Beyond Hunting and Herding

Beyond Hunting and Herding: Humans, animals, and the political economy of the Vina period.
David Clive Orton Darwin College

3rd June 2008
Thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy
Faculty of Archaeology and Anthropology University of Cambridge
For Professor Clive Orton (dad)
Declaration of original research and content

This thesis is the result of my own work and includes nothing that is the outcome of work undertaken in collaboration except where specifically indicated in the text.
Statement of Length
This thesis does not exceed the word limit as stipulated by the Degree Committee for the Faculty of Archaeology and Anthropology.
ii
Abstract
This thesis is concerned with the role of animals within Neolithic society, specifically in the later Neolithic/Copper Age Vina culture of the central Balkans. There are two central themes: (a) the continuing role of wild animals in the Neolithic and the relationship of this to the place of domesticates; and (b) the role of animals within the political economy and their implication in processes of social change. Various authors have developed models of social change during the Vina period, revolving round sedentism, intensification of production and the emergence of the household rather than the community as the primary unit of social organization. I use faunal data from two sites studied at first hand, coupled with published reports, to address the possible roles of animals within these changes. At the same time, by documenting the continuing importance of wild animals throughout the Neolithic in the central Balkan region I hope to contribute to the recognition that the period cannot be characterized purely in terms of the domestic. The thesis starts with a review of the archaeology and chronology of the Vina period, and a theoretical consideration of human-animal relations, including problematization of the central concept of domestication. It is concluded that the wild:domestic distinction has an analytical validity if couched in terms of (a) incorporation of live animals into human property relationships, and (b) the possibility of direct relationships between individual people and animals. Domesticates constitute sentient property. Since both of these aspects are variable, the relevance of a wild:domestic distinction is context-dependent. The study itself consists of three parts. Firstly, all relevant faunal data as of April 2008 are assembled, along with the available dating evidence, and used to document trends in relative taxonomic representation. Statistically, proportions of wild fauna increase slightly from the earliest Neolithic to the earlier half of the Vina period, with a very small decrease into the later half, but this is dwarfed by variation within each phase. Numbers of wild animal remains decrease at several large sites, while the proportion of cattle amongst domesticates increases. Neither change is seen at two small, relatively wilddominated sites. The potential roles of both wild and domestic species are discussed, and it is concluded that the role of the former cannot be reduced to seasonal hunting or risk-buffering. Evidence for extraction of tangible secondary products from domesticates is mixed, but it is argued that various social roles have a similar effect on the value of live animals. The second part involves biometric analysis of fauna from Gomolava and Petnica, along with basic site/faunal reports for each site. The significance of feature types in terms of depositional practices is discussed at some length. The third part presents a detailed taphonomic-contextual study of fauna from the two sites. This involves a five-stage taphonomic methodology, working back from the observed assemblage to the deposited fauna but at each stage gleaning information about human processes of processing, consumption and deposition. Consumption was generally very intensive, even for largebodied species, and there is little difference between wild and domestic groups in this respect. Certain taphonomic variables indicate differential treatment of taxa according to size and only secondarily to wild/domestic status It is argued that domestic animals were slaughtered primarily for specific social events, and that the same was often true for wild animals notably for pig in the earlier phases at Gomolava. Increasing numbers of cattle are suggested to represent rising social importance for the species, linked to the tension between household and community proposed for the period. Cattle, it is argued, were particularly powerful as inalienable possessions, their breeding and exchange allowing a form of accumulation that was also cohesive through the establishment of interlinked human and animal social networks, and thus did not immediately conflict with a generally egalitarian ethos. This may have led to more extensive herding, with implications for the mobility of herders and for the role of animals in defining human identity. Importantly, it is argued that all these changes occurred on a site rather than a regional level.
iii
Contents
List of tables ________________________________________________________vii List of figures _______________________________________________________ ix Acknowledgements__________________________________________________ xiv Chapter 1 Introduction _______________________________________________1 Chapter 2 Archaeological background and aims __________________________3
2.1 The Neolithic: a brief history __________________________________________ 3 2.2 Hunting versus herding _____________________________________________ 5 2.3 The Vina culture ___________________________________________________ 7
2.3.1 introduction____________________________________________________________ 7 2.3.2 chronology ___________________________________________________________ 10 2.3.3 models of social change _________________________________________________ 13 2.3.4 sedentism ____________________________________________________________ 16
2.4 Statement of aims __________________________________________________ 18
Chapter 3 Theoretical perspectives on animals in the Neolithic______________19

3.1 Animals in archaeology: a critique ____________________________________ 19 3.2 Attitudes to animals_________________________________________________ 22
3.2.1 hunter-gather human-animal relations ______________________________________ 3.2.2 - domestication__________________________________________________________ 3.2.3 implications for the political economy ______________________________________ 3.2.4 implications for the wild:domestic dichotomy ________________________________ 24 28 36 38
3.3 Animals as food ____________________________________________________ 42

3.3.1 sharing_______________________________________________________________ 43 3.3.2 communal consumption and feasting _______________________________________ 44 3.3.3 disposal / discard / deposition _____________________________________________ 45
3.4 Animals in the Vina period: a new model ______________________________ 46

3.4.1 competition / cohesion / communal consumption______________________________ 46 3.4.2 wild and domestic in the Vina period ______________________________________ 47
3.5 Statement of hypotheses _____________________________________________ 49
Chapter 4 Outline of methodology _____________________________________50

4.1 Introduction _______________________________________________________ 50 4.2 Regional review of zooarchaeological data ______________________________ 51 4.3 Standard zooarchaeological methodologies _____________________________ 51
4.3.1 recording systems ______________________________________________________ 4.3.2 identification __________________________________________________________ 4.3.3 quantification _________________________________________________________ 4.3.4 demographic data and seasonality__________________________________________ 51 52 54 56
4.4 Taphonomic and contextual analysis___________________________________ 60

4.4.1 theory and scope of taphonomy ____________________________________________ 60 4.4.2 social zooarchaeology _________________________________________________ 62 4.4.3 development of taphonomic methodology ___________________________________ 63
4.5 Summary and structure of results _____________________________________ 77
Chapter 5 Regional trends in faunal data _______________________________78

5.1 Introduction _______________________________________________________ 78 5.2 Regional and chronological trends in species abundance __________________ 81
iv
5.2.1 the sample ____________________________________________________________ 81 5.2.2 dated sites ____________________________________________________________ 82 5.2.3 recovery bias __________________________________________________________ 87 5.2.4 wild versus domestic taxa ________________________________________________ 89 5.2.5 taxonomic representation ________________________________________________ 95 5.2.6 taxonomic relationships ________________________________________________ 104
5.3 Discussion ________________________________________________________ 108

5.3.1 roles of wild species ___________________________________________________ 108 5.3.2 roles of domestic species _______________________________________________ 116 5.3.3 temporal trends _______________________________________________________ 121
5.4 Summary ________________________________________________________ 123
Chapter 6 Results from Gomolava and Petnica__________________________124

6.1 Introduction ______________________________________________________ 124 6.2 Biometry _________________________________________________________ 124
6.2.1 Sus scrofa ___________________________________________________________ 6.2.2 Bos taurus / Bos primigenius ____________________________________________ 6.2.3 dogs and wolves ______________________________________________________ 6.2.4 deer ________________________________________________________________ 6.2.5 biometry summary ____________________________________________________ 6.3.1 introduction__________________________________________________________ 6.3.2 phasing and chronology ________________________________________________ 6.3.3 Block VII ___________________________________________________________ 6.3.4 feature types _________________________________________________________ 6.3.5 the faunal sample _____________________________________________________ 6.4.1 introduction__________________________________________________________ 6.4.2 phases ______________________________________________________________ 6.4.3 features _____________________________________________________________ 6.4.4 the faunal sample _____________________________________________________ 125 134 145 146 148 150 152 157 160 176 204 205 208 210
6.3 Gomolava ________________________________________________________ 150
6.4 Petnica __________________________________________________________ 204
6.5 Summary ________________________________________________________ 228
Chapter 7 Taphonomic analysis at Gomolava and Petnica ________________229

7.1 Evidence for attrition ______________________________________________ 229
7.1.1 completeness of carpals and tarsals _______________________________________ 229 7.1.2 correlation of abundance with bone density _________________________________ 229 7.1.3 stage 1 summary ______________________________________________________ 242
7.2 Evidence for peri-depositional damage ________________________________ 242

7.2.1 overall frequency and impact on fragmentation ______________________________ 242 7.2.2 comparison between contexts, taxa and phases ______________________________ 244 7.2.3 stage 2 summary ______________________________________________________ 250
7.3 Breakage and fragmentation ________________________________________ 250

7.3.1 overall levels _________________________________________________________ 7.3.2 pot-sizing and marrow extraction________________________________________ 7.3.3 contextual differences __________________________________________________ 7.3.4 stage 3 summary ______________________________________________________ 250 251 254 256
7.4 Visible human modification _________________________________________ 256

7.4.1 burning _____________________________________________________________ 256 7.4.2 butchery marks _______________________________________________________ 262 7.4.3 stage 4 summary ______________________________________________________ 270
7.5 Body-part representation ___________________________________________ 270

7.5.1 correlation between utility and abundance __________________________________ 270 7.5.2 body-part profiles _____________________________________________________ 271 7.5.3 stage 5 summary ______________________________________________________ 278
7.6 Spatial distribution at Gomolava _____________________________________ 279

7.6.1 distribution of taphonomic modifications ___________________________________ 279 7.6.2 distribution of taxa and body parts ________________________________________ 279
7.7 Discussion ________________________________________________________ 285 7.8 Conclusions ______________________________________________________ 291
Chapter 8 Synthesis and interpretation ________________________________292

8.1 Introduction ______________________________________________________ 292 8.2 Human relationships with specific taxa________________________________ 292
8.2.1 domestic pigs ________________________________________________________ 8.2.2 wild pigs ____________________________________________________________ 8.2.3 cattle _______________________________________________________________ 8.2.4 aurochsen ___________________________________________________________ 8.2.5 red deer _____________________________________________________________ 8.2.6 roe deer _____________________________________________________________ 8.2.7 caprines _____________________________________________________________ 8.2.8 dogs________________________________________________________________ 292 296 297 300 302 303 304 305
8.3 Explaining changes ________________________________________________ 307

8.3.1 animals and the (political) economy _______________________________________ 307 8.3.2 sedentism and settlement histories ________________________________________ 311
8.4 Wild:domestic? ___________________________________________________ 312
Chapter 9 Conclusions _____________________________________________316 References ________________________________________________________318
vi
List of tables
Table 2.1: summary of relative and absolute chronology for the Vina period. _________________ 11 Table 4.1: definitions of diagnostic zones. ______________________________________________ 55 Table 4.2: birth seasons for main taxa. ________________________________________________ 58 Table 4.3: components of taphonomic analysis, with potential cultural information. _____________ 66 Table 4.4: published bone density data for mammalian taxa. _______________________________ 68 Table 4.5: definition of element portions with corresponding scan sites. ______________________ 70 Table 4.6: calculation of Freshness Index (FI). Scores are summed to give final value.___________ 73 Table 4.7: definitions of anatomical regions (Stiner 1991)._________________________________ 75 Table 5.1: studied faunal assemblages from Vina and neighbouring cultures. _________________ 79 Table 5.2: earlier Neolithic faunal assemblages from the Central Balkans. ____________________ 80 Table 5.3: body size distribution by sample at Obre I and Opovo. ___________________________ 89 Table 5.4: unrotated component loadings for Central Balkan Neolithic faunal assemblages. _____ 105 Table 5.5: suggested seasonal foci of hunting at Vina sites. ______________________________ 110 Table 5.6: sex ratios for wild species in Vina assemblages._______________________________ 112 Table 5.7: age-at-death data for hunted animals on Vina sites.____________________________ 114 Table 5.8: sex ratios for domestic species on Vina/contemporary sites. _____________________ 120 Table 6.1: relative chronology of Gomolava phases in later publications (e.g. Brukner 1988). ____ 153 Table 6.2: evidence for pit-dwelling at Gomolava. ______________________________________ 165 Table 6.3: numbers of fragments and of identified specimens from Gomolava by feature type and phase. _________________________________________________________________________ 178 Table 6.4: specimen counts at Gomolava by feature._____________________________________ 178 Table 6.5: taxonomic frequencies at Gomolava by phase NISP.___________________________ 180 Table 6.6: frequencies of main taxa at Gomolava by phase MNE. _________________________ 181 Table 6.7: frequencies of main taxa at Gomolava by phase - DZ. ___________________________ 181 Table 6.8: taxonomic frequencies (%NISP) in Gomolava Block VII compared with Clason 1979.__ 182 Table 6.9: taxonomic frequencies of main species at Gomolava by feature type NISP. _________ 182 Table 6.10: frequency (%DZ) of main taxa at Gomolava and Opovo by feature type. ___________ 185 Table 6.11: taxonomic frequency (%NISP) at Gomolava by phase and feature type. ____________ 186 Table 6.12: raw and corrected distribution of Payne age stages for domestic cattle at Gomolava. _ 188 Table 6.13: frequencies of immature and mature roe deer mandibles at Gomolava by feature type and phase. _________________________________________________________________________ 192 Table 6.14: age attributions for caprine mandibles from Gomolava Block VII. ________________ 196 Table 6.15: fusion data for cattle and pigs at Gomolava. _________________________________ 197 Table 6.16: sex determinations at Gomolava. __________________________________________ 198 Table 6.17: distribution of sexes amongst pigs and cattle at Gomolava by main feature types. ____ 198 Table 6.18: sieving yields from Petnica 1991 (where known).______________________________ 211 Table 6.19: comparison of sieved, hand-collected and unknown-recovery material at Petnica.____ 211
vii
Table 6.20: NSP and NISP at Petnica by phase. ________________________________________ 212 Table 6.21: recovered densities of faunal remains in Vina levels at Opovo (Russell 1993), Gomolava and Petnica. ____________________________________________________________________ 212 Table 6.22: (a) NSP and (b) NISP in Vina phases at Petnica by feature type._________________ 212 Table 6.23: NSP & NISP by feature at Petnica._________________________________________ 213 Table 6.24: taxonomic frequencies at Petnica by phase (NISP). ____________________________ 217 Table 6.25: frequencies of main taxa at Petnica (%NISP) from the current study and Greenfield (1986, 284). __________________________________________________________________________ 218 Table 6.26: taxonomic frequencies at Petnica by phase (MNE). ____________________________ 220 Table 6.27: taxonomic frequencies at Petnica by phase (DZ).______________________________ 220 Table 6.28: taxonomic frequencies (%NISP) at Petnica by context type. _____________________ 222 Table 6.29: age estimates for caprines from Petnica. ____________________________________ 225 Table 6.30: sex attributions at Petnica. _______________________________________________ 227 Table 7.1: percentage of small bones recorded as complete or almost complete at Gomolava and Petnica.________________________________________________________________________ 229 Table 7.2: correlations between MAU and median portion density at Gomolava. ______________ 230 Table 7.3: correlation coefficients and sample sizes for corrected NISP vs. density at Gomolava. _ 231 Table 7.4: correlation coefficients and sample sizes for MAU vs. density at Petnica.____________ 237 Table 7.5: correlation coefficients and sample sizes for corrected NISP vs. density at Petnica.____ 237 Table 7.6: frequencies and percentages of weathering and gnawing at Gomolava and Petnica. ___ 243 Table 7.7: gnawing and weathering rates by taxon at Gomolava and Petnica._________________ 246 Table 7.8: weathering and gnawing rates by phase and feature type for three main taxa at Gomolava. ______________________________________________________________________________ 249 Table 7.9: observed and expected frequencies of weathering and gnawing by phase at Petnica. ___ 249 Table 7.10: overall fragmentation levels at Gomolava and Petnica._________________________ 251 Table 7.11: evidence for post-burning breakage of long-bones at Gomolava. _________________ 252 Table 7.12: overall frequencies of burning at Gomolava and Petnica. _______________________ 258 Table 7.13: burning rates by taxon and context type at Gomolava and Petnica.________________ 261 Table 7.14: burning on long-bone shafts versus articular ends at Gomolava and Petnica. _______ 262 Table 7.15: distribution of cut-marks at Gomolava and Petnica. ___________________________ 263 Table 7.16: frequency of cut-marks by taxon and phase at Petnica. _________________________ 269 Table 7.17: head, axial and appendicular specimens at Gomolava by phase and context type. ____ 278 Table 7.18: correlations between frequencies of major taxa in individual pits, by (a) %NISP at Gomolava and (b) %DZ at Opovo.___________________________________________________ 284 Table 7.19: correlations between right and left elements in pits at Opovo.____________________ 284 Table 7.20: summary of taphonomic evidence from Gomolava and Petnica. __________________ 287
viii
List of figures
Figure 2.1: map of the central Balkan region with place names mentioned in the text. ____________ 8 Figure 2.2: estimated start and end dates for Vina phases. ________________________________ 12 Figure 3.1: the relationship between epistemologies and approaches to animals in archaeology.___ 22 Figure 3.2: taxonomies of human-environmental relations set out by (a) Descola and (b) Plsson. _ 23 Figure 3.3: Ingold's representation of Western anthropological (top) and hunter-gatherer (bottom) 'economies of knowledge'. From Ingold 1996, 127. _______________________________________ 27 Figure 3.4: domestication as process and as state in four common approaches. ________________ 30 Figure 3.5: a hypothetical model of the relationships between humans and animals (a) without, and (b) with domesticates. ________________________________________________________________ 41 Figure 4.1: targets of taphonomic research. ____________________________________________ 61 Figure 4.2: structure of taphonomic analysis. ___________________________________________ 64 Figure 4.3: expected patterns of household faunal waste resulting from several hypothetical scenarios. _______________________________________________________________________________ 77 Figure 5.1: locations of Central Balkan Neolithic sites with published faunal assemblages. _______ 78 Figure 5.2: radiocarbon dates from Divostin. ___________________________________________ 84 Figure 5.3: radiocarbon dates from Selevac.____________________________________________ 84 Figure 5.4: radiocarbon dates from Anza.______________________________________________ 85 Figure 5.5: radiocarbon dates from Obre I & II . ________________________________________ 86 Figure 5.6: percentage contribution of small mammals by recovery standard.__________________ 88 Figure 5.7: distribution of wild components in Early Neolithic versus Vinca/contemporary sites ___ 90 Figure 5.8: range of variation in wild component of assemblages by period ___________________ 90 Figure 5.9: contributions of wild and domestic taxa to (a) pre-Vina and (b) Vina/contemporary assemblages _____________________________________________________________________ 91 Figure 5.10: contributions of wild and domestic taxa to (a) Vina A/B (and contemporary) and (b) Vina C/D (and contemporary) assemblages____________________________________________ 93 Figure 5.11: percentage contributions of wild taxa to assemblages by approximate date _________ 94 Figure 5.12: representation of wild species over time at multi-phase sites. ____________________ 95 Figure 5.13: proportions of domestic species (%NISP) in (a) Early Neolithic, and (b) Vina/contemporary assemblages. ____________________________________________________ 97 Figure 5.14: proportions of domestic species(%NISP) in (a) Vina A/B and (b) Vina C/D assemblages._____________________________________________________________________ 98 Figure 5.15: percentages of cattle amongst domesticates at Central Balkan Neolithic sites. _______ 99 Figure 5.16: percentages of cattle amongst domesticates in multi-phase assemblages. ___________ 99 Figure 5.17: percentages of pigs identified as wild at Central Balkan Neolithic sites. ___________ 100 Figure 5.18: proportions of wild species (%NISP) in (a) Early Neolithic, and (b) Vina/contemporary assemblages.____________________________________________________________________ 102 Figure 5.19: proportions of wild species(%NISP) in (a) Vina A/B and (b) Vina C/D assemblages. 103 Figure 5.20: component loading plots. _______________________________________________ 106 Figure 5.21: faunal assemblages plotted by scores on the 2nd and 3rd components. ____________ 108 Figure 5.22: variation in proportions of wild taxa per unit at Gomolava and Petnica. __________ 111
ix
Figure 5.23: ubiquity against unit %NISP for major species at Gomolava and Petnica. _________ 112 Figure 5.24: harvest profiles for (a) cattle, (b) caprines, and (c) pigs at central Balkan sites. _____ 118 Figure 5.25: caprine harvest profile at Opovo by phase.__________________________________ 120 Figure 6.1: Sus distal humerus measurements from Gomolava, Petnica and other Vina or related sites. __________________________________________________________________________ 126 Figure 6.2: Sus distal tibia measurements for Gomolava, Petnica and other Vina or related sites. 126 Figure 6.3: Sus astragalus measurements from Gomolava, Petnica and other Vina and related sites. ______________________________________________________________________________ 127 Figure 6.4: distribution of Sus LAR measurements from Gomolava, Petnica, Opovo and Vina.___ 128 Figure 6.5: Sus lower tooth measurements from Gomolava and Petnica. _____________________ 130 Figure 6.6: breadth of Sus lower M1 versus M2. ________________________________________ 131 Figure 6.7: histogram and kernel density plot for mean Sus postcranial log-ratio values by specimen at Gomolava and Petnica ____________________________________________________________ 131 Figure 6.8: mean Sus postcranial log-ratio values from Gomolava and Petnica by previously-assigned domestication status. _____________________________________________________________ 132 Figure 6.9: log ratio values for Sus lower and upper permanent teeth at Gomolava and Petnica by sex. ______________________________________________________________________________ 133 Figure 6.10: Bos distal tibia measurements from Gomolava, Petnica and other Vina and related sites. ______________________________________________________________________________ 136 Figure 6.11: Bos calcaneus measurements from Gomolava, Petnica and other Vina and related sites. ______________________________________________________________________________ 136 Figure 6.12: Bos pelvis measurements from Gomolava, Petnica and Vina, showing sex where attributed. ______________________________________________________________________ 137 Figure 6.13: Bos metacarpal measurements from Gomolava, Petnica and other Vina and related sites. ______________________________________________________________________________ 138 Figure 6.14: Bos metatarsal measurements from Gomolava, Petnica and other Vina and related sites. ______________________________________________________________________________ 139 Figure 6.15: Bos lower dP4 & M3 measurements from Gomolava and Petnica. _______________ 140 Figure 6.16: histograms and kernel density estimates for Bos metacarpal measurements from Gomolava. _____________________________________________________________________ 141 Figure 6.17: histograms and kernel density estimates for Bos metatarsal measurements from Gomolava. _____________________________________________________________________ 142 Figure 6.18: Bos horn core measurements from Gomolava, Divostin and Petnica, with summary results from Obre II (shown as ovals in chart a). _____________________________________________ 143 Figure 6.19: distributions of Bos post-cranial mean log-ratio values at Gomolava and Petnica.___ 145 Figure 6.20: distributions of Bos log-ratio values at Gomolava and Petnica by previously assigned domestication status. _____________________________________________________________ 145 Figure 6.21: mean postcranial Canis log-ratio values from Gomolava and Petnica ____________ 146 Figure 6.22: selected Canis mandible measurements from Gomolava and Petnica. _____________ 147 Figure 6.23: Cervus acetabular lengths at Gomolava and Petnica by sex. ____________________ 148 Figure 6.24: Cervus metapodial measurements from Gomolava and Petnica. _________________ 149 Figure 6.25: location of Gomolava.__________________________________________________ 150 Figure 6.26: plan of excavation campaigns at Gomolava. ________________________________ 151 Figure 6.27: plan of Gomolava Ia in Blocks I-VI, plus 1967-1968 riverside area. ______________ 154
Figure 6.28: plan of Gomolava Iab in Blocks I-VI_______________________________________ 155 Figure 6.29: plan of Gomolava Ib in 1967-1977 excavation areas. _________________________ 156 Figure 6.30: available radiocarbon dates from Gomolava.________________________________ 159 Figure 6.31: constrained calibration of GrN-7376 (Gomolava Ia). _________________________ 159 Figure 6.32: layout of Block VII. ____________________________________________________ 160 Figure 6.33: reconstruction of features in Block VII at Gomolava by phase. __________________ 161 Figure 6.34: evidence for activities in pits and on the surface in Gomolava Block VII. __________ 167 Figure 6.35: double dog burial in Pit H, Gomolava Block III. _____________________________ 169 Figure 6.36: cattle skull placement in Pit x.3, Gomolava Block I.___________________________ 169 Figure 6.37: probable and possible Gomolava Iab structures in Block VII. ___________________ 172 Figure 6.38: size distribution of all recorded fragments from Gomolava and diagnostic specimens from Petnica. ___________________________________________________________________ 177 Figure 6.39: distribution of faunal sample across pits in Gomolava Block VII. ________________ 179 Figure 6.40: representation of domestic cattle, red deer and pigs by phase at Gomolava. ________ 184 Figure 6.41: distributions of fragment size in pits and the cultural layer at Gomolava. __________ 185 Figure 6.42: taxonomic frequency (%NISP) of major species at Gomolava by phase and feature type. ______________________________________________________________________________ 187 Figure 6.43: raw and corrected distribution of Payne age stages for domestic cattle at Gomolava. 188 Figure 6.44: survivorship curve for domestic cattle at Gomolava. __________________________ 189 Figure 6.45: domestic cattle survivorship curves from Gomolava by (a) phase, (b) feature type and (c) excavation area. _________________________________________________________________ 190 Figure 6.46: aurochs mortality at Gomolava. __________________________________________ 191 Figure 6.47: survivorship curves for wild and domestic pigs from Gomolava. _________________ 191 Figure 6.48: domestic pig survivorship curves from Gomolava by (a) phase, (b) feature type and (c) excavation area. _________________________________________________________________ 193 Figure 6.49: wild pig survivorship curves from Gomolava by (a) phase, (b) feature type and (c) excavation area. _________________________________________________________________ 194 Figure 6.50: red deer mortality at Gomolava by (a) phase and (b) feature type. _______________ 195 Figure 6.51: distribution of pig postcranial log-ratio values by main feature types at Gomolava.__ 199 Figure 6.52: distribution of cattle postcranial log-ratio values by main feature types at Gomolava. 199 Figure 6.53: Grant (1982) MWS scores for (a) domestic and (b) wild pigs at Gomolava. ________ 203 Figure 6.54: mandible wear scores (MWS) against individual tooth wear scores (TWS) for pigs at Gomolava. _____________________________________________________________________ 203 Figure 6.55: location of Petnica. ____________________________________________________ 204 Figure 6.56: location of trenches and approximate extent of settlement at Petnica. _____________ 205 Figure 6.57: trench layout at Petnica. ________________________________________________ 206 Figure 6.58: sample section through deposits at Petnica. _________________________________ 206 Figure 6.59: outline plan of Petnica by phase. _________________________________________ 207 Figure 6.60: AMS date from Petnica 2. _______________________________________________ 208 Figure 6.61: partially excavated oven from Petnica 2, Trench 10. __________________________ 210
xi
Figure 6.62: distribution of finds in Petnica 2. _________________________________________ 214 Figure 6.63: distribution of finds in Petnica 3 (below House 3).____________________________ 215 Figure 6.64: distribution of faunal remains in Petnica 3 (House 3 floor level). ________________ 216 Figure 6.65: frequency of main taxa by phase at Petnica (%NISP)._________________________ 219 Figure 6.66: trends in frequencies of main taxa at Petnica by (a) %NISP, (b) % MNE and (c) %DZ.221 Figure 6.67: taxonomic frequencies (%NISP) at Petnica by context type. ____________________ 222 Figure 6.68: domestic cattle mortality in Petnica 2 & 3 (Payne stages). _____________________ 223 Figure 6.69: survivorship curve for domestic cattle from Petnica 2 & 3. _____________________ 223 Figure 6.70: survivorship curves for pigs at Petnica. ____________________________________ 224 Figure 6.71: caprine mortality at Petnica._____________________________________________ 225 Figure 6.72: red deer mortality at Petnica. ____________________________________________ 225 Figure 6.73: percentage of domestic cattle specimens fused in each fusion group at Petnica, by phase. ______________________________________________________________________________ 226 Figure 6.74: percentage of caprine specimens fused in each fusion group at Petnica. ___________ 227 Figure 7.1: relationship between MAU and median density for domestic pigs at Gomolava. ______ 230 Figure 7.2: median density against corrected NISP for bovids and cervids at Gomolava. ________ 232 Figure 7.3: mean density against corrected NISP for pigs at Gomolava. _____________________ 233 Figure 7.4: density vs. % MAU for pigs from Gomolava. _________________________________ 234 Figure 7.5: relationship between NISP and density by feature type for (a) cattle and (b) red deer at Gomolava. _____________________________________________________________________ 235 Figure 7.6: relationship between NISP and density by feature type for (a) wild pig and (b) domestic pig at Gomolava. ___________________________________________________________________ 236 Figure 7.7: median density against MAU for bovids and cervids at Petnica. __________________ 238 Figure 7.8: median density against corrected NISP for bovids and cervids at Petnica. __________ 239 Figure 7.9: relationship between NISP and density by feature type for (a) wild pig and (b) domestic pig at Petnica. _____________________________________________________________________ 240 Figure 7.10: relationship between NISP and density by feature type for (a) cattle, (b) red deer, (c) roe deer and (d) caprines at Petnica. ____________________________________________________ 241 Figure 7.11: median long bone fragment length versus weathering/gnawing rates in individual units (NSP > 9) at (a) Gomolava and (b) Petnica. ___________________________________________ 244 Figure 7.12: percentage of specimens gnawed in individual units (NISP > 9) at (a) Gomolava and (b) Petnica, split by context type._______________________________________________________ 245 Figure 7.13: rates of gnawing and weathering by taxon in pits and the cultural layer at Gomolava. 247 Figure 7.14: rates of gnawing and weathering by taxon in house remains and the cultural layer at Petnica.________________________________________________________________________ 248 Figure 7.15: length distributions for long bones at Petnica. _______________________________ 253 Figure 7.16: length distributions for long bones at Gomolava. _____________________________ 254 Figure 7.17: long-bone completeness by body size and element at (a) Petnica and (b) Gomolava. _ 255 Figure 7.18: measures of fragmentation by taxon and context type at Gomolava. ______________ 257 Figure 7.19: measures of fragmentation by taxon and context type at Petnica. ________________ 257 Figure 7.20: burning rates at Gomolava (a) by unit (NSP > 9) and (b) by pit (NSP > 19). _______ 260
xii
Figure 7.21: anatomical distributions of burning at Gomolava. ____________________________ 264 Figure 7.22: anatomical distributions of burning at Petnica. ______________________________ 265 Figure 7.23: distribution of cut-marks on cattle-sized mammals at Gomolava and Petnica. ______ 267 Figure 7.24: distribution of cut-marks on pig-sized mammals at Gomolava and Petnica. ________ 269 Figure 7.25: scatter plots showing strongest observed correlations between MAU and FUI at Gomolava and Petnica. ___________________________________________________________ 271 Figure 7.26: body-part profiles for the main species at Gomolava. _________________________ 274 Figure 7.27: body-part profiles for aurochs and dog at Gomolava. _________________________ 275 Figure 7.28: body-part profiles for beaver, aurochs and dog and Petnica.____________________ 275 Figure 7.29: body-part profile for wild carnivores at Petnica. _____________________________ 276 Figure 7.30: body-part profiles for the main species at Petnica.____________________________ 277 Figure 7.31: body-part profiles for cattle and red deer in houses at Gomolava.________________ 278 Figure 7.32: selected taphonomic variables by feature in Gomolava Block VII.________________ 281 Figure 7.33: (a) taxonomic and (b) anatomical (cattle-sized only) profiles by pit at Gomolava, by %NISP. ________________________________________________________________________ 283
xiii
Acknowledgements
A great many people have contributed to this thesis over the three-and-a-half years from conception to completion. The research was undertaken independently rather than under the auspices of any wider project, but it nonetheless depended heavily on support from various quarters. Altogether, about eight months of my time was spent in Serbia, a country of which I had no prior experience but with which I quickly became fascinated. Inevitably, my experiences during fieldwork were not limited to the academic field, but I have restricted these acknowledgements mainly to those who provided help or advice in an archaeological capacity. There are many others not listed below who helped to make my time in Serbia so enjoyable, and to whom I am also very grateful. In Serbia, a first mention must go to Marko Pori, who has been a constant source of constructive discussion, obscure references and practical help. During my time in Novi Sad studying the Gomolava faunal material, Svetlana Blai of the Muzej Vojvodine went out of her way to assist me, facilitating access to the collection and providing me with the space, reference material and expert advice to study it, not to mention helping me to find my feet in the city. Thanks also go to all those at the museum who made my stay a pleasant one, but especially to Tijana Pesterac and Jovan Koledin: without those long morning chats in the library over rocket-fuel coffee my sojourn at the Muzej Vojvodine would have been far less interesting, if possibly shorter. The enormous task of extracting my sample from the vast stores in the museum attic was undertaken with the aid of Andrea Galambos, James May, and Teodora Matkovi over one exhausting week. Andrej Duda Starovi offered me access to the Petnica collection, and in so doing introduced me to the remarkable institution that is the Istraivaka Stanica Petnica (ISP), or Petnica Science Center. I am greatly indebted to all the staff, students and saradnici of ISP from the director, Vigor Maji, downwards. Special mention must go to the programme heads in archaeology during my association with the centre: Bogdana Opai, Voja Filipovi, and Vlada Pecikova; and of course to eljko Mitrovi, who kept me sane during the longer gaps between seminars when the centre resembled a ghost town. eljko Je, director of the earlier excavations at Petnica, provided additional background information about the site and the region. Laboratory assistance during the Petnica study was provided variously by Sonja Vukovi, Marko Pori and Teodora Matkovi, all working purely as volunteers. I am particularly grateful to Sonja, who agreed to help me with the final leg at three days notice, and proceeded to go way beyond the call of duty in helping me to finish the recording under tight time pressure.
xiv
Nenad Tasi kindly took me to Vina-Belo Brdo during my first visit to Belgrade, while Vesna Dimitrijevi provided me with unpublished faunal data from the type-site, along with invaluable information and advice She also helped with some of the trickier identifications from Petnica, and shared her electronic reference manual and recording system with me. I am grateful to Duan ljivar of the National Museum in Belgrade for showing me the excavations at Plonik and the zoomorphs from Belovode. Thanks also go to Julka KuzmanoviCvetkovi and her family for hosting me in Prokuplje during that visit, and to Ksenija Celner for arranging it all. Nikola Krsti and his colleagues at the Belgrade Veterinary Faculty very kindly radiographed some of the mandibles from Gomolava for me free of charge. I am again grateful to Sonja Vukovi for making the necessary connections with the Veterinary Faculty, and for translation. A few people contributed to the success of my time in Serbia in ways which were not strictly academic but which nonetheless demand acknowledgement: to James May and Milo Petrik in Belgrade, and to Nick Gilbert, Milica Vasi and Lazar Strievi in Novi Sad, thank you. Finally, special thanks are due to Marija Kalabi for tolerance of my horrendous accent during Serbian lessons. Kemajl Luci and Tomor Kastrati of the Pritina museum were very tolerant of my speculative visit and vague questions, and provided me with useful information and literature. I would also like to thank Milot, Jim and the others at the YIHR who made my brief stay in Kosovo possible. Back in Cambridge, thanks are obviously due to my supervisor, Preston Miracle, and the project would never have got off the ground without contacts and advice provided by Duan Bori. I am also grateful to John Robb and especially to Lindsey Friedman for commenting on chapter drafts. Invaluable help regarding ArcGIS, OxCal and Illustrator was offered by Rachel Opitz, Cameron Petrie and Chris Stimpson, respectively. Peter Biehl and Mariana Egri kindly provided useful Romanian literature. I would like to thank all my fellow members of the Grahame Clark Laboratory for zooarchaeology, but especially Tony Legge for amusing stories and cautionary tales regarding his own work in the Balkans; Krish Seetah for several conference paper opportunities; and of course Jessica Rippengal. I am also grateful to Peter Biehl for involving me in the atalhyk project. My time in the field faunal laboratory there so far has been a stimulating and enjoyable experience, providing not only a welcome break from this PhD but also numerous ideas to take back to my Balkan data at the end of each season. I would therefore like to thank everyone I worked with in the atalhyk faunal team, but especially Nerissa Russell, who also provided useful information regarding Opovo and its fauna.
xv
The project was funded by a Millennium Scholarship from the University of Cambridge Board of Graduate Studies, supplemented on occasion by small but extremely helpful travel grants from Darwin College. Welcome additional funding (and experience) was provided by Simon Stoddart and Corinne Roughley, who sent bone work and teaching opportunities my way respectively. Some final, miscellaneous thanks: to Clive Orton for everything; to Chris Pearson for rentfree accommodation when funds ran low; to Andrew Shapland and Louise Martin for coorganising the UCL Attitudes to Animals seminar and for introducing me to the work of Philippe Descola; to Jacqui Mulville and Adrienne Powell for the Walk of the Wild Side session at ICAZ 2006; to Lszl Bartosiewicz for comments on the resulting paper, and to both Lszl and Alice Choyke for hosting me in Budapest (not to mention introducing me to the work of Emir Kusturica); to Miljana Radivojevi and Vida Rajkovaa for encouragement and local archaeological knowledge; to Wietske Prummel for information on the Gomolava bones in Groningen; to Ruth Tringham and Mira Stevanovi for plans and information relating to Opovo; and to Naomi Farrington for keeping me going during the write-up.
xvi
Introduction
Chapter 1 Introduction
This project grew out of two concerns regarding approaches to animals in later prehistory. Firstly, the legacy of evolutionist thinking within archaeology had, to my mind, led to a general neglect of research on hunting within farming societies, with few publications on this topic per se set against a vast literature on stock-raising and its implications in the Neolithic and metal ages. Accordingly, I proposed a project with the working title Hunting in Farming Societies, aiming to explore the possible explanations that might be offered for the phenomenon, to set out their likely (zoo)archaeological correlates, and to apply a methodology based on these to one or more case studies. Needless to say, the title soon changed. As I started to conduct preliminary research on the topic, it rapidly became clear that what I had initially considered an unproblematic distinction between wild and domestic animals was in fact the subject of intense debate, which would have to be engaged with in some depth. While I believe that the distinction retains considerable utility, as I hope I demonstrate in Chapter 3, the arguments leading to this conclusion caused me to rethink certain other aspects of the planned study. The idea of cataloguing explanations for Neolithic or metal age hunting was itself the product of the notion that it is somehow out of place, that it needs to be accounted for. Related to this was the realization that to focus the project on wild species alone would be to take an artificially narrow view, discarding a great deal of potential for considering their roles vis--vis domesticates. Instead, the aim of the project (stated in full in Chapter 2) became to consider human engagements with all mammals in a case where both wild and domestic species were zooarchaeologically abundant, while avoiding as far as possible the usual assumptions about the centrality of the latter. In fact, there is probably more material on domestic than on hunted animals in the finished thesis. The initial choice of case study also changed dramatically within the first few months, several dead-ends prompting a shift of several thousand miles and several thousand years to the later Neolithic of the central Balkans, and specifically the Vina culture. This proved a sound move. The Neolithic archaeology of Serbia is fascinating, and the social changes which appear to have taken place across the course of the Vina period provide the basis for what I hope is a productive case study in social zooarchaeology. Access was secured to the fauna from two very different Vina sites, Gomolava and Petnica, the analysis of which forms the core of the thesis. The second original concern motivating the study concerned forms of interpretation. Zooarchaeologists have tended to lag behind the rest of the discipline with regard to
Introduction
theoretical developments, and interpretations of animal remains are still often naively materialist. At the same time, the potential symbolic potency of animals has long been recognised, and in recent years has come to the fore within zooarchaeology. These two strands of interpretation materialist and idealist; economic and symbolic sit uncomfortably together in the literature, especially given the increasingly widespread recognition that the division between economic and symbolic realms in prehistory is artificial. I hope to reconcile these two positions by taking what I consider a realist stance: animals certainly are material resources, and they certainly are symbols. But they are also animals: living beings with which people interact in a variety of ways. This view is elaborated in Chapter 3. Developing a methodology to address such aims is easier said than done. That applied here is something of an ungainly chimera, but I believe it serves its purpose adequately. Traditional economic studies have tended to focus on production, while more socially-oriented zooarchaeologists often approach their interests through consumption and deposition. I tackle the problem from both directions, moving from consideration of coarse-scale economic issues at a regional level down to a detailed taphonomic study of post-mortem treatment and deposition. The interpretation draws on both of these aspects, along with basic ecological information and a little ethnographic analogy. The structure of the thesis is fairly conventional. Chapter 2 sets the scene in terms of the Vina culture, and of narratives of the Neolithic more widely, before setting out the project aims. Theoretical approaches to animals are reviewed in Chapter 3, with particular attention devoted to problematizing the wild:domestic dichotomy, while Chapter 4 outlines the methodological structure and discusses analytical techniques. The study proper begins with Chapter 5, a top-down review of regional-level trends in taxonomic representation, combining all available published and unpublished faunal data as of April 2008 with the latest absolute dating evidence. The latter part of the chapter brings in demographic data to assess likely economic roles played by both wild and domestic species. Chapter 6 moves to a finer scale, introducing the case study sites of Gomolava and Petnica, and their faunal assemblages. These are really two separate site/faunal reports, but are gelled together by the biometric analysis a critical if unglamorous element of the overall study. Chapter 7 continues the reduction in scale, presenting a detailed taphonomic study of the fauna from the two sites. The top-down and bottom-up aspects of the study are brought together in Chapter 8 in order to assess the changing forms of human relationships with different species in the Vina period. Finally, conclusions are summarised in Chapter 9.
Archaeological background
Chapter 2 Archaeological background and aims

2.1 The Neolithic: a brief history
The term Neolithic was coined by Lubbock (1865) to refer to the later part of the Stone Age. This was conceived as a definite chronological period within a universal evolutionary framework, analogous to a geological era (Thomas 1993, 362) and defined by its position vis-vis other periods as much as by its specific attributes. The term rapidly changed in meaning over the ensuing decades, however, coming to be used as an adjective more than a noun: a descriptor of a way of life rather than the name of a single epoch. More-or-less comprehensive histories of the concept are available elsewhere (e.g. Barker 2006, 1-41; Pluciennik 1998; Thomas 1993; Trigger 2006); the aim of the brief outline given here heavily biased towards Europe is to draw out two recurring themes. The phrase Neolithic package has been attributed to Chris Chippindale while an undergraduate at Cambridge in the 1970s (Sherratt 2005, 145), but the definition of Neolithic in terms of a common suite of technological and cultural traits can be traced back much further. Writing within the explicitly universalist and colourfully value-laden paradigm of 19th century evolutionism, Hodder Westropp (1872) equated it with a pastoral stage on the path to higher development, defined both by ground stone tools and by domestic animals. Pluciennik (1998, 62) places the origin of the Neolithic package in the same year with Gabriel De Mortillet, who added agriculture, megaliths and pottery to the mix none actually included in Lubbocks original concept (Thomas 1993, 362-363). By the 1920s the quartet of cultivated plants, domestic animals, pottery and ground stone was fairly well established (e.g. Burkitt 1921, 157). Of these, Childe (1925) stressed the centrality of food production, a decade later (Childe 1936) making it the linchpin of his Neolithic Revolution:
The steps by which mans control [over Nature] was made effective have been gradual, their effects cumulative. But among them we may distinguish some which stand out as revolutions. The first revolution that transformed human economy gave man control over his own food supply Childe, 1936, 66
Although careful to hang neither sedentism nor ceramic technology upon it too closely, Childe clearly saw the adoption of food production as the shift in economic base upon which both social changes and further technological innovations were founded, culminating in his second (Urban) Revolution (1936, 105-106). Both the materialism and the stepped nature of social change inherent in Childes work were in keeping with his Marxist theoretical leanings, as was his rejection of 19th century ideas of progress (1936, 1). This materialism was retained
by Grahame Clark. Gradual processes were emphasized over sudden shifts the Neolithic Revolution was neither a revolution nor was it Neolithic (Clark 1969, 72) but the pairing of domestication and sedentism, in that order, remained the crucial first steps setting in motion a rapid development in the sphere of culture (1969, 70) ultimately culminating in what Clark could still describe as the achievement of civilization (1969, 94). Civilization was replaced by complexity in the more mechanistic neo-evolutionism of the North American New Archaeology, but the universalizing classificatory schemes set out by Sahlins (1968) and Service (1962) never adequately separated the specific, multilinear evolution implied by an adaptationist or ecological approach from the general, unilinear notion of progress inherited from the Victorians (Trigger 2006, 389). The possibility of reversals in such schemes was acknowledged, and value-judgements largely excluded, but by the late 1960s a shift away from food production could still be referred to as devolution (e.g. Lathrap 1968, 29). Across the Atlantic, the Cambridge Palaeoeconomy school was arguably more successful in this respect, advocating a highly case-specific approach to economies and particularly attacking the dichotomy between foraging and food production upon which concepts of the Neolithic were founded (Higgs & Jarman 1972; Jarman 1972; Jarman & Wilkinson 1972). As Barker (2006, 32) notes, however, palaeoeconomists generally reverted to mainstream New Archaeological models to explain the development of the mixed farming regime entailed by most formulations of the Neolithic package. In keeping with wider trends in archaeology, there has been a recent shift from materialist to idealist concepts of the Neolithic. Earlier emphasis on new technologies and changing modes of subsistence has been replaced in some circles by a focus on the creation of new ways of seeing the world. For Thomas (1991, 13), farming is not always the crucial component in a Neolithic defined rather as an integrated cultural system involving the expression of a fundamental division of the universe into the wild and the tame. In effect, the economic package has been replaced by an ideological one (Pluciennik 1998, 73-74). A nuanced variant of the wild:tame theme underlies Hodders The Domestication of Europe (1990). Drawing on the works of Cauvin (e.g. 1972; Cauvin & Cauvin 1984), Hodder sees food production not as the cause but simply as one expression of changing ways of thinking. Specifically, he proposes the formation of the domus, a conceptual sphere relating to the home and to control of the wild, and lying at the heart of Neolithic life. The agrios, concerned with the wild or extra-social, takes shape in contrast. This scheme cannot be mapped directly onto wild:domestic: the domus is as much metaphor as domain, involving the processes of conceptual domestication rather than the domestic per se; accordingly, the position of domestic animals is ambiguous (Hodder 1990, 69). Nonetheless, the structuralist roots are
clear: Hodders Neolithic essentially represents the origin of the familiar nature:culture dichotomy (Pluciennik 1998, 72; Whittle 2003, 93). Whittle offers a less rigidly idealist approach, emphasizing contingency and history within a disparate Neolithic: there was no uniform process, no single history. The Neolithic period is itself a series of becomings rather than the spread of something already formed (1996, 9). More recently (2003; 2005) he has focused on the physicality of everyday experience in Neolithic contexts, a move towards the reconciliation of material and ideal also apparent albeit very differently expressed in Hodders (2006) embracing of material entanglement. James Barrett (2007) has suggested that such developments in archaeology as a whole represent a new, more sophisticated materialism; I see them rather as a welcome trend towards a balanced realist stance.
2.2 Hunting versus herding

The above potted history of the Neolithic concept has deliberately skirted around the questions of why and how the Neolithic occurred and spread, since this project is not directly concerned with origins, either globally or locally. Even the issue of its coherence and analytical validity/utility is peripheral here. Rather, I am interested in the implications that narratives of the Neolithic have for human-animal relations in later prehistory, and particularly for the interpretation of zooarchaeological data. To this end two recurring themes can be noted. Firstly, there is the idea of humans emerging from and coming to control their environments: from Hodder Westropp to Ian Hodder excepting the palaeoeconomists the onset of the Neolithic represents the crucial moment in a pervasive archaeological metanarrative concerning the separation of nature and culture (Pollard 2006, 136). The domestication of plants and perhaps more emotively animals has inevitably been a crucial strand in this narrative. The second, related, theme is evolutionism. Despite the widespread rejection of overtly social-evolutionist views, there remains a deeply ingrained tendency to view foraging and farming as alternatives, the latter replacing the former. Unilinearity may (arguably) have been banished, but uni-directionality remains: rather than seeing wild resources as technologically obsolete we increasingly see them as conceptually obsolete, rendered peripheral by the creation of a domestic sphere. While no-one denies that wild and domestic resources are often exploited alongside each other, prehistoric societies are typically characterised as either foragers or farmers. This essentialist categorisation serves not only to imply a uni-
directional progression, but also to elevate subsistence to the position of primary determinant in group identity (Bori 2005, 16-18; see also Pluciennik 1998, 67). One result of these two themes is that while the causes and mechanisms of domestication are subject to endless debate, the concept of domestic itself is sometimes taken for granted, by zooarchaeologists at least. This issue is explored in some detail in 3.2.2. A second result is that human involvement with wild animals in the Neolithic and later periods appears out of place, especially in areas where domesticates were apparently introduced: if the Neolithic is about domesticates then wild species must be secondary, their frequent presence requiring special explanation. This is perhaps an unfair caricature, since the Mesolithic-Neolithic transition is in fact the subject of much subtle debate (e.g. Bori 2005; Kotsakis 2005; Tringham 2000b; Zvelebil 1998). Moving beyond the leading edge of the Neolithic, however, there remains a widespread if by no means universal reluctance to challenge the peripheral place of wild species. In small numbers they are frequently ignored; in large numbers often explained away as special cases. An apparent lag in the adoption of domesticates in parts of Europe led Zvelebil & RowleyConwy (1984; 1986, see also Harris, 1996; Zeder 2006) to propose a model in which the transition to food production is seen as a three-stage process (availability, substitution, consolidation) rather than an instantaneous change. A model for declining use of wild animals in particular was later put forward in a very similar vein (Zvelebil 1992, see 5.3.1). These models remain strictly evolutionary and uni-directional, however (Pluciennik 1998, 68), leaving no room for reversals. In fact, hunting within the Neolithic cannot always be placed within a pattern of long-term decline. In some areas the contribution of wild animals to faunal assemblages increases significantly over the course of the Neolithic, in others it remains steady throughout the period, and in some the predicted pattern of decline is observed. In still other cases wild species almost drop out of the record with the appearance of domestic animals but re-appear later. Given the range of trajectories observed, prescriptive models of a one-way transition however gradual from Mesolithic hunting to Neolithic stock-keeping clearly cannot be sustained. The focus of this project is not, in fact, specifically on wild animals. Rather, it aims to investigate the changing economic and social roles of all animals in one particular Late Neolithic context where wild species appear to retain a considerable importance, while avoiding, as far as possible, the assumptions and prejudices stemming from the narratives outlined above. The case study is introduced below.
2.3 The Vina culture

