) that the search,
differed from those noted for fresh bone. Marrow and cancellous tissue interpretati<
removal also created diagnostic damages. In addition, he argues that detailed logical rec01
analyses of.bone modifications are essential to the interpretation of patterns in that considE
the frequencies of bone elements because these frequencies can be the result of storage, coo:
a number of different processes (e.g., transport by humans and/ or destruction Gifford-G
of low density bones). I agree wholeheartedly with his assertion that more analytical r
experimental research is needed to identify cooking-related damages because ·Cooking pr.
I am interested in the preparation of meat and how food preparation changed evolution c
through time. Although post-consumption processes, such as disposal, scav- addressed i
enging by dogs, and trampling by humans, may obscure earlier n1odifications butchering c
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 1
elements: One productive line of research for Midwestern zooarchaeologists sharing pla:
would be to examine changes in food preparation related to the transition notes that '
Social Interaction I 265
.s closely as I from roasting to stone boiling and eventually to boiling in ceramic vessels.
reading since Along these same lines, Gifford-Gonzalez convincingly argues that "what is
missing from recent ethnoarchaeological analyses is serious treatment of the
r this session, culinary end of the processing spectrum and its major influence ori the struc-
he impacts of ture of faunal assemblages." Based on her studies of the Dassanetch of north-
to synthesize ern Kenya, she found that culinary tactics were crucial to decisions about
1e their appli- dismemberment, selective transport, and bone breakage. High frequencies of
jagged, transverse fractures on bones from a Dassanetch camp were ulti-
mately 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
liver note the about how the meat, marrow, and other useful parts of the dead animals will
assemblages. be stored or processed for consumption." Thus, the "processing technology
reparation for available at the destination site" and the "ultimate form or forms the animal
notes that the products will take" should be added to the long list of factors affecting field
l used for this butchering and transport.
y carcass size According to Gifford-Gonzalez, the research emphasis on uncooked bone
ed with body reflects an early focus on 1nass kill assemblages and mass butchering where
fmarrow and meat is likely to be removed prior to cooking and bones are likely to be frac-
after boiling. tured ra/v, if at all, and a focus on early hominid sites that presumably predate
transport, to the use of fire. She also attributes the pattern to "unconscious androcentric
u a decision- bias" within the field of anthropology, which favored studies of male hunting
~ansport, field and butchery of large animals as central to the understanding of past adapta-
tions. At many Holocene archaeological sites, such as the ones I deal with,
ms related to animals of all sizes entered the site and were processed in relatively small
as for the pot numbers, and the meat, the bones and the meat, and/ or the bones were cook-
.rchaeological ed through roasting, stone boiling, and/ or boiling in ceramic pots depending
J visible dam- on the time period under investigation. Clearly, early hm.ninid sites and mass
~r, roasting, as kill sites do not provide the best analogues for many sites, and emphasis on
~d edges) that the search, pursuit, and processing of large game will not lead to holistic
:ellous tissue interpretations of subsistence practices or of the formation of the archaeo-
; that detailed logical record. Gifford-Gonzalez calls for a product-focused research strategy
of patterns in that considers all the tasks and factors that affect faunal remains, including
1e the result of storage, cooking, and refuse disposal.
::>r destruction Gifford-Gonzalez offers more than a cautionary tale because she shows how
~on that more analytical models can influence ethnographic research and observations.
1ages because Cooking practices, although clearly important to our understanding of the
ition. changed evolution of human subsistence and nutrition, have been inadequately
lisposal, scav- addressed in many studies because of an emphasis on search, pursuit, and
modifications butchering of large animals.
Juld certainly
breakage and Foo~ Sharing
patterns is no
ing of faunal The paper by Fiona Marshall emphasizes the important role food
trchaeologists sharing plays in the dispersal of bones within and between sites. Marshall
the transition 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 netw1
sharing." Yet Marshall found food sharing to be the the single major factor the other facl
patterning early stages of the formation of assemblages on Okiek sites in the be :supportec
high altitude rain forests of Kenya. "In 12 out of 15 observed cases" animals animals am01
were shared between 2 and 7 households separated by several kilometers." tial transport
Secondary and tertiary distribution of meat often followed primary distribu- ered, as are t
tion. The factors governing food sharing included hunting success, social attention nee
relationships, method of capture, size of the animal, and the context of con- between site1
sumption. There was a lot of variability between households in species com- cause of pattE
position, 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 E1
variability in assemblage composition and spatial distributiOI\ of taxa between
households within and between sites, presence of complementary element 0:
profiles at differenfhouseholds or different sites, bones from different house- ences. He de
holds or camps that could be refitted, and paired bones from individual groups at sitE
animals at different households or sites as recognized through osteometric Lese horticul
analyses. She also points out that although many of the differences between processing si
assemblages were due to differences in hunting skills and social relationships the presence
they might have been erroneously attributed to differential transport of car- each other et:
cass parts or site fun~tion. , at cooperath
She rightfully notes that "recognition of food sharing in archaeological sites include Efe c
will depend on exceptional conditions of preservation." Although Marshall is that that site ·
optimistic that sharing can be detected in the archaeological record, I have my the spatialla~
doubts that this will be easy to accomplish. Many zooarchaeologists assume tion, the num
that food sharing or at least some form of meat distribution was a major dis- and the exter
persant of bones across sites. However, it is rarely possible to attribute faunal and element
remains, especially in secondary refuse deposits, to particular households, group; thus, 1
much less to link elements from a single animal across a site or bet~een sites. ethnic group
The effects of post-consumption processes, for example, refuse/ disposal, their villageE
burning, scavenging by dogs, and weathering, serve to disperse and damage carcass-procE
bones, making refitting of breaks or the recognition of the bones from an be used to re(
individual animal difficult. The relatively gross temporal resolution of the ties of dried
archaeological record compounds the problem. sites. Thus, tl
In highly structured archaeological sites, for example, late prehistoric vil- of the struct,
lages where house basins, pits, and refuse deposits can be associated, it may sometimes d]
be possible to reconstruct food-sharing units. However, even for highly the basis of tl
structured sites where refuse can be tentatively associated with households, Attempts 1
potential evidence cited by Marshall cannot be exclusively linked to food follow a coar.
sharing. As she observes for the Okiek, "some of the more typical body part is, the struct
profiles resulting from sharing are similar to those resulting from the logistics differences in
of carcass transport, scavenging, and the destruction of bone during site data. Howev1
formation processes." As noted by Marshall, complementary body part pro- presence of c
files at two camps could reflect e:ither sharing or differences in settlement interaction w
function. Bone fragments that can be refitted suggest potential contempo- would undo1
raneity of deposits, but the agent of dispersal is not necessarily sharing. It will discrete grou:
be difficult to find unequivocal evidence for the reconstruction of food- disperse bonE
j
:1ce for food sharing networks in archaeological contexts. However, if we can control for all
major factor the other factors cited above, then perhaps the food-sharing explanation can
k sites in the be supported. As Marshall argues, given "the ubiquity of sharing of large
tses, animals animals among contemporary hunter-gatherers, ... the possibility of differen-
kilometers." tial transport of elements through food sharing should be routinely consid-
ary distribu- ered, as are transport based on anatomical·utility and transport costs." More
ccess, social attention needs to be paid to patterning in species and body parts within and
ntext of con- between sites and food-sharing needs to be considered as a potential major
species com- cause of patterning in the record.
f the sharing
tg would be Ethnic Differentiation -

:axa between
tary element Only the paper by John Fisher attempts to document ethnic differ-
=eren t house- ences. He develops criteria for recognizing the presence of different ethnic
1 individual groups at site, in this case for seminomadic Efe Pygmy foragers arid sedentary
osteometric Lese horticulturalists in the context of cooperative elephant butchering and
.ces between processing sites. He addresses the basic question of whether we can identify
relationships the presence of foragers and horticulturalists who are widely different from
sport of car- each other, ethnically, culturally, and socially on the basis of material remains
at cooper.~tive carcass-processing sites. Those short-term processing camps
ological sites include Efe dwellings, Lese dwellings, drying racks, and fires. He concludes
l1 Marshall is that that site type can be differentiated from residential sites of both groups in
·d, I have my the spatial layout, the disposal areas for discarded bones, the species composi-
gists assume tion, the number of individual animals, the paucity of debris other than bones,
a major dis- and the extensive breakage of the bones. However, the distribution of bones
ribute faunal and elements precludes their association with a single dwelling or ethnic
households, group; thus, the faunal data do not contribute to the differentiation of the two
etween sites. ethnic groups. The horticultura1 Lese store large quantiHes of dried meat at
.se disposal, their villages and consequently build large drying racks at the cooperative
and damage carcass-processing sites. The remains of the racks were detectable and could
nes from an be Used to recognize their presence. The Efe foragers do not store large quanti-
1
.ution of the ties 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
ehistoric vil- of the structural features, for example, spatial organization of houses and
lated, it may sometimes different house forms and different-sized drying racks, but not on
1 for highly the basis of the distribution of bones or body parts.
households, Attempts to recognize different groups at archaeological sites generally
ked to food follow a coarse-grained approach and at le,ast initially use the other data, that
al body part is, the structural features and the ·artifacts rather than faunal data. Ethnic
the logistics differences in prehistory are generally reconstructed from these other types of.
during site data. However, subsistence data may play an important role. For example, the
dy part pro- presence of cultigens would certainly argue for the presence of, or at least
:1 settlement interaction with, ;horticulturalists. Archaeologists dealing with older records
1 contempo- would undoubtedly face serious problems in linking faunal remains with
aring. It will discrete groups at short-term camps bec~use of the wide variety of agents that
.on of food- disperse bones .
.
1-
2681 B. W. Styles
For archaeologieal issues, a comparison of faunal records from residential Although:

camps associated with the foragers and farmers would also provide an inter- tiona! or inci<
esting case study for comparing fau11al signatures at sites associated with and weatheri
foragers and farmers. However, these specialized cooperative sites do provide sumption, I ·
a good opportunity to look at (1) the evidence for forager-farmer interactions tribute to ou
and (2) the underlying differences in subsistence-settlement strategies that patterning iJ
ultimately shape the material record. ethnoarchaec
pdst-transpOI
see at many'
Conclusions uct we see at
to place morE
These studies all provide potential insights into archaeological ravages of til
faunal assemblages. However, the degree to which it is possiple to deal with mental conte:
these "social issues" in the prehistoric archaeological record varies. Although cooking,. and
Fisher could not discriminate the presence of Efe foragers and Lese horticul- understand t
turalists 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
I us to worry a
means of historic analogues, faunal remains from historic archaeological sites l R

in the eastern United States have been used to interpret socioeconomic status
mediated by settlement function (see Reitz and Zierden 1991) based on the
l
:! Binford, L. R.
:t
presence of particular meat cuts. Such techniques have been less successful for 1967 Smu
prehistoric sites where it would be difficult to rank the value of cuts other lI Reasonin;
than on the basis of food utility. Disproportionately high quantities of high 1981 Bone
food utility cuts of deer meat have been recorded for prehistoric ceremonial :1
1984 Butc
sites in the eastern United States, and it should be noted that in at least one logical Arc
Brain, C. K.
case the body part representation does not correlate with bone density (e.g.,
Styles and Purdue 1991). j 1967 BonE
of Science
Although I think there is little to no doubt that food sharing ~ccurred 1969 The
throughout prehistory, I am pessimistic about unequivocal demonst,tations of AustralOJ
food-sharing or the reconstruction of food-sharing networks, given the rela- Research t
tively coarse temporal resolution of the archaeological record and the diffi- Gould, R. A.
culty of linking food refuse with particular household units at many sites. 1969 Subf
Complementary body part profiles at two sites could reflect settlement func- Oceania 31
tion, for example, transport of parts from a hunting camp to a residential base, 1971 The
provisioning of a residential base by a logistical camp, or transport of high American
food utility parts from a village to a ritual site for ceremony or tribute. Isaac, G. L.
However, Marshall is correct in her assertion that food sharing, which is 1967 Tow
ubiquitous in the ethnographic record, should be explicitly considered as a Observat:
possible explanation for patterns in bone distributions within and between Lee, R. B., and
1968 Man
archaeological sites. If sharing followed set rules and was patterned by hunt-
Reitz, E. L anc
ing success and social relationships as noted for the Okiek, and if the rules 1991 · Catt
and these relationships endured over a long enough period of time, and if Bobwhites
bones were preserved and could be associated with particular households or Purdue,'
social subsets within a site, then these patterns might be evident in the Scientific
archaeological record. More detailed studies of within-site faunal distributions Styles, B. W., c
are needed to document potential evidence for food sharing. 1991 Ritu
1m residential Although post-consumption processes, such as garbage disposal, inten-
vide an inter- tional or incidental burning, scavenging by carnivores, trampling by humans,
3ociated with and weathering, serve to obscure evidences of butchering, cooking, and con-
es do provide sumption, I think more systematic studies of bone modifications can con-
~r interactions tribute to our understanding of the agents and processes responsible for
trategies that 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 prod-
uct 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
rchaeological ravages of time affect the evidence of the earlier, more interesting environ-
~ to deal with mental contexts and human activities such as hunting, transport, butchering,
·ies. Although cooking, and consumption. You do not have to love garbage, but it helps to
Lese horticul- understand the agents and processes that made it, modified it, and left it for
~emains alone us to worry about.
:~.1 remains do
rentiation. By
~ological sites References
Jnomic status
based on the Binford, L. R.
successful for 1967 Smudge Pits and Hide Smoking: The Use of Analogy in Archaeological
of cuts other Reasoning. American Antiquity 32(1):1-12.
1tities of high 1981 Bones: Ancient Men and Modern Myths. Academic Press, New York.
·ic ceremonial 1984 Butchering, Sharing, and the Archaeological Record. Journal of Anthropo-
.n at least one logical Archaeology 3:235-257._
Brain, C. K.
density (e.g.,
1967 Bone Weathering and Problem of Bone Pseudo-Tools. South African Journal
of Science 63(3):97-99.
·ing occurred 1969 The Contribution of Namib Desert Hottentots to an Understanding of
onstrations of Australopithecine Bone Accumulations. Scientific Papers of the Namib Desert
;iven the rela- Research Station 39:13-22.
and the diffi- Gould, R. A.
Lt many sites. 1969 Subsistence Behavior among the Western Desert Aborigines of Australia.
:tlement func- Oceania 39(4):253-274.
'idential base, 1971 The Lithic Assemblage of the Western Desert Aborigines of Australia.
.sport of high American Antiquity 36(2):149-169. ·
1y or tribute. Isaac, G. L.
ing, which is 1967 Towards the Interpretation of Occupation Debris: Some Experiments and·
nsidered as a Observations. The Kroeber Anthropological Society Papers 37.
Lee, R. B., and I. DeVore (editors)
and between
1968 Man the Hunter. Aldine, Chicago.
rned by hunt- Reitz, E. J., and M.A. Zierden
td if the rules 1991 Cattle Bones and Status from Charleston, South Carolina. In Beamers,
f time, and if Bobwhites, and Blue-points: Tributes to the Career of Paul W. Parmalee, edited by J. R.
1ouseholds or Purdue, W..E. Klippel, and B. W. Styles, pp. 394-405. Illinois State Museum
vident in the Scientific Papers 23. Springfield..
l distributions Styles, B. W., and J. R. Purdue .
1991 Ritual and Secular Use of Fauna by Middle Woodland Peoples of Western
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. IV.
Harvard University Press, Cambridge.
1977 Archaeological Approaches to the Present: Models for Reconstructing the Past.
-Il
I
I-
·eer of Paul W.
., pp. 420-436.
Unpublished
:tl Perspective.
re, pp. 47-72.
IV. Noncultural Processes:
cting the Past.
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 carni-
vores 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 dual-
patterned 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 zooarchaeolog-
ical assemblages. His work, along with Isaac's (1967) lesser-known experimen-
tation, 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 assem-
blages. 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

reserved. ISBN 0-88104-076-2. .
273
274 I R. ]. Blumens~hine and C. W. Marean
proclaim hominids as the dominant agent of bone assemblage formation. The produced b~
research effort has produced a detailed understanding of the composition and colleagues ("
condition of bone assemblages cre~ted by either carnivores or humans of whole ar
(Andrews and Cook 1985; Binford 1978, 1981, 1984; Binford and Bertram 1977; spotted hyeJ
Blumenschine 1988; Blumenschine and Selvaggio 1991; Bonnichsen 1973, 1983; and Marear
Brain 1967, 1969, 1981; Bunn 1981, 1983, 1990; Bunn et al. 1988; Capaldo 1990; include axil
Casteel1971; Crader 1983; Gifford 1980; Gifford-Gonzalez 1990; Haynes 1980, hyenas, as l
1982a, 1982b, 1983; Henschel et al. 1979; Hill1979a, 1979b, 1983, 1990; Hill and Selvaggio (J
Behrensmeyer 1984; Hughes 1954; Johnson 1985; Ke~t 1981; Klein 1975, 1980; tooth-marki
Lyon 1970; Maguire et al. 1980; Marean and Spencer 1991; Mareah et al. 1992; stone tools.
Morlan 1984; Myers et al. 1980; Noe-Nygaard 1977, 1989; O'Connell et al. 1988, archaeologit
1989; Potts and Shipman 1981; Read-Martin and Read 1975; Richardson 1980; strably more
Sadek-Kooros 1972, 1975; Scott and Klein 1981; Selvaggio 1990; Shipman and The purpc
Phillips 1976; Shipman and Phillips-Conroy 1977; Shipman and Rose 1983; published s~
Simons 1966; Skinner et al. 1980; Sutcliffe 1970; White 1952; Yellen 1977). can affect th
Archaeologists have met the "carnivore challenge" by applying those patterns of c
models of single-agent bone accumulations to fossil assemblages. Yet many of ary access t<
the Pleistocene archaeological bone assemblages that have been used to and distribu
interpret hominid behavior preserve clear evidence of udual-patterning" sity and lo1
(Capaldo 1990) by carnivores and hominids. This is seen most conspicuously produced b:
in the presence of both tool marks and carnivore tooth marks on bones from size, nutritic
the same assemblage and, occasionally, on the same bone specimen. of assembla!
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 I
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 j
portions. The premise is justified by several studies, subsumed under the ened menu 1
rubric of the hunting-scavenging debate. They have begun to shpw how on a whole
carnivores, in reducing feeding opportunities for hominids on a cavcass, will hominids, t
alter the condition and range of parts available for hominid butchery and refuse woul<
transport (e.g., Binford 1984; Blumenschine 1986, 1991; Potts 1983; see also axial bones.
Bunn 1986; Gifford-Gonzalez 1990). to includes
Relatively little experimental attention has been given to carnivores cranial proc
scavenging from hominid-butchered assemblages. This is ironic, given both posed marrc
the contention and importance surrounding the sequence and degree of carni- animals whE
vore and hominid involvement with Pleistocene archaeological bone assem- remaining f•
blages (e.g., Binford 1986, 1988; Bunn and Kroll 1986, 1988) and the fact that bone-crushil
Brain's and Isaac's work that posed the "carnivore challenge" was, purposely hominid-bu
or not, modeling a hominid-followed-by-carnivore sequence of taphonomic attractive to
agents. Crader (1983) showed the effects of carnivores scavenging large A carnivc
mammals butchered by the Valley Bisa in an attempt to assess the validity of butcher.ed a~
Pleistocene "butchery" sites. Blumenschine (1988) and Blumenschine and carcass. Can
Selvaggio (1991) showed how the epiphyseal:shaft fragment ratios and decreasing r
segmental distribution of tooth-marking and percussion-marking on long defleshed b)
bone assemblages experimentally broken with a hammerstone and subse- order nutriti
quently scavenged by spotted hyenas (Crocuta crocuta) differed from those nutritional '
I
.J
r
A Carnivore's View I 275
·mation. The produced by carnivores alone and hammerstone breakage alone. Binford and
position and colleagues (1988) similarly experimented with the dispersal and modification
or humans of whole and hammerstone-broken buffalo (Syncerus caffer) long bones by
ertram 1977; spotted hyenas. More recently, Marean (1991), Marean and colleagues (1992),
n 1973, 1983; and Marean and Spencer (1991) have expanded these later experiments to
:tpaldo 1990; include axial and compact (tarsals, carpals) bones using captive spotted
!aynes 1980, hyenas, as has Capaldo (1990), using a variety of free-ranging carnivores.
190; Hill and Selvaggio (1990) also examines the frequency and anatomical patterning of
1 1975, 1980; tooth-marking on bones defleshed and fragmented by carnivores and/ or
n et al. 1992; stone tools. All of these studies aim to improve the realism of models for
:11 et al. 1988, archaeological bone assemblages whose predepositional history is demon-
1rdson 1980; strably more complex than that described by single-agent taphonomic studies.
hiprnan and The purpose of this paper is to integrate and expand upon our previously
l Rose 1983; published studies that demonstrate the manner in which hominid butchery
i977). can affect the nutritional attractiven~ss of parts to carnivores, thus influencing
>lying those patterns of damage and deletion that are unique to carnivores having second-
Yet many of ary access to bones. Specifically, we will illustrate how the different amount
~en used to and distribution of food on a hominid-butchered assemblage affects the inten-
·patterning" sity and location of damage and the skeletal part and portion profiles
mspicuously produced by carnivore ravaging. We will also explore the effect of carcass
l bones from size, nutritional quality, and competition among carnivores on these measures
:n. of assemblage disturbance.
1ot sensitive
)f carnivores
vi ties of one Premise
d, hence, the
es and bone A hominid-butchered bone assemblage offers a carnivore a short-
d under the ened menu of parts and a reduced nutrient yield compared to that available
>show how on a whole carcass. Prior to· the advent of grease-rendering technology by
carcass, will hominids, the skimpiest Plio-Pleistocene menu offered by hominid food
utchery and refuse would have featured mainly grease from defleshed limb epiphyses and
183; see also axial bones. If butchery was incomplete, the menu may have been expanded
to include some flesh, especially the near-bone flesh around vertebral and
> carnivores cranial processes that is hard to remove with a stone tool, and some unex-
·, given both posed marrow cavities from long bones and axial parts. Except for very large
;ree of carni- animals where prior hominid feeding might not have depleted all flesh, the
bone assem- remaining fare· would be marginal and would entice only a carnivore with
the fact that bone-crushing abilities (mainly hyaenids and canids). Despite the low yield of
lS, purposely hominid-butchered fare, our studies demonstrate that it is still extremely
taphonomic attractive to scavenging carnivores (contra Binford 1981:286). ·
~nging large A carnivore's usual selection of parts for consumption from a hominid-
le validity of butchered assemblage should also differ from its preferred choices on a whole
nschine and carcass. Carnivores can be expected to select parts in a sequence that follows
t ratios and decreasing nutr~ent yield. This is true for the sequence by which bones are
:ing on long defleshed by a variety of carnivores (Blumenschine 1986). However, the rank
! and subse- order nutritional value of bones on a whole carcass differs from the rank order
l from those 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 mar-
row 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 hominid-
butchered 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 experi-
mental 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 con-
ducted, 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
A Carnivore's View 277 I
Jtili ty Index, Standardized Grease. Index Modified General Utility Index
)nes / reveals D FEMUR \ 1§11
(Figure 16-1; P FEMUR
P TIBIA
STERNUM
~-~
RIB
erdelin 1989) PELVIS
lage to differ D TIBIA

THORACIC
ecialist felid. SCAPULA
not so much P HUMERUS I

D HUMERUS
r diaphyses. CERVICAL
~-?0?~;~~
LUMBAR
te incidental ASTRAGALUS '=
of defleshed CALCANEUM I
TARSALS
se most mar- P METATARSAL
P RADIUS-ULNA
f long bones, D METATARSAL
~s a nutrient D RADIUS-ULNA -~ .~~

CARPALS w~~~~~J!~fiJ
.yses-that a MANDl B LE % .. -
2D PHALANX
~equen tl y on 3D PHALANX
1ST PHALANX
P METACARPAL ~~ ~4fi2~~.;i@'*tNflJJJS~~;;'I,J
tat still offer D METACARPAL
~ ~
AXIS
:oun tered by ATLAS
I I T I I I I I I
y of skeletal
30 25 20 15 10 5 0 5 10 15 20 25 30
.si ty of bone
. a hominid-
Is played no 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. ~
lsturbance to
~ts of experi- usually disarticulated after total defleshing. Carpals and tarsals were usually
tschine 1988; disarticulated as a unit and left attached by the posterior tendon bundle of a
Marean and ~etapodial to the phalanges, which were not broken. All long bones were
1ew analyses broken, using a hammerstone-on-anvil technique, and to an extent sufficient
perimen ts as to remove all marrow with the aid of a thin, spatulate scoop. To the extent
ere designed possible, like bones were broken in a similar manner by placing hammerstone
~d-butchered blows at the same locations.
D continue to On the day of breakage, all fragments produced by hammerstone breakage
te-only" and were placed in simulated sites in a variety of settings so they could be dis-
turbed freely by bone-crushing carnivores. Most were placed in Acacia wood-
he Serengeti land or plains settings. Two that were placed in riparian woodlands were not
8; see also disturbed after as many as 13 days; they were recovered .intact and are here
s were con- included in the hammerstone-only sample described below.
)ne forelimb, Most simulated site bones derived from animals in good condition prior to
3on' s gazelle death. The marr?w (and grease) in these bones was fat-rich, as determined by
la (Aepyceros visual inspection (Sinclair and Duncan 1972) and additionally by oven-drying
i (Damaliscus experiments. Two simulated sites were composed of bones from a wildebeest
~ments were (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 possible ow
hyenas live (
simulation I
Size 1 & 2 Size3 &4 TOTAL
·N Assem NISP N Assein NISP
simulated s
Site Type N Assem NISP
collected aft.
came to lean
Simulated sites While the
Serengetia data analysi
Fat-rich 6 287 4 267 10 554 marking anc
Fat-depleted 0 2 44 2 44
Berkeley 1b 98 424 sites (refern
9 326 4 13
Berkeley 2c,d 33 1162 0 33 1162 Serengeti sat
also summa
Carnivore onlya 3 91 6 140 9 231 previously 1:
for 33 assen
Hammerstone onlya 6 303 1 24 7 327 overlap wit:
Marean and
TOTALd 57 2169 15 573 74 2742 includes 21
whichhyenc
Two sets c
Note: Carcass size groups are defined in the text. Number of assemblages and number of
identified specimens (NISP) are given. analyses. B
anata reported by Blumenschine (1988). conducted i
bNew analyses. assemhlageE
cnata reported by Marean and Spencer (1991) and Marean et al. (1992). bones (withe
dFive sites (NISP = 229) overlap with Berkeley 1. animal that
stone-only '
l composition
simulated sites are therefore divided into a fat-rich sample (10 assemblages)
and a fat-depleted sample (2 assemblages; Table 16-1). -! The carni vm
to be produt
The Serengeti simulated sites were probably disturbed mainly l]y spotted I whole long[
hyenas. This was not confirmed by direct observation since all d;isturbance j Thus, our
occurred at night and sites were not monitored continuously. However, data sets th
hyenas were heard at a number of the sites, and many preserved hyena spoor free-ranging
after disturbance were complete. Black-backed jackals (Canis mesomelas), com- stone-brokeJ
mon in the Seronera area, probably also contributed to assemblage disturb- hammers ton
ance, as Capaldo's (1990) ongoing experiments amply suggest. ed butcher 1
Disturbance was complete after the first night, as indicated by the usual from carnivc
total deletion of limb epiphyses. All bones remaining in the vicinity of the the no carnivore
original site (within a 50 m radius) were collected and cleaned by boiling for
•1
subsequent analysis. l
A second group of simulated sites was created at the Berkeley Spotted ~-
I
Hyena Colony, Berkeley, California,. between 1988 and 1989 (Marean 1991; 1
Marean and Spencer 1991, Marean et al. 1992). Limb bones and some post-
cranial axial bones of sheep (Ovis aries; size 1 /2) and cow (Bas taurinus; size c
3I 4) obtained from a local butcher. were used. portions. Er
The experiments at Berkeley afforded complete observations of hyena dis- distal articu·
turbance and strict control of their number and dominance rank. Such exper- ity' s sake, pr
imental control, which could not be achieved for the Serengeti sites, was 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). Pre-
simulation preparation of bones was similar to that used for the Serengeti
TOTAL
simulated sites. All remaining fragments from the Berkeley sites were
ssem NISP
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 tooth-
10 554 marking and epiphyseal:shaft fragment ratios for 13 of the Berkeley simulated
2 44
13 424 sites (referred to as "Berkeley 1" in Table 16-1). Comparable data from the
33 1162 Serengeti samples is presented in this bivariate form for the first time here. We
also summarize information on skeletal part and portion profiles reported
9 231 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
7 327 overlap with the above sample of 13. In Marean and Spencer (1991) and
Marean and colleagues (1992), this sample of 33 hyena-ravaged assemblages
74 2742 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
1d number of
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 hammer-
stone-only assemblages provide a pre-simulation control on the original
composition of the simulated sites in a way described by Blumenschine (1988).
ssemblages) 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
r by. spotted whole long bones of animals they killed or scavenged fron1 other carnivores.
disturbance Thus, our total experimental sample analyzed to date includes an array of
'. However, data sets that provide the complementary benefits of observations in both
hyena spoor free-ranging and captive settings. The sample comprises defleshed hammer-
melas), com- stone-broken bones scaven.ged by carnivores in a natural setting, defleshed
:tge disturb- hammerstone-broken bones scavenged by hyenas in a captive setting, deflesh-
ed butcher bones scavenged by hyenas in a captive setting, bones derived
>y the usual from carnivore kills in a natural setting, and hammerstone-broken bones with
ty of the the no carnivore involvement.
r boiling for
ley Spotted Definitions and Identification of Carnivore