2.3.1 introduction
The Vina culture is a later Neolithic phenomenon in the central Balkans spanning some seven centuries in the approximate period 5300-4600 cal. BC (see below) and defined on the basis of ceramic typology. The later part of the period is sometimes placed within the Copper Age (e.g. Sherratt 1984; Tasi 1998) on the basis of evidence for copper mining, production and use (Jovanovi 1971a; 1990; Jovanovi & Ottaway 1976; ljivar 1996). However, with the earliest metallurgy being pushed ever further back in time (Bori & Jovanovi in press; pers. comm. M. Radivojevi) most of the period could now be called Copper Age on this basis. The entire Vina period is here labelled Neolithic since this is the term used most commonly in the literature and is, after all, only a label. Named after the type-site of Vina-Belo Brdo on the Danube just outside Belgrade, the Vina culture extends over a large area of the Balkans, including all of Central Serbia; Kosovo; the southern part of the Vojvodina; the Romanian Banat and parts of Hungary immediately north of the Maros; the Iron Gates (erdap) region; parts of Oltenia; the far north-west of Bulgaria; northern Macedonia; and the easternmost areas of Slavonia and Bosnia1 (Chapman 1981; Markoti 1984). In addition, the Turda variant of early Vina clusters around the Mure in Transylvania. Within the main area of extent, settlements are concentrated along the floodplains and lower terraces of the Danube, Sava, Morava and their tributaries, and on the Kosovo and southernmost Great Hungarian Plains (see Figure 2.1 for locations of regions and rivers). Chapman (1981) has defined numerous regional and local groups within the overall Vina complex. Whilst acknowledging that typologically-defined cultures cannot be taken to represent meaningfully bounded entities, I nonetheless consider the recognized spatial and temporal limits of the Vina culture preferable to a purely arbitrary case study area. Across most of its extent the Vina period is preceded by the Starevo/Krs/Cri complex, which represents the earliest Neolithic in the region. Apart from the erdap area and various cave sites in the Dinaric Alps, the Mesolithic of the region is very poorly known (Bori 2005, 18-21). Neolithic sites appear earliest in the south, reaching southern Serbia by about 6200 cal. BC, and are known in the southern Hungarian Plain by 6000 cal BC (Whittle et al. 2002; 2005). The mechanisms by which food production and ceramic technology reached the region remain a matter of debate (e.g. Bori 2005; Tringham 2000b; Whittle et al. 2002), especially with regard to the Iron Gates (e.g. Bonsall et al. 2000; Bori & Dimitrijevi 2006; Bori &
No political statement is intended by choices of names and spellings. Likewise, modern national and provincial boundaries are omitted from all maps as irrelevant to prehistory. Apart from the Danube, rivers are named in the language of the country in question in each instance.
1
Miracle 2004), where the traditional association of pottery and permanent architecture with domesticates cannot be sustained. The issue is peripheral here, but one can note that neither a pure migrationist nor a strict diffusion/indigenous adoption model seems likely for the region as a whole.
Figure 2.1: map of the central Balkan region with place names mentioned in the text. (Thanks to Duan Bori for base map).
The long-standing argument that the Vina culture represents a further wave of external influence (e.g. Garaanin 1984; Miloji 1949) has been widely rejected in favour of indigenous development from the preceding Starevo group (Lekovi 1990; Makkay 1990; Srejovi 1988, 12), although the idea of an Anatolian origin has had something of a revival (Garaanin 1998, 65-67; zdoan 1993; 1997). Recently, Tringham (2000b, 51-53) has suggested that the essentially foraging communities of the later Lepenski Vir culture in
erdap contributed significantly to the emergence of what she sees as fully agricultural, sedentary Vina settlements, through various forms of interaction with Starevo farming communities. Vina sites vary considerably in size, with the largest settlements, such as Selevac, emerging in the middle part of the period (Vina B-C see below), followed by population dispersal and reduction in average site size (Chapman 1990; Tringham 1992). Smaller sites exist throughout the period. Tell settlements are less common than in the eastern Balkans (Bailey 2000, 161). The implications of settlement and other evidence for mobility are discussed in 2.3.4. The anatomy of Vina sites is considered in some detail in Chapter 6 in the context of Gomolava, but in general settlements consist of larger rectilinear wattle-and-daub buildings interpreted as houses and a few smaller structures whose function is unknown but might involve storage or animal penning (6.3.4.3). The initial phase of many sites is characterized by a paucity of apparent dwellings and an abundance of large pits (Tripkovi 2003, 450), rather resembling the typical Starevo case. The resultant interpretation of the largest cutfeatures as semi-subterranean dwellings is not accepted here (see 6.3.4.2). Houses vary in size both between and within sites, and some very large examples are known. A structure with elaborate internal subdivisions and numerous clay animal heads and cattle skulls has been interpreted as a shrine at Para (Lazarovici 1989; Lazarovici et al. 2001) and is taken by Tripkovi (2003, 455) to indicate authority above the level of households. Organization at a communal level is also implied by the massive ditches at Uivar (Schier 2005; 2008), less substantial fortifications being known from a few other sites (Markoti 1984, 17). The overall impression, however, is an absence of communal structures or central foci. Certain spaces within houses are often interpreted as cult areas on the basis of artefact associations and, occasionally, bucrania (e.g. Petrovi 1993, 10). Most known Vina houses were destroyed by fire, although since these are usually easier to recognise than unburnt structures this may be misleading (Schier 2008, 56-57). Houseburning certainly becomes more common towards the end of the period, as is seen clearly at multi-phase sites (Tringham & Stevanovi 1990, 114-117; Tripkovi 2003, see 6.3.4.3), accompanying increasingly substantial construction (Tringham & Krsti 1990b, 583-584). Several theories have been put forward to explain the burnt house phenomenon, from catastrophic accidents (McPherron & Christopher 1988, 477-478) to marauding IndoEuropeans (Gimbutas 1974) and deliberate burning as a ritual of house death (Stevanovi & Tringham 1998; Tringham 1991; 2005; Tringham & Stevanovi 1990, 116). The interpretation rests on two principal questions: (a) were house-burnings individual or en masse, and (b) were they deliberate or accidental? Stevanovi (1996; 1997) has shown in
my view conclusively that burnings were both individual and deliberate. Most importantly, the temperatures involved would have been very hard to achieve intentionally, let alone by accident. This is not in itself proof of the house death hypothesis, but the abundant presence of artefacts in burnt houses McPherrons and Christophers principal argument against the theory certainly favours a ritual context. Tringhams and Stevanovis explanation is accepted here. One of its more important, if more prosaic, implications is that house inventories cannot be taken as a direct representation of typical house contents. In terms of subsistence, Vina sites feature the typical range of domestic cereals and legumes found in the temperate European Neolithic (Hopf 1974; Jovanovi 2004), but the scale and manner in which these were cultivated, and the contribution of wild taxa, are matters for debate. Likewise, in the absence of stable isotope data it is impossible to state the relative importance of plant foods and animal products. Borojevi (2006, 149) argues that fields around Opovo where wild plants are also very important were probably small and dispersed, perhaps as much as three kilometres away from the site, and that crop rotation is a possibility. Chapman (1990, 37-39) proposes that successive phases at the larger site of Selevac saw a change from shifting garden plots amongst the houses to fixed fields adjacent to the settlement, but this is based on an as yet undemonstrated early adoption of the ard. Whatever the cultivation regime, clearance was probably very limited: anthropogenic impact on the environment appears minimal at this point, both around the large Vina site of Gomolava (van Zeist 2002, 112) and in the wider Balkan region (Gardner 1999; Willis & Bennet 1994). Large concentrations of molluscs snails and mussels are found at sites such as Gomolava (personal observation; Clason 1979b), Uivar (Beigel & Kuhn 2005, 46), and Vina (pers. comm. M. Pori) and must have been at least periodically important in the diet. Bird and fish bones are rarely reported in any numbers but may be victim to a common lack of sieving; fishing equipment is known amongst the bone artefacts (e.g. Russell 1990, 530; Tringham 1971, 180-183). The data concerning human use of and engagement with mammals are considered in the remainder of this thesis.
2.3.2 chronology
The chronology of the Vina period is primarily reckoned according to ceramic typology, extrapolated from the deep stratigraphic sequence at the type-site for all but the latest periods. Of abundant typological schemata (e.g. Holste 1939; Lazarovici 1981; Schier 1996), the most popular are those of Garaanin (1951; 1979) and Miloji (1949). The latter, used here throughout, consists of stages A-D with further subdivisions (Table 2.1).
10
Garaanin 1974
Miloji 1949 Start
Chapman 1981, 18; b.c. End Range (BC)* 4260-4240 5450-5100
Tringham & Krsti 1990, 54; BC Start End ~5250 5000
Lenneis & Stadler 1995, 7; BC Start End 5300 Start** 5317-5267
New estimates see below; BC End** Range*** 5276-5214 5295-5250
Tordo I Tordo II Gradac Plonik I Plonik IIA
A B1
4500-4450
4260-4240 B2 C1 C2 D1 3950-3850 D2 4100
4100
5100-4975
5000
4700
5300-5205
5000-4904
5255-4950
3950-3850
4975-4720
4700
4500
5039-4919
4886-4740
4970-4800
Plonik IIB
3300
4720-4020
4500
4000
4500
4931-4860
4635-4553
4870-4600
Table 2.1: summary of relative and absolute chronology for the Vina period. *approximate calibrated ranges based on Chapmans b.c. figures. **68.2 % HPDs for phase boundaries. ***between boundary modes.
Reviewing the earliest available radiocarbon dates, Chapman (1981, 19) argued that Vina D covers roughly the same span as the preceding three phases, proposing a simplified Early/Late Vina division in reaction. His estimated calendar age ranges for phases A-D are given in Table 2.1 along with other authors. More recent reviews of the Vina radiocarbon chronology are largely restricted to its overall span (Breunig 1987; Ehrich & Bankoff 1992; Lenneis & Stadler 1995), although Schier (1996; 2000) uses AMS dates to support a revision of the typological sequence at the type-site, and Glser (1996) discusses these in the wider Vina context. Obeli et al. (2004) give estimated ranges for Vina A/B, B/C and D, based on one site each. These papers judge date spans either informally or using the OxCal sum function, both of which techniques systematically overestimate duration (Bayliss et al. 2007, 8-10). If trends in faunal data are to be assessed, it is clearly necessary to reassess the absolute chronology here. Of 154 available Vina culture dates, 113 are attributed to a specific phase (Appendix 1; the Banat variant is included here but Turda and Dudeti-Vina excluded). All were included in a simple model using OxCal 4.0.5 (Bronk Ramsey 1995; 2001; 2007; calibration data from Reimer et al. 2004), apart from ten having > 100, and eleven from Uivar (Schier 2008) which were not available in time. Each of phases A to D was treated as a bounded group, ignoring subdivisions. D1 and D2 would ideally have been taken separately but few dates derive from the latter, while many sources are not specific. As products of typological seriation the phases are expected to overlap and are only constrained to start after the preceding phase starts and end before the succeeding phase ends (see Buck et al. 1996, 217233). Intra-site stratigraphic information was not included as potential pitfalls were judged to outweigh likely benefits. Extreme outliers were identified from agreement indices and
11
excluded. Unfortunately, this reduces the Vina A sample to eight, all but three from Belo Brdo. The calculated boundaries for each phase are given in Figure 2.2. Vina A is surprisingly sharply-defined, although this must relate to (a) most samples coming from the same site, and (b) a pronounced wiggle in the calibration curve at 5200 cal. BC. The transition from B to C is also fairly well-defined. The considerable overlap seen between C and D relates almost entirely to samples from phase Ib at the site of Gomolava, which is described as Vina C in earlier publications (Brukner 1980a) but D more recently (Brukner 1990). Removing these samples eliminates the overlap but obviously creates a false impression of the reliability of the ceramic chronology. Placing them in Vina C simply shifts the period of overlap, confirming Gomolava Ibs position on the border between the two phases. The boundary for the end of Vina D is quite narrow but must be treated with caution since few samples are associated with D2 assemblages. The Central Balkan Copper Age after Vina D1 is very poorly dated overall, so the end of the Vina culture remains relatively obscure.
Figure 2.2: estimated start and end dates for Vina phases.
Estimated start and end points for each phase are summarised in Table 2.1. In general overlap is surprisingly limited, especially since many dates are on charcoal. In at least some cases these are likely to overestimate age significantly, presumably blurring boundaries between phases. AMS dates are currently only included from Schier (1996) and Bori and Jovanovi
12
(in press), but funding has been secured for 26 further bone samples from Opovo, Gomolava and Petnica. In the meantime the ceramic chronology appears adequate at least to divide the period into A-B and C-D, and these groupings are used here throughout.
2.3.3 models of social change

Over the last twenty-five years, a series of works has been published on the subject of social change in the Vina period. Although some of the authors might disagree, I see most of these as developments and/or refinements of a model set out by Tringham et al. in 1985, with roots in a paper by Kaiser and Voytek two years earlier. While the language has changed and successive formulations have drawn on different theoretical approaches, the same essential processes are described. Kaisers and Voyteks original paper argued for intensification in non-subsistence (i.e. ceramic and lithic) production on Vina sites in response to increasing sedentism and population growth (Kaiser & Voytek 1983). The changing social relations argued to be bound up with this intensification were not spelt out in detail, but the importance of production at a household level was stressed (1983, 342-345). Tringham et al. (1985, 427) build on this model in suggesting that the period sees the emergence of the household as the primary unit of social and economic organization, at the expense of the community. This theory is elaborated in later works (Tringham 1991; Tringham & Krsti 1990b, 602-615), where Tringham argues that the increasing independence of households was as much a stimulus as a result of sedentism. Evidence cited in support of the model includes larger sites, increasingly substantial houses, and greater frequency of house burning, as well as the artefact-based argument from Kaiser and Voytek. Russell (1993, 10-14, 472-482) emphasises changing property relations, suggesting that the Vina period saw the development of accumulation at the household level and of ideas of private property, resulting in incipient social ranking that was not fully realised until subsequent periods. She applies her variant of Tringhams model to human use of animals, arguing, amongst other things, that intensification of production should apply only to those species for which it can be restricted, i.e. domesticates (1993, 12-13). The emphasis on accumulation and the causal sidelining of sedentism is reflected also by Chapman (1992, 312), who sees the quantity and diversity of apparent household possessions in the period in terms of status negotiations in the domestic arena. This created a tension between the accumulation of wealth and an egalitarian ethos, a growing contradiction which eventually prompted a shift to the mortuary arena in final Vina and subsequent Copper Age contexts. This specific argument is undermined somewhat by the fact that a single Vina D cemetery is
13
known, at Gomolava, while the only other Vina example, at Boto, dates to the earlier part of the period (Chapman 1981, 55-58). The basic theory of increasing focus on the house and probably the household is accepted for the 5th mill. BC Balkans in general by Bailey (2000, 165). Hodder reformulates the hypothesis of intensified domestic production in terms of his domus:
As social and economic competition and intensification increased they did so in terms of the domus concept. The latter reinforced the domestic locus of production while at the same time becoming reinforced and elaborated itself. Hodder 1990, 53
Similar arguments to Tringhams have been advanced with respect to the Greek later Neolithic and earliest Bronze Age by Paul Halstead and colleagues (Halstead 1999; 2007; Pappa et al. 2004). As in Tringhams model, emergent household autonomy features highly. Interestingly, Halstead (1999, 86) reverses Russells argument regarding wild animals, suggesting that increasing quantities of wild taxa in the Final Neolithic represent a response to relaxed sharing obligations. As Hamilakis (2003, 242) notes, however, this assumes a privileging of material gain over social relationships that can hardly be sustained for the period. Increasing competition between households is likely to have intensified rather than suppressed exchange. The potential of exchanges and/or communal consumption to act as cohesive mechanisms is stressed in later works (Halstead 2007, 26-27; Pappa et al. 2004, 16). The idea of tension between household- and community-level social organization in both the central and southern Balkans has been addressed through studies of household variability at Banjica (Tripkovi 2003) and in Thessaly (Souvatzi 2008), respectively. In the latter case, the author accepts considerable household autonomy but argues that dependence on communal organization actually increases over time. One might relate this to the incipient development of institutionalized authority around the Final Neolithic-Bronze Age transition in Greece (see e.g. Hamilakis 2003). By contrast, internal differentiation remains limited on Vina sites, with no apparent foci save at Para, and settlements typically become smaller and more dispersed after around Vina C2 (Brukner 2003). Tringham (1991, 114, 1992; see also Bankoff & Greenfield 1984, 17-18) argues that the social tensions and contradictions created by
increasingly autonomous and competitive households within expanding communities could have been dealt with either by the development of a centralizing dominance structure or by fissioning, i.e. disaggregation of households into smaller communities (1992, 141). The latter option was apparently chosen in the Vina case. There are obviously differences of emphasis and detail amongst the authors cited thus far, but the development of a broad consensus can be traced. A dissenting voice comes from Whittle,
14
who questions (a) the role (and nature) of sedentism, (b) the applicability of the notion of household, particularly its equation with physical houses, and (c) the interpretation of abundant material culture in terms of competition rather than of social relationships and transactions, involving exchange, gift-giving and the provision of hospitality, designed to consolidate a sense of community (1996, 105). These are pertinent criticisms, but are not as irreconcilable with the Tringham model as they might initially appear. Sedentism is in fact only a crucial cause of social change in Kaiser and Voyteks formulation, but is nonetheless clearly central to the debate, and is discussed in the next section. The danger of equating house with household is frequently expressed, if rather less often heeded. Tripkovi (2003, 447-448) notes that multiple households may dwell in a single building, and one might add that the converse is also eminently plausible. The search for households matching a particular historically- or anthropologically-derived form is probably misguided both theoretically and practically, but given the size of some Vina communities smaller corporate groups must have existed, and these are likely to have been associated in some way with buildings or groups thereof. Tringham and Krsti (1990b, 603) state that household is intended as shorthand for a kin-based co-resident domestic group smaller than a village. Less prescriptively still, Souvatzi (2008, 18) defines household as a social group cooperating in a sphere of social, economic and ideological practices consisting minimally of production, distribution/consumption and social reproduction, although co-residence is also implicit in her analysis. It is in this sense that I use the term. The impossibility of identifying such groups with certainty does not negate their existence, and indeed Whittle (1996, 105) allows for varying interests within a settlement. Tringhams and Chapmans arguments entail an increasing tension between the authority of the community and of households as thus defined, certainly not the complete autonomy of the latter. Finally, with regard to rival interpretations of abundant Vina material culture as either competitive or cohesive, I would argue that this is a false dichotomy unless one posits commodification of artefacts. In setting out his fragmentation model of social relations in the Balkan Neolithic and Copper Age (NCA), Chapman (2000a, 47-48) argues that the south-east European Neolithic was characterized by gift-exchange involving inalienable objects, creating enchained relations between people. Such exchange is nothing if not competitive, but the competition revolves around the social relations materialised by objects rather than the objects themselves, and so also has a cohesive potential. As the value of objects per se starts to exceed the significance of these relationships, Chapman continues, accumulation emerges as an alternative way of relating. This is seen as essentially a Copper Age (i.e. post-Vina) development, but tension between personal or household accumulation
15
and corporate kinship relations exists throughout the Neolithic, the former being kept under control by the levelling mechanism of enchained relations (2000a, 47). The apparent intensification of production during the Vina period could be taken as a heightening of this tension, with increasingly intense gift-exchange being a corollary of incipient accumulation. Chapmans specific arguments regarding breakage, re-use, and deposition of ceramic artefacts have been criticised from several directions (see Chapman & Gaydarska 2007, 6-8), particularly on grounds of excavation and recovery quality (Bailey 2005, 111-112). Having worked with material from large-scale Balkan excavations, I must confess to sharing Baileys scepticism. Nonetheless, the underlying social theory derived primarily from Strathern (1988) and Weiner (1985; 1992) is sound and represents, in my opinion, a fruitful way of looking at social relations in the Balkan NCA.
2.3.4 sedentism
Sedentism is a difficult topic in archaeology, due to both terminological (see Rafferty 1985) and evidential ambiguities. There has been a traditional equation of the Mesolithic with mobility and the Neolithic with sedentism (Whittle 1996, 6), one that has its roots in evolutionist thinking (Milner 2005, 32; Tringham 2000a, 120), although opinions have varied as to the direction of causality. However, just as it is now recognised that foraging does not necessarily imply residential mobility (Kelly 1992), so the connection between food production and settling down in the European Neolithic has been challenged (e.g. Bailey 1999b; Whittle 1996; 1997). Considerable confusion is introduced by the different scales at which the terms sedentary and mobile are used. Firstly, there is the question of temporal scale: does sedentism imply year-round presence at a site, year-to-year occupation, or inter-generational fixity of residence? Conflation of sedentism with permanence would be a mistake: people might lead very sedentary lives for several years between wholesale translocations (sedentism without permanence); conversely they might spend only a small part of each year at a site, but do so for generations (permanence without sedentism). Likewise, the social scale of mobility may vary. Even if a site is occupied by some people at any given point in the year this does not mean that the lives of all, or even most, of its inhabitants can be considered settled. Segments of Neolithic populations may have spent varying lengths of time engaged in activities away from a central site, perhaps in smaller camps, as amongst logistically mobile foragers (Kotsakis 2005, 12). Some such camps may themselves have had a degree of year-toyear permanence. On a different scale, individual people must sometimes have moved residential focus between sites, not least for purposes of exogamy. In response to this potential variability, Whittle (1997, 21-22) lists six loose, overlapping forms of
16
mobility/sedentism, of which residential or circulating mobility (#1) and embedded sedentism (#6) represent the traditional Mesolithic and Neolithic stereotypes respectively. Given its generally small, short-term sites, there is a degree of consensus that the earliest Neolithic of the Central Balkans involved considerable mobility on some or all of these scales (Bailey 2000, 57; Barker 2006, 352-357; Greenfield & Jongsma 2008; Tringham 2000b, 4041; Whittle 1996, 69), although the exact patterns will never be easy to discern. The existence of larger, long-term settlements, and particularly tells, in subsequent periods as throughout the Neolithic in the eastern and southern Balkans stimulates rather more disagreement. Such sites indicate permanence of a sort, although truly continuous occupation is practically impossible to demonstrate, but as noted above settlement permanence and sedentism are not coterminous. The range of interpretations put forward for long-term Balkan sites therefore varies from fully sedentary villages, with some mobility around them (Halstead 2005; Tringham 2000b, 51) to anchor points in a pattern of radiating mobility (Whittle 1996). The disagreement is clearly a matter of degree rather than of kind. Probably the most extreme antisedentary argument is put forward by Bailey, who argues that eastern Balkan tells in particular were foci for periodic congregations of people engaged in very specific activities, and thus represent visible statements alluding to a permanence of place that did not in reality exist (1999b, 97). While valuably provocative, this entails no less of an assumption than does the traditional equation of tells with mixed farming and year-round occupation, the latter seeming more parsimonious in some cases (Halstead 2005, 49). In the Vina case, Baileys argument is hard to sustain. Tells are rare, sharing the landscape with both shorter-term sites and long-term flat settlements. Even if one proposed a special function for tells such as Vina, Gomolava and Uivar, it is not clear where extensive sites such as Selevac or Plonik with their presumed lateral displacement of occupation would fit in. The increasing investment in residential structures could represent a rather disingenuous statement of commitment to place, with actual residence by all or most members of the domestic group sporadic, but there is simply no reason to believe that this is the case. It is also notable that very substantial structures are seen on smaller sites (e.g. Petnica 3: see 6.4.3), and continue to be built in Vina D (Divostin II, Banjica), after the disaggregation of the largest sites. In any case, a wide range of activities is attested on Vina sites, and the sheer rate of cultural layer (see 6.3.4.1) deposition would be hard to reconcile with seasonal or otherwise periodic aggregations. Having said this, I would certainly not preclude considerable mobility by segments of the population on a seasonal or other basis. Some of the main activities that are likely to have occurred in the landscape at a greater or lesser distance from settlements involve the tending of animals. Accordingly, the recent unsettling of the Neolithic has led to suggestions of highly mobile herding sometimes 17
labelled pastoral in the Balkan early Neolithic especially (e.g. Bailey 2000, 133-134; Barker 2006, 354; Greenfield & Jongsma 2008). While the literal meaning of pastoralism is broad enough to render it practically useless, the word has several unfortunate connotations, namely (a) large-scale, extensive herding, (b) specialization, and (c) functional separation of herding and arable farming. Greenfield and Jongsma base their case on the dominance of mobile domestic fauna, i.e. caprines (mostly sheep) and cattle (2008, 120). Potential for confusion between scales resurfaces here: while cattle and caprines would probably be easier to move between sites than pigs, sheep can hardly be considered suitable for extensive herding in the still heavily-wooded environment of the Neolithic; the mobility implicated here is moving on rather than moving around, in Whittles (1997) terms. The model of intensive garden cultivation set out by Bogaard (2004; 2005; see also Halstead 2006), in which domestic animals are few and sheep at least are grazed primarily on cleared land, currently seems most plausible for the Balkans as a whole, at least in the Early Neolithic. While large-scale herding of the kind implied by pastoralism is thus unlikely, there remains considerable scope for more mobile, extensive herding of cattle, which are relatively common in the area between the Hungarian Plain and the Ove Pole. This issue is considered further in Chapters 5 and 8.
2.4 Statement of aims

The intention of this thesis is to investigate the changing forms of human-animal relations within the Vina period, as one example of a later Neolithic case in which both wild and domestic taxa appear from previous studies to have been of considerable importance. Humananimal relations obviously cannot be divorced from the sphere of human-human interactions, and two primary aims can thus be set out: To document changes (or lack thereof) in the importance of different mammalian species across the course of the Vina period, assessing the nature of human use of/engagements with them as far as is possible. To relate these changes to the arguments for social developments outlined above, with particular attention to the roles of wild vis--vis domestic species. These aims are pursued through a zooarchaeological study of published and unpublished material, starting from a review of all available data across the region then zooming in to the first-hand study of faunal remains from two very different Vina sites, and finally to a comprehensive taphonomic study exploring post-mortem treatment and deposition. The different scales of analysis are then brought together for interpretation in the final chapter. First, however, it is necessary to review theoretical approaches taken to animals in archaeology, and the relevance of these to Vina period social change.
18
Theoretical perspectives on animals
Chapter 3 Theoretical perspectives on animals in the Neolithic

3.1 Animals in archaeology: a critique
Human societies engage with animals in a bewildering variety of ways. At different times and in different places animal species have played the roles of companion, of resource and of competitor or threat to human groups. They have provided material resources such as meat, milk and fur, as well as less tangible benefits including labour and protection. They have participated in human social relations as prestige items or units of wealth and have constituted a rich source of symbolism. Equally importantly, they form an active and often highly visible element of the physical environment in which humans live. The importance of these myriad engagements has quite rightly meant that animals have featured highly within archaeological studies, especially since the move towards archaeology as anthropology from the 1960s onwards. However, archaeological approaches to animals have often been under-theorised. Prehistorians, and especially those working directly with faunal data, have traditionally interpreted evidence for human-animal relations in strongly materialist terms. Whether wild or domestic, animals are typically viewed as material resources to be exploited by human groups. This is perhaps unsurprising given that zooarchaeological method has its roots in the processual movement (Grant 2002, 17), but is nonetheless unhelpful. The materialist view can be criticised from a number of directions. Firstly, to focus exclusively on animals physical products is to adopt an extremely narrow conception of the role they play for human groups. Not only are the animals themselves reduced to a passive resource, but their social and symbolic significance for humans is ignored, and a whole realm of prehistoric experience is thus rejected as unworthy of study. Secondly, the overwhelmingly utilitarian emphasis in traditional zooarchaeology has very often been expressed through ideas of rational maximization or optimization. In studies of hunter-gatherer societies these have most often been borrowed from evolutionary ecology and Optimal Foraging Theory (OFT), while for farming societies the language of microeconomics is more evident. The fundamental opposition of these superficially similar bodies of reasoning has been demonstrated by Ingold (1996b, 48), but both typically suffer from the same basic flaw: the assumption that we can talk in universal terms of what is rational, i.e. that the utility of resources is a cross-cultural constant. The shortcomings of OFT highlighted by Ingolds analysis are mirrored by the economic reasoning prevalent within zooarchaeological work on farming societies, although in the latter
19
case the assumption of rational maximization is rarely stated explicitly. Rather, it is implicit in the analysis of kill-off patterns, body part profiles and so on, an analysis which tends towards the formalist in the sense that little recognition is made of the socially embedded nature of both production and exchange. The terminology and concepts employed in the analysis of zooarchaeological data do admittedly become increasingly substantivist as ones research interests move from production to consumption, and indeed the development of an archaeology of food rather than merely of calories can itself be seen as a step forward. However, this interest in the social context of animal use has made little progress back through the production cycle and remains a marginal concern within zooarchaeology. In fact, the situation within traditional zooarchaeology is worse even than a substantivist critique might suggest, since the emphasis on individual decisions (real or metaphorical) at the heart of both OFT and formalist economics is often lost. Perhaps the most serious shortcoming of many avowedly economic approaches to prehistoric human-animal relations is a strong implicit assumption that decisions about appropriate activities are made on a site or even regional level, as if by some peculiarly rational guiding deity, rather than by individuals or small-scale corporate groups (but see Bogucki 1993; Halstead 1999; Russell 1998 for worthy exceptions). This scale of analysis betrays a functionalism that removes human motivations from consideration the emphasis on rational optimization remains, but it is unclear where this rationality is supposed to reside. Economic man, for all his theoretical shortcomings, is replaced by the yet more dubious concept of the economic settlement. More recently, interpretations of animal remains in terms of symbolic behaviour always more popular within social anthropology and cultural geography have come to the fore in archaeology. Numerous variants on Levi-Strausss classic dictum that animals are bon penser (1962) some more pithy than others have been offered as (zoo)archaeologists scramble to recognise the truism, long accepted amongst anthropologists (e.g. Tambiah 1969), that human use of animal species will be influenced by their culturally-specific symbolic associations as well as their physical and behavioural properties. This is clearly a welcome development in breaking the dominance of rationalist interpretations of prehistoric humananimal relations. However, the emphasis on animals as symbols with the anthropological literature raises the danger of replacing a materialist reductionism with an idealist one (Knight 2005b, 1); animals may become simply a mirror for human society (Mullin 1999). The dichotomy between (formalist) economic and symbolic approaches to animals is clearly problematic given the increasing realisation within archaeology, recently articulated by Bradley (2005), that we cannot make a distinction between economic and symbolic behaviour, just as we cannot presuppose a separation of domestic and ritual spheres. Animals
20
in prehistory find themselves caught between two rival forms of generalising interpretation: on the one hand the shared and generalised structures inherent in many symbolically-driven interpretations such as Hodders (1990), and on the other hand the top-down economic reasoning of traditional zooarchaeology (see Shanklin 1985, 378-379). Whether viewed as walking larders or as walking symbols, the animals themselves run the risk of being objectified to the extent that we lose any concept of their basic animality what Whittle (2003, 90) has referred to as their ubiquitous physicality or of peoples actual experiences of, and relationships with, them. In the words of Yannis Hamilakis, environmental archaeologists, including zooarchaeologists, are caught between the rock of empiricism and scientism, and the hard place of social constructivism (2001, 30). The objectification of animals in archaeology has been criticised by Ray and Thomas (2003), who argue that the dominant approaches to animals in prehistory are founded on an ontological separation between humans and non-humans in the Cartesian worldview, a separation which is unlikely to have characterised the experiences of Neolithic people. Accordingly, Pollard (2006, 136-137) calls for recognition of the fluidity of human-animal relations, noting that the cosmological status of animals will have been influenced as much by their involvement in social relations as by any abstracted symbolic scheme. This problem reflects a more fundamental dilemma in how we approach human/non-human interactions, between relativist and universalist epistemologies. The materialist reasoning which has dominated traditional zooarchaeology ignores the role of culture in influencing how people interact with their environments, implying that all employ the same categorizations and logics in their dealings with the non-human world. Since these logics derive from the cultures of the zooarchaeologists themselves, such analyses can inevitably be criticised as universalizing. Conversely, social-constructivist approaches to human-animal relations, whether inspired by structuralism or other schools of anthropology, deny the innate characteristics of the animals themselves and tend towards an idealist relativism. Apart from rendering understanding of archaeological cultures all but unattainable, save possibly through methods such as Hodders (1986) contextualism, this position presupposes a separation between nature and culture which is itself part of our own mode of engagement with the environment (Descola 1996, 84-85; see also Ingold 1996a, 118-120). A productive zooarchaeology must negotiate the pitfalls of both these positions, identifying likely commonalities in human-animal relations while upholding the importance of cultural variations. We must try to establish a realist approach, in which animals are seen as animals (Figure 3.1).
21
Figure 3.1: the relationship between epistemologies and approaches to animals in archaeology.
3.2 Attitudes to animals

A contrast is often drawn in the anthropological and archaeological literature between huntergatherer attitudes to their environments and those prevalent in the modern West. Huntergatherers, it is argued, typically do not see animals and the non-human world in general as being radically separated from humans; animals are rather perceived as social beings, sharing an essential kinship with their hunters (see 3.2.1). In the modern Western cultures from which most students of European prehistory originate, a very different attitude prevails: the dominant view of animals rests on an ontological discontinuity between humans and the nonhuman world (e.g. Descola & Plsson 1996; Ingold 1996a). Whether considering them commodities to exploit or treasures to protect, we typically place animals in a domain of nature which is independent of subordinate to a human/cultural realm. Given the sharp distinction between hunter-gatherer societies and the modern West, one might ask where non-Western societies who practise food production including those of the Neolithic fit in. Is it hunter-gatherers who are marked out from farming peoples, or the West that differs from the rest of the world? Alternatively, is this a contrast between opposite ends of a spectrum, masking considerable scope for intermediate situations? Anthropologists interested in exploring a contrast between hunter-gatherer ontologies and Western viewpoints have rarely engaged in detail with the position of animals, and especially wild species, in non-Western or pre-industrial societies with domesticates. The taxonomy of human-environmental relations set out by Descola (1996) is silent on nonWestern farming societies: animism and totemism dominate amongst hunter-gatherers while naturalism, founded on the belief in an independent nature, characterises Western societies (Figure 3.2a). Likewise, Plsson (1996) distinguishes orientalism and paternalism as modernist phenomena, opposed to the communalism of hunter-gatherers (Figure 3.2b).
22
Arguments concerning the emergence of an ontological discontinuity between humans and animals accordingly fall into two broad camps: those placing the most fundamental shifts with domestication and the onset of the Neolithic, and those citing rather the Enlightenment or the Industrial Revolution. In the latter camp, Ingold (1994, 17) stresses that commodification or objectification of animals is not a result of domestication but a much later product of industrial livestock management: when hunters became pastoralists they were not taking the first steps on the road to modernity. Likewise, Tapper (1988) sees domestic animals as being engaged in social relations with humans in all contexts other than urban-industrial societies, where they become reduced to machines.
Figure 3.2: taxonomies of human-environmental relations set out by (a) Descola and (b) Plsson. Diagram (a) is after Shapland (2006), based on Descola 1996; diagram (b) is directly from Plsson 1996.
23
On the other hand, we have already seen (2.1) that many archaeological narratives concerning the Neolithic imply a radical change in the form of human engagement with non-humans. This view is not restricted to archaeologists. Serpell (1986, 174-176), for example, describes the Neolithic transition as a fall from grace and a journey of no return: once domestication has taken place farmers are compelled to act in opposition to nature. This colourful language in particular the biblical overtones of the Fall is echoed by popular romanticization of hunter-gatherers vis--vis farmers (e.g. Brody 2000). More tempered arguments for a Neolithic disjuncture in human-environmental relations are found in the anthropological literature. Goodman (1992, 19), for example, states that the development of horticulture entails a fundamental break with hunter-gatherer attitudes. Morris (1998, 1-6) is unusually direct. Following a robust critique of simplistic contrasts between the West and most other societies with the former characterised by domination over nature he instead argues for a marked contrast between hunter-gatherers and agriculturalists attitudes towards animals, concluding that an ethic of control began not with the Enlightenment but with the Neolithic: the advent of farming has had a profound effect on the way humans relate to the natural world, especially towards animal life (Morris 1998, 4). At the same time, he concedes that beliefs in essential kinship between humans and animals are not restricted to hunter-gatherers, stressing rather the complex, ambivalent and frequently contradictory attitudes to mammals that exist within farming societies. In fact, Morriss position is not so far removed from Ingolds, with its recognition that the establishment of dominion over animals need not reduce them to objects. The apparent opposition between viewpoints may reflect a conflation of modes of identification and relation, to use Descolas terminology. Both Ingold and Morris implicitly disjoin the question of ontological continuity between humans and animals (i.e. the mode of identification) from their terms of engagement (the mode of relation). The following sections consider the implications of the adoption of farming specifically of animal husbandry for the place of animals in human society. To this end it is necessary to problematize the phenomenon of domestication itself (3.2.2). First, however, a selective review of human-animal relations amongst hunter-gatherers is presented.
3.2.1 hunter-gather human-animal relations

The above discussion makes clear the need to preface consideration of domestication with a review of approaches to animals outside the context of farming societies. This offers a way to break down certain preconceptions about the ways in which people relate to non-humans. Since Neolithic peoples were farmers like us, it is all too easy to project our own
24
understandings of animals back onto them. Comparison with hunter-gatherer societies helps to highlight these preconceptions, paving the way for a more broad-minded approach to early farmers and their engagements with their environments. Whatever models of Neolithization one prefers, the phenomenon ultimately grew out of non-farming societies. While the processes by which it did so are beyond the scope of this thesis, to assume a priori an absolute disjunction between Neolithic and pre-Neolithic ways of life is to perpetuate the narrative described in 2.1. Closely linked to this narrative, at least in its earlier manifestations, is the idea that huntergatherers exist closer to nature than do farming peoples. This notion underlies a longstanding moral ambivalence in Western society regarding hunter-gatherers, one that can be traced back to the 17th century contrast between Drydens noble savage and Hobbess remark that the natural life of man is solitary, poor, nasty, brutish and short. Within archaeology, the idea of hunter-gatherers as somehow natural has prompted the famous observation that in the literature as a whole, successful farmers have social relations with one another, while hunter-gatherers have ecological relations with hazelnuts (Bradley 1984, 11). The description of non-farming peoples as close to nature raises both ethical and epistemological objections. If hunter-gatherers exist in a state of nature then we, farming peoples, are implied to be partially outside that nature; in order for us to intervene in or to domesticate it, we must first be separate from it (Williams 1972, 154). Far from being a given, however, the idea of an independent nature has increasingly been recognized as a social construct (e.g. Descola 1996, 82; Dwyer 1996b; Sahlins 1976). The nature:culture dichotomy has thus attracted considerable criticism as an element of a culturally specific primarily Western and post-Enlightenment ontology, linked closely to the conceptual separation between humans and non-humans, and ultimately to the Cartesian mind:body duality. Anthropology has contributed to this realization by demonstrating the incommensurability of such dualisms with the ontologies of certain non-Western people (e.g. MacCormack & Strathern 1980). The social-constructivist position dominant within much current anthropology holds that peoples environments are culturally constituted, and that the designation of part of the environment as nature is likewise (Ingold 1996a, 118). Descola compares the dualist naturalism of the modern West with monist animist or totemist ontologies among amongst hunter-gatherers:
the very existence of nature as an autonomous domain is no more a given of experience than are talking animals or kinship ties between men and kangaroos Descola 1996, 88
25
An example of animism is provided by Bird-David (1990), who cites forest is as parent as the principal metaphor on which hunter-gatherer perceptions of their environments are founded: the environment is seen as responsive and giving. The same unconditional sharing expected among families or small communties Sahlinss (1972) generalized reciprocity is perceived between people and the forest (Bird-David 1992). She contrasts this with neighbouring horticulturalists, for whom the environment is like an ancestor, engaged in balanced reciprocity with people. It is notable, however, that the horticulturalists environment is not seen as entirely passive or inert, simply more distanced from humans. The idea of generalised reciprocity between humans and non-humans is also stressed by Plsson (1996, see Figure 3.2b). Ingold accepts the essence of Bird-Davids analysis but criticises the idea of metaphor within it, neatly demonstrating that the claim nature is a cultural construct presupposes the very construction it sets out to expose (1996a, 117-122). If forest is as parent, this attributes primacy to a social sphere over an ecological; a concept from the realm of human-human relations is projected onto the realm of human-environment relations. Ingold argues that people are socialized simultaneously into human society and into their environments, and that a rigid separation between the two is a product of the nature:culture dichotomy: one might just as well state parent is as forest (Ingold 1996a, 50). Whereas anthropologists see humans as straddling two domains, one ecological and one social, hunter-gatherer relations with other people and with their environments take place on a single plane (Figure 3.3). Here, Ingold upholds Bird-Davids claim that hunter-gatherers do not ascribe into the nature of things a division between the natural agencies and themselves They view their world as an integrated entity (1992, 29-30). Ingolds diagram makes clear that the division between nature and society rests on the Cartesian separation between mind and body within humanity: we are seen as both person and organism. The generally rigid modern division between humans and animals has the same foundation: humans are seen as animals plus, our personhood being a uniquely human attribute lain on top of a shared organic nature (see also Ingold 1988, 1-6; Ingold 1996a; Midgley 1983). But if we step outside the dualist framework, there is no reason why animals and indeed other elements of ones environment should not be persons. This is not the place for a comprehensive definition of personhood; suffice to say that it broadly rests on the attribution to an entity of both individuality and intentionality. Huntergatherers do indeed commonly consider animals to be persons in the sense of autonomous social beings with their own thoughts and intentions, typically on an equal level to humans (Hallowell 1960; Ingold 1994; 1996a; 2000, 90-92; Nadasdy 2007; Tanner 1979; see also
26
Viveiros de Castro 1998).
In this conception, hunting becomes a social rather than a
technological activity, through which relations of reciprocity between human persons and non-human persons are established and maintained. These relations are not seen as radically separated from those between humans.
Figure 3.3: Ingold's representation of Western anthropological (top) and hunter-gatherer (bottom) 'economies of knowledge'. From Ingold 1996, 127.
A well-known example of the attribution of personhood to animals is the notion amongst various hunter-gatherer groups that animals give themselves up to the hunter as part of a system of reciprocity, and that hunting is thus essentially non-violent (Ingold 1998; Nadasdy 2007). We have already seen that this kind of contention should not be seen in terms of metaphor, which would assume the primacy of a purely human social realm. It might alternatively be dismissed as anthropomorphism: human qualities are read onto animal behaviour by the Cree hunter, for example, in the same manner that one might attribute personhood to a family pet. Anthropomorphism, however, entails the projection of uniquely human qualities onto animals. It therefore requires the analytical assumption that animals categorically do not possess those qualities linked to personhood. While the existence of
27
conscious intentionality in non-human animals is an imponderable, animals do certainly seem to have intentions, both to the layperson and to the broad-minded animal psychologist (Griffin 1976; see also Midgley 1983). Likewise, the existence of individual personalities amongst non-human animals is increasingly accepted by evolutionary biologists (see Bell 2007 for a brief review). Philosophically speaking, our attribution of mind to other humans is on no firmer ground, and Milton (2005, 263) proposes the term egomorphism to replace anthropomorphism: it is not the human mind which we project onto non-human animals, but rather our own individual mind that we project alike onto humans and (potentially) nonhumans. To draw a rigid, qualitative line between the two in this regard relies neither on logic nor experience but on a priori beliefs. The actual status of non-human animals vis--vis humans is not at issue here the reader is welcome to maintain a dualist episteme if they so wish, and I certainly would not go as far as Nadasdy (2007, 32-34), who suggests that all of Northern hunters beliefs about animal personhood should be considered reality. The superficial glance at hunter-gatherer humananimal relations presented here may be counted successful if it has conveyed the following three points. Firstly, that animals and other non-human entities may be perceived not merely described as persons. Secondly, that hunter-gatherers engage in relationships with such persons that need not be seen as discontinuous from their relations with each other but rather operate on the same plane. Finally, that these relations are typically characterised by ideas of trust and reciprocity.
3.2.2 - domestication
The idea of animal domestication is certainly neither straightforward nor taken for granted, as testified by a substantial discursive literature on the subject within archaeology and related disciplines. However, more of this is dedicated to the mechanisms and context of early domestication than to the nature of the category domestic itself. Beyond the leading edge of the Neolithic the dichotomy between wild and domestic species is rarely problematized, despite the near-complete analytical primacy it assumes within zooarchaeology. The range of human-animal relations grouped under the heading is extraordinarily diverse, as indeed is that classified as wild. What within this diversity of relationships allows us to tease out two overarching categories to which we assign such a priori importance? The following discussion reviews approaches to this question (3.3.2), drawing on several schools of thought to formulate the understanding of domestication used in the remainder of the thesis (3.3.3). It should be stressed at the outset that this position is not intended as definitive; rather, it is developed with the specific aims and context of the present study in mind.
28
Numerous definitions of domestication can be found in the literature, a comprehensive review of which is provided by Russell (2002). A useful starting point is Ducoss (1989, 28-29) distinction between domestication as a process or event and domestication as a condition or state of being. The following discussion reviews four broad schools of thought on domestication, assessing what each says about the concept both as an event and as a condition. I do not mean to imply that domestication necessarily involves a discrete event followed by a static condition, although this is often implicit in treatments of the topic. Rather, some authors have sought to describe how domestication occurs, while others have tried to define what it is to be domestic. On the surface, the first of these aims might seem the more worthy, with the latter at risk of becoming a mere semantic exercise. For the current purposes, the opposite is true. The processes, mechanisms and causes of domestication represent a worthwhile field of study in their own right. They are likely to differ widely between species and between regions, however, and should not be assumed to follow a single common pattern. Seeking to define an essence (Bknyi 1989, 22) of domestication in these terms thus risks imposing a general model on a potentially very varied record. Moreover, most models see animal domestication as a discrete process of change, taking place around the start of the Neolithic for the majority of relevant species. For archaeological purposes beyond this period, therefore, it is the condition of domestication that we are really interested in. Pinning down exactly what we mean by domestic need not be a matter of dealing in definitions for their own sake: the understanding of domestication developed below is relevant in the context of the Balkan Neolithic, where I argue it has important implications for the political economy. For other times, places and research questions different ideas of domestication may be more appropriate, or the term may not be useful at all.
3.2.2.1 classical definitions

The emergence of distinct phenotypic characteristics at the population level implies at least partial genetic separation from a parent population, and this lies at the heart of many classic models of domestication. While individual animals may be tamed, they are usually only referred to as domesticated if they belong to a reproductively isolated population, with successive generations existing in a different relationship with humans than do other representatives of the species. But what is the nature of this special relationship that qualifies some populations as domestic? After all, reproductive isolation of one form or another is fundamental to all speciation events and is certainly not restricted to ecological relationships classed as domesticatory. Here, authors differ, but the unifying feature of what I term the classical approach is an emphasis on the role of deliberate human control of breeding (e.g.
29
Clutton-Brock 1989, 7): animals are domestic if their relationship with humans involves intentional restrictions on their movement and reproduction. This is the view enshrined in Bknyis oft-quoted definition of domestication:
... the capture and taming by man of animals of a species with particular behavioral characteristics, their removal from their natural living area and breeding community, and their maintenance under controlled breeding conditions for profit (Bknyi 1969, 219, emphasis added)
This is reminiscent of the progressivist attitude to the Neolithic criticised in Chapter 2. Humans are (or rather man is) seen as removing animals from nature, presumably into a cultured sphere, through deliberate and directed action for profit (later changed to for mutual benefits Bknyi 1989, 22). The use of the word capture emphasises the dominance of humans in this process, implying (futile) resistance on the part of animals. As Ducos (1989, 28) notes, this approach to domestication is primarily about process, specifically the capture of animals by humans. One might suggest reproductive isolation as an associated condition of domestication but this has little meaning unless one specifies human control (Figure 3.4).
Figure 3.4: domestication as process and as state in four common approaches. Paler words in brackets represent vague and/or problematic elements in understandings of domestication.
3.2.2.2 ecological approaches

Truly deliberate artificial selection is generally believed to have been a relatively recent phenomenon (e.g. Leach 2007). Some zooarchaeologists stress rather the role of unconscious
30
selection a phrase first coined by Darwin (1859, 34) in causing the morphological and behavioural changes associated with domestication (e.g. Zohary et al. 1998). If significant genetic change and effective reproductive isolation could have arisen as unintended consequences of changes in selection pressures due to increasingly protective human management of wild herds, we need not invoke any deliberate or absolute breeding control until a much later stage. A second approach to domestication thus avoids anthropocentrism by playing down the role of human intentionality. This view attributes to humans and non-human animals a more-or-less equal role in determining the form of their interactions: human choices are only part of the picture and the behaviour of animals played a role in their own domestication (Budiansky 1992). A common physiological basis for many of the human-animal relationships subsumed under domestication has recently been proposed. Noting the lack of a concrete mechanism accounting for the rapid physical and behavioural changes accompanying the domestication of many species, Crockford (2006) proposes one in the form of changing thyroid hormone rhythms. According to her model, only relatively stress-tolerant individuals of a given species are able to colonise anthropogenic environments, causing what is essentially a founder effect. Since various aspects of stress tolerance are governed by the same thyroid hormone release patterns which control growth, coat colour and reproductive seasonality, plus behavioural traits such as docility, this process can very quickly cause the establishment of discrete populations of various species with a distinctive suite of physical and behavioural characteristics, living in close proximity to humans (Crockford 2006, 99-102). It remains to be seen whether this model will be accepted by mainstream biology, but Crockford presents a compelling argument. Under her model, domestication would represent a specific form of symbiosis characterised by a common package of physical and behavioural traits, doing away with the need for capture in a definition of domestication. If the process of domestication from an ecological viewpoint thus becomes a matter of niche colonization and potentially speciation, then the condition of being domestic is one of symbiosis in its various forms (Figure 3.4). Budiansky (1992) rejects any suggestion that domestication is unnatural, comparing it with numerous examples of symbiosis between nonhuman animals. He nonetheless retains the term, treating it as coherent category throughout his argument. OConnor (1997) goes rather further. By analysing human-animal relationships in terms of their costs and benefits for both partners, he demonstrates that the idea of domestication as a coherent form of ecological relationship falls apart when ideas of commensalism, mutualism and so on are applied too closely: the specifics of human-animal relationships in these terms simply cross-cut wild and domestic as we understand them.
31
Where Budiansky seems reluctant to abandon the concept altogether, the clear implication of OConnors paper is that domestication is simply not a useful label for categorizing humananimal relationships. From a purely ecological point of view, then, domestication is something of a red herring, and attempts to form a concrete definition are thus arbitrary and academic. Following this reasoning, one might call for the wild:domestic dichotomy to be rejected altogether, as previously advocated on slightly different grounds by the Cambridge palaeoeconomic school (Higgs & Jarman 1972; 1975; Jarman & Wilkinson 1972). Zeder (2006a, 107), on the other hand, argues for a reconciliation of the ecological approach with ideas of human intentionality, seeing domestication as a distinctive process but one which does not profit from strict demarcation.
3.2.2.3 anthropological approaches

From an anthropological point of view, domestic animals are distinguished by their ability to act as vehicles of relationships between humans while alive (e.g. Ingold 1980; 1984). In its simplest form an anthropological approach thus posits the idea of animals as property as the essence of domestication. The idea that domestication is founded on changes in humanhuman relationships formed the essence of Ducoss classic definition almost thirty years ago:
Domestication can be said to exist when living animals are integrated as objects into the socio-economic organization of the human group, in the sense that, while living, those animals are objects for ownership, inheritance, exchange, trade etc. Ducos 1978, 54
More recently, Russell (2002, 294; see also 2007, 39) has argued that the appropriation of animals as property is a quantum shift in human-animal relations that we cannot ignore, a difference not only of degree but also of kind. In an early article on the subject, Ingold (Ingold 1984, 4) defines domestication as the social incorporation or appropriation of successive generations of animals into human society. A wild animal thus becomes one that is not, in the living state, engaged by the structure of social relations in the human community. The idea of animals as property is central, with a principle of shared access replaced by one of divided access: we describe as domestic any population of animals which are subject to property relations while still alive, and as wild any population which represents a shared resource up to the moment of death. More strictly, one could point to the extension of property relations to the potential offspring of individual animals, thus ruling out one-off tamings.
32
Here, we are clearly dealing with a definition of domestication as a state of being: for animals, to be domestic is to be owned by humans while alive (Figure 3.4). In terms of the process of domestication one might suggest Ingolds term appropriation, but this is very vague and gives no indication of the actual events or mechanisms involved. Such fence-sitting is not a problem for the present purposes, however: one may emphasize either deliberate capture or behavioural co-evolution as the driver of initial domestication, but when dealing with the later Neolithic the status of domestic animals as property is, in this view, their most important common characteristic.
3.2.2.4 human-animal sociality

The final school of thought reviewed here also falls largely within the bounds of anthropology, but recognises that over-emphasis on animals as property risks a return to the anthropocentrism of the classical approach: there is more to relationships between humans and domesticates than the role of the latter as tokens of relationships between the former. Some authors have focused on the form of direct relations between humans and domesticates, returning to the recognition of animals as persons common in hunter-gatherer anthropology. In his later work, Ingold characterizes domestication as a change in the terms of engagement between humans and animals, a transition from trust to domination (1994, 18). Humans gain the power of determination over animals, compromising the status as autonomous social beings that the latter typically hold for hunter-gatherers. Ingold stresses, however, that this need not involve the outright objectification of animals: traditional societies with domestic animals typically still view them as subject-persons rather than objectthings (1994, 17). In this way he distances himself from Ducoss position in which domestic animals are objects and to an extent from his own earlier work. For Marx (1964, 102) domestic animals were unambiguously a form of capital mere tools through which an owner might act. Ingold (1980, 88) suggests that they should rather be seen as labour, and his trust to domination model develops this argument to draw an explicit parallel between domestication and slavery (Ingold 1994). Neither human slaves nor domestic animals are passive tools, but both have their autonomy compromised by the exercise of force on the part of human agents. The distaste which such a comparison might provoke in modern Western society is, in Ingolds view, another product of dualist thought: since humans are considered qualitatively apart from above non-human animals, any suggestion which would place certain humans on the same footing as mere animals is hard to swallow (see also Clark 1988, 30-31). But as Ray and Thomas (2003, 38) note, the idea of a mere animal would have been incomprehensible for Neolithic people. Ingold agrees: traditional
33
pastoralists2 may rank animals hierarchically below freemen, but they are not assigned to a separate domain of being (Ingold 1994, 17-18). Tapper (1988) takes a similar view but looks more closely at specific forms of human-animal relationship, phrased in terms of relations of production. While he sees working animals including the dogs kept by some hunter-gatherers as slaves, pastoral herds are described rather in terms of a reciprocal transaction analogous to that in feudalism. Livestock thus become serfs, provided with care and protection in return for a rent. The extraction of secondary products is here seen as an important factor in the establishment of social rather than purely ecological relations between herder and herded (Tapper 1988, 53). This conception of pastoralism reintroduces the ideas of trust and reciprocity to the human-animal relationship, although it remains a hierarchical relationship in which one side holds the power. This is somewhat reminiscent of Ingolds earlier (1980, 96) description of domestic animals as like dependents in the house of a patriarch, suggesting a relationship which is lacking in symmetry but not entirely in trust or care. We should perhaps be thinking here in terms of a shift from generalized to balanced reciprocity. The recognition of trust and reciprocity within pastoral relationships has been taken further by Knight (2005a, and articles therein) in a volume with two central themes. The first is that, as Ingold suggests, pastoralists and mixed farmers have direct, personal, social and emotional relationships with their animals. The second is that these relationships are frequently characterised by trust, care and genuine affection. Milton (2005, 262-263) concentrates on the importance of intersubjectivity in coming to know oneself, and asks why this concept should not be applied between as well as within species. Campbell (2005) emphasises the personal nature of human-animal relationships, and their inseparability from those between humans, among Tamang speaking communities in Nepal:
Animals are sustained by and sustain dense community relationships: by looking after animals people are also looking after themselves and each other It is not a domination pure and simple, but a convivial domestic distributional hierarchy based on practices of interpersonal attentiveness, that characterises this milieu. Campbell 2005, 83
The interconnectedness of domestication and kinship relations has also been stressed by Russell (2007, 33-37). In a study of recent small-scale farming in Greece, Theodossopoulos (2005) strikes many of the same chords but stresses a sense of order within the domestic economy and the manner in which the eventual death of farm animals is interpreted as a reciprocation of the care they receive during life.
2
Throughout this chapter, pastoral and related words are used in the broadest sense, indicating the keeping of domestic animals. No implication of scale, specialization or mobility is intended.
34
Knight uses these arguments to attack the trust to domination model, drawing on the distinction between types and individual animals. Hunter-gatherers may attribute personhood to animals, but if one considers individuality as a prerequisite of personhood then they do not in fact experience them in this way (Knight 2005b, 3-5). Since encounters during hunting are episodic and unrepeated, the intimacy established between hunter and prey can only be with animals as types, not as individuals. Under many forms of domestication, by contrast, continuous daily association between humans and animals provides the temporal and spatial conditions for human-animal intimacy to emerge (Knight 2005b, 5). To see domesticates as objects, he argues, is to conflate process with outcome. Oma (2006, 42-43) uses this reasoning to turn Ingolds model around. The lack of intimacy between individual hunters and individual animals precludes reciprocal trust, she argues, while close relationships between humans and domestic animals are founded upon it: docility and co-operation are expected in return for protection, food and care.
3.2.2.5 discussion: animals as sentient property

There are useful elements in all four of these approaches. Classical definitions of domestication, in my view, assume too much regarding specific human motivations. The ecological school provides an important counterpoint to this, and is unassailable on its own terms, but risks losing sight of the aspects of domestication of most interest to anthropologists and archaeologists. Both approaches, however, have advanced the understanding of morphological changes commonly associated with domestication, through emphasis on reproductive isolation and on the changing selection pressures implied by colonization of anthropogenic niches. These ideas underpin the concept of morphological domestication upon which zooarchaeological studies including the present work necessarily rely. In order fully to appreciate the implications of animal domestication for human societies one must turn to the second two approaches. Before doing so, however, it is worth asking whether a watertight definition of domestication is really either necessary or helpful. The emphasis on wild versus domestic species in archaeology has obscured numerous other distinctions in the ways in which humans engage with animals, and the issues highlighted by the latter two approaches to domestication are only two amongst many dimensions of variation upon which classifications of human-animal relationships could be based. Nonetheless, I believe that they are useful in that taken together they (a) define domestic in a way which I think would be roughly coterminous with most archaeologists intuitive understandings, and (b) have particular relevance to Neolithic archaeology and especially to the models of social change proposed for the Vina period (see 2.3; 3.4)
35
I shall introduce this composite understanding of domestication with an example. Pigs are chosen due to the existence of thought-provoking ethnographic studies, but the same arguments could be applied to any species. Pigs are renowned for their versatility, both in terms of habitat and of their possible relationships with humans, as a result of which there exists a considerable ambiguous area between straightforward hunting and close husbandry (Albarella et al. 2006a). Ethnographically, pigs are often kept under only very loose control, and interbreeding with wild populations is common (e.g. Dwyer & Minnegal 2005). Rosman and Rubel (1989) document New Guinean cases in which the male offspring of domestic sows are castrated and the sows allowed to breed only with wild boars, clearly removing reproductive isolation from the equation. Both male and female pigs are generally allowed to roam freely. While the difference between boar and barrow under this regime is painfully clear, one might ask how a distinction can be drawn between wild and domestic sows. Albarella et al. (2006b, 221) hint at two answers to this question: the domestic pigs (a) at least nominally belong to someone and (b) have direct intersubjective relationships with their owners. They recognize them, return to be fed, and generally behave differently around humans than do their wild counterparts, despite the lack of discrete breeding populations. There are thus twin elements to a social conception of domestication. The first concerns the role of domestic animals as vehicles of human-human relations, the second deals with humananimal relations in the strictest sense. In fact, these can be seen as part of the same picture. The human-human and human-animal relations involved in domestication are parts of a single social system or, if one prefers, a single ecological system, a rigid distinction between the two terms being unhelpful. Taking the twin elements together, domestic animals can be seen as sentient property.
3.2.3 implications for the political economy

Domestic animals may constitute property, but property itself is by no means a straightforward concept. Meadow (1989, 81) notes that domestication entails a change of focus on the part of humans from the dead to the living animal. While pastoralists recognize rights over live animals, hunters only do so for dead ones, with far-reaching implications for exchange. The sharing of meat is common in hunter-gatherer societies, but the relations that it entails only last between killing and eating. With domestication, these relations can potentially be extended not only throughout the life of individual animals, but also through generations of herds. Thus, animal resources can become vehicles of enduring social relations in a pastoral society in a way that they cannot in a hunting society (Ingold 1980, 144, emphasis added). The involvement of animals in transactions between individuals or households may lead to the establishment of extensive networks of social relations every
36
herd comes to embody an aggregate of separate but overlapping interests (Ingold 1980, 175). Ingold notes that such networks depend on the use of secondary products since this permits borrowing and lending of animals for labour (see also Bogucki 1993) or distribution of their tangible products. Carnivorous pastoralism, by contrast, he associates with accumulation and a lack of obligations to distribute animals after death. However, the services of breeding males may be exchanged even where animals are raised purely for meat especially where herds are small creating the type of obligations to which Ingold refers. In addition, while carnivorous pastoralism may be characterised by accumulation this is not necessarily about accumulating food so much as animals per se. Drawing primarily on African ethnography, Russell (1998) shows that the role of some domestic animals as units of wealth can be motivation enough for the keeping of herds, with exchanges of livestock, for example as bridewealth, being crucial to the maintenance of social ties. In effect, wealth and prestige are secondary products. The accumulative aspect common in carnivorous pastoralism does not preclude the establishment of overlapping interests in herds which might be played out in the distribution of meat. That animals represent sentient property, able to form social ties both with and between humans, has major implications for the political economy. The idea that exchange objects can play an active role in social relations is well established and can be traced back to Mausss (1954) magnum opus. A more recent formulation based on Gell (1998) has become popular within archaeology, making the explicit claim that artefacts have agency. If we can accept, at least for the purposes of debate, that inanimate objects are active, it is ironic that we cannot do the same for animals, which palpably are active in the most literal of senses (Nadasdy 2007, 35; see also Mleku 2007, 267; Oma 2006, 40-41). Gell himself includes animals in his argument but does not see them as qualitatively different from material culture (1998, 17). Chapman emphasises the inalienability of objects in prehistory, stating that an indissoluble link exists between all owners or users of an artefact and the artefact with its distinctive biography (2000a, 5). This idea, based on Stratherns (1988) enchainment, underpins his fragmentation model (see 2.3). Things, according to the theory, take on some of the social form of persons (1988, 134) and develop their own biographies as they move between people, all the while establishing complex networks of relations. This is reminiscent of Ingolds arguments concerning the exchange of animals, outlined above, and indeed Strathern provides an example involving pigs (1988, 163). Chapman also extends the theory to animals, suggesting that, along with lithics, they constitute the original fractal object par excellence (2000a, 7). Whether or not one accepts Chapmans hypothesis of deliberate artefact breakage, the post-mortem butchery and distribution of domestic animals can probably be taken as a
37
given in prehistory, and would surely have reflected ties and obligations built up during animals lives. But animals are not only objects they are also subjects. A biography is a life story and animals have lives in a very real sense that pots and figurines do not. For the latter biography can only ever be metaphorical; for the former it is literal (Oma 2006, 220). Moreover, animals have genealogies in the real sense: Kula valuables may have a history of social engagements but they certainly do not have sexual partners and rivals; Vina figurines may become fragmented but they do not produce and care for offspring. Nor can this be dismissed as a purely structural issue, with dispersal, fragmentation and formation of sets broadly replacing most of the structures of biological reproduction, for as we have seen animals are capable of entering into direct intersubjective relations with humans. Quite simply, animals are a qualitatively different order of being from artefacts. In placing the two on the same level Chapman risks perpetuating a dualistic view of personhood which sits uneasily with the sophistication of his study. This is not to say that Stratherns and Chapmans arguments should not be applied to animals, only that there is an added dimension when the media of exchange are animate. Transactions involving animals do not only create ties amongst people, they also create and alter social relations between people and animals, and amongst animals. A tension thus arises between animals as social beings in their own right, and animals as objects subject to property relations and able to act as vehicles of relations between humans, a tension that may be played out differently between species, cultures and social contexts. Following this logic, Ray and Thomas (2003, 41) argue not only that cattle in Early Neolithic Britain were inalienable possessions rather than commodities, but that they were tied intimately into social life, as social beings. The kinship relations of humans and domestic animals become intertwined it is impossible to speak of [Tamang] kinship without reference to their livestock (Campbell 2005, 96).
3.2.4 implications for the wild:domestic dichotomy

It should immediately be noted that the two criteria of domestication stated in 3.2.2.5 jointly constitute an analytical definition. While demarcating a logically coherent and (arguably) heuristically useful set of the domestic, they do not necessarily reflect a meaningful category for Neolithic people. The coherence of the domestic as an emic category concerning animals would rely on these two factors being privileged above a whole range of others. The appropriation of animals, in terms of the extension of property relations to them, necessarily entails their conceptual placement below humans. In one of his earlier works,
38
Ingold (1986, 253) suggests that the hierarchical relationship between animal and spirit master in many by no means all hunter-gatherer cosmologies provides the model for domestication; mastery just needs to be transposed from the non-human master to humans. Whether or not there is any truth in this theory, domestication surely requires the removal of animals from a position of equality with humans. At the same time, the establishment of direct intersubjective relations promotes greater intimacy between humans and animals, working against any tendency for conceptual estrangement as a result of the subordination of the latter by the former. Rather than reducing animals to possessions, domestication sets up a tension between their affinity with and separation from humans (see Tambiah 1969, 454-456), underlain by their status as social beings and as property respectively. This tension is dynamic in the sense that the strength of the two opposing forces varies: decreased interpersonal contact between humans and domesticates, as in loose herding systems, may weaken the sense of intimacy or kinship, while increasingly competitive or accumulative human-human relations might stress the object aspect of domestic animals to the detriment of their subjecthood. There is much of Tappers (1988) argument in this (see 3.2.2.4), although I would avoid his overtly Marxian tone with its implications of discrete and deterministic modes of relation: the depersonalization of human relations with domesticates reaches an extreme with industrial farming and also with ranching, in which animals are owned yet essentially hunted, but these two systems certainly do not entail similar attitudes to animals. Both can in any case be probably considered modern phenomena; Ingold (1980, 2) notes that even amongst the Skolt Laps ranching is a recent development from pastoralism due to the influence of neighbouring markets and that it relies on divided access to uncultivated land. Within the Neolithic, at least, we should not therefore expect the outright objectification of domestic species: domestication represents a change in the mode of relation but not necessarily in the mode of identification. Thus while Siberian Buriat herders, for example, perceive a hierarchical, protective relationship between themselves and cattle, the animals are still conceived of as sharing an essence with their masters (Hamayon 1990, 605-704). Descola (1996, 95) cites this as an example of a protective mode of relation operating within an animic mode of identification. Such protective animism is as good an a priori model as any for human-animal relations in the Neolithic, at least as regards domesticates. The position of wild animals in such a scenario is more problematic. Animals which are not incorporated into the social system must, in a way, become distanced from it, but they need not lose their hierarchical equality with humans. They thus assume a position which can be seen as the reverse of that occupied by domesticates, illustrated schematically (and
39
hypothetically) by Figure 3.5. At the same time, physical, behavioural and life-cycle similarities between the two groups must have been recognised by Neolithic people (Whittle 2003, 90-93), especially where wild and domestic populations of the same species were present. The conceptual position occupied by wild species will surely have been influenced by the form and geography of human encounters with them and the extent of perceived similarities with related domestic populations, as well as by the nature of human relationships with the latter. It is thus impossible to predict how wild animals were viewed in comparison to domesticates, beyond saying that it must have been highly contingent and species-specific. From this perspective the idea of a unified concept of the wild in the Neolithic appears unlikely even where a coherent domestic domain can be established. Whittle (2003, 91) has pointed out that factors such as taste, habitat, behaviour, sociality and size might have been just as important as domestication in Neolithic peoples experience of animals. The status of animals may have been negotiable, with domestic animals taking on wild characteristics in certain contexts as has been suggested for bull-leaping in Neopalatial Crete (Shapland in press), or for the morphologically domestic pigs apparently shot with arrows at Durrington Walls (Albarella & Serjeantson 2002, 43-44; although see Studer & Pillonel 2007). While nothing in the understanding of domestication and human-animal relations outlined here calls for the existence of coherent domains of domestic and/or wild in prehistory, one cannot rule out their development. Oma (2006, 35) argues that a conceptual separation between wild and domestic animals was likely in European Bronze Age farming societies based on the differing geographies and temporalities of human encounters with each group. Wild animals were encountered beyond the home-sphere and are thus likely to have provoked different reactions from people than their domestic counterparts. Hamilakis (2003, 244) extends this argument to the Final Neolithic in Greece, proposing a concept of wildness founded on otherness akin to that traced through more recent Western history by Cartmill (1993). This may be a valid argument for specific cases, but should not be taken as a general model for prehistory. Distance, both conceptual and physical, from human society will surely have been an important factor in the perception and perhaps classification of animals, but cannot be assumed to correlate with wild and domestic in the senses used here. Variations in degree of control over animals, from close husbandry to very loose management, may have blurred the boundaries (Pollard 2006, 144). Some domestic animals could have been herded at a considerable distance from sites, resulting in relatively low visibility for those not involved in their care. Meanwhile, garden hunting (see Cook & Ranere 1989; Linares 1976) raises the possibility that encounters with certain wild species may have taken place very much within
40
both the temporality and the geography of the domestic sphere. Associations between wild and distant are thus subject to the same arguments for contingency and species-specificity made above. While individual species may have been perceived as wild in something approaching the emotive sense of the word in English, there is simply no reason to believe that encounters with undomesticated animals in general in the Neolithic were rare, geographically remote, or in any way otherworldly. This is especially true where their remains are common in the zooarchaeological record.
Figure 3.5: a hypothetical model of the relationships between humans and animals (a) without, and (b) with domesticates. Based partially on Ingold's (1986, 253) suggestion that domestication in some cases represents the transposition of mastery over animals to humans.
Even where a separate wild domain does exist, it need not be dualistically opposed to a human, social realm (Pollard 2006, 137), as implied by the common conflation of wild and domestic with nature and culture. This conflation is highly problematic and merits further consideration. Concepts broadly equivalent to wild and domestic are common in small-scale
41
farming communities but are not necessarily rigid dualisms (Morris 1998, 125) and need not indicate the presence of an ontological nature:culture divide (Strathern 1980; see also Descola 1996, 84). One might ask which group of animals in Figure 3.5 represents an independent nature: the wild species, whose position relative hierarchically equal to humans is unchanged, or the domestic species which find themselves distanced hierarchically from humans yet which are more intimately tied into social life. In this sense wild:domestic certainly cannot be equated with nature:culture, and in fact the superficial homology between these two dichotomies has been shown to be problematic, even contradictory, in general (Strathern 1980, 202), domestic being variously equated with both culture (e.g. Barth 1975, 194-195) and nature (e.g. Ortner 1974). Hodders (1990) The Domestication of Europe argues for the creation in the Neolithic of conceptual domains labelled domus and agrios. In so doing, it has been accused of perpetuating a narrative concerning the separation of nature and culture (Pollard 2006, 136) that can be traced back through successive generations of Neolithic scholars (see Barker 2006, chapter 1, my section 2.1). Hodder himself is careful not to phrase his arguments in this way: rather than representing nature and culture, or even wild and domestic, the domus and agrios are two different aspects, or perhaps loci, of the transformation of wild into domestic (Hodder 1990, 68-87). But of course this both presupposes the existence of a wild:domestic dichotomy and does so in a way which unavoidably entails the idea of an independent nature ready to be brought into a human sphere. While to some extent circumventing the ambiguities highlighted by Strathern through his use of the twin concepts, Hodder nonetheless effectively equates wild:domestic with nature:culture; Pollards criticism is valid. Even if one accepts Hodders concepts of domus and agrios, they should not be expected to map onto wild and domestic animals in any straightforward manner. In fact, animals have a highly ambiguous status within the model: hunting is unequivocally associated with the agrios but animals in general are also tentatively linked with this domain, while related objects such as bucrania and zoomorphic figurines fall in its intersection with the domus (Hodder 1990, 69). The intention here has not been to deny that coherent concepts of wild and domestic were ever relevant in prehistoric societies. Rather, I hope to have shown that one cannot assume their significance a priori, and that human-animal relationships in the Neolithic are better approached species-by-species in the first instance.
3.3 Animals as food

While the arguments set out thus far should have established that animals represent much more than food to human societies, the vast majority of archaeological faunal remains do 42
derive from consumption. It is this post-mortem aspect of human engagements with animals that is most directly implicated in zooarchaeological data. The following sections review some considerations surrounding sharing, communal consumption and disposal respectively, with particular attention to possible lines of difference in treatment between wild and domestic species.
3.3.1 sharing
The sharing of food can probably be considered a ubiquitous element of human society. Food has a particular power to convey meaning and create social ties that probably derives from its universal necessity (Counihan & Esterik 1997), and meat in particular has been argued to be especially rich in social significance (Fiddes 1991, 11-19). The anthropological study of sharing and exchange is founded on ideas of reciprocity as set out by Mauss (1954) and later by Sahlins (1972). These are well established, and need not be reiterated here. Instead, I shall base this section around some ideas from Ingolds earlier work that relate particularly to the sharing of meat. Concerned with different forms of sharing amongst hunter-gatherers and pastoralists, Ingold draws a distinction between sharing in and sharing out (Ingold 1980, 173-174; 1986, 233). Within hunter-gatherer communities, meat sharing has typically been seen in terms of generalized reciprocity: when a carcass is brought home, the successful hunter has an obligation to share the meat with all, regardless of previous such transactions. This is in line with the principle of collective access. As Dowling (1968, 505) points out, the animal only becomes the hunters property at the point of death and effectively ceases to be so when distributed possession confers the right to distribute, not the right to consume. Reciprocal obligations are not incurred and all that accrues to the hunter is influence (Ingold 1980, 158159). Sharing in this context is therefore not a risk reduction strategy but rather a mechanism for ensuring equality (Woodburn 1998). This is what Ingold refers to as sharing out. The slaughter of domestic stock is a different matter. If one conceives of domestication as underpinned by a principle of divided access then there is no generalized obligation to share the carcass (Ingold 1980, 172). And yet meat often is distributed following slaughter, even where it could be consumed in a short period or preserved. Ingold refers to this as sharing in, since the recipients are claiming shares in the animal which were established during its life through transactions between the owner and claimant: overlapping interests held in herds are mapped out in the distribution of meat (1980, 175, see 3.2.3). Updating Ingolds model somewhat, the links between owner and claimant can be thought of in terms of
43
enchainment rather than simple balanced reciprocity, focusing attention on the biographies of individual animals. There is a danger here of drawing too strict a line between hunter-gatherers and farmers. For the purposes of the current thesis, one might ask what forms of sharing should be expected where both hunting and herding are practised. Dowling (1968, 503) suggests that sharing obligations for wild animals might persist alongside other systems for domesticates. Ingold makes no predictions, but notes that both modes of relation can apply simultaneously to different animals within the same society (1984, 5; 1994), loosely supporting Dowlings remark. Morris (1998, 93) describes a system of sharing for hunted meat in Malawi that is very reminiscent of the literature on hunter-gatherers, but he omits details of how domestic carcasses are distributed. The account given by Marshall of sharing among the Okiek of Kenya is problematic for the arguments set out above since the same rules apply for wild and for domestic taxa. Sharing is obligatory in both cases: social relationships and patterns of sharing are not upset by individual ownership of domestic animals (Marshall 1993, 232). The discussion almost exclusively relates to wild animals, unfortunately, but the distribution of these does not seem quite as egalitarian as that described elsewhere. Although meat is distributed to people with no specific claim in it, the largest portions go to the successful hunter and others involved in the hunting and butchery (Marshall 1993, 233). This is perhaps an example of a situation where the sharing of both wild and domestic stock has elements of both sharing in and sharing out, but should certainly be taken as a cautionary tale regarding applying the categories envisioned by Ingold too rigidly.
3.3.2 communal consumption and feasting

The topic of communal consumption has acquired a certain fashionable status within archaeology, and feasting in particular has become something of a buzzword. Both concepts become rather nebulous on closer inspection, however. After all, very little consumption in the sense of eating of food occurs alone, and the distinction between a meal and a feast is fuzzy to say the least. Nonetheless, a useful operational definition of feast or the less loaded term communal consumption event is a meal which is significantly out of the ordinary in terms of scale, whether of participation, provision, or both. This will often be a quantitative rather than qualitative difference (Marciniak 2005a, 71-72). I shall not list common characteristics of feasts, nor their archaeological correlates, since I do not believe the term to have analytical, rather than heuristic, validity. Suffice to say that such a list is provided by Dietler and Hayden (2001, 3-10). It is worth considering the possible functions of unusual large-scale consumption events, however, and three categories
44
of feast have been defined by Hayden (1996, 128-131). The first of these is termed celebratory, and is concerned with the establishment and maintenance of social bonds. Hayden sees this form of feast as distinctive to hunter-gatherer societies. The second is the reciprocal-aid feast, associated with mutual-aid work parties, and the third is labelled as commensal and further subdivided. The primary distinction in Haydens scheme is between cohesive and competitive forms, but Dietler (2001) notes that feasts may simultaneously perform both functions. As for sharing more generally, we can ask whether wild and domestic species are likely to have played different roles with regard to feasting events in farming societies. One might, for example, suggest that wild animals as a shared resource are more likely to be associated with the cohesive side of feasting, while domestic animals would be particularly useful as a resource for competition. On the other hand, feasting involving domesticates could also have a socially cohesive role since, as we have seen, they can become intimately tied into kinship relations. In fact, since the occurrence and nature of feasting will relate to the specific social tensions that obtain in a given society at a given time, it is probably unwise to make predictions for the general case.
3.3.3 disposal / discard / deposition

The final stage in the material side of animal use is that of disposal, discard or deposition. No unproblematic definition of this phenomenon has yet been formulated, for reasons that will become clear, but it can loosely be thought of as what happens to animal remains at the end of their use-life. However, the popular notion that archaeological material consists primarily of passive, valueless rubbish, disposed of as a matter of practical imperative, has been heavily attacked (Martin & Russell 2000). The critique of rubbish disposal involves two related realizations. Firstly, ethnoarchaeological studies (notably Hodder 1987) have shown that the discard of objects is not purely functional in any universal sense, but rather combines symbolic and functional elements, the distinction between the two being of little heuristic value (Marciniak 2005a, 75). This realization has given rise to the concept of structured deposition a term coined by Richards and Thomas (1984) and later used to great effect by Hill (1995) which stresses that even everyday practices of disposal are often structured by cultural ideas and categories, blurring the boundary between mundane and ritual deposits. Leading on from this, it is now widely accepted that items which have been used, consumed or broken are not necessarily devoid of meaning or significance; the process by which material enters the archaeological record is not necessarily a process by which it leaves the social world. The second realization, then, is that practices of discard and deposition can be discursive rather than merely a passive reflection of cultural categories (e.g. Chapman 2000c).
45
This has been an extremely brief account, since the very nature of the problem dictates that it cannot be tackled in the abstract: there can be no general theory of discard (Martin & Russell 2000, 59). Some possibilities regarding the significance of depositional practices on Vina settlements are discussed in 6.3.4.
3.4 Animals in the Vina period: a new model

Having reconsidered the position of animals within archaeology in general, and the nature of domestication in particular, we can now return to the particular issues of change in the Vina period. A focus on domestication in terms of property and of interrelated changes between human-animal and human-human relations becomes particularly relevant in this context. As established in 2.3, it has been argued that the period sees a shift in focus from the community to the household as the principal unit of social organization, involving increased competition within communities. Given the likely centrality of animals in Neolithic life, the realm of human-animal relations must have been a key arena in which these changes were both expressed and negotiated. Yet relatively little has been said about the social role of animals within the Vina period. The following sections fill this gap by developing a model linking human-human and human-animal relations with regard to Vina period social trends.
3.4.1 competition / cohesion / communal consumption

Given the tension between household and community posited for Vina society, and the large size of the main species, communal consumption is likely to have been important, both in the broad sense of everyday sharing and the more specific sense of discrete large-scale consumption events. Referring to similar models of change in the Greek Neolithic, Halstead (2007, 26-27; Pappa et al. 2004, 16) suggests a role for feasting as a social cohesive mechanism working against household isolation, but one might equally expect a competitive element. Tension between cohesion and competition may to some extent have been paralleled by that between animals as social beings and as property. It would probably be oversimplistic, however, to think of wild animals as emblematic of the community and domestic species of the household. Indeed if domestic animals were herded on a communal level they may have retained a cohesive function even where individuals were the property of specific households. The involvement of domestic animals in enchainment might be expected to give them a role in competitive gift exchange while hunting, still founded on a principle of shared access, played no such role or even acted against it through obligatory, non-reciprocal sharing. On the other hand, one should not think of competition in enchained relations as anti-cohesive. Quite the opposite. Chapman talks of a tension between personal or household accumulation and
46
corporate kinship relations that was played out throughout the Balkan Neolithic and Copper Age (2000a, 47). While enchained relationships are competitive and far from egalitarian, they may nonetheless serve to restrict accumulation, even acting as levelling mechanisms emphasising horizontal, corporate kin relations (2000a, 5-6). Accumulative property relations are seen as resulting from their breakdown:
the loss of the direct relationship between person and object, in which objects were created out of persons, in favour of a representation of an abstract value, such as wealth, by the object now devoid of its most intimate personal connotations Chapman 2000a, 48
Chapman sees this as essentially a Copper Age development (c.f. Russell 1993, 7-8), but in any case I would argue that the unique properties of animals prevent this aspect of his model from being extended to them. Direct personal relations between humans and animals obtain in any system other than ranching or industrial farming. In the Neolithic and Copper Age animals thus remained subjects as well as objects in human relationships and cannot have become estranged from human persons in the way that Chapman describes. Following this logic, domestic animals remain inalienable property and sharing of their meat simultaneously has cohesive and competitive aspects. While representing tensions within society, these aspects are both fundamental to the practice itself. Accordingly, one should expect domestic animals to be involved in asymmetrical sharing and in large-scale consumption events, while wild meat may have been shared more evenly or less likely not shared at all.
3.4.2 wild and domestic in the Vina period

It was argued above that domestic animals, as sentient property, embody a tension between personhood and possession (3.2.3), and that the relative importance of each aspect of their status will vary according to the particular social conditions and husbandry regimes that apply (3.2.4). Even within the set of domestic animals there may be considerable differences in intensity of interaction with humans. Meanwhile, as the form of human-human relations involving animals changes reflecting wider social trends the idea of animals as property may resonate to a greater or lesser degree. Within the Vina period, the proposed social changes might be expected to promote an increasing emphasis on ownership and control of animals, which in turn might enhance the validity both of the wild as a conceptual unit representing that which cannot be controlled or possessed, and of its converse the domestic. Any wild category is unlikely to have the associations of mystery and distance from society with which we are familiar, however, since hunting does not seem to have been an unusual activity in the Vina period and the three main
47
wild species are all potential crop-raiders (see Chapter 5). At the same time, domestic animals may have been herded at varying distances from human habitation. The use of secondary products is an important issue here, since it provides additional opportunities for the establishment and maintenance of relations through transactions centring on domestic stock, and their extraction is also likely to increase the level of interpersonal contact between humans and individual animals. Secondary products also make animals more valuable as a living resource, potentially changing the way in which meat is viewed. It has already been argued (3.3.5) that even in a purely carnivorous system living animals are likely to be highly valued and entwined in human social relations, but the adoption or intensification of exploitation for secondary products would surely reinforce both aspects of domestication, setting domesticates more clearly apart from wild species. Secondary products are not thought to have played a major role in the Balkan Neolithic, although they are unlikely to have been entirely unknown. This is evaluated in Chapter 5. Both cattle and caprines may have provided milk, creating a conspicuous similarity between them, although Russell (1998) notes that amongst the European Neolithic domesticates only cattle meet Ingolds (1980, 224-226) criteria for acting as a unit of wealth. Cattle also appear to have particular symbolic significance on Vina sites (see 8.2.3). Pigs, on the other hand, provide neither tangible secondary products nor labour. Coupled with their potential for relatively loose management, this has major implications for how they were perceived in relation to other domesticates. To conclude this section, the validity of a wild:domestic dichotomy as regards animals depends on the status of domesticates as sentient property being privileged above various other factors (3.2.4). The form of direct human relationships with domestic animals is likely to have remained varied throughout the Vina period, providing little basis for a straightforward contrast with a similarly varied set of wild species. The proposed changes in human social relations, however, will have served increasingly to emphasise the status of domesticates as property and thus to stress their coherence as a group. Any increase in the use of secondary products would have had a similar effect, and in fact would presumably have been bound up with social changes. Since conceptions and categorizations of animals are often reflected in their post-mortem treatment, one might expect to see increasingly structured modification and deposition of domestic versus wild species over the course of the period.
48
3.5 Statement of hypotheses

Having considered various theoretical aspects of human relations with animals, both in the abstract and in the context of the Vina period, we can return to the project aims. These are restated below. To document changes in the importance of different species during the Vina period. To relate these to arguments for social change, with particular attention to the roles of wild vis--vis domestic species. It is now possible to suggest some patterns in the faunal data that might be expected to accompany increasing tension between household and community. These should not be considered as hypotheses in any proper, Popperian sense: their falsification cannot be taken as a refutation of the underlying social models, although it might weaken them. The derivation of some tentative predictions simply serves as a useful guide for the development of an appropriate methodology and the collection of data relevant to the second aim. Domestic animals will have been preferentially involved both in large-scale consumption events and in structured (i.e. non-random) everyday sharing. A division between wild and domestic will have become increasingly meaningful in the later Vina period; this will have been expressed through increasingly structured differences in treatment, consumption and disposal of wild and domestic species. The methodology developed to address both of the project aims is set out in the next chapter.
49
Methodology
Chapter 4 Outline of methodology

4.1 Introduction
As with any research project within the humanities, a major challenge lies in bridging the gap between theory and method. While it has been argued that there is no meaningful difference between the two (Robb & Dobres 2005, 160-162) there is nonetheless a great danger of disjunction between the questions posed and the data and analytical techniques brought to bear. One might ask how the analysis of fragmentary animal remains can hope to test the assertions made in Chapter 3 regarding the nature of human-animal relations, and of domestication in particular. This would miss the point somewhat. The aim is not to prove that the important aspect of domestication from an anthropological point of view lies in the convergence of human-animal and human-human relationships, nor that rigid categories of wild and domestic cannot be assumed universal, nor even that animals will have been an important part of Neolithic life. The first of these points is an opinion; the second two are taken as givens. Rather, the aim is to address the economic and social roles of different animal species within the Vina period without presupposing the overarching significance of a wild:domestic dichotomy, and prejudicing neither the material nor the symbolic aspects of human-animal relations. The theoretical considerations reviewed above serve to shape the hypotheses derived and the types of interpretations and explanations sought rather than to define the technical methods employed. The project aims necessitate consideration of (a) broad-scale spatial/chronological patterns in human use of animals, and (b) fine-scale evidence for their distribution, consumption and deposition, since both reflect important aspects of peoples interactions with and experiences of animals. A three-part research structure is employed here, zooming in from the regional to the contextual: 1. Regional-scale analysis of published and unpublished zooarchaeological data (Chapter 5). 2. Zooarchaeological reports from Gomolava and Petnica (Chapter 6). 3. Taphonomic-contextual study of the fauna from Gomolava and Petnica (Chapter 7).
50
Methodology
4.2 Regional review of zooarchaeological data

The regional study is methodologically very straightforward, consisting of two parts. Firstly, all available zooarchaeological data are assembled and compared at several chronological resolutions, taking advantage of absolute dating where available, in order to detect gross trends in taxonomic representation over time. Fragment count (NISP) data are used exclusively due to both availability and technical advantages (see 4.3.3), and efforts are made to account for the impact of differential excavation and recovery techniques. Apart from visual inspection of the data, Principal Components Analysis (PCA) is employed to detect relationships between taxa. In the second part, other aspects of the economic roles of animal use are discussed, making reference to published demographic and seasonality data. Relevant data from Gomolava and Petnica are included in both sections.
4.3 Standard zooarchaeological methodologies

4.3.1 recording systems
Faunal material from Gomolava and Petnica was studied at first hand by the author. Specimens from Petnica were recorded using an MS Access database adapted from that used by Vesna Dimitrijevi. The database consists of tables for individual specimens and for units, the latter including summary data for specimens designated undiagnostic (see below). The recording protocol falls broadly within a genealogy that can be traced back to Meadows (1978) Bonecode. As well as species, element, and side, fields are included for maximum length, diagnostic zones (Watson 1979, see below), element portion, weathering, burning, gnawing, digestion, breakage type, fusion, sex criteria and pathology, with additional sections for recording measurements, human modification, and tooth development/wear. The authors own additions included a system for recording diagnostic zones following Dobney and Reilly (1988) and a field for burning location. A new database was built from the ground up for the Gomolava assemblage. This system rationalized various additions made during the Petnica study and normalized the database structure, while maintaining compatibility through the use of the same core codes. Feature and Unit tables were also custom-built to fit the structure of the site documentation. Measurements were taken wherever possible, mainly following von den Driesch (1976) but with some additional dimensions defined. In a departure from the protocol advocated by von den Driesch, juvenile specimens were measured and flagged as such in an associated field. Similar flags were used for other factors potentially affecting measurements, such as burning, surface erosion or minor extra bone growth.
51
Methodology
4.3.2 identification
Specimens from Petnica were initially sorted into diagnostic and undiagnostic categories following a protocol adapted from Miracle (Miracle & Pugsley 2006, 260). Bones are considered diagnostic if they meet any of the following criteria: All teeth and tooth fragments except for entirely un-diagnostic enamel pieces. Complete bones. Bones with intact articular surfaces. All fragments with a maximum dimension greater than 5 cm. Long bone fragments with nutrient foramina. Bones with signs of human modification. Any other bone which could be identified at first glance.
Undiagnostic specimens were counted by body size category and element type (rib, long bone, cranial, pelvis/scapula, antler/horn core, indeterminate), and the totals recorded at the unit level. Due to poor recovery and possibly selective curation, very few specimens from the Gomolava assemblage would have fallen into the unidentifiable category. Accordingly, all specimens from the site were recorded individually. The Petnica assemblage was studied at the Petnica Science Centre (ISP), under whose auspices the site was excavated (see 6.4.1). Several students from Belgrade University acted as laboratory assistants, but identification and recording of diagnostic specimens were solely the authors work. A reference collection assembled by Greenfield during his earlier study was available, including modern dog, sheep, goat and roe deer skeletons, a few cattle elements, and much relatively intact archaeological material that could be used to identify more fragmentary specimens. Some problematic specimens were identified in Belgrade with the help of Vesna Dimitrijevi, but there were many more that could have been identified to a lower taxonomic level given better reference material. The Gomolava assemblage was studied in laboratory space at the Museum of Vojvodina kindly provided by Svetlana Blai, who also made her excellent comparative collection available Taxonomic identification was aided by a substantial collection of literature (Gromova 1950; 1960; Hillson 1992; 2005; Lawrence & Brown 1967; Schmid 1972, and others cited below). Identification challenges in the temperate European Neolithic include separation of sheep and goats, distinction between similarly sized Cervids and Bovids, and discrimination between wild and domestic forms of pigs, cattle and Canis. The former problem was mitigated by multiple reference specimens in each collection, but published references were helpful nonetheless (Boessneck 1969; Halstead et al. 2002; Payne 1985; Prummel & Frisch 1986).
52
Methodology
Even with the aid of these resources the majority of caprines were assigned to an indeterminate sheep/goat status. This was classed as identification to taxon for the purpose of calculating identification rates. Separation between cervids and bovids in this case Cervus versus Bos and Capreolus versus Ovis/Capra can be problematic on certain elements. The main references employed were Helmer and Rocheteau (1994) for the smaller species and Prummel (1988) for the larger. While this allowed most relevant specimens to be identified to taxon, a substantial number were assigned to the categories Ovis/Capra/Capreolus and Bos/Cervus, especially to the latter at Petnica where reference material for larger species was limited. Three of the species present on Vina sites are expected to have both domestic and local wild forms, namely pigs, cattle and dogs/wolves. While demographic criteria are commonly used to detect early domestication at the population level (e.g. Collier & White 1976; Monchot et al. 2005; Zeder 2005), distinction between wild and domestic individuals where both co-exist in an assemblage is primarily a matter of size, although non-metric morphological criteria do exist for certain elements, notably horn cores (Armitage & Clutton-Brock 1976). The use of size to determine domestication status rests on the observation that most species reduce in size considerably following domestication, creating a differential between domesticated individuals and their wild counterparts (Higham 1968b). Wild and domestic populations rarely show completely discrete size ranges, however, and sexual dimorphism often confuses the matter further (e.g. Payne & Bull 1988). Trimodal size distributions may emerge in assemblages containing both wild and domestic individuals since wild females and domestic males are often of a similar size. In most cases a substantial number of specimens must be assigned indeterminate status. Given that the size ranges observed within both wild and domestic populations depend on a large number of environmental and genetic factors, their potential for separation varies between regions and periods. There can be no universal metrical criteria for the identification of wild or domestic individuals. Rather, individual specimens can only be identified as wild or domestic by reference to the overall distribution of measurements, and only then when these measurements fall into relatively clear groups. Sample size is thus of crucial importance. Where limited numbers of data points are available for particular measurements, two techniques can be employed to improve sample size. Firstly, data from similar sites in the same region and period may be superimposed. Secondly, the data from different measurements may be combined by transforming them into log-ratio scores relative to a standard animal (Simpson et al. 1960). For each measurement on an archaeological specimen the natural logarithm is taken, and that of the equivalent measurement on the standard animal
53
Methodology
is subtracted to produce a measure of size difference between archaeological specimen and standard, allowing for comparability between elements. This increases the overall sample against which individual specimens can be compared, but ignores variation in anatomical proportions, effectively assuming that all individuals are the same shape as the standard. Logratio values are employed here alongside raw measurements on individual dimensions, the latter being considered more reliable when samples are respectable. Many specimens provide more than one measurement and thus more than one log-ratio value, in which case the mean is used as a generalized size indicator. The Ullerslev cow (Degerbl & Fredskild 1970) is used as the standard for Bos, along with specimens from the Hungarian Agricultural Museum for Sus and Canis (Russell 1993, 140; 163). Unmeasurable specimens can often be identified as wild or domestic either by noting their size relative to more complete elements (e.g. slightly larger than specimen x with length y) or, in extreme cases, by subjective judgement.
4.3.3 quantification
Methods of zooarchaeological quantification have been debated at length since the beginnings of the discipline without any consensus on their relative merits being reached (see Grayson 1984; Ringrose 1993 for reviews). The most basic measure of abundance, the Number of Identified Specimens (NISP) or fragment count, is still the most frequently used. It is employed here throughout, with the minor qualification that bones which can be shown to conjoin or articulate are treated as a single specimen (cf. Chaplin 1971, 65). NISP counts suffer from potential biases relating to differential fragmentation, recovery and identification rates (Bknyi 1970; Chaplin 1971; Grayson 1984; Ringrose 1993; White 1953) and to varying numbers of skeletal elements between taxa (Shotwell 1958). In addition, a lack of independence between data renders inferential statistics based on NISP all but impossible (Grayson 1984, 22-23; Payne 1972). The system of Diagnostic Zones (DZ) proposed by Watson (1979) circumvents inter-taxon anatomical differences by counting only certain key parts of the skeleton. This also potentially reduces fragmentation bias since zones are counted only if more than half is present, preventing any single zone from being counted twice. On the other hand this bias may simply be reversed since readily-fragmented portions will rarely cross the 50% threshold even where frequently identifiable. Watsons zones were recorded here alongside NISP, with some modifications following Dimitrijevis (pers. comm.) protocol (Table 4.1).
54
Methodology
Element Maxilla Mandible Atlas Axis Scapula Humerus Radius Ulna Metacarpal III* Pelvis Femur Tibia Astragalus Calcaneum Metatarsal III* prox. dist. prox. dist. prox. dist. prox. dist. prox. dist. prox. dist.
Criterion Alveolus for (d)P4 Alveolus for (d)P4 > > of body > of glenoid fossa > of metaphysis > of metaphysis > of articulation > of metaphysis > of semi-lunar articulation > of articulation > of metaphysis > of acetabulum > of metaphysis > of metaphysis > of metaphysis > of metaphysis > > > of articulation > of metaphysis
Table 4.1: definitions of diagnostic zones. (Dimitrijevi, pers. comm.)
Minimum number counts go further in avoiding redundancy. Minimum Number of Individuals (MNI) was initially heralded as a more reliable method than fragment counts for estimating relative taxonomic abundance (e.g. Chaplin 1971, 69-70) and, by extension, dietary contribution (White 1953). By using the most abundant anatomical part from each taxon as a predictor of that taxons frequency, MNI reduces bias due to differential fragmentation and prevents sample inflation. A considerable body of literature has accumulated critiquing MNI (Chaplin 1971; Grayson 1978; Grayson 1979; Grayson 1984; Klein & Cruz-Uribe 1984; Ringrose 1993; Watson 1979). While NISP may to a considerable extent reflect fragmentation and recovery, MNI may become little more than a measure of sample size: less abundant taxa become progressively over-represented as the total sample decreases in size (Grayson 1981). Nor does MNI fully avoid fragmentation bias, since even the most abundant elements will be subject to this effect (Grayson 1984, 25). MNI is also dependent on the level of aggregation. Since minimum number counts are not additive between contexts or units of analysis, totals depend heavily on the resolution of the faunal study: the finer the scale at which analysis is conducted, the greater the total MNI counts for the site (Grayson 1973, 1984; Watson 1979). Given the sample size effect mentioned above, this will also alter relative taxonomic abundance.
55
Methodology
The archaeological relevance of MNI estimates is also questionable. While it may be crucial to the interpretation of discrete special deposits, the number of individuals contributing to an assemblage is otherwise rather an arcane and abstract datum. Total meat weight by taxon is not necessarily the primary goal of zooarchaeological study (cf. Daly 1969, 148), and in any case its calculation assumes that each individual represented in an assemblage, howsoever defined, was fully consumed in relation to that assemblage. This assumption is clearly problematic below and in many cases even at the site level, especially if one acknowledges the possibility that deposition was structured. MNI counts are not used here. The Minimum Number of Elements (MNE) is rather more useful, despite technically suffering from all the problems of minimum number counts. The primary purpose of MNE is for the comparison of element representation between contexts and/or taxa, reducing but not eliminating fragmentation bias. MNE can also be totalled as a measure of species abundance less vulnerable than MNI to over-representation of certain elements, in which case it becomes roughly equivalent to DZ. Marean et al. (2001) review methods used to calculate MNE, but omit that used here: diagnostic zones sensu Dobney and Reilly (1988). Each major element is divided into a number of zones, and the presence or absence of each is recorded for individual specimens using the same 50% presence as for Watsons DZ. Minimum numbers may then be calculated from this data. Whilst not providing the maximum possible figures, this technique allows for the easy re-calculation of MNE from the database alone. When assessing element representation, MNEs are normed according to the frequency of each element in the body to produce Minimum Animal Unit scores (MAU, sensu Binford 1978).
4.3.4 demographic data and seasonality 4.3.4.1 age estimation

Estimation of age-at-death in zooarchaeological assemblages serves a number of purposes, primarily related to hunting or herding strategies via analysis of kill-off patterns. The substantial literature (e.g. Pike-Tay 2001; Ruscillo 2006; Wilson et al. 1982) focuses primarily on epiphyseal fusion and especially dental development and attrition. Dental eruption and attrition were recorded routinely in the present study following Grants (1975; 1982) scheme for caprines, cattle and pigs. Cattle deciduous fourth premolars with only the mesial-most accessory pillar joined to the main crown by visible dentine were given the designation ja, between j and k. Paynes system (1973; 1987) was used alongside Grants stages for caprines. For red and roe deer, a more complex scheme developed by Brown and Chapman (1991b) was employed. Whole-mandible eruption/wear states were also assigned using Paynes (1973) stages A-I for cattle, pigs and caprines, ensuring compatability
56
Methodology
with published data (Arnold & Greenfield 2004; 2006; Greenfield 2005; Legge 1990; Russell 1993). Paynes (1973) method of distributing incomplete mandibles across possible stages pro rata was employed, but loose teeth were excluded. Grants (1982) MWS scores were calculated for large samples. Data used to estimate ages from eruption and wear states included that from Aitken (1975 - roe deer), Legge and Rowley-Conwy (1988 - red and roe deer), Payne (1973 - caprines), Jones (2006 - caprines), Halstead (1985 - cattle), Hambleton (1999 - pigs), Magnell (2006 - wild pigs), Silver (1969 - various species), and sources cited in Ervynck (1997, 71 - wild and domestic pigs). Where large samples are available, harvest profiles are constructed following Payne (1973). These can be compared using 2-sample Kolmogorov-Smirnov (K-S) tests, since mandible fragments assigned Payne scores are complete enough to ensure that single mandibles are not counted twice, although there is still the risk of counting both left and right mandibles from the same individual. This is typically dealt with by considering only the most common side but a different technique is employed here. Since the true sample size will inevitably lie somewhere between the total number of scored fragments and the number of those from the most common side, K-S scores are calculated for both a best-case (i.e. all mandibles unpaired; n = total) and a worst-case (all mandibles paired; n = total / 2) scenario. Probability values are quoted for each of these. Crown heights were measured on dP4 and the molars following Klein and Cruz-Uribe (1983). In principle, crown heights have significant advantages associated with interval level data over ordinal stages, but they are undermined by unknown starting heights (see e.g. Enloe & Turner 2006, 131-132), and were barely used in the analysis. As an alternative source of age data, epiphyseal fusion is beset by problems (reviewed in OConnor 2003), most notably differential destruction. Used alongside dental age indicators, fusion data avoids the danger of relying exclusively on a single body part for age estimation, and the two techniques are broadly compatible (Zeder 2006b). Where appropriate, specimens from Gomolava and Petnica were recorded as unfused, fused or fusing, the latter describing any specimen with an externally visible fusion line.
4.3.4.2 seasonality
Inferences from mammalian remains regarding seasonality rely either on antlers or on a combination of tight, well-known birth seasons and sufficiently precise age estimates. In the latter case, the feasibility of determining season of death for particular species is debatable (e.g. OConnor 1998, 7). The seasonality of interest here is that of hunting and herding activities per se. The problems associated with attempting to extrapolate such evidence out to
57
Methodology
issues of site occupation and seasonal mobility are well known (e.g. Milner 2005; Monks 1981). Birthing seasons are restricted amongst many ungulates, including sheep (Williams 1968), red deer (Fletcher 1974) and roe deer (Kozdrowski et al. 2005). Cattle, however, only show a strong seasonal specifically spring concentration of births (Hall 1989, 215; Hammond et al. 1971, 48), although it is reasonable to expect that this will be encouraged and accentuated in domestic populations, at least outside market economies. The autumn calving common in modern herds relies on sophisticated modern farming methods including hormone treatment (Lamming 1975). Wild and domestic pigs are more problematic still, since they can theoretically breed all year round (Albarella et al. 2007b; Pond 1983), making double and even triple farrowing possible in certain circumstances (Lauwerier 1983). Table 4.2 shows the estimated seasons of birth used here.
Species Red deer Roe deer Wild pig Caprines Cattle Pig
Birth season May-June (Bknyi 1972; Clutton-Brock and Guinness 1975; Fletcher 1974) May-June (Bknyi 1972; Habermehl 1985) March-May (Briedermann 1990; Habermehl 1985; Mauget 1982) Late winter-late spring: varies with latitude and altitude (Harker 1968; O'Connor 1998) Mainly March-May, but not strictly seasonal (Hall 1989; Hammond, et al. 1971) No season assumed
Table 4.2: birth seasons for main taxa. (Bknyi 1972; Clutton-Brock & Guinness 1975; Habermehl 1985; Harker 1968; Mauget 1982).
Observation of tooth eruption and occlusal wear on juvenile specimens may in ideal circumstances provide sufficiently precise age estimates for inferences regarding seasonality. This is the case with Legges and Rowley-Conwys (1988) analysis of the roe deer mandibles from Star Carr, which uses comparative data from three modern populations. The potential for variation in wear rates between archaeological and comparative populations casts some doubt on the reliability of this technique for seasonality studies. Mandibular radiography potentially provides more precise age estimations. Since the roots of cheek teeth continue to develop after the crowns come into wear (Carter 1975, 231) they provide a basis for age estimation which is under more direct biological control than attrition but which extends further into life than eruption for any given tooth. Radiography also allows for the observation of un-erupted teeth in situ, typically increasing the number of teeth which can be employed simultaneously in the estimation of age for a single mandible. Detailed radiographic molariform development schemes accompanied by known-age data have been developed for deer (Brown & Chapman 1991a; Carter 1997; 1998; 2001; 2006), sheep (Orton
58
Methodology
2003) and cattle (Orton 2004). Radiographic scores have a stronger relationship with age than attrition scores (Carter 1997, 497-498). All cattle, caprine, red deer and roe deer mandibles from Gomolava with two or more of the main cheek teeth in situ and uncompromised were radiographed at the Veterinary Faculty in Belgrade under the supervision of Prof. Nikola Krsti, in order to supplement conventional age estimation. Radiographs were taken at 46-50kV and 6-16 mAs, with exposure reduced for very young specimens. The ages of radiographed specimens were estimated using data from Orton (2003; 2004, for sheep and cattle respectively) and Carter (2006, for deer). Following Carter, ages estimates were assigned for each developmental state according to the range of ages seen in the reference sample. This procedure introduces age-mimicry if used to construct mortality profiles (Konigsberg & Frankenberg 1992), but is acceptable for attributing rough ages to individuals, given a well-constructed reference sample. Since antlers develop and are shed following a yearly cycle, the presence on a site of cervid skulls with antlers either attached or clearly shed implies hunting during a certain speciesspecific season. Such finds are typically relatively rare, however, and shed antlers may have been brought to the site as a raw material (Caulfield 1978; Grigson 1981; Legge & RowleyConwy 1988; Pitts 1979). Shed or indeterminate specimens are therefore uninformative regarding seasonality.
4.3.4.3 sex determination