~arean 1991;
Tooth Marks
some post-
zurinus; size Our sample of fragmented long bones is divided into the following
portions. Epiphyseal fragments preserve at least some of the proximal or
f hyena dis- distal articular surface, along with varying lengths of the shaft. For simplic-
Such exper- ity's sake, proximal metapodials are lumped with epiphyses even though they
:i sites, was derive from the metaphyseal growth area of the bone's diaphysis. Near-
l
1
,
l
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. By
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 --- 1Q(
comparable to the term limb shafts used by Bunn (e.g., 1986). "0 '
rn
Tooth ·marks were identified only after bones had been thoroughly mascer- 8(
ated by boiling to produce clean, grease-free surfac~s. Identification followed
+
'-"'
+-'
procedures and criteria of Blumenschine and Selvaggio (1988, 1991). A c
Q) 6(
16-power hand lens aided all identifications. Tooth marks on cortical and (J
L
Cl)
medullary surfaces occur as roughly circular to equilateral depressions (pits), 0...
or more linear marks (scores; see also Binford 1981; Bunn .1981; Potts and c 4(
0
Shipman 1981; Shipman and Rose 1983). The internal cross-sectional topogra- Cl)
::::;E
phy is U-shaped for scores (following Bunn 1981) and usually bowl-shaped 2(
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'c-
tures .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 l b. By
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
I 1Ql
light with the aid of a hand lens. Scanning electron microscopy is unnecessary, J ,......._
and optical microscopy reveals little of diagnostic value that the hand lens I '"0
(/) 81
does not. Our experience with control assemblages shows that thesff obscure T'""'
marks are as accurately diagnostic of carnivore action as marks associated +

...__,
with heavy gnawing. +-' 61
c
(l)
u
L
Tooth-Marking (l) 4
0...
c
Figure 16-2a shows the incidence of tooth-marked long bone frag- 0
(l) 2
ments for three sets of simulated sites and the Serengeti carnivore-only ~
sample. Blumenschine (1988) has reported the data upon which the carni-
vore-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.
1
I
a portion of
ivation from a. By Bone Portion · - carnivore only
t:s:sl Serengeti simulated site (fat-rich)
~rouped into IZZl Serengeti simulated site (fat-depleted)
md Spencer ~ Berkeley simulated site
1s bone. The
nents and is ,...... 100
"'0 .·
(f)
~hl y mascer- + 80
.on followed '-.J
+'
m, 1991). A c
0)
60
cortical and u
L
0)
ssions (pits), Q._
1; Potts and c 40
0
nal topogra- 0)
L
)Owl-shaped 20
rks, internal
>ressure was 0
·ores). Punc- EPI NEF MSH TOTAL
r and linear Bone Portion
1lar bone of
re much less
eep pits and
~ad, they are b. By Size Group
-pth), which,
.istinguished
.ow-incident 100 -Size 1 & 2
1nnecessary, ~
t:s:sl Size 3 & 4
te hand lens "0
(/) 80
lese obscure or-
s associated +
"-J
+-' 60
c
(!)
0
L.
(!)
(l_
40
c
g bone frag- 0
(!) 20
~ni vore-only L
:h the carni-
he sample is 0
rmatively as CARNIVORE SERENGETI BERKELEY
Lriability can ONLY SIMULATED SIMUlATED
SITES SITES
wwn for all
r-epi physeal
ted as being Figure 16-2. Proportion of long bone fragments that bear at least one
the marks it tooth mark, by (a) bone portion: EPI =epiphyseal fragment, NEF =near-
:onspicuous epiphyseal fragment, MSH = midshaft fragment, and (b) carcass size
group (see text for definitions).·
Carnivores extracting marrow and grease from whole long bones typically epiphyses tr
produce a high (mean = 84%, min. = 67%, max. = 100%) proportion of frag- consuming 1
ments with at least one tooth-mark (Blumenschine 1988). Epiphyseal frag- tion would E
ments that are not consumed completely are invariably tooth-marked, while a sampled by
slightly lower proportion of near-epiphyseal fragments (85%) and midshaft Capaldo (19~
fragments (83%) are tooth-marked. · et al. 1988).
The segmental distribution and overall incidence of tooth-marking in the Interestin~
carnivore-only sample contrasts strongly with those produced by carnivores marking at
scavenging hammerstone-broken long bones. The overall incidence of tooth- marking inc
marking in all three simulated site samples is significantly lower~ with sample Serengeti (f
means not exceeding 20%. The highest incidence, obtained for one fat-rich contrasts wi1
Serengeti site, is 45%. The substantially lower values for the simulated sites larger anima
are primarily due to the very low incidence of tooth-marked mi_dshafts. Here, (Figure 16-21
carnivores typically ignore hammerstone-generated midshaft fragments be- In summa
cause they have been deprived of virtually all nutrient value (Blumenschine ultimately d
1988). Our direct observations of site disturbance at the Berkeley Colony Carnivores '
confirm this pattern: only rarely, a hyena may briefly "mouth" a midshaft but 84% of bon
will invariably reject it without biting. Those simulated site midshafts that are epiphysis t<
tooth-marked are therefore derived exclusively from hammerstone-generated broken long
epiphyseal fragments through consumption of cancellous ends by hyenas. all bone spe1
Hyenas focus on epiphyseal fragments because they are the only bone to midshaft,
portions in hammerstone-generated assemblages of long bones that contain marks. The~
nutrients (grease; Blumenschine 1988). In destroying epiphyses, hyenas will secure distin
produce tooth-marked near-epiphyseal fragments in high frequencies. An access to ass~
interesting exception to this rule is provided by the two simulated sites com-
posed of fat-depleted bones. The absence of tooth-marked near-epiphyseal I
and midshaft fragments in the two assemblages (Figure 16-2a) is a result of an - J I
absence of on-site epiphyseal destruction. Remaining epiphyses are com-
monly tooth-marked in this sample, but only with tooth punctures inlo cancel-
lous bone, and not pits or scores ori cortical surfaces. The hyenas disturbing
I epiphyses t<
T
these sites apparently needed to use their sense of taste to sample the nutrient shaft fragmE
value of grease in the epiphyses' cancellous bone before rejecting it as a whole (near-epiph)
unit. tify carnivor
Variability in the incidence of tooth-marked fragments among the simu- sity of their 1
lated sites is probably also based on competition. Competition among scav- Loss of ha
engers of simulated sites will influence whether destruction of hammerstone- on-site destr
generated epiphyseal fragments occurs on-site, or only after epiphyses are to levels pro
transported away from the site to minimize competition while feeding. On- Such severe
site epiphyseal destruction will generate maximal levels of tooth-marked l exceeds by a
l.
fragments, while complete deletion prior to destruction will leave the sample. Los:
simulated site with no tooth-marked fragments but nevertheless depleted composed o:
numbers of epiphyses. This premise can only be supported at this time (Figure .16-3)
qualitatively. Nonetheless, in experiments at Berkeley using a single hyena, Variabilit)
we have commonly observed it to destroy many epiphyses at or near the site, sample is h:
unless a hot sun forced it to seek shade or an infestation of wasps forced a tencies in bo
retreat from the main concentration of bones. However, when two or more one assembli
hyenas are present, subordinate individuals will commonly remove any a (size 1) T
nes typically epiphyses they are able to acquire to a remote corner of the enclosure before
·tion of frag- consuming them (Marean etal. 1992). In free-ranging conditions, such dele-
'hyseal frag- tion would effectively remove any resulting fragments from the area typically
·ked, while a sampled by excavation, a premise that is being investigated systematically by
nd midshaft Capaldo (1990) with free-ranging carnivores in the Serengeti (see also Binford
et al. 1988).
Lrking in the Interestingly, bone size does not seem to influence the incidence of tooth-
y carnivores marking at the simulated sites. There is virtually no difference in tooth-
nee of tooth- mar king incidences between size 1 I 2 and size 3 I 4 animals for both the
with sample Serengeti (fat-rich) and the Berkeley simulated sites (Figure 16-2b). This
one fat-rich contrasts with the carnivore only sample, where the more robust diaphyses of
nulated sites larger animals seem to promote a higher incidence of tooth-marked fragments
shafts. Here, (Figure 16-2b).
·agments be- In summary, the frequency and distribution of carnivore tooth-marking is
lumenschine ultimately determined by the presence or absence of within-bone nutrients.
eley Colony (::arnivores consuming whole bones produce tooth marks on an average of
midshaft but 84% of bone fragments, and the frequency decreases only slightly from
1afts that are epiphysis to midshaft. In contrast, carnivores scavenging hammerstone-
le-generated broken long bones generate tooth marks on only an average of about 20% of
·hyenas. all bone specimens, and the frequency dwindles significantly from epiphysis
e only bone to midsWaft, with less than approximately 10% of the midshafts bearing tooth
that contain marks. These overall and segmental patterns of tooth-marking provide a
hyenas will' secure distinction between primary carnivore access and secondary carnivore
uencies. An access to assemblages of long bones.
~d sites com-
r-epiphyseal
1 result of an Long Bone Epiphysis: Shaft Fragment Ratios
;es are com-
; into cancel- The above shows that carnivores typically :onsume or destroy
LS disturbing epiphyses to access grease but ignore nutritionally unattractive, ungreasy
: the nutrient shaft fragments. This principle suggests that the ratio of epiphyseal to shaft
it as a whole (near-epiphyseal plus midshaft) fragments (NISP values) can be used to iden-
tify carnivore involvement with a bone assemblage and to measure the inten-
lg the simu- sity of their ravaging.
among scav- Loss of hammerstone~generated epiphyseal fragments through deletion or
trnmerstone- on-site destruction by carnivores can reduce epiphyseal:shaft fragment ratios
Jiphyses are to levels produced by carnivores destroying whole long bones (Figure 16-3).
feeding. On- Such severe epiphyseal loss in the fat-rich Serengeti simulated site sample
>oth-marked exceeds by an order of magnitude the mean value for the hammerstone-only
11 leave the sample. Loss of epiphys~al fragments was less severe in the simulated sites
~ss depleted composed of fat-depleted bones and in those created at the Berkeley Colony
at this time (Figure 16-3). ·
ingle hyena, Variability in epiphyseal:shaft fragment ratios for the hammerstone-only
~ear the site, sample is high ..Part of this variability results from unavoidable inconsis-
sps forced·a tencies in bone fracturing methods. However, much of the variability is due to
:wo or more one assemblage (NAA39G) compos.ed ~f forelimb and hindlimb long bones of
remove any a (size 1) Thomson's gazelle (Figure 16-3). Bones of such small animals
l "
2841 R. ]. Blumenschine and C. W. Marean
1
!
1.6
(f) ~Size 1 & 2
+'
4--
0
lZZl Size 3 & 4 0.
...c D NM39G (size 1)
(/) 1.2 (f)
" ;;-
+'
4--
~ 0
({) (/) ...c
(/) ,.......
~ ~ 0.8 "~
Q)
·- 0 .
(/}
CJ)
0... 0 (/} 0
w ~ >. >
L ..._; ...c 0
0... L
Q)
...0 0.4 ··a_ 1
E w (f)
::J L 0 0.
z Q) 0...
...0 ......_,_
E
::J
fat-rich fat-depleted · Berkeley z
CARNIVORE SIMULATED SITES HAMMERSTONE
ONLY 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
dt
typically require only one blow to access all marrow with a probe. This mini-
mizes the production of shaft fragments in comparison to bones of larger
of
ra
animals where addi tiona! blows toward the proximal and distal ends are
needed to remove all marrow. If the resulting very high epiphyseal:shaft frag-
ment ratio (1.56) of this assemblage is excluded, the hammerstone-only mean destroyed b)
drops to .36, a value still greater than the me;;tns for any of the simulated site , more severe
samples. stone-on!y as
Aside from the above example, carcass size does not seem tci have a fragments ari
consistent effect on epiphyseal: shaft fragment ratios. In fact, epiphyseal loss is nally thoughi
similar for assemblages using size 1/2 and size 3/4 bones across the spectrum The two rr
of site formation scenarios examined (Figure 16-3). independent
A tight (r = -.94) negative relationship exists between the post-ravaging always resuli
epiphyseal:shaft fragment ratio and the observed percentage change in the relationship ~
original number of epiphyseal fragments (Figure 16-4).1 This shows that the be described
epiphyseal:shaft fragment ratio accurately measures depletion of epiphyseal fairly immun
fragments by carnivores. Percentage change in the number of epiphyseal ber and seqw
fragments was calculated using the following formula: Variability
probably a p
b-a (1991) has qt
%change=-- x 100 analyses sug
b duced is in(
decreases~ Gi
where b is the number of epiphyses before disturbance, and a is the number seal:shaft fra:
remaining at or near the site after disturbance. Values forb were recorded for the Serengeti
all simulated sites upon breakage, while those for the carnivore-only sample than that exp
were obtained by simply multiplying the estimated number of whole bones 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_
w (/)
l <>
L Q 0.2 <>
Q) o_
_Q ......,
E y = .32- .0031x
:J
z r = .94
MERSTONE
ONLY 0.0+-~~~~-,--~.-~--.-~-.--~-.~~w
-40 -20 0 20 40 60 80 100
e epiphyses as % Change in Number of Epiphyseal Fragments
experimental
[ (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
1e. This mini- of epiphyseal fragments, and post-ravaging epiphyseal:shaft fragment
nes of larger , ratios.
;tal ends are
!al:shaft frag-
Le-only mean destroyed by two. Increasingly positive percentage change values indicate
[mulated site more severe epiphyseal loss. Negative values, obtained for most hammer-
stone-only assemblages (Figure 16-4), reflect the post-boiling addition of shaft
n to have a fragments arising from the removal of periosteum that bound fragments origi-
hysealloss is nally thought to represent a single hammerstone-broken fragment.
the spectrum The two measures of epiphyseal deletion/ destruction are not completely
independent. A 100% change in the number of epiphyseal fragments will
ost-ravaging always result in an epiphyseal:shaft fragment value of zero. Nonetheless, the
hange in the relationship shows that the intensity of epiphyseal removal by carnivores can
.OWS that the be described by a single measure-epiphyseal:shaft fragment ratios-that is
>f epiphyseal fairly immune to differences in carcass size, fat content of bones, and the num-
f epiphyseal ber 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 intro-
duced is increased (i.e., increased competition), survival of epiphyses
decreases. Given this relationship, one can infer that the lower mean epiphy-
; the number seal:shaft fragment ratios obtained for the fat-rich simulated sites created in
recorded fbr the Serengeti results from a higher level of competition among site scavengers
-only sample than that experienced by the Berkeley sites. This is not surprising, given that
whole bones 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 j 10
i
likely to promote less competition than those which can characterize the site l
disturbance of free-ranging hyenas.
In summary, carcass size does not ·severely affect the proportionate survival I ~
......., 8
of limb epiphyses and shafts. Epiphyseal:shaft fragment ratios correlate well I (/)
t
with the depletion of epiphyses by carnivores across the spectrum of site
formation scenarios examined here. Hence, this simple ratio value of fragment
l
~
<
I
(/) 6
0
counts (NISPs) can provide an accurate estimate of epiphysis loss through w
~
(Y
carnivore agencies. Variation in epiphysis:shaft ra~ios reflects the intensity of < 4
:::::E
carnivore destruction, and in a way that appears to be sensitive to levels of I
I
competition among carnivores. 1-
0 2
0
1-
Epiphyseal Destruction and the Incidence of

Tooth:..Marking
More thorough bone destruction by carnivores should produce
higher proportions of fragments that are tooth-marked. For assemblages of
long bones, low epiphyseal:shaft fragment ratios should be accompanied by a
relatively high proportion of tooth-marked bones. Further, at comparably low
epiphyseal:shaft fr&gment ratios, tooth-marking should be higher when carni-
vores destroy whole long bones than when their consumption is restricted to
grease in epiphyses from a hammerstone-generated assemblage.
These predictions are borne out in a general manner by our experimental
second set<
data. Figure 16-5 shows epiphyseal destruction plotted against the proportion
and one frc
of shaft fragments (near-epiphyseal and midshaft) that are tooth-marked. The
portion oft
relationship is monotonic negative and incorporates the three simulated site
loss of epip
samples and the hammerstone-only and carnivore-only samples. As expected,
The devi<
the simulated sites show a negative trend, as indicated by the lecrst-squares
sites contaiJ
regression line in Figure 16-5. At low levels of tooth-marking (i.e., minimal
suggest tha
disturbance), many simulated sites show epiphyseal:shaft fragment ratios
plished thi
expected by their derivation from hammerstone-only assemblages. As well,
epiphyses f
the carnivore-only assemblages preserve evidence of epiphyseal destruction
consumptic
as intense as that witnessed for the most competitive simulated sites, but at
ducing neg;
levels of tooth-marking substantially higher. Again, this difference is due to
deviations
the fact that tooth-marked fragments in the carnivore-only assemblages result
destruction
from the extraction of both epiphyseal grease and diaphyseal marrow.
that negath
When the simulated sites are examined alone (filled symbols in Figure 16-.
engers is hi
5), deviations from the expected relationship between epiphyseal loss and
model is se·
tooth-marking become more apparent. This is reflected in the low but still
sample of:
significant correlation coefficient (r = -.40, p = .05) for the regression equation
hypothesis'
describing all simulated sites. One set of exceptions includes three Berkeley
sites and the two fat-depleted sites that show strong negative deviations from
the expected relationship. These assemblages contain no tooth-marked shaft
fragments but preserve epiphyseal:shaft fragment ratios lower than expected
from hammerstone breakage alone. Indeed, all of the sites are known to have Our disc
experienced some epiphyseal loss as a result of carnivore disturbance. A destruction
1ndi tions are 100 0 0 = Hammerstone only

0 0 = Carnivore only
~rize the site
0
Simulated Sites:
1ate survival 80 • = Serengeti, fat-rich
(/) 0
Jrrelate well t Y = Serengeti, fat-depleted
trum of site
<( 0 A. = Berkeley
I
(/) 60
~ of fragment 0
.oss through w
~
n:::
~intensity of <(
::2
40 • ... •
~ to levels of I
I
1-
g 20 ... ...
1- ...
0 Q
... ...
II
0+-~b-~~~~~~--~-~--~~~rr-.~~
/7 0
-"'-J.... - ~ ...
0 0.1 0.2 0.3 '
0.4 0.5 1.5 1.6
uld produce NUMBER EPIPHYSES/SHAFT

.emblages of
1panied by a Figure 16-5. Relationship between the extent of epiphyseal removal
tparably low (number epiphyseal/shaft fragments) and the proportion of shaft frag-
when carni- ments that bear at least one tooth mark. The dashed line is the least
restricted to squares regression line for simulated sites (filled symbols) only.
~xperimen tal
second set of simulated sites, including two from the Serengeti fat-rich sample
e proportion
and one from Berkeley, shows a strong positive deviation, preserving a pro-
marked. The
portion of tooth-marked shafts higher than that expected by their incomplete
mulated site
loss of epiphyseal fragments.
As expected,
The deviations are not based on carcass size. Both deviant sets of simulated
east-squares
sites contain assemblages made from size 1/2 and size 3/4 bones. Rather, we
.. e., minimal
suggest that the deviations are related to whether epiphyseal loss was accom-
;ment ratios
plished through on-site destruction of epiphyses or through removal of
~es. As well,
epiphyses from the site prior to consumption. Removal of epiphyses prior to
. destruction
consumption will minimize production of tooth-marked shafts on-site, pro-
sites, but at
ducing negative deviations from the regression line. Conversely, the positive
tee is due to
deviations may result if a majority of epiphyses were lost through on-site
blages result
destruction· rather than removal from the site area. In summary, we propose
·ow.
that negative deviations may be expected when competition among site scav-
n Figure 16-
engers is high, while positive ones may result when competition is low. This
ealloss and
model is set out in schematic fashion in Figure 16-6.2 Our currently analyzed
low but still
sample of sites from the Berkeley Colony does not allow us to test that
lon equation
hypothesis at this time.
ree Berkeley
ia tions from
ctar ked shaft Skeletal Part Profiles
tan expected
own to have Our discussion of a carnivore's selection of bones for deletion from or
turbance. A destruction at a site has thus far considered only NISP values of the various
288 I R. f. Blumenschine and C. W; Marean 1
j
EPIPHYSEAL:SHAFT FRAGMENT RATIO j

I
j 12(
ij
LOW HIGH
j
i (/)
I l 1 Q(
(/)
l ~
I z<(
t<( I
~ HIGH
lower
competition
I
I
I
I
not
applicable I 0:::
lL
8(
0 I 0
w
~
I
I I L
::J
6(
0:::
<(
-~--------------+---------------
1 . l (/)
2
I
I
I
I 4(
higher I no 1
~ LOW competition
I
I competition j
2(
0
0
I (ct·O% tooth-marking) lj (
I
f-
I
I
~ I
Fi
Figure 16-6. Qualitative model of the degree of competition among
j se,
scavengers of assemblages of hammerstone-broken long bones. i fn
15
1j
portions of unspecified long bones. Here, we summarize data reported by ·l

Marean and colleagues (1992) on selection and attendant deletion/ destruction J
! ratios. Bones
of specific skeletal parts by hyenas at the Berkeley Colony. Sites included in i destroy and c
the analysis are all from sheep and include either one, two, or three hindlimbs I
I
The ultimc::
I
plus a section of the vertebral column (either atlas to third thoracic, or fifth
and sixth lumbar plus the sacrum). A hindlimb included one innominate, a .l Berkeley colo
seen in Figur,
l
femur, a tibia, a metatarsal, pne set of t9-rsals, and the first phalanges.,All
bones were defleshed as thoroughly as possible/In the simulated s.i'tes, only l parts used in
represented :
long bones were fractured for their marrow; pelves, vertebrae, and carpals/
tarsals were left intact. Il
abundances i
phalanges sh1
The sequence by which skeletal parts are chosen for consumption can be diate values.
expressed by summing the rank order removal of like parts from individual Thus, hyen
sites (see Marean et al. 1992 for analytical details). Rank order of removal
could be determined confidently, by continuous observation of site distur-
I dramatic effe,
bones or bonE
bance, for 17 simulated sites in this sample. Vertebrae were removed first in ribs very rar
16 of the 17 assemblages, while pelves were removed typically second, and at compact bon
latest third, in 16 of 17 experiments. An intermediate group includes the distal numbers if tal
and proximal femur, proximal tibia, and proximal metatarsal (which were left
articulated to the tarsals and connected by tendon to a phalangeal mass). Parts
most likely to be chosen last for consumption are the distal metatarsal and Dj
distal tibia: the former was the last or penultimate bone to be removed in 11 of
the 17 assemblages, while the latter occupied the last two positions in 16 of the Th
17 experiments (Marean et al. 1992): implications f
The sequence by which skeletal parts are chosen for consumption is pretations of :
positively and significantly related to bulk density of the part (Figure 16-7; contain dual-I
Marean et al. 1992).3 Less dense bones have higher cancellous:cortical bone applying the
·~,
~~·''r. .-...
. . .······.···.l·:..
A Carn1vore' s View I 289