Sex determination involved both metrical and non-metrical techniques. All antler fragments can be assumed to be male (reindeer being absent from the region), but this does not allow for estimation of sex ratios. Cervid frontal bones with antlers clearly present or shed are more useful, but rare. Sus canines are readily sex-able, and in many cases even damaged canines permitted sex determination. Since mandibular measurements are important in distinguishing wild and domestic populations, the ability to sex some specimens provides an important check on sexual dimorphism. Only a few caprine pelvises were sufficiently intact to provide confident sex determinations from overall morphology, following Boessneck (1969) and Prummel and Frisch (1986). For the larger species the acetabular region had to be relied upon, in particular the lateral part of the pubis and the medial wall of the acetabulum, between pubis and ileum. Female cattle have a defined ridge running between ileum and pubis just to the superior of the acetabular rim, while males feature a much deeper, flatter acetabular wall (Grigson 1982, 8). The same
59
Methodology
morphological distinction was observed on red deer and assumed to relate to sex in the same way. Cattle horn cores are commonly used in sex determination, and potentially relevant variables including curvature, torsion and tip shape were recorded following Armitage and CluttonBrock (1976). The standard measurements of outer length, basal circumference and minimum/maximum basal diameter were taken. Ultimately, however, potential interpopulation differences were judged to prevent reliable sex determination (Sykes & Symmons 2007). Metrical sex determination was only attempted for deer, sexual dimorphism being particularly pronounced amongst Cervus in particular. Measurements from a range of skeletal elements, notably distal metapodials, were inspected for evidence of discrete size groupings.
4.4 Taphonomic and contextual analysis

4.4.1 theory and scope of taphonomy
Taphonomy is a sub-discipline of both zooarchaeology and palaeontology, concerned with the transition of animal remains from the biosphere to the lithosphere (Efremov 1940, 85; see Lyman 1994, 1-40). The interpretive role assigned to it differs widely, constituting a principal methodological difference distinguishing workers with different research agendas and theoretical orientations. By its very nature quantitative, taphonomic reconstruction in archaeology grew out of, and is often associated with, the more avowedly scientific branches of the discipline. However, a focus on consumption and deposition in recent more socially-oriented perspectives brings taphonomic concerns to the fore, and they have never been more central than in the latest formulations of social zooarchaeology. In current usage, taphonomy refers to the study of all processes intervening between a live community of animals and the records in an analysts database. The discipline is concerned with understanding these various processes, primarily through actualistic studies, and using this understanding to assess their impact on assemblages. Flow-charts illustrating the taphonomic history of a generic assemblage are a useful heuristic device here (e.g. Davis 1987, 22; OConnor 2000, 21; Reitz & Wing 1999, 111), consisting of series of populations, from life-assemblage to published data. Each population is a sample of the previous one; where sampling is non-random we have taphonomic bias. Accordingly, the sequence is often seen as a progressive loss of information content (e.g. Clark & Kietzke 1967, 117): the total
60
Methodology
amount of information declines over time, obeying the law of entropy (Hesse & Wapnish 1985, 19). Bias is a relative term, dependent on the questions being asked (Lyman 1994, 32). Indeed, many taphonomic inputs represent the addition of information to the assemblage, providing evidence regarding the processes which have taken place. This not only helps us to understand the likely biases in the observed data, but may be of interest in its own right. Taphonomy therefore has two aspects: 1. Reconstructive: controlling for bias and working back from the observed assemblage. 2. Descriptive: detecting and describing events and processes which are of interest in and of themselves. The balance between these depends on ones research interests, as does the point in the sequence towards which ones reconstructive efforts aspire (Figure 4.1).
Life assemblage
mortality
Palaeontological interest
Production / procurement
Death assemblage
primary butchery, transport
On-site remains
secondary butchery, distribution, processing, primary disposal
Archaeological interest
Consumption / deposition
Deposited remains
ravaging, weathering, disturbance, secondary disposal
Buried remains
diagenesis: chemical and mechanical changes in the ground
(Sub)fossil assemblage
excavation strategy, recovery techniques
Recovered assemblage
curation, identification, research interests, funding
Figure 4.1: targets of taphonomic research.
61
Methodology
Palaeontologists are usually concerned ultimately with the life assemblage and possibly causes of mortality so the reconstructive aspect predominates. Selective hunting by humans might therefore be seen as a distortion, whereas for the archaeologist it is quite the opposite. In theory, archaeologists are interested in all the human-mediated processes, but again differences in research agenda apply. Destructive processing might be a hindrance to analysts with an interest in procurement strategies, but for those pursuing questions of consumption it is the very object of study. Close parallels can be drawn here with wider archaeological debates regarding formation processes, especially within the processual movement (e.g. Ascher 1968; Binford 1981; Collins 1975; Schiffer 1976; 1983). Notably, Binford rejected the notion that Schiffers C-transforms constitute distortions, arguing rather that they are direct evidence for cultural activity. In reality, of course, they are both, and an archaeology that did not attempt to reconstruct areas of human activity prior to waste disposal would be a restricted discipline indeed. In archaeology as a whole, therefore, we are not interested in any one sampling population but in a whole section of taphonomic history. Archaeological taphonomy is thus unavoidably a simultaneous process of reconstruction and interpretation, and these two aspects cannot be separated sequentially.
4.4.2 social zooarchaeology

Much work on the archaeology of food is fundamentally taphonomic (e.g. papers in Miracle & Milner 2002). The most overtly taphonomic formulation of social zooarchaeology, however, is set out by Marciniak (1999; 2001; 2005a; 2005b), who advocates that inferences about human-animal relations be made via the reconstruction of contexts of consumption and differential practices of treatment and disposal. Marciniaks (2005a) landmark study of animals in the Polish Neolithic is an inspiration for the present work, but has considerable scope for methodological refinement. His six-step methodology is outlined below. 1. Qualitative taphonomy (surface modification, butchery and breakage) 2. Correlation of NISP with density 3. Correlation of NISP with MGUI (Binford 1978) 4. Correlation of NISP with MI (Binford 1978) 5. Interpretation of body part profiles by context 6. Interpretation of taxonomic composition by context
62
Methodology
The first stage is aimed at elucidating patterns of human processing, consumption and disposal, while steps 2-4 fall firmly within the realm of quantitative, reconstructive taphonomy, aimed at identifying the primary determinants of observed element profiles. Stage 5 only applied where preservation bias is judged to be minor returns to the more interpretive site of taphonomy, while stage 6 is a standard element of zooarchaeological analysis. Marciniaks approach is not adopted here wholesale, primarily due to concerns with its structure, which effectively divides analysis into qualitative (descriptive) and quantitative (reconstructive) portions. The former (stage 1) thus risks becoming anecdotal, while the latter is essentially a matter of working back to a desired point (2-4) before commencing interpretation (5). The role of utility indices is unclear. The potential for understanding the interaction between the various anthropogenic and non-anthropogenic factors in assemblage formation is thus limited.
4.4.3 development of taphonomic methodology

My own methodology combines Marciniaks aims and interpretive paradigm with the rigorous multivariate taphonomy advocated by Bar-Oz and Munro (2004). Here, multivariate refers not to statistical techniques, but to an integrated approach in which numerous taphonomic variables are brought to bear. Quantitative taphonomy inevitably suffers from apparent equifinality (sensu lato see Rogers 2000b, 721), and Bar-Oz and Munro argue that this is best dealt with by applying a wide range of analyses in a structured fashion. They propose a three-part methodology: 1. Descriptive phase. Overall frequencies of various taphonomic variables are summarised and tabulated in a quantitative version of Marciniaks stage 1. 2. Analytical phase. Possible factors shaping the assemblage are worked through in broadly reverse-chronological order: in situ attrition, density-mediate attrition, fragmentation, human transport and disposal. This is aimed at identifying influential agents of attrition as with Marciniaks stages 2-4 but brings in the information on breakage and bone modification from phase 1. 3. Comparative phase. The impact of particular taphonomic agents is confirmed by comparing assemblage sub-groups (taxa, elements, age-groups, context types etc.) which are expected to be affected differentially.
63
Methodology
This is methodologically more rigorous and analytically deeper than Marciniaks scheme, but is primarily reconstructive and not geared towards questions of distribution, consumption and deposition. The structure employed here incorporates elements of both schemes. Five stages are defined working broadly backwards through possible taphonomic processes, and the descriptive, analytical and comparative phases are included informally within each, where appropriate. Each analysis informs subsequent stages, but in many cases also provides direct information on human activities, and there is also a degree of feedback since stages inevitably overlap in reality. The structure is summarised in Figure 4.2 and Table 4.3, with details of specific analyses given in the following sections.
1.
Evidence for densitymediated attrition
2.
Evidence for peridepositional damage
refuse disposal practices
3.
Breakage & fragmentation
butchery, intensity of processing, body part selection (for processing)
4.
Visible human modification
butchery, cooking practices, refuse disposal
5.
Assessment of element representation
body part selection (for processing and/or disposal), sharing and distribution, differential access
Figure 4.2: structure of taphonomic analysis. Each stage informs the interpretation of the next, but there may be a degree of feedback, represented by the smaller arrows.
4.4.3.1 evidence for density-mediated attrition

Comparison of element frequencies with bone density and various measures of utility has been used widely since its introduction by Binford (1978). The underlying principle is that the observed proportions of elements in an assemblage should reflect the principal agents in its formation. The phenomenon of density-mediated destruction of bone by gnawing, weathering and diagenesis is well known if imperfectly understood (e.g. Binford & Bertram 1977; Lyman
64
Methodology
1984; Munson & Garniewicz 2003). Certain parts of the skeleton are consistently better preserved in archaeological assemblages and structural density is currently the best predictor of survival (Lyman 1994, 235). In theory, an assemblage in which skeletal part frequency strongly correlates with density is likely to owe its composition primarily to density-mediated destruction, rendering inferences regarding human selection unreliable. This approach suffers from one fundamental problem: equifinality. An element profile matching that expected from density-mediated destruction of complete carcasses does not necessarily indicate that complete carcasses were ever present, nor even that differential preservation was the primary factor in determining the observed profile. In fact, human selection and density-mediated destruction will both have been important in the formation of most zooarchaeological assemblages, and element profiles reflect this in that they rarely conform closely to ideal models. This is compounded by a weak but significant negative correlation between density and utility (Lyman 1985). A recent attempt to tackle equifinality uses maximum likelihood to estimate various parameters of assemblage formation (2000a; 2000b; Rogers & Broughton 2001), including degree of attrition and importance of utility-mediated transport. Through successful application of his technique to simulated assemblages, Rogers (2000b) argues that the supposed equifinality problem is simply an artefact of inadequate statistical treatment. Specifically, the pertinent criticism is made that correlation with density and subsequently utility applies bivariate analysis to a fundamentally multivariate situation (Rogers 2000a, 112). While statistically sophisticated, Rogerss technique has extremely limited archaeological utility. The whole enterprise rests on the assumption that ones primary aim is discrimination between kill sites and base camps, a simplistic notion for hunter-gatherer zooarchaeology and certainly not the case here. The technique also requires detailed assumptions about the manner in which agents of accumulation operate. Density-mediated attrition can in theory be well understood from actualistic studies, but human selection is a different matter. Rogers (2000a, 122-123) recognises the complexity of transport decisions but suggests that they might eventually be modelled, neglecting the effects of cultural variation. Selection of body parts for reasons other than transport efficiency is not considered, despite the fundamental importance of structured sharing and distribution in the formation of element profiles, even amongst dispersed hunter-gatherers (Marshall 1993).
65
Stage 1
Analysis Evidence for density-mediated attrition Correlation between element abundance and density Completeness of carpals and tarsals
Potential cultural information
2 Refuse disposal practices
Evidence for peri-depositional damage Frequency/severity of gnawing Frequency/severity of weathering Comparison with fragment size Food preparation (specific practices) Food preparation ('pot-sizing') Intensity of human use
Breakage and fragmentation Qualitative assessment Size distributions Percentage completeness Identification rate Index of breakage freshness
Visible human modification Frequency of burning Location of burning Butchery marks
Food preparation, refuse disposal Food preparation vs. refuse disposal Food distribution and preparation Transport, primary vs. secondary butchery deposits Transport, selective deposition Differential access, sharing/distribution, selective deposition
Assessment of element representation Correlation with Food Utility Index (FUI) Visual assessment of element profiles Comparison of anatomical/taxonomic profiles between features (Gomolava)
Table 4.3: components of taphonomic analysis, with potential cultural information.
(Behrensmeyer 1975; Binford & Bertram 1977; Boaz & Behrensmeyer 1976; Brain 1969; Carlson & Pickering 2004; Elkin 1995; Lam et al. 1999; Lyman 1984; Pickering & Carlson 2002; Stahl 1999; Symmons 2004; Willey et al. 1997)
66
Methodology
The traditional correlation method is used here, but in a critical manner. Element profiles are treated with caution where appreciably correlated with bone density, but are not necessarily rejected out of hand. Additional control for density-mediated effects is obtained by (a) assessing fragmentation presumably post-depositional amongst small dense bones such as carpals and tarsals (Marean 1991), and (b) comparing the abundance of skeletal parts with differing structural densities that are unlikely to have been treated independently by humans. Correlation between abundance and density raises numerous technical issues, concerned with choosing the best (a) measure of density, (b) measure of abundance, (c) level of anatomical aggregation and (d) method of assessing relatedness. Reviewing published bone density measurements, Lam (Lam & Pearson 2005; Lam et al. 2003) scores studies A to D on the basis of increasingly accurate incorporation of the crosssectional area of each scan site. Quantitative Computed Tomography (QCT) is considered the most accurate methodology. Table 4.4 summarises the available data for medium- and largebodied mammals. The choice of data for archaeological studies inevitably involves a compromise between measurement quality, sample size, and taxonomic relevance. Lam et al. (2003, 1706) state unequivocally that shape-adjusted data are preferable even if such data are only available for a distant relative of [the] species of interest. Data in class A are to be avoided at all costs. The best dataset for bovids is provided by Symmons (2002; 2004; 2005). Apart from sheer sample size, Symmonss use of two perpendicular radiographs at each scan site allows readings to be adjusted for the thickness of scanned bone. The study does not feature in Lams reviews but would fall between C and D by their criteria. Since Kreutzer (1992) highlights similarity in patterns of density between bovids and cervids, Symmonss data are used here for both families. Pig density measurements are provided by Ioannidou (2003) and Pugsley (2002). The latter are used here, being superior in both methodology and sample size. Zooarchaeological quantification was discussed above (4.3.3), but is revisited here with regard to suitability for correlation with bone density. Marshall and Pilgram (1991) use both uncorrected NISP and MNI (actually MNE), producing inconsistent results. Since the underlying principle is to compare observed element frequencies to an ideal complete profile, MNE-derived measures are theoretically more appropriate than NISP, and Binfords (1978) pioneering study employed MAU. Marciniak (2005a, 115) uses NISP for reasons of availability, conceding that MNE is a far better method of skeletal part frequency measurement. Others suggest that NISP is an acceptable proxy for MNE (Grayson & Frey 2004; 1993, 326-327), while Ringrose (1993) advocates using multiple measures to explore
67
Methodology
the data more thoroughly. While MAU (i.e. corrected MNE) is preferable in principle, it is used here alongside corrected NISP.
Source Brain 1969
Taxon Capra
n Technique 1 Mass / volume
Notes Class* N/A Long bones only; split into proximal and distal with cortical and trabercular bone measured together
Behrensmeyer 1975
Hippopotamus Equus burchelli Redunca Damaliscus Hylochoerus Ovis
1 1 1 Mass / volume 1 1 1 ? Mass / volume
N/A
Cortical and trabercular bone measured together
Boaz & Behrensmeyer Homo 1976 Binford & Bertram 1977 Lyman 1984 Ovis Rangifer Odocoileus Ovis Antilocapra Bison
N/A
35 specimens from an unspecified number of individuals. Unclear whether ends measured alone or with shaft Multiple scan sites. Assumes rectangular crosssection
3 Mass / volume 1 13 Photon 3 absorptiometry 1 12 Photon absorptiometry 1 1 Photon densitometry 2 32 Photon absorptiometry 4 4 QCT 2 2 3 Dual energy X-ray 2 absorptiometry 5 95 Digital photodensitometry 3 Photon densitometry 1 11 QCT 2 Photon 1 absorptiometry 6
N/A
Kreutzer 1992
As Lyman, but thickness estimated more accurately Volume measured by immersion and used to norm densitometry readings. Scan sites not published. Assuming circular rather than rectangular crosssection shown to improve prediction of survival
Elkin 1995
Vicugna Lama glama Lama guanicoe Homo Connachaetes Rangifer Equus burchelli Equus przewalskii Lama pacos Lama glama Lama sp.
Willey et al . 1997 Lam et al. 1999
Technique provides detailed control for cross-sectional shape.
Stahl 1999
Uses approximate cross-sections for each scan site Includes detailed analysis of variation with age, sex and individual Later refined by shape-adjustment (Carlson & Pickering 2004) 5 wild boar and 6 domestic pigs Follows Lyman's method.
Symmons 2002, 2004 Ovis Pickering & Carlson 2002 Pugsley 2002 Ioannidou 2003 Papio Ovis Sus Bos Ovis Sus
C+ A/C D A
Table 4.4: published bone density data for mammalian taxa. * following Lam et al. 2003
A related debate involves the appropriate anatomical level at which body part representation should be compared with density. Two rival approaches have been expounded in the literature. Stiners (1991; 2002) Anatomical Regions Profiling (ARP) technique groups elements into nine regions and calculates MNE at this level. With the exception of the neck
68
Methodology
and axial skeleton, Stiner (2002) argues that the higher densities within each region are similar, making the MNE values directly comparable at least between the limbs. Pickering et al. (2003) argue that ARP (a) entails discarding information that could be used to differentiate between taphonomic agents, and (b) rests on an assumption of uniform attrition that is contradicted by actualistic studies. Instead, they advocate careful identification and MNE calculation for all body parts, including long bone shafts. Accordingly, the present study operates at an element/sub-element level. Since density differs within each bone, MNEs are calculated separately for portions of major elements using Dobney and Reilly zones. Table 4.5 shows defined portions and corresponding scan sites. The same portions are used for NISP, although those on the scapula and ulna are combined. Element/portion abundance is compared with density for each major taxon, both at the site level and for specific context types, using Spearmans rank and Pearsons product-moment correlation coefficients. It is unwise to rely on these without assessing the relationship graphically (Ringrose 1993, 147), preferably plotting density ranges rather than average values (Symmons 2005, 89). Accordingly, scatter plots are produced and inspected wherever appreciable correlations are evident, using interquartile ranges for bovids and cervids and 50% confidence intervals for Pugsleys pig data. Medians are used where possible for the correlation itself to reduce the effect of outliers (Symmons 2002, 224), but only means are available for pigs.
4.4.3.2 peri-depositional damage

Weathering and gnawing provide direct information regarding formation processes and, by extension, depositional practices: both sub-aerial weathering and pig/carnivore access imply that remains were exposed for some time before final burial. In addition, the frequency and severity of marks can assist in the interpretation of element frequencies by indicating major agents of destruction. Weathering was recorded on a four point scale from unweathered to heavily weathered. Where it affected only part of the bone this was noted in the comments field. Gnawing was recorded in a single field, with options for rodent gnawing, carnivore punctures and general (carnivore/pig) gnawing. At Gomolava, categories were introduced within the latter, allowing gnawing to be entered as possible, marks, serious damage or portions destroyed. Digestion was recorded on the same field. Neither weathering nor gnawing was recorded for teeth.
69
Methodology
Zones Element Mandible Portion rostrum alveolar heel artic. Atlas Axis Scapula Humerus glenoid collum prox. shaft dist. Radius prox. shaft dist. Ulna Metacarpal* prox. artic. prox. shaft dist. Pelvis acetabulum ileum pubis Femur prox. shaft dist. Tibia prox. shaft dist. Astragalus Calcaneus Metatarsal* artic. shaft prox. shaft dist. Phalanx 1** Phalanx 2** Phalanx 3 prox. dist. prox. dist. (Dobney & Reilly 1988) 7 1 6 4,5 1,2 1 1,2,3 4,5 1,2,11 9,10 3,4,5,6,8 1,2 6,7,8 3,4,9,10,J A,B C,D 1,2 5,6 3,4,7,8 1 5 8 1,4,5 6 7,8,9,10,11 1,2,3,4 8,9 5,6,10 1,2,3,4 3,5 2 1,2 5,6 3,4,7,8 1 3 1 3 1 1 3 2 1 3 2
Bovid/Cervid scan site (Symmons 2002)
Pig scan site (Pugsley 2002) Dn-1 Dn-4 Dn-6 Dn-8 At-3 Ax-1
P Scap P Hum S Hum D Hum P Rad S Rad D Rad
Sp-1 Sp-2 Hu-2 Hu-3 Hu-5 Ra-1 Ra-3 Ra-4 Ul-1 Ul-2
P MetaC S MetaC D MetaC P Pelv
Mc-1 Mc-3 Mc-5 Ac-1 Il-2 Pu-1
P Fem S Fem D Fem P Tibia S Tibia D Tibia
Fe-2 Fe-4 Fe-6 Ti-2 Ti-3 Ti-4 As-1 Ca-3 Ca-2
P MetaT S MetaT D MetaT Phal 1
Mt-1 Mt-3 Mt-5 P1-1 P1-3 P2-1 P2-2 P3-1
Table 4.5: definition of element portions with corresponding scan sites. prox. = proximal; dist. = distal; artic. = articulation * Bovid/cervid metapodial zones on left, pig zones on right ** phalanx zones follow diagram rather than text in Dobney & Reilly
4.4.3.3 - a note on the validity of hypothesis testing in taphonomic analysis

Formal hypothesis tests are employed here to evaluate the significance of certain observed trends. This is controversial, since their shortcomings are well documented (e.g. Shennan 1997). The principal problem, clearly demonstrated by Grayson (1984, 22-23), is sample inflation: since zooarchaeological specimens are typically fragments their independence 70
Methodology
cannot be assumed and true sample size remains unknown. The degree of significance accorded to any test statistic is thus dependent as much on the degree of fragmentation in an assemblage as it is on either the strength of association or the original number of specimens contributing to the sample. In Graysons example, a 2 test for Bos versus Ovis frequency in two strata only becomes significant when each bone is broken into a number of pieces: effectively, the sample has been inflated. This is a matter of scale of analysis. Sampling issues in zooarchaeology are particularly complex since fragmentation takes place on two levels: (a) that of individual carcasses into elements; and (b) that of elements into fragments (Orton 2000, 53). In Graysons example, where frequencies of species are compared, the individual carcass becomes the unit of analysis and the sample thus suffers from inflation at both levels. In cases where the element is the unit of analysis, for example when comparing anatomical profiles, only fragmentation process (b) applies. In such cases formal significance testing would be valid provided that all elements were complete an extremely unlikely scenario. Minimum number estimates are no solution as they simply introduce new complications: MNE values are themselves estimates with an error term which is non-random, uni-directional and correlated with the (unknown) true sample size. Certain taphonomic analyses arguably shift the unit of analysis to a third level, that of the individual fragment. When analysing weathering, for example, the datum of interest is not the number of individuals whose remains underwent sub-aerial weathering, nor even necessarily the number of elements, but rather the number of specimens in whatever state of fragmentation obtained at the commencement of the weathering process. Of course, this state is both variable and unknowable. The validity of formal hypothesis testing for comparing rates of weathering and other taphonomic variables between assemblage sub-groups thus depends on the amount of fragmentation which occurred subsequent to the commencement of the process in question. Strictly speaking, any further fragmentation entails sample inflation and thus invalidates significance values. If one accepts that apparent sample size is always an overestimate of the true value, however, and errs on the side of caution accordingly, formal significance testing may remain a very useful tool, especially if one assesses computed p values rather than relying slavishly on arbitrary cut-offs (see Cowgill 1994, 8-9). Very highly significant results would require an improbable degree of sample inflation in order to be invalidated, while patterns which prove statistically insignificant even on the basis of the apparent sample size would certainly not be significant were the true frequencies known.
71
Methodology
4.4.3.4 breakage and fragmentation

The aim of comparing breakage and fragmentation data is to highlight differences between taxa and/or depositional contexts in terms of pre-depositional damage, specifically that which might be related to differential processing and different contexts of consumption. For example, Miracle (2002, 81) uses percentage identifiability and fragment size to suggest a decrease over time in the intensity with which carcasses were used at Pupiina. Breakage at Petnica was recorded in a single field, primarily allowing for distinction between spiral and angular breaks. The latter are often taken to indicate that bones were broken when fresh, presumably to extract marrow (e.g. Biddick & Tommenchuck 1975; Johnson 1985). Where both angular and spiral breaks were present, the breakage was recorded as spiral. In addition, recent breaks were noted. A more detailed breakage recording system was developed for Gomolava, inspired partly by Outrams (2001; 2002) Fracture Freshness Index (FFI). A separate breakage table allowed more than one type of break to be recorded in detail from the same specimen, but was only used for long-bones and metapodials. Fields were defined for breakage type (equivalent to Outrams fracture edge texture), direction (broadly equivalent to fracture outline), location and comments. Fracture angle was excluded since it is both time consuming and the weakest part of the FFI (Outram 2002, 59-60). Records were created for each break which appeared clearly to represent a discrete breakage event. For shaft breaks, breakage type was recorded as smooth, rough, intermediate or recent, and direction as transverse, longitudinal, diagonal or spiral. For breakages through articular ends, only direction was recorded. Fragmentation levels are compared between species, units and context types at each site. The relative merits of various measures are discussed by Outram (2001), and three are employed here: Percentage identification rate. A crude but simple measure of fragmentation at the unit level. Fragment maximum dimension in millimetres. This is useful for testing the impact of taphonomic processes on fragmentation, and for detecting possible pot-sizing. Percentage completeness. While both of the above are valid for comparison between sites or context types, neither is comparable between taxa or elements.
72
Methodology
Percentage completeness was therefore calculated for long-bones using the formula described by Morlan (1994):
PP / NISP PD
Where PP represents Portions Preserved and PD is Portions Defined. The portions used here are Dobneys and Reillys diagnostic zones (see 4.4.3). To assess the relative importance of different pre/post-depositional agents of fragmentation, these data are accompanied by a measure of breakage freshness. For Petnica this consists simply of the percentage of long-bone/metapodial fragments with spiral breaks. At Gomolava, a simple Freshness Index (FI) based on breakage type and direction scores (Table 4.6) was calculated for each breakage record, and mean values quoted by context type and taxon.
Breakage type Description Score Rough 0 Intermediate 1 Smooth 2 Recent Excluded
Breakage direction Description Score Transverse 0 Longitudinal 1 Diagonal 1 Spiral 2
Table 4.6: calculation of Freshness Index (FI). Scores are summed to give final value.
4.4.3.5 visible human modification

Leaving aside artefacts produced on bone, visible human modification refers to butchery marks and burning. Burning may sometimes be a direct result of food preparation, but the majority of recorded cases are probably too severe to reflect ordinary cooking practices (Kent 1993, 348) and may tell us more about depositional processes. The effects of heat treatment vary considerably with temperature (Shipman et al. 1984), duration and heating type (Nicholson 1993; Outram 2002; Pearce & Luff 1994), not to mention prior bone condition and subsequent (post)depositional processes (Bennett 1999; Svensson & Wendel 1965). Two fields were used to record burning at Petnica and Gomolava, for degree and location. Several studies have addressed the effect of heating and burning on bone colour. Shipman et al. (1984, 312) describe a pathway from yellows through reds and purples to diverse neutral hues, while Nicholson (1993) documents a sequence from brown to black, grey then white.. Since colour is affected by bone condition before heating and by depositional environment afterwards, no universal scheme can be drawn up. The colour categories used for burnt or probably burnt specimens from Gomolava and Petnica were refined through observation of
73
Methodology
the material itself, a process which occasionally entailed returning to previously studied units. For example, a bright yellow colour and plastic-like texture were commonly observed at both sites but especially Petnica. At first it was not clear whether this represented burning or simply unusually fresh bone, but it soon became apparent that the phenomenon very frequently coincided with clearly charred patches. Accordingly, the yellow appearance was assumed to represent burning or heating to a level lower than that required to cause blackening or perhaps under a different set of conditions and the category burnt yellow was defined to record it. Cut- and chop-marks, impact scars and chop surfaces were all recorded with a detailed description in the databases. Cut-marks are quantified as fragment-count rather than cutmark-count data as defined by Abe et al. (2002), and are presented as percentage NISP where sample size allows.
4.4.3.6 assessment of element representation

Correlation with utility indices has most often been employed in studies of hunter-gatherer groups, either to distinguish kill or primary butchery sites from camps, or to elucidate changes in selection and transportation strategies (e.g. Miracle 2002, 77-79). The latter approach is equally relevant for hunted animals in a Neolithic context, or indeed for domestic animals which may have been slaughtered off-site. The technique may also be applied to contexts or context types within a settlement, in which case the distribution of high and low utility parts may reflect patterns of structured deposition and/or differential access to resources. A fundamental shortcoming of this approach is that it assumes a universal conception of utility. Various indices have been calculated based on different combinations of resources such as meat, marrow and grease (see below), but in each case one must ignore the likelihood that factors other than weight and raw nutritional content play a part in selection, whether for transport, consumption or sharing. While correlation between element abundance and the chosen index may provide evidence for selective use and discard on the basis of nutritional value, lack of correlation does not imply that the distribution of elements is not structured by conscious human action. That action may simply not have followed a predictable, utility driven pattern. Body-part representation should always therefore be assessed graphically to identify any recurring patterns. Correlation with utility raises technical issues analogous to those of correlation with density, and the same solutions are employed. Binfords (1978) original Modified General Utility Index (MGUI), based on sheep and caribou, has been simplified and refined into a Food
74
Methodology
Utility Index (FUI) by Metcalfe and Jones (1988), who also offer a Complete Bones FUI, averaging proximal and distal ends. The only available figures for deer (Odocoileus Madrigal & Holt 2002) treat long bones as a single unit. There is some doubt over the appropriateness of splitting long bones into proximal and distal ends when dealing with field butchery, but since the main issue at stake here is the on-site distribution of portions which may very feasibly have been reduced to this extent, the basic FUI is employed for bovids and cervids. For pigs, whose morphology is substantially different, FUI values are taken from Pugsley (2002). More importantly, a subjective impression of body-part representation is gained by comparing bar-charts between sites, taxa and context types. At this stage, elements are aggregated to the level of Stiners (1991) anatomical regions, defined in Table 4.7. The MAU for a region is simply the highest element/portion MAU from that region.
# 1 2 3 4 5 6 7 8 9
Region Horn core/antler Head Neck Axial Upper forelimb Lower forlimb Upper hindlimb Lower hindlimb Feet
Consituent elements Horn core, antler Cranium, mandible Atlas, axis, other cervical vertebrae Thoracic, lumbar, sacral and caudal vertebrae; pelvis Scapula, humerus Radius, ulna, carpals, metacarpal(s) Femur Tibia, fibula, tarsals, metatarsal(s) Phalanges
Table 4.7: definitions of anatomical regions (Stiner 1991).
4.4.3.7 distribution and sharing

Sharing not only represents a primary determinant of faunal assemblage composition, potentially confounding other factors (Kent 1993), but is increasingly a topic of interest in itself. For the purposes of this project, differences in patterns of distribution between taxa are of particular interest. Sharing is notoriously difficult to identify in the zooarchaeological record, however. The most direct evidence comes from refits, pair-matching and articulations: if particular features or discrete concentrations of bones can be associated with different dwellings, and specimens from the same individual can be shown to be spread between two or more, then this is solid evidence for the distribution of body parts (e.g. Enloe & David 1989). This approach has been taken for red deer at Opovo (Prizer 1994), but has two drawbacks. Firstly, it requires good preservation, thorough recovery, and high chrono-stratigraphic resolution for useful results.
75
Methodology
Secondly, it is extremely time-consuming. Refits and articulations were recorded when observed at Gomolava and Petnica but no systematic study was undertaken. A second approach involves evidence for differential access to meat, the other side of the sharing coin. This is far from straightforward, however, even if deposits can convincingly be attributed to separate domestic groups. While structured asymmetrical sharing should create differentials in deposited food remains, sharing which is less structured, broadly symmetrical or simply averaged over a considerable length of time may serve rather to obscure any differences in the initial availability of carcasses (e.g. Kent 1993, 349-350). This is a gross oversimplification, since different sharing ethics may operate for different taxa, at different times and so-on. In addition, it does not account for sharing in the form of communal consumption or feasting, with the remains deposited either in a special context or in association with a host dwelling. Ideally one would hope to be able to identify such deposits through other aspects of taphonomy, but this is far from certain. Put simply, sharing will tend to smooth differences in species availability, at least to an extent, but may or may not create structured variation in skeletal part representation. While several possible scenarios lead to general uniformity (Figure 4.3), divergence from this may be very informative. Non-random variation in species composition strongly suggests that differential access to carcasses exists and is not entirely mitigated by sharing ethics. Differing element representation once functional variation between the deposits in question has been ruled out points either to asymmetrical sharing, perhaps relating to status, or to a structured pattern of distribution operating on differentially available carcasses. It is worth noting that, a priori, neither uniform hunting success nor equal access to domestic stock seems likely in any society. Marshall (1993, 234-235) notes that amongst the Okiek a combination of differential hunting success and the social relationships which structure carcass distribution leads to some variation in species composition but much more in body part representation. A very subtle line of evidence for sharing is applied to the Opovo fauna by Russell (1993; 2000). As well as calculating correlations between pairs of species and of elements in different pits at the site essentially quantifying variation in species composition and element representation respectively she correlates frequencies of left and right specimens of each element. The reasoning is that carcass distribution introduces the possibility of a random side imbalance. Since this will be averaged out over time, strong positive correlation tells us little, while weak or negative correlation provides circumstantial evidence for sharing. Finally, Russell looks at correlations between adjacent elements, on the basis that these should be higher for elements which are likely to be distributed together. Neither faunal sample studied here is amenable to this treatment.
76
Methodology
Figure 4.3: expected patterns of household faunal waste resulting from several hypothetical scenarios.
4.5 Summary and structure of results

This study has two main components: a regional review of all available published and unpublished faunal data, in order to assess overall trends and to discuss the economic roles of various taxa on a gross scale (Chapter 5); and the detailed first-hand study of faunal material from Gomolava and Petnica. The latter is presented in two parts. Firstly, Chapter 6 consists of basic faunal reports from each site, including background information, in-depth discussion of context types, and standard zooarchaeological data. This provides a frame of reference for the taphonomic-contextual study in Chapter 7, in which evidence for differential postmortem treatment of animal species is gleaned from five stages of taphonomic analysis. The two scales of analysis are drawn together by the discussion of Vina period humananimal relations in Chapter 8. This starts with a species-by-species consideration of the evidence, taking into account the known characteristics of each taxon, before moving on to consider more general themes.
77
Regional and chronological patterns
Chapter 5 Regional trends in faunal data

5.1 Introduction
I have advocated a fine-scale contextual/taphonomic approach to faunal remains as a means of moving beyond the top-down economic/ecological analyses of traditional zooarchaeology. This is not to say that the latter have no value. On the contrary, trends in species representation and related variables on both the site and regional level are an essential element in studies of past human-animal relationships. The roles of wild and domestic species within Vina society cannot be understood in the absence of coarse-scale data on their changing representation any more than they can be apprehended on the basis of such data alone. Chronological and regional trends are thus assessed here, prior to the in-depth case-studies in Chapters 6 and 7.
Figure 5.1: locations of Central Balkan Neolithic sites with published faunal assemblages. (a) Early Neolithic; (b) Vina and contemporary. Numbers refer to Tables 5.1 and 5.2.
Tringham (1971, 180) states that the importance of wild animals increased during the later Vina period, but virtually no data were available when the statement was made. The first part of this chapter (5.2) presents all available data in order to test this assertion and to assess patterns in taxonomic representation more broadly drawing on recent published and unpublished results, my own research, and the latest absolute dating. In the second part (5.3) various theories regarding the respective roles of wild and domestic animals in the later Neolithic are reviewed.
78
Table 5.1: studied faunal assemblages from Vina and neighbouring cultures. Taxonomic representation is shown as %NISP *indeterminate cattle and pigs distributed between wild and domestic pro rata for charts and calculations. **one box indicates figures for broad developmental groups; two boxess indicate dental data following Payne or Grant. ***faunal data only available for upper layers.
79
Table 5.2: earlier Neolithic faunal assemblages from the Central Balkans. *** full data from Foeni-Sala became available too late for inclusion, but are listed above for completeness. Other notes as Table 5.1.
(Babovi 1986; Bartosiewicz et al. 2001; Blai 1985; Blai 1992; Blai 2005; Bknyi 1969; Bknyi 1974a; Bknyi 1974b; Bknyi 1976; Bknyi 1984; Bknyi 1988; Bknyi 1992; Clason 1980; El Susi 1995a; El Susi 1995b; El Susi 1996a; El Susi 1996b; El Susi 1996c; El Susi 1998; El Susi 2000; El Susi 2003; Greenfield 1993; Greenfield & Draovean 1994; Greenfield & Jongsma 2008; Jongsma & Greenfield 1996; Jovanovi et al. 2003; Jovanovi et al. 2004; Lazi 1988; Moskalewska & Sanev 1989; Nicolescu-Plopor 1980; Pipe n.d.; Schwartz 1976; Schwartz 1992; Vrs 1980)
80
5.2 Regional and chronological trends in species abundance

5.2.1 the sample
A surprising number of Vina faunal assemblages have been studied, although detailed publications are few (Table 5.1; Figure 5.1b). The Banat and Turda groups are treated as variants of early Vina here. Obre II and Foeni-Cimiturul Ortodox, from the neighbouring Butmir culture and Foeni group respectively, are also included as large samples from sites geographically close to and chronologically contemporaneous with Vina. Taxonomic representation is known from twenty-two sites, excluding Bardhosh and Rast. Earlier Neolithic primarily Starevo assemblages are also considered in order to detect any longerterm patterns (Table 5.2; Figure 5.1a). Post-Vina sites are excluded since their relevance for trends during the preceding periods would imply teleological reasoning. The geographical limits of the later Neolithic sample are based on the maximum extent of the Vina culture: sites lying within this area and contemporary with the culture are included, plus Obre. The early Neolithic sample is defined in the same way, including all pre-Vina assemblages. It thus covers only part of the Starevo-Krs-Cri complex. Some conspicuous gaps exist. The area between the Zapadna Morava and the Ove Pole is devoid of data apart from a tiny sample from the Vina site of Bardhosh, near Pritina (Pipe n.d.; T. Kastrati, pers. comm.). This is due not to a lack of sites but simply of faunal studies. More problematically, coverage shifts between periods: a concentration of studied Early Neolithic sites in northern Banat and Baka is replaced by a spread of Vina settlements further south in the Banat and around the Danube-Sava confluence. Again, this probably reflects research rather than settlement patterns: to my knowledge, there are simply no faunal studies from later Neolithic sites in the north of the Vojvodina or Romanian Banat. The Maros is taken as an arbitrary northern limit for earlier sites. umadija and erdap, meanwhile, are fairly well represented in both samples. Collating and interpreting the data is far from straightforward since numerous variables, from environmental differences to excavation strategy, influence observed variation. Percentage NISP is used as the measure of relative taxonomic abundance throughout, since (a) it is most commonly available and (b) the manifold problems of MNI are amplified when comparing data between analysts (see 4.3.3). However, this increases the potential for preservation and recovery biases between sites. Little can be done to control for the former, but the latter can at least be evaluated through comparison of sites with known recovery strategies.
81
Chronological resolution also varies: while some assemblages are fairly well dated, others are simply described as Vina. Data are therefore assessed at three levels. Firstly, they are grouped into pre-Vina and Vina/contemporary, allowing inclusion of all assemblages except Rug Bair, which straddles the division. Each group is subdivided to give Early Neolithic I, Early Neolithic II, Vina A/B and Vina C/D (including contemporary sites in the latter cases). The division between EN I and EN II is set arbitrarily at 5800 cal. BC. Multiphase assemblages are divided or aggregated as appropriate for each scale, and the resulting figures are given in Appendix 2. Finally, radiocarbon-dated assemblages are plotted on an absolute chronological scale. This also allows changes over time in multi-phase sites to be tracked, providing some control for regional/environmental variations.
5.2.2 dated sites

The following sections review the dating of sites with later Neolithic fauna and a reasonable complement of radiocarbon determinations. Models are kept simple since (a) stratigraphic information is usually limited and/or ambiguous, (b) frequent use of dates on charcoal undermines direct relationships between samples, and (c) the precision required here does not justify the effort or risks involved in detailed modelling. For pre-Vina assemblages other than Divostin I, Anza I-III and Obre I, I rely upon published discussions (Bori & Miracle 2004; Whittle et al. 2002; 2005).
5.2.2.1 Gomolava
The dating of Gomolava is discussed in 6.3.2. The first phase, Ia, is poorly dated but falls around probably a little after 5000 BC, typologically placed as Vina B2/C. Boundaries for the third phase, Ib (excluding the cemetery), bracket roughly 300 years, from perhaps 4900 to 4600 BC, although the old-wood effect is probably at work. Typologically, Ib is currently attributed to early Vina D (Brukner 1988; 1990), although there is some ambiguity (see e.g. Brukner 1980a). The second phase, Iab, is undated and poorly stratigraphically distinguished from Ia, but must be very early 5th millennium BC.
5.2.2.2 Divostin
Divostin features Starevo (I) and Vina D (II) phases, each with substantial faunal assemblages. A shortage of secure samples (see McPherron et al. 1988), is mitigated by three AMS dates from house floors (Bori & Jovanovi in press). Figure 5.2 shows the available dates grouped only by phase. Divostin I lies at the start of the 6th millennium BC, while Divostin II covers roughly the same span as Gomolava Ib. Again, old wood may be implicated, especially since the latest four dates are also the only non-charcoal samples. On
82
the other hand, these all come from within destroyed houses and might genuinely be pre-dated by charcoal samples from pits. Occupation within the period 4800-4600 seems most likely, but a later start cannot be excluded.
5.2.2.3 Selevac
Selevac is divided into four Stratigraphic-Architectural (S-A) phases, the question of continuity between which is left open (Tringham & Krsti 1990a). Eight samples were dated altogether from S-A I-III, with a further three from grain in a silo loosely assigned to S-A II. One of the latter (Z-233) was impossibly early compared to the others and rejected. The two dates from S-A III are considerably later than those from I-II (Figure 5.3) but several centuries too late for the typological placement of S-A III in Vina B/C. Descriptions of both samples reveal some doubt over their phasing (Tringham & Krsti 1990a, 50) so it is possible that they are intrusive from S-A IV or even later. The dates from S-A II suggest occupation in the last century of the 6th and/or the first century of the 5th millennium, and the lone date from S-A I is unsurprisingly a little earlier.
5.2.2.4 Anza
Both faunal assemblages and radiocarbon dates are published for Aegean Early Neolithic (I), Starevo (II/III) and Vina B (IV) phases at Anza (Bknyi 1976; Gimbutas 1976). From the typology and stratigraphy, Gimbutas suggests a possible hiatus between I and II, while the absence of dates from IVa forces a hiatus between III and IVb which may or may not be genuine. Ia-Ib and II-III are considered continuous. The dates generally fit with this model (Figure 5.4), although one sample from Ib (LJ-2333) was rejected due to very low agreement. Under this model, the duration of Anza I is very short, probably restricted to the last century of the 7th millenium. The Starevo phases probably span 5900-5500 BC, with a transition from II-III loosely around 5600. Anza IVb can only be narrowed down to roughly the last three centuries of the 6th millennium, matching its Vina B designation.
5.2.2.5 Opovo
Opovo boasts the most detailed Vina faunal study to date (Russell 1993). Its three building horizons (BH1-3) are dated primarily typologically to Vina C. Two unpublished dates (Tringham, pers. comm.) are consistent with this, having 68.2% Highest Probability Densities (HPDs) entirely within the first two centuries of the 5th millennium when treated as a phase. Since both are from the earliest phase (BH3), the duration of occupation requires independent checking. Funding has been secured for nine AMS dates, but in the meantime the estimate of not more than 200 years is accepted (Tringham et al. 1992, 365). This would put BH1 towards the later limit of Vina C.
83
Figure 5.2: radiocarbon dates from Divostin. All date references given in Appendix 1.
Figure 5.3: radiocarbon dates from Selevac.
84
Figure 5.4: radiocarbon dates from Anza.
85
Figure 5.5: radiocarbon dates from Obre I & II .
86
5.2.2.6 Belovode
The fauna from Belovode is published without information on context or phasing (Jovanovi et al. 2003; 2004). Nine AMS dates cover almost the entire range of the Vina culture, but until stratigraphic details are published (Bori & Jovanovi in press) these cannot be brought to bear on the faunal sample.
5.2.2.7 Vina
The relationships between relative and absolute chronology at the type-site are now wellstudied (Schier 1996; 2000). Unfortunately, all faunal data are from Vina D, while NISP counts are only given for material from the current excavations (Dimitrijevi forthcoming). No samples in association with this material have yet been dated.
5.2.2.8 Obre I & II

These two Central Bosnian sites produced large assemblages attributed to Starevo (Obre I, A), Kakanj (Obre I, B-C), Butmir I (Obre II, I) and Butmir II (Obre II, II-III). Although modelled independently (Figure 5.5), occupation at Obre II appears to start as Obre I finishes. The Starevo and Kakanj dates cover a millennium, and the final phase of Obre I appears contemporaneous with Vina A/B. Chronological resolution is better for Obre II site, with phase I dated to roughly 5000-4800 BC parallel with Vina C and the Butmir II phases to the third and fourth centuries of the 5th millennium roughly equivalent to Vina D.
5.2.2.9 Para
In the absence of stratigraphic data the Para dates are treated as a single phase, giving a likely range of 5400-5050 cal. BC. The three faunal assemblages are been spread arbitrarily through this range for the sake of graphical representation, but are treated as relatively dated.
5.2.3 recovery bias

Figure 5.6 shows the percentage of small mammal i.e. caprine, dog and roe deer specimens in each assemblage, grouped according to the extent of sieving. The some category varies from probably over half of units at Petnica (see 6.4.4.2) to limited flotation at Vina, the exact regime rarely being detailed. At first glance, sieving appears to be the major factor, with small mammals only exceeding 40% in completely sieved or unknown assemblages. The two Anza samples, however, are from the semi-arid zone south of the Danube watershed where a predominance of caprines is typical (Greenfield 1991); while
87
sieving will undoubtedly have inflated the small mammal counts from Anza, they are likely to have been amongst the highest anyway.
Ludo Budak
R
Anza I-III
75
R
Donja Branjev inaR Anza IV
Mihajlov ac-Knjepite R
Small mamma ls (%NISP)
Rszke-Ldv r Gy llart-Szilgy i
R R
50
R
Madari
Obre I R
R R
Nosa
R R R R R
25
Selev ac
R R R R R R
R R R R R R R R R R R R R
Petnica Div ostin I Obre II Div ostin II
R R R
Gomolav a R
R
All
Some
None
Unknown
Units sie ve d
Figure 5.6: percentage contribution of small mammals by recovery standard.
The highest values in the unknown category mostly relate to sites from the semi-arid zone (Madari) or belonging to the Krs group in the southern Great Hungarian Plain. The status of caprines as the main Early Neolithic domestic species in this area in contrast to areas south is well established (Bartosiewicz 2005; Bknyi 1974a, 26; Vrs 1980), and there is no reason to believe that systematic differences in recovery are responsible. Donja Branjevina can perhaps also be included in this category. 88
That recovery techniques introduce bias is undeniable. Payne (n.d., cited in McPherron 1988, 6) demonstrates a bias against small bones at Divostin, and the same is true at Petnica (6.4.4.2). The potential severity of the problem is illustrated by comparison between the Yugoslav and more slowly excavated American sondages at Obre I, or between unsieved and partially-sieved samples from Opovo (Table 5.3). However, this bias does not appear to be the main source of variation amongst the assemblages reviewed here, and one might note differences between Obre I & II, excavated by the same teams. In fact, since the majority of the unknown assemblages probably enjoyed little if any sieving, it is actually the wellcollected samples Selevac, Anza, Petnica and perhaps Opovo which obstruct comparison. NISP is a biased proxy-measure of species abundance even under perfect conditions; for purposes of inter-site comparison the most important requirement is for similar biases to apply.
Obre I (Bknyi 1974b) 'American' 'Yugoslav' 48.3 72.7 9.7 6.8 39.4 19.1 Opovo Partly sieved* Unsieved** 63.1 70.3 18.5 17.7 18.4 12.0
Large Medium Small
Table 5.3: body size distribution by sample at Obre I and Opovo. * Russell 1993 **Greenfield 1986
5.2.4 wild versus domestic taxa

Figure 5.7 shows the overall distribution of wild components in pre-Vina versus Vina/contemporary assemblages. Wild here includes deer, wild pig3 and cattle, equids, beaver, lagomorphs, and carnivores other than dogs. Small rodents and insectivores are probably intrusive and are not considered. Fish, birds, reptiles, amphibians and invertebrates are excluded due to recovery problems and incommensurable quantification, not to mention the very different nature of activities involved in their procurement. It is immediately apparent that an overall increase or decrease in the importance of wild taxa cannot be supported at this scale: both periods feature a range of variation from almost no wild specimens to nearly three-quarters, with no clear shift in mode. If anything, the data show a reduction in diversity more Vina/contemporary assemblages cluster in the centre of the range. Nonetheless, the median does appear to increase slightly, and this becomes more apparent at the next scale (Figure 5.8): a clear increase between EN II and Vina A/B is followed by a marginal decrease into Vina C/D. However, these shifts are dwarfed by the variation observed within each period.
3
Wild pig is used in preference to wild boar throughout this thesis. Boar, where used, refers to any male member of the genus Sus.
89
Early Neolit hic 4
Frequency
3 2 1
Vinca/contemporary 4
Frequency
3 2 1 0 25 50 75 100
Wild taxa (%N ISP)
Figure 5.7: distribution of wild components in Early Neolithic versus Vinca/contemporary sites
80
n = 13 n=7 n = 10
n = 12
Wild taxa (%NISP)
60
40
20
0 Early Neolithic I Early Vinca A/B Neolithic II Vinca C/D
Figure 5.8: range of variation in wild component of assemblages by period
Since much variation probably relates to sub-regional and/or environmental differences (Lazi 1993), contributions of wild and domestic species are plotted according to geographic location in Figure 5.9. There is some evidence for regional differences in the Early Neolithic: wild animals are poorly represented in the Macedonian sites, and unsurprisingly common both in erdap and at Golokut (#45), situated in the ideal hunting grounds of the Fruka Gora. Much variation cannot be accounted for in these terms, however, as illustrated by Ludo Budak (#36) and Nosa (#39), both situated just north of modern Subotica in Baka yet with very different assemblages.
90
Figure 5.9: contributions of wild and domestic taxa to (a) pre-Vina and (b) Vina/contemporary assemblages
91
Moving into the Vina period, the nature of change is unclear. A Srem/Danube-Sava-Morava confluence cluster is characterized by fairly large wild components, especially if one includes Opovo (#9). Boljevci (#5) stands out here and is unusual overall, with cattle and aurochs jointly constituting 93.4%. Poor recovery is no doubt a factor, but cannot account for the nearabsence of red deer. Since this area features only a single Early Neolithic assemblage (Starevo, #31), albeit with relatively few wild specimens, one cannot safely talk of an increase. Rather, the generally domestic-dominated Krs sites are replaced by clusters along the Tami and around Belgrade, and it is thus impossible to distinguish geographical from chronological trends from the available data. Where there is continuity, no such increase is seen: in erdap there is an apparent decrease in wild specimens, while Anza (#8, 28), Divostin (#1, 27), Obre I & II (#24, 25, 29) and the two sites at Foeni (#23, 46) change little. Figure 5.10 shows the data at the next level of resolution, comparing Vina A/B with Vina C/D. There is rather more evidence for change here, with domestic counts increasing substantially at Selevac (#3), Gomolava (#6), between Obre I C (#25) and Obre II (#24), and at the sites north of the Tami, (#21-23). Petnica (#10) and Liubcova (#13) barely change, however, while the appearance of two new sites with large wild components in the southern Banat, coupled with the abandonment of Anza, largely counters any decrease in wild specimens in the overall figures. Nor does any trend become apparent at the third scale (Figure 5.11). Evidence for systematic change over time is clearly limited, mostly involving successive phases at individual sites. Multi-phase assemblages are useful in that they provide a control for sub-regional variations allowing one to compare like with like in environmental terms and usually also for archaeological and zooarchaeological methodology. They may introduce problems or opportunities of their own, however, since it is feasible that prolonged occupations at the same locale might result in trends in the exploitation of wild species over and above any regional-level patterns, perhaps due to resource depletion (see e.g. arguments in Speth & Scott 1989). Figure 5.12 shows changes over time in tolerably well-dated multiphase assemblages. The near-absence of pre-Vina multi-phase sites none within the core area of the sample becomes apparent on comparison with Figure 5.11, and is probably more than a research bias. The relatively low levels of wild animals at Anza and Obre I compared to those at shorter-lived sites, however, probably reflect geographical and cultural differences as much as they do an effect of long-term occupation.
92
Figure 5.10: contributions of wild and domestic taxa to (a) Vina A/B (and contemporary) and (b) Vina C/D (and contemporary) assemblages
93
80
60
40
20
6000
5500 Approximate date (cal. BC)
5000
Some absolute dating; bars represent uncertainty
0 4500
Well dated; bars represent likely start and end points
Figure 5.11: percentage contributions of wild taxa to assemblages by approximate date
Secondly, two of the larger Vina sites, Gomolava and Selevac, show a similar trend away from wild taxa, albeit from different starting levels. A more gradual shift is seen slightly earlier at Snandrei, a relatively large site in the Banat (Jongsma & Greenfield 1996). Shorterlived sites such as Obre II and Opovo show little overall change, and it might be noted that these are in areas of fairly low agricultural potential, as is Liubcova. Para, however, is not, and any consistent connection between large permanent sites and declining wild resource exploitation may be undermined by Vina: although faunal data are only available from the end of the long sequence at the type-site, wild species make up 41.8% by NISP (Dimitrijevi forthcoming). The question of settlement continuity is left open at Selevac in any case (Tringham & Stevanovi 1990, 54-55), although it does seem likely at Gomolava (see 6.3.2). At Petnica, where discontinuities between phases are perhaps more likely, very little change occurs.
94
Wild taxa (%NISP)
Well dated; bars represent likely start and end points Some absolute dating; bars represent uncertainty Date estimated purely on relative grounds
80
Opovo Para Petnica Liubcova Snandrei Gomolava Obre I 20 Divostin Anza 6000 5500 Approximate date (cal. BC) 5000 Obre II Selevac 0 4500 Wild taxa (%NISP) 60
40
Figure 5.12: representation of wild species over time at multi-phase sites.
In summary, there is no single overarching trend in the relative representation of wild and domestic taxa in the Central Balkans either between the Early and later Neolithic or within the Vina period. Indications of a slight increase in hunted taxa at the start of the Vina period are confounded by a geographical shift in the available sample, and their representation actually decreases in some consistently-documented areas. Within the period, a very slight apparent decrease in wild taxa is largely driven by changes at some established sites, while several new Vina C/D assemblages have substantial wild components. Gaps in the data complicate matters at both scales, but there is considerable variation between sub-regions and individual sites. The continuing importance of hunting throughout the Neolithic is undeniable, however, with wild taxa constituting more than 25% of most Vina faunal assemblages.
5.2.5 taxonomic representation

Percentage wild is an extremely simplistic means of characterizing faunal assemblages. The wild:domestic split has considerable significance for the role of animals in the human political economy, as argued in 3.2.3, but important differences in likely forms of human-animal relationship also exist within each group. Differences amongst domestic species have particular relevance for human social organization, but physical characteristics such as body
95
size not to mention the evident similarities between wild and domestic pigs and cattle may transcend the dichotomy in their implications for human use of and attitudes towards animals. Figure 5.13 shows proportions of different domestic species at Early and later Neolithic sites in the region. There is a shift from caprines to cattle, although much of this can be attributed to the cluster of early (Krs) sites around Subotica and Szeged. These stand out even in the Early Neolithic and cannot be accounted for in terms of environment, which is neither greatly differentiated from sites to the south and west in the Vojvodina and Romanian Banat nor especially suited to caprines (Bartosiewicz 2005, 51). Whittle (2005, 68) suggests that Krs sheepherding may reflect an association of caprines with the collective past. Given the ubiquity of animals in everyday Neolithic life, and their intimate involvement in human social relations, particular forms of human-animal relationship with their attendant spatial and temporal rhythms must have been at least as important as material culture for community identity in certain times and places. The dominance of cattle in the Vina period is striking, especially in the lowland areas of the Banat, Srem and Danube-Sava confluence, although there are few earlier sites for comparison. Change is less apparent within the period (Figure 5.14). The contribution of cattle increases at Snandrei, Gomolava and Selevac along with that of domestic taxa taken together. The data from the latter site are problematic, however, since Legge does not give relative proportions of wild and domestic cattle by S-A phase, and these had to be estimated pro rata. The increase in representation of cattle here may be illusory if the wild:domestic ratio amongst pigs dropped significantly in the later phases. Meanwhile, cattle become less abundant at Liubcova, and no change is seen at Petnica. Figure 5.15 shows percentages of cattle amongst the main domesticates on an absolute chronological scale. These data are subject to considerable preservation and recovery noise, but an overall increase is apparent. Taking only multi-phase assemblages (Figure 5.16), major increases are seen at Selevac, Snandrei and Para. The increase at Gomolava slows after caprines drop out of the poorly recovered assemblage almost completely. A distinct decrease is seen at Liubcova, and less clearly at Obre II, Opovo and Petnica.
96
Figure 5.13: proportions of domestic species (%NISP) in (a) Early Neolithic, and (b) Vina/contemporary assemblages.
97
Figure 5.14: proportions of domestic species(%NISP) in (a) Vina A/B and (b) Vina C/D assemblages.
98
100 Cattle as percentage of domesticates, excluding dog (NISP)

Cattle as percentage of domesticates, excluding dog (NISP)
80
60
40
20
6000
5000
0 4500
Figure 5.15: percentages of cattle amongst domesticates at Central Balkan Neolithic sites. Squares represent sites with regular sieving.
Well dated; bars represent likely start and end points Some absolute dating; bars represent uncertainty Date estimated purely on relative grounds
100
Gomolava Obre II Opovo Snandrei Obre I Divostin Selevac Para 40 Liubcova Petnica 60 80
20 Anza
6000
5000
0 4500
Figure 5.16: percentages of cattle amongst domesticates in multi-phase assemblages.
99
The increase in abundance of domestic pig remains in the Vina period is clear, and they are present in appreciable numbers at only two earlier sites. Since wild pigs are widespread in both periods this requires careful consideration, and the trend could be seen as a shift from wild to domestic (Figure 5.17). The potential grey area between unambiguous hunting and close husbandry amongst pigs is well established (e.g. Albarella et al. 2006a; Dwyer & Minnegal 2005; Rosman & Rubel 1989), but Balkan Late Neolithic specimens typically separate very clearly into wild and domestic groups (6.2.1). The few large samples available from earlier sites are contradictory. At Obre I almost all measured pigs match the wild group at Obre II and elsewhere (Bknyi 1974b, 76-79), but a single cluster at Starevo identified mainly as wild (Clason 1980) falls squarely between the wild and domestic groups at later sites, prompting suggestions of local domestication (Legge 1990, 222-224). At Divostin, meanwhile, clear groups emerge despite the pooling of measurements separated by a millennium, and in fact proportionally more specimens are identified as domestic in Divostin I than II (Bknyi 1988). Fully morphologically domestic specimens were thus present in the region before the possible local domestication at Starevo.
100
80
60
40
20
6000
5000
0 4500
Figure 5.17: percentages of pigs identified as wild at Central Balkan Neolithic sites.
100
Percentage of pigs identified as wild (NISP)
This is particularly interesting given recent European mtDNA evidence. Larson et al. (2007) demonstrate a Neolithic introduction of Near-Eastern domestic pigs, replaced by populations with exclusively European haplotypes by at least the 4th millenium. Their Balkan Neolithic sample derives from four Romanian sites dated to around 5500 cal. BC, but it is plausible that the Starevo pigs represent an early example of the replacement process. This would have two implications: firstly, the morphological status of pigs from Central Balkan sites roughly contemporaneous with Starevo (my EN II) would be a poor indicator of their actual engagement with humans; secondly, most if not all Vina period pigs would be descended from local wild populations, losing their exotic status. Without direct genetic evidence this remains speculative, but in any case two essentially reproductively isolated populations clearly existed by the start of occupation at Selevac. Figure 5.18 and Figure 5.19 compare wild taxa between periods, showing little change. Proportions of wild species are of course particularly dependent on environment but are certainly not a passive reflection of it; people do not hunt by driftnet. Red deer are in general the prevalent species, but are replaced by aurochs in a few cases and by wild ass at two Krs sites (#36, 39). Roe deer are usually present in small numbers, only exceeding 25% at Golokut (#6) and in the latest phase at Snandrei (#22). As the smallest major wild species, their relative paucity may be an artefact of recovery in some cases. Unsurprisingly, they are marginally more common in lowland areas. Wild pigs are also nearly ubiquitous, although their distinction from domestic pigs may not always be straightforward, especially in my EN II phase (see above). Aurochsen vary from 1% of wild specimens to over 25% at a few sites, and these are not always where one would expect on environmental grounds: while Boljevci (#5) and especially At (#11) may have been near prime habitat for aurochs, this is less likely at Divostin, Anza and Madari. Apart from At, however, these all have few wild remains overall, so the absolute number of aurochs specimens is still small. Aurochsen thus appear relatively independent of other wild species, confirmed by much higher correlation with domestic cattle (raw NISP: Pearsons r = 0.9360) than with all other wild species combined (Pearsons r = 0.4451). Other game is present in limited numbers. Excluding the sites with abundant wild ass, other mostly consists of small carnivores mustelids, fox and wildcat hare and beaver, the latter reaching 1.2% of total NISP at Petnica. Larger carnivores wolf, brown bear and occasionally lynx are also represented, especially at upland sites and in erdap.
101
Figure 5.18: proportions of wild species (%NISP) in (a) Early Neolithic, and (b) Vina/contemporary assemblages.
102
Figure 5.19: proportions of wild species(%NISP) in (a) Vina A/B and (b) Vina C/D assemblages.
103
5.2.6 taxonomic relationships

The above discussion hints at relationships between the frequencies of certain taxa which are difficult to investigate due to the seldom-noted compositional nature of faunal data. The absence of any independent reference scale effectively closes faunal datasets; the frequency of any taxon can only be measured relative to a value of which it is itself a component total NISP. It is thus impossible to say whether the frequency of one taxon has increased or that of another decreased, nor can correlations between them be taken at face value: by default taxa should be correlated positively if raw counts are compared or negatively if percentages are used. Comparing NISP/(volume of sediment) potentially alleviates these problems by introducing an independent referent, but such data are rarely available. Two statistical approaches have been used to assess relationships between variables in compositional datasets. The correct method uses log-ratios to remove the inbuilt negative correlations from percentage data (Aitchison 1986). The resulting covariance structures are difficult to interpret, however, and their empirical benefits are not uncontroversial (e.g. Baxter 1993; 2006; Tangri & Wright 1993). Secondly, a datasets correlation matrix can be simplified using conventional multivariate statistics. This is theoretically unsound for true compositional datasets, but may be appropriate for faunal data where the closure is introduced by lack of outside control rather than by underlying sampling conditions. Multivariate statistics reduce the dimensionality of datasets, extracting a manageable number of dimensions (components or factors) while retaining as much variability as possible. In the case of Principal Components Analysis (PCA), these dimensions are by definition orthogonal. The first component is simply the axis through the multidimensional data cloud which accounts for the maximum possible variance; the second component is the line orthogonal to this which accounts for the maximum of the remainder. The third is orthogonal to both of these, and so-on. Readers are directed to Baxter (1994) for discussion of multivariate statistics in archaeology. Bartosiewicz (2005) uses Factor Analysis4 with Varimax rotation to investigate interspecific relationships on Neolithic sites from the Great Hungarian Plain, showing that most variation can be expressed by two factors. He labels these hunting and animal keeping, although an environmental distinction between forest and grassland is more apparent in his interpretation and loading plots, with cattle and domestic pig moderately loaded on each factor.
4
Probably actually PCA, which is subsumed under Factor Analysis in some popular software packages.
104
Table 5.4 shows the results of PCA on taxonomic composition (raw NISP) in the Central Balkan sample (50 assemblages from 33 sites). Three components were extracted, of which the first accounts almost entirely for variation in sample size (correlation between Component 1 and n: Pearsons r = 0.9369), representing 58.4% of total variance. Applying rotation would remove this, but at the cost of reduced transparency (Baxter 1994, 90). Lower loadings on this component for caprines, roe deer and other (mostly small game) probably reflect a body size effect, with differential recovery intervening to weaken relationships between smaller taxa and overall sample size. Roe deer are less affected than caprines, while the apparent immunity of dogs probably reflects very low fragmentation (personal observation at Gomolava and Petnica). The second component is of most interest, apparently representing a distinction between rough wild and domestic categories (Figure 5.20a), discussed below. This accounts for 18.8% of variance, a considerable amount given the nature of the first component. The third component is of limited worth since its eigenvalue is below the usual cut-off of 1, but similar loadings for both wild and domestic pigs and cattle are interesting, suggesting a weak negative relationship between these taxa and caprines, with deer and dog somewhere in-between (Figure 5.20b).
Caprines Wild cattle Pig Cattle Other Dog Red Wild pig Roe Eigenvalue % Variance
1 0.5237 0.8294 0.8483 0.8432 0.5701 0.8324 0.8174 0.8376 0.6921 5.2596 58.44
Component 2 0.5420 0.4309 0.4150 0.4086 0.1091 -0.3030 -0.4317 -0.4304 -0.6275 1.6877 18.75
3 0.3284 -0.1976 -0.2037 -0.2387 0.7213 0.0540 -0.0127 -0.1549 0.0721 0.7979 8.87
Kaiser-Meyer-Olkin Measure of Sampling Adequacy: 0.819 Bartlett's Test of Sphericity: p < 0.001
Table 5.4: unrotated component loadings for Central Balkan Neolithic faunal assemblages.
105
a)
1.0
Caprines 0.5 W.Cattle Pig Cattle
Component 2
Other 0.0
Dog W.Pig -0.5 Roe Red
-1.0 -1.0 -0.5 0.0 0.5 1.0
Component 1
b)
1.0
Caprines 0.5 Cattle Pig W.Cattle
Component 2
Other 0.0
Dog W.Pig -0.5 Roe Red
-1.0 -1.0 -0.5 0.0 0.5 1.0
Component 3
Figure 5.20: component loading plots.
106
Returning to the second component, two clusters include (a) cattle, caprines, pig and aurochs, and (b) dog, wild pig and deer, with other falling between. These can be seen as packages of statistically co-occurring taxa, but what exactly this means is a matter of interpretation, PCA being a purely descriptive technique. The results do indicate that wild and domestic species with some exceptions form two alternative complexes, justifying their separate analytical treatment within reason. They do not cannot address underlying causes, however, leaving the question of a perceived wild:domestic dichotomy open (Bartosiewicz 2005, 53). Similarity with Bartosiewiczs results is unsurprising given that the samples overlap. A clearer-cut division in my case presumably relates to the lack of major wild grassland species, wild ass and hare being rare and subsumed under other. Otherwise, the main difference is that aurochsen fall in the wild group for Bartosiewicz, but are strongly related to domestic cattle here. This relationship has already been noted (5.2.5) and indicates either that (a) wild cattle are hunted more where domestic cattle are most important which would be extremely interesting or (b) many aurochsen are actually large domestic cattle. The latter seems more likely given the biometric difficulties (see 6.2.2) but it is surprising that clarity of separation should differ so much between overlapping samples, especially since Bknyi was responsible for many identifications in each case. Dogs fall in the wild group in both samples. Figure 5.21 plots each assemblages scores on the second and third components. An overall shift from sheep to cattle and pig (plus their wild counterparts) is seen in the neardisappearance between EN II and Vina A/B of Component 3 values greater than zero. The high loading on this component for other is here revealed to relate to the two Krs sites with abundant wild ass. A slight overall reduction in wild taxa (i.e. Component 2 scores) is also seen. However, both these trends result from the disappearance of the northernmost Early Neolithic sites; several other early sites lie within the densest cluster of Vina/contemporary points. Changes between Early and Late Neolithic are thus confounded by geographical sample bias, but there is some indication of diversification in taxonomic make-up between earlier and later Vina sites.
107
Divostin II
1.5
Anza IV Ludo Budak
Divostin I
0.5
Nosa
-2 Component 2
-1 -0.5
1
Vina
-1.5
Opovo (BH2)
Opovo (BH1)
-2.5 EN I EN II Vina A/B Vina C/D (EN) (Vina)
-3.5
Opovo (BH3)
-4.5 Component 3
Figure 5.21: faunal assemblages plotted by scores on the 2nd and 3rd components.
5.3 Discussion
5.3.1 roles of wild species
Perhaps the most striking feature of the faunal data from the Central Balkan Neolithic is that, despite numerous subtle changes, a range of wild taxa remains of considerable importance throughout the period. Even by the later Vina phases often considered part of the Copper Age this shows no signs of changing. The view is taken here that the presence of considerable proportions of wild taxa in later prehistoric assemblages does not require explanation per se. The notion that it does is founded on an essentialist conception of the Neolithic itself, one that stems from the periods long history as a pivotal stage within evolutionist schemes of human development (see 2.1). Wild resources frequently play a major part in the subsistence of Neolithic communities (Boyle 2006), and in many cases one could quite reasonably turn the question of why farmers hunt around to ask why Neolithic hunter-
108
agriculturalists also chose to keep domestic animals. We seek to explain Neolithic hunting since it is a practice which is out of place according to our own imposed framework. Of course, wild animals do all but drop out of the faunal record at the onset of the Neolithic in some regions, for example in Poland and Greece. Such an extreme shift is hard to explain in pure ecological terms, seeming rather to indicate deliberate choices regarding appropriate and inappropriate activities (Hodder 1990, 117; Marciniak 2005a, 205; Perls 2001, 152). This is an interesting phenomenon and deserves study as such, possibly indicating a conceptual distinction between wild and domestic in such cases. However, it should certainly not be taken as a template for the Neolithic in general. In other regions a gradual decline in the contribution of wild taxa is apparent. This fits the prediction of a number of more economically-driven models (e.g. Harris 1996). Zvelebil (1992) predicts steadily decreasing dietary importance for hunted meat as agriculture becomes established, with a residual risk-buffering role also fading away as reduced mobility and agricultural intensification decrease the availability of game. Wild animals may retain social significance for some time but will eventually come to be replaced by domestic species in terms of symbolism, after which they are largely seen as a commodity, with hunting aimed at the procurement of specific tradable resources such as fur (Zvelebil 1992, 8-10). Theoretically, one might object to the unilinear evolutionary nature of such schemes, or to their relegation of the social and cultural significance of animals to epiphenomena. Empirically, they cannot account for a later Neolithic increase in the contribution of wild species seen in many regions, including Greece (Halstead 1999), Poland (Marciniak 2005a), the Hungarian plain (Bartosiewicz 2005), and arguably in the current case study (5.2.4). Moreover, there is something rather teleological in accounting for the relative abundance of wild species in terms of a trajectory towards eventual obsolescence, rather than focusing on their roles at the point in question. Wild animals are a central feature of Neolithic life in many regions, and this should not require special pleading. This is not to say that the plethora of models proposed for their role within farming societies is of no value. Rather, I wish to cast these as ways of characterizing how, not explaining why, wild species were important. Several models are reviewed briefly below in the Vina context.
109
5.3.1.1 micro-regional specialization

High percentages of wild animals at certain sites may be explained in terms of the local environment; either sites were located specifically for this purpose in which case one might expect an extreme dominance of hunted species or hunting was simply a common supplementary activity in areas with abundant game and more modest agricultural potential. The latter argument can perhaps be applied to Early Neolithic sites in erdap and the Fruka Gora, with typically higher wild:domesic ratios than those in Central Serbia or on the Hungarian Plain. By the Vina period, however, a majority of assemblages are over 25% wild by NISP. Environment will surely have been a factor in the balance of taxa represented, but we are not dealing with isolated special cases. Interestingly, wild animals actually cease to dominate at Vina sites in erdap.
5.3.1.2 a seasonal resource

Given that the availability of most food resources, both domestic and wild, follows seasonal cycles, it is almost inevitable that hunting, where practised, will have a seasonal aspect. Relevant data from Vina sites are extremely thin, but do not suggest a strongly seasonal basis. Suggested seasons, based primarily on mandibles, are necessarily tentative, in no way implying that all hunting took place during the months in question. Certain species appear to have played seasonal roles at certain sites, but these are not consistent (Table 5.5). At Opovo the largest sample red and roe deer were probably hunted to some extent year-round. One might in any case question whether a seasonal nature for hunting would necessarily place it on the periphery of the economy, let alone of society; agriculture and husbandry also have strongly seasonal cycles of labour input and return but are nonetheless typically considered the core of Neolithic subsistence.
Suggested season Roe deer Late winter/early spring Year-round Late spring-autumn Late autumn-early spring
Site Gomolava Opovo Petnica Selevac
Red deer ? Year-round ? Autumn/winter?
Wild pigs Autumn/winter? ? ? ?
Reference 6.3.5.8 Russell 1993 6.4.4.8 Legge 1990
Table 5.5: suggested seasonal foci of hunting at Vina sites.
5.3.1.3 risk-buffering
Hunting in agro-pastoral societies potentially serves as a risk-buffering strategy (e.g. Halstead 1989, 72-73). Wild resources may become important when crops fail, or when populations of domestic animals are hit by disease. A background level of hunting, the argument goes,
110
allows for the maintenance of knowledge and skills that may occasionally become vital for subsistence. Such bad years must have occurred during the Vina period, and wild resources will surely have played a mitigating role. It would be a stretch to characterize hunting during the period in these terms, however, given the sheer quantity of wild remains recovered. Their ubiquity also militates against a primary risk-buffering role for hunting, which should result in occasional short-term increases in the proportion of wild species. Wild species are present in 99% of units at Gomolava with a NISP of ten or more. Coarse stratigraphic resolution may mask fluctuations but this is less likely at Petnica, where wild species appear in 98% of the generally much smaller excavation units, with the NISP threshold reduced to five. Figure 5.22 shows the range of variation within units at Gomolava and Petnica, using the same NISP thresholds as above. Given the small samples involved, inter-quartile ranges are remarkably tight. When ubiquity defined as the proportion of excavation units containing each taxon is plotted against NISP at either site (Figure 5.23), wild and domestic species fall on the same curves. No Vina assemblages have the requisite chronological resolution to compare taxonomic representation with longer-term climatic fluctuations (c.f. HsterPlogmann et al. 1999).
1.0
Percentage wild (NISP)
0.8
0.6
0.4
0.2
0.0 Gomolava Petnica
Figure 5.22: variation in proportions of wild taxa per unit at Gomolava and Petnica.
111
Gomolava
1 0.9 0.8 0.7 Ubiquity Ubiquity 0.6 0.5 0.4 0.3 0.2 0.1 0 0 10 20 %NISP 30 40 50 Pig Aurochs Roe deer Sheep / goat Dog Wild pig 0.25 0.2 0.15 0.1 0.05 0 0 Pig Wild pig Roe deer Sheep / goat Red deer 0.3 Cattle 0.4 0.35
Petnica
Red deer Cattle
Dog Aurochs 10 20 %NISP 30 40 50
Figure 5.23: ubiquity against unit %NISP for major species at Gomolava and Petnica.
5.3.1.4 raw materials

Hunting aimed at specific resources such as pelts seems likely a priori for certain species in Vina assemblages, and is supported by butchery marks on beaver and small carnivore mandibles at Opovo (wildcat: Russell 1993, 156-157) and Petnica. Carnivores and small game constitute a very small part of the recovered wild fauna. Red deer antler was also used widely, with tools common at Gomolava, Petnica, Opovo (Russell 1993, 468) and Selevac (Russell 1990). While this no doubt constituted one reason for hunting red deer, variable sex ratios indicate that it was not a primary motivation (Table 5.6). Male roe deer dominate at Opovo although their antler is much less commonly used (Russell 1993, 381).
Site Gomolava Petnica Opovo Divostin Vina Opovo Gomolava Petnica Opovo Source Various Various Various Unspecified Biometry Various Dentition Dentition Various Female 27 17 407 Male 28 9 501 Ratio (F/M) 0.96 1.89 0.81 ~0.33 0.46 0.67 0.6 0.73 Reference 6.3.5.7 6.4.4.7 Russell 1993, 357 Bknyi 1988, 427 Dimitrijevi forthcoming Russell 1993, 357 6.3.5.7 6.4.4.7 Russell 1993, 356
Red deer
"approximately equal" 90 10 6 97 194 15 10 132
Roe deer Wild pig
Table 5.6: sex ratios for wild species in Vina assemblages.
5.3.1.5 garden hunting

The three main wild species on Vina sites wild pig, red and roe deer are all potential crop raiders, raising the possibility that their killing followed Linaress (1976) garden hunting
112
model. Cultivated land may have attracted wild animals during the growth seasons, creating both a concentrated source of meat and a necessity to protect the crops. Wild animals doubtless threatened crops around Vina sites on occasion, and protective or opportunistic hunting on this basis is very likely to have taken place. However, the admittedly limited seasonal data show no sign of the expected spring/summer concentration, militating against this being the primary mode. Age-at-death data can also be brought to bear on the question of opportunistic versus directed hunting. Such evidence is only patchily recorded for wild species in the Vina/contemporaneous sample, with no unified relative age scheme, (Table 5.7). Extremely adult-dominated samples at Snandrei almost certainly include postcranial data, and should probably be discounted given poor preservation at the site (Jongsma & Greenfield 1996). Red deer ages at Opovo and Gomolava suggest a combination of opportunistic and selective hunting: sub(adults) are perhaps over-represented, and seniles rare, but a substantial number of young deer are also present. The proportion of adult specimens decreases over time at Gomolava in parallel with the overall number of deer, possibly representing a reduction in the frequency of targeted hunting. At Opovo percentage deer maturity rises overall, although it peaks in BH2 for red deer. This may indicate increasingly selective hunting practices alongside a more encounter-based mode perhaps related to crop protection. A bias towards male deer at most sites could reflect selective hunting with males preferred for antler, meat weight or prestige or alternatively garden hunting, since males are also more frequent crop raiders (Russell 1993, 407). A prime-dominated age-profile accompanied by a bias against males at Petnica suggests selection for females, perhaps related to hunting techniques. A tentative contrast between heavily selective hunting at Petnica and a more opportunistic or protective mode at Gomolava would certainly make sense given their respective environments and agricultural potentials, as would the position of Opovo somewhere in-between. Separate wild pig age data are not provided for Opovo, but at Gomolava the same shift towards immature specimens is noted as for red deer. Aurochs from Gomolava are chiefly juvenile or senile. There will be a slight bias against subadults here since dP4 and M3 are useful for identifying wild cattle, but the high senile:adult ratio indicates a focus on slower and/or weaker individuals.
113
Opovo BH3 Opovo BH2 Opovo BH1 Opovo total Selevac Snandrei (Banat) Snandrei (Vina C) Gomolava Ia Gomolava Iab Gomolava Ib Gomolava total Petnica Opovo BH3 Opovo BH2 Opovo BH1 Opovo total Snandrei (Banat) Snandrei (Vina C) Gomolava Petnica Snandrei (Banat) Snandrei (Vina C) Gomolava Ia Gomolava Iab Gomolava total Petnica Gomolava
Immature Mature Infantile Juvenile Subadult Adult Senile 28.9 23.7 0.0 47.4 10.0 20.0 70.0 0.0 22.2 14.8 1.9 61.1 23.0 20.4 56.0 0.5 13.8 13.8 17.2 41.4 13.8 0.0 0.0 40 60 0.0 0.0 21 79 15.1 24.1 14.9 45.9 37.0 22.2 0.0 40.7 28.6 57.1 0.0 14.3 24.7 26.7 7.7 40.8 0.0 17.4 30.4 52.2 19.0 7.1 8.3 13.1 15.5 14.3 8.3 13.1 25 10 47.1 33.3 18.0 17.0 29.8 16.1 27.9 35.0 0.0 62.1 67.9 75.0 67.2 75 90 38.2 66.7 76.0 67.0 43.5 24.6 34.3 0.0 18.2 6.7 6.3 5.4 10.0 54.5 3.5 10.7 8.3 6.6
n 97 40 54 191 29 ? ? 37 27 7 71 23 58 28 36 122 ? ? 34 15 ? ? 30 32 74 10 11
% Mature 47.4 70.0 61.1 56.0 72.4 100.0 100.0 60.8 40.7 14.3 48.6 82.6 62.1 67.9 75.0 67.2 100.0 100.0 85.3 100.0 94.0 84.0 80.0 46.9 67.6 45.0 72.7
Reference Russell 1993
Red deer
Legge 1990 Jongsma & Greenfield 1996 6.3.5.5
6.4.4.5 Russell 1993
Roe deer
Jongsma & Greenfield 1996 6.3.5.5 6.4.4.5 Jongsma & Greenfield 1996 6.3.5.5
Aurochs
0.0 0.0 0.0 0.0 0.0 0.0 0.0 10.0 0.0
14.7 0.0 6.0 17.0 20.0 53.1 32.4 45.0 27.3
Wild pig
6.4.4.5 6.3.5.5
Table 5.7: age-at-death data for hunted animals on Vina sites. Figures based on mandibles in all cases except Snandrei, where unspecified.
5.3.1.6 social roles

The roles of hunting outlined thus far err on the materialist side of interpretation. This is a factor of the scale and nature of the data assessed thus far, and no apology is made: animals are, amongst other things, material resources. To describe the contribution of hunting practices to subsistence is not to suggest that such concerns are the only, or even necessarily the main, factor determining their form, however. Human-animal relationships are important in the construction of identities, and conversely issues of identity must contribute to individuals decisions regarding their engagements with animals. The argument that sheep were central to Krs identity (Whittle 2005) has already been mentioned, and this need not have been epiphenomenal to their economic importance. Indeed, the relative unsuitablility of caprines on the Hungarian Plain (Bartosiewicz 2005, 51) suggests that their very real contribution to subsistence was conditioned by their centrality to Krs experience and identity as much as vice versa. Such arguments apply equally to hunting: red deer and wild pig in particular may have been an important aspect of identity among Vina communities.
114
This raises the question of participation. Nothing in the limited age- or season-at-death data suggests periodic large-scale potentially communal hunting, and the ubiquity of wild taxa also militates against this. The main hunted taxa are not in any case large herd animals under normal conditions, but are more suited to small-group hunting, although adult wild pig and especially aurochs presumably required hunting parties. A statistical association of dogs with wild taxa (5.2.6) may indicate their involvement. Hunting might have been a specialist activity, with certain individuals or households relying on it to a greater extent than others. This could simply be a matter of differential knowledge and skills or could reflect varying access to domestic resources. Given the probability of encounter hunting around crops, rigid specialization does not seem likely, but nor does equal involvement throughout the community. Involvement in and/or skill at hunting may have been an important resource for the construction of identity within communities, possibly drawing on symbolic associations of hunting or the hunted taxa. An exhaustive list of possible associations is impossible since symbolic systems are by their nature highly contingent, but two themes are worth mentioning. Firstly, hunting plays a widespread part in the negotiation of gender roles, specifically legitimation of the male position in society (Kent 1989). Aspects of such negotiation are hinted at by real and imitation deer canines in burials at the Herply site of PolgrCsszhalom (Bartosiewicz 2005, 58; Choyke 2001). In the absence of faunal mortuary data from Vina sites I will not speculate on the gender issue besides noting its probable importance. Secondly, Hamilakis (2003, 244) argues that hunting in Late Neolithic and Bronze Age Greece derived ideological importance from its association with the other, making it a valuable resource for the generation and legitimation of authority by an incipient elite. Hunting is ascribed an importance which is not only independent of its potential subsistence value but which derives from its very position outside both the routine and geography of the domestic economy. This location of the wild conceptually and/or geographically apart from the domestic may be valid in the Greek case, but cannot be supported with regard to the ubiquitous and probably frequently opportunistic hunting of the Vina period.
5.3.1.7 hunting: summary

Hunting was a ubiquitous aspect of life in the Vina period, not restricted to specialized sites, specific seasons, or lean agricultural periods. It may often have taken place on an opportunistic and/or protective basis, but was probably a planned, targeted activity in at least
115
some times and places. Of the possible motivations for hunting outlined above, most were probably important at certain times and in certain places, while few are mutually exclusive.
5.3.2 roles of domestic species

Debate regarding animal husbandry in the Neolithic has revolved around the use or otherwise of secondary products. This has a relevance well beyond traditional questions of economic prehistory since the extraction of secondary products, where practised, is central to direct human relations with domestic animals (3.4.2). Regular contact through milking or yoking for traction promotes the establishment of intimacy between people and domestic animals that is to say the formation of intersubjective relationships and the recognition of individual characteristics. The existence and form of such contact are fundamental to human experiences and perceptions of animals. Since intimacy is one strand of the understanding of domestication advocated in 3.2.2.5, animals providing secondary products can in a sense be considered more domestic that is, tied more closely to humans and human society. Secondary products also have a bearing on the second aspect of domestication, the ability of animals to act as vehicles of relationships between humans. Their use expands the range of possible transactions involving animals, increasing the variety of products that can be shared, exchanged or gifted, and also introducing the possibility of loans for labour (Bogucki 1993) as well as for breeding. The involvement of animals in the human political economy is thus potentially intensified, and the overlap between human and animal sociality deepened. Animals become living assets with a value increasingly independent of their eventual slaughter at the same time that intimacy in interspecific relations is promoted. The tension inherent in their status as sentient property both subject and object is thus heightened. The extraction of secondary products is therefore much more than a matter of subsistence base. Sherratts original (1981) model of the Secondary Products Revolution (SPR) saw it as a unified post-Neolithic shift from hoe cultivation to plough agriculture and transhumant pastoralism. Working primarily from settlement patterns, Chapman (1982; 1988; 1990) argues for a less coherent package, placing the adoption of the ard in the Central Balkans earlier, during the earlier part of the Vina period. Greenfield upholds Sherratts view in a series of studies of faunal age profiles from Central Balkan Neolithic-Bronze Age sites (Arnold & Greenfield 2004; 2006; Greenfield 1988; 1989; 1991; 2005). His data are shown in Figure 5.24, supplemented by Legges (Selevac), Russells (Opovo) and my own (Gomolava). These are taken to show post-Neolithic changes in exploitation of cattle and caprines but not of pig, which serves as a control species with no
116
practical secondary products (Greenfield 2005, 23-28). Caprines are the most convincing: most post-Neolithic curves lie to the right of the Neolithic group, suggesting a move towards wool production following the standard models. Eneolithic Petnica, meanwhile, lies close to the ideal milk curve. Less distinct diversification is seen for cattle. One should note that the presence of woolly sheep in Neolithic Europe is unlikely (Ryder 1982), although caprine hair may have been used for some purposes. Frequent loom weights and spindle whorls do not necessarily indicate wool use, as implied by Bailey (2000, 183), since spun and woven flax is known to have been used in the Vina period (Borojevi 2006, 12). Greenfield concludes a shift between the Neolithic and Eneolithic from primary products (i.e. meat) to a balance of primary and secondary. This interpretation is problematic. The inclusion of loose teeth is questionable, but the unacceptably small size of many samples is critical. If those having n < 25 (dotted in Figure 5.24) are omitted, patterns change dramatically: a) Neolithic post-Neolithic differences for caprines are greatly reduced. b) The only remaining post-Neolithic cattle sample falls squarely in the Neolithic range. c) By far the clearest change is seen for pigs, supposedly unaffected by the SPR. A large part of the apparent diversification is thus probably caused by reduction in average sample size, although there may be a genuine increase in typical slaughter ages for caprines. The pig result must be taken as a refutation either (a) of the methodology, or (b) of the premise that slaughter strategies only reflect the extraction of tangible products. I take the latter view, calling the whole enterprise into question. The definition of secondary products should not be limited to the material resources of milk, wool and labour. The use of any of these products increases the value of the living animal, but even without secondary products sensu stricto animals frequently serve as units of wealth (Ingold 1980; Russell 1998), prestige items (Reid 1996), or stores of food (OShea 1989, 5859), and the potential importance of livestock for identity has already been mentioned with regard to Krs sheepherding. The social value of live animals might have created the conditions for the adoption and refinement of dairying and traction as much as vice-versa. Although clearly a simplification, there is thus much merit to the suggestion that domestication represents a change in focus from the dead to the living animal (Meadow 1984). The use of wool, traction and/or milk represents a continuation in this direction rather than a qualitative change.
117
a)
100 % survival to start of stage
EN Fo eni-Sala (n = 1 7) EN B lago tin (n = 78) M N Stragari (n = 29)
80
M N P etnica (n = 22) LN Vina (n = 1 4) LN Opo vo (n = 1 4) LN P etnica (n = 1 9) Eneo B lago tin (n = 1 2)
60
40
B A Ljuljaci (n = 1 2) B A Livade (n = 1 6)
20
B A Vina (n = 1 6) IA Kadica B rdo (n = 47)
0 A B C D E F G H I Payne stage (cattle)
LN Selevac (n = 78; Legge 1 990) LN Go mo lava (n = 1 6.3.5.5) 94;
b)
100 % survival to start of stage
EN Fo eni-Sala (n = 61 ) EN B lago tin (n = 1 37) M N Stragari (n = 26) LN Vina (n = 34)
80
LN P etnica (n = 1 9) Eneo P etnica (n = 1 3)
60
Eneo No vaka uprija (n = 1 0) B A Ljuljaci (n = 1 3)
40
B A Livade (n = 1 ) 1 B A Vina (n = 23)
20
B A P etnica (n = 22) B A No vaka uprija (n = 48) IA Kadica B rdo (n = 263)
0 A B C D E F G H I Payne stage (caprines)
LN Opo vo (n = 1 Russell 1 03; 993, 31 4) M /LN Selevac (n = 75; Legge 1 990)
c)
M N P etnica (n = 1 3)
100
LN Opo vo (n = 1 ) 1
% survival to start of stage
80
LN Vina (n = 33) LN P etnica (n = 1 3) Eneo P etnica (n = 1 0) B A Ljuljaci (n = 31 )
60
40
B A Livade (n = 27) B A Vina (n = 20)
20
B A P etnica (n = 1 0) B A No vaka uprija (n = 1 0)
0 A B C D E F G H I Payne stage (pigs)
IA Kadica B rdo (n = 43) LN Go mo lava (n = 229; 6.3.5.5)
Figure 5.24: harvest profiles for (a) cattle, (b) caprines, and (c) pigs at central Balkan sites. Dotted lines denote samples below 25. Data from Greenfield 2005 unless specified otherwise.
118
Ironically, the same arguments which stress the importance of secondary products utilization undermine the basis of its detection. If the value of the living animal goes beyond the provision of milk, wool, labour and offspring then the idealized demographic models set out by Payne (1973) cannot be relied upon (Craig 2002, 99-100). Apart from the impact of taphonomic processes (e.g. Munson & Garniewicz 2003), additional disincentives to slaughter disrupt the underlying logic. Variation in herd growth rate also affects kill-off structure, while Paynes models assume static populations (Hesse 1988). The problem is amplified if herding decisions are made on a level below the settlement: even when overall numbers of animals are stable, fluctuations in individual herds will affect slaughter patterns. In fact, herd size alone is a factor: management must diverge from the optimum as numbers decrease, since fractions of livestock cannot be slaughtered. Herd security is also an issue (Dahl & Hjort 1976; Greenfield 1988, 576; Russell 1993, 397-398). Cooperation and coordination between individuals or households may mitigate these effects but their extent cannot be assessed. The upshot of these complications is that we can comment on shifts in age profiles but cannot draw detailed conclusions from their precise form. Larger numbers of adults versus subadults may imply greater value was placed on the living animal. Age structures cannot reveal the specific reasons for this, although coupled with sex data they may provide clues. Variable slaughter patterns also disjoin death assemblages from life assemblages. Since herd sizes have practical limits, any factor increasing the value of adult animals and their presence in Neolithic lives may paradoxically decrease the osteological representation of the species in question (Hesse 1988). This is one possible explanation for the apparent increase in wild species in some areas during the Vina period. A post-Neolithic intensification along the lines of the SPR would not in any case mean that Neolithic livestock were valued purely as future carcasses, nor indeed that tangible secondary products were not known to and sometimes used by Neolithic people (Isaakidou 2006). There is even a tentative indication that average caprine slaughter ages increased between earlier and later Neolithic samples in the Central Balkans (Figure 5.24b). Of the multi-phase sites, only Opovo and Gomolava have respectable mandible samples from individual layers. Remarkable uniformity is seen in kill-offs for cattle and caprines at the former (Figure 5.25), and for cattle and pigs at the latter (see 6.3.5.5). Sex ratios for cattle are remarkably similar at Opovo and Gomolava, with significantly lower figures at Petnica, Obre II, and apparently Divostin (Table 5.8). Samples from individual phases are too small too be reliable at Gomolava but appear very stable at Opovo. Unsurprisingly, no sites show the extreme dominance of females predicted by intensive
119
dairying, but the 1:2.4 ratio at the lowland sites is consistent with a fairly generalist strategy. More even ratios elsewhere indicate that adult male cattle were valued, whether for traction supporting Chapmans argument or for social reasons. Since the sites are broadly contemporaneous but divide along topographical lines with Selevac intermediate an environmental explanation cannot be excluded. However, the heavy chernozems around the two Vojvodinian sites (Borojevi 1988) would benefit from ploughing as much as the smonicas of Petnica (Greenfield & Je n.d.) or Divostin (Barker 1975), and more than the brown forest soils of Selevac (Chapman 1990). If genuine, differences in cattle herds are probably underlain by social factors. Pig sex ratios are very similar at Gomolava and Petnica, and in line with expectations for efficient meat production.
100
BH3 (n = 33) BH2 (n = 36) BH1 (n = 29)
% survival to start of stage
80
60
40
20
0 A B C D E F G H I
Payne stage (caprines)
Figure 5.25: caprine harvest profile at Opovo by phase. Data from Russell 1993, 313-314.
Cattle Male 57 48 46 151 1 6 0 7 6 11 Domestic pig Male
Opovo
Gomolava
BH3 BH2 BH1 Total Ia Iab Ib Total
Petnica Divostin Obre II Selevac
Female 121 124 123 368 5 7 3 17 8 17 -
F/M 2.12 2.58 2.67 2.44 5.00 1.17 2.43 1.33 ~1 1.55 ~2
Female
F/M
Reference Russell 1993
11 18 3 36 7
6 2 0 10 2
1.83 9.00 3.60 3.50
6.3.5.7
6.4.4.7 Bknyi 1988 Bknyi 1974b Legge 1990
Table 5.8: sex ratios for domestic species on Vina/contemporary sites. Gomolava and Petnica data based on pelvises (cattle) and in situ canines (pigs). Opovo data includes metric and non-metric traits. Sources not specified for other sites.
120
In summary, zooarchaeological arguments regarding secondary products are called into question if one recognises the importance of intangible products in husbandry decisions: age profiles may hint at changes in the value attached to live animals but cannot reveal its basis. Age data thus have limited potential for testing the classic SPR model. In the present case, there is some sign that caprines were typically kept alive longer in the Vina period than the Early Neolithic, while sex ratios provide tentative evidence that adult male cattle were highly valued. Whether this reflects social factors or their use in traction supporting Chapmans early chronology for secondary products is a moot point.
5.3.3 temporal trends

Although various trends in taxonomic abundances were documented in 5.2, no clear overall shift was seen in the relative proportions of wild and domestic taxa. Amongst the latter category a shift from caprines to larger livestock is probably genuine, but confounded by geographical coverage. The only really clear patterns are seen where faunal data are available from multiple phases at the same site. At Gomolava, Selevac and Snandrei the contribution of wild taxa decreases as that of cattle amongst the domesticates increases. At Liubcova the opposite change occurs, while Opovo and Petnica show no clear trend in either case. A dramatic increase in cattle at Para accompanies a modest overall rise in wild taxa. Changes of this kind can potentially be explained in at least three ways. Firstly, Legge (1990, 221-222) explains the decline in wild species at Selevac in terms of increasing human disruption of habitats in the vicinity of the site. This is plausible, although anthropogenic impact on the Neolithic landscape of southeastern Europe is now believed to have been much more subtle and complex than a simple forest-clearance model would imply (Gardner 1999), and the pollen record shows no detectable deforestation around Gomolava (van Zeist 2002, 112), where the same trend is seen. In fact, Chapmans (see also 1989; Chapman) model of change from gardens within the settlement to plots nearby would surely increase potential for garden hunting, provided that plots remained relatively small and dispersed. Depletion of the local wild fauna is perhaps more likely, although only Snandrei shows the relative switch to smaller game expected in such circumstances (e.g. Winterhalder 1981, although see Speth & Scott 1989). The decreasing age of wild animals at Gomolava and very tentatively Snandrei could be taken as a sign of depletion, but more plausibly suggests that a reduction in active, selective hunting underlies the trend, i.e. that it is driven by demand rather than supply.
121
An alternative possibility in the realm of settlement histories relates to Tringhams (1992) proposal that Vina sites were formed by one or more households budding-off from established settlements. Being at an early stage in their household cycles, these groups would presumably have had limited numbers of domestic animals. Assuming an impetus to increase herd sizes as much for social ends as for meat supply one might expect an initial period of intensive growth and hence limited slaughter. Whether this increased incentives to hunt or simply reduced the amount of meat consumed altogether, the effect would be to raise the relative representation of wild specimens in assemblages. As settlements grew and herds approached the limits of available labour and/or grazing land, more animals would have become available for slaughter. This could also explain the trend towards cattle, through the principle that herders should trade up when herds are increasing (Mace & Houston 1989; cited for Opovo in Russell 1999). No support for this possibility exists in the domestic age profiles. The pattern of declining wild taxa and increasing cattle applies to some of the larger and longer-lived settlements within the sample, and both lines of argument outlined above assume continuous settlement. Relative lack of change at Petnica and Opovo should probably be seen in the context of their small size, limited agricultural potential and (especially for Petnica) possibly discontinuous occupation. It is noteworthy that these are the two most wilddominated Vina assemblages yet studied. Prime-dominated age profiles at Petnica have already been noted, and are increasingly matched by those at Opovo, which was also particularly reliant on wild plants (Borojevi 1988; 1998; 2006). Finally, changes may be linked to the general intensification of production and interhousehold competition during the period (2.3.3). Since domestic animals are likely to have acted as units of wealth, vital for social transactions as well as for prestige (Russell 1998), such intensification presumably increased the value of owning livestock. If status negotiation involved communal consumption, however, then the demand for deadstock will also have risen, leading to rapid herd turn-over and hence higher representation of domestic species. Potentially irrational selection of valuable animals for slaughter also makes age profiles hard to interpret (Russell 1999, 154-155). The specific increase in cattle remains suggests a particularly important social role. This explanation is accepted here as a working hypothesis, and has two interesting implications. Firstly, occupation at sites such as Gomolava and Selevac need not have been continuous. Secondly, the lack of observed changes at sites such as Petnica and Gomolava implies that different social conditions obtained, i.e. that the proposed intensification did not
122
apply equally across Vina sites. It is worth noting that both Kaiser and Voytek (1983) and Chapman (2000a, 42) tie increased social tensions to sedentism. Faunal data from earlier (pre-Vina D) phases at Vina-Belo Brdo would be extremely useful for testing this hypothesis.
5.4 Summary
Several key points can be extracted from the preceding discussion. Firstly, wild animals are a ubiquitous component of Central Balkan Neolithic faunal assemblages. There is no single trajectory to their representation, although a slight overall increase is seen in the later Neolithic, with a very minor drop in the final period. Principal Components Analysis
suggests that wild and domestic species can be seen as two alternative packages of taxa, forming an axis along which assemblages are spread. Aurochs, however, fall within the domestic group, and dogs within the wild. Red deer, roe deer and wild pig are the main wild species and were hunted in different circumstances at different sites, with both selective and garden hunting probably playing a part. Secondly, the Vina period sees a shift from sheep to pigs and especially cattle amongst the domestic species. There is only weak evidence for secondary products exploitation during the Neolithic, but this does not preclude the limited use of animals for purposes other than meat, as suggested by cattle sex profiles at some sites. Changes in overall figures are partly driven by marked trends at some multi-phase sites, and a tentative distinction emerges between (a) large, long-lived, increasingly domesticatedominated settlements (Selevac, Gomolava, Snandrei), and (b) smaller sites with a consistent focus on hunting (Opovo, Petnica), although not all sites conform to these categories. Changes in faunal make-up in the first group are taken as expressions of social changes originally proposed by Kaiser and Voytek (1983) and since developed by others (Russell 1993; Tringham 1992; Tringham & Krsti 1990b), most radically Chapman (1996; 2000a). The following chapters examine the faunal assemblages from one site in each group: Gomolava and Petnica.
123
Biometry
Chapter 6 Results from Gomolava and Petnica

6.1 Introduction
This chapter presents basic data from the two sites studied at first hand by the author, Gomolava and Petnica. Some of these results have already featured in Chapter 5, and they will form the background for the detailed taphonomic study in Chapter 7. Biometrical analysis of the material is presented in 6.2, then the sites and their fauna are described individually in some detail (6.3 and 6.4).
6.2 Biometry
Since biometry plays an important role in the identification of wild and domestic specimens, it must necessarily precede any other analyses. As noted above (4.3.2), measurements are usually best interpreted by superimposing data from similar sites in order to increase effective sample sizes. Accordingly, measurements from Gomolava and Petnica are plotted on the same graphs, together with data from other sites. The same approach is taken with regard to sex attribution, where applicable. Three species may potentially be represented by wild and domestic forms on Balkan Neolithic sites pigs, cattle and dogs. In these cases biometric data is vital both for assessing the relative proportions of each group and for identifying individual specimens. Measurable fragments thus identified may then be used as a guide for specimens lacking preserved measurement points. Bivariate plots were constructed for each element with a reasonable number of paired measurements from Gomolava, Petnica and other relevant sites. Histograms and kernel density plots are also employed where there are many unpaired measurements or the scatter plots are unclear. Only the most numerous and/or useful measurements for each species are presented here; data for all measured elements are provided in Appendix 3. Since many measurements are too scarce to be informative on an element-by-element basis, this approach is supplemented by the log-ratio method (4.3.2), combining most common measurements by reference to a standard animal. In rare cases of disagreement between element-by-element and log-ratio results the former is given preference.
124
Biometry
6.2.1 Sus scrofa 6.2.1.1 postcranial measurements

Previous studies of fauna from Vina sites have noted that wild pigs are often abundant and that a surprisingly clear metrical separation exists between these and the domestic form (Dimitrijevi forthcoming; Russell 1993, 139). The data from Gomolava and Petnica largely corroborate this impression. Clear size clusters are observed for most common elements and more-or-less match those previously identified as wild and domestic groups on comparable sites including Obre II, Divostin, Gomolava Block I and Vina (respectively, Bknyi 1974b; Bknyi 1988; Clason 1979b; Dimitrijevi forthcoming). Figure 6.1 and Figure 6.2 show measurements on the distal humerus and distal tibia as exemplars of this clustering. The scapula, ulna, distal humerus and proximal radius all show the same pattern, with most specimens from Gomolava and Petnica falling into the larger, presumably wild, cluster. In general, separation seems clearer for individual sites than for the combined data. While this must in part relate to sample size, it also suggests some size variation between sites, whether due to local ecological factors, contingent differences in the genepool or simply varying sex ratios. For some measurements, including the distal tibia, more inter-site variation is apparent amongst the smaller, presumed domestic, group than the larger, suggesting that different patterns of management are partially responsible. The presumed wild group, on the other hand, often shows more variation overall. Not all elements show such clear clusters. Measurements on the distal radius, metapodials, astragalus (Figure 6.3) and calcaneus at Gomolava and Petnica all form single groups, albeit with occasional outliers. For the latter two elements smaller groups are observed in the data from comparable sites, and in all cases log-ratio results confirm that the single groups represent wild individuals. A similar absence of clustering amongst the phalanges could not be dealt with in this way since phalanx measurements are not known for the standard animal. Figure 6.4 shows the distribution of pelvis LAR measurements. Two groups emerge again but the larger appears split in two. The strongly sexually dimorphic nature of pigs and of the pelvis presents a possible explanation: the smaller sub-cluster might relate to domestic males or more likely wild females, or indeed to both.
125
Biometry
65
60
55
50 Dd (mm)
45
40
35
30
Obre II Vina Opovo Gomolava Petnica Divostin

25 30 35 40 45 Bd (mm) 50 55 60 65
25
Figure 6.1: Sus distal humerus measurements from Gomolava, Petnica and other Vina or related sites. (Obre II - Bknyi 1974b; Divostin - Bknyi 1988; Gomolava - Clason 1979b; Vina - Dimitrijevi forthcoming; Opovo - Greenfield 1986). Bd measurement follows von den Driesch (1976).
40
35
30 Dd (mm) 25
20
Obre II Vina Opovo Gomolava Petnica Selevac Anza IV Divostin