)
120
/
(/) /
~
100 /
z<( >/
PFEM ....
0::: 80 /
8 PTIB / 0
LL
0 DFEM e PMTM
(with tarsals &
6
2 60 1st phalanges)
:J ;...-
(/)
40 e PEL
/ VERT
20
0.15 0.25 0.35 0.45 0.55
BULK DENSI1Y
Figure 16-7. Relationship between bulk density (Lyman 1984) and the
;titian among sequence by which skeletal parts of sheep were observed to be removed
es. 'from 17 Berkeley simulated _sites by spotted hyenas (after Marean et al.
/ 1992).
reported by
L/ destruction ratios. Bones chosen first by hyenas are therefore those that are easier to
:; included in destroy and contain more grease, giving them a higher net yield.
·ee hindlimbs The ultimate survivorship of bones subjected to hyena ravaging at the
racic, or fifth Berkeley colony shows a pattern that follows the sequence of removal. This is
1nominate, a seen in Figure 16-8, which compares the pre- and post-ravaging MNEs of all
1alanges. All parts used in the Berkeley experiments (after Marean et al. 1992). Long bones,
~d sites, only represented mostly by midshafts, show high survivorship, remaining in
and carpals/ abundances identical with or· close to original MNEs. Axial parts and first
p4alanges show very .low survivorships, while compact bones show interme-
.ption can be diate values.
m individual Thus, hyena scavenging of hammerstone-broken and unbroken bone has a
r ofremoval dramatic effect of skeletq.l element survival. Hyenas choose for consumption
,f site distur- bones or bone portions with the lowest bulk density, such that vertebrae and
toved first in ribs very rarely survive consumption while roughly half of pelves and
~cond, and at compact bones do. Limb bones survive at levels comparable to .original
des the distal numbers if tallies are based on midshaft fragments.
1ich were left
l mass). Parts
etatarsal and Discussion and Implications
.oved in 11 of
ts in 16 of the The experimental results reported and summarized here have
implications for the methods of archaeological bone analysis and for inter-
.sumption 'is pretations of site formation and the nature of hominid behavior at sites that
(Figure 16-7; contain dual-patterned bone assemblages. The implications are illustrated by
cortical bone applying the experirpental results to data from the ric;h and well-preserved
290 I R. f. Blumenschine and C. W. Marean
mediate to tl
Atlas 6 ~ 2
(Bl umenschi1
Axis 6 ~ 1
30~2
tooth-markec
Cervical
Thoracic 18~ is closely cc
Lumbar 56 S~rengeti sirr
Sacral 140 5 lous only wi
Rib 36~1 substantially
Pelvis 49 cracking qur
Femur 50 result has be
Tibia 50 tion agents a
Fibula 50
the single- c
Astragalus 50
available (Bll
Calcaneum 50
Navicular-Cuboid 50
tion, indudii
External Cuneiform 50 followed by
Internal Cuneiform 50 Selvaggio (IS
Metatarsal 50 An additic
First Phalange 100 fragments e)
nature of epi
150 100 50 0 50 100 150
Pre-ravaging MNE. Post-ravaging MNE
mental sam]
changes in t
consumption
Figure 16~8. Pre- and post-ravaging MNE estimates for 33 experimental epiphyseal:sl
assemblages at Berkeley, using sheep bones (after Marean et al. 1992). seal destruct
. mental relatit
assemblage from FLK I level 22 (Zinjanthropus level), Bed I, Olduvai Gorge, model shoulc
that havebeen reported by Bunn and Kroll (1986) and Blumenschine and primary accu
colleagues (1990). More than any other site, it is the FLK Zinjanthropus level ratio of .14 f(
upon which debates on the nature of Plio-Pleistocen,e hominid beha~ior have (Blumenschil
been based (e.g., Binford 1981, 1986, 19~8; Bunn and Krol11986, 1988; Leakey yields a pred
1971). ' of 58% from
It has been argued that the overall incidence and segmental distribution of degree of rav
tooth-marked long bone fragments are sensitive to the relative timing of Such a higl
carnivore and hominid influence on an assemblage (Blumenschine 1988). A at FLK Zinjat
high (> 67%) overall proportion and even segmental distribution of tooth- quantified m
marked fragments in the carnivore only sample was proposed to signal initial content of th
access by carnivores to marrow cavities. Conversely, an overall incidence of< counts. Ratic
45%, where additionally most tooth-marking is restricted to epiphyseal and have been ust
near-epiphyseal fragments, was sugges'ted to signal a restriction of access by subject to a n
carnivores to hammerstone-generated epiphyseal fragments (Blumenschine portions) an(
1988). The Berkeley simulated site sample has been shown here (Figure 16-2a) epiphyseal:sl
to be consistent with this original model and to substantiate it. · by carnivore
Tooth-mark data acquired during a new analysis of long bones from the MNEs based
FLK Zinjanthropus assemblage are still being analyzed (Blumenschine et al. Kroll's (1988:'
1990). Nonetheless, the preliminary analysis reveals patterns that can be Zinjanthropus
discussed qualitatively at this time. plus shafts is
The FLK Zinjanthropus assemblage preserves an overall incidence of a minimum j
tooth-marked long bone fragments from size 1-4 mammals that is inter- whole femor;
!
i
,.,i,l.
mediate to the carnivore-only and simulated site scenarios of site formation

(Blumenschine et al. 1990; refer to Figure 16-2a). However, the proportion of
tooth-marked epiphyseal and near-epiphyseal fragments at FLK Zinjanthropus
is closely comparable to the means for these portions obtained for the
Serengeti simulated sites. Indeed, the FLK Zinjanthropus assemblage is anoma-
lous only with regard to midshafts, which are tooth-marked at frequencies
substantially higher than that expected for an assemblage created by a bone-
cracking carnivore destroying hammerstone-generated epiphyses alone. This
result has been used to suggest that the sequence and identity of site forma-
tion agents at FLK Zinjanthropus were more complex than those modeled by
the single- or dual-patterned scenarios of assemblage formation currently
available (Blumenschine et al. 1990). More complex sequences of site forma-
tion, inCluding carnivore defleshing followed by hominid marrow extraction
followed by carnivore destruction of epiphyses, are being investigated by
Selvaggio (1990).
An additional source of variability in the incidence of tooth-marked shaft
fragments explored for the first time here (Figure 16-5) is the extent and
nature of epiphyseal loss. A strong correlation was obtained from the experi-
0 150
g MNE
mental sample between epiphyseal:shaft fragment ratios and observed
changes in the number of epiphyseal fragments resulting from carnivore
consumption of marrow and/ or grease (Figure 16-4). This suggests that the
3 experimental epiphyseal:shaft fragment ratio can be used to estimate the degree of epiphy-
t al.1992). seal destruction by carnivores at an archaeological site. Because the experi-
mental relationship included all experimental sets, the predictive value of the
duvai Gorge, model should be high regardless of whether hominids or carnivores were the
~nschine and primary accumulator or consumer of carcass parts. Given an epiphyseal:shaft
!thropus level ratio of .14 for size 1-4 mammal long bones from the FLK Zinjanthropus site
~ehavior have (Blumenschine et al. 1990), solution of the regression equation in Figure 16-4
1988;Leakey yields a predicted percentage change in the number of epiphyseal fragments
of 58% from that originally present at FLK Zinjanthropus. It indicates a high
istributiqn of degree of ravaging of the assemblage by carnivores.
ve timing of , Such a high predicted rate of epiphyseal deletion/ destruction of epiphyses
1ine 1988). 'A at FLKZinjanthropus is important for two·reasons. First, it provides a readily
ion of tooth- quantified measure of carnivore-induced change in the long bone epiphysis
, signal initial content of the FLK Zinjanthropus assemblage that is based on simple NISP
ncidence of < counts. Ratios of proximal:distal humeri and proximal tibae:distal femora
iphyseal and have been used to identify carnivore ravaging (e.g., Binford 1981), but they are
L of access by
subject to a number of biases (e.g., differential access and transport of skeletal
,lum~nschine portions) and do not indicate the degree of ravaging. The accuracy of the
Figure 16-2a) epiphyseal:shaft fragment ratio for measuring long' bone epiphyseal deletion
by carnivores can be tested using Bunn and Kroll's (1988) calculations of
nes from the MNEs based on epiphyses alone versus those that also use shafts. Bunn and
1schine et al. Kroll's (1988:Table 3) MNE estimate for size 1-3b femora and tibiae from FLK
that can be Zinjanthropus based on epiphyses alone is 33, while that based on epiphyses
plus shafts is 53. The more accurate.MNE estimate based on shafts means that
incidence of a minimum of 106 epiphyses (53 x 2) .were originally present at the site if
that is inter- whole femora and tibiae wereintroduced. Hence, 73 epiphyses (106-33) are
I I
292 I R. J. Blumenschine mid C. W. Marean
1
missing according to Bunn and ;Kroll's estimates. This yields a percentage i pattern has c
change value of 68.9%, which is close to our e_stimate o£'58% obtained for all j and Kroll 19.
long bones. Apparently, the long bo~e epiphyseal:shaft fragment ratio pro- limbs and he
vides an accurate and easily obtained qt~antitative estimate of the degree to et al. 1988). :
which carnivores distorted the original composition of an arc;haeological bone (1992) that w
assemblage. Perhaps this will not apply to assemblages that have experienced I and destroy
severe post-depositional compaction, such as many cave sites where NISP generated lor
values are inflated relative to ~I values (Klein and Cruz-Uribe 1984). ated and exJ
Second, the extent of epiphyseal loss at FLK Zinjanthropus can be used to ranging earn
evaluate the incidence of tooth-marking at the site if one assumes that the that a long-b<
dual-patterning of the FLK Zinjanthropus assemblage reflects a sequence of ity, such as F
site formation agents modeled by the simulated site scenario (cf. Bunn et al. of post-crani1
1980; Bunn and Kroll 1986; Isaac 1983; Leakey 1971). Given this assumption, negative rela
the tooth-marking and epiphyseal:shaft fragment values at FLK Zinjanthropus Zinjanthropus
group it with the simulated sites whose rate of tooth-marking is higher than (Marean et al
that expected from their degree of epiphyseal loss. According to the model required exp
schematized in Figure 16-6, FLK Zinjanthropus would represent an assemblage dominated b1
disturbed in a context of low competition (such as by a solitary hyena in a This prese;
riparian setting), where one would expect all epiphyseal destruction to occur transport by l
on-site. This inference would further suggest that FLK Zinjanthropus was not can produce-
the "hotbed" of contpetition among carnivores and hominids suggested by thJrd process
Potts's (1984) stone cache model of site formation. FLK Zinjanth
Finally, our estimate of the extent of epiphyseal loss at FLK Zinjanthropus scavenging b
has a bearing on the methods by which archaeologists should estimate long row, brain) o
bone MNEs for archaeological assemblages demonstrably affected by tion toward 1:
carnivores. Bunn and Kroll (1986) and Bunn (1986) contend that accurate broken throu
MNE estimates for long bones from Plio-Pleistocene archaeological sites that tion must als·
have been ravaged by carnivores can only be determined on the basis of limb bution of star
shafts and, hence, that those obtained by Binford (1981) on the/basis of tooth marks.
epiphyses alone are flawed. An experimental and quantitative basis for this Zinjanthropus
contention has been shown by Marean and Spencer (1991) to be true of ing of carniv·
midshafts only. At the Berkeley Colony, hyena ravaging of whole or hammer- tissues consu:
stone-broken femora, tibiae, and metatarsals reduces MNEs by as much as an
average of 85% if they are based on epiphyses alone (see Figures 3-5 in
Marean and Spencer 1991). Furthermore, they have shown that the reduction c
is highly variable among epiphyses of the three long bones used, and closely
correlated to bulk density. MNE estimates based on proximal or distal shafts A
of these bones underestimate original ~Es to a lesser (by as much as 40%) where recon~
but still variable extent among skeletal parts. Rather, only midshaft fragments standing of t
survive carnivore ravaging in frequencies sufficient to reconstruct the original, Our experimc
pre-ravaging ~Es of an assemblage accurately. to show that
The experimental results summarized here on the rank-order removal and crucial for in
ultimate survivorship of skeletal parts from simulated sites (Figures 16-7 and nature of hon
16-8) have implications for the interpretation of skeletal part profiles at FLK Zooarchaec
Zinjanthropus and other sites. Many Pleistocene archaeological sites, including record data il
FLK Zinjanthropus, preserve bone assemblages dominated by long bones and ment with an
heads. Vertebrae, pelves, and ribs are poorly represented. This recurrent to bones in m
ITI
A Carnivore's View 1293
a percentage pattern has often been ascribed to the schlepp effect (e.g., Bunn 1986; Bunn
>tained for all and Kroll 1986), an assemblage bias produced by the selective transport of
ent ratio pro- limbs and heads by hominids from kill sites to camp sites (but see O'Connell
: the degree to et al. 1988). However, the experimental data from Marean and colleagues
~ological bone (1992) that was summarized here show that carnivores preferentially delete
2 experienced and destroy vertebrae, pelves, and ribs before consuming hammerstone-
:where NISP generated long bone epiphyseal fragments. This patterning is being substanti-
1984} ated and expanded to include other skeletal parts by Capaldo, using free-
an be used to ranging carnivores. As such, Marean and colleagues (1992) have suggested
tmes that the that a long-bone-and-head-dominated assemblage evidencing carnivore activ-
a sequence of ity, such as FLK Zinjanthropus, can result from selective deletion/ destruction
£ Bunn et al. of post-cranial axial parts by carnivores. This is supported by the significant,
; assumption, negative relationship between the MAU representation of parts from FLK
Zinjanthropus Zinjanthropus and the survivorship of these parts obtained experimentally
s higher than (Marean et al. 1992). Selective transport of limbs by hominids to sites is not a
to the model required explanation for dual-patterned archaeological assemblages that are
n assemblage dominated by these parts.
7 hyena in a This presents a classic case of equifinality, where two processes (selective
:::tion to occur transport by hominids and selective post-butchery deletion by site scavengers)
·opus was not can prodpce the same, limb-dominated pattern of assemblage composition. A
suggested by third process has also been shown to be capable of producing this pattern at
FLK Zifzjanthropus and another Bed I Olduvai site, FLKN 1/2: unselective
Zinjanthropus scavenging by hominids of carcass parts remaining with edible contents (mar-
~stimate long row, brain) on defleshed, abandoned felid kills will bias skeletal part acquisi-
affected /by tion toward limbs and heads (Blumenschine 1991). The equifinality cannot be
that accurate broken through recourse to skeletal part data alone. Data on bone modifica-
:ical sites that tion must also be consulted, particularly the frequency and anatomical distri-
basis of limb bution of stone tool cut marks, hammerstone percussion marks, and carnivore
the basis of tooth marks. This was the emphasis of the reanalysis of long bones from FLK
basis for this Zinjanthropus (Blumenschine et al. 1990), which should help to assess the tim-
:o be true of ing of carnivore versus hominid access to the bones and the types of edible
e or hammer- tissues consumed by each:
$ much as an
gures 3-5 in
the reduction Conclusions
i, and closely
~ distal shafts A carnivore's view of an archaeological bone assemblage is one
nuch as 40%) where reconstruction of their feeding activities is prerequisite to an under-
aft fragments standing of the particular hominid behaviors encoded in the bone remains.
t the original, Our experimental results.and applications to.FLK Zinjanthropus are intended
to show that an understanding of carnivore behavior at archaeological sites is
removal and crucial for interpreting the role of hominids in site formation, including the
1res 16-7 and nature of hominid foraging, feeding, and social behavior.
ofiles at FLK Zooarchaeologists analyzing dual-patterned assemblages must consistently
tes, including record data indicative of the timing, degree, and nature of carnivore involve-
:1g bones and ment with an assemblage. Estimates of .the relative timing of carnivore access
his recurrent to bones in an archaeological assemblage can be made if data are collected on
1
2941 R. ]. Blumen~chine arid C. W. Marean .
the segmental distribution and overall incidence of tooth-marked long bone species may 1
fragments. Tooth-mark identifications cannot be limited to punctures, fur- port, and pre
rows, and intense areas of gnawing,. a limitation that, we strongly suspect, reconstructin
characterizes most zooarchaeological analyses. Rather, zooarchaeologists on dual-pattE
must also seek the far more numerous, less conspicuous, bu~ equally diagnos- seek signs no
tic marks that occur as shallow and isolated pits and scores on cortical and a hungry can
medullary surfaces. Many of the tooth-marked fragments in our experimental
samples preserved only a single, inconspicuous pit or score. Nevertheless,
.they can be identified with high accuracy and preci~ion with the aid of a hand A
lens and strong, low-incident light, thus overcoming the prohibitive time and T1
cost of scanning electron microscopy on whole assemblage analysis. proceededw
The degree of carnivore involvement with an assemblage, including the Research Ins1
amount of epiphyseal loss attributable to carnivore ravaging, can be assessed warden of th
if epiphyseal:shaft fragment ratios are recorded. As shown through applica- and the staff
tion to FLK Zinjanthropus, this ratio, based on NISP values, provides a simple in the Seteng
estimate of long bone epiphyseal loss that provides comparable results to that of the Seren~
derived from the tedious calculation of MNEs based on shafts and epiphys~s. bones. Our l
As such, the epiphysis:shaft fragment ratio can be used to evaluate the degree fited from 01
to which reliance on epiphyses alone will bias MNE estimates of long bones. Hudsonforu
In combination with tooth-mark data, epiphyseal:shaft fragment ratios might
also be sensitive to tne competitiveness of the site setting.
Finally, the nature of carnivore involvement with an archaeological bone N
assemblage must be ascertained. In particular, one must determine whether 1. The equal
carnivores, as opposed to hominids, were responsible for producing dispro- regression lim
portionate skeletal part profiles. When epiphyseal loss through carnivore blages arise fr
ravaging or other density-dependent processes is detected, accurate skeletal (NAA39G) at;
part profiles for long bones can be produced only when MNEs are based on 2. The lowE
epiphyses and shafts. The resultant skeletal part profiles might, then be attrib- Figure 16-5, th
utable to hominid behavior if the frequency qnd anatomical patterning of regression lin,
tooth marks, percussion marks, and cut marks can be used to distinguish positive residt
selective post-butchery deletion of parts by site scavengers from either differ- with a high in
ential availability to hominids of edible parts at a death site or selective trans- carnivore and
port of carcass parts by hominids. 3.Sequence
each of the 17
It is no surprise that controversy rather than consensus pervades interpre-
sitometer sca1
tations of site formation and hominid behavior at many Pleistocene sites. Very miscellaneous
few archaeological bone assemblages have been analyzed and reported in the PFEM = proxi
ways prescribed above. Evidence of carnivore activity must be given more metatarsal, P:tv
than lip service so that anthropocentric interpretations of assemblage forma- relationship bl
tion do not result. Conversely, undue reliance on single-agent taphonomic sal, which is r1
models derived from carnivore studies should be avoided so that "carne- was left attach
centric" interpretations do not arise. Rather, the complex predepositi'onal
histories of archaeological bone assemblages for which carnivores and homi-
nids were the dominant biological agencies can only be modeled by experi-
R
mentation that involves both sets df agents. Andrews, P., a
The studies reported here, while more realistic than single-agent studies, 1985 Nah:
represent only the simplest of a diversity of possible dual-patterned scenarios Binford, L. R.
where multiple episodes of carnivore involvement by different carnivore 1977 Olor
pological E
·ry~·' ·:'• ·;.·,_~.· ,
.• .·.'···;·
I"
A Carnivore's View 1295
~d long bone I species may have been embedded within the carcass part procurement, trans-
1ctures, fur- port, and processing strategies of hominids. We are ultimately interested in
.gly suspect, reconstructing hominid behavior from bones. This can only be accomplished
chaeologists on dual-patterned archaeological assemblages if our description and analysis
:tll y diagnos- seek signs not only of hominid activity but are also sensitive to the concerns of
. cortical and a hungry carnivore. ·
~xperimen tal
Jevertheless,
tid of a hand
Acknowledgments
ive time and The studies upon which this paper is based could not have
is. proceeded without the cooperation o~ Karim Hirji of the Serengeti Wildlife
lcl uding the Research Institute, the director of the TanzaniaN ational Parks, the chief park
l be assessed warden of the Serengeti National Park, and Steve Glickman, Laurence Frank,
ugh applica- and the staff at the Berkeley Spotted Hyena Colony. Charles Saanane assisted
des a simple in the Serengeti experiments, Marie Selvaggio and Sal Capaldo in the analysis
~suits to that of the Serengeti bones, and Lillian Spencer in the analysis of the Berkeley
d epiphyses. bones. Our understanding of "dual-patterned" bone assemblages has bene-
:e the degree fited from ongoing discussions with Sal Capaldo. We are grateful to Jean
: long bones. Hudson forI useful comments on the final draft.
ratios might
I
logical bone Notes

Line whether 1. The equation and Pearson's correlation coefficient is shown for the least squares
tcing dispro- regression line. Negative percentage change values for the hammerstone only assem-
~h carnivore blages arise from the effect of boiling (see text). One hammerstone only assemblage
Lra te skeletal (NAA39G) at X= -7.7, y =1.56 is not shown (see text).
ue based on 2. The lower right cell is represented by the hammerstone only assemblages in
len be attrib- Figure 16-5, the lower left cell by the simulated sites that deviate negatively from the
~atterning of regression line in Figure 16-5, and the upper left by the simulated sites showing
distinguish positive residuals in Figure 16-5. A high epiphyseal:shaft fragment ratio combined
2i ther differ- with a high incidence of tooth-marked fragments is not modeled by any sequence of
lective trans- carnivore and hominid agencies presented here.
3. Sequence of removal is measured by the sum of ranks of a part's removal from
each of the 17 assemblages. See Marean and colleagues (1992) for the photonden-
:les in terpre- sitometer scan sites of Lyman (1984) that were used for bulk density. VERT ~
te sites. Very miscellaneous vertebrae (see text), PEL = one-half of the pelvis, DFEM = distal femur,
Jorted in the PFEM = proximal· femur, PTIB = proximal tibia, DTIB = distal tibia, DMTM = distal
given more metatarsal, PMTM = proximal metatarsal. Solid line is the best-fit line describing the
blage. forma- relationship between all points. Dashed best-fit line excludes the proximal metatar-
taphonomic sal, which is removed earlier than expected because the tarsals/first phalange mass
that "carne- was left attached to it.
depositional
$and homi- References
d by experi-
Andrews, P., and J:Cook
~ent studies, 1985 Natural Modifications to Bones in a Temperate Setting. Man 20:675-691.
.ed scenarios Binford, L. R.
n t carnivore 1977 Olorgesailie Deserves More Than the Usual Book Review. Journal of Anthro-
pological Research 33:49~502.
l
296j R.]. Blumenschine and C. W. Marean
1978 Nunamiut Ethnoarchaeology. Academic Press, New York Bunn,H. T.

1981 Bones: Ancient Men and Modern Myths. Academic Press, New York. 1981 At
1984 Faunal Remains from Klasies River Mouth. Academic Press, New York. fromK,
1986 Comment on Bunn and Kroll's "Systematic Butchery by Plio-Pleistocene 1983 Cc
Hominids at Olduvai Gorge." Current Anthropology 27:444-446. Gathen
1988 Fact and Fiction about the Zinjanthropus Floor: Data, Arguments and Inter- DenNt
pretations. Current Anthropology 29:123-135. Prey, ec
Binforci, L. R., and J. B. Bertram Series 1
1977 Bone Frequencies and Attritional Processes. In For Theory Building in 1986 Pa
Archaeology, edited by L. R. Binford, pp. 77-153. Academic Press, New York. Olduvz
Binford,, L. R., M. G. L. Mills, and N. M. Storie 15:673-
1988 Hyena Scavenging Behavior and Its Implications for the Interpretation of 1990 Di
Faunal Assemblages from FLK 22 (the Zinj Floor) at Olduvai Gorge. Journal of Eviden~
Anthropological Archaeology 7:99-135. 316. Ce
Blumenschine, R. J. Bunn,H. T.,
1986 Carcass Consumption Sequences and the Archaeological Distinction of 1988 Vz
Scavenging and Hunting. Journal of Human Evolution 15:639-659. ing and
1988 An Experimental Model of the Timing of Hominid and Carnivore Influence Bunn, H. T.,
on Archaeological Bone Assemblages. Journal of Archaeological Science 15:'483- K. Behrensrr
502. 1980 Fx
1991 Hominid Carnivory and Foraging Strategies, and the Socio-Economic 12:109-
Function of Early Archaeological Sites. Philosophical Transactions of the Royal Bunn,H. T.,
Society (London) B 334:211-221. 1986 Sy
Blumenschine, R. J., H. T. Bunn, M. M. Selvaggio, and C. Saanane Tanzan
1990 Site Formation and Hominid Behavior at FLK Zinjanthropus: Preliminary 1988 Re
Refinements Based on Long Bone Modification. Paper presented at the Sixth ogy 29:1
International Conference of the International Council for Archaeozoology, Capaldo, S. l
Washington, D.C. 1990 Di
Blumenschine, R. J., and M. M. Selvaggio Camivc
1988 Percussion Marks on Bone Surfaces as a New Diagnostic of Hominid Confen
Behaviour. Nature 333:763-765. j Casteel, R. V
1991 On the Marks of Marrow Bone Processing by Hammerstones/and Hyenas: 1971 Di
Their Anatomical Patterning and AFchaeological Implications. In Cultural Begin- 469.
nings: Approaches to Understanding Early Hominid Life-ways in the African Savanna, Crader, D. C
edited by J. D. Clark, pp. 17-32. Dr. Rudolf Habelt GMBH, Bonn. 1983 Re
Bonnichsen, R. the Arc
1973 Some Operational Aspects of Humq.n and Animal Bone Alteration. In Prey, ec
Mammalian Osteo-archaeology: North America, edited by B. M. Gilbert, pp. 9-26. Series '1
University of Missouri Archaeological Society, Columbia. Gifford, D. F
1983 The Broken Bone Controversy: Some Issues Important for the Study of 1980 EtJ
Early Archaeological Sites. In Carnivores, Human $cavengers, and Predators: A Fossils 1
Question of Bone Technology, edited by G. Lemoine and A. S. MacEachern, pp. Univen
271-284. University of Calgary Archaeological Association, Calgary. Gifford-Con
Brain, C. K. 1990 EtJ
1967 Hottentot Food Remains and Their Bearing on the Interpretation of Fossil from Ec
Bone Assemblages. Scientific Papers of the Namib Research Station 32:1-7. pp. 179-
1969 The Contribution of the Namib Desert Hottentots to an Understanding of Haynes, G.
Australopithecine Bone Accumulations. Scientific Papers of the Namib Research 1980 Ev
Station 39:13-22. Paleobio
1981 The Hunters or the Hunted? An Introduction to African Cave Taphonomy. 1982a Pn
University of Chicago Press, Chicago. Bone Si
Bunn,H. T.
r'ork. 1981 Archaeological Evidence for Meat-Eating by Plio-Pleistocene Hominids
vYork. from Koobi Fora and Olduvai Gorge. Nature 291:547-577.
io-Pleistocene 1983 Comparative Analysis of Modern Bone Assemblages from a San Hunter-
Gatherer Camp in the Kalahari Desert, Botswana, and from a Spotted Hyena
ents and Inter- Den Near Nairobi, Kenya. In Anir.nals and Archaeology, val. 1: Hunters and Their
Prey, edited by J. Clutton-Brock and C. Grigson, pp. 143-148. BAR International
Series 163. Oxford.
ry Building in 1986 Patterns of Skeletal Representation and Hominid Subsistence Activities at
.Jew York. Olduvai Gorge, Tanzania, and Koobi Fora, Kenya. Journal of Hur.nan Evolution
15:673-690.
:erpretation of 1990 Diagnosing Plio-Pleistocene Hominid Activity with Bone Fracture
)rge. Journal of Evidence. In Bone Modification, edited by R. Bonnichsen and M. H. Sorg, pp. 299-
316. Center for the Study of the First Americans, Orono, Maine.
Distinction of 1988 Variability in Bone Assemblage Formation from Hadza Hunting, Scaveng-
ing and Carcass Processing. Journal of Anthropological Archaeology 7:412-457.
.vore Influence Bunn, H. T., J. W. K. Harris, G. Isaac, Z. Kaufulu, E. Kroll, K. Schick, N. Toth, and A.
3cience 15:483- K. Behrensmeyer
1980 FxJj 50: An Early Pleistocene Site in Northern Kenya. World Archaeology
>cia-Economic 12:109-136.
f!S of the Royal Bunn, H. 1 ., and E. M. Kroll
1986 Systematic Butchery by Plio/Pleistocene Hominids at Olduvai Gorge,
Tanzania. Current Anthropology 27:431-452. ,
s: PreliminarY-" 1988 Reply to Fact and Fiction about the Zinjanthropus Floor. Current Anthropol-
~d at the Sixth ogy 29:135-149.
:haeozoology, Capaldo, S. D.
l990 Differential Treatment of Axial and Appendicular Elements by Scavenging
Carnivores at Simulated Archaeological Sites. Paper presented at the Biennial
c of Hominid Conference of the Society for Africanist Archaeologists, Gainesville, Florida.
Casteel, R. W.
~s and Hyenas: 1971 Differential Bone Destruction: Some Comments. Ar.nerican Antiquity 36:466-
Cultural Begin- 469.
frican Savanna, Crader, D. C.
·1983 Recent Single-Carcass Bone Scatters and the Problem of "Butchery'' Sites in
the Archaeological Record. In Animals and Archaeology, vol. 1: Hunters and Their
Alteration. In Prey, edited by J. Clutton-Brock and C. Grigson, pp. 107-141. BAR International
Jert, pp. 9-26. Series 163. Oxford. ·
Gifford, D. P.
· the Study of 1980 Ethnoarchaeological Contributions to the Taphonomy of Human Sites. In
:d Predators: A Fossils in the Making, edited by A. K. Behrensmeyer and A. Hill, pp. 93-106.
tcEachern, pp. University of Chicago Press, Chicago.
y Gifford-Gonzalez, D.
1990 Ethnographic Analogues for Interpreting Modified Bones: Some Cases
:ation of Fossil from East Africa. In Bone Modification, edited by R. Bonnichsen and K M. Sorg,
.:1-7. pp. 179-246. Center for the Study of the First Americans, Orono, Maine.
ierstanding of Haynes, G. .
Jar.nib Research 1980 Evidence of Carnivore Gnawing on Pleistocene and Recent Mammal Bones.
Paleobiology 5:341-351. ·
ue Taphonor.ny. 1982a Prey Bones and Predators: Potential Ecologic Information from Analysis of
Bone Sites. Ossa 7:75-97.
I
l -
2981 R. J. Blumen$chine and C. W. Marean.
1982b Utilization and Skeletal Disturbances of North American Prey Carcasses. Lyon,P.
Arctic 35:266-281. · 1970 Dif
1983 A Guide for Differentiating Mammalian Carnivore Taxa Responsible for uity 35:2
Gnaw Damage to Herbivore Limb Bones. Paleobiology 9:164-172. Maguire, J. N
Henschel, J. R, R Tilson, and F. von Blottnitz · 1980 Th1
1979 Implications of a Spotted Hyena Bone Assemblage in the Namib Desert.
South African Archaeological Bulletin 34:127-131. .-, or Hillv.
Hill, A. Marean, C. V'
1979a ·Butchery and Natural Disarticulation: An Investigatory Technique. Ameri- 1991 Me
can Antiquity 44:739-747. Assemb
1979b Disarticulation and Scattering of Mammal Skeletons. Paleobiology 5:261-274. Marean, C. V'
1983 Hyenas and Early Hominids. In Animals and .Archaeology, vol. 1: Hunters and 1991' ImJ
Their Prey, edited by J. Clutton-Brock and C. Grigson, pp. 87-92. BAR Interna- Abunda
tional Series 163. Oxford. Marean, C. V'
1990 Bone Modification by Modem Spotted Hyenas. In Bone Modification, edited 1992 Ca·
by R Bonnichsen and E. M. Sorg, pp. 169-178. Center for the Study of the First Olduvai
Americans, Orono, Maine. · Morlan,R. E.
Hill, A., and A. K. Behrensmeyer 1984 To1
1984 Disarticulation Patterns of Some Modem East African Mammals. Paleobiol- Alterati1
ogy 10:366-376. ·
1985 Natural Disarticulation and Bison Butchery. American Antiquity 50:141-145. Myers, T. P.,
Hughes, A. R 1980 Spi
1954 Hyaenas Versus Australopithecines as Agents of Bone Accumulation. Antiqu#
American Journal af Physical Anthropology 12:467-486. Noe-Nygaarc
Isaac, G. L. 1977 Bul
1967 Towards the Interpretation of Occupation Debris: Some Experiments and ical Dep
Observations. Kroeber Anthropological Society Papers 37:31-57. 1989 Ma
1983 Bones in Contention: Competing Explanations for the Juxtaposition of O'Connell, J.
Early Pleistocene Artifacts and Faunal Remains. In Animals and Archaeology, vol. 1988 Ha
1: Hunters and Their Prey, edited by J. Clutton-Brock and C. Grigson, pp. 3-19. Implic~l
BAR International Series 163. Oxford. 1989 Rec
Johnson, E. nal of Ar
1985 Current Development in Bone Technology. In Advances in Archaeological
Method and Theory, vol. 8, edited by M. B. Schiffer,pp. 157-235. Acaq~mic Press, Potts, R.
NewYork. ' 1983 POI
Kent, S. Tanzani
1981 The Dog: An Archaeologist's Best Friend or Worst Enemy. The Spatial Clutton·
Distribution of Faunal Remains. Journal of Field Archaeology 8:367-372. 1984 Ho
Klein, R. G. Potts, R. B., a
1975 Paleoanthropological Implications of the Non-Archaeological Bone 1981 Cu
Assemblage from Swartlip I, Southwestern Cape Province, South Africa. Quater- Nature'2
nary Research 5:275-288. Read-Martin,
1980 The Interpretation of Mammalian Faunas from Stone-Age Archaeological 1975 Au
Sites, with Special Reference to Sites in the Southern Cape Province, South Faunal.1
Africa. In Fossils in the Making, edited by A. K. Benrensmeyer and A. P. Hill, pp. Richardson, J
223-246. University of Chicago Press, Chicago.
Klein, R. G., and K. Cruz-Uribe 1980 Ca1
1984 The Analysis of Animal Bones from Archeological Sites. University of Chicago Paleonto,
Press, Chicago. Sadek-Kooro
Leakey, M. D. 1972 Pri
1971 Olduvai Gorge, vol. 3, Excavations in Beds I and II, 1960-1963. Cambridge 37:369-3
University Press, Cambridge. 1975 Inb
Lyman, R. L. Studies,,
1984 Bone Density and Differential Survivorship of Fossil Classes. Journal of , Scott, L., and
Anthropological Archaeology 3:259-299. 1981 AI
2y Carcasses. Lyon,P.
1970 Differential Bone Destruction: An Ethnographic Example. American Antiq-
sponsible for · uity 35:213-215.
Maguire, J. M., D. Pemberton, and H. M. Collett
1980 The Makapansgat Limeworks Grey Breccia: Hominids, Hyaenas, Hystricids
amib Desert.
or Hillwash? Paleontologia Africana 23:75-9-8.
Marean, C. W.
nique. Ameri- 1991 Measuring the Postdepositional Destruction of Bone in Archaeological
.Assemblages. Journal of Archaeological Science 18:677-694.
·gy 5:261-274. Marean, C. W., and L. M. Spencer
.:Hunters and 1991 Impact of Carnivore Ravaging on Zooarchaeological Measures of Element
BAR Interna- Abundance. American Antiquity 56:645-658.
Marean, C. W., L. M. Spencer, R. J. Blumenschine, and S. D. Capaldo
cation, edited 1992 Captive Hyena Bone Choice and Destruction, the Schlepp Effect, and
ly of the First Olduvai Archaeofaunas. Journal of Archaeological Science 19:101-121.
Morlan, R. E.
1984 Toward the Definition of -Criteria for the Recognition of Artificial Bone
lals. Paleobiol- Alterations. Quaternary Research 22: 1'60-171.
Myers, T. P., M. Voorhies, and R. G. Corner
If 50:141-145.
1980 Spiral Fractures and Bone Pseudotools at Paleontological Sites. American
:cumulation. Antiquity 45:483-490.
Noe-Nygilard, N.
1977 'Butchering and Marrow Fracturing as a Taphonomic Factor in Archaeolog-
eriments and ical Deposits. Paleobiology 3:218-237.
1989 Man-made Trace Fossils on Bone. Human Evolution 4:461-491.
taposition of O'Connell, J. F., K. Hawkes, and N. Blurton Jones
:haeology, vol. 1988 Hadza Hunting, Butchering, and Bone Transport and Their Archaeological
on, pp. 3-19. Implications. Journal of Anthropological Research 44:113-161.
1989 Reanalysis of Large Mammal Body Part Transport among the Hadza. Jour-
nal of Archaeological Science 17:3"01-316.
4.rchaeological Potts, R.
1demic Press,
1983 Foraging for Faunal Resources by Early Hominids at Olduvai Gorge,
Tanzania. In Animals and Archaeology, vol. 1: Hunters and Their Prey, edited by J.
. The Spatial Clutton-Brock and C. Grigson, pp. 51-62. BAR International Series 163. Oxford.
'2. 1984 Home Bases and Early Hominids. American Scientist 72:338-347.
Potts, R. B., and P. Shipman
ogical Bone 1981 Cutmarks Made by Stone Tools on Bones from Olduvai Gorge, Tanzania.
.frica. Quater- Nature 291:577-580. ·
Read-Martin, C. E., and D. W. Read
rchaeological 1975 Australopithecine Scavenging and Human Evolution: An Approach from
·vince, South Faunal Analysis. Current Anthropology 16:359-368.
1\. P. Hill, pp. Richardson, P. R. K.
1980 Carnivore Damage to Antelope Bones and Its Archaeological Implications.
y of Chicago Paleontologia Africana 23:109-125.
Sadek-Kooros, H.
1972 Primitive Bone Fractures: A Method of Research. American Antiquity
3. Cambridge 37:369-382. .
1975 Intentional Fracturing of Bone: Description of Criteria. In Archaeozoological
Studies, edited by A. T. Clason, pp: 139-150. Elsevier, New York.
es. Journal of Scott, L., and R. G. Klein ·
1981 A Hyena-accumulated Bone Assemblage from Late Holocene Deposits at
I
I
___l__
Deelpan, Orange Free State. Annals of the South African Museum 86:217-227.
Selvaggio, M. M.
1990 Scavenging from Carnivores c;m the Serengeti: Implications for Early
Hominid Behavior. Paper presented at the Biennial Conference of the Society for
Africanist Archaeologists, Gainesville, Florida.
Shipman, P. L., and J. Phillips
1976 On Scavenging by Hominids and Other Carnivores. Current Anthropology
17:170-172. 17.
Shipman, P. L., and J. Phillips-Conroy
1977 Hominid Tool-making Versus Carnivore Scavenging. American Journal of
Physical Anthropology 46:77-86.
Shipman, P. L., and J. Rose
1983 Early Hominid Hunting, Butchering and Carcass-processing Behaviors:
Simons, J. W. J
1966 The Presence of Leopard and a Study of Food Debris in the Leopard Lairs
of the Mount Suswa Caves, Kenya. Bulletin of the Cave Exploration Group, East
Africa 1:50-69.
Sinclair, A. R. E., and P. Duncan
1972 Indices of Condition in Tropical Ruminants. East African Wildlife Journal
10:143-149.
Skinner, J. D., S. Davis, and G. Ilani
1980 Bone Collecting by Striped Hyenas, Hyaena hyaena, in Israel. Paleontologia
Africana 23:99-104.
Sutcliffe, A. J.
1970 Spotted Hyena: Crusher, Gnawer, Digester and Collector of Bones. Nature
227:1110-1113.
Werdelin, L.
1989 Constraint and Adaptation in the Bone-cracking Canid Osteoborus
(Mammalia: Canidae). Paleobiology 15:387-401.
White, T. E. 1
1952 Observations on the Butchering Techniques of Some Aboriginal Peoples: I.
American Antiquity 17:337-338.
Yellen, J. E.
1977 Cultural Patterning in Faunal Remains: Evidence from the !Kung Bushmen.
In Experimental Archaeology, edited by D. Ingersoll, J. E. Yellen, and W.
Macdonald, pp. 271-331. Columbia University Press, New York.
domestic de
fiable anim
gatherer grc
74% to 97%
in a residen
loss. In a st
(1984, 1985)
notes a size
From Bones to
Faunal Remain
Paper No. 21
reserved. ISH!'
.17-227.
ns for Early
he Society for
Anthropology
17. The Impacts of Domestic Dogs on
-:an Journal of Bone in Forager Camps; or,
The Dog-Gone Bones
.g Behaviors:
'gy 2:57-98. Jean Hudson
-eopard Lairs
n Group, East
Abstract: Domestic dogs have played a role in the destruction and distri-
bution of bone in hunter-gatherer camps for thousands of years, con-
Tildlife Journal tributing to the modification of bone assemblages recovered archaeo-
logically. Ethnoarchaeological research among a modern foraging group,
the Aka of central Africa, provides a controlled situation in which the
. Paleontologia impacts of domestic dogs on zooarchaeological measures, such as MNI,
NISP, and MNE, can be monitored. Results for the set of assemblages
tested indicate that (1) the loss of bone is considerable; (2) both MNI and
Bones. Nature NISP are reasonably successful at ranking taxa with live weights of 1 kg
or more; and (3) MNE measures do not accurately reflect original values
per element, although a ranked pattern of survival of broader categories
d Osteoborus of body parts may be characteristic of canid-ravaged assemblages. This
coarser ranking suggests that heads survive best, limb elements (partic-
ularly limb shafts) survive moderately well, and vertebrae and ribs
1al Peoples: I. survive least well.
mgBushmen. Introduction
len, and W.
In many of the cultural contexts that interest archaeologists,
domestic dogs can be a significant factor affecting the preservation of identi-
fiable animal bone. Ethnoarchaeological research with a modern hunter-
gatherer group, the Aka of the Central African Republic, documents a loss of
74% to 97% of the elements of four key species known to have been deposited
in a residential camp, with smaller-bodied taxa showing a greater degree of
loss. In a study conducted with Alyawara dogs in central Australia, Walters
(1984, 1985) documents a loss of approximately 97% of all bone by count and
notes a size bias in the recovery rate per taxon, with larger species, such as

301
l
3021 ]. Hudson
kangaroo, showing better survival than smaller species, such as lizard and remains de
chicken. Describing the effects of dogs in a Wachipaeri village in Peru, Lyon cui tural bel
(1970) notes that small-bodied animals, such as fish, birds, and small mam- redistributi<
mals, are almost completely removed fron1 the archaeological record and that basis while
larger game survive better than medium-sized game. donment o:
The magnitude of the destructive impact of dogs and its potential effects on niques. Thrc
the relative preservation of different taxa and different skeletal elements has After aba
important implications for archaeological interpretations. Interpretations of layer provE
human subsistence strategies typically derive from-measures of the relative encoun tere<
abuneiance of taxa, such as MNI (Minimum Number of Individuals) and NISP compacted;
(Number of Identified Specimens), and from measures of the relative abun- of the. sites
dance of different skeletal elements, such as MNE (Minimum Number of the bone we:
Elements) and MAD (Minimum Animal Units). Interpretations of settlement organic deb
patterns, such as distinctions between kill sites and residential sites, often extended fr
build upon these measures also. density per
These interpretive impacts have considerable temporal depth and broad unit was ex
geographic relevance, given the an.tiquity of the relationship between dogs situ. All resi
and humans. Evidence from several sites in Europe and the Middle East Bone wa~
suggest that dogs were domesticated during the Upper Paleolithic, around comparativ1
10,000 to 12,000 years ago. Remains identified as dog were recovered from the American 1
Natufian site of Palegawra in Iraq and dated to at least 12,000 years B.P. History. WJ
(Turnbull and Reed 1974). The skeleton of a puppy was found in association three medit
with a human burial at the site of Mallaha in Israel (Davis and Valla 1978). leucogaster, ·
Analysis of dental attributes shows the puppy to have attributes intermediate tion of 2,18L
between those of local wolves and dogs, suggesting that domestication was only, leavin
underway by 10,000 to 12,000 B.P. The context of the find suggests that the or species is
relationship between the human and the canid was 11affectionate rather than Fragment
gastronomic" (Davis and Valla 1978:610). Benecke's (1987) discrhninant func- and evaluat
tion analysis of canid remains from Upper Paleolithic and Mesolithit sites in calculated l
Europe, including Mezin, Dobntz-Kniegrotte, Bonn-Oberkassel, Senckenberg, rating them
and Star Carr, suggests that domestication of the dog was under way by ly evaluatir
10,000 to 12,000 B.P. in Europe as well. 1\1NI repre5
The aim of this paper is to examine the nature of the destructive effects of highest elen
dogs on bone in residential camps. Their impacts on the relative abundance of Excavatic
taxa and the relative frequency of skeletal parts are evaluated, using the depositiona
known ethnographic frequencies to test the accuracy of archaeological mea- paper that 1
sures.1 This case study is then compared with the results of other controlled resident d01
studies of canid damage to assess what, 'if any, archaeologically useful gener- wild canids
alizations can be made. cane rats, T}
scavenger c
patterns an
Background on the Data Set marks on th
very little i~
The data discussed here are the result of an ethnoarchaeological cal forests c
project carried out among the Aka Pygmies, a modern hunter-gatherer group wild scaver
living in the tropical forests of central Africa (Hudson 1990). The primary while locaL
research objective of the project was to document links between faunal abundant i1
Dogs in Forager Camps I 303
; lizard and remains deposited in temporary residential camps and various aspects of
Peru, Lyon cultural behavior involved in their deposition. Hunting, butchering, meat
small mam- redistribution, cooking, consumption, and discard were monitored on a daily
)rd and that basis while the camp was occupied, and all bone remaining after the aban-
donment of the camp was recovered, using standard archaeological tech-
.al effects on niques. Three residential camps were documented in this way .
lem~nts has After abandonment each site was excavated completely. The occupation
'reta tions of layer proved to be very shallow; sterile, undisturbed soil was typically
the relative encountered within the first five centimeters. The soil was fine-grained and
s) andNISP compacted and did not permit much downward movement of the bone. None
lative abun- of the sites had been previously occupied in the recent past. The majority of
Number of the bone was recovered on the surface or from above-surface accumulation of
,f settlement organic debris. A meter grid was set up across the site, and excavations were
sites, often extended from the center of the site outwards, continuing beyond the high-
density peripheral midden until the meter units ceased to yield bone. Each
Land broad unit was excavated by trowel, and the majority of the bone was recovered in
tween dogs situ. All residue was dry-screened through 1/4" mesh.
vfiddle East Bone was identified to the most specific taxonomic level possible, using
:hie, around comparative materials collected while in the field as well as collections at the
~ed from the
Americah Museum of Natural History and the National Museum ofNatural
0 years B.P. Historl Where species level identification was not possible, as between the
. association three medium-sized duiker species, Cephalophus callipygus, C. dorsalis, and C.
Valla 1978). leucogaster, broader categories were used. Fifty-five percent of the total collec-
ntermediate tion of 2,182 fragments was identified to genus or species and 40% to size class
tication was only, leaving 5% of the bone unidentified. Only the bone identified to genus
~sts that the
or species is included in this analysis.
rather than Fragments were also identified according to element, portion, side, and age,
ctinant £unc- and evaluated for evidence of burning, butchering, and gnawing. MNEs were
i thic sites in calculated by laying out all fragments of the element at the same time, sepa-
enckenberg, rating them according to side, size, and age, attempting refits, and then visual-
der way by ly evaluating the minimum number of separate elements represented, The
IMNI represented by the element was calculated at the same time, and the
ve effects of highest element value was taken as the MNI for the taxon.
)Undance of Excavation was begun within a month after site abandonment, and post-
l, using the depositional destructive processes were thus limited. It is a key premise of this
ogical mea- paper that the _primary agents operating to destroy discarded bone were the
r controlled resident domestic dogs. I base this assumption on (1) the lack of hyaenids and
seful gener- wild canids in this tropical forest ecosystem; (2) the apparent local absence of
cane rats, Thryonomys spp.1 the animal identified by informants as the principal
scavenger of bone; (3) observation of the interaction between Aka discard
patterns and dog scavenging behavior; and (4) the frequency of dog gnaw
marks on the bone recovered (14% to 25% of all bone).2 It should be noted that
very little is known of the natural taphonomic processes at work in the tropi-
:haeological cal forests of the West Congo Basin. The study area lacks many of the larger
herer group wild scavengers typical of more open and arid environments. The cane rat,
'he primary while locally notorious for its bone-gnawing habits, did not appear to be
reen faunal abundant in the vicinity of the sites. Its absence was substantiated by the
.r
_j_
3041 f. Hudson
consistent failure of bone-baited traps placed at the sites by Aka at the time of study with
abandonment and by the extremely low incidence of rodent gnaw marks on tive methoc
the bone recovered (one case observed).
Because I was resident in camp throughout the period of occupation, I had
ample opportunity to observe both Aka discard habits and .the scavenging
behavior of their dogs. Aka dogs are relatively small and are used in hunting
to flush game. A typical camp has one or two dogs. They range freely within reliability c
the camp ·and scavenge for their food. Most of this scavenging occurs while :MJ\JI, NISP,
the Aka are eating, as they toss bones to the ground. The dogs will then either graphically
gnaw the bone where it is tossed or remove it to the midden area behind the variation in
huts. In camps occupied for more than a week, the Aka typically sweep the in carcass si
areas inside and to the front and sides of their huts, and most larger bone All three
fragments eventually end up in the midden that rings the camp. game const
The three camps studied were small residential sites occupied by an aver- blages are c
age of six families for a duration of one week to two and a half months. All between ac
faunal remains introduced into tne sites were the direct or indirect result of MNI and r-
Aka hunting. During the dry season, between November and June, coopera.- assemblage
tive net hunts are the primary hunting technique and duikers (Cephalophus hunting cm
monticola, C. dorsalis, C. callipygus, C. leucogaster, and C. silvicultor) are the key was less aff
game species. During the rainy season, from July to October, trapping is the duce. Thos
dominant form of hunting and the catch is more diversified; duikers continue reliability c
to be important, but rodents such as porcupine (Atherurus africanus), forest rat MNE.
(Cricetomys emini), and squirrels (Protoxerus stangeri, Funisciurus pyrrhopus, and
others) are the game caught most frequently.
All of the animals procured during the study period, with the single excep-
tion of one yellow-backed duiker (Cephalophus silvicultor), weighed less than
25 kg and so were small enough to be carried back to camp as complete
carcasses. The Aka camp assemblages thus are not subject to the type of body two measm
part selection that might be expected to occur between kill sites and fesiden- of the anal,
tial camps in the case of larger animals. mize sampi
Other cultural behaviors that result in selective body part transport are at are compar
work in Aka camps, however. Chief among them are redistribution rules different ta:
about cooperatively caught game and patterns of meat trade with neighboring sample of 1
agriculturalists. Meat trade tends to remove high utility body parts from these 16 ca:
camp, although this varies with the size of the game animal (Hudson 1990). actual indh
Redistribution results in a general juggling of parts of varying utility among
cooperating camps. Both meat trade and redistribution rules have greater
effect on body part distribution in net-hunting camps than in trapping camps.
Other factors contributing to the nature of the assemblage of bone discarded
on site by the Aka include their use of metal butchering tools and their Figure 17-1
modern habit of cooking most meat in metal pots (i.e., boiling rather than abundance
roasting). The use of metal tools influences the ways in which bone is varies.· Of tl
butchered, and that, in conjunction with the habit of lengthy stewing, may are documE
affect the relative attractiveness of different bones once discarded and the ease been lost.
with which dogs can destroy bone (see Blumenschine 1988 and Blumenschine In spite c
and Marean, this volume, for a discussion of fat-rich versus fat-depleted bone survival, a
and destruction by wild carnivores). Future comparison of the results of this the correlat
the time of study with other ethnoarchaeological data documenting the effects of alterna-
v marks on tive methods of butchering and cooking would be useful.
:ttion, I had
scavenging
Structure of the Analysis
in hunting The focus of the analysis is on the impacts that dogs have on the
eely within reliability of some of the zooarchaeological measures most commonly used:
:curs while MNI, NISP, and l\1NE. The predictability of the relationships between ethno-
then either graphically documented values and archaeological values is examined, as is
behind the variation in rates of survival and their potential causes in terms of differences
sweep the in carcass size and sample size. .
arger bone All three of the study sites produced controlled data pairing the actual
game consumed with the bone recovered, and data from the separate assem-
JY an aver- blages are combined during analysis of the predictability of the relationships
nonths. All between actual numbers of individuals and the archaeological measures of
~ct result of MNI and NISP. Certain other aspects of the analysis are restricted to the
.e, coopera- assemblage from camp Kl. Because K1 was a trapping camp rather than a net-
:ephalophus hunting camp, it involved the exploitation of a greater diversity of taxa and
are the key was less 1affected by meat trade and the body part biases meat trade can pro-
'Ping is the duce. T,hose two characteristics make it more suitable for illustrating the
rs continue reliability of MNI in ranking taxa and for evaluating the impacts of dogs on
;) , forest rat MNE.
rhopus, and
ngle excep- Relative Taxonomic Abundance