20 25 30 Bd (mm) 35 40 45
15
Figure 6.2: Sus distal tibia measurements for Gomolava, Petnica and other Vina or related sites. (Anza IV - Bknyi 1976; Selevac - Legge 1990, other sources as previous figures)
126
Biometry
a)
40 35 30 25
Bd (mm)
20 15 10 5 0 -5 -5 15 35 GLl (mm) 55 75 Obre II Vina Gomolava (A.C.) Gomolava Petnica Selevac Divostin
b)
45 40 35 30 25
Dm (mm)
20 15 10 5 0 -5 -5 5 15 25 35 45 55 65 GLm (m m )
Gomolava (A.C.) Gomolava Petnica
Figure 6.3: Sus astragalus measurements from Gomolava, Petnica and other Vina and related sites. Sources as previous figures.
127
Biometry
5 4 3 2 1 5 4 3 2 1 5 4 3 2 1
Frequency
All
Frequency
Gom olava
Frequency
Pe tnica
10
20 30 LAR (mm)
40
50
Figure 6.4: distribution of Sus LAR measurements from Gomolava, Petnica, Opovo and Vina. Sources as previous figures.
6.2.1.2 cranial measurements

Clustering apparent in the post-cranial data is assumed here to represent wild and domestic pig populations, following determinations by previous researchers. Since pigs are highly sexually dimorphic, however, one cannot rule out the possibility that either sex or a combination of sex and domestic status is responsible for the groupings. Cranial measurements are helpful since pig mandibles and maxillae can often be sexed on the basis of their canines, or the alveoli thereof. Payne and Bull (1988, 30) also report much lower dimorphism for tooth measurements than for the postcranial skeleton. Comparative data are largely restricted to third molars (e.g. Obre, Divostin, Anza), which is unfortunate since (a) sex is rarely known for the distalmost teeth, and (b) they can only reveal the size distribution of mature specimens. Since the latter problem is shared with most postcranial elements, the remainder of the dentition provides a valuable opportunity to assess wild:domestic ratios amongst younger individuals. For the lower dentition, a very large sample of measurements collected by the author from Blocks I to VI at Gomolava (see 6.3.5) makes up for the lack of comparative data. Measurements from this sample are restricted to dP4 and M3, plus breadth from M1 and M2 and various mandibular measurements, while all possible measurements were taken from the main Block VII sample. Definite groupings emerge for each tooth (Figure 6.5), although for M1 and M2 these are only clear from breadths (Figure 6.6). The relative number of wild specimens increases from mesial to distal, suggesting that domestic individuals were typically killed at an earlier age than their wild counterparts. The presence of two size groups, with the
128
Biometry
larger being more common amongst older individuals, is corroborated by measurements on upper teeth and the mandibles themselves (Appendix 3.1.9; 3.1.10).
a) Sus lower dP4 measurements 12 G: male G: female G: indet. P: female P: indet. G (A. Clason)
Breadth (mm)
10
6 14 16 18 Length (mm) 20 22 24
b) Sus lower M1 measurements

16 15 14 13 Breadth (mm) 12 11 10 9 8 7 6 8 10 12 14 16 Length (m m ) 18 20 22 24
G: male G: female G: indet. P: female P: indet. G (A.Clason)
129
Biometry
c) Sus lower M2 measurements

24
G: male
22 20 Breadth (mm) 18 16 14 12 10 8 16 18 20 22 Length (mm ) 24 26 28
G: female G: indet. P: female P: indet. G (A.Clason)
d) Sus lower M3 measurements 25 G: male G: female G: indet. Breadth (mm) 20 P: indet. G (A.Clason)
15
10 20 25 30 35 Length (mm) 40 45 50
Figure 6.5: Sus lower tooth measurements from Gomolava and Petnica.
6.2.1.3 the log-ratio method

Measurable specimens that could not be attributed wild or domestic status on an element-byelement basis due to rarity or lack of clustering were investigated using the log-ratio method, (4.4.2). The standard was a mature wild female from the Hungarian Agricultural Museum (Russell 1993, 140). Following Uerpmanns (1979) recommendation, cranial elements were not included due to the tendency for greater size change in the postcranial skeleton. Separate
130
Biometry
log-ratio calculations were instead performed for the upper and lower jaws using two mandibles and two maxillae from Gomolava, one adult and one juvenile in each case.
22
20
G: male G: female G: indet. P: female P: indet. G (A.Clason)
M2 Breadth (mm)
18
16
14
12
10
10
11
12
13
14
15
M1 Breadth (mm)
Figure 6.6: breadth of Sus lower M1 versus M2.
Figure 6.7: histogram and kernel density plot for mean Sus postcranial log-ratio values by specimen at Gomolava and Petnica
Figure 6.7 shows the distribution of postcranial log-ratio values at Gomolava and Petnica. The distribution appears to be composed of two superimposed broadly normal distributions, the smaller with its mean at approximately -0.35, and the larger centred around +1.5. This
131
Biometry
corroborates most individual element measurements, suggesting a large sample of wild specimens and a smaller number of domestic ones, with some overlap between them. Plotting separate histograms for specimens identified as wild, domestic and indeterminate on the basis of individual element scatter graphs (Figure 6.8) reveals strong agreement between the two methods: there is little overlap in mean log-ratio values between the wild and domestic groups. Separation between the groups is less clear in the log-ratio data than for many individual elements, but this is to be expected since wild-domestic size differences are not expressed equally in all measurements: shape as well as size changes following domestication. The lowest point between the two peaks falls in the range -0.15 to -0.10 at Gomolava and -0.20 to -0.15 at Petnica, in both cases more-or-less consistent with the -0.17 reported from Opovo (Russell 1993, 139).
Frequency
40 30 20 10
Domestic
Frequency
40 30 20 10
Indeterminate
Frequency
40 30 20 10
Wild
- 1.0
- 0.5
0.0
0.5
Mean log -ratio value
Figure 6.8: mean Sus postcranial log-ratio values from Gomolava and Petnica by previouslyassigned domestication status.
For individual elements the domestic group was typically less dispersed than the wild, as expected given the reduced sexual dimorphism characteristic of domestication populations (Zohary et al. 1998, 133). For the log-ratio data the reverse is true. This may reflect morphological differences between the wild and domestic populations: since the standard animal is wild it is likely to be closer in relative proportions to its Neolithic counterparts than to the domestic specimens. Differences between skeletal portions in the degree of size reduction following domestication may have inflated the variability in log-ratio values seen within the domestic group.
132
Biometry
The indeterminate group in Figure 6.8 overlaps much more with the wild than with the domestic distribution, primarily due to various elements with measurements forming a single group: the atlas, axis, metapodials, astragalus and calcaneus. From the log-ratio values, these specimens almost all appear wild (Appendix 3.1.11). Cranial log-ratio values also show bimodality, but as with the individual measurements this is less clear than for post-cranial elements. When the values are presented by sex (Figure 6.9) males are typically larger than females, at least for the lower teeth, but both large females and small males are also present. At face value the female and male peaks, lining up as they do with the modes of the overall distribution, suggest that these modes reflect sexual dimorphism rather than domestication status. However, since the range of sizes for each sex is almost identical, sexual dimorphism is unlikely to be the major factor, and the peaks must relate more to wild and domestic populations. The lower peak for females thus presumably represents their dominance amongst the (sub-)adult domesticates a likely pattern in most husbandry regimes although the sample of sexed individuals is too small for reliable conclusions.
Lower Female Upper Female
12
Frequency
Lower M ale
Upper Male
12
Frequency
Lower Indet erminat e
Upper Indet erm inate
12
Frequency
0 -0.5
0.0
0.5 -0.5
0.0
0.5
Mean log-ratio value
Figure 6.9: log ratio values for Sus lower and upper permanent teeth at Gomolava and Petnica by sex.
133
Biometry
6.2.1.4 Sus scrofa conclusions

The data presented here demonstrates a bimodal size distribution for pigs at Gomolava and Petnica, with the two modes so far assumed to represent wild and domestic populations. This assumption must be justified, especially given the sexual dimorphism observed amongst pigs. Given the clear clustering seen in most common measurements, only four scenarios are possible: 1. Groups represent wild and domestic, with sexual dimorphism a minor factor. 2. Groups represent sexes, with dimorphism dwarfing wild/domestic differences. 3. The lower group represents all domestic individuals plus wild females. 4. The upper group represents all wild individuals plus domestic males. The presence of both males and females in each group precludes options 3, 4, and especially 2. In addition, option 3 would imply an improbable number of wild males, while 4 would require higher dimorphism in the domestic than the wild population. Option 1 is thus the only plausible explanation.
6.2.2 Bos taurus / Bos primigenius

Aurochsen are typically present in Vina assemblages in much smaller numbers than wild pigs, presenting rather different identification challenges. Since there also appears to be a greater size overlap between wild and domestic cattle, it is rare to see distinct groups in the bivariate plots. Identifying wild individuals is thus a matter of noting any especially large specimens lying outside of the main spread. In addition, specimens lying in the upper tail of size distributions are treated as Bos sp. since they may represent smaller wild females rather than large domestic males. The line between domestic and indeterminate is inevitably arbitrary to an extent. There also appears to be greater inter-site variation in size for cattle than for pigs, suggesting that the former are more sensitive to differing ecological conditions and/or husbandry regimes. This reduces the utility of overlaying data from different sites. Given the overlap in size between the wild and domestic populations of cattle and the degree of sexual dimorphism, amongst aurochsen especially, metrical analysis must consider sex alongside domestication status. Russell (1993, 127-135) typically divides measurements from Opovo into at least three groups usually wild, domestic male and domestic female on
134
Biometry
the basis of visual clustering and previous authors determinations. Such detailed interpretation of the metrics should be approached with caution and is not attempted here since the relevant clusters do not emerge from large samples at individual sites, while intersite variation limits recourse to published authorities. The possible presence of castrates complicates matters still further in that one cannot even be sure how many populations (in the statistical sense) contribute to the overall distribution. Only the more informative plots for each element are included here, with the remainder in Appendix 3.2.
6.2.2.1 postcranial measurements

The majority of elements show a single broad spread of measurements with a few upper (and occasionally lower) outliers. This is exemplified by the distal tibia and calcaneus (Figure 6.10 and Figure 6.11), but a similar pattern is seen for the scapula, radius, astragalus, navicularcuboid, and second and third phalanges. In each case the main cluster is assumed to consist overwhelmingly of domestic individuals, while extreme upper outliers are taken to be aurochsen. Specimens in the upper tail of the main spread or lying slightly above it are identified as Bos sp.. Bknyis determinations from various sites are sometimes used as a guide to separate indeterminate from probable domestic individuals, but specimens are never recorded as definitively wild unless they clearly stand out within the data from their own site. Clearer clusters are seen amongst ulna and humerus measurements, although this may be attributable to small sample sizes. Full results and discussion for these elements are provided in Appendix 3.2.1-3.2.8. Bos pelvises can often be sexed according to non-metric criteria on the pubis and/or the medial wall of the acetabulum. Since the most common measurements are also on the acetabulum there is considerable scope for sex data to inform metrical analysis. Figure 6.12 plots these measurements for Gomolava and Petnica, showing sex where determinate. The upper cluster of three specimens is interpreted as wild, while the two females towards the top of the main range raise the possibility that some of this cluster may be small wild females. These two, and the slightly larger indeterminate specimen from Gomolava, are treated as Bos sp. while the rest of the main cluster is assumed to be domestic. As with pigs, the presence of both sexes in each cluster means that sexual dimorphism cannot be the main factor underlying the observed pattern. Metapodial measurements are considered in detail here since they are (a) numerous and (b) often used to investigate both sex and domestication status. Worryingly, Clasons proximal measurements for both sets of metapodials fall mostly between the two clusters observed in the authors own data, while her distal measurements largely fall below the range of those
135
Biometry
presented here (Figure 6.13 and Figure 6.14). This must be a matter of measurement protocols although it is surprising that both breadth and depth are affected and her measurements are thus not treated as comparable with my own.
70
65
60
55
Dd (mm)
50
45
40
35
Gomolava Petnica Vina Selevac Gomolava (A.C.) Obre II - wild Obre II - domestic Divostin Anza IV
45 50 55 60 65 70 75 80 85 90
30
Bd (mm)
Figure 6.10: Bos distal tibia measurements from Gomolava, Petnica and other Vina and related sites. Sources as previous figures.
70
65
60
GB (mm)
55
50
45
40
Gomolava Petnica Vina Gomolava (A.C.) Obre II - wild Obre II - domestic Divostin Anza IV
125 135 145 155 165 175 185
35 115
GL (mm)
Figure 6.11: Bos calcaneus measurements from Gomolava, Petnica and other Vina and related sites. Sources as previous figures.
136
Biometry
All four measurement pairs show two overlapping clusters with some upper outliers, especially when sites are taken alone. At Selevac, the central cluster of distal metacarpals was taken to represent a combination of wild females and domestic males (Legge 1990, 224-226). Following Degerbls and Fredskilds analysis of Danish cattle and aurochsen (1970, 104), Legge distinguishes these groups on the basis of shape, with wild females relatively deeper than domestic males. Given the lack of a clear shape distinction within this cluster at any of the Balkan sites, however, such extrapolation from Danish data cannot be supported. While the presence of some small wild females in the central group is likely, it is nonetheless assumed primarily to represent domestic males. Most of the other elements including the proximal metatarsals - show a single main spread along with very few outliers, some of which presumably wild males are very substantially larger, while others perhaps female aurochsen are closer to the presumed domestics. It would be surprising if the division of the main spread into two amongst the sex-sensitive distal metapodials represented an unusually large number of small wild females rather than simply greater sex separation within the domestic population.
85 75 65 55
LAR (mm)
45 35 25 15 5 -5 -5 5 15 25 35 45 55 65 75 85 95
G: male G: female G: indet. P: male P: female G: (A. Clason) Vina
LA (mm)
Figure 6.12: Bos pelvis measurements from Gomolava, Petnica and Vina, showing sex where attributed. Sources as previous figures.
137
Biometry
a) proximal metacarpal measurements

55
50
45
Gomolava Petnica Vina Selevac Gomolava (A.C.) Obre II - wild Obre II - domestic Divostin
Dp (mm)
40
35
30
25 45 50 55 60 65 70 75 80
Bp (mm)
b) distal metacarpal measurements

50
45
40
Dd (mm)
35
30
25
20 45 50 55 60 65 70 75 80 85 90
Bd (mm)
Figure 6.13: Bos metacarpal measurements from Gomolava, Petnica and other Vina and related sites. Sources as previous figures.
138
Biometry
a) proximal metatarsal measurements

70
65
60
Gomolava Petnica Vina Selevac Gomolava (A.C.) Obre II - wild Obre II - domestic Divostin
Dp (mm)
55
50
45
40
35 35 40 45 50 55 60 65 70 75 80
Bp (mm)
b) distal metatarsal measurements

45 43 41 39 37
Dd (mm)
35 33 31 29 27 25 45 50 55 60 65
70 75 80
Bd (mm)
Figure 6.14: Bos metatarsal measurements from Gomolava, Petnica and other Vina and related sites. Sources as previous figures.
139
Biometry
Figure 6.16 and Figure 6.17 show histograms and kernel density estimates for metacarpals and metatarsals respectively. These generally show three groups, the frequency of specimens in which decreases as size increases. This supports the suggestion that wild females and domestic males overlap, but unless the female:male ratio is extremely high plausible for domesticates but unlikely for aurochsen the central group must be composed mainly of domestic bulls. First phalanges show a similar pattern to the metapodials.
6.2.2.2 - cranial measurements

Only the lower dentition was measured for cattle, and this shows less potential than that of pigs for separating wild from domestic specimens. Only for M3 is a single clear outlier observed, along with a cluster of possible wild specimens (Figure 6.15). The single spread seen for dP4 may reflect an absence of juvenile aurochsen, although four specimens from Gomolava could be wild.
a) Bos lower M3 measurements

24
Gomolava
22 20
Petnica
Breadth (mm)
18 16 14 12 10 30 35 40 45 50
Length (mm)
b) Bos lower dP4 measurements

17 Gomolava 16 Breadth (mm) 15 14 13 12 11 24 26 28 30 32 Length (mm) 34 36 38 40 Petnica
Figure 6.15: Bos lower dP4 and M3 measurements from Gomolava and Petnica.
140
Biometry
Figure 6.16: histograms and kernel density estimates for Bos metacarpal measurements from Gomolava. All measurements in mm.
141
Biometry
Figure 6.17: histograms and kernel density estimates for Bos metatarsal measurements from Gomolava. All measurements in mm.
142
Biometry
Figure 6.18 shows horn core measurements from Gomolava, Petnica and Divostin along with the rough limits of clusters within the huge sample from Obre II (Bknyi 1974b, fig. 4). The Obre II data include a main group presumably domestic cows with maximum diameter less than around 70 mm (marked a), followed by a looser cluster (b) that Bknyi divides roughly 50:50 between wild females and domestic males. Finally, three extremely large specimens (maximum diameter > 95 mm) are taken to represent wild males.
a)
Minimum basal diameter (mm)
75 70 65 60 55 50 45 40 35 30 30 40 50 60 70 80 Maximum basal diameter (mm) 90 100 Gomolava Gomolava A.C. Petnica Divostin
b)
280 260
Basal circumference (mm)
Gomolava 240 220 200 180 160 140 120 40 50 60 70 80 Maximum basal diameter (mm) 90 100 Gomolava A.C. Petnica Divostin
Figure 6.18: Bos horn core measurements from Gomolava, Divostin and Petnica, with summary results from Obre II (shown as ovals in chart a). Sources as previous figures.
143
Biometry
The clear division into two groups at Divostin is suprising, but the Gomolava data are as indistinct as expected. Those falling firmly within ellipsoid a are considered domestic, probably female, and if any points had fallen well above b they would have been taken as wild males. Those within b, however, may represent either domestic bulls or wild females. Following Armitage and Clutton Brock (1976), Russell (1993, 118-126) uses a circularity index to distinguish between sexes from horn cores at Opovo, but this does not appear sexrelated in a disparate sample of early-modern cattle (Sykes & Symmons 2007, fig. 4). Interpopulation variation is an issue here: circularity may be a good guide to sex within a single Neolithic population, but in the absence of known-sex Vina reference data this remains conjecture. In any case, if wild individuals are present and morphologically distinct then we are by definition dealing with multiple populations whose dimorphism need not follow the same pattern. Since sheer size as measured by basal circumference and diameter shows little overlap between the sexes even in the mixed-breed early-modern sample, sex would appear an unavoidable confounding factor when distinguishing wild and domestic.
6.2.2.3 the log-ratio method

Only postcranial cattle elements were standardised, due to the apparent lack of potential for separating wild and domestic teeth even from the most common measurement pairs. The standard used was the Ullerslev cow (Degerbl & Fredskild 1970), a relatively large wild female from Denmark. Since Danish aurochsen are typically larger than their south-east European counterparts (Grigson 1969), Russell (1993, 113) points out that the Ullerslev specimen should fall well within the wild range for Vina sites. Figure 6.19 shows the overall distributions of log-ratio values at Gomolava and Petnica. While these are clearly unimodal with modes of around -0.175 and -0.125 respectively that for Gomolava shows a slight positive skew, hinting at the presence of a much smaller secondary population with a mean greater than that of the main group. The same is true at Opovo (Russell 1993, 116), where the mode is approximately -0.13. Plotting log-ratio values by domestication status as assigned on an element-by-element basis supports the original assignations, although there is an appreciable overlap (Figure 6.20). The dividing line, such as there is, falls around the zero mark. While it is certainly not possible to identify every specimen, one can say with considerable confidence that aurochsen are rare at both sites.
144
Biometry
Gomolav a
75
Petnica
Frequency
50
25
0 -1.0
-0.5
0.0
0.5 -1.0
-0.5
0.0
0.5
Figure 6.19: distributions of Bos post-cranial mean log-ratio values at Gomolava and Petnica.
Frequenc y
100 75 50 25 0
Domestic
Frequenc y
100 75 50 25 0
Indeterminate
Frequenc y
100 75 50 25 0 - 1.0
Wild
- 0.5
0.0
Mean log -ratio value
Figure 6.20: distributions of Bos log-ratio values at Gomolava and Petnica by previously assigned domestication status.
6.2.3 dogs and wolves

Dogs are the first domestic animals known in the Central Balkan region, being present at Vlasac (Bknyi 1978), while wolves remain to this day in the more remote areas. Canis specimens are relatively rare at Gomolava and especially Petnica, and are biased towards cranial elements. Figure 6.21 shows the distribution of log-ratio values, using the Hungarian Agricultural Museums specimen 73.4 as the standard (from Russell 1993, 163). Since this is a fairly small female it can be expected to lie towards the lower limit for wolves.
145
Biometry
Gomolav a
Petnica
Frequency
-1.0
-0.5
0.0
0.5 -1.0
-0.5
0.0
0.5
Figure 6.21: mean postcranial Canis log-ratio values from Gomolava and Petnica
All measurable specimens from both sites proved smaller than the standard, whereas at Opovo a cluster of log-ratio values between 0 and 0.2 is assumed to represent wolves (Russell 1993, 161). Several metapodials two of which articulate and an astragalus from Petnica were preliminarily identified variously as wolves and Canis sp. but none was measurable. If wolves are represented at all amongst the post-cranial Canid remains at Petnica it is thus only by elements which could have arrived in association with skins rather than entire carcasses. Wolves were present in Gomolava Blocks I to VI in very small numbers (Clason 1979a, 64). The red lines in Figure 6.22 show the maximum values from the sample of 27 measured Canis mandibles from Obre II and eleven from Divostin, all identified as domestic (Bknyi, 139-140; 1988, 444). Again, no specimens from Gomolava or Petnica stand out as wild.
6.2.4 deer
Red deer are highly sexually dimorphic, with males reaching up to 1.7 times the bodyweight of females (Clutton-Brock et al. 1982). Russell (1993, 279-291) identifies sex groups at Opovo from measurements on the scapula, humerus, radius, ulna and astragalus, but no clear clusters are seen among these elements at Gomolava and Petnica (see Appendix 3.3). Differential weight distribution between the sexes in cattle renders sexual dimorphism greater for metacarpals than for metatarsals (Bartosiewicz 1987, 48). This appears to hold true for red deer at Vina sites. Metacarpal measurements from Gomolava and Petnica line up remarkably well with Russells (1993, 281) data for the distals, and quite well for the proximals (Figure 6.24a, b). Sex can be assigned fairly securely in most cases. The same cannot be said for the metatarsal datasets, each of which forms a single spread (Figure 6.24c, d). Red deer pelvises can be sexed through non-metric traits of the pubis and acetabulum, providing an independent check on the assumption that size groups represent sexes. While
146
Biometry
there is little clear water between the size ranges of male and female specimens at Gomolava and Petnica (Figure 6.23) there is a gratifying absence of overlap between the sexes. One possible explanation for the lack of clear sex groups at Gomolava and Petnica compared to Opovo is that conditions were more favourable for deer at the latter, since dimorphism is greatest in prime conditions (Mitchell et al. 1981). The overall size range of the deer at the various sites provides only weak support for this hypothesis, however; while most measurements attain their greatest value at Opovo this is usually only very slightly higher than the equivalents at Petnica and Gomolava, and there is no overall impression of larger deer at the former site.
a) mandible me asurements 1 14 13 12 11 Breadth in front of M1 (mm) 10 Breadth of M1 (mm) 9 8 7 6 5 4 3 2 1 0 -1 -2 0 2 4 6 8 10 12 14 16 18 20 22 24 He ight in front of M 1 (mm) c) mandible me asurements 3 45 40 35 Length of premolar row (mm) 30 25 20 15 10 5 8 7 6 5 4 3 2 1 0 -1 -2 0 2 4 10 9
b) mandible me asureme nts 2
Gomolava Petnica
Gomolava Petnica
10
12
14
16
18
20
22
24
26
Length of M 1 (mm)
Gomolava
0 -5 -5 5 15 25 35 Le ngth of molar row (mm)
Petnica
Figure 6.22: selected Canis mandible measurements from Gomolava and Petnica. Dashed red lines show maximum values from Obre II and Divostin (Bknyi 1974; 1988).
147
Biometry
Frequency
Gomolava female
Petnica female
0 3
Gomolava male
Frequency
Petnica male
0 3
Frequency
Gomolava in det.
Petnica indet.
20
40
60
20
40
60
LA (mm)
LA (mm)
Figure 6.23: Cervus acetabular lengths at Gomolava and Petnica by sex.
Roe deer are only weakly dimorphic, with males about 1.1 times the size of females (CluttonBrock et al. 1982). Given the limited success in distinguishing red deer sexes metrically at Gomolava and Petnica, and the smaller samples of roe deer, the potential for separating male and female specimens is low. Likely sex groups were not observed for any elements (Appendix 3.4).
6.2.5 biometry summary

Biometric analysis gives a good separation between wild and domestic populations of pig in the Balkan Neolithic. The situation for cattle and dogs is rather less clear, but both species are apparently dominated by domestic individuals at Gomolava and Petnica. Biometric sex attribution amongst deer met with very limited success. Measurable specimens are a sub-sample of the whole assemblage, and one which is likely to be biased in favour of larger, more robust specimens, i.e. those from wild and/or male individuals. The wild:domestic and male:female ratios seen in the above figures should thus be treated with caution. That said, attributions based on biometry relatively rarely necessitated changes from the field identifications. While a good number of specimens originally classed as indeterminate were moved into the wild or domestic categories and a smaller number changed in the opposite direction extremely few wild specimens were reclassified as domestic or vice versa. This provides good grounds for confidence in the field identifications of un-measurable specimens.
148
Biometry
a) proximal metacarpal measurements
b) distal metacarpal measurements
40 38 36 Dp (mm) 34 32 30 28 26 40 42 44 46 48 Bp (mm) 50 52 54 Gomolava Petnica Approx. division at Opovo Dd (mm)
36 35 34 33 32 31 30 29 28 27 26 40 42 44 46 48 Bd (mm)
d) distal metatarsal measurements
Gomolava Petnica Approx. division at Opovo 50 52 54 56
c) proximal metatarsal measurements
50 48 46 Dp (mm) 44 42 40 38 36 34 36 38 40 Bp (mm) 42 44 46 Gomolava Petnica Dd (mm)
38 36 34 32 30 28 26 42 44 46 48 50 Bd (mm) 52 54 56 58 Gomolava Petnica
Figure 6.24: Cervus metapodial measurements from Gomolava and Petnica. Opovo data from Russell (1993, 285).
149
Gomolava
6.3 Gomolava
6.3.1 introduction
Gomolava is one of few large tell sites in the Vina culture, located on a meander of the Sava in Srem, near the village of Hrtkovci (Figure 6.25). Due to rapid erosion, only an estimated 50% (Brukner 1988, 20) or even 25% (Petrovi 1984, 7) of the tell survived at the start of excavations. Much of the remainder has been destroyed since their cessation in 1985. The original extent of the mound is thus unknown, but was estimated at 13,800 m2 in 1954 (Brukner 1988, 20). Of this, approximately 5000 m2 has been excavated in a series of campaigns. The known portion of the tell consisted of broad, flat, mounds to the north and south, almost divided by a central col.
Figure 6.25: location of Gomolava. After Clason 1979.
No detailed environmental reconstruction work has been carried out around the site, although palynological analysis (Bottema & Ottaway 1982, 236) indicates that the Neolithic vegetation in the sites vicinity was relatively similar to that of the present day. Petrovi (1984, 7) suggests that the Neolithic site was surrounded by swamp or the river on all sides, while Clason (1979a, figure 4) shows marshes only along the banks of the Sava to the south, with arable land abutting the settlement to the north and east. The latter view fits better with
150
Gomolava
Borojevis (1988, 114) conclusion, based on plant micro- and macro-fossils, that Gomolava was situated on less fertile but drier chernozem than the site of Opovo: less marshy, and hence more suitable for cereal cultivation. In Chapmans (1981, 90) terms, the site is in an area of medium arable potential. Van Zeist (2002, 87) notes that the site was probably founded at some distance from the river, which has since shifted its course. A riverside strip just over 200 m long and between 3 m and 18 m in breadth was excavated during 1953-1957 (Figure 6.26). Further excavations between 1966 and 1968 continued this approach, working back into the mound as the river encroached.
Figure 6.26: plan of excavation campaigns at Gomolava. Partly based on Brukner 1980, figs 2 & 3. The small trench marked S is the 1972 sondage published by Bottema and Ottaway (1982).
Truly systematic fieldwork started in 1969, directed by Prof. Bogdan Brukner, Dr. Nikola Tasi, Dr. Borislav Jovanovi and Jelka Petrovi. Almost the entire surviving northern portion of the mound was excavated, defined as Blocks I to VI. A large number of structures from several building phases within the Vina layers was revealed, along with a cemetery apparently within the bounds of the settlement. A new trench measuring 40 m by 20 m was opened in 1978 on the southern portion of the tell and designated as Block VII. Excavation reached the Vina levels in 1980 and continued until 1985, revealing a particularly dense concentration of building remains and pits. Despite the enormous scale of the excavations indeed perhaps because of it no definitive monograph was produced for the Vina phases. Several brief articles came out of the 1950s campaign (Dimitrijevi 1956; Giri 1960; 1965; Na 1957; 1960; Raajski 1954; Sekere
151
Gomolava
1961), and regular field reports were published in the 1960s and 1970s (Brukner 1965; 1967; 1968; 1972; 1973; 1974; Brukner & Petrovi 1977; Jovanovi 1965a; 1969; Petrovi 1978; 1979; 1982; Tasi & Brukner 1975; 1976). These are supplemented by a number of comments on the tell's stratigraphy (Brukner 1966; Jovanovi 1965b; 1971b), various specialist reports, and a booklet (Petrovi 1984). Detailed plans for Blocks I to VI and the 1967-1969 riverside area are available, along with notes on individual features (Brukner 1980a), but published plans for Block VII (Brukner 1988) are largely restricted to the final Vina phase and are not accompanied by detailed description. The vast corpus of documentation is thus largely unpublished, and the analysis here is an exercise in making the most of limited information, relying on a combination of published material and original plans, sections, unit lists and diaries studied by the author in the Museum of Vojvodina. Permission to introduce the latter into the public domain is not established at the time of writing and I must request that no direct reference be made without consultation. The stratigraphy and chronology are summarised here, with more detailed information in Appendix 4.
6.3.2 phasing and chronology

The earliest artefacts at Gomolava belong to the Early Neolithic Starevo culture, but no pits or structures were identified from this phase (Brukner 1988, 20). Early stratigraphic analyses indicated two Vina horizons (e.g. Brukner 1971), but by the start of Block VII the existence of three phases was well established, referred to as Gomolava Ia, Iab and Ib. (Brukner 1980a). Gomolava Ia is dominated by pits (Figure 6.27), with a single house from Block VII assigned to the phase by the excavators (Brukner 1988, 21). Petrovi (1984, 17) suggests that numerous deep pits may either have been borrow-pits for clay or temporary shelters inhabited prior to the completion of overground structures. According to Brukner (1988, 2122), at least seven pit-dwellings were excavated from Gomolava Ia across all sectors. These interpretations are discussed in 6.3.4.2. The second Vina phase, Gomolava Iab, is dominated by the remains of presumed houses (Figure 6.28). The majority are represented only by postholes and foundation trenches, presumably because the remains of their superstructures were destroyed by later phases of building activity (Brukner 1988, 23). Some patches of burnt building material are attributed to Iab, but it is not always clear whether these are partial remains of structures preserved in situ or secondary dumps of rubble. Some Gomolava Iab houses are extremely large (see 6.3.4.3 buildings). Pits are considerably less common in this phase, although the upper fills of some of the larger pits from Ia appear to have been deposited in Iab.
152
Gomolava
Building remains from the third Vina phase, Gomolava Ib are particularly numerous, (Figure 6.29). A possible 25 structures were excavated, most interpreted as houses (Brukner 1988, 21). These are typically smaller than in Iab (Petrovi 1984, 18) but are substantially built. As at many Vina sites, the Gomolava houses were destroyed by fire. The burnt remains of some Ib houses have suffered very little subsequent disturbance, allowing considerable insight into construction and contents (see Petrovi 1992; 1993). Pits are relatively rare in this phase and small where present. The Gomolava Ib cemetery in Block III is not only the first known from the later part of the Vina culture, but the first discovered within a settlement from any Vina phase. It is close to the edge of the tell, however, and Petrovi (1984, 20) suggests that it is from the very latest part of Gomolava I, post-dating the known Ib structures. It could therefore have been in use after the abandoment of settlement on the tell, or might alternatively relate to a final building phase restricted to part of the mound now destroyed by the Sava. It is also interesting that the numerous grave-goods do not include the figurines, decorative altars or bucrania often found in pits and/or houses. Given the existence of around 2 m of Vina-period deposits across most of the site, the phases at Gomolava clearly cannot have been one-off construction events, but rather periods of occupation. Indeed, plans for Gomolava Iab show at least two episodes of building in the case of House 2/77. It is unclear whether the three phases are envisaged as discrete episodes of occupation or more-or-less arbitrary divisions within a continuous accumulation of deposits. This issue is closely related to their typological phasing and is subject to changing interpretations over the course of the excavations. Phases Ia and Iab can be fairly confidently described as continuous (see Appendix 4.1) indeed there may well be some overlap, with the earliest features assigned to Iab probably contemporaneous with the latest Ia pits while the possibility of a hiatus between Iab and Ib cannot be entirely excluded. Table 6.1 shows the typological phasing from the most recent publications.
Gomolava phase Ia Iab Ib Cemetery Miloji phase B2- start of C C D1 D2
Table 6.1: relative chronology of Gomolava phases in later publications (e.g. Brukner 1988).
153
Figure 6.27: plan of Gomolava Ia in Blocks I-VI, plus 1967-1968 riverside area. Modified from Brukner 1980, figs 3, 4 & 10. Riverside section includes both Ia and Iab remains.
154
Figure 6.28: plan of Gomolava Iab in Blocks I-VI Modified from Brukner 1980, figs 3 & 11.
155
Figure 6.29: plan of Gomolava Ib in 1967-1977 excavation areas. Modified from Brukner 1980, figs 3, 4 & 10. *Two structures are listed as house 6/1972 presumably a typographical error.
156
Gomolava
The radiocarbon dates from Gomolava are of limited use since almost all derive from Ib, and continuity is in any case notoriously hard to demonstrate from absolute chronology (Bayliss & Orton 1994). Figure 6.30 shows the spread of dates for the site, constrained using an appropriate stratigraphic model (see Appendix 4.2). The lone date from Ia under this model is given in Figure 6.31, while the boundaries for the Ib building horizon are as follows (68.2% HPDs): Start: 4929-4859 cal. BC End: 4694-4605 cal. BC At face value the dates suggest a 300-year span for Ib, starting shortly after Ia. Since most are on charcoal, however, some of the Ib determinations may be misleadingly early.
6.3.3 Block VII

Block VII is divided into eight 10 x 10 m squares, each composed of four quadrats (Figure 6.32). Excavation was extended to the river in front of squares A, B and part of C, the portion in front of D and the remainder of C having already been excavated in 1969. These areas are referred to here as A5, A6, B5, etc. The same three Vina phases were identified as elsewhere on the site. A plan of Gomolava Ib houses from the block is available (Brukner 1988, 37), but only for squares D and H was a composite plan produced, along with a vertical projection of the main features from the two squares onto the NW-SE section of the block (Brukner 1988, 35 & 38 respectively). For the present purposes the basic stratigraphy of Block VII had to be reconstructed from published sources and from documentation held in the Museum of Vojvodina (Figure 6.33; see Appendix 4.3). This is a provisional attempt to grasp the stratigraphy a necessary prerequisite for the faunal analysis rather than a definitive interpretation of features in the block.
157
Gomolava
158
Gomolava
Figure 6.30: available radiocarbon dates from Gomolava. See Appendix 4.2 for details.
Figure 6.31: constrained calibration of GrN-7376 (Gomolava Ia).
159
Gomolava
Figure 6.32: layout of Block VII.
6.3.4 feature types 6.3.4.1 the cultural layer

On most Vina sites and in much of the Balkan Neolithic and Copper Age (NCA) discrete features such as buildings and pits exist within a matrix of anthropogenic sediment which is typically very rich in artefacts and organic material, constituting a cultural layer that varies widely in depth. Due to its ubiquity this phenomenon is commonly taken for granted by archaeologists of the region, but in fact its nature and formation are very poorly understood. The cultural layer is most prominent on tell sites such as Gomolava. The classical NearEastern tell-formation process, in which remains of collapsed or destroyed houses are used as the basis for subsequent construction, has been demonstrated geomorphologically at Sitagroi (Davidson & Thomas 1986, 39) and may well apply to other densely occupied mounds in the Balkan NCA. This process alone cannot explain the cultural layer phenomenon on Vina sites, however. Structures at Gomolava are neither arranged densely enough nor apparently rebuilt with sufficient frequency for their erosion, collapse and/or firing to be the primary source for the 2 m of sediment built-up during the Vina period. Moreover, appreciable cultural layers are found even on short-lived flat sites, and their high organic and artefactual content suggests a different mechanism.
160
Gomolava
Figure 6.33: reconstruction of features in Block VII at Gomolava by phase.
161
Gomolava
Chapman (2000b, 62-63; 2000c, 355-356) follows Russell (1993, 72) in arguing that much of the cultural layer derives from waste material being deposited directly on the ground within the settlement. Central to this argument is a critique of the concept of rubbish as applied to prehistory (see 3.3.3). In the Balkan NCA, Chapman argues, material which we might class as rubbish was in fact subject to structured deposition. Specifically, objects produced and/or used within a household did not lose their links to that household at the end of their uselives but rather were deposited nearby in a way that minimises the distance between the living people and their material culture (Chapman 2000b, 63). The cultural layer phenomenon is thus the result of a specific attitude to deposition on Balkan NCA sites:
a walk around the settlement involved avoiding the larger, if not sharper, materials lying on the ground and was dominated by the smells of decomposing human faeces, vegetal and animal matter what most twentieth-century archaeologists would call a refuse tip. Chapman 2000c, 356
This phenomenon has certain implications. Firstly, the gradual nature of accumulation renders stratigraphic excavation extremely difficult, and is partially responsible for the continuing popularity of arbitrary spits and grids. As a result, much material from Vina sites typically derives from arbitrary or at best insecure units. Worse still, the lack of clear stratigraphy within the cultural layer obscures relationships between discrete features set into it, giving the impression that they are floating in a sea of homogeneous sediment and necessitating an overreliance on pottery typology and absolute depths for phasing features which do not directly overlie each other. This is clearly a problem at Gomolava. Secondly, Chapmans concentration principle (2000b, 82) tying objects to households provides a theoretical foundation for the common assumption that deposits near a structure are likely to be informative regarding the activities, possessions and/or status of its occupants. The spatial distribution of artefacts and ecofacts within the cultural layer is thus of considerable importance. To make the most of this principle greater precision is required than the 5 m quadrats used in Gomolava Block VII, but even this resolution confirms the concentration of animal bones around houses at Divostin, matching the pattern for pottery (McPherron & Gunn 1988). The use of single-context recording in current excavations at Belo Brdo will hopefully shed light on the anatomy of a cultural layer. Unit 830, a very large, homogeneous, artefact- and bone-rich deposit from the 2006 season, was excavated as a single unique feature but is now suspected to be one of the constituent lenses of the cultural layer, poorly differentiated from the surrounding units and perhaps formed over an appreciable period (M. Pori, pers. comm.). Ironically, for extensive units such as 830 the abandonment of an arbitrary grid in
162
Gomolava
favour of single-context recording results in the loss of the very spatial data which is important under Chapman's model.
6.3.4.2 pits
Pits jame are a common if highly varied feature on Vina sites. Often containing the highest concentrations of artefacts and ecofacts, their nature and purpose are not well established. Suggested functions include storage, extraction of clay, refuse disposal and occupation. Not all of these are mutually exclusive, and it is clear that no single function can account for all Vina period pit digging and pit filling. The interpretation of larger pits on Vina and Starevo sites as semi-subterranean dwellings zemunice is common (e.g. Bailey 1999a; Bogdanovi 1988; Garaanin 1949; Greenfield & Jongsma 2006; 2008; Vasi 1910) but not without critics (e.g. Ehrich 1977, 65; Lichter 1993, 24-27; Markoti 1984 92-93; Peri 1998; Tripkovi 2003, 449) both amongst local specialists and western prehistorians (contra Greenfield & Jongsma 2008). Several Gomolava Ia pits are interpreted in this way (Brukner 1980a; 1988, 21), partly to account for the scarcity of evidence for other forms of occupation in the phase. Petrovi (1984, 17) suggests that they may have been temporary shelters prior to the construction of the surface houses, but this would imply a very short duration indeed for Phase Ia. Chapman (2000b, 86) lists three common lines of argument for pit-houses. Firstly, structural features within the pits themselves may indicate occupation. In some cases small postholes suggest the presence of a roof, although as Chapman observes this does not necessarily mean that a feature was lived in. Noting that the post holes are almost always inside the pit he cites experimental work by D. Monah demonstrating the unsuitability of pit-houses constructed in this way for human habitation (Chapman 2000b, 87). Hearths and ovens within pits have also been taken as proof of occupation (e.g. Jongsma 1997, 136), although unequivocally outdoor pyro-features at many Vina sites (Tringham 2005, 99), including Gomolava (Brukner 1988, 22), undermine this argument. Moreover, such features are rarely at the base of the pit as would be expected (McPherron & Christopher 1988, 470-471). Likewise, Tringham (1971, 84) points out that no floor has ever been recovered from the base of a Balkan Neolithic pit. Jongsma (1997, 83-95, 123-128) neglects secondary deposition in arguing for pit-houses from the presence of daub fragments. Secondly, concentrations of artefacts in pits are sometimes interpreted as in situ occupation waste. Since dense and/or unusual concentrations may be found at various levels within the fill, however, Chapman (2000b, 87) argues that they represent a later phase in the pits lifecycle. None of this excludes the possibility of pit dwelling, but the third line of argument
163
Gomolava
ethno-historic and archaeological analogy militates strongly against it. Examples of semisubterranean dwellings are not especially rare, but the Vina and Starevo zemunice do not compare favourably. Chapman cites 19th century pit-houses in Romania and Hungary while McPherron and Christopher (1988, 470-471) refer to Neolithic Jomon dwellings, but in both cases the same features are listed: rectangular shape, straight sides, flat floors and a solid, regular pattern of supporting posts. The majority of putative pit-houses on Starevo or Vina sites has none of these features, while Na (1960, 112) notes that those at Gomolava are more similar to modern borrow/refuse pits on the edges of villages in the region. Table 6.2 summarises the evidence for pit-dwelling at Gomolava. Eleven pits are interpreted as zemunica but justification is only given in five cases, and none matches the expected form of pit-houses outlined above. Only Pits 11 and possibly 15 are plausible candidates for occupation, but there is no particular reason to believe this was their purpose. Overall, pitdwelling at Gomolava seems very unlikely. The role of the pits thus remains to be explained. Given the importance of cereals for Vina subsistence (Borojevi 2006, 7-8; Hopf 1974), storage of grain as indeed of other organic materials must have been important. Smaller pits have been interpreted in this way at Vina and Starevo sites (e.g Banjica: Todorovi & ermanovi 1961), especially when roughly cylindrical (e.g. Divostin I, McPherron & Christopher 1988, 470-471) or pear-shaped in profile (e.g. Belo Brdo, Vasi 1936, 145-155). At Selevac, shallow pits with a clay lining are assumed to be storage facilities, although domed clay silos are also found (Tringham & Stevanovi 1990, 114). Few of the pits from Gomolava meet these descriptions. Of the pit plans and sections viewed by the author only Pit 16 (Iab) is roughly cylindrical, and none shows any kind of lining. Pit 16 also has the most homogeneous fill. Some Block VII pits attributed to Gomolava Ib are both small and round in plan, but they do not appear in unit lists and specific documentation has yet to be seen by the author. Likewise, the smallest pits from the earlier blocks, such as L and J in Ia (Figure 6.27) or the feature just north of House 3/72 in Ib (Figure 6.28), could feasibly be storage facilities. For now this remains speculation.
164
Block Ash, charcoal, animal bones, 'finds' Ia Ia (& Iab?) Ia (& Iab?) Ia Ia Ia Ia (& Iab?) 1.97 No ? 1.48 ? ? ? 1.66 Round No ? ? 2.58 No ? ? 1.69 No No No 1.94 No No No 0.50 ? ? ? "A few figurines", lots of ceramics ? Ceramics and a figurine No Fig. 6.27 Fig. 6.27 Fig. 6.27 Fig. 6.27 Fig. 6.27 Fig. 6.27 Fig. 6.27 No No
Pit
Interpretation
Cited grounds
Contents noted
Phase
Approx. depth (m) Plan
Regular plan
Vertical sides Flat base
Photograph
67-68
x.1
'zemunica'
II
x.2
'zemunica ?'
Brukner 1980, Ph.2 Brukner 1980, Ph.3
III
'jama-zemunica'
Postholes towards SW edge
VI
'zemunica-jama'
VI
'zemunica-jama'
Postholes
VI
'zemunica-jama'
In situ hearth
II
'zemunica-jama'
VII
'zemunica'
Broken figurine, lithics (2 axes, 2 cores, Ia & Iab 3 blades), bone awl, lots of animal bones, ash & charcoal lenses Lithic core, lots of ceramics inc. large vessel, "red earth" (ochre?), lots of animal bones, ash & charcoal lenses Bone needle, ceramics, lithics, lots of animal bones, ash & charcoal lenses Lots of ceramics and animal bone, small patches of charcoal Ia (with Iab recut) Ia 0.60 Ia (& Iab?) 1.00 No
1.26
No
No
No
Fig. 6.33a
VII
10
'zemunica'
Large patch of burnt daub
No
Fairly
Fig. 6.33a
VII
11
'zemunica'
No
Yes
Yes
Fig. 6.33a
VII
15
'zemunica'
Postholes skirting one side
1.30
No
Fairly
Partly?
Fig. 6.33a
Table 6.2: evidence for pit-dwelling at Gomolava. See text for explanation of Serbian terms in interpretation column. Data on Block VII from unpublished documentation; data on previous areas from Brukner 1980.
165
Gomolava
Storage away from buildings would be interesting, especially in Ib. Concentration of storage facilities inside houses may be a corollary of social changes (e.g. Bailey 2000, 163, see 2.3) and by later phases of the Vina culture storage seems mainly to have involved fired-clay bins and pithos-like storage vessels, as at Divostin II (Bogdanovi 1988) and the most recent levels at Selevac (Tringham & Stevanovi 1990, 114). At Gomolava, one room of House 4/75 is interpreted as a grain storage area (Jovanovi 2004; Petrovi 1993). Markoti (1984, 96) refers to storage pits within houses at Banjica and Crnokalaka Bara, but none is recorded from Gomolava. The typically large size and amorphous form of Vina and Starevo pits are more consistent with their interpretation as borrow-pits for building materials and/or potters clay (e.g. Petrovi 1984; Russell 1993). As McPherron and Christopher (1988, 470) point out, however, their frequency does not generally correlate with the intensity of daub use. More damningly, pits are often cut into deposits of no use for construction. Neolithic pits have also been interpreted as working areas. Lithic reduction is most commonly implicated, but the occasional presence of ovens or hearths suggests a wider range of uses. At Gomolava, Pit 6 (Iab) is interpreted as a lithic workshop (Brukner 1984, 32) based on the presence of 48 cores, 108 blades and 203 debitage flakes. This is plausible although the number of finished blades is problematic but alone is an unlikely explanation for the pits creation, especially given that at least one surface-level lithic-working area exists. Concentrations of mussel and/or edible snail shells are found in Pits 6, 20 and x.9 (Clason 1979b, 64) at Gomolava and are reported from three pits at Uivar (Beigel & Kuhn 2005, 46), suggesting mollusc processing as another possible activity. Again, however, such concentrations exist within the cultural layer and in any case need not reflect in situ activity. Figure 6.34 shows the evidence for activities in pits and on the surface in Block VII. It would be hard to account for the digging of pits purely in terms of activities which also took place elsewhere. The same argument applies to the role of pits in refuse disposal. Material which we might class as rubbish was certainly deposited in pits, but was also routinely deposited on the surface. It is difficult to explain the decision sometimes to dispose of the same classes of material in discrete pits in functional terms. The contents of pits are, however, often quantitatively and structurally different from those of the surrounding cultural layer (Chapman 2000b, 63).
166
Gomolava
Figure 6.34: evidence for activities in pits and on the surface in Gomolava Block VII. Serbian terms for pyro-features are discussed in 6.3.4.5.
Each pit at Gomolava may have served one or more of the purposes outlined above during its life-cycle. Since the majority of evidence regarding the use of a pit its contents relates to its filling, however, to study pits is to study deposition. Chapman sees this life-cycle as a continuous, meaningful process rather than a sequence of expedient activities pit digging cannot conceptually be separated from pit-filling, even if the immediate purpose is not depositional. Exploring the implications of pit digging, he makes an initial distinction between pits dug into the natural and those cut into a pre-existing cultural layer. While both categories involve the eventual placement of new cultured material into the ground, in the latter case this replaces ancestral material which has been removed for current use:
Far from being simply a neutral means of disposing of unwanted refuse, pitdigging can be seen as an exchange with the ancestors or of new material for old when the pits are dug into earlier cultural layers Chapman 2000b, 64
Even the former kind of pit may be seen as establishing links between the present and the more distant/abstract past. Chapman notes that pits are set apart from the cultural layer by: the limitation of horizontal (spatial) relations and the expansion of vertical (temporal) relations (2000b, 64). They thus have a parallel with tells, and it is perhaps not purely a matter of practicality that they are more common on flat sites. If the basic principle of tells is living where ones ancestors have lived, then surface deposition may have had a greater resonance, rendering direct ancestral exchange through pit digging less necessary to stress continuity. It might conversely have been of greater significance when it did occur. This trend may explain the decrease in
167
Gomolava
frequency of pits between phases at Gomolava, since the site was not a tell in its earliest phases. All Iab pits and many from Ia are nonetheless cut into or at least through a cultural layer, with all the attendant implications. Whatever ones views on these arguments, the evidence for structured deposition in Vina pits is compelling. Chapman (2000b, 65) defines two very broad categories of pits foundation deposits (associated with buildings) and open air. The former is rare in the western Balkan NCA, although a dog skull with articulated mandibles immediately under the floor of House 2 at Opovo is an exception (Russell 1993, 167-168). At Gomolava a bone concentration underneath Structure 11/81 may represent a foundation deposit. Chapmans second category is more important in the Vina context. Clear cases of special deposits are again rare, but several exist in the Gomolava pits. Two articulated dog skeletons were discovered in Pit H (Figure 6.35) while a complete Bos skull with horns was placed at the bottom of Pit x.3 (Figure 6.36). Both features are from Ia. A connection between cattle skulls and primary pit-fill may be more common than reported: the Gomolava collection includes a number of more-or-less intact skulls of both wild and domestic cattle which may have derived from similar placements, while in Block VII horn cores are preferentially found in the lower fills of pits. Chapman (1981, 73) mentions a complete aurochs skull from a possible ritual pit, possibly Pit x.3. A single aurochs horn core was found near the base of Feature [pit] 45 at Opovo in association with a complete pot (Russell 1993, 84). An articulated juvenile dog skeleton was also found at Opovo, in Feature [pit] 41, while pits at both sites contained surprisingly intact dog skulls. These cases notwithstanding, the primary evidence for structured deposition in Vina pits concerns the commonly heterogeneous nature of the fill, often with alternating layers of burnt/unburnt or otherwise contrasting deposits, sometimes associated with different artefact classes (Chapman 2000b, 70-73; e.g. Selevac Tringham & Stevanovi 1990, 100). Evidence for structured deposits is heavily dependent on excavation technique. Chapman infers likely structured deposition from rescue excavations at Kompolt-Kistr on the basis of complex, heterogeneous pit sections and analogy with the more carefully excavated site of Endrd 119 (2000b, 76-78). Similarly, pit sections from Gomolava Block VII reveal some fairly complex fills, permitting comparison with the better documented features at Opovo. The latter is less than 60 km from Gomolava and was almost certainly inhabited during the tells occupation, while similarities in admittedly rare special deposits have already been noted.
168
Gomolava
Figure 6.35: double dog burial in Pit H, Gomolava Block III. Photo: Museum of Vojvodina.
Figure 6.36: cattle skull placement in Pit x.3, Gomolava Block I. Photo: Museum of Vojvodina.
The Opovo pits are discussed in detail by Russell (1993, 80-86, 428-461) and Borojevi (1998; 2006; see also Tringham et al. 1992), focusing on animal and plant remains respectively. The larger pits fills tend to be multi-layered, consisting of garbage animal
169
Gomolava
bones, pottery sherds and plant remains intermingled with multiple lenses of ash, charcoal and other burnt material. In several cases some of the burnt lenses typically in the upper fill seem to represent in situ burning. Russell (1993, 428-429) proposes two possible reasons for this: burning down of the contents for practical reasons, or ritual burning to end the use-life of pits, mirroring that suggested for houses (Stevanovi 1997; Tringham et al. 1992). Since many pits are capped with burnt house rubble, Russell suggests that their closure might be directly related to house burnings. On the other hand, pits often include fragmentary burnt daub within the fill, while some received substantial deposits of rubble more than once. Tringham (2005, 108) argues that inclusion of material from long-since destroyed houses could also have been meaningful. Russell argues that animal bones in at least some pits derived from feasting (1993, 428-429; 1999). This conclusion is based on internal stratigraphy combined with circumstantial faunal evidence for different modes of consumption and the fact that cereal grains were almost entirely restricted to pits (see also Borojevi 2006):
One can at least imagine a scenario where feasting remains are placed in garbage pits, which are then finished off with ritual burning, perhaps even including the burning and dismantlement of a house Russell 1993, 429
Communal consumption events associated with pits need not have been restricted to ritual closures, and the tentative life spans assigned by Russell (1993, 429-437) from season-ofdeath estimates around six months for smaller pits and perhaps a few years for the largest suggest a more gradual build-up of material. Feasting constitutes another possible explanation for pit digging: while a general scatter of detritus was the norm for inhabitants of Vina settlements, this attitude may not have stretched to substantial piles of waste from large-scale consumption events (Russell 1993, 439). Such practical concerns do not preclude a meaningful nature for deposits. If pits were dug to provide building materials for houses, became filled with materials deriving from activities in those houses possibly including feasts hosted by the relevant household and in some cases were eventually closed in rituals associated with the death of those houses, this would represent a tangible example of Chapmans abstract concentration principle: structured deposition of daub in pits would have linked house narratives with pit stories (Chapman 2008, 73). Some Gomolava pits show parallels with the more complex Opovo cases (see Appendix 4.3.3). There is little evidence for rubble capping, but this is unsurprising given the lack of burnt house remains in Gomolava Iab and especially Ia. Ashy lenses and small pieces of burnt daub are common, however. In situ burning cannot be inferred securely from pit sections alone but is mentioned in the diaries, although few of the pits are discussed in detail. Multiple
170
Gomolava
discrete fills are fairly common, often with burnt remains at the interface. Two Ia pits have clear re-cuts assigned to Iab, the fills of which show some of the best evidence for in situ burning, with substantial areas of burnt earth forming complete layers. Pit 15 also has part of a Iab structure probably re-deposited immediately above it. By contrast, the fills of new Iab pits are among the simplest where sections are available, and it may be significant that the re-cuts delve entirely into cultural deposits. There is no apparent pattern in the spatial distribution of pits with more complex fills. While the processes and motivations involved in pit digging and filling at Gomolava remain unclear, there is good evidence for structured deposition in at least some cases. Chapman argues that this should follow the same concentration principle as surface deposition, and links it to his fragmentation theory (Chapman 2000a). Since we have already seen (3.2.3) that animals, where owned, constitute both the inalienable possession and the fractal artefact par excellence, it is to be expected that their deposition in pits will have had considerable significance, even if carried out on an everyday basis. Differences in depositional patterns between owned and hunted species are thus of particular interest. Russells feasting hypothesis, also constituting a form of structured deposition, would have different implications for interpretation depending on the social context of feasts. Both models are complementary, however, in stressing the potentially meaningful nature of deposition in pits and the likely links to particular households and to ideas of continuity of dwelling on a site.
6.3.4.3 buildings
This is not the place for detailed discussion of Vina houses (c.f. Chapman 1981, 60-68; Lichter 1993; Markoti 1984, 93-98), but a brief description is necessary. There is very little evidence of construction in Gomolava Ia. Brukner (1990, 81) suggests that Gomolava was little more than a farmstead at this point, perhaps with other scattered dwellings on the destroyed part of the mound (Brukner 1988, 21). Blocks I-VI contain between six and eight posthole structures from Iab (Brukner 1980a, see Figure 6.28), but the situation in Block VII is less clear. The evidence is discussed in Appendix 2.3.4, but Figure 6.37 shows tentative outlines of probable and possible structures. These are provisional pending more detailed study, but the key point here is the presence of multiple structures in dense association and probably not built in a single episode. At least 25 burnt structures are known from Ib, including ten from Block VII (see Figure 6.33). The buildings seem to form three clusters within the excavated area, of which the densest lies at least partially in Block VII (Brukner 1988, Abb. III). As in Iab, the buildings
171
Gomolava
on the northern part of the mound are oriented NNW-SSE and those on the southern part mostly NW-SE. The equation of structure with house is problematic, both in theory and on the basis of observed variation. While most Vina buildings are of a size appropriate for occupation by a nuclear or extended family (Chapman 1981, 61-62), smaller structures also exist. Chapman sets a cut-off between houses and ancillary buildings at 10 m2. A posthole structure associated with a layer of carbonized grain on its floor at Selevac may be a granary (Tringham & Stevanovi 1990, 104-106), while House 3 at Opovo could represent a well-house (Russell 1993, 78-79). Several structures at Gomolava may not be residential, but with the available data this remains speculation.
Figure 6.37: probable and possible Gomolava Iab structures in Block VII. (As reconstructed by the author).
Certain changes over time are worth mentioning. Firstly, the average size of buildings decreases considerably from Iab to Ib, attributed by Brukner (1990, 81) to a shift from extended to nuclear families as the basic co-resident unit. Conversely, Chapman (1981, 6162) reports increasing house size in the Late Vina period as a whole, interpreted as a move towards extended families. The Gomolava trend would not be inconsistent with the model of social change set out in 2.3, but a decrease in size differentiation between phases is more surprising. Chapman also notes that where larger buildings occur alongside smaller as in
172
Gomolava
Gomolava Iab their inventories do not support a qualitatively different function. The large Gomolava Iab posthole structures show a tripartite division, but this is retained in the much smaller Ib buildings. Some of these (Houses 4/75 & 3/72) appear to have two pyro-features, while House 6/80 has four small rooms. Differences in preservation obstruct discussion of building construction, but also suggest developments. The abundance of preserved house remains in Ib may relate to lack of subsequent disturbance compared to Iab. Given the proximity of the Baden/Kostolac layers to Gomolava I often cutting Ib features equally likely explanations are (a) increasingly substantial house construction, with greater use of daub rather than wood, or (b) house burning rather than dismantlement becoming the norm. Similar arguments have been proposed for Selevac (Tringham & Stevanovi 1990, 110). The amount of burnt daub and other material found in the cultural layer and pits suggests that house burning was common from at least Iab, however, so a combination of changes in construction and decreased disturbance/re-deposition is probably responsible. Within the earlier phases there seems to be a shift from postholes and temeljni ukopi to foundation ditches (Brukner 1980a, 24), indicated by the superimposition of Houses 3/80 and x.1, and of the two phases of House 2/77. House Ib structures are typically built on plots levelled with a layer of clay (Petrovi 1992, 20). Increasingly compact floors and walls thus seem to be accompanied by less intrusive foundations. This is perhaps a parallel for the decrease in pit digging, given Chapmans arguments for the significance of cut-features on tells. There is also considerable continuity at Gomolava. In at least two cases (Houses 1/85, x.1 and 3/80; Houses 2/77 and 4/85) three structures are directly superimposed. Most houses are not built directly on previous structures, but Petrovi (1992, 20) suggests that Ib houses were erected adjacent to their predecessors. On a larger scale, the three building clusters in Ib are arguably already present in Iab, albeit shifted slightly. Differences in orientation between these clusters persist between phases, while the largest structures in both Iab and Ib are arguably found in the central part of the tell (see Brukner 1988, Abb. III). House inventories only available for Gomolava Ib are not discussed here, save to note that they are consistent with other Vina sites (c.f. Chapman 1981, 62-68; Markoti 1984, 95-98; Petrovi). Bucrania are worth mentioning, however. These are primarily a later Vina
phenomenon and have been attributed ritual or cult significance (e.g. Chapman 1998, 125). It has been suggested that they were attached to the outsides of houses (e.g. Jovanovi & Glisi 1960; Petrovi 1984, 18) although most are actually found within buildings (Chapman
173
Gomolava
1998, 125), sometimes apparently mounted on a clay pedestal (e.g. Kormadin: Chapman 1981, 64; Opovo: Russell 1993, 78). At least four bucrania were found in Gomolava Ib, two each from Houses 4/75 and 6/80 notably the two best preserved buildings. Their absence in earlier phases might reflect house preservation, although one would still expect to find examples in secondary contexts. It is possible that bucrania were treated as heirlooms or otherwise curated, but given their concentration in later Vina phases elsewhere it is most likely that they were simply not present prior to Ib. Both bucrania from House 4/75 are attached to sections of daub, one being found in the central room and the other just outside the building. Petrovi (1992; 1993, 10) therefore suggests that the first formed part of a cultic area inside the house while the second was mounted on the external wall. Since cattle skull/horn core placements in pits are probably restricted to earlier phases, the Ib bucrania could be taken as an example of the concentration of activities in houses towards the end of the Vina period. In addition, a bucranium featuring red deer antlers is reported from the earlier excavations, without detailed provenance (Na 1957, 404).
6.3.4.4 pyro-features
A number of features at Gomolava are described as oven/kiln (pe), hearth (ognjite) or grill (rotilj). Both pe and ognjite often refer to domed heating structures which would be called ovens in English. Pe seems to be used preferentially for large pyro-features within houses, and ognjite more for smaller structures and those outside. The meaning of rotilj in an archaeological context is unclear, but it is applied to some particularly dense clusters of burnt material in the pits and cultural layer. Pyro-features are found both inside and outside of houses on Vina sites (Chapman 1981, 6364; Markoti 1984, 97), and there is some evidence of a shift towards the former over time. Vina houses usually contain at least one pyro-feature, at least on later sites, although this varies. Only one building at Opovo contained an oven in the excavated area (Russell 1993, 77-80), while several at Divostin had multiple pyro-features, with four in House 14 (Bogdanovi 1988, 79-84). External pyro-features receive less attention, but Chapman (1981, 63) suggests that they represent shared facilities at Banjica. Where sufficiently well-preserved, Gomolava Ib houses typically contain at least one pyrofeature. House 3/72 has two horseshoe-shaped ovens (Brukner 1980a, 26), House 4/75 has a large domed oven in one room and a smaller hearth in another (Petrovi 1992), and ovens
174
Gomolava
are found in at least four Ib houses in Block VII. The 1967-1968 area includes two apparently external Ib pyro-features (Brukner 1980a, fig. 5), but the first may actually derive from Gomolava II (Brukner 1980a, 8) while the second appears internal if plotted on a more complete plan (see Brukner 1988, Abb. III). There is more evidence for external pyro-features in earlier phases. Apparently outdoor collapsed hearths are noted in Block VII D3, E3 and F4 (Figure 6.34 and Figure 6.37) as well as in Pit I, Block VI. Pits 13 and 19 contain potentially re-deposited parts of ognjita. A Iab pyro-feature immediately beneath the NW part of House 4/80 described as pe in the unit lists but ognjite in the diary is probably external. This was the only pyro-feature from which animal bones were recorded.
6.3.4.5 from pits to houses