d less than
s complete The effects of Aka dogs on bone is discussed first as it relates to
rpe of body two measures of relative taxonomic abundance, MNI and NISP. For this part
1d residen- of the analysis, data from three different Aka camps a;e combined to maxi-
mize sample size. Controlled data on actual numbers of individual carcasses
;port are at are compared with MNI and NISP values recovered archaeologically for 11
ution rules qifferent taxa; several of these taxa occur at more than one site, yielding a total
leighboring sample of 16 cases. Figures 17-1 and 17-5 indicate the distribution of values for
parts from these 16 cases and the regression line predicted by the relationship between
dson 1990) . actual individuals and the zooarchaeological measures.3 .
.lity among
:tve greater MNI
ing camps.
e discarded Figure 17-1 illustratesthe relationship for MNI. It is clear from
; and their Figure 17-1 and Table 17. . 1 that recovered MNI does underrepresent the actual
~ather than abundance in most of the 16 cases and that the degree of underrepresentation
:h bone is varies. Of the total149 individual animals originally represented at the site, 77
·wing, may are documented by MNI; knowledge of 48% of the original game record has
nd the ease been lost.
lmenschine In spite of that degree of loss and the fact that taxa have different rates of
)leted bone survival, a linear model provides a very good fit between the variables, and
:ults of this the correlation is statistically significant (r = .963, propability of error < .001,
306 IJ. Hudson
Site
Kl
Kl
Kl
Kl
Kl
M6
M6
Kl
M10
Kl
Kl
MlO
Kl
Kl
Kl
MNI MlO
Figure 17-1. Plot of the relationship between MNI and the actual
number of individuals observed ethnographically. Data from three sites:
M6, M10, and Kl. variability i
be influenc:
rates. Exam:
df = 14, N = 16; rs = .918, probability of error< .01). As illustrated at the site K1 MNI survh
(Figure 17-2), although some error would result if 11NI were used to i;nterpret blage by on
subsistence strategies, overall representation is fairly accurate in te~§ of rank by 1v1NI, re,
order and the taxonomically diversified, rather than specialized, nature of the MNI is calo
procurement strategy. one, any ide;
The major source of error apparent in Figure 17-2, the underrepresentation same elemE
of squirrel, may be error of a predictable type, that is, the underrepresentation probability
of very small fauna in assemblages where dogs are operating as agents of carcasses in'
destruction. Controlled feeding experiments conducted by Payne and Munson
(1985) have documented very poor survival of squirrel bone fed to dogs.
However, the Aka data document simil8:rly poor survival for medium duiker,
which are among the largest of the taxa exploited at these sites (Table 17-1).
The picture is further complicated by the higher degrees of survival exhibited
by other taxa of intermediate size. Rodents with live weights somewhat and actual 1
greater than the squirrels' 100 to 600 grams, such as forest rat (1 kg) and to numbers
porcupine (3 kg), show better MNI representation than medium duiker (18 measure of
kg). The relationship between carc~ss size and degree of destruction by dogs contributin!
warrants further examination. representinl
The relationship between animal size, as measured by live weight, and 11NI loss of ideni
survival is plotted in Figure 17-3. It is essentially unpredictable in simple per taxon is
linear terms. Species with live weights of less than 10 kg show extreme N = 16), as i:
Table 17-1. Comparison of MNI and Actual Number of Individuals
Live
Weight Actual Percent
Site Taxon (kg) Individuals MNI Survival
K1 Mongoose 3 1 1 100
K1 Monkey 4.5 1 1 100
K1 Lizard 10 1 1 100
K1 YB duiker 60 1 1 100
K1 Civet 3 2 2 100
M6 Mdduiker 18 3 1 33
M6 Srn duiker 5 3 3 100
K1 Tortoise 1.5 5 2 40
M10 Porcupine 3 8 7 88
K1 Mdduiker 18 9 3 33
K1 Srn duiker 5 10 4 40
M10 Mdduiker 18 13 4 31
K1 Porcupine 3 14 7 50
K1 Squirrel 0.3 15 4 27
K1 Forest rat 1 16 9 56
M10 Srnduiker 5 47 27 57
i the actual
n three sites:
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
t the site K1 :MNI survival, such that species represented in the original discard assem-
to interpret blage by only one or two individuals are more likely to be fully represented
~rms of rank by MNI, regardless of body size. This is not surprising, given the fact that
.ature of the :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
>res entation same element must survive from each. of the original carcasses, and the
>res entation probability that this will occur can be expected to decrease as the number of
$agents of carcasses increases .
.ndMunson
ed to dogs.
ium duiker, NISP
Table 17-1).
al exhibited Figure 17-5 and Table 17-2 illustrate the relationship between NISP
somewhat and actual numbers of individuals. Although NISP is not theoretically linked
(1 kg) and to numbers of individuals in the same way that MNI is, it is often used as a
L duiker (18 measure of relative taxonomic abundance. Of the 149 animals originally
lon by dogs contributing to the assemblage, 898 identifiable fragments were recovered,
representing an average of 6 fragments per carcass. In spite of the degree of
lt, and MNI loss of identifiable bone, the correlation between NISP and actual individuals
e in simple per taxon is statistically significant (r =·.921, probability of error< .001, df = 14,
>w extreme N = 16), as is rank order (rs = .786, probability of error oe:: .01).
308 I J. Hudson
30
25
20
1(
15 !B.
ca
10 ::::J
"'C
:~
"'C
Forest Squirrel Porcupine Blue MediLJll Tortoise Civet Monkey Yellow- Mongoose Lizard
-coc:·
Rat Oui ker Oui k.er backed ::::J
Dui ker .....
0
<(
Figure 17-2. Comparison of percentage contribution of each taxon based ::;;;;.
on MNI and aCtual number of individuals observed ethnographically. z
~
Data from the site Kl.
MNI Survival vs.Body Size I

(n=16)
s
8
100 cvaD LZ YD
(/)
MGMK
ca 90 pp
NISP sur
::::J animals are
"'C /
80 in Figure 17·
:~ in the in teq
"'C
c: 70 1-
overreprese:
co::::J 60 1-
duiker, and
Tortoise is c
+'"" FRBD
0 fragments iJ
<( 50 1- pp
::;;;;. over the otn
z 40 1- TR BD ship betwee
:E MD ability, and
MD
30 1-
MD (r 2 = .162).
SQ
To sumrn
I I
20 NISP demo:
0 20 40 60
actual num~
Live Weight in Kg destructive ,
considerablE
Figure 17-3. Plot of the relationship between MNI survival and carcass a tendency f
size as measured by live weight. both measu
overreprese1
documented
MNI Survival vs.Sample Size

(n=16)
LZ
100 - YD BD
~Kcv
en MG
n; 90- pp
:::::J
"'C 80 r-
:~
"'C
>Se Lizard
-c:: 70 r-
60 r-
FR BD
50- pp
z taxon based
ographically. 40- TR BD
MD
MD MD
30 ;-
sa
I L I I L
20
0 10 20 30 40 50
Sample Size (No. of Actual Individuals)
Figure 17-4. F'lot of the relationship between MNI survival and sample
size as measured by the actual number of individuals observed ethno-
graphically for each taxon.
YD
NISP survival does appear to covary with body size, such that larger
animals are likely to be better represented. This is illustrated to some degree
in Figure 17-6 and more clearly in Figure 17-7, where the chief sources of error
in the interpretation of subsistence strategies at the site K1 result from the
overrepresentation of the two largest taxa, medium duiker and yellow-back
duiker, and the dramatic. underrepresentation of the smallest taxa, squirrel.
Tortoise is also overrepresented, but that is the result of including carapace
fragments in the NISP, a practice that gives tortoise a structural advantage
over the other taxa and can easily be avoided. Figure 17-:S plots the relation-
ship between NISP survival and sample size. Small samples show high vari-
ability, and the overall pattern is not well predicted by a simple linear model
(r 2 = .162).
To summarize, for the sample of 16 cases examined here, both MNI and
60
NISP demonstrate a statistically significant degree of correlation with the
actual numbers of individuals documented ethnographically prior to the
destructive effects of scavenging by dogs. The amount of bone destroyed is
considerable and the accuracy of representation is variable between taxa, with
1and carcass a tendency for artimals of squirrel size to be significantly underrepresented by
both measures, and a further tendency for relatively large animals to be
overrepresented by NISP. It was also noted that, for the range of sample sizes
documented here, v~ry small original numbers of individuals are more likely
_l
.
310 I]. Hudson
1
40
Site
!
:2
30 K1
K1
~ K1
ttl K1
2(.) 20 K1
<(
M6
M6
K1
10 M10
K1
K1
M10
100 200 300 K1
K1
NISP K1
M10
Figure 17-5. Plot of the relationship between NISP and the actual num-
ber of individuals observed ethnographically. Data from three sites: M6,
M10, and Kl.
sections of
to be better represented by :MNI than are larger numbers of individuals and, subdivision:
in general, small samples show greater variability in degree of representation often more
by both :MNI and NISP. / elements an
best survive:
proximal h1
Skeletal Part Frequency tibia. Limb ·
of the limb,
To examine the impact of dogs on skeletal part representation in fragments
an Aka residential camp, four taxa of varying size from the trapping camp K1 (Blumensch
were compared. They were medium duiker (Cephalophus callipygus, C. dorsalis, this volume
and C.leucogaster), with an average live weight of 18 kg; small duiker (the blue are in keepi
duiker, Cephalophus monticola), with an· average live weight of 5 kg; brush- assemblage
tailed porcupine (Atherurus africanus), with a live weight of 3 kg; and giant (Binford 19E
forest rat (Cricetomys emini), with a live weight of 1 kg. Ethnographic observa- CompariE
tions were used to establish the actual numbers of -each element discarded on smaller taxc
site for each of the four taxa, and these were com pared with the MNE values medium du
recovered archaeologically. Element frequencies are summed per taxon in ments of tl-
Table 17-3. · density ofF
The first level of evaluation of body part frequency is based on fairly between sm
detailed distinctions of some 40 skeletal parts, with quantification of parts of ial element~
elements most commonly identified, such as proximal, distal, and shaft the rodent:
Dogs in Forager Camps 1311
Table 17-2. Comparison of NISP and Actual Number of Individuals
Live Survival:
Weight Actual NISP per
Site Taxon (kg) Individuals NISP Individual
K1 Mongoose 3 1 3 3
K1 Monkey 4 1 3 3
K1 Lizard 10 1 7 7
K1 YB duiker 60 1 26 26
K1 Civet 3 2 7 3.5
M6 Mdduiker 18 3 26 8.7
M6 Smduiker 5 3 40 13.3
K1 Tortoise 1 5 46 9.2
M10 Porcupine 3 8 56 7
K1 Mdduiker 18 9 67 7.4
K1 Sm duiker 5 10 51 5.1
M10 Mdduiker 18 13 118 9.1
K1 Porcupine 3 14 77 5.5
K1 Squirrel 0 15 7 0.5
K1 Forest rat 1 16 106 6.6
M10 Smduiker 5 47 258 5.5
1ctual num-
·e sites: M6,
sections of limb bones, four sections of the pelvis, and other similar
lduals and, subdivisions (Table 17-4). The overall degree of loss is marked for most parts,
res entation often more than 50%. There is a considerable degree of variation among
elements and among taxa. In general, cranial and mandibular parts show the
best survival while carpals, tarsals, and phalanges survive poorly, as do the
proximal humerus, distal radius and ulna, and distal femur and proximal
ti8ia. Limb bone shaft fragments are typically the most well represented part
of the limb, offering further support to the claim that identification of shaft
entation in fragments can have a significant effect on skeletal profile studies
tg camp K1 (Blumenschine 1988; Blumenschine and Marean, this volume; Bunn 1986 and
. C. dorsalis, this volume; Bunn and Kroll1986; Marean and Spencer 1991). These patterns
er (the blue are in keeping with what has been documented previously for canid-ravaged
kg; brush- assemblages: differential destruction of the articular ends of limb bones
; and giant (Binford 1981; Brain 1981).
.ic observa- Comparison of the survival of medium duiker elements with that of the
scarded on smaller taxa suggests that carcass size does influence survival, with several
[NE values medium duiker elements surviving at higher rates than the comparable ele-
~r taxon in ments of the smaller taxa. Structural variation in both identifiability and
density of particular elements appears responsible for some of the variation
:l on fairly between small duiker and the two rodents. Artiodactyl metapodials and cran-
of parts of ial elements, for example, show better preservation and identifiability, while
and shaft the rodent limb shaft fragments generally have an a~vantage over those of
3121 ]. Hudson
NISP Survival vs.Body Size 30

(n=16)
28 , . - - - - - - - , - - - - - - - - - - - - - - - - - - - - - - · - - - ------------ ----
25
26 YD
20
24
(/) 15
22
co::s 20
"0 .10
:2!
"0
18
r::: 16
co::s 14
BD
Forest
+"" 12 Rat
0
~ 10
TR
MD
MD
a. 8 PP
~ MD
FR BD LZ
z 6
PPBD
4 CV MK
2 MG
0
sa I I
0 20 40 60 point of vie
" Live Weight in Kg brae and ri1:
and nutrien
Figure 17-6. Plot of the relationship between NISP survival and carcass difficulty o1
size as measured by live weight. greater deg1
While noj
tion of the 1
small duiker. For all three of the smaller taxa, vertebrae and ribs, along with baseline ex!
carpals, tarsals, and phalanges, show very low survival rates. dog-ravage<
To evaluate whether differences in the numbers originally deposited affects
rates of survival, sample size was plotted against percent survival p~f element
for each of the four taxa (Figure 17-9). Small samples are more variable in their
representation, but large samples, although more consistent, are generally
poorly represented. It is certainly not the case that a larger original number of
elements improves the rate of survival.
camps, doc1
In an effort to identify consistent and generalizable patterns in body part
noting the- 1:
representation, elements were collapsed into five body part units, head, chest,
be compare'
back, forelimb, and hindlimb, and the highest survival value for any con-
The dame
stituent element was used to determine the body unit's survival rate (Figure
. several con1
17-10). Although not perfect, there is a general pattern of ranked survival,
wolves (Bin
whereby the head survives best, the fore and hind limb survive moderately
ethnoarchaE
well and to a comparable degree, and the chest and back survive least well bone assem·
and to a comparable degree.
Hudson 199
Anatomy, nutrition, biomechanics, and identifiability may be the factors
and dens (B
behind that patterning. The presence of teeth in cranial and mandibular
All the stud
elements may give the head, as a unit, an advantage in terms of density-
of artiodac1
mediated preservation, distinctiveness of separate MNEs and MNis based on
comparable
aging of dentition, and relative unattractiveness to dogs from a nutritional
variation in
Dogs in Forager Camps j313
Forest Squirrel Porcupine Blue MedilJJl · Tortoise Civet Monkey Yellow- Mongoose Lizard
Rat Dui ker Oui ker backed
Duiker
Figure 17-7. Comparison of percentage contribution of each taxon based

on NISP and actual number of individuals observed ethnographically.
Data from the site Kl.
60 point ot view. Limb bone shafts would have similar advantages over verte-
brae a11d ribs in terms of density and nutrition. The high proportion of fragile
and nutrient-rich cancellous bone in vertebrae and ribs, combined with the
•al and carcass difficulty of identifying these elements to particular taxa, may explain their
greater degree of loss when scavenging canids are present.
While not consistent enough to allow retrodiction of the original composi-
tion of the pre-dog assemblage, this pattern does permit the generation of a
s, along with baseline expectation of the ranking of these body parts for small animals in a
dog-ravaged assemblage. "
osi ted affects
1per element
riable in their
Comparison with Other Studies of Canid
:tre general! y Damage to Bone
:al number of Having briefly summarized- the effects of dogs on bone in Aka
camps, documenting the degree of loss at the level of taxon and element and
in body part noting the biases associated with prey size, the results of this case study can
:, head, chest, be compared with other studies of canid damage.
for any con- The damage done by canids to artiodactyl bone has been documented in
l rate (Figure
several controlled studies. They include feeding experiments with dogs and
<ed survival, wolves (Binford and Bertram 1977; Payne and Munson 1985; Snyder 1988),
~ moderately
ethnoarchaeological studies that combine the effects of humans and dogs on
ve least well bone assemblages (Binford 1978; Binford and Bertram 1977; Brain 1967, 1969;
Hudson 1990)1 and documentation of natural assemblages from wolf kill sites
,e the factors and dens (Binford 1981) and from areas scavenged by foxes (Stallibras 1984).
mandibular All the studies dted provide data on the relative survival of various elements
s of density- of artiodactyl skeletons when acted upon by canids. A combined plot of
Nis based on comparable data from five of those assemblages indicates that the range of
a nutritional variation in percen~ survival per element tends to be l~rge and that patterns of
314j ]. Hudson
NISP Survival vs.Sample Size

(n=16)
28 c - - - - - . - - - - - - - - - - - - - - - - - - - - - - - - - -
26 YD
Taxon
24
f/)
22
ca:s 20
Medium dui
"C Small duiker
:~ 18 Porcupine
"C 16 Forest rat
c
ca:s
-
14
BD
12
( .)
~ 10 TR varies, as 1
a. MD MD
8 MD assemblage,
~ LZ PP FR Density i
z 6 BD pp BD
degree of c
4 MK
CV contexts. B
2 MG
sa survival pe
'---'----'--~-'----'-- L _ .I _____ __j ______ __l_
0
0 10 20 30 40 50 models exi:
(1977) and
Sample Size (No. of Actual Individuals)
weight and
Figure 17-8. Plot of the relationshipbetween NISP survival and sample tion. The th
size as measured by the actual number of individuals observed ethno- although L:
graphically for each taxon. Lyman's bu
isons betwE
element, an~
covariation between studies are not immediately obvious, beyond a broad were recorc
generalization that vertebrae show poor survival relative to crania a11d limbs value of an)
(Figure 17-11). / Lyman's
A wide range of variables are incorporated in those studies, some 'of which consistency
may be responsible for the degree of variation seen. Among them: (1) prey correlation,
size varies between small duiker (@5 kg) and caribou (@100 kg); (2) canid size est correlati
varies between fox and wolf; (3) access to skeleton varies between whole of small du:
carcasses and selected parts; (4) the role of humans varies from considerable and Snyder
(butchering and cooking) to none; (5) there is little control of the age composi- (rs = .608 ar
tion or nutritional status of the prey group; (6) the sample size of prey car- suggesting
casses and body parts varies; and (7) there are differences in the methods used between thE
to calculate survival percentages. Given'the degree of variation, it is perhaps Given th
surprising that some of the studies show as high a correlation as they do in documente<
ranking the survival of elements. retrodicting
Spearman's rank order correlation shows the strongest correlation (rs = .859) in the way
to occur between Snyder's controlled feeding of whole deer to wolves and (1985), ·seen
Brain's study of the effects of dogs on goat in Hottentot sites. Brain's study relatively C(
also correlates reasonably well with Stallibras' s natural assemblage of fox controls rna
scavenging on sheep Crs = .803) and with Binford and Bertram's study of the and differer
effects of dogs on caribou (rs= .746). The causal factors behind these similari- Are there
ties in survival patterns are not obvious, since the size of canid and artiodactyl damage in t
Table 17-3. Actual Number of Elements Versus MNE for Four Taxa
Live Weight Adual Number Percent

Taxon (kg) of Elements MNE Survival
I Medium duiker
Small duiker
18
5
637
1323
167
124
26
9
Porcupine 3 4026 124 3
Forest rat 1 4035 220 5
n'
varies, as does the involvement of humans in the accumulation of the

assemblages.
Density is a reasonable candidate for a common factor that could link the
degree of destruction of various elements to a range of canids, prey, and
contexts. Binford and Bertram (1977) use it with some success to model
survival percentages of elements of both caribou and sheep. Several density
I.
)0 models exi~t, including tp.ose derived separately by Binford and Bertram
(1977) and; by Behrensmeyer (1975); which are based on measurements of
weight and volume, and that derived by Lyman (1984), using photon absorp-
1d sample tion. The three approaches differ in their exact methods and in their results,
r;d ethno- although Lyman's and Behrensmeyer's show similarities in overall pattern.
Lyman's bulk density measures are used in this analysis. To facilitate compar-
isons between his density measures-~- which were taken at several sites per
element, and the controlled experimental and ethnoarchaeological data, which
a broad were recorded per element, each element was assigned the highest density
nd limbs value of any of its sites.
Lyman's approximation of bulk density, based on deer and checked for
Jf which consistency with several other artiodactyl species, was tested for rank order
(1) prey correlation with the six canid-damaged assemblages reviewed here. The high-
anid size est correlation occurs between this density measure and the Aka assemblage
:n whole of small duiker bone ravaged by dogs Crs = .788). Brain's goat-and-dog study
;iderable and Snyder's deer-and-wolf study also show some correlation with density
:omposi- (rs = .608 and rs = .599, respectively), but the remaining studies show little,
Jrey car- suggesting that density is not a major cause of the similarities that do occur
ods used between the various canid studies.
perhaps Given the range of variation in percent survival of particular elements
ey do in documented by these six cases of canid damage to artiodactyls, attempts at
retrodicting skeletal part frequencies that characterize pre-canid assemblages,
:rs = .859) in the way modeled by Gilbert and Singer (1982) and addressed by Walters
lves and (1985), seem premature. Better control of the many variables at work, even in
1's study relatively controlled studies, and repeated studies within this framework of
~e of fox controls may, in the future, clarify the reasons for the patterns of similarity
iy of the and difference seen in studies conducted to date.
similari- Are there some useful generalizations that can be made regarding canid
tiodactyl damage in the meantime?
1
316j ]. Hudson
Table· 17-4. Survival per Element and Part for Four Taxa
Percent Surviving Archaeologically
Elementa Partb
Medium
Duiker
Small
Duiker Porcupine
Forest
Rat
Mean
Survival
Rate -
'U
Q.)
'Ci)
0
a.
Q.)
CRN ANY 67 75 33 20 49 0
MND ANY 50 38 27. 43 40
~
TIH ANY 35 29 12 19 24
m
ATV ANY 50 12
.!:
AXV ANY 50 7 14 0)
ocv ANY 40 10 ·;::
THV ANY 21 3 2 7 0
LMV ANY 15 7 1 6 (/J
:I-
SAC ANY 33 8 Q.)
CAV ANY .c
SCA GLN 50 15 4 7 19 E
::s
SCA BLD 50 23 15 13 25 z
HUM PRX
HUM SHF 67 8 11 20 27
HUM DST 47 15 4 17 13
RAD PRX 31 22 27 20
RAD SHF 17 48 22 23 28
RAD DST 4 3 2
ULN PRX 33 23 11 43 28
ULN SHF 23 11 40 19
ULN DST 3 1
CRP ANY
MTP PRX 18 16 7 10
MTP SHF 27 36 1 8 18
MTP
PHL
INN
INN
INN
DST
ANY
ACE
ILl
ISH
6
40
12
8
1
7
11
4
8
20
17
20
5
2
7
17
8
-
"0
Q)
'Ci)
0
a.
Q)
INN PUB 8 4 23 9 0
FEM PRX 4 7 3 ~
FEM SHF 17 15 20 13 m
t:
FEM DST
0')
TIB PRX ·;::
TIB SHF 40 17 33 40 33 0
TIB DST 4 20 6 U)
FIB ANY 11 3 :I-
Q.)
TRS ANY 3 1 1 .c
RIB ART 24 9 3 2 10 E
RIB BLD 74 13 2 2 23 ::::::1
z
acRN = cranial, MND = mandible, TIH = tooth, ATV = atlas, AXV = axis, OCV = other cervical
vertebra, THV = thoracic vertebra, LMV = lumbar vertebra, SAC = sacrum, CA V = caudal vertebra,
SCA = scapula, HUM = humerus, RAD = radius, ULN =ulna, CRP = carpal, MTP = metapodial, PHL =
phalanx, INN = innominate; FEM = femur, TIB =tibia, FIB = fibula, TRS = tarsal, RIB = rib.
b ANY = any part (unspecified), GLN = gle~oid, BLD = blade, PRX = proximal, SHF = shaft, DST =
distal, ACE = acetabulum, ILI = ilium, ISH = ischium, PUB = pubis, ART = articular (vertebral)
extremity.
'li0
.~ ~
I
r
I
Dogs in Forager Camps j317
Sample Size vs. Percent Survival (MNE}

Medium Duiker
120
110 RIB
Mean "0
Q)
Survival ..... 100
Rate 'Cii
0 90
0.
I
Q)
49 0 80
40
~ 70
24 '@ PHL
12 I· .5 60
14 0)
10 'i: 50
7 0 THV
6 f/) 40
~ CRP·
8 Q)
.a 30 TRS /
19 E 20 LMV CEV
::I
25 z 10 MTP HUM
FIB FEM RAD ULN iNN TIB CRN SCA
27 O L _ _ L_ __ L_ __ J_ _ _ _~~~--_L_ __ L_ __ J_ _ _ _L __ _~_
13 0 0.2 0.4 0.6 0.8
20
28 Proportion Surviving Archaeologically
2
28
19 Sample Sizevs.Percent Survival (MNE}
1 Porcupine
700~-------------------------------------
10
18 PHL
5
2
-g..... 600
7 'Cii
0
17 0. 500
8 Q)
9 0
3 ~ 400 -
13 m
c:
RIB
0)
'i: 300
33 0 MTP
6 f/)
~
3 Q) 200 THV
1 .a CRP
10 E TRS
23 ::l 100 CEV
z LMV FJB HUM INN
1er cervical ULN SCA FEM RAD TIB CRN j __
0
al vertebra, 0 0.1 0.2 0.3 0.4 0.5
,dial,PHL=
Proportion Surviving Archaeologically
haft, DST =
(vertebral)
Figure 17-9. Plots of the relationship between MNE survival and sample
size for each of four taxa at the site Kl. Sample size measured by the
actual number of elements observed ethnographically.
~
\1
,1)
318j ]. Hudson
· Sample Size vs. Percent Survival (MNE) 100

Blue Duiker 90
180,----------------------------
80
RIB
160 70
-
'0
Cl)
PHL 60
'Ci)
140 50
0
c.
Q) 40
0 120
30
THV
?
coc 100 20
10
0')
·;::
0 He
en 60
J...
Cl)
.c 40
E CEV LMV
:l
z MTP
20 HUM
FIB INN TIB ULN RAD
FEM SCA
0 0.2 0.4 0.6
Proportion Surviving Archaeologically
Sample Size vs.Percent Survival (MNE)

Forest Rat
800,----------------------------
CRN MNO
PHL
-
"0 700
(!)
'Cii
0 600
c.
Cl)
c 500
?
co 400
.5
0')
·;:: RIB
0 300
f/)
J...
Cl) TRS
.0 200 CAP THV
E MTP
:l
z 100 CEV
LMV RAD and such e
0
FIB SCA FEM HUM INN TIB ULN
' - - - - L - - - - - _ _ _ _ l __ _ _ _ _L . __ _ _ _---L.-_ _ _ _ _ L
CAN
_ _ _ _ _J_ _ _ _ J. __ _ tions. Ofter
0 0.2 0.4 0.6 our method
Proportion Surviving Archaeologically The Aka
bone in an
ing the imr
Figure 17-9.-Continued. fauna can·
1
Ir
100
90
80
70 ~Medium Duiker
60 c:::::===:::J Bl ue Du i ke r
50 -Porcupine
40 ~Forest Rat
30 -Mean
20
10
Head Chest Back Foret irrb HincH imb
Figure 17-10. Comparison of percent survival of each of five broad body

part categories (head, chest, back, forelimb, and hindlimb) for four taxa.
CRN
_L
0.8
CRN MND ATV AXV CEV THV LMV RIB SCA HUM RAD ULN MTC INN FEH TIB MTT PHL
- o - - Goat/fox -+-small Duiker/Dog -D--Goat/Dog
~Hediun Duiker/Dog - X - caribou/Dog -o--- Deer;~ol f
Figure 17-11. c;omparison of percent survival per element in canid-

ravaged assemblages of artiodactyl bone. Data derived from Brain (1967,
1969); Binford and Bertram (1977); Hudson {1990); Snyder (1988); and
Stallibras (1984).
Conclusions
Controlled studies are well suited to producing "cautionary tales,"
and such cautions are important in helping us avoid unwarranted assump-
CRN tions. Often, however, they stop short of providing guidelines for improving
0.6 our methods to achieve more useful results.
The Aka data suggest that while dogs can do a great deal of damage to the
bone in a residential camp, :MNI and NISP still work reasonably well in rank-
ing the hnportance of taxa with live weights of 1 kg or more. Since smaller
fauna can be expected to be underrepresented, alternative approaches to
320 I J. Hudson
recognizing and quantifying their contribution, such- as fine-mesh sampling gnaw dame:
and flotation, should be pursued. Those methods improve the odds of recov- question co1
ering any small, identifiable elements that have survived. . and collectil
The degree of variation in skeletal part survival suggests that archaeological 3. The fol
frequency values, such as MNE, should not be viewed as direct reflections of Taxa
original skeletal profiles in canid-ravaged assemblages. However, based on BD =blued
the Aka example and other controlled studies, some general patterns of CV =civet (
destruction can be predicted. They can be used to suggest a baseline of FR =forest 1
expected survival frequencies on an ordinal scale: the preferential survival of LZ =lizard
heads, the loss of the articular ends of limb bones, and of carpals, tarsals, MD=medil
phalanges, and the underrepresentation of vertebrae and ribs. Observed devi- MG=mong
ations from this baseline may strengthen arguments that other agents, such as MK=monk
particular aspects of cultural behavior, were at work. PP = porCUf
Given the expected loss of articular ends, additional attention to the identi- SQ = squirn
fication of shaft fragments is likely to lead to more accurate representation of TR = tortois
the relative frequency of appendicular elements. Interpretations of site func- YD = yelloVI
tion and patterns of transport and- trade may also ·be improved by lumping Elements
individual elements and parts of elements into larger body part units. ATV =atlas
Contrasts of heads, backs, and limbs or of axial ·and appendicular parts, when ATX =axis·
based on the most abundant constituent element, may prove more robust and CAV = cauc
better serve to document cultural behavior patterns. CRN-= cran
Finally, it should be kept in mind that some of the ideas presented here are CRP =carp<:
derived from a single case study and so are in need of testing against a larger FEM=femt
population of controlled assemblages. FIB= fibula
HUM=hur:
INN= innOJ
Acknowledgments LMV=luml
MND=mru
I would like to thank the many people who made this project MTP=metc:
possible. The fieldwork was funded by the Leakey Foundatiod and the
University of California at Santa Barbara. Permission to conduct the research
was granted by the Ministere de la Recherche Scientifique et Technique of the
Central African Republic. Faunal identification at the American Museum of Behrensmey
Natural History was supported by a grant from the museum. And a great deal 1975 Tl
of amused goodwill was granted me by the Aka, who allowed me to share blages
their life and their garbage. I would also like to thank Mark Aldenderfer, Kay 146(10)
Behrensmeyer, and the two anonymous reviewers for their helpful comments Benecke,N.
on an earlier version of this paper. 1987 St
ological
Binford, L. I
Notes 1978 N1
1981 Bo
1. Other studies that have focused on the general problem of the accuracy of these Binford, L. I
measures, without special attention to the role of dogs, include Breitburg's (1991) 1977 Be
controlled comparison of historically documented frequencies with archaeological Buildin·
MNI and NISP and Grayson's (1978, 1979, 1981, 1984) investigations of sample size Blumenschb
bias and the relationship between MNI and NISP. 1988 A1
2. Humans are also potential gnawers of bone, as Kay Behrensmeyer pointed out on Arc
during the From Bones to Behavior conference. No controlled data on the kinds of 502.
rI
I
~'
Dogs in Forager Camps J321
sampling gnaw damage the Aka themselves may do is currently available. However, this
; of recov- question could be addressed in the future by intervening in the Aka discard process
and collecting bone after consumption but before the dogs are given access to them.
aeological 3. The following is a key to the abbreviations used in Figures 17-1 through 17-11:
lections of Taxa
based on BD =blue duiker (Cephalophus monticola)
atterns of CV = civet (Nandinia binotata)
aseline of
,urvival of
ls, tarsals,
II
.. l
FR = forest rat (Cricetomys emini)
LZ =lizard (Varanus niloticus)
MD= medium duikers (Cephalophus callipygus, C. dorsalis, C. leucogaster)
rved devi- I MG = mongoose (Bdeogale nigripes)
ts, such as MK =monkey (Cercopithecus cephus)
PP = porcupine (Atherurus africanus)
the identi- SQ = squirrel (Protoxerus stangeri, Funisciurus pyrrhopus)
TR = tortoise (Kinixys erosa)
~ntation of
YD =yellow-backed duiker (Cephalophus silvicultor)
·site func-
r lumping Elements
~art units. ATV = atlas vertebra OCV =other cervical vertebra
uts, when ATX = axis vertebra PHL = phalanx
·obust and CAV = caudal vertebra RAD= radius
CRN = crahial vertebra RIB= rib
d here are CRP =carpal SAC=sacrum
lSt a larger FEM =femur SCA = scapula
FIB= fibula THV = thoracic vertebra
HUM = humerus TIB =tibia
INN = innominate TRS =tarsal
LMV = lumbar vertebra TTH =tooth
MND = mandible ULN=ulna
lis project- MTP = metapodial
1 and the
e research References
que of the
fuseum of Behrensmeyer, A. K.
great deal 1975 The Taphonomy and Paleoecology of Plio-Pleistocene Vertebrate Assem-
e to share blages East of Lake Rudolf, Kenya. Bulletin of the Museum of Comparative Zoology
.erfer, Kay 146(10):473-578.
comments Benecke, N.
1987 Studies on Early Dog Remains from Northern Europe. Journal of Archae-
ological Science 14:31-49.
Binford, L. R.
1981 Bones: Ancient Men and Modern Myths. Academic Press, New York.
tcy of these Binford, L. R., and J. R. Bertram
trg's (1991) 1977 Bone Frequencies and Attritional Processes. In Background Studies for Theory
haeological Building, edited by L. R. Binford, pp. 77-153. Academic Press, New York
;ample size Blumenschine, R. (
1988 An Experimental Model of the Timing of Hominid and Carnivore Influence
)Ointed out on Archaeological Bone Assemblages: Journal of Archaeological Science 15:483-
he kinds of 502.
3221 J. Hudson
Brain, C. K. Snyder, Lyn
1967 Hottentot Food Remains and Their Bearing on the Interpretation of Fossil 1988 A
Bone Assemblages. Scientific Papers of the Namib Desert Research Station 32:1-11. tailed l
1969 The Contributions of Namib Desert Hottentots to an Understanding of letter 1
Australopithecine Bone Accumulations·. Scientific Papers of the Namib Desert Knox vi
Research Station 39:13-22. Stallibras, S.
1981 The Hunters or the Hunted. University of Chicago Press, Chicago. 1984 Tr
Breitburg, E. Post-gli
1991 Verification and Reliability of NISP and MNI Methods of Quantifying Europe,
Taxonomic Abundance: A View from Historic Site Zooarchaeology. In Beamers, nationa
Bobwhites, and Blue-Points, edited by J. Purdue, W. Klippel, and B. Styles, pp. Turnbull, P.
153-162. Illinois State Museum Scientific Papers 23. Springfield. 1974 Tb
Bunn,H. T. AnthroT
1986 Patterns of Skeletal Representation and Hominid Subsistence Activities at Walters, Ian
Olduvai Gorge, Tanzania, and Koobi Fora, Kenya. Journal ofHuman Evolution 1984 Gc
15:673-690. . site. Ma
Btinn, H. T., and E. M. Kroll 1985 Bo
1986 Systematic Butchering by Plio/Pleistocene Hominids at Olduvai Gorge, 642-64~
Tanzania. Current Anthropology 27:431-452.
Davis, S. J. M., and F. R. Valla
1978 Evidence for Domestication of the Dog 12,000 Years Ago in the Natufian of
Israel. Nature 276:608-610.
Gilbert, A. S., and B. H. Singer
1982 Reassessing Zooarchaeological Quantification. World Archaeology 14:21-40.
Grayson, D. K.
1978 Minimum Numbers and Sample Size in Vertebrate Faunal Analysis. Amer-
ican Antiquity 43:53-65.
1979 On the Quantification of Vertebrate Archaeofaunas. In Advances in Archae-
ological Method and Theory, val. 2, edited by M. B. Schiffer, pp. 199-237. Academic
Press, New York.
1981 The Effects of Sample Size on Some Derived Measures in Vertebrpte Faunal
Analysis. Journal of Archaeological Science 8:77-88.
1984 Quantitative Zooarchaeology. Academic Press, New York.
Hudson, J.
1990 Advancing Methods in Zooarchaeology: An Ethnoarchaeological Study among the
Aka Pygmies. Ph.D. dissertation, Univ~rsity of California at Santa Barbara.
University Microfilms, Ann Arbor.
Lyman, R. L.
1984 Bone Density and Differential Survivorship of Fossil Classes. Journal of
Lyon,P.
1970 Differential Bone Destruction: An Ethnographic Example. American Antiq-
uity 35:213-215.
Marean, C. W., and L. M. Spencer
Abundance. American Antiquity 56(4):645-658.
Payne, S., and P. Munson
1985 Ruby and How Many Squirrels? In Paleobiological Investigations: Research
Design, Methods, and Data Analysis, edited by N. R. J. Fieller, D. D. Gilbertson,
and N. G. A. Ralph, pp. 31-48. American Elsevier, New York.
n•. ." .
r'
II
I Snyder, Lynn
1988 A Controlled Feeding Study Involving Gray Wolf (Canis lupus) and White-
on of Fossil
'1 32:1-11. I tailed Deer (Odocoileus virginianus). Tennessee Anthropological Association News-
3tanding of letter 13(3):1-9. Department of Anthropology, University of Tennessee,
;zmib Desert Knoxville.
Stallibras, S. .
1984 The Distinction Between the Effects of Small Carnivores and Humans on
Post-glacial Faunal Assemblages. In Animals and Archaeology, vol. 4: Husbandry in
2uan tifying Europe, edited by C. Grigson, and J. Clutton-Brock, pp. 259-269. BAR Inter-
In Beamers, national Series 227. Oxford.
Styles, pp. Turnbull, P. F., and C. A. Reed
1974 The Fauna from the Terminal Pleistocene of Palegawra Cave. Fieldiana
Anthropology 63(3):81-146.
A.ctivities at Walters, Ian
zn Evolution 1984 Gone to the Dogs: A Study of Bone Attrition at a Central Australian Camp-
site. Mankind 14(5):389-400. ,
1985 Bone Loss: One Explicit Quantitative Guess. Current Anthropology 26(5):
1vai Gorge, 642-643.
Natufian of
/
1j 14:21-40.
1ysis. A mer-
:s in Archae-
7. Academic
'rate Faunal
fy among the
.ta Barbara.
s. Journal of
~rican Antiq-
. of Element
ms: Research
l
I
Gilbertson,
I
I
I
I
which our k
of skeletal p
1992).
Binford (1
sure of the 1
dactyl taxa«
18. Density-Mediated Attrition of and, especi2
innovative 1
'Bone Assemblages: N_ew Insights sively empl<
emulated dt
R. Lee Lyman Lyman 198!:
frequencies;
the prehistoj
strategy.
Abstract: Previous research showing that frequencies of skeletal parts
often correlate with bone density and thus suggesting that density- Another E
mediated attrition has occurred is substantiated by correlations between structural d
bone density and skeletal part frequencies in 20 assemblages. Half of the taphonomic
87 total assemblages that have now been s1:1bjected to such analysis seem to remove t
to have undergone some form of density-mediated attrition, and only often from'
about 32% of those assemblages correlate with and thus might be parts with h
explained by the MGUI. Carnivore-gnawed assemblages do not always be argued th
correlate with density, suggesting that gnawing must attain a threshold between b01
of damage for bone frequencies in carnivore-ravaged assemblages to responsible
correlate with bone density. Amounts of within-bone nutrients such as muted or m1
grease and marrow appear to account for greater destruction of large weaklymutl
ungulate bones than of small ungulate bones in African assemblages. Several yE
The use of maximum bulk density values for skeletal parts instead of that have hi
traditionally used values does not change previous results. Future stud-
structural d,
ies aimed at explaining skeletal part frequencies should focus on more
than just correlations of those frequencies with bone density or the tend to also
MGUI. is that if higl
of bones wi1
correlation t
Introduction between bor
case of equi:
Zooarchaeologists have been 'making significant strides toward and must d
gaining insights to prehistoric economics in recent years. Many of those transport of
insights have come as a result of Binford's (1978) broadening of our knowl- destruction
edge about how modern hunter-gatherers differentially exploit and transport cesses (see L
portions of animal carcasses. Simultaneously we have learned the treachery of To determ
inferring past human behaviors from a zooarchaeological record that may previously E
only reflect past human actions dimly as a result of the myriad processes that bone parts i1
contribute to the creation of that record (e.g., Lyman 1984). These two ways in that study, I
the zooarch<
From Bones to Behavior: Ethnoarchaeologicaf and Experimental Contributions to the Interpretation of blages may,
Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional posi tiona! dE
Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights the archaeol
reserved. ISBN 0-88104-076-2. blages (37%)
324
Density-Mediated Attrition 1325
which our knowledge has increased meet head-on when varying frequencies
of skeletal parts of animals used as food are the subject of study (Lyman 1985,
1992).
Binford (1978) derived the modified general utility index (MGUI), a mea-
sure of the different food utility that various skeletal elements of two artie-
dactyl taxa displayed, in order to monitor how efficiently people were using
and, especially, transporting portions of the carcasses of prey animals. That
innovative technique for deriving such measurement scales has been exten-
ohts
I sively employed as an interpretive device in archaeological settings and been
emulated during the derivation of utility indices for other taxa (see reviews in
+
l
Lyman 1985, 1992). The MGUI can serve as an explanation of skeletal part
frequencies; for example, if parts of high utility are found in abundance, then
I the prehistoric people under study are said to have practiced a gourmet-utility
strategy.
~letal
parts
Another explanation of different abundances of carcass parts concerns the
1t density-
:lS between
structural density of· skeletal parts. Experimental work shows that some
Half of the taphonomic processes, such as carnivore ravaging and fluvial transport, tend
:1lysis seem to remove bone parts with low structural densities more rapidly and more
t, and only often from an assemblage of bones than those same processes remove bone
might be parts with nigh structural densities (see review in Lyman 1984). Thus, it might
not always be argued that v\r,hen a positive and statistically significant correlation is found
1 threshold between bone P,'i'lrt frequencies and the structural density of bone parts, the
nblages to responsible taphonomic process was one (or more) whose effects are greatly
1ts such as muted or mediated by the high structural density of some bone parts but are
m of large weakly muted or mediated by the low structural density of other bone parts.
3emblages. Several years ago I (Lyman 1985, 1992) observed that many skeletal parts
instead of that have high food values as measured by the MGUI also tend to have low
1ture stud-
structural. densities and that many skeletal parts that have low food values
ts on more
;ity or the
tend to also have high structural densities. The implication of this observation
is that if high frequencies of bones with low utility values and low frequencies
of bones with high utility values are found, the analyst will find a negative
correlation between the MGUI and bone frequencies and a positive correlation
between bone density and bone frequencies. The analyst is here faced with a
case of equifinality (the production of similar results by different processes)
~s toward and must determine if the bone frequencies are the result of differential
r of those transport of high~utility skeletal parts away from the site, density-mediated
ur knowl- destruction of high-utility /low-density skeletal parts, or perhaps both pro-
transport cesses (see Lyman 1991a and Rapson 1990 for suggestions).
2achery of To determine how often we might be faced with such cases of equifinality, I
that may previously examined the statistical relationships between the frequencies of
:esses that bone parts in 67 assemblages, bone density, and the MGUI (Lyman 1991a). In
'O ways in that study, I argued that density-mediated attrition might be quite frequent in
the zooarchaeological record and showed that 28 (41.8%) of the 67 assem-
;rpretation of blages may well have undergone such attrition. I also suggested that postde-
0Occasional positional destruction might accountfor the fact that proportionately more of
All rights
o
the archaeological assemblages (45%) than of the ethnoarchaeological assem-
blages (37%) were significantly correlated with bone density. Finally, I found
326IR. L. Lyman
that 68.7% of those 67 assemblag~s were not correlated with the MGUI, indi:.. the late Mi
eating that the then available MGUI values were of limited help in explaining in age, anc
skeletal part frequencies. . derived.
In this chapter, I examine 20 additional assemblages to further assess the I correla
patterns I originally detected. This 30% increase in the sample of assemblages values des
is sufficient to warrant a reevaluation of my original analysis and to add to the caribou M<
growing discussion on such issues in zooarchaeological research. Importantly, coefficient
the reevaluation leads to several new insights into the issue. ordinal sea
out. Follb"'
set to.one <
Materials and Methods tions are tl
significan tl
Methods used are identical to those described in Lyman (1991a). assemblag~
Briefly, assemblages of artiodactyl bones for which the number of identified 1991a). I cc
specimens (NISP), minimum number of elements or element portions (MNE), blages and
or minimum number of animal units (MAU) values were available were used. in the 87 cc
NISP values were treated as MNE values if the latter were not available, and
all MNE (for long bones these are actual! y MNEs of proximal or distal ertd
portions) values were converted to MAU values (Binford 1984). Occasionally,
only MNI (minimum number of individuals) values per skeletal element had
been reported, and in those cases the values were treated as equivalent to
MAU values. Because of variations in how data were reported, the number of (see Figure
MAU categories that could be correlated with the MGUI and bone density not surpris
values varies between assemblages. logical rese
The effect on my analysis of treating the quantitative units in the manners to have un
just described is unknown, although I suspect that it is minimal for the simple lated with
reason that NISP is often a very good predictor of MNE and MNI, and MNE correlated
likewise is a good predictor of MNI in many assemblages (Grayson 1984). It when reve
should be noted as well that it is the frequencies of proximal and dystal ends problem oJ
of long bones that typically show evidence of density-mediated destruction, tween a se
not MNE values for complete bones (e.g., Marean and Spencer 1991). Further, density val
the MGUI was originally built around the former coun'ting units, not the transport <
latter. Thus, statements that numbers of long bone shaft fragments provide observed b
better estimates of abundances of limb bones (e.g., Marean and Spencer 1991), 10% of the
while perhaps correct, cannot serve as a logical warrant for use of complete problem o
long bone :MNE values in analyses like those presented here. Obviously, the ubiquitous.
analytic goals and thus the analytic units are different. Garvin('
Assemblages were selected for analysis from literature available in my mented tht
personal library and in my university's library. An attempt was made to use bones. The
assemblages such that as many time-space contexts as possible were repre- density (p =
sented. Given the nature of the literature available to me, several contexts are 1984, 1991;:
better represented than others, and some contexts are not represented. Since I bone assen
am monolingual, the references I used were restricted to those published in (1980) doe~
English. To date, I have performed the analyses described here on 60 North remaining.
American, 1 South American, 22 African, and 4 European assemblages. Geo- and collea~
graphic contexts such as Australia, China, and the Near East are not repre- a strong po
sented in the sample of assemblages I consider here. One assemblage dates to (Lyman 19
ri
I Density-Mediated Attrition I 327
l
GUI, indi- I the late Miocene or early Pliocene, 11 date to the Pleistocene, 45 are Holocene
=xplaining
I
l
in age, and 30 are ethnoarchaeologically documented or were experimentally
derived.
assess the I correlated the MAU values for each assemblage with the volume density
semblages values described in Lyman (1984) and with both the sheep MGUI and the
add to the caribou MGUI described by Binford (1978);· each assemblage thus has a three-
lportantly, coefficient set. Because there are good reasons to believe MAU values are
ordinal scale (Grayson 1984; Hudson 1990), I use Spearman's rho (r5 ) through-
I
!
out. Following Grayson's (1988) lead, I assigned each assemblage's coefficient
set to one of nine possible classes (Figure 18-1) as follows. Significant correla-
-I tions are those for which p $; 0.05. If the MAU values for an assemblage were
m (1991a). I significantly correlated with either the sheep MGUI or the caribou MGUI, that
assemblage was classed as being correlated with the MGUI values (Lyman
identified
ns (MNE),
l 1991a). I consider individual coefficient sets for several of the 20 new assem-
blages and then tun~ to some general observations about the patterns evident
Nere used. in the 87 coefficient sets. I conclude with some general observations.
ilable, and
distal end
casionally, . Results of Correlation Analysis
ement had
livalent to / Assemblage-specific coefficient sets are described in Table 18-1
number of (see Figure 18-1 for the classification system used in Table 18-1). It is perhaps
ne density not surprising that one of the assemblages of bones that started zooarchaeo-
logical research down the taphonomic road, the Makapansgat bovids, appears
e manners to have undergone significant density-mediated attrition (is positively corre-
the simple lated with bone density). The fact that that assemblage is also negatively
and MNE correlated with the MGUI illustrates how the problem of equifinality arises
m 1984). It when reverse utility curves are derived (Grayson 1989; Lyman 1985). The
iistal ends problem of equifinality arises because finding a significant relationship be-
estruction, tween a set of MGUI values and a set of MAU values, and a set of bone
l). Further, density values and that same set of MAU values, means that either differential
ts, not the trqnsport or differential destruction, or both, may be responsible for the
ts provide observed bone frequencies. Such a pattern of coefficients is found in just over
ncer 1991), 10% of the 87 total assemblages I have examined, suggesting that while the
f complete problem of equifinality thus identified is a frequent occurrence, it is not
iously, the ubiquitous.
Garvin (1987) fed domestic cattle (Bas sp.) limb bones to canids and docu-
tble in my mented the progressive weight loss of proximal and distal ends of those
.ade to use bones. The bone ends that lost the most weight are also those with the lowest
rere repre- density (p = 0.06), as migh~ be predicted given previous analyses (e.g., Lyman
)ntexts are 1984, 1991a). Given those preceding analyses, it is quite surprising that the
:ed. Since I bone assemblage recovered from a hyena den and reported by Richardson
rblished in (1980) does not correlate with bone density, nor do the frequencies of bones
1 60 North remaining from f9ur deer (Odocoileus sp.) carcasses fed to wol~es by Klippel
ages. Geo- and colleagues .(1987). Given most previous research, I would have predicted
not repre- a strong positive relationship should haye been found. In my earlier analysis, I
ge dates to (Lyman 1991a) implied such unpredictable results might be obtained from
3281 R. L. Lyman
M.AU:BONE DENSITY
NEGATIVE POSITIVE
SIGNIFICANT INSIGNIFICANT SIGNIFICANT
WI-
>Z CLASS 3 CLASS 2 CLASS 1 Bone Asser
-<(
1-u
<(_
and Ta
CJlL E=1 E=S E=1 (Referen
w-
z~ A=O A=3 A=B
<f) CLASS 1
Makaparu:
CLASS 6 CLASS 5 CLASS 4 (Brain 198
E=1 E=9 E = ·s Hatiuhpul

A=O A= 21 A= 20 deer-size'
(Lyman 1~
Lost Terra
1-
CLASS 9 CLASS 8 CLASS 7 antelope (
z Fisher 199
w<C
>~ E=1 E=4 E=O
-LL Head-Smc:
1--
-z A=O A=5 A=O
(/)CJ (Brink anc
o_
Q_(/)
Eden-Fars
E = ethnoarchaeologiceJ 8.3semblages pronghorr
A= archae ol o gicaJ (prehistoric, pal eont ol ogicaJ) 8.33 emblag es (Frison 19:
Figure 18-1. Classification of 87 coefficient sets. CLASS2