Links between houses and pits are obviously limited at Gomolava, where the latter largely predate the former. This is a common phenomenon on Vina sites (Tripkovi 2003, 450-451). Referring to pit-sites in the Balkan Early Neolithic, Chapman (2008, 76) proposes three possible explanations for the absence of houses: (1) pit-sites are exclusively loci of deposition, with people actually dwelling elsewhere; (2) over-ground dwellings existed but left no trace; and (3) houses were in unexcavated areas. Tripkovi (2003 451) favours the latter for the Vina site of Banjica, but the extent of excavation at Gomolava militates against such an explanation. Besides, shifting on-site use of space coupled with limited excavation cannot account for the fact that pit-dominated horizons on Vina sites are always at the start of occupation. Explanation (2) seems most likely for Gomolava Ia, although (1) may also be of some relevance if construction of more ephemeral dwellings reflects repeated short-term presence on-site rather than permanent occupation. The earliest buildings at Selevac are suggested to have been built predominantly of wood, with sparse shallow postholes (Tringham & Stevanovi 1990, 108), and the first houses at Banjica are less substantial than in subsequent horizons (see Tripkovi 2003, 450-451). Following Brcks (1999) arguments, Chapman (2008, 73) takes the absence of houses as evidence for a lack of investment in domestic structures or domus-based ideology. Looked at slightly differently, peoples attachments to place may have been expressed more through pit digging/filling than through construction of substantial dwellings. This expression may have operated on an intra-site level between groups approximating to households in the manner implied by the concentration principle, even where domestic structures themselves do not provide durable foci. Place-marking through depositional practices would certainly be less exclusionary in tone than the construction of wattle-and-daub houses, especially if some of the deposition took place in supra-household social contexts. We may be seeing the
175
Gomolava
demarcation of space through hospitality rather than exclusion; through often inclusive practices around temporary dwellings rather than exclusive activities within them. It is notable that pits decrease in frequency at Gomolava once post-frame houses begin to be built, but do persist around them. Likewise, houses supplement rather than replace pits in many of Chapmans Early Neolithic case studies. The social change implied by increased investment in house-building relative to pit digging/filling would fit well with arguments for a region-wide shift from community to household as the primary social unit (see 2.3), apart from one small problem: the timing is wrong. The pit-site of Gomolava Ia was founded long after numerous settlements with post-frame structures. The change from pit-site to housesite must be understood in terms of individual settlement histories rather than regional trends.
6.3.5 the faunal sample 6.3.5.1 previous work

Clason reports 7010 specimens from Block I, amounting to 43% of the Vina bones from 1973. She estimates these to represent ca. 10.75% of the Vina bones that could have been collected in 1/3 of the original mound (1979b, 64). Her study continued until around 1990 (pers. comm.) without further publication save on the equids (Clason 1988). Species counts from field identifications are appended to some excavation diaries (Brukner 1980b; 1981; 1982) but are mostly presented by layer rather than phase or context.
6.3.5.2 recovery and curation

The lack of routine sieving at Gomolava is immediately apparent from the assemblage. Wetand dry-sieved samples from 1973-1976 apparently produced few additional bones (Clason 1979b, 64), implying an efficiency of hand collection not borne out by the material. Most of the units studied here were already sorted, raising the possibility of an additional curation bias: some boxes of scrap fragments were found but their numbers are suspiciously few. Figure 6.38 compares fragment size distribution with Petnica. Only diagnostic specimens are included for the latter site (see 4.3.2) to minimize curation effects. While the assemblages have the same modal size, the mean is significantly lower for Petnica (t-test or Mann-Whitney U: p < 0.0001), indicating that substantial recovery bias in Block VII did affect potentially identifiable specimens. This is considered in more detail in 7.1.
6.3.5.3 sample size and distribution

The faunal collection from Gomolava is extremely large, necessitating selection for the present study. Block VII was chosen because none of the faunal material had previously been
176
Gomolava
published. Prior sorting by species and element made sampling a major undertaking, a team of students being employed for a week to extract the Block VII material. Time constraints necessitated further reduction of the sample, so bones from the cultural layer in squares A, B, E and F were set aside. All material from discrete features was analysed. While the vast majority of diagnostic recovered material from sorted units must have been extracted and analysed, a few unsorted units are missing. In addition, a small number of specimens were taken to Groningen and have yet to be returned (W. Prummel, pers. comm.). The main Block VII sample is supplemented by age data from all ungulate mandibles collected at Gomolava between 1972 and 1982.
Gom olava
750
Frequency
500
250
0 750
Pe tnica
Frequency
500
250
50
100
150
200
250
Le ngth (m m )
Figure 6.38: size distribution of all recorded fragments from Gomolava and diagnostic specimens from Petnica.
Table 6.3 shows NSP and NISP by feature type and phase. There is remarkably little material given the area sampled. The 3790 fragments from the NE half of Block VII an area of 400 m2 compare with 46,431 from the 320 m2 trench at Opovo (Russell 1993, 185), which also had a much thinner cultural layer. Assuming a similar actual density of fragments, collection at Opovo was well over ten times as efficient. A total of 850 DZ from the NE area is much closer to the 2409 recorded at Opovo, confirming the recovery bias against smaller, nondiagnostic fragments at Gomolava . This bias presumably carries over to body size, rendering inter-site comparison of taxonomic frequencies problematic (see 5.2.3).
177
Gomolava
Given the sample structure, bones from the cultural layer dominate overall, closely followed by pits. Only one oven was recorded as containing bones, while the material listed under Structure comes from several context types related to buildings. The clear shift from pits (Ia) to cultural layer (Iab) jeopardises comparison between phases. Since this pattern holds across the entire excavated area at Gomolava it is taken as a genuine change in depositional practices rather than a sampling bias. It does reduce sample sizes for comparisons between context types within each phase, however. Identification rates are slightly lower in the cultural layer than in the pits, hinting at possible preservation and/or recovery biases between feature types, explored in Chapter 7. Table 6.4 lists specimen counts by feature, shown graphically in Figure 6.39. Given the likelihood of missing units this is intended merely to outline the structure of the sample.
Ia Cultural layer Pit Structure Oven TOTAL NSP Cultural layer Pit Structure Oven TOTAL NISP 214 2005 2219 46.7% 181 1436 1617 49.9%
Iab 1397 366 4 76 1843 38.8% 921 249 4 47 1221 37.7%
Iab/Ib 220 220 4.6% 120 120 3.7%
Ib 289 180 469 9.9% 161 121 282 8.7%
TOTAL 2120 2371 184 76 4751 1383 1685 125 47 3240 44.6% 49.9% 3.9% 1.6%
42.7% 52.0% 3.9% 1.5%
Table 6.3: numbers of fragments and of identified specimens from Gomolava by feature type and phase.
Feature House 2/80 House 4/80 House 8/80* House 10/81 Structure 11/81* Oven in C3 Pit 1 Pit 2 Pit 3 Pit 4 Pit 5 Pit 6 Pit 7 Pit 8 Pit 10 Pit 11 Pit 12 Pit 13 Pit 14
NSP 34 7 6 15 117 75 158 172 98 108 86 65 99 42 104 96 199 83 90
NISP 26 3 3 13 76 46 126 126 66 82 63 40 76 25 60 62 140 53 53
Feature Pit 15 Pit 16 Pit 17 Pit 19a Pit 19b Pit 21 Pit 22 Pit 23 Pit 24 Pit 25 Pit 26 Pit 28 Pit 29 Pit 31 Pit 32 Pit 34 Pit 35 Pit 36 Pit 37
NSP 127 43 28 1 26 56 31 54 37 46 1 27 75 20 133 32 111 94 29
NISP 97 34 27 1 19 38 17 30 33 35 1 22 52 11 93 19 89 73 22
Table 6.4: specimen counts at Gomolava by feature.
178
Gomolava
6.3.5.4 taxonomic frequencies

Summary data on the frequencies of major taxa were included in Chapter 5. They are presented here in full by NISP (Table 6.5), MNE (Table 6.6) and DZ (Table 6.7). The range of species is limited, probably due in part to poor recovery. Clason (1979b; Brukner 1981, 134) reports additional taxa including wolf, hare, marten, weasel, bear, tortoise, goose, and several species of fish, all in very small numbers. Table 6.8 compares my data with Clasons, revealing notable differences. Cattle and red deer dominate much less in Clasons sample although their relative proportions are similar to my own results while pigs and caprines are more common, dramatically so in the latter case. Figures for dogs and roe deer are fairly consistent. These discrepancies may relate to recovery and bias, although there is no reason to believe that recovery standards dropped between 1973 and 1980. In any case, the inconsistencies cannot be explained purely in terms of body size.
Figure 6.39: distribution of faunal sample across pits in Gomolava Block VII.
Assuming correct identifications on the part of both analysts and it is very difficult to mistake a pig for a cow or deer the inconsistencies must relate either to biases in terms of phasing and feature types or to genuine differences between excavation areas. Unit lists show that most material excavated in the relevant blocks during 1973 was from Gomolava Ia, while Clason (1979b, 68) reports that the fauna studied primarily derived from pits. Comparing her
179
Gomolava
data with that from Ia pits in Block VII certainly improves the match but discrepancies remain, especially for caprines and red deer. One can only assume differences between areas, providing a cautionary tale concerning extrapolation from spatially limited samples. Inconsistencies within the cattle, pig and caprine groups are more worrying, since they may reflect differences in identification techniques. As well as identifying twice as many wild cattle as Clason, I also attribute almost 9% to Bos sp.. Clason includes a category for Aurochs/Domestic ox in her table but places no specimens in it, suggesting rather less caution in identifications, as befits a more experienced analyst. On the other hand, her measurements do seem (a) clearer cut and (b) smaller on average than my own (see 6.2.2), so the difference may be genuine.
Taxon # Cattle
Domestic Wild Indet.
Ia % 41.3
35.6 2.2 3.5
Iab # 710
605 39 66
Iab/Ib % 58.1
49.5 3.2 5.4
Ib % # 185
163 6 16
Total % 65.6
57.8 2.1 5.7
# 72
68 2 2
# 1635
1411 83 141
% 50.5
43.5 2.6 4.4
668
575 36 57
60.0
56.7 1.7 1.7
Pig
332
70 194 68
20.5
4.3 12.0 4.2
176
59 84 33
14.4
4.8 6.9 2.7
31
12 10 9
25.8
10.0 8.3 7.5
35
13 14 8
12.4
4.6 5.0 2.8
574
154 302 118
17.7
4.8 9.3 3.6
Caprine
Goat Sheep Indet.
43
3 16 24
2.7
0.2 1.0 1.5
13
2 6 5
1.1
0.2 0.5 0.4
3
2 1
1.1
0.7 0.4
59
5 24 30
1.8
0.2 0.7 0.9
Red deer Roe deer Dog Equid Badger Beaver Fox TOTAL ID: Domestic Wild Indet. Roe deer/Caprine Cattle/Red deer Canid Bird Mammal Large Large-medium Medium Medium-small Small Size indet. TOTAL NON-ID:
445 86 39 2 1 1 1617 727 765 125 7 48 4 543 340 116 83 3 1 602
27.5 5.3 2.4 0.1 0.1 0.1 100.0 45.0 47.3 7.7
287 19 15 1
23.5 1.6 1.2 0.1
15 2
12.5 1.7
51 5 3
18.1 1.8 1.1
1221 692 430 99 2 34 1 585 373 122 75 12 3 623
100.0 56.7 35.2 8.1
120 82 27 11 6
100.0 68.3 22.5 9.2
282 182 76 24 2 12
100.0 64.5 27.0 8.5
798 110 59 2 1 1 1 3240 1683 1298 259 11 100 4 1 1394 901 287 181 18 2 5 1511
24.6 3.4 1.8 0.1 0.0 0.0 0.0 100.0 51.9 40.1 8.0
94 65 18 7 2 2 100
172 123 31 16 1 1 186
Table 6.5: taxonomic frequencies at Gomolava by phase NISP.
180
Gomolava
Taxon Cattle Domestic Wild Indet. Pig Domestic Wild Indet. Caprine Red deer Roe deer Dog TOTAL
Ia # % 239 37.2
204 23 31 31.8 3.6 4.8
Iab # % 254 55.9

212 19 37 46.7 4.2 8.1
Ib # % 70 59.8
61 5 8 52.1 4.3 6.8
143
42 92 26
22.3
6.5 14.3 4.0
70
28 33 18
15.4
6.2 7.3 4.0
23
10 10 6
19.7
8.5 8.5 5.1
23 164 46 27 642
3.6 25.5 7.2 4.2
8 107 8 7 454
1.8 23.6 1.8 1.5
2 22
1.7 18.8
117
Total # 563 477 47 76 236 80 135 50 33 293 54 34 1213
% 46.4 39.3 3.9 6.3 19.5 6.6 11.1 4.1 2.7 24.2 4.5 2.8
Table 6.6: frequencies of main taxa at Gomolava by phase MNE. Elements with Dobney and Reilly zones only see 4.3.3. Note that totals column shows sum of phase values, since fragments from different phases should derive from separate elements.
Taxon Cattle
Ia # 206
168 16 22
% 35.0
28.5 2.7 3.7
Iab # % 197 52.7

157 13 27 42.0 3.5 7.2
Iab / Ib % 23 52.3
21 1 1 47.7 2.3 2.3
Ib # 42
36 3 3
% 53.8
46.2 3.8 3.8
Pig
157
41 103 13
26.7
7.0 17.5 2.2
66
22 36 8
17.6
5.9 9.6 2.1
16
8 7 1
36.4
18.2 15.9 2.3
17
10 4 3
21.8
12.8 5.1 3.8
Caprine
Goat Sheep Indet.
18
2 11 5
3.1
0.3 1.9 0.8
6
5 1
1.6
1.3 0.3
2
1 1
2.6
1.3 1.3
Red deer Roe deer Dog TOTAL
141 44 23 589
23.9 7.5 3.9
82 15 8 374
21.9 4.0 2.1
4 1 44
9.1 2.3
12 1 4 78
15.4 1.3 5.1
Total # 468 382 33 53 256 81 150 25 26 2 17 7 239 60 36 1085
% 43.1 35.2 3.0 4.9 23.6 7.5 13.8 2.3 2.4 0.2 1.6 0.6 22.0 5.5 3.3
Table 6.7: frequencies of main taxa at Gomolava by phase - DZ.
Clason also identifies three goats and two sheep amongst her caprines, compared to my five goats and twenty-four sheep, although most are Ovis/Capra in both cases. Methodological biases are probably responsible: seven of my sheep identifications, for example, use mandibular criteria published since Clasons study (Halstead et al. 2002; Payne 1985). Most problematic are identifications as wild and domestic amongst pigs: Clason identifies well over half her pigs as domestic, compared to only 26% of my sample. Identifications based on metrics are in little doubt, and indeed Clason (1979b, 77-83) identifies more measured specimens as wild than domestic (137 versus 127, with 38 indeterminates). Measurement data relate overwhelmingly to adult specimens, however, while domestic pig age-at-death profiles are weighted towards immature individuals (6.3.5.5). Lower dP4
181
Gomolava
measurements (Figure 6.5) verify this, with a substantial majority apparently domestic. It is thus possible that the discrepancy between areas relates to systematic inter-analyst differences in assigning immature specimens. I was very cautious in this regard, assigning many juveniles to Sus sp., but Clason may have securely identified most immature specimens as domestic, thus avoiding the age bias. There are not enough indeterminate specimens in my sample to account for the difference alone, however, and in any case my own analysis of mandibles from Blocks I-VI found a wild:domestic ratio of 64:217, compared to 53:47 in Block VII.
Taxon Cattle
A.C. '79 (all) 38.3

37.0 1.3 0.0
Ia 41.3
35.6 2.2 3.5
D.O. Iab 58.1

49.5 3.2 5.4
Ib 65.6
57.8 2.1 5.7
Total 50.6
43.7 2.6 4.4
D.O. Ia, pits only 39.3

34.1 2.1 3.1
Pig
29.1
16.3 10.2 2.6
20.5
4.3 12.0 4.2
14.4
4.8 6.9 2.7
12.4
4.6 5.0 2.8
17.8
4.8 9.3
21.7
4.3 13.0 4.4
Caprine
Goat Sheep Indet.
8.0
0.1 0.1 7.8
2.7
0.2 1.0 1.5
1.1
0.2 0.5 0.4
1.1
0.7 0.4
3.7 1.8
0.2 0.7 0.9
2.7
0.2 1.1 1.4
Red deer Roe deer Dog Wolf Equid Badger Beaver Fox Hare Goose TOTAL ID NON-ID
17.5 4.8 1.5 0.1 0.1 0.1 0.0 0.2 0.2 0.0 2663 4367
27.5 5.3 2.4 0.1
23.5 1.6 1.2
18.1 1.8 1.1
24.7 3.2 1.8 0.1 0.0 0.0 0.0
27.6 5.6 2.7 0.1 0.1 0.1
0.1 0.1 0.1
1615 602
1215 623
282 186
3112 1411
1436 569
Table 6.8: taxonomic frequencies (%NISP) in Gomolava Block VII compared with Clason 1979. (A.C. = Anneke Clason; D.O. = David Orton)
Taxon Cattle
Cultural layer # % 868 62.5

740 47 81 53.3 3.4 5.8
Pit # % 658 39.2

571 32 55 34.0 1.9 3.3
House # 83
76 2 5
% 66.4
60.8 1.6 4.0
Pig
206
79 80 47
14.8
5.7 5.8 3.4
360
72 217 71
21.4
4.3 12.9 4.2
7
2 4 1
5.6
1.6 3.2 0.8
Caprine Red deer Roe deer Dog TOTAL
8 289 13 4 1388
0.6 20.8 0.9 0.3
49 467 93 53 1680
2.9 27.8 5.5 3.2
1 29 3 2 125
0.8 23.2 2.4 1.6
Total # 1609 1387 81 141 573 153 301 119 58 785 109 59 3193
% 50.4
43.4 2.5 4.4
17.9
4.8 9.4 3.7
1.8 24.6 3.4 1.8
Table 6.9: taxonomic frequencies of main species at Gomolava by feature type NISP.
182
Gomolava
Returning to the Block VII material, the three quantification methods give broadly similar results when frequencies of the main taxa are plotted by phase (Figure 6.40). Representation of cattle increases dramatically over time while that of red deer decreases steadily. Differences between quantification methods emerge for pigs. NISP data show a steady decline in wild pigs with little change in percentage presence of domestic pigs, while the MNE results have an increase in both populations between Iab and Ib. Quantified by DZ, wild pigs continue their decline into the final phase while their domestic counterparts increase markedly in relative frequency. Since specimens contributing to DZ counts are often measurable, the pig data in Figure 6.40c includes the most reliable attributions to population. Relying solely on DZ raises the possibility of fragmentation biases between the pig populations, but it would be hard to account for an over-representation of countable domestic specimens in these terms, and in any case this would mean a reversal of the situation in Ia. This is explored further in 7.3, but for now the trends seen in Figure 6.40c are taken to be genuine. Roe deer, caprines and dogs are present in smaller numbers and virtually disappear in the later phases. This is surprising: while caprines are never the main taxon on Vina sites north of Macedonia, they are typically present in appreciable numbers (5.2.5). Likewise, roe deer are often much more evident. Sadly, the paucity of remains from these taxa at Gomolava probably reflects poor recovery. More alarmingly, a strong differential exists between pits and the cultural layer for these small taxa (Table 6.9), raising the prospect of recovery bias between context types: it is quite feasible that pits should have been excavated more carefully than the surrounding matrix. On the other hand, domestic pigs are actually considerably less common relative to wild pigs in the pits than in the layer despite their smaller size and greater percentage of juveniles. Since there is no significant difference in fragment sizes between pits and the cultural layer (Figure 6.41; t-test: p = 0.6890; Mann-Whitney U: p = 0.2803) a genuine discrepancy in taxonomic representation between feature types is accepted as a working hypothesis. Table 6.10 compares taxonomic frequencies between Gomolava and Opovo by DZ. Cattle are overrepresented in the layer at both sites, while deer are more common in pits. Dogs and caprines show opposite patterns at the two sites, however. Russell does not provide data distinguishing wild and domestic pigs but notes that the vast majority are wild. The overrepresentation of wild pigs in pits at Gomolava is thus not matched at Opovo. Differences in taxonomic composition between feature types are potentially confounded by temporal trends, and vice-versa. Table 6.11 gives taxonomic frequencies by both feature type and phase, shown graphically in Figure 6.42. The main differences are clearly a matter of
183
Gomolava
feature type rather than phase, although a gradual decline in red deer NISP does seem to occur within the cultural layer material. Two (quantitatively) different patterns of deposition appear to apply.
70 60 a) percentage NISP 50 Cattle 40 30 20 10 0 Ia 60 50 b) percentage MNE 40 30 20 10 0 Ia 50 45 40 c) percentage DZ 35 30 25 20 15 10 5 0 Ia Iab Ib Iab Ib Iab Ib Red deer Domestic pig Wild boar
Figure 6.40: representation of domestic cattle, red deer and pigs by phase at Gomolava. Quantified by (a) NISP, (b) MNE, and (c) DZ.
184
Gomolava
Cultural layer
300
Frequency
200
100
Pit
300
Frequency
200
100
100
200
300
400
Length (mm)
Figure 6.41: distributions of fragment size in pits and the cultural layer at Gomolava.
Taxon Cattle
Cultural layer* Op. 20.2

18.8 1.4
Pit Op. 17.4

16.0 1.4
House Gom. 32.1

25.6 2.5 4.0
Gom. 57.1
46.9 3.8 6.4
Op. 10.6
10.2 0.4
Gom. 47.1
47.1 0.0 0.0
Pig
21.3
19.9
9.7 7.7 2.4
20.3
27.4
6.0 19.1 2.3
17.3
11.8
5.9 5.9 0.0
Caprine
Goat Sheep Indet.
10.4
0.6 3.1 6.7
0.9
0.0 0.2 0.7
4.5
0.8 2.0 1.7
3.5
0.3 2.7 0.5
3.1
0.0 1.8 1.3
2.9
0.0 0.0 2.9
Red deer Roe deer Dog TOTAL N
30.3 11.2 4.2 348
19.7 2.0 0.4 452
34.2 18.1 3.2 776
23.2 8.5 5.2 598
45.3 14.2 4.0 213
29.4 0.0 8.8 34
Table 6.10: frequency (%DZ) of main taxa at Gomolava and Opovo by feature type. *cultural layer at Gomolava compared with midden at Opovo (Russell 1993, 187).
185
Gomolava
This does not necessarily undermine the evidence for temporal trends, however, since the shift from pits to cultural layer is probably not a sampling bias: pits almost disappear across the whole excavated area from Gomolava Iab onwards. Since the faunal composition of the cultural layer does not seem to change in response to this, three possible scenarios can be set out: The apparent absence of pits in later phases is a sampling bias, with a concentration of pits existing in the unexcavated portion. This is unlikely given the extent of excavations and a paucity of evidence for spatial differentiation on other Vina sites. Deposition in pits virtually ceases but the underlying practices continue in modified form, a portion of the fauna now being deposited off-site. No sudden shift is thus seen in taxonomic frequencies within the cultural layer, which in fact become even less like those seen in pits. The practices underlying deposition in pits cease (almost) altogether and the overall frequency of each taxon shifts accordingly, implying that the mode of deposition of animal remains is bound up with the decisions involved in hunting/slaughter. The mechanism linking the two remains to be seen, with one possibility involving Russells feasting hypothesis.
Taxon Ia # Cattle
Pit Iab % 39.5

34.2 2.1 3.1 490 30 45
Cultural layer Ia % 93
81 2 10
Iab % 56.9
47.0 3.3 6.6
Iab/Ib % 63.8
54.0 3.8 6.1
Ib % # 103
88 4 11
# 103
85 6 12
# 588
497 35 56
# 72
68 2 2
% 64.0
54.7 2.5 6.8
565
37.5
32.7 0.8 4.0
60.0
56.7 1.7 1.7
Pig
311
62 186 63
21.7
4.3 13.0 4.4
49
10 31 8
19.8
4.0 12.5 3.2
21
8 8 5
11.6
4.4 4.4 2.8
122
47 51 24
13.2
5.1 5.5 2.6
31
12 10 9
25.8
10.0 8.3 7.5
30
12 10 8
18.6
7.5 6.2 5.0
Caprines
Goat Sheep Indet.
39
3 16 20
2.7
0.2 1.1 1.4
10
6 4
4.0
2.4 1.6
2.2
2
2
0.2
0.2
2
2
1.2
1.2
2.2
Red deer Roe deer Dog TOTAL
397 81 39 1432
27.7 5.7 2.7
70 12 14 248
28.2 4.8 5.6
48 5 181
26.5 2.8
202 6 1 921
21.9 0.7 0.1
15 2 120
12.5 1.7
23 2 1 161
14.3 1.2 0.6
Table 6.11: taxonomic frequency (%NISP) at Gomolava by phase and feature type.
186
Gomolava
Pits
70 60 50 Percentage NISP 40 30 20 10 0
Cultural layer
Ia
Cattle
Iab
Red deer
Ia
Iab
Domestic pig
Iab/Ib
Wild boar
Ib
Figure 6.42: taxonomic frequency (%NISP) of major species at Gomolava by phase and feature type. Values are percentages of the total for these four taxa only.
6.3.5.5 age-at-death: mandibles

Apart from the main Block VII sample, tooth eruption and wear states were recorded on all Vina cattle, pig and red deer mandibles from Blocks I-VII with two or more molariform teeth present, excluding (d)P2. Unfortunately, caprine and roe deer specimens could not be found for the 1970s seasons. The large samples thus assembled make sub-group comparisons worthwhile, and these are aided by 2-sample Kolmogorov-Smirnov (K-S) tests. Cattle Table 6.12 shows the distribution of domestic cattle mandibles across Paynes (1973) relative age stages. Counts are corrected by allocating partially aged specimens to their possible stages pro rata as proposed by Payne. This increases sample size while reducing the problem of some stages being more easily identified than others, at the risk of offending the readers sensibilities by dealing in fractions of specimens. The mortality distributions before and after correction (Figure 6.43) are the same overall shape but noticeably distinct. The difference is arguably statistically significant (K-S best-case: 0.05 > p > 0.025; worst-case: p > 0.1; see 4.3.4.1 for explanation), and correction is applied in all subsequent cases. Since the lengths of stages vary, these data must be treated with caution. The absence of specimens at stage A is at least partially taphonomic, although it is also a very short stage. Conversely, the apparent peak at G-H is largely maybe entirely an artefact of these stages
187
Gomolava
long duration. Figure 6.44 shows the survivorship curve for domestic cattle scaled according to Halsteads (1985, 219) suggested ages. Scaling flattens out the peak at C/D these are relatively long stages compared to B and E and instead shows a steady decline through the first three years, with 30% of individuals apparently surviving to adulthood. This is probably inflated somewhat by age-correlated preservation (Munson & Garniewicz 2003), but is nonetheless surprisingly high.
Corrected count Stage A B C D E F G H I Total 123 Count 0 18 30 17 6 2 29 18 3 0.0% 14.6% 24.4% 13.8% 4.9% 1.6% 23.6% 14.6% 2.4% 1st 0 18 33 24 6 6 2 33 8 34 5 8 177 8.95 194 8.95 194 4.6% 42.11 50.33 25.9% 38.63 17 46.18 23.8% 2nd 0 18 33 24 6 6.31 Final count 0 18 33 24 6 7.54 0.0% 9.3% 17.0% 12.4% 3.1% 3.9%
Table 6.12: raw and corrected distribution of Payne age stages for domestic cattle at Gomolava.
30% Uncorrected (n = 123) Percentage frequency 25% 20% 15% 10% 5% 0% A B C D E F Payne stage G H I Corrected (n = 194)
Figure 6.43: raw and corrected distribution of Payne age stages for domestic cattle at Gomolava.
188
Gomolava
100 80 60 40 20 0 0 6 12 18 24 30 36 Suggested age (months) 42 Young 54 48 adult 60 Adult 66 72 78 84 90 Old adult Senile
Percentage survival
Figure 6.44: survivorship curve for domestic cattle at Gomolava. Based on Payne stages, with A-E scaled according to Halsteads suggested ages.
Figure 6.45 compares survivorship for assemblage sub-groups. Un-scaled curves are used for simplicity and because they reflect the ordinal nature of the K-S tests. Raw data for this and subsequent survival curves are given in Appendix 3. No significant differences exist (p > 0.1 in all cases, even assuming independence) between phases, feature types or blocks. This consistency in age structure is re-assuring considering the limited excavations from which most other samples are drawn. Aurochsen Figure 6.46 shows aurochs mortality at Gomolava. The sample is rather small but the pattern is striking. Only two specimens are prime-aged, with the bulk falling in Halsteads 1-8 months, old adult or senile categories. This suggests opportunistic hunting of weaker and/or slower individuals, in keeping with the low overall numbers of aurochsen. Pigs Unsurprisingly, there is a significant difference in survivorship between wild and domestic pigs (K-S best-case: p < 0.001; worst-case: 0.001 < p < 0.005), with the former typically older at death. Figure 6.47 shows the survivorship curves scaled by suggested age following Hambleton (1999, 65). Such curves have little meaning for hunted populations since they rest on the assumption that the recovered sample represents the entire death-population, but are convenient for comparative purposes. Given that wild pigs appear to reach dental maturity later than domestic ones (Bull & Payne 1982, 56) the difference between populations is probably greater than it appears. The vast majority of domestic pigs died within the first two years.
189
Gomolava
a)
100% 80%
Ia (n = 97) Iab (n = 67)
Survival
60% 40% 20% 0% A B C D E Payne stage F G H I
b)
100% 80%
Cultural layer (n = 142) Pits (n = 48)
Survival
c)
100% 80%
Blocks I-VI (n = 161) Block VII (n = 33)
Survival
Figure 6.45: domestic cattle survivorship curves from Gomolava by (a) phase, (b) feature type and (c) excavation area.
190
Gomolava
Stage 5 4 Frequency 3 2 1 0 A B C D E Payne stage F G H I I Total 11 A B C D E F G H
Count 0 3 0 0 0 1 1 4 2 0.0% 27.3% 0.0% 0.0% 0.0% 9.1% 9.1% 36.4% 18.2%
Figure 6.46: aurochs mortality at Gomolava.
100% Percentage survival 80% 60% 40% 20% 0% 0 10 20 30 40 Adult 50 Suggested age (months)
Domestic (n = 229) Wild (n = 74)
Old 60 adult
Senile 70
Figure 6.47: survivorship curves for wild and domestic pigs from Gomolava. Stages A-F scaled according to Hambletons suggested ages.
Figure 6.48 compares sub-groups amongst domestic pigs. The curves are remarkably consistent between phases and feature types, but an arguably significant difference is seen between areas (K-S best-case: 0.01 < p < 0.025; worst-case: p > 0.1). The mandibles from Block VII are typically older, a phenomenon which might help to explain the higher percentage of wild pigs in my data than in Clasons (6.3.5.4). A taphonomic bias is possible, but preservation would be more likely to vary between pits and the cultural layer, where no difference is seen.
191
Gomolava
Figure 6.49 shows the equivalent data for wild pigs. No significant differences are seen (p > 0.05 in all cases, even assuming independence) but this is partly due to lower sample sizes. Interestingly, given the results for domestic pigs, there is virtually no difference between excavation areas. Specimens from the pits are typically older than those from the cultural layer, although a larger sample would be required to state this with confidence. Red deer Age attribution in deer is hampered by variability amongst comparative data and the absence of standard recording schemes, so only broad relative age classes are used here. Clason (1979b, 69) simply states a ratio of 1:3 for animals younger than versus older than two years at Gomolava. Figure 6.50 shows the proportions of relative age categories from my own work by phase and feature type. Immature (infantile/juvenile) specimens are proportionally more common in pits than in the cultural layer. This may be an artefact of differential preservation but is not significant using either K-S or 2 (subadult omitted) even ignoring the possibility of pairing. Surprisingly, this is also true of an apparent decrease in mature specimens between Gomolava Ia and Iab. Roe deer Ageing roe deer is particularly problematic since eruption is completed within the first year of life (Aitken 1975, 17) and beyond this point no wear-scoring system is well established. Table 6.13 shows numbers of specimens assigned to immature (< 1 year), young adult (1-2 years) and adult (3+ years), primarily following Aitken (1975). The vast majority are at least dentally mature. Clason (1979b, 69) notes a similar ratio of mature to immature specimens from her 1973 sample.
Pits Immature Young adult Adult Cultural layer Immature Young adult Adult Total Immature Young adult Adult
Ia 2 5 5 2 2 4 5 7
Iab 1 1 1 5 3 1 7 4
Ib
All 2 6 6 3 8 7 5 16 13
1 2
1 2
Table 6.13: frequencies of immature and mature roe deer mandibles at Gomolava by feature type and phase.
192
Gomolava
a)
100% 80%
Ia (n = 76) Iab (n = 91) Ib (n = 28)
Survival
60% 40% 20% 0% A B C Payne stage D E
b)
100% 80%
Survival
c)
100% 80%
Survival
Figure 6.48: domestic pig survivorship curves from Gomolava by (a) phase, (b) feature type and (c) excavation area.
193
Gomolava
a)
100% 80%
Ia (n = 30) Iab (n = 32)
Survival
b)
100% 80%
Survival
c)
100% 80%
Survival
Figure 6.49: wild pig survivorship curves from Gomolava by (a) phase, (b) feature type and (c) excavation area.
194
Gomolava
a) 60%
Ia (n = 37) 50% 40%
Frequency
Iab (n = 27) Ib (n = 7)
30% 20% 10% 0% Infantile Juvenile Subadult Adult
b) 50%
Pits (n = 25) 40%
Frequency
Cultural layer (n = 47)
30% 20% 10% 0% Infantile Juvenile Subadult Adult
Figure 6.50: red deer mortality at Gomolava by (a) phase and (b) feature type. Counts have been corrected using the same principle as for Payne stages among cattle.
Caprines Only seven ageable caprine mandibles were recovered from Block VII, of which five are sheep and two indeterminate. In addition, two loose sheep M3s were found. This sample is too small to produce meaningful mortality/survivorship curves, but Table 6.14 shows age estimates based on eruption (Silver 1969), eruption/wear (Jones; Payne 1973) and radiographic development (Orton 2003). The latter are given as point estimates and as approximate one-sigma confidence intervals. As far as can be detected from nine specimens, mortality appears fairly evenly spread across age categories, with the absence of very young specimens probably attributable to preservation.
6.3.5.6 age-at-death: epiphyseal fusion