/ Dead Indi,
r'
samples that were not representative of the deposit. While that may be the bighorn sl
(Fisher 191
case for Richardson's paleontological hyena den, such an explanation does not
work for Klippel and colleagues' neotaphonomic assemblage of deer bones. CLASS4
The bone assemblage from the Nossob porcupine den reported by Brain
(1981) does not correlate with bone density or the MGUI, nor does the bison Dead Indi•
(Bison bison) bone assemblage from Shield Trap Cave reported by Oliver deer (Fish(
(1986). Both of the latter assemblages were probably accumulated without
regard for the humanly perceived food utility of particular skeletal parts. Bugas-Hol
Thus, it is not surprising that neither of the latter two assemblages correlates (Rapson 1~
with the MGUI.
It is evident that as carnivores gnaw on bones, those bones become progres- Swartkran
sively smaller and thus progressively less identifiable. It is thus reasonable to bovid size
ask if assemblages gnawed to different degrees of intensity all correlate with (Brain 198.
bone density. To address that question, I used Garvin's (1987) data on weight
Swartkran
loss of long bone ends gnawed over a 22-day period by carnivores. Garvin
bovid size
listed the percent of weight loss of each bone part every two days. Those data III (Brain 1
indicate that it took nearly three weeks before those bone parts were gnawed
I
330 R. L. Lyman .
Table·18-1.-Continue4 sufficientl"
are less id
indicate tl
Bone Assemblage Number With With With
significan
and Taxa of Bone Sheep Caribou
until som·
(Reference) Categories Density MGUI- MGUI
bones in a
are not ne·
CLASS5 Klippel
carcasses·
Bugas-Holding the wolve
bighorn sheep 28 0.357 -0.188 -0.255
The coeff:
(Rapson 1990) p = 0.06 0.33 0.18
waited an
Lower Pittsburgh due to ''be
Landing deer-size remains a:
artiodactyls (Lyman 29 0.201 0.117 0.178 the hyena
1991b) p = 0.29 0.54 0.35 dent gna1
tween bm
Salishan Mesa deer- surprising
size artiodactyls 29 0.247 -0.207 -0.133 blages we
(Lyman 1990b) p = 0.19 0.27 0.49 not be cor
study of"
Nossob porcupine den 20 -0.206 -0.253 -0.332 ence/absE
(Brain 1981) p = 0.37 0.27 0.14 proportio:
sary, as d
Shield Trap bison 29 -0.177 0.315 0.047
fragments
(Oliver 1986) p =0.35 0.09 0.79
Marsha
Dog-gnawed bovid nutrients'·
bone weight loss 8 0.714 -0.683 -0.683 frequenciE
(Garvin 1987) p = 0.06 0.07 0.07 significan
amount o
Hyena den 12 -0.14 0.115 0.025 Pearson's
(Richardson 1980) p = 0.65 0.704 0.89 found tha
grease per
Wolf-gnawed deer is used (r5
carcasses (Klippel 26 0.318 0.213 0.213 and Dawe
et al. 1987) p = 0.11 0.287 0.286 after 22 di
erred is w
Krorndraai A, bovid
"grease" i
size class n 27 0.189 0.014 0.098
(Brain 1981) p = 0.34 0.90 0.62 volume (i
measure 1
CLASS8 index and
the pore~
45KI285-IT deer and the 1
(Chatters and 25 0.009 0.433 0.306 Marshall'
Zweifel1987) p = 0.92 0.03 0.13 rather beli
In my E
r
I
! Density-Mediated Attrition j331
sufficiently to correlate with density (Table 18-2). Insofar as smaller fragments

are less identifiable than larger pieces of the same skeletal element, the results
indicate that carnivore gnawing will not necessarily produce a statistically
With significant relationship between bone density and skeletal part frequencies
Caribou
until some threshold amount of damage is reached. This means that when
MGUI
bones in an assemblage have gnawing marks, the observed bone frequencies
I
l are not necessarily a result of carnivore gnawing.
Klippel and colleagues (1987:157) collected the remains of the four deer
·j
I
I
carcasses they fed to wolves (Canis lupus) "when all meat was consumed and
the wolves showed no further interest in the remains (usually 4 to 7 days)."
-0.255
The coefficients in Table 18-2 suggest that if Klippel and colleagues had
0.18
I waited another 15 days the bones would have been gnawed more (perhaps
due to ''boredom gnawing"; Haynes 1980) and a correlation between the deer
remains and bone density would have resulted. If the bone assemblage from
0.178
0.35
I the hyena den described by Richardson (1980) had not been subjected to suffi-
cient gnawing to reach the damage threshold, the lack of a correlation be-
tween bone density and the frequencies of bones in that assemblage is not
surprising. Similarly, the Nossob porcupine lair and Shield Trap Cave assem-
-0.133 blages were not extensively carnivore gnawed and thus should also, perhaps,
0.49 not be correlated with bone density. If correct, then clearly we need detailed
study of ,\rariables other than just the frequency of skeletal parts or the pres-
-0.332 ence/ absence (or frequencies) of gnawing marks or both. Data such as the
0.14 proportion of weight lost by bone parts with gnawing damage seem neces-
sary, as do data on the influence of fragment size on the identifiability of
0.047
0.79
fragments.
Marshall and Pilgram (1991) suggested that the amount of "within-bone
nutrients" such as grease and marrow is a key factor in determining the
-0.683 frequencies of skeletal parts represented in some assemblages. They found no
0.07 significant relationship between bone density and a fiat measure of the
amount of bone grease in a bone part (the volume [ml] of that bone part):
0.025 Pearson's r = -0.39, p = 0.11 (rs = -0.29, p = 0.23). However, I (Lyman 1992)
0.89 found that relationship to be strongly negative when the actual amount of
0.213
iI grease per part as determined for bison (Bison bison) by Brink and Dawe (1989)
is used (r5 = -0.986, p = 0.0015). In fact, the bone grease index derived by Brink
I and Dawe (1989) for bison is correlated with Garvin;s (1987) bone weight loss
0.286
I after 22 days' data (r5 = 0.81, p = 0.03). Where I believe Marshall and Pilgrim
erred is with their measure of grease amount (volume of bone part) because
0.098
0.62
i "grease" is perhaps not distributed in a manner strongly related to the total
volume (including grease-filled pore space) of a bone part, which their fiat
I measure necessarily pre$umes. Of course, the reason that the bone grease
index and bone density are inversely correlated is because the grease occupies
I the. pore space in trabeculated bone; the more pore space, the more grease,
and the less the volume density of a bone part. Thus, I do not dispute
0.306 f Marshall and Pilgram's claim for the importance of within-bone nutrients but
0.13
I
l
rather believe they are correct for the wrong reason.
In my earlier analysis I suggested postdepositional destruction might be
I
Il
I
Il
i
.l
I
3321 R. L. Lyman
Table 18-2. Correlation Coefficients (r J Between Percent Weight
Loss of Carnivore-gnawed Long Bone Ends and Bone Density
over Time
Time (days) Gnawed rs p

Coefficient
2 0.333 0.38 Set Class
6 0.429 0.26
10 0.595 0.11
14 0.595 0.11 Class 1 (d~
18 0.691 0.06 Class 2
22 0.714 0.06 Class 3
Class 4 (dE
Class 5
Class 6
Class 7 (dE
held accountable for a greater proportional abunda,nce of prehistoric assem- Class 8
blages (18 of 40) being positively correlated with bone density and a lesser Class 9
proportion of ethnoarchaeological assemblages (10 of 27) being positively
correlated with bone density. On one hand, only 1 of 3 (33%) ethnoarchaeo-
logical assemblages in the set of 20 new assemblages is positively correlated Note: SeeT
with bone density, and that is Garvin's (p = 0.06). On the other hand; 9 of the
17 (53%) prehistoric assemblages are positively correlated with bone density.
Omitting the Shield Trap Cave bison and Richardson's hyena den assem- the fact tr
blages because of their surface contexts, 9 of 15 prehistoric assemblages (60%) such as c
are positively correlated with bone density. Those results suggest density- from oveJ
mediated postdepositional destruction has taken place, whether the respon- ential trar
sible taphonomic process involves the weight of overlying sediment crushing a bone a~
bones into pieces too small to identify (e.g., Lyman and O'Brien 19871 or some MGUI.
other (pre-burial?) process. 1
Consid'
The 20 new assemblages produce results similar to those derived with the logical" a~
original 67 assemblages (Table 18-3). While it appears that proportionally as 28 of tl
many more of the 20 new assemblages fall in Class 1, the Kolmogorov- ly correla
Smirnov two-sample test suggests that difference is statistically insignificant sta tisticall
(D = 0.19, p > 0.05). The fact that eight of the nine classes are represented by results in<
the sample of 67 assemblages whereas only five of the nine classes are repre- more freq
sented by the sample of 20 assemblages no doubt reflects the tight statistical assembla1
relationship between sample size and richness of classes represented (Grayson blages to'
1984). These considerations suggest that it is reasonable. to add the results of A final
the two samples, and I do so in the following. ' assembla~
Figure 18-1 summarizes the results of the analysis of the total 87 assem- not none:x
blages. Approximately 41% of the 87 assemblages are positively correlated assemblaE
with bone density (Classes 1, 4, 7). Only 20.7% are negatively (Classes 1, 2, 3) but also (
and 11.5% positively (Classes 7, 8, 9) correlated with MGUI. Density-mediated MGUiwi·
attrition outstrips MGUI as an explanation of variability in bone frequencies in density
shown by this set of 87 assemblages. Part of the reason for that may reside in density, it
87 assemb
[
I
!
I
I Density-Mediated Attrition j333
zt I
l Table 18-3. Summary Distribution of Correlation Coefficients for
20 New Assemblages and 67 Previously Analyzed Assemblages
~
II
p I
I,
It
%of %of %of
Difference
Between
67 and 87
I Coefficient 20 Additional 67 Original 87 Total Assemblage
0.38 I Set Class Assemblages Assemblages Assemblages Samples
0.26 I
0.11 !
0.11 t
l
Class 1 (destruction) 25.0 6.0 10.3 4.3
I Class 2 5.0 10.5 9.2 -1.3
0.06
l Class 3 0.0 1.5 1.2 -0.3
0.06
Il Class 4 (destruction) 20.0 35.8 32.1 -3.7
Class 5 45.0 31.3 34.5 3.2
Class 6 0.0 1.5 1.2 -0.3
Class 7 (destruction) 0.0 0.0 0.0 0.0
c assem- Class 8 5.0 11.9 10.3 -1.6
a lesser
JSitively
I Class 9 0.0 1.5 1.2 -0.3
I
arch a eo-
>rrelated
, 9 of the I
I
Note: See Tjiible 18-1 for details and Figure 18-1 for definitions of coefficient classes.
density.
1 assem- I the fact that density-mediated attrition subsumes many taphonomic processes
such as carnivore gnawing, fluvial transport, fragmentation, and crushing
;es (60%)
density-
I from overburden weight, whereas the MGUI is meant to monitor only differ-
~ respon-
crushing
I
l
ential 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
or some MGUI.
I Considering the total sample of 87 assemblages, 9 of the 30 "ethnoarchaeo-
with the i logical" assemblages (30%) are positively correlated with bone density, where-
rtionally
wgorov-
I' as 28 of the 57 "archaeological" or prehistoric assemblages (49%) are positive..,
ly 'correlated with bone qensity. That difference in proportions is weakly
gnificant I statistically significant (arcsine transformation ts = 1.75, 0.1 > p > 0.05). These
results indicate that there is some potential for postdepositional destruction to
2nted by
re repre- more frequently result in: significant positive correlations ·of prehistoric bone
,tatistical
:Grayson
I assemblages with bone density than for ethnoarchaeological bone assem-
blages to correlate positively with density.
'esults of
I A final comment seems warranted regarding the absence of any Class 7
assemblages (Figure 18-1). That such assemblages should be extremely rare, if
7 assem- I 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
)rrelated ~ but also correlate positively with density. Given that the correlation of the
~s 1, 2, 3) !-
nediated J MGUI with density indicates bones that rank high in utility tend to rank low
quencies I 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
reside in
Il 87 assemblages I have examined.
I
J
l
j
i
I
~
!
!i
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 Hyena dE
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 Klasies R
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 Ngamurii
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) corre-
lated those bone frequencies with density; the coefficients are presented in
Table 18-4. The pattern that emerges is clear. Skeletal part frequencies of small avalues ar
bovids (< 84 kg live weight) do not correlate with bone density, whereas bcoefficieJ
skeletal part frequencies of large bovids (> 84 kg live weight) do correlate with to 84 kg, II
bone density (chF = 10.07, p = 0.005). Granting that density is inversely corre- from Klein
lated with the amount of grease in a bone, the coefficients in Table 18-4 ccoefficier
suggest more frequent and/ or more intensive exploitation of that grease in dcoefficie:
(1978) 90-n
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 Mem-
consider'
ber 2 size class II bovid remains, and the Kromdraai A size class II bovid remains·
remains as listed in Table 18-1 can be added. As well, I recalculated MAU
a passivE
frequencies for large and small bovids for Makapansgat from Dart (1957) and mula ted
for large and .small bovids for Richardson's (1980) hyena den data. Addition remains,
of those seven assemblages does not significantly alter the result obtained by
On the ·c
using only those coefficients listed in Table 18-4 (Table 18-5; chF = 8.00, p =
fromNoJ
0.005).
all but or
There seem to be two ways to pursue this issue. First, if the density-
ungulate
mediated attritional agents that produced the pattern of coefficients in Tables
lope (Ani
18-4 and 18-5 were in fact extracting within-bone nutriel'\ts, I would expect
considen
more percussion marks (Blumenschine and Selvaggio 1988), flake scars two in T;
(Lyman 1987), or gnawing marks on the bones of larger taxa than on the bones that·fall;
of small taxa. However, such data are not available for these assemblages. eluded w
The second way to evaluate the more intensive exploitation of within-bone those thr
nutrients in larger bovids is to examine other assemblages. Relevant data are,
positive!~
however, unavailable. On one hand, bone frequency data for four bovid taxa
can data
collected by Andrew Hill and described by Binford (1981:214-215) were
sizes of 1
rI
i
i
~
i
l
Density-Mediated Attrition I 335
I Table 18-4. Correlation Coefficients (r 5) Between Bone Frequencies
I for Bovid Size Classes and Bone Density
~ncies at
~in 1989;
II
! Bovid Size Cla$S rs p
Summed
MNia
hall and
al site of I Hyena denb (Binford 1981) Grysbok (I) 0.39 <0.05 143
991) are ! Springbok (II)
Reedbuk (ll)
0.25
0.25
> 0.15
> 0.15
117
141
~as large
tts. This
1m more
1'
a
Topi (Ill)
Wildebeest (Ill)
0.65
0.58
< 0.005
< 0.005
85
232
1 Kudu/ eland (Ill/IV) 0.64 112
; (grease < 0.005
Klasies River mouthc Small (I) 0.14 0.46 258
loitation
Small-medium (ll) 0.25 0.18 245
Large-medium (ill) 0.53 0.005 313
~ncies of Large (IV) 0.73 544
0.0001
ity; they Very large (V) 0.55 0.003 140
ncy data Ngamuriakd Caprine (ll) 0.09 0.72 299
n. Those Cattle (lli-IV) 0.43 0.07 140
a) corre-
ented in
of small avalues are sum of MNI for each skeletal part included.
whereas 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
late with from Klein (1975).
ly corre- ccoefficients from Klein (1989).
ble 18-4
dcoefficients recalculated from corrected data in Marshall and Pilgram (1991). Binford's
;rease in (1978) 90-month-old sheep weighed 44.9 kg.
tt. Is this
hat it is,
1.s Mem- considered in my earlier analysis (Lyman 1991a). However, because those
II bovid remains were from surface contexts, represent modern animals, and represent
~dMAU a· passive accumulation a~ross a large landscape rather than an actively accu-
957) and mulated assemblage like the other assemblages considered in Table 18-4, the
'\ddition remains were deemed inappropriate for consideration in Tables 18-4 and 18-5.
:tined by On the other hand, while there are many assemblages- of bones recovered
8.00, p = 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
density- ungulate in North America that is< 84 kg live weight is the pronghorn ante-
n Tables lope (Antilocapra americana) whoseremains make up two of the assemblages
d expect considered ih Table 18-1.-Because the three assemblages (one in Lyman 1991a~
ke scars two ih Table 18-1 here) are the only North American assemblages with taxa
he bones that fall in the small size class, comparisons like that in Table 18-S are pre-
ges. cluded with those assemblages. It is perhaps important to note, however, that
1in-bone those three North American small ungulate assemblages are all correlated
data are, positivgly with density and thus suggest that the pattern shown by the Afri-
1vid taxa can data (Table 18-5) may be a function ofthe wide range of available body
5) were sizes of ungulates ~n that continent. That is, African organisms (whether
3361 R. L. Lyman
Table 18-5. Classification.of 20 Correlation Coefficients Between subjectec

Bone Density and Small (< 84 kg live weight) and Large(> 84 kg the 95 sc
live weight) African Bovids skeletal I
extent of
dictated
Live Weight Size <84kg >84kg scan site
ysis of tl
Number correlated with bone density 3 9 150% of
much gn
Number not correlated with bo.ne densitya 7 1 proxima]
ful analy
aChi2 = 8.00, p = 0.005.
tion and
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 be-
cause they had several size classes to choose from, whereas North American and Sorg
organisms had fewer size classes from which to choose. · on tapho
1991; Pu:
Greater Density Values data two
ical and c
When I originally chose the bone density values that were to be not been
correlated with the MGUI and MAU values, my choices were somewhat this anal·
"subjective" and geared toward "averages" for the anatomical region being Europe,.
quantified (Lyman 1984:287). Recent discussions of those choices (Rapson Only 6 o:
1990; Gifford-Gonzalez, personal comm~nication) have suggested that the original,
choices do not represent realistic ones and that perhaps the greatest density relates tc
value for an anatomical region would be a better choice (see also LymaJil-1992). _ (1988:12~
Following that suggestion, I recalculated correlation coefficients betw'een the among t:
87 assemblages and the greatest density value for particular anatomical por- prehistor
tions. That resulted in only six (7%) of the assemblages having to be reclassi- for exa1
fied: two prehistoric assemblages changed from Class 4 to Class 5, three pre- (1986:641
historic assemblages changed from Class 5 to Class 4, and one ethnoarchaeo- rion [for
logical assemblage (Klippel et al.'s wolf-gnawed deer) changed from Class 5 data prm
to Class 4. These minimal changes are no doubt due to the fact that the density hominidE
values I have consistently used are quite tightly correlated with the greater Regard
density values others have suggested sho,uld be used (r5 = 0.88, p < 0.0001, N = what iss
31). The net result of using the greater-density values is that two assemblages various u
are removed from Class 5 and added to Class 4. and Ree\
Thus, while the results of my original analysis do not change significantly reported,
when the greater density values are used, the lessons here should be clear. sented or
First, density-mediated destruction is a very real phenomenon that must be are left v.
dealt with during the analysis of skeletal part frequencies. That is so because alternativ
with the maximum density values, 39 of the 87 assemblages (45%) are posi- are identj
tively correlated with density, which, while only a 3% increase over the results are now c
when the traditionally utilized density values are employed, indicates that as l'v1NE val
many as half of the assemblages that we might study may well have 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
84kg scan site on the radius, for instance, may not mean that the proximal diaph-
ysis of the radius was also destroyed, as the latter has a bulk density that is
9 150% 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
1 proximal end is missing but the proximal shaft is not, then, could be a power-
ful analytic tool that allows us to disentangle the effects of differential destruc-
tion and differential transport.
Discussion
r taxa
Lts be- With the publication of at least two edited volumes (Bonnichsen
erican 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 archaeolog-
ical and ethnoarchaeological, that I could add to my sample. That simply has
:to be not been the case in the literature I have examined, although I note that since
ewhat this analysis I have found a half dozen assemblages from the Near East and
being Europe, all published prior to 1990, which could be added to the analysis.
apson Only 6 of the 20 new data sets !.examine here appeared since I compiled the
,at the original data set (Marshall and Pilgram's 1991 data excluded). Perhaps this
ensity relates to a stronger belief in the mid to late 1980s that, to paraphrase Potts
1992). (1988:122), skeletal element frequency data are not adequate to distinguish
en the among the diverse attritional and accumulational agents that acted upon
1l por- prehistoric bone assemblages. That, however, does not seem likely because,
classi- for example, we find statements such as this by Blumenschine
~e pre-
(19.86:641): "Body part [frequencies] seem to represent the most telling crite-
:haeo- rion [for distinguishing hunted from scavenged faunal assemblages. Such
]ass 5 data provide] unambiguous evidence for the timing of access to carcasses by
.ensity hominids versus carnivores."
;reater Regardless of whether one occupies the Potts or the Blumenschine position,
11,N= what is surprising is that analysts still interpret bone frequencies, producing
blages 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
cantly reported, save in a graphed or other less than explicit form. The values repre-
clear. sented on a bivariate scatterplot are often extremely difficult to derive, so we
ust be are left with a graph that represents an interpretation but no "raw" data. An
~cause
alternative is to find the %:MNI per skeletal element listed, but whether those
~ posi- are identical to what Binford (1978) originally labeled %MNI values and what
results are now called (%)MAU values is seldom clear. Another alternative is to read
:hat as MNE values per complete skeletal element (e.g., Blumenschine 1986, 1989),
~been
3381 R. L. Lyman
which masks the important data <;m the frequencies of proximal and distal Blumensc
ends of long bones. While I am not implying that I want standardized data 1988
recordation, I do want to make it exp~icit that, whatever one's taphonomic Beha
a.nalysis might entail, the basic raw data .must be included to allow other Bonnichsl
researchers to evaluate that data's concordance with the desired goal of anal- 1989
y-sis and to allow those other researchers the option of using. it in their own Mair
work Few of us have the resources necessary to reanalyze all the collections Brain, C. J
1981
that would help us in our search for taphonomic regularities. Thus, several Brink, J., c
ctuantitative measutes-NISP, :NINE, MNI, MAU, an9- the minimum number 1989
of each scan site-should be regularly reported for individual skeletal · Albe1
elements and for individual taxa, along with explicit definitions of each Chq.tters,.
measure. 1987
Perhaps the correct belief is that obvious middle-ground position that skele- Centt
tal part frequencies alone are difficult to interpret in terms of the taphonomic logic
history involved. In my initial study of this problem I suggested that Dart, R. A
"multiple lines of evidence that bear on the destruction versus transport issue, 1957
not just the MGUI, should be examined" (Lyman 1985:234). That suggestion Muse
has not always been followed, even by me, perhaps because I was not explicit Davis, L. l
about what those other lines of evidence should involve. In several places in 1990
my discussion here I have indicated w~at I think some of the other kinds of Davis, L. l
1990
evidence are that, in conjunction with bone frequency data, might be impor- in th
tant to record. Those other kinds of data will, I hope, help us to analytically edite
disentangle the complex taphonomic histories of our bone assemblages. That Fisher, J. \
is, I think, the next step. It will show us either the limits to our present course 1984
or what we need to change about how we go from details about bones to CreeA
details of human behavior. Wyor
Wyo~
Frison, G.
Acknowledgments 1971
/ WyOJ
I must thank Jean Hudson; this analysis never would have taken Garvin, R.
place without her cajoling, pleading, and browbeating. Diane Gifford- 1987
Gonzalez, Lee Kreutzer, Curtis Marean, Fiona Marshall, James Savelle, and Scave
Mary Stiner were important sources of intellectual inspiration during the ofCa
preparation of this paper. Grayson, 1
1984 . I
1988
References can 1\
1989 l
Binford, L. R. ArchCi
1978 Nunamiut Ethnoarchaeology. Academic Press, New York Haynes, G
1981 Bones: Ancient Men and Modern Myths. Academic Press, New York 1980 J
1984 Faunal Remains from Klasies River Mouth. Academic Press, New York Bone!
1989 Response to Turner. Journal of Archaeological Science 16:13-16. Hudson,J.
Blumenschine, R. J. 1990
1986 Carcass Consumption Sequences and the Archaeological Distinction of Aka ..
Scavenging and Hunting. Journal of Human Evolution 15:639-659. Micrc
1989 A Landscape Taphonomic Model of the Scale of Prehistoric Scavenging Klein, R. C
Opportunities. Journal of Human Evolution 18:345-371. 1975 ]
distal Blumenschine, R. J., and M. M. Selvaggio

i data 1988 Percussion Marks on Bone Surfaces as a New Diagnostic of Hominid
1omic Behaviour. Nature 333:763-765.
other Bonnichsen, R., and M. H. Sorg (editors)
·anal- 1989 Bone Modification. Center for the Study of the First Americans, University of
Maine at Orono.
r own
Brain, C. K.
ctions 1981 The Hunters or the Hunted? University of Chicago Press, Chicago.
~venil Brink, J., and B. Dawe
tmber 1989 Final Report of the 1985 and 1986 Field Season at Head-Smashed-In Buffalo Jump,
eletal Alberta. Archaeological Survey of Alberta Manuscript Series No. 16. Edmonton.
:each Chatters, J. C., and M. K. Zweifel
1987 Archaeology of Eight Sites in the Multipurpose Range Complex, Yakima Firing
skele- Center, Washington. Archaeological Report 87-2. Central Washington Archaeo-
nomic logical Survey, Central Washington University, Ellensburg.
l. that Dart, R. A.
issue, 1957 The Osteodontokeratic Culture of Australopithecus prometheus. Transvaal
estion Museum Memoir No. 10. Pretoria.
)(plicit Davis, L. B., and B. 0. K. Reeves (editors)
1990 Hunters of the Recent Past. Unwin Hyman, London.
~ces in
Davis, L. B., ~d J. W. Fisher, Jr ..
1.ds of 1990 A;Late Prehistoric Model for Communal Utilization of Pronghorn Antelope
mpor- in the Northwestern Plains Region, North America. In Hunters of the Recent Past,
tically edited by L. B. Davis and B. 0. K. Reeves, pp. 241-276. Unwin Hyman, London.
:. That Fisher, J. W., Jr.
:ourse 1984 Medium-sized Artiodactyl Butchering and Processing. In The Dead Indian
nes to Creek Site: An Archaic Occupation in the Absaroka Mountains of Northeastern
Wyoming, edited by George C. Frison and Danny C. Walker, pp. 63-82. The
Wyoming Archaeologist 27{1-2).
Frison, G. C.
1971 Shoshonean Antelope Procurement in the Upper Green River Basin,
Wyoming. Plains Anthropologist 16:258-284.
taken Garvin, R. D.
Hord- 1987 Research in Plains Taphonomy: The Manipulation of Faunal Assemblages by
e, and 'Scavengers. Unpublished Master's thesis, Department of Archaeology, University
1g the of Calgary, Alberta.
Grayson, D. K.
1984 Quantitative Zooarchaeology. Academic Press, Orlando.
1988 Danger Cave, Last Supper Cave, and Hanging Rock Shelter: The Faunas. Ameri-
can Museum of Natural History Anthropological Papers 66(1).
1989 Bone Transport, Bone Destruction, and Reverse Utility Curves. Journal of
Haynes, G. .
1980 Evidence of Carnivore Gnawing on Pleistocene and Recent Mammalian
Bones. Paleobiology 6:341-351.
Hudson, J. L.
1990 Advancing Methods in Zooarchaeology: An Ethnoarchaeological Study among the
.ion of Aka. Ph.D. dissertation, University of California at Santa Barbara. University
Microfilms, Ann Arbor.
2nging Klein,R. G.
1975 Paleoanthropo.logical Implications of the Nonarch~ological Bone Assem-
I
>40 R. L. Lyman
blage from Swartklip I, South-Western Cape Province, South Africa. Quaternary Purdue,J.
Research 5:275-288. . ·
1989
1989 Why Does Skeletal Part Representation Differ Between Smaller and Larger
Upla
Bovids at Klasies River Mouth and Other Archaeological Sites? Journal of Archae-
Miss<
ological Science 16:363-382. · Rapson, C
!Clippel, W. E., L. M. Snyder, and P. W. Parmalee 1990
1987 Taphonomy and Archaeologically Recovered Mammal Bone from South- Resea
east Missouri. Journal of Ethnobiology 7:155-169. New
Lyman,R. t.. Richardso
1984 Bone Density and Differential Survivorship of Fossil Classes. Journal of 1980
Anthropological Archaeology 3:259-299. Paleo1
1985 Bone Frequencies: Differential Transport, In Situ Destruction, and the Stiner, M.
MGUI. Journal ofArchaeological Science 12:221-236. 1990
1987 Archaeofaunas and Butchery Studies: A Taphonomic P.erspective. In Pleist
Advances in Archaeological Method and Theory, vol. 10, edited by M. B. Schiffer, pp. Univ~
249-337. Academic Press, San Diego. Turner, A.
1990a Zooarchaeology. In Archaeological Data Recovery at Hatiuhpuh, 45WT134, 1989
Whitman County, Washington, edited by D. R. Brauner, pp. 98-138. Oregon State Partie:
University Department of Anthropology report to the Walla Walla District Klasi1
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
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
Density-Mediated Attrition 1341
ternary Purdue, J. R., B. W. Styles, and M. C. Masulis
1989 Faunal Remains and White-Tailed Deer Exploitation from a Late Woodland
Larger Upland Encampment: The Boschert Site (23SC609), St. Charles County,
Archae- Missouri. Midcontinental Journal of Archaeology 14:146-163.
Rapson, D. J.
1990 Pattern and Process in Intra-Site Spatial Analysis: Site Structural and Faunal
South- Research at the Bugas-Holding Site. Unpublished Ph.D. dissertation, University of
New Mexico, Albuquerque.
Richardson, P. R. K.
Irnal of 1980 Carnivore Damage to Antelope Bones and Its Archaeological Implications.
Paleontologia Africana 23:109-125.
md the Stiner, M. C.
1990 The Ecology of Choice: Procurement and Transport of Animal Resources by Upper
:ive. In Pleistocene Hominids in West-Central Italy. Unpublished Ph.D. dissertation,
ffer, pp. University of New Mexico, Albuquerque.
Turner, A.
WT134, 1989 Sample Selection, Schlepp Effects and Scavenging: The Implications of
m State Partial Recovery for Interpretations of the Terrestrial Mammal Assemblage from
District Klasies River Mouth. Journal of Archaeological Science 16:1-11.
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,
interpret
human-g
\ could pre
brigade,
cerning n
anthropo
19. Discussion: Noncultural Processes menton 1
Anna K. Behrensmeyer
vide disti
Introduction cultural,
The theme of chapters 16 through 18 is how to identify and control cautioh n
for noncultural processes in order to untangle them from cultural ones that Thepa
are of interest to archaeologists and anthropologists: Researchers who pursue they dea:
evidence of human behavior often regard noncultural taphonomic processes Althougl
as troublemakers that need to be "controlled." A paleobiologist takes a more grap!llc s
positive viewpoint on these processes as intriguing records of any and all tially,gen
biological and physical agencies that affect the vertebrate fossil record. Frus- the value
tration levels are typically much lower for the paleobiologist, who is happy to Blume1
be able to relate features of a bone assemblage to any one of many possible scavenge
causal processes. The anthropological focus has, however, catalyzed a great human-b
deal of productive taphonomic research and is helping to define the limits for on thee>
inferring human behavior (or that of any other single agent) from bones. define be
Identifying specific taphonomic processes that interacted with a given bone identify i
assemblage is becoming easier with the growth of actualistic studies that show assembla
what various agencies, such as wolves, hyenas, and trampling, do /o bones. tion beca
The zooarchaeologist has a new and improved set of tools and guidelines for ing resul·
identifying who or what has affected such assemblages. Controlling for the be recorc
noncultural components, that is, "removing the taphonomic overprint/' is segments
much harder and remains a difficult problem for ethnoarchaeological and ments co
fossil assemblages alike because (1) most assemblages reflect multiple pro- hyena b1
cesses, (2) many different processes produce similar effects, and (3) the same Serengef
bones may be affected by different processes in varying sequence. Ideally the ecosyster
archaeologist would like to be able to say exactly which features in an assem- entire Pli
blage were caused by humans and whic-h were not, but with increased appre- multicorr
ciation for the complexity of this problem, we are left wondering whether strong si1
even unequivocal qualitative analysis is possible. _ icked by1
The role of taphonomic research over the past decade has often been Hudso
characterized as that of a "wet blanket/' continually showing that simplistic small ani
Republic
From Bones to Behavior: Ethnoarchaeologicalcmd Experimental Contributions to the Interpretation of units are
Faunal Remains, edited by Jean Hudson. Center for Archaeological Investigations, Occasional resolutioJ
Paper No. 21. © 1993 by the Board of Trustees, Southern Illinois University. All rights contribu·
reserved. ISBN 0-88104-076-2. Blumens(
342
Noncultural Processes 1343
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 con-
cerning noncultural processes, as well as a theoretical treatment of the issue of
anthropologically oriented hope versus this reality. First, however, I will com-
;ses ment on the three papers in Part IV.
Comments on Written Contributions

The papers in Part IV, and many others in the volume as well, pro-
vide distinct glimmers of hope for distinguishing at least some of the effects of
cultural versus noncultural processes. They also provide expected words of
i control caution regarding interpretations of archaeological bone assemblages.
'nes that The papers by Blumenschine and Marean and by Hudson are similar in that
) pursue they deal with the effects of scavengers on human-generated assemblages.
~rocesses
Although they derive information from specific experimental and ethno-
sa more graphic situations, they distill from them some interesting ideas about poten-
·and all tially general patterns that could be tested with further work. This underlines
~d. Frus-
the valud of case studies that relate to a larger theoretical framework.
lappy to Blumenschine and Marean aim their study at documenting the behavior of
possible scavengers (specifically, the spotted hyena) when these animals encounter
l a great
human-butchered bones and use their observations to test predictions based
imits for on the expected effects of scavengers on bone assemblages. Their goal is to
es. define bone assemblage and bone modification features that could be used to
ren bone identify the "hyena after hominid" dual-component effects in archaeological
1at show assemblages. They point out that these two agents cannot be studied in isola-
:o bones. tion because the activities of one affect those of the other. One of the interest-
~lines for
ing results of the study is that competition levels in a hyena population may
s for the be recorded in the types of damage and degree of removal of edible bone
>rint," is segments. There are several critical points that apply to this paper: (1) experi-
;ical and ments conducted in the artificial hyena colony reflect only a limited view of
~ple pro-
hyena behavior, some of which may not be entirely "natural"; (2) the
:he same Serengeti experiments, .while much less artifical, still represent only one
eally the ecosystem at one point in time and should not be taken as models for the
n assem-
entire Plio-Pleistocene of Africa; and (3) most "real" situations will involve
d appre- multicomponent rather than dual-component taphonomic systems, and even
whether strong signatures imposed by hominds and hyenas may be masked or mim-
icked by other processes qffecting archaeological bone assemblages.
·en been Hudson presents a controlled study ofbone assemblages from relatively
implistic 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
>Jretation of units are used to .show the effects and are deemed more informative than finer
)ccasional resolution skeletal part tallies. The dominance of heads and limb fragments
All rights 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 At Death
25 kg body weight. Hudson raises the issue of a prey body-size threshold in 1. Initii
2. Idim
scavenging effects, an interesting possibility that needs further study. As in
3. Prey
the paper by Blumenschine and Marean, an important problem involves how 4. Prey
the assemblage characters Hudson documents would be further modified by 5. Heal
other processes en route to becoming an archaeological site. 6. Seas1
The paper by Lyman is quite different from the first two in its emphasis on 7. Tim~
the general empirical observation that bone density often explains much of the 8. Habi
patterning in modern and archaeological bone assemblages. Lyman's contri- 9. Pres1
-bution is important because it underlines the fact that differe11t processes can 10. Timi
result in similar patterns because of the density effect. His data show that
modern carnivores do not always cause density-dependent patterning, which Pre-Buria
is interesting and somewhat contrary to expectation. He also suggests that 1. Tran
there may be a threshold for density effects, and that might help to explain 2. Win!
3. Wea1
why some of the carnivore assemblages do not show this pattern. The dat'a
4. Buric
also show that archaeological assemblages are more often correlated with
density than modern ones, and Lyman speculates that it may be partly a Post-Buri
consequence of burial and, compaction. There is a need for more rigorous 1. Plan!
process-based explanations for the observations presented in this paper. 2. Insec
Otherwise, there is no theoretical framework for predicting under what cir- 3. Soil c
cumstances density-dependent effects seriously comprise our ability to distin- 4. Com
guish cultural signals in bone assemblages. 5. Diag;
The empirical studies discussed above are informative and heuristic
research efforts. They contribute to the growing body of data that provides a - Post-Diso
foundation for interpreting noncultural processes in bone assemblages. Such 1. Exca'
studies are essential building blocks that help to generate as well as test new 2. Meth
ideas and hypotheses. 3. Ident
4. Meth
Realities of Noncultural Taphonomy

The more research that is done on taphonomic processes, the effects OI
greater our appreciation for the complexity of what happens between the assemble
death of an animal and the final preserved state of its skeletal remains. affected 1
Unraveling taphonomic history and isolating the record of human activity are the stren
challenging and sometimes impossible-tasks. There is often great disparity relative b
between what we want to know and what we can know, given the limitations The pr
of the data. This is less discouraging, however, if we see increased under- cultural1
standing of complexity as a natural phase in the evolution of the relatively tap honor
new field of zooarchaeological taphonomy, to be followed by improved inter- through:
pretative frameworks and questions better suited to the available evidence. in single-
The complexity involved in taphonomic analysis of noncultural processes relatively
can be illustrated using one of the currently popular variables in zooarchaeol- not humi
ogy: skeletal part frequency. Table 19-llists processes, agents, and characteris- question
tics of the bones themselves that are known or suspected to have differential Hope is 1~
Noncultural Processes j345
scaveng- Table 19-1. Variables Known to Affect Skeletal Part Frequencies

o that of
severely
At Death or Shortly Thereafter
tls under
1. Initial agent of carcass modification (species of predator or scavenger)
~shold in
2. Idiosyncratic behavior of carnivores or scavengers (within same species)
iy. As in 3. Prey species morphology
lves how 4. Prey size
iified by 5. Health of prey
6. · Season (affects predatorI scavenger behavior)
)hasis on 7. Time of day (as for #6)
_ch of the 8. Habitat
's contri- 9. Presence of additional resources (e.g., a shade tree)
~sses can 10. Timing of access by multiple predator /scavengers
lOW that
g, which Pre-Burial
ests that 1. Trampling and kicking
) explain 2. Winnowing and transport
3. Weathering
The data 4. Burial potential (a function of size, shape, density)
ted with
partly a Post-Burjal
rigorous 1. Plant roots
.s paper. 2. Insects
Nhat cir- 3. Soil chemistry
to distin- 4. Compaction
5. Diagenesis (mineralization)
h.euristic
·ovides a Post-Discovery
~es. Such 1. Excavation techniques
test new 2. Methods of counting
3. Identification of fragmentary parts
4. Method of quantitative analysis
sses, the effects on skeletal part survival. Not all of these variables would affect every
veen the assemblage, of course; but most excavated samples probably would be
remains. affected by at least several of them. Isolating one agent or process depends on
tivity are the strength and distinctiveness of its overprint on skeletal part frequencies
:l.isparity relative to the effects of other processes.
nitations The present state of affairs regarding hope (i.e., probability) of unraveling
i under- cultural components in bone assemblages versusthe reality of noncultural
elatively taphonomic overprints can be simply portrayed in graphic form (Figures 19-1
·ed inter- through 19-3 ). The chances for distinguishing a cultural signature are greatest
lence. in single-component assemblages (i.e., those formed by a single process), for
)rocesses relatively simple questions (Figure 19-1). Such a question might be whether or
Hchaeol- not hu:rnan activity affected the assemblage at all, whereas a more complex
lracteris- question would involve the nature of the activity or the number of humans.
fferential 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
HOPE QUESTIONS UN
OF I
UNRAVELING I NFI
CULTURAL (E
COMPONENT
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 results ir
DUAL or multic
HOPE
t COMPONENT reach.tliE
overwhe
There
OF case stw
UNRAVELING / theory-b<
CULTURAL behavior
COMPONENT growth'
growth i:
from nea
might bt
INTENSITY OF TAPHONOMIC PROCESS ~ Often ho
Figure 19-2. Schematic plot showing the possible relationship between edge oft
hope (as defined in Figure 19-1) and the intensity of a particular taphon- bolsterec
omic process, such as human' butchery, for single-component, dual- ecologic'
component, and multicomponent bone assemblages (see text for further numbers
explanation). hopes fa]
emerge 1
amplituc
two or more processes. Another factor to consider, however, is the relative long run,
impact of a process on the bone assemblage (Figure 19-2).1 In single- the who]
component systems, hope of identifying the taphonomic agent or process is perhaps'
high at relatively low intensity but may fall off with time because most (althoug:
destructive processes eventually eliminate fragile, low-density parts. This would fif
Noncultural Processes 1347
THEORY
/{OFTEN BORROWED)
.--; DATA ON
, VARIABILITY
f
HOPE
I
f
I
OF f
UNRAVELING
CULTURAL
INFORMATION
(BEHAVIOR)
KNOWLEDGE OF "REAL WORLD" ~

(ACTUALISTIC & ARCHAEOLOGIC DATA)
Figure 19-3. Schematic plot suggesting changes in hope (probability) of

'-JTS
unraveling cultural information with increased knowledge of the com-
·iphering a plexity of taphonomic processes. The dashed lines represent the optimistic
~ersus the versus pessimistic initial outlook on the problem based on minimal
emblage. 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 knowl-
'p between edge of the complexity of the evidence. This optimistic point of view may be
ar taphon- bolstered by theory, which is often borrowed initially from another field. The
ent, dual- ecological theory of resource optimization is one example. With increased
cor further 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
e relative long run, although new questions or new general theories can appear to reset
n single- the whole system. At present, the field appears to be in the first valley and
::Jrocess is perhaps beginning to climb up the next hill as new hypotheses are developed
use most (although, of course, individual researchers vary greatly as to where they
arts. This 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
20.
part frequencies, species body size, breakage, and bone surface modifi-
cation.
2. Comparative analysis of archaeological sites, such as demonstrated in
the work by M. Stiner, and also more deliberate multivariate compar-
isons between ethnoarchaeological and archaeological assemblages.
3. Use of computer-assisted simulations based on theoretical predic- made b,
tions, ethnoarchaeological, or noncultural taphonomic assemblages to through
provide comparative data for multicomponent and time-averaged presentE
archaeological assemblages. related·
4. Testing of different scales of analysis, such as advocated by Hudson major tll
and Stiner. For some questions, general categories of body parts may b'e points o
more informative and also more statistically robust than finer catego- tal studi
rizations using individual bones or segments. itably tc:
ference 1
alogical
Note Jo Wats
1. Different processes may generate increasingly similar features in a bone assem- Behrens:
blage the longer or more intensely they act upon it, thereby reducing the hope of Grays
unraveling cultural components. There is perhaps more chance in multicomponent some of
situations of having at least one process that continues to leave distinctive traces with most etl
increased intensity, which is why the nhope line" remains slightly higher for multi- differen
component than for dual-component assemblages in this diagram. I 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
20. Concluding Discussion: The Role
nodifi- of Actualistic Studies
ated in
)lnpar-
::s. The following discussion represents an integration of comments
predic- made by various participants during the discussion sessions that occurred
Iges to throughout the two-day conference in 1991. Much more was said than is
eraged 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
Iudson major themes of the conference. The aim has been to capture some of the key
maybe points of dialogue concerning the role of ethnoarchaeological and experimen-
catego- tal studies and the future directions that such actualistic research might prof-
itably takf. Don Grayson served as an important catalyst early in the con-
ference by questioning the utility of some recent approaches to ethnoarchae-
ological research. Other participants whose comments are cited here are Patty
Jo Watson, Diane Gifford-Gonzalez, John Yellen, Susan Kent, and Kay
Behrensmeyer.
assem-
Grayson: I would like to begin by saying that I am not as enthusiastic about
wpe of
lponent some of this ethnoarchaeological work as I used to be. It seems to me that
es with most ethnoarchaeology that has been published to date has made two very
~multi- 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 contribu-
tions 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 in-
ductive 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 interest-
ing and helpful. They help define what we cannot or do not know; they pre-
vent 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

Paper No. 21. © 1993 by the Board of Truste~s, Southern Illinois University. All rights
349
I
350 Concluding Discussion·
driven and represents particularistic responses to archa,eological analyses. GiffoJ

There is a second kind of ethnoarchaeology that I find both far more in- decoupl
triguing and of longer-term value. I will take O'Connell's work as an example. to addre
While he has been busy issuing important cautionary tales, it is also true that lar, prec
most of the ethnoarchaeological work that he has done has been embedded in systems
a. particular theoretical approach, which has as its goal understanding why tures on
particular cultural systems work the way they work One may or may not like Yelle1
the optimization models that O'Connell employs, but what is important to me useful aJ
vance.F
is that his work does have a clear theoretical context.
essential
Watson: It seems to me that what is at issue here is the perennial debate in Andr
many fields about which should be uppermost-particularizing or general- _ perspect
izing goals: cautionary tales are particularizing and optimization theory is often sej
generalizing. If we are doing science, then of course generali.zing should be own exF
stressed: explanation of groups or classes of data has priority over explanation exercise
()f individual cases. ·But you cannot achieve and maintain good generaliza- animall
tions without good particularistic work conditio:
Don [Grayson] has suggested that a lot of ethnoarchaeological research is Simil2
inductive, a reaction to current concerns, and will be of passing interest only. groups;
In what sense is it inductive, and why is it of short-lived importance? fuller raJ
Don may mean that research of the cautionary tale variety is inductive be- If wear
aspects_(
cause it does not flow deductively from a specific theoretical framework. He
cally.
may feel it is of short-lived importance because the results are often inconclu-
Kent:
sive, or irrelevant to archaeological types of data, or because researchers may sions is·
make a pass at a subject and then drop it for another. that prol
It seems likely to me that what is happening is basic science at a relatively archaeol
early developmental stage: constructing an observational method to get at the alone.
issues of general concern-past human beings, human societies, and human Behre
cultures. To use Kay Behrensmeyer's phrasing, the effort is aimed at bringing multiple
hope and reality closer together through taphonomic studies, Bifl.fordian use a gn
middle-range theory, and site formation processes. Results may be short-lived particulc
for a very good reason: because a particular question or set of questions has of assoc
been answered, or has been shown to be a deadend, and it is time to extend words tc
inferential boundaries in new directions. nation o:
Yeller
O'Connell's research is predicated upon cost-benefit analysis of specific
often rer
aspects of a particular human society in one part of the world. Presumably, data are
his research is motivated by a belief that optimization theory can be used with blagemc
considerable profit by himself and others doing "real archaeology" (as Kevin con tinge
Jones phrased it) on analogous archaeological data. But to do real archaeology Behre:
about human optimization you would certainly like to be able to sort out spot- time-ave
ted hyena from hominids, and faunal assemblages created by schlepp effects, Yelle1
food-sharing, or culinary processes from those created by noncultural factors needed.
such as wind, water, dogs, and earthworms. months,
Both particularistic and generalizing approaches are essential and, in fact, understc:
inseparable. By using both, we learn a lot along the way and progress toward the factc
the goal that I think everyone does have in mind: figuring out how the world theoretic
of humankind works, regardless of which world is meant-the human past,
the human present, the archaeological record, or all three.
Concluding Discussion 1351
rses. Gifford-Gonzalez: While there is a danger with middle-range research of

more in- decoupling it from general theory, I think that we do need actualistic studies
~xample. to address uniformitarian processes. We need to investigate and identify regu-
true that lar, predictable relationships between dynamic processes and the states of
2dded in systems that generate them on the one hand, and their archaeological signa-
ling why tures on the other.
r not like
Yellen: Particularistic studies in response to archaeological questions are
useful and important. They typically do involve research issues of broad rele-
mt to m:e
vance. For example, I would say that many of the papers presente<:i here are
essentially concerned with the question of how humans got smart.
lebate in And having raised that issue, I would like to add that I find the prevalent
general- perspective on hunting tends to be somewhat simplistic and ethnocentric. We
:heory is often see hunting as a physical skill; a contest with a dangerous animal. My
1ould be own experience with the !Kung suggests to me that hunting is primarily an
)lanation exercise in intelligence, a combination of observational skills, knowledge of
neraliza- animal behavior, assessment of the probability of success given the particular
conditions, and evaluation of the costs and benefits involved.
search is Sim1larly, our notion of cooperative hunting is simplistic, focusing on
·est only. groups actively involved in coordinated pursuit of game rather than on a
fuller range of cooperative social behavior, such as the sharing of information.
If we are interested in how humans got smart, we need to consider these
ctive be-
aspects ;of intelligence and if and how they might be reflected archaeologi-
vork. He cally.
nconclu- Kent: One of the other issues that has been raised repeatedly in our discus-
1ers may sions is the problem of equifinality. I would suggest that we need to address
that problem and that one way to approach it is to better integrate all types of
elatively archaeological data and avoid building our interpretations on faunal data
~et at the alone. ·
i human Behrensmeyer: I think we should also place more emphasis on integrating
bringing multiple lines of evidence available in the bone assemblages themselves. To
nfordian use a grammatical analogy, we have been accumulating words, in the sense of
ort-lived particular types of analysis focused on certain types of data and certain types
:ions has of associated behaviors and processes, and it is time to start putting those
J extend words together to build sentences that will tell us something about the combi-
nation of behaviors and processes represented.
Yellen: I think we also need to remember that an archaeological assemblage
specific often represents a compression of multiple events, while ethnoarchaeological
;umably, data are often collected with a single event focus. The archaeological assem-
sed with blage may reflect a mix of responses to a variety of social and environmental
:.ts Kevin contingencies over time.
1aeology Behrensmeyer: There are ways to approach that problem, which is one of
Jut spot- time-averaging. Simulation modeling is one such approach.
J effects, Yellen: I think long-term field studies, spanning 20 or 30 years, are also
Ll factors needed. Much of our data comes from small windows in time-several
months, a year-and we do not get a sense of variability over time. A better
in fact,
~., understanding of complexity, when focused on the range of variability and
'toward the factors to which that variability responds, leads to the ability to address
le world theoretically important issues.
tan past,
Contributors
Barry W. Baker, doctoral candidate, Department of Anthropology, Texas

A & M University, College Station, Te~as
Laurence E. Bartram, Jr., doctoral candidate, Department of Anthropology,

University of Wisconsin-Madison, Madison, Wisconsin
Anna K. Behrensmeyer, research scientist, Department of Paleobiology,

National Museum of Natural History, Washington, D.C.
Robert ]. Blumenschine, associate professor, Department of Anthropology,

Rutgers lilniversity, New Brunswick, New Jersey
Henry T. Bunn, associate professor, Department of Anthropology, University

of Wisconsin-Madison, Madison, Wisconsin
Alice M. Emerson, research associate, Center for Northwest Anthropology,

Department of Anthropology, Washington State University, Pullman,
Washington
]ames G. Enloe, assistant professor, Department of Anthropology, University

ofiowa,IowaCity,Iowa
John W. Fisher, Jr., assistant professor, Department of Sociology, Montana

State University, Bozeman, Montana
Diane Gifford-Gonzalez, associate professor, Board of Studies in Anthropol-

ogy, University of California, Santa Cruz, California
Donald K. Grayson, professor, Department of Anthropology, University of

Washington,- Seattle, Washington
Jean Hudson, director, Zooarchaeology Laboratory, Institute of Archaeology,

University of California, Los Angeles, California
Kevin T. ]ones, assistant state archaeologist, Antiquities Section, Division of

State History, Salt Lake City, Utah
//
353
i
cl
I
I
3541 Contributors ii
Susan Kent, associate professor, Anthropology Program, Old Dominion

University, Norfolk, Virginia
I
Heidi Knecht, adjunct assistant professor, Department of Anthropology,
University of Miami, Miami, Florida I
R. Lee Lyman, associate professor, Department of Anthropology, University
I
of Missouri, Columbia, Missouri
Curtis W. Marean, assistant professor, Department of Anthropology, State

University of New York at Stony Brook, Stony Brook, New York
Fiona Marshall, assistant professor, Department of Anthropology, Washing-

ton University, St. Louis, Missouri
]ames F. 0 'Connell, professor, Department of Anthropology, University of

Utah, Salt Lake City, Utah
]ames S. Oliver, research associate, Anthropology Section, Illinois State

Museum, Springfield, illinois
Anne Pike-Tay, assistant professor, Department of Anthropology, Vassar

College, Poughkeepsie, New York
D. Gentry Steele, professor, Department of Anthropology, Texas A & M

University, College Station, Texas
Mary C. Stiner, assistant professor, Department of Sociology and Anthropol-

ogy, Loyola University, Chicago, illinois
Bonnie W. Styles, director of sciences, Illinois State Museum, Springfield,

Illinois
Patty ]o Watson, professor, Department of Anthropology, Washington

University, St. Louis, Missouri
John E. Yellen, director, Archaeology Program, National Science Foundation,

Washington, D.C.

Hudson (Ed) (From Bones To Behavior Ethnoarchaeological and Experimental Contributions To The Interpretation of Faunal Remains 1993) PDF

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Hudson (Ed) (From Bones To Behavior Ethnoarchaeological and Experimental Contributions To The Interpretation of Faunal Remains 1993) PDF

Încărcat de

Drepturi de autor:

Formate disponibile

From Bones to Behavior

Visiting Scholar Conference Volumes

Emergent Horticultural Economies

Tracing Archaeology's Past:

Between Bands and States

Processual and Postprocessual Archaeologies:

Quandaries and Quests:

Center for Archaeological Investigations

Edited by Ruth N .- Kissell .

I. Subsistence Strategies: Meat Acquisition and

3. The Place of Hominids among Predators: Interspecific

4. Uncovering Technological, Organizational, and Seasonal

5. Ethnoarchaeology of Marrow Cracking: Implications for the

IJ. Settlement Patterns: Site Function and Duration

7. Perspectives on Skeletal Part Profiles and Utility Curves

8. The Role of Body Part Utility in Small-scale Hunting.under 18. Densi

9. Bone Assemblages at Base Camps: A Further Consideration 19. Discu

III. Social Interaction: Food Sharing, Preparation Contr

11. Gaps in the Zooarchaeological Analyses of Butchery: Is

12. Carcass Processing by the Hadza: Bone Breakage from

13. Food Sharing and the Faunal Record

14. Foragers and Farmers: Material Expressions of Interaction

15. Discussion: Social Interaction

IV. Noncultural Processes: Carnivore Scavenging

16. A Carnivore's View of Archaeological Bone Assemblages

17. The Impacts of Domestic Dogs on Bone in Forager Camps

18. Density-Mediated Attrition of Bone Assemblages: New

19. Discussion: Noncultural Processes

2-1. Predatory strategies concerning number of predators

3-1. Geographic distribution of four Middle Paleolithic

3-2. Examples of the two anatomical representation patterns

3-3. /Anatomical patterns for medium ungulate remains in

3-5. The relative frequenCies of horn parts (HORN /L) to head

3-6. Comparison of (H+H)/L medians and ranges for predator

3-7. Comparison of CH+H)/L medians and ranges for predator

3-8. Comparative data on total fat to total protein content

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

4-5. Seasons.:.of-death of red deer as determined by ~ementum 8-8. E

8-2. Comparative general utility model values for bison with

8-3. Total fat utility models for bison 143 12-2. c

8-6. Frequency of unit transport as indicated by scalogram 13-2. SJ

8-7. Frequency of unit transport as indicated by scalogram 14-1. L1

8-9. Relationship between bison nonskeletal fat yield and a

130 11-2. Model showing animal processing activities likely to occur

141 11-3. Model showing bone modifications and discards likely to

142 12-1. Axial and limb bone breakage frequencies 208

143 12-2. Comparison of breakage frequencies for cooked and uncooked

13-2. Spatial distribution of the faunal remains of a widely shared

16-1. Comparison of the rank order of 29 skeletal parts based on 17-7.

16-3. Size-specific destruction/ deletion of long bone epiphyses 284

16-6. Qualitative model of the degree of competition among 18-1.

17-2. Comparison of percentage contribution of each taxon based

17-3. Plot of the relationship between MNI survival and carcass

17-4. Plot of the relationship between MNI survival and sample

2-1. Examples of Hunting Strategies for Selected Social Mammalian

2-2. Selected Examples of Taxa Taken by Humans in Multiple

3-1. Expected MNE Values 42

3-2. SummaryStatistics for (H+H)/L by Predator Type 50

3-3. , Results of Rank-Ordered Comparisons of Predators 52

5-1. Bone Splinters from Three Nunamiut Samples 87