Epiphyseal fusion is generally of less use for ageing than dental development due to typically lower preservation potential of younger post-cranial bones. It is nonetheless worth considering as a check on the cranial data. Table 6.15 shows fusion data for cattle and pigs from Gomolava. The figures for cattle suggest a very low level of juvenile mortality, with
195
Gomolava
over 83% of individuals apparently surviving to maturity (the final group, vertebrae, is omitted since these were not normally identified to species). Combining fusion ages (Silver 1969, 285-286) with Halsteads suggested ages for mandibles, this should be roughly equivalent to the proportion reaching stage F, but the latter figure is only 58.2%. The discrepancy can presumably be put down to preservation bias, mandibles being relatively resistant to age-mediated destruction.
Payne (1973) Phase Ia Ia Ia Ia Ia Ia Iab Iab Iab Feature Pit 2 Pit 4 Pit 2 Pit 35 Pit 14 Pit 2 Pit 5 Pit 1 Pit 1 Species Sheep Sheep/goat Sheep Sheep Sheep Sheep Sheep/goat Sheep Sheep Silver (1969) 3-24 9-24 18+ 21+ 21+ 18+ 18+ 18+ 18+ Stage C+ D+ G D+ D+ H E G H Suggested age 6+ 12+ 48-72 12+ 12+ 72-96 24-36 48-72 72-96 Gb
Jones (2006) Stage 3+ 10+ 54-(108) 10+ 10+ 84+ E3+ Ga 22-36 54-(78) 84+ Age*
Orton (2003)** Estimate 7 21 38 44 56 1 range 3-10 18-23 37-41 41-47 52-60
(loose M3) 31 27-35 33 29-37 (loose M3)
Table 6.14: age attributions for caprine mandibles from Gomolava Block VII. All ages in months. * Figures in brackets are estimated upper limits. ** Point estimates and confidence limits rounded to nearest month.
Greater discrepancy is seen between cranial and post-cranial ageing evidence for domestic pigs: fusion suggests that 50% of pigs survive to around 2 years, yet only 3.2% of mandibles are at stage E suggested as 21-27 months or beyond. Likewise, 100% of group 2 elements from wild pigs are fused but only around 40% of mandibles are at E or beyond, with 15% reaching F. For both populations, the percentage fusion in group 2 is actually greater than that in group 1, presumably reflecting unreliably small samples rather than misleading fusion data or differential depositional processes. The difference between dental and fusion data is clear however, and almost certainly results from greater age-correlated destruction amongst postcranial material. This is particularly problematic for the pigs since both dental and epiphyseal data show very different mortality structures for the wild and domestic populations. If there is heavy selection against immature specimens then this must disproportionately have affected the domestic sample, inflating the apparent importance of wild specimens. At the same time, mature specimens are more likely to be positively identified as wild or domestic only 2 of 81 fused specimens in the table were assigned to Sus sp. compared with 9 out of 32 unfused. Other criteria including antler/horn core development and cranial sutures have been used to age individual specimens, but these results are essentially anecdotal and reveal little about the age structure of the sample.
196
Gomolava
Cattle
Group 1 2 3 Group 1 2 3 4 Group 1 2 3 4 Group 1 2 3 4
Unfused 1 16 14 Unfused 7 4 7 Unfused 1 0 1 3 Unfused 12 4 1 15
Fusing 2 3 7 Fusing 0 2 0 Fusing 0 2 0 4 Fusing 0 6 0 4
Fused 133 105 64 Fused 5 2 0 Fused 29 16 11 16 Fused 35 18 11 17
% fused* 99.3% 87.1% 83.5% % fused* 41.7% 50.0% 0.0% % fused* 96.7% 100.0% 91.7% 87.0% % fused* 74.5% 85.7% 91.7% 58.3%
Domestic pigs
Wild pigs
All pigs
Table 6.15: fusion data for cattle and pigs at Gomolava. Groupings follow OConnor (2003, 167) for cattle and Bull and Payne (1982, 67) for pigs. * includes fusing specimens.
6.3.5.7 sex
Table 6.16 shows all sex determinations from Block VII, plus those on pig mandibles from other areas. Biometric sex attributions were made very cautiously and are only given for red deer. The results thus rely on canines for the pigs and pelvises for the remaining species, while a few deer crania could be sexed on the basis of presence or absence of antlers. Antlers are otherwise excluded since (a) they have no female equivalent and (b) they could be collected. Sex determinations from horn core morphology are not considered reliable for prehistoric populations (see 6.2.2.2). Based on pelvises alone, domestic cattle appear just over two-thirds female, while the tiny sample of sexable aurochsen is mostly male. Domestic pigs are similarly biased towards females, with wild individuals split fairly evenly between sexes. The predominance of females amongst domestic pigs and cattle is predictable given that neither species can be sexed reliably unless relatively mature. The sex ratio amongst red deer is remarkably even, while the three sexable roe deer specimens were all male.
197
Gomolava
F Cattle Domestic Wild Indet Pig Domestic Wild Indet Caprine Red deer Roe deer
Pelvis M 17 1 2 7 4 1
Mandible F M
Maxilla F M
Cranium F M
Biometry M
Total M 7 4 1 10 15 1 28 3
17 1 2 32 9 5 9 10 0 4 1 2 1 5 1 0 0 2 2 22 18 36 10 7 27 0
1 5 0
0 8 1
Table 6.16: sex determinations at Gomolava. Note that data for pigs include Block I-VI material.
Tabulating sex frequencies by feature type (Table 6.17) reveals an interesting pattern for pigs. Taking both populations together, males are clearly more common in pits and females in the cultural layer. On closer inspection it seems that wild females and domestic males i.e. midsized individuals are fairly evenly represented, although the samples are small, while the (large) wild males are mainly from pits and the (small) domestic females are overwhelmingly found in the layer. Put another way, a pig specimen from a pit at Gomolava is more likely to be wild than domestic, and in either case is more likely to be male than female. It seems that sheer size rather than either sex or domestication status alone is the main factor here. This is supported by post-cranial log-ratio data (Figure 6.51), and by the tendency for older pigs in pits (see Figure 6.49b). The meaning of this pattern is considered in 7.7. No such size trend is seen for cattle (Table 6.17; Figure 6.52) or for red deer.
Wild Male Female Domestic Male Female Total Male Female
Pigs Cattle Pit Layer Wild Pit Layer 11 3 Male 0 4 5 5 Female 0 1 Pit Layer Domestic Pit Layer 5 4 Male 0 7 4 32 Female 3 14 Pit Layer Total Pit Layer 17 7 Male 0 11 12 40 Female 3 15
Table 6.17: distribution of sexes amongst pigs and cattle at Gomolava by main feature types.
198
Gomolava
Cultural layer
Percentage frequency
40% 30% 20% 10%
Pit
40% 30% 20% 10% -1.0 -0.5 Mean log-ratio value 0.0 0.5
Figure 6.51: distribution of pig postcranial log-ratio values by main feature types at Gomolava.
Cultural layer
25% 20% 15% 10% 5%
Pit
25% 20% 15% 10% 5% -1.0 -0.5 Mean log -ratio value 0.0 0.5
Figure 6.52: distribution of cattle postcranial log-ratio values by main feature types at Gomolava.
6.3.5.8 seasonality
The difficulty of determining seasonality of occupation is well known (e.g. Milner 2005; Monks 1981). Balkan NCA tells such as Gomolava are reasonably assumed to represent longterm year-round occupation (but see Bailey 1999b), although this does not preclude mobility by group segments or individuals on a seasonal or other basis (Halstead 2005, 39). The aim here is rather to detect seasonality of activities.
199
Gomolava
Cattle Estimating season of death is difficult for cattle since (a) the standard eruption/wear schemes achieve low-resolution even in the first year, and (b) birth is not strictly seasonal (4.3.4.2). Twenty-one Gomolava specimens are at Payne stage B, but given a main calving season of March-May and Halsteads suggested age of 1-8 months, their deaths can only be narrowed down to April-January. A range of Grant wear states is seen within this group, however, with dP4 ranging from d to ja. The latter is a long-lived stage, persisting well into the wear of M1 and even M2. The three specimens at this point but with M1 still erupting can be placed at the very end of stage B, i.e. in late autumn or winter. The bulk of the jaws at stage B shows moderate dP4 wear, suggesting death between late summer and early winter, but specimens younger than this are conspicuously absent. This may be due to poor preservation of infantile specimens or may indicate that very few specimens were killed in the first few months of life. The earliest nine stage C specimens with M1 at a or b probably died during the first winter, but M1 Grant states are evenly distributed beyond this point. No further inferences are made save for the observation that a lack of clustering implies slaughter was not strongly seasonal, at least amongst young animals. Eight cattle mandibles from Gomolava were radiographed and scored following Orton (2004, see 4.3.4.2, Appendix 3.2). In seven cases the age brackets assigned are small enough to suggest seasons of death, and these cover the entire first year of life. Deer Only ten red deer mandibles provided estimates for season of death, three of these using radiographs and data from Carter (2006) and the remainder based on eruption timings. These cover the whole year as evenly as can be expected from such a small sample, giving no indication of a seasonal focus for hunting. Of five roe deer specimens with incompletely erupted dentition, three were radiographed and scored following Carter, giving two estimates of 8-9 months and one of 8-10. The remaining two specimens are closely consistent with these in terms of eruption, Brown and Chapmans (1991b) wear scores and M1 crown height, and are assumed to be approximately the same age. There follows a conspicuous gap, with no specimens featuring M3 in late eruption or early wear the lowest Brown and Chapman score assigned to a third molar was 16 then a group of fifteen individuals roughly matching published descriptions of yearlings (1975; Legge & Rowley-Conwy 1988, 24). Radiographs from six of these placed one at around 14 months with the others estimated variously at 13+ and 14+. Most of these yearlings are rather more worn on the molars than both the ~14 month specimen and Aitkens November kills, raising the possibility that the bulk of this group at Gomolava are around a year older
200
Gomolava
than the 8-10 month cluster. Wear on the third cusp of M3 in most cases suggests animals slightly older an average than Legges and Rowley-Conwys (1988, 25) 17-19 month group. This would imply a broad concentration of roe deer hunting in late winter/early spring, with the ~14 month specimen an outlier. Antlers provide problematic seasonality evidence due to their utility as a raw material whether the aim is to determine the seasonality of hunting or of deposition and only crania with antlers clearly shed are reliable (4.3.4.2). Only one such red deer frontal was recorded in the Block VII sample, compared to six unshed antlers. Since only three red deer antler specimens were definitely shed, it seems unlikely that many of the unshed fragments were collected from chance finds of carcasses rather than taken from hunted individuals. Russell (1993, 381) notes that roe deer antler was rarely worked at Opovo, using this to justify its tentative use for seasonality. This argument holds at Gomolava, where no worked specimens were found, and is supported by the fact that six out of ten specimens are unshed, with three indeterminate and only one demonstrably shed collection of loose antlers would appear to have been rare. A single cranium with shed antlers was recovered. A shed:unshed ratio of 1:6 amongst crania (1:3 even including the shed specimen) suggests a MarchNovember period of roe deer hunting that is at odds with the more reliable mandibular data. Pigs The birthing season amongst pigs is not technically restricted but births are typically concentrated in the spring, 90% falling between March and May in one study of German wild pigs (Briedermann 1990). The ten domestic and one wild pig at Paynes stage A would thus represent late March-early July deaths using Hambletons (1999, 65) suggested ages. The 56 domestic specimens at stage B should have died between May and December, while the 22 wild specimens are likely to be slightly older an average since eruption appears to be accelerated in the domestic populations on which the suggested ages are based (see Bull & Payne 1982, 56). Both groups show apparent continuity within stage B as indicated by the wear of dP4, and the same can be said for the succeeding stage C, theoretically corresponding to a ten-month period from October to July. Beyond this point Paynes stages are not useful for estimating season of death. At this resolution there is no hint of seasonality in either population. Figure 6.53 shows Grants (1982) MWS stages in the format advocated by Ervynck (1997; 2005). Apparent mortality peaks are seen for domestic pigs at MWS 3, 8, 17, 26 and 31. Ervynck (1997) has interpreted a similar set of peaks at Wellin at MWS 2, 10, 19 and 28 as a group of very young piglets followed by four seasonal culls.
201
Gomolava
The emergence of a similar pattern among the wild and domestic pigs is suspicious, raising two worrying possibilities. Firstly, that these might not be separate populations at all, but perhaps just males and females. While some misidentifications are probably inevitable in an assemblage of this size, the biometry is hard to refute overall (6.2.1.4). Secondly, the peaks and troughs might be an artefact of unequal MWS lengths (OConnor 2003, 158-159). This possibility has been explored thoroughly and rejected by Ervynck (2005). The moving average should reduce the impact of variation between individual stages, but a general slowdown once each tooth comes into full wear is more problematic. On the basis of comparison between MWS and individual tooth scores, Ervynck argues that the combination of molars in play at a given time avoids the slow-down effect. His graph is reproduced with Gomolava data as Figure 6.54, confirming that there are no points when multiple teeth are in their most rapid stage of development. The first three troughs do each correspond to the period of eruption for one molar, however. The peaks do not clearly correspond to seasons if cross-referenced with Payne stages. Using Magnells (2006) comparative data for wild pigs, however, the first two lie within 6-12 months and 16-24 months respectively and may well thus be seasonal (autumn or winter). The later peaks overlap substantially but certainly do not preclude a seasonal hypothesis. The pattern seen for wild pigs is similar, but given the smaller sample size must be treated with caution. It would of course be extremely interesting if both populations were typically killed in the same season, as appears to be the case. Caprines Of the caprines with radiographic age estimates in Table 6.14, only those under three years are reliable enough for estimation of season. Since there are only three of these, nothing can be said about the seasonality of caprine herding at Gomolava. Seasonality by feature Given large samples and tight season estimates, inferences can be made about the seasonality of deposition within particular features, although this can never be certain. The caution with which seasons have been estimated here precludes useful results from such analysis; most estimates of season at death are greater than six months in duration, and no feature contains more than six such estimates.
202
Gomolava
15
a)
10 5
Frequency
0 5
10 15 20
Moving average Raw data 10 20 30 40 47
b)
2
Frequency
Moving average Raw data

6
10
20
30
40
47
Figure 6.53: Grant (1982) MWS scores for (a) domestic and (b) wild pigs at Gomolava. Data smoothed using a three-point moving average (following Ervynck 1997; 2005).
20 18 16 14 TWS 12 10 8 6 4 2 0 0 10 20 MWS 30 40 50
M1 M2 M3
Figure 6.54: mandible wear scores (MWS) against individual tooth wear scores (TWS) for pigs at Gomolava. Peak MWS scores are shown by blue dotted lines; troughs by red.
203
Petnica
6.4 Petnica
6.4.1 introduction
Petnica is a small site in NW Serbia, around 7km from Valjevo at the head of a small valley (Figure 6.55). The site is situated at the foot of a north-facing cliff, just outside two wellknown caves. A small stream, the Banja, issues from the lower of these and passes the site, eventually joining the Kolubara. Palaeolithic remains are reported from the caves (Vlahovi & Miloevi 1971, although see Greenfield & Je n.d., 7), but for present purposes Petnica refers to the open-air site, Petnica naselje ispred male peine (Figure 6.56). The site is much smaller than Gomolava, covering an estimated 1ha (Greenfield 1986, 111). In contrast to the flat plains of Srem, Petnica is on a moderate slope, surrounded by cliff-face to the south and rolling hills to the north. It lies on the boundary between the karstic uplands which rise into central Bosnia and much gentler terrain sloping down to the Kolubara valley and eventually to the Pannonian plain (Starovi, pers. comm.; Greenfield 1986, 111). There are thus a wide variety of ecotopes in the surrounding area. Petnica is a mid-altitude site in Arnolds and Greenfields (2004) classification. Surface survey suggests that it was the only major middle/late Neolithic site in the Banja valley (Greenfield & Je n.d., 6). .
Figure 6.55: location of Petnica. Aerial photo: ISP
204
Petnica
Figure 6.56: location of trenches and approximate extent of settlement at Petnica. (A. Starovi/ISP)
Excavations were very limited compared to Gomolava. Early sondages confirmed the presence of a Vina settlement, although few records survive. Systematic excavations by eljko Je during the 1980s were continued by Andrej Starovi in 1991, 1997 and 1999. Since this was a training project for students of Petnica Science Centre (ISP), excavation proceeded unusually slowly and carefully. Figure 6.57 shows the layout of trenches. The current study is largely restricted to Trenches 10 and 11. No definitive publication exists. Summary accounts of the earlier seasons are available (Greenfield 1986, 111-115; Greenfield & Je n.d.; Je 1985), and a student article from ISP (Stevanovi & Mihaljevi 1996) includes basic descriptions of two houses. The present study relies primarily on archived material and personal communication with Andrej Starovi.
6.4.2 phases
Three Vina occupation phases have been identified and are labelled Petnica 1-3. These are overlain and often disturbed by Eneolithic (Petnica 4), Early Iron Age (Petnica 5) and Late Roman (Petnica 6) deposits (Figure 6.58). All layers include considerable slope-wash from the hill immediately to the south, including some large rocks visible in the sections.
205
Petnica
Figure 6.57: trench layout at Petnica. Trench 5 is incorporated into Trench 1. The exact location of Trench 12 (1999; 2 x 2 m) is unknown to the author.
Figure 6.58: sample section through deposits at Petnica. (A. Starovi/ISP)
Petnica 1 belongs typologically to Vina B2. Remains of two structures are known but neither has been extensively uncovered: excavation ceased just after House 1 was encountered in Trench 10, while House 4 lies mostly outside the excavated area (Figure 6.59a). A bucranium was found around 50 cm outside House 1 in Trench 8, on a burnt packed-earth surface. This surface, described as a large open-air food-processing and activity area (Greenfield & Je n.d., 12), extends through most of Trenches 1, 7 and 8, featuring pits, pyro-features, numerous artefacts and a remarkable concentration of over 300 bone fragments, close to the bucranium. Je (1985, 48) suggests that a pit in Trench 1 containing a hearth may be part of a zemunica.
206
Petnica
Figure 6.59: outline plan of Petnica by phase. Based on diaries, photographs, discussion with A. Starovi, and Stevanovi & Mihaljevi 1996.
207
Petnica
Petnica 2 is attributed to Vina C. Patches of floor and collapsed walls in Trenches 3, 4 and 6 were initially taken as a row of houses (Greenfield & Je n.d., 14) but later reinterpreted as badly-preserved parts of House 2 (Starovi, pers. comm., Stevanovi & Mihaljevi 1996, 1112, Figure 6.59b). This structures outline is anything but well-defined, with further patches of floor and wall in Trenches 1 and 2. A surface in Trench 10 assigned to Petnica 2 features a well-preserved outdoor oven (pe) with associated pit (Pit 3), a lithic workshop and a single copper bead. A ruminant bone close to the latter provides the only radiocarbon date, calibrated to 5060-4940 BC (68.2% HPD. Bori & Jovanovi in press, Figure 6.60). Petnica 3 Vina D1 is dominated by House 3, covering Trench 10 and parts of 1, 2, 3, 7, 8, 9 and 11. The phase is often disturbed by Eneolithic activity, but its interface with Petnica 2 is relatively clear. Typologically, the three phases should be broadly comparable with Gomolava Ia, Iab and Ib.
Figure 6.60: AMS date from Petnica 2.
6.4.3 features
The anatomy of the Vina layers at Petnica is rather different to Gomolava. The cultural layer is much thinner, if of similar composition, and shows relatively clear boundaries between phases 1-3. The same broad arguments regarding depositional practices apply, however. Bones appear much more densely concentrated at Petnica, but this may relate entirely to recovery and curation (see 6.4.4.2). The excavated area at Petnica does not feature pits comparable with those at Gomolava, apart from the possible zemunica in Trench 1. Of two pits within Trench 10, both are less than 1 m
208
Petnica
in diameter. Storage is the working hypothesis for Pit 2, dug from floor level within House 3. Pit 3 may simply be an unusually deep ash-pit pepelite serving the oven, although it was not ultimately filled with ash. Outdoor storage pits from Petnica 1 containing charred cereal grains are mentioned by Greenfield and Je (n.d., 12). House 2 is outside the scope of this study, while Houses 1 and 4 were barely excavated. House 3 is extremely large an estimated 20 x 11 m (Stevanovi & Mihaljevi 1996, 12) finding a parallel in Banjica House 4 (Markoti 1984, 93). At Gomolava only Houses 1/73 and 3/73 are clearly comparable, although the largest Ib buildings may be a similar size if remains from Block V and the 1957 area are from the same structures (see 6.3.4.3, Brukner 1988, Abb. III). The building is poorly preserved but the substantial nature of its construction is apparent from the sheer quantity of burnt rubble. A central row of at least four posts apparently provided support for the roof. There may have been a small porch area at the east end, and much of the surrounding area had a pavement of river cobbles. A cluster of loomweights with associated small postholes has been interpreted as remains of an in situ loom (Starovi, pers. comm., 1997, 43). As with all buildings at Petnica, House 3 seems to have been destroyed by fire. Excavation units from among the collapsed walls are of limited use since most finds presumably post-date destruction. There is substantial Eneolithic intrusion in places, but in most cases all finds relate to Petnica 3, suggesting that Vina-period occupation did not end with the destruction of the house. Units from the house base are problematic since material may derive from the floor or simply be part of the make-up. The uppermost layer of floor remains immediately beneath the wall rubble including internal features such as ceramic clusters and the loom, and some units specifically noted to be trapped between walls and floor is distinguished from the base proper for analytical purposes. No oven or hearth associated with House 3 was found in Trenches 10 or 11, and none is reported from earlier seasons. Sizable portions remain unexcavated, however. The substantially built and well-preserved outdoor oven from Phase 2 (Figure 6.61) is suggested to have served House 2 (A. Starovi, pers. comm.), and there is also a hearth 5 m SW of this. Neither contained bones apart from four un-diagnostic fragments in the oven substructure.
209
Petnica
Figure 6.61: partially excavated oven from Petnica 2, Trench 10. Pit 3 in foreground on left (photos: ISP).
6.4.4 the faunal sample 6.4.4.1 previous work

Haskel Greenfield studied faunal remains from the 1980-1986 seasons, publishing those from 1982 (Greenfield 1986). This material derives from open air middens and pits primarily the former (Greenfield 1986, 63-64; 1988, 577) within Trenches 1, 3 and 4. Some 2377 Vina specimens are reported, of which 571 were identified to species. No detailed contextual or taphonomic information is available.
6.4.4.2 recovery and curation

Recovery standards were much higher than at Gomolava, with excavation primarily by trowel. Most 1991 units were sieved (10 mm) and the fractions bagged separately, revealing that sieving increased yields by 74% overall and 38% for diagnostic specimens (Table 6.18). Comparison with other units in which fractions were combined indicates that sieving was common but not ubiquitous. The percentage of specimens considered diagnostic falls mid-way between the total for known sieved units and the hand-collected fraction alone (Table 6.19). All material was curated at ISP. Units relate primarily to arbitrary layers and 1 m squares but also respect stratigraphic features, such that a single arbitrary block may have several units. Unit, OS and square
210
Petnica
numbers were copied from the bags. Stratigraphic information was subsequently entered by A. Starovi for most units, and extracted from diaries and other documentation by the author for the remainder.
Recovery Unknown Hand Sieve
'Undiagnostic' (%) < 20 mm 20-50 mm 1.1 34.4 2.7 25.7 6.1 57.5
'Diagnostic' (%) 64.5 71.6 36.4
Table 6.18: sieving yields from Petnica 1991 (where known).
Known sieved units
Hand Sieve Total
All other units
Unlabelled (%) < 20 mm 20-50 mm 2.7 25.7 6.2 57.2 4.2 39.1 1.1 34.4
Labelled (%) 71.6 36.6 56.7 64.5
Table 6.19: comparison of sieved, hand-collected and unknown-recovery material at Petnica.
6.4.4.3 sample size and distribution

This study includes all material from the last four seasons (1989, 1991, 1997, 1999), plus a portion of the 1988 assemblage. Table 6.20 presents the sample by phase. The vast majority including 10,439 Vina specimens derives from Trenches 10 and 11. Controlling for area, sample size falls between Gomolava and Opovo (Table 6.21). Recovery should if anything be better at Petnica than Opovo where < 50% of units were sieved (Russell 1993, 71) while occupation was probably briefer at the latter site, suggesting a genuine difference in density of faunal remains. Table 6.22 shows the Vina sample by phase and feature type. The assemblage is dominated by the cultural layer in Petnica 2 but split between this and House 3 in Petnica 3. Most Petnica 3 cultural layer units actually predate House 3, being beneath its base, although the exact interface between these and the floor make-up is problematic. Unsurprisingly, mixed Vina remains are predominantly from the cultural layer. Petnica 1 is poorly represented but includes remains from both main feature types. Other, incorporating pits and pyro-features, includes very few bones. Identification rates are very consistent, suggesting relatively
uniform preservation. Sample sizes for individual features are given in Table 6.23.
211
Petnica
Phase 1 Vina B2 2 Vina C 3 Vina D1 Mixed Vina 4 Eneolithic 5 Halstatt 6 Roman Medieval Mixed post-Vina Mixed
Unlabelled 237 1244 1936 627 13 10 8 8 233 492 4808
Labelled 455 2060 3089 956 31 17 6 22 373 791 7800
Total NSP 692 3304 5025 1583 44 27 14 30 606 1283 12608
NISP 255 1187 1710 514 21 15 5 10 214 465 4396
Table 6.20: NSP and NISP at Petnica by phase.
Sample Site Opovo Gomolava* Petnica** (NSP) 46431 3790 10439
Area (m2) 320 400 132
Density (NSP/m2) 145.1 9.5 79.1
Table 6.21: recovered densities of faunal remains in Vina levels at Opovo (Russell 1993), Gomolava and Petnica. * Block VII, squares C, D, G & H ** Trenches 10 and 11
(a) Cultural layer House* Other TOTAL NSP (b) Cultural layer House* Other TOTAL NISP
1 424 268 692 6.5% 141 114 255 7.0%
Petnica phase 2 3 3287 2634 2325 17 66 3304 5025 31.2% 47.4% 1185 866 821 2 23 1187 1710 32.4% 46.6%
Mixed 1523 60 1583 14.9% 495 19 514 14.0%
TOTAL 7868 74.2% 2653 25.0% 83 0.8% 10604 2687 954 25 3666 73.3% 26.0% 0.7%
Table 6.22: (a) NSP and (b) NISP in Vina phases at Petnica by feature type. * Includes all house-related contexts except Pit 2.
Figure 6.62 and Figure 6.63 show the spatial distribution of faunal remains, worked stone and pottery in Petnica 2 and 3. Excavation of Petnica 1 was too limited for such illustration. Each dot represents a single find, but find-spots are randomized within 1 m squares. Petnica 2 shows a strong concentration of all material types towards the west of Trench 10 and another in the NE of the excavated area. These are unlikely to be random since 30-40 cm of deposits are represented, and may thus represent foci of activity. It is harder to account for pottery distribution in this way, however, while there is no particular clustering around the pyrofeatures as might be expected. Informal foci of deposition are another possibility, but this
212
Petnica
remains speculation. The Petnica 3 cultural layer beneath House 3 is similar, also having a concentration in the NE. The distribution in Trench 10 is patchier, with two small concentrations towards the south. Apart from a single cache of lithics in the NE of Trench 10, the material types all have similar distributions. Figure 6.64 shows the distribution of faunal remains at House 3 floor level. No data on ceramics or lithics were available. Bones are sparse within the house compared to at least part of the surrounding cultural layer in Trench 11. Their distribution is patchy but the detection of any patterns is hampered by the limited areas and the number of squares in which the precise floor level could not be ascertained.
NSP Petnica 1 Cultural layer House 1 (walls) House 4 Walls Floor/base Petnica 2 Cultural layer Hearth Oven Pit 3 Petnica 3 Cultural layer House 3 Walls Interior features Floor Base Pit 2 Foundation trench Mixed Vina Cultural layer 424 16 252 142 110
NISP 141 8 106 62 44
3759 1 4 12
1335 0 0 2
2147 2466 838 97 621 829 66 15
711 868 265 29 234 312 23 5
1523
495
Table 6.23: NSP and NISP by feature at Petnica.
213
Petnica
N
Figure 6.62: distribution of finds in Petnica 2.
214
Petnica
N
Figure 6.63: distribution of finds in Petnica 3 (below House 3).
215
Petnica
Figure 6.64: distribution of faunal remains in Petnica 3 (House 3 floor level). Note that locations of internal features are extracted from diaries and photographs and are therefore approximate
6.4.4.4 taxonomic frequencies

Taxonomic frequencies at Petnica are shown by NISP in Table 6.24. A surprisingly wide range of taxa is present given the relatively small sample size. This is not simply a product of extensive sieving since the small carnivores are often represented by relatively large specimens, while bear and beaver are also present in unusual numbers. Most of the rarer
216
Petnica
species are fur-bearing, raising the possibility that specialized hunting for pelts took place at Petnica. The sites location is certainly suitable for such activity, but ungulates nonetheless dominate the assemblage. Fur-bearing species are considered further in 6.4.4.9. No bird or fish bones are present in my sample, but Greenfield (1986, 121) reports two Aves specimens from Petnica 1 and a few small fish vertebrae from Petnica 2, the latter recovered by flotation.
Taxon
Petnica 1 # % 84
80 4
Petnica 1/2 # % 103

98 5
Petnica 2 # % 373
354 19
Petnica 2/3 # % 51
48 3
Petnica 3 # % 533
487 14 32
Total # 1144
1067 14 63
% 31.3
29.2 0.4 1.7
Cattle
32.9
31.4 1.6
30.5
29.0 1.5
31.4
29.8 1.6
30.5
28.7 1.8
31.2
28.5 0.8 1.9
Pig
49
19 17 13
19.2
7.5 6.7 5.1
62
12 32 18
18.3
3.6 9.5 5.3
226
47 110 69
19.0
4.0 9.3 5.8
23
5 14 4
13.8
3.0 8.4 2.4
344
86 150 108
20.1
5.0 8.8 6.3
704
169 323 212
19.3
4.6 8.8 5.8
Caprine
Goat Sheep Indet.
8
2 6
3.1
0.8 2.4
11
2 3 6
3.3
0.6 0.9 1.8
64
14 50
5.4
1.2 4.2
12
2 10
7.2
1.2 6.0
129
4 37 88
7.5
0.2 2.2 5.1
224
6 58 160
6.1
0.2 1.6 4.4
Red deer Roe deer Canis

Dog Wolf Indet.
86 22 3
2 1
33.7 8.6 1.2

0.8 0.4
118 24 4
3 1
34.9 7.1 1.2

0.9 0.3
387 94 10
8 1 1
32.6 7.9 0.8

0.7 0.1 0.1
63 8 1
37.7 4.8 0.6
537 122 16
13 2
31.4 7.1 0.9

0.8 0.1 0.1
1191 270 34
26 4 4
32.6 7.4 0.9

0.7 0.1 0.1
0.6
Fox Wildcat Lynx Beaver Badger Otter Marten* Brown bear Hare Hedgehog TOTAL ID: Domestic Wild Indet. Roe deer/Caprine Cattle/Red deer Canid Cervid Mammal Large Large-medium Medium Medium-small Small Size indet. TOTAL NON-ID:
3 3 1 0.4 5 1 3 2 1 1 338 124 190 24 7 35 0.9 1.5 0.3 0.9 0.6 0.3 0.3 100.0 36.7 56.2 7.1 1 17 1 1 5 4 1 1187 473 625 89 26 83 1 762 446 156 140 11 4 5 872
0.3 2 0.1 1.4 0.1 0.1 0.4 0.3 0.1 100.0 39.8 52.7 7.5 5 1.2 3.0
2 1 17 1 2 4 2 1710 715 854 141 43 160 1 9 1157 656 228 202 52 5 14 1370
0.1 0.1 1.0 0.1 0.1 0.2 0.1 100.0 41.8 49.9 8.2
1 1 255 109 129 17 6 16 1 175 101 39 33 1 1 198
0.4 0.4 100.0 42.7 50.6 6.7
1 1
0.6 0.6
5 6 1 45 2 2 12 11 5 1 3657 1486 1892 279 88 311 1 11 2457 1407 502 437 73 12 26 2868
0.1 0.2 0.0 1.2 0.1 0.1 0.3 0.3 0.1 0.0 100.0 40.6 51.7 7.6
167 65 94 8 6 17
100.0 38.9 56.3 4.8
236 138 46 40 4 3 5 278
127 66 33 22 5 1 150
Table 6.24: taxonomic frequencies at Petnica by phase (NISP). * One specimen is Martes martes; the others Martes sp.
Table 6.25 compares my results with Greenfields study, revealing gratifyingly similar overall proportions. Differing numbers of pig remains represent an identification bias.
217
Petnica
Proportionally, Greenfield assigned nearly twice as many specimens to Bos/Cervus(/Equus) or Ovis/Capra/Capreolus as I, the main references for distinguishing Bovids and Cervids not having been published (Helmer & Rocheteau 1994; Prummel 1988). The importance of the readily-identifiable pigs is thus inflated. The increase in caprines is probably due to improved sieving and again recently published identification guides are implicated in differences within the category. Aurochsen are extremely rare in both samples, and most of my Bos sp. specimens are probably domestic. The only serious discrepancy is for pigs: Greenfield identifies almost four-fifths as domestic compared to my roughly one quarter but notes that, with few measurable specimens, his identifications are largely subjective (Greenfield 1986, 118). Having an appreciable measured sample, I was more content to assign juvenile and/or fragmentary specimens to Sus sp. This category is relatively large at Petnica, due ultimately to good recovery standards. Even if the vast majority of these are domestic not impossible given that many are juveniles they would still total little over half. Given the difficulty of separating wild and domestic pigs from small samples, I consider my data to be the more reliable. Trends in species abundance are compared in Figure 6.65. Greenfields data show a small increase in domestic species (Greenfield 1986, 121) at each body size, with cattle, caprines and pigs better represented in Petnica 3 than Petnica 1, at the expense of deer. My own data loosely support this trend for the smaller species but leaving aside Petnica 2/3 with its small sample size show remarkably consistent contributions of cattle, red deer, and pigs. Pleasing as Greenfields trend would be, the overall impression is of uniformity.
Taxon # Cattle Domestic Wild Indet. Pig Domestic Wild Indet. Caprine Goat Sheep Indet. Red deer Roe deer
Orton % 1144 1067 14 63 704 169 323 212 224 6 58 160 1191 270 3533 32.4 30.2 0.4 1.8 19.9 4.8 9.1 6.0 6.3 0.2 1.6 4.5 33.7 7.6 #
Greenfield % 176 171 5 143 110 30 3 17 2 2 13 173 43 552 31.9 31.0 0.9 0.0 25.9 19.9 5.4 0.5 3.1 0.4 0.4 2.4 31.3 7.8
Table 6.25: frequencies of main taxa at Petnica (%NISP) from the current study and Greenfield (1986, 284).
218
Petnica
40 35 30 %NISP 25 20 15 10 5 0 Petnica 1
n = 245 (n = 152) n = 313
Cattle Pigs Red deer Roe deer Caprine
Petnica 2
n = 1125 (n = 288) n = 154
Petnica 3
n = 1619 (n = 107)
Figure 6.65: frequency of main taxa by phase at Petnica (%NISP). Dotted lines and sample sizes in brackets show data from Greenfield (1986, 284).
Table 6.26 and Table 6.27 show taxonomic frequencies by MNE and DZ respectively, compared graphically with NISP in Figure 6.66. In contrast to Gomolava, quantification method affects the overall picture: the consistency seen for NISP is not matched by MNE and especially DZ. This must partly reflect smaller sample sizes for the latter methods which introduce fluctuations rather than clear trends but certain points are interesting. Cattle are much less frequent by DZ, presumably because their large but not especially robust bones are relatively rarely preserved sufficiently intact to score in Watsons scheme. Their MNE and NISP figures are similar, however, suggesting that fragmentation is not extreme. The robust nature of wild pig articular ends accounts for their relative abundance by DZ. The clearer gap between larger and smaller species for NISP suggests that fragmentation has exaggerated differences in abundance. The order of species is little changed, however, while on closer inspection fluctuations are often amplified rather than appearing from nowhere as one moves from (a) to (c). One could argue that NISP is the most robust measure or conversely that it masks genuine patterns; without independent control data is it impossible to evaluate these alternatives. If MNE and/or DZ are trusted over NISP, the main trend is a decrease in cattle over time, matched by an increase either in wild pig (MNE) or a range of species (DZ). Remarkably little difference in taxonomic frequency is observed between the main feature types (Table 6.28, Figure 6.67). Floor-level features within House 3 include almost 50% red deer, but from a total of only 29. Roe deer are most common in contexts with fewer red deer, and vice-versa. The same phenomenon occurs across phases, where these species curves are
219
Petnica
mirror-images whichever measure is used (Figure 6.66). Only beaver and wild carnivores, taken together, show appreciable differences between context types in relative terms: almost all are found in the cultural layer or amongst house rubble.
Taxon Cattle Domestic Wild Indet. Pig Domestic Wild Indet. Caprine Red deer Roe deer TOTAL
1 # % 33 33.3
32 2 32.3 2.0
1/2 # % 42 32.6
40 3 31.0 2.3
2 # % 107 31.2
100 9 29.2 2.6
2/3 # % 16 28.6
14 3 25.0 5.4
3 # % 141 26.9
133 6 11 25.4 1.1 2.1
22
9 7 7
22.2
9.1 7.1 7.1
29
6 20 3
22.5
4.7 15.5 2.3
76
28 41 18
22.2
8.2 12.0 5.2
12
3 8 1
21.4
5.4 14.3 1.8
132
31 83 30
25.2
5.9 15.8 5.7
4 31 9 99
4.0 31.3 9.1
5 46 7 129
3.9 35.7 5.4
22 98 40 343
6.4 28.6 11.7
5 19 4 56
8.9 33.9 7.1
41 174 36 524
7.8 33.2 6.9
Total # 339 319 6 28 271 77 159 59 77 368 96 1151
% 26.9 25.4 1.1 2.1 25.2 5.9 15.8 5.7 7.8 33.2 6.9
Table 6.26: taxonomic frequencies at Petnica by phase (MNE).
Taxon
Petnica 1 # % 13
13
Petnica 1/2 # % 15
14 1
Petnica 2 # % 54
51 3
Petnica 2/3 # % 4
4
Petnica 3 # % 62
54 3 5
Total # 148
136 3 9
% 20.6
18.9 0.4 1.3
Cattle
25.5
25.5
16.5
15.4 1.1
22.7
21.4 1.3
14.8
14.8
19.9
17.3 1.0 1.6
Pig
14
4 8 2
27.5
7.8 15.7 3.9
18
3 15
19.8
3.3 16.5
51
10 38 3
21.4
4.2 16.0 1.3
4
4
14.8
0.0 14.8
78
9 59 10
25.0
2.9 18.9 3.2
165
26 124 15
22.9
3.6 17.2 2.1
Caprine
Goat Sheep Indet.
1
1
2.0
2.0
4
1 2 1
4.4
1.1 2.2 1.1
9
5 4
3.8
2.1 1.7
4
1 3
14.8
3.7 11.1
25
1 16 8
8.0
0.3 5.1 2.6
43
2 25 16
6.0
0.3 3.5 2.2
Red deer Roe deer Canis

Dog Wolf Indet.
14 7 1
1
27.5 13.7 2.0

2.0
36 3 1
1
39.6 3.3 1.1

1.1
66 33 3
2 1
27.7 13.9 1.3

0.8 0.4
8 2 1
29.6 7.4 3.7
99 24 9
8 1
31.7 7.7 2.9

2.6 0.3
223 69 15
12 2 1
31.0 9.6 2.1

1.7 0.3 0.1
3.7
Fox Wildcat Lynx Beaver Badger Otter Marten Brown bear Hare Hedgehog TOTAL
3 4 3 1 3 2 1 51 100.0 91 4.4 3.3 1.1 3.3 2.2 1.1 100.0 238 1 7 1 1 5 4
1.3 2 0.4 2.9 0.4 0.4 2.1 1.7 1 7.4 3.7
0.3
8 1 3 2
2.6 0.3 1.0 0.6 100.0
2.0
3.7
4 6 1 19 2 2 13 6 2 1 719
0.6 0.8 0.1 2.6 0.3 0.3 1.8 0.8 0.3 0.1 100.0
100.0
27
100.0
312
Table 6.27: taxonomic frequencies at Petnica by phase (DZ).
220
Petnica
40 35 30
a)
Cattle Red deer Domestic pig Wild boar Roe deer Caprines
%NISP
25 20 15 10 5 0 Petnica 1 Petnica 2 Petnica 3
40 35 30
b)
%MNE
25 20 15 10 5 0 Petnica 1 45 40 35 30 Petnica 2 Petnica 3
c)
%DZ
25 20 15 10 5 0 Petnica 1 Petnica 2 Petnica 3
Figure 6.66: trends in frequencies of main taxa at Petnica by (a) %NISP, (b) % MNE and (c) %DZ.
221
Petnica
Taxon Cattle
Cultural layer # % 842 31.5

788 8 46 29.5 0.3 1.7
Base # 99
88 3 8
House Floor % 32.2

28.7 1.0 2.6
# 89
86 1 2
% 29.1
28.1 0.3 0.7
Rubble # % 105 31.4

96 2 7 28.7 0.6 2.1
Pig
520
115 248 157
19.4
4.3 9.3 5.9
51
12 24 15
16.6
3.9 7.8 4.9
61
15 25 21
19.9
4.9 8.2 6.9
67
26 26 15
20.1
7.8 7.8 4.5
Caprine Red deer Roe deer Dog Other* TOTAL
153 869 197 17 76 2674
5.7 32.5 7.4 0.6 2.8
18 116 18 3 2 307
5.9 37.8 5.9 1.0 0.7
25 102 23 3 3 306
8.2 33.3 7.5 1.0 1.0
28 93 31 3 7 334
8.4 27.8 9.3 0.9 2.1
Table 6.28: taxonomic frequencies (%NISP) at Petnica by context type. * wild carnivores and beaver.
40 35 30
Cultural layer House base House floor House rubble
%NISP
25 20 15 10 5 0 Cattle Pig Caprine Red deer Roe deer Dog Other*
Figure 6.67: taxonomic frequencies (%NISP) at Petnica by context type. * wild carnivores and beaver.
6.4.4.5 age-at-death: mandibles

Cattle Mandibles were assigned to Payne stages as at Gomolava and can be compared with Arnold s and Greenfields (2006) results from 1980-1986. Almost opposite patterns emerge, even when only Petnica 2 and 3 (Arnolds and Greenfields Late Neolithic) are included (Figure 6.68) very few of my specimens derive from Petnica 1 (Middle Neolithic). Methodological bias is probably responsible, presumably relating to loose teeth since protocols are otherwise identical. Fifteen of Arnolds and Greenfields nineteen specimens are loose, while I include only mandibles. Loose M1s and M2s are hard to distinguish, while
222
Petnica
loose dP4s can only be attributed to Payne stage A or B+ (realistically, B-D). M3s are unmistakeable and can be assigned securely to D, E, F, G/H or I. A priori, including loose teeth should favour these categories at the expense of B and C. Correlation between percentage at stage E and percentage loose (Pearsons r = 0.7858; n = 12) across Arnolds and Greenfields study confirms that loose teeth are problematic, although the number at stage D cannot be explained in the same way (r = 0.2010). My own mandibular data are nonetheless considered more reliable and are used for the survival curve in Figure 6.69. Samples for individual phases are too small for any confidence, but specimens are generally older in Petnica 3. A single aurochs mandible with an in situ M3 was at stage I.
Stage 7 6 5 Frequency 4 3 2 1 H 0 A B C D E Payne stage F G H I I Total 17
Count 0 3.4 5 1.2 0.5 0.5 4 1 1 0.0% 31.2% 48.8% 10.9% 4.5% 4.5% 36.4% 9.1% 9.1%
Orton (n = 17) Arnold & Greenfield 2006 (n = 19)
A B C D E F G
Figure 6.68: domestic cattle mortality in Petnica 2 and 3 (Payne stages).
100 80 60 40 20 0 0 6 12 18 24 30 36 Suggested age (months) 42 Young 54 48 adult 60 Adult 66 72 78 84 90 Old adult Senile
Percentage survival
Figure 6.69: survivorship curve for domestic cattle from Petnica 2 and 3.
223
Petnica
Pigs Arnoldss and Greenfields pig data include very few loose teeth. Their mortality profiles fit fairly well with my own data for wild and domestic pigs combined, but not with domestic taken alone (Figure 6.70). Their data probably include wild specimens (see 6.4.4.4); samples cannot therefore be pooled to produce a viable dataset. Both wild and domestic pigs have very young profiles but samples are too small to trust, while older individuals of both populations are represented among the maxillary specimens. Little can be said about pig mortality at Petnica.
Domestic (n = 11) Wild (n = 10) Percentage survival 80% 60% 40% 20% 0% 0 10 20 30 40 Adult 50 Suggested age (months) Old 60 adult Senile 70 Arnold & Greenfield 2006 (n = 26)
100%
Figure 6.70: survivorship curves for pigs at Petnica.
Caprines Few caprine mandibles can confidently be assigned to Payne stages (Table 6.29), but these cluster at C/D and F/G, concentrated into D and G by Paynes reallocation system. In fact, the younger specimens mostly fall on the C/D boundary using Joness sub-stages, while the elder appear more widely spread. Given the sample sizes and the presence of fourteen loose teeth in their sample, Arnolds and Greenfields results are not inconsistent with this (Figure 6.71). Nonetheless, it is hard to make any inferences from such thin data. Deer Greenfield (1986,119-120) notes that both red and roe deer are mostly adults in the Vina levels. My observations for red deer support this (Figure 6.72), although detailed age estimation is hampered by fragmentary mandibles. There are no obvious differences between phases or context types. Of fifteen ageable roe deer mandibles, five are young adults and the remainder adults (see 6.3.5.5).
224
Petnica
Phase 1 1/2 2 2 2 2 2 2 2/3 3 3 3 3 3 3 3 3 3 3 3 3 3
Feature Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer Cultural layer House 3 floor House 3 floor House 3 walls House 3 walls House floor
Species Sheep/goat Sheep/goat Sheep Sheep Sheep Sheep Sheep/goat Sheep/goat Sheep/goat Sheep Sheep Sheep Sheep Sheep Sheep Sheep Sheep/goat Sheep Sheep Sheep Sheep/goat Sheep/goat
Silver (1969) 9-24 3-24 <24 9-24 21+ 21+ 3-24 <24 3-24 21+ 3+ <24 3-24 21+ 21+ 18+ 21+ <24 3-24 21+ 21+ 21+
Stage D+ C/D D F/G F/G C+ C/D D+ C+ C/D G E+ G G C/D D+ F/G
Payne (1973) Suggested age 12+ 6-24 12-24 36-72 36-72 6+ 6-24 12+ 6+ 6-24 48-72 24+ 48-72 48-72 6-24 12+ 36-72
Jones (2006) Stage* Age** 10+ ('C6+'/'D1/2') 8-13 10-24 30-(78) 30-(108) 3+ 3-24 10+ 3+ ('C6+'/'D1/2') Gb 8-13 54-(108) 20+ 54-(108) 54-(108) 8-13 10+ 30-(108)
F/Ga F/Gb
Gb ('C6+'/'D1/2') F/Gb
Table 6.29: age estimates for caprines from Petnica. * Stages in brackets are almost certain from Joness inter-tooth comparisons. ** Numbers in brackets are estimated upper limits.
Orton (n = 9) Arnold & Greenfield 2006 (n = 19)
Corrected count
0 A B C D E Payne stage F G H I
Figure 6.71: caprine mortality at Petnica.
14 12 Corrected count 10 8 6 4 2 0 Infantile Juvenile Subadult Adult
Figure 6.72: red deer mortality at Petnica.
225
Petnica
6.4.4.6 age-at-death: epiphyseal fusion

The combination of (a) smaller mandibular samples and (b) better post-cranial recovery at Petnica than at Gomolava potentially makes fusion data valuable. The sample for cattle is large enough for comparisons between phases, revealing similar curves but with many fewer individuals apparently surviving to maturity in Petnica 1 than Petnica 2, with Petnica 3 intermediate (Figure 6.73). These results must be treated with caution since mandibular data indicate much lower survival, suggesting a taphonomic bias against unfused postcranial specimens. Nonetheless, comparisons between phases should be valid if one assumes this bias is constant. Fusion data are problematic for pigs since unfused specimens are much more likely to be recorded as Sus sp. Most of these are probably domestic, but to treat them as such would be to pre-suppose the age-at-death results. Morans and OConnors (1994) fusion groups are used for caprines as Zeders (2006b) finer scheme produces unworkably small samples. Even so, group 4 (n = 1, fused) is excluded. The curve features a reversal between groups 2 and 3 (Figure 6.74), indicating that mortality in the relevant portion of the animals lives is low enough to be obscured by sampling error. At face value, the data suggest that around 70% survive to (sub)adulthood. This is roughly equivalent to the number exceeding Payne stage D: 64% in my sample and 60% for Arnold and Greenfield. Cranial and post-cranial results are thus broadly consistent.
100% Percentage fused/fusing 80% 60% 40% 20% 0% 0 1 2 Group 3

1 (n = 24) 2 (n = 52) 3 (n = 80) Total (n = 177)
Figure 6.73: percentage of domestic cattle specimens fused in each fusion group at Petnica, by phase. Includes Bos sp.
226
Petnica
100%
Percentage fused/fusing
80%
60%
40%
20%
0% 0 1 Group 2 3
Figure 6.74: percentage of caprine specimens fused in each fusion group at Petnica.
6.4.4.7 sex
Table 6.30 shows all sex attributions from the Vina levels at Petnica, using the same techniques as for Gomolava (see 6.3.5.7). The sex ratio for cattle is surprisingly even, but that for domestic pigs is female biased as one might expect. Wild pigs appear fairly balanced until loose canines are considered, most of which are male. This may be a recovery bias or might relate to collection and/or curation of tusks. Red deer were sexed from pelvises and metacarpal measurements, giving a female:male ratio of almost 2:1. Only a single unshed antler was recovered, along with a single female frontal. Unshed antlers are the main form of evidence for roe deer and suggest a bias towards males. The sample is small, however, and it is always possible that unshed antlers were collected from carcasses, or that male skulls were preferentially brought to the site from kills.
Pelvis F M Cattle Domestic Indet. Pig Domestic Wild Indet. Red deer Roe deer 7 1 6 0
Mandible F M
Maxilla F M
Loose teeth F M
Cranium F M
Biometry F M
Total F 7 1 M 6 0 2 10 7 9 5
1 4 0 8 1 6 0
1 2 2
3 0 1
0 1 0
3 2 2
1 7 5 1 0 1 5 8 2
7 6 3 17 1
Table 6.30: sex attributions at Petnica.
227
Petnica
6.4.4.8 seasonality
Seasonal data are extremely limited at Petnica. The peak for cattle at Payne stages B and C spans one-and-a-half years, and the sample is too small for the range of observed dP4 or M1 states to be informative. For pigs, MWS stages could only be assigned to four wild (10, 20, 24, 25) and two domestic (12, 18) specimens. The wild MWS scores all fall more-or-less in the troughs formed by the Gomolava data: if hunting was seasonal at both sites then it took place in opposite seasons. The peak at the end of C and start of D for caprines is suggestive of a seasonal cull, but from Joness data falls a little later in the year than might be expected. The sample is very small, in any case, while the second peak at F/G is at least two years later and probably spread over a considerable period. Almost all red deer at Petnica have complete dentition, preventing estimation of season at death. Likewise, the intact roe deer mandibles have all permanent teeth in wear, and there is no clustering in the Brown and Chapman (1991b) wear scores. Two young adult specimens missing the distal end of the alveolar might feasibly have M3 still in eruption, placing them at a little under one year, but by comparison of wear stages with other specimens are more likely to be dentally mature. Neither definitely shed antlers nor crania with antlers shed are present, while one red deer and five roe deer specimens have unshed antler. No credence is given to the single Cervus specimen, but the Capreolus antlers provide moderately good evidence for a MarchNovember focus of hunting.
6.5 Summary
This chapter has presented the two case-study sites and their faunal samples. This provides the necessary background for the more detailed taphonomic-contextual analysis of Chapter 7, as well as being the source for some of the data already synthesised in Chapter 5. Discussion of the implications of this material has been minimal here but is re-visited in Chapter 8, where the various scales of analysis of Chapters 5-7 are brought together to characterize human engagements with animal species.
228
Taphonomic analysis
Chapter 7 Taphonomic analysis at Gomolava and Petnica

7.1 Evidence for attrition
The taphonomic structure employed here was set out in 4.4.3. Analyses proceed in a broadly reverse-chronological order, working back from observed assemblage to inferred behaviour in five steps. At each stage apart from the first, information is gleaned regarding human activities as well as formation processes and preservation biases. The first stage involves testing for the effects of density-mediated typically post-depositional destruction processes on the observed element profiles. The core of this analysis is the assessment of correlation between element abundance and density.
7.1.1 completeness of carpals and tarsals

Fragmentation amongst small, compact bones i.e. carpals and tarsals provides an independent check on post-depositional attrition, since they are unlikely to have been broken up by direct human action or carnivore ravaging (Marean 1991). Table 7.1 shows the percentage of each of these recorded as either complete or almost complete at Gomolava and Petnica. Fragmented specimens may simply have been missed at Gomolava given the lack of sieving, but even the relatively large navicular-cuboids are rarely broken. The much better recovered assemblage from Petnica also features very high levels of complete or almost complete small bones.
Gomolava % complete 100 100 100 100 100 96 Petnica % complete 100 90 100 100 93.75 90 100 81.48 100
Radial Intermedium Ulnar Accessory 2nd & 3rd C. 4th & 5th C. Lateral malleolus Navicular-cuboid 4th & 5th T.
n 8 5 3 3 3 25 -
n 17 10 9 1 16 10 3 27 5
Table 7.1: percentage of small bones recorded as complete or almost complete at Gomolava and Petnica.
7.1.2 correlation of abundance with bone density

The methodology used to assess density-mediated attrition is detailed in 4.4.3.1. Density is plotted against both MAU and NISP by species, excluding aurochsen and caprines due to small samples, and canid due to their different skeletal anatomy. Table 7.2 summarises correlation coefficients for MAU versus density in all Vina units together and in the main two context types. Significance
229
Taphonomic analysis
levels are not quoted since they would be spurious; the technical sample size is the number of data points fixed by the number of portions defined rather than the actual number of elements or fragments contributing to the figures. An r value of 0.5 would be attributed the same significance whether the total MAU is 20 or 200. Significance must thus be judged informally with the aid of scatter plots, which are included in the text where appropriate and otherwise available in Appendix 6. In general, there is very little evidence for density-mediated effects on MAU values. Domestic pigs are the exception, but even here the scatter plot reveals no clear relationship (Figure 7.1).
Context type All
Species Bos taurus Cervus elaphus Capreolus capreolus Sus (all) Sus (wild) Sus (domestic)
Spearman's 0.2087 0.0872 -0.1171 0.2382 0.0386 0.3444 0.1023 0.0444 -0.1171 0.2424 0.1441 0.3046 0.2363 0.0742 0.2991 -0.132 0.3858
Pearson's 0.1642 0.0955 -0.0447 0.3078 0.1181 0.353 0.072 0.0434 -0.0447 0.3116 0.201 0.366 0.1404 0.1041 0.2693 -0.0503 0.298
Total MAU* Data points 217 134 22 211 105 68 93 79 22 132 78 34 118 52 73 23 32 21 21 21 39 39 39 21 21 21 39 39 39 21 21 39 39 39
Pit
Bos taurus Cervus elaphus Capreolus capreolus Sus (all) Sus (wild) Sus (domestic)
Cultural layer
Bos taurus Cervus elaphus Sus (all) Sus (wild) Sus (domestic)
Table 7.2: correlations between MAU and median portion density at Gomolava. * includes portions involved in correlation analysis only; rounded to nearest whole unit.
12 11 10 9 8 7 MAU 6 5 4 3 2 1 0 -1 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Density Spearman's r =0.3444
Figure 7.1: relationship between MAU and median density for domestic pigs at Gomolava. Data from Pugsley 2002; whiskers represent 50% confidence intervals. For visual clarity, points have been jittered by adding or subtracting negligeable random values to/from MAU scores.
230
Taphonomic analysis
Correlation results for corrected NISP and density are given in Table 7.3. Moderate correlations for cattle, aurochsen and red deer suggest that density-mediated destruction played a significant but not overwhelming role in the formation of the assemblage as a whole at Gomolava. Inspection of the scatter plots confirms this (Figure 7.2): the least dense elements are all present in small numbers. Considerable variation remains amongst the denser parts, however, indicating that other factors also operate. Pigs appear much less affected (Figure 7.3). The discrepancy between NISP and MAU results can be interpreted two ways: 1. MAU is more robust than NISP with regard to density-mediated processes, making it more reliable for interpreting body part profiles in terms of pre-depositional factors. 2. Minimum number calculations obscure variations between elements whatever their origin by exaggerating the abundance of rarer bones, especially in small samples. Both points are probably partly true. One could reasonably argue that calculations based on NISP are preferable for small samples, for example individual features, while MAU is more meaningful for larger entities.
Context type All
Species Bos primigenius Bos taurus Cervus elaphus Capreolus capreolus Caprines Sus (all) Sus (wild) Sus (domestic)
Spearman's 0.5343 0.7054 0.4938 0.1671 0.283 0.3359 0.3029 0.3085 0.309 0.4893 0.5203 0.1537 0.2451 0.3469 0.3793 0.2906 0.4893 0.663 0.5052 0.1691 0.0528 0.2525
Pearson's Total NISP* 0.5408 0.6817 0.5985 0.3439 0.3691 0.2616 0.2557 0.2191 0.2928 0.4802 0.5906 0.3317 0.3132 0.2873 0.3763 0.2568 0.5327 0.6553 0.5603 0.1753 -0.0025 0.1203 33 492 304 39 21 249 134 45 10 185 174 38 16 164 100 26 19 252 111 64 27 18
Data points 21 21 21 21 21 39 39 39 21 21 21 21 21 39 39 39 21 21 21 39 39 39
Pit
Bos primigenius Bos taurus Cervus elaphus Capreolus capreolus Caprines Sus (all) Sus (wild) Sus (domestic)
Cultural layer
Bos primigenius Bos taurus Cervus elaphus Sus (all) Sus (wild) Sus (domestic)
Table 7.3: correlation coefficients and sample sizes for corrected NISP vs. density at Gomolava. * includes portions with density values only; rounded to nearest whole unit.
231
Taphonomic analysis
a)
6 r = 0.5343 5
b)
60 50
r = 0.7054
Corrected NISP
Corrected NISP
40 30 20 10 0
0 0 0.1 0.2 0.3 Density 0.4 0.5 0.6
0.1
0.2
0.3 Density
0.4
0.5
0.6
c)
9 8 7 r = 0.1671
d)
4.5 4 3.5 r = 0.2830
Corrected NISP
5 4 3 2 1 0 0 0.1 0.2 0.3 Density 0.4 0.5 0.6
Corrected NISP
3 2.5 2 1.5 1 0.5 0 0 0.1 0.2 0.3 Density 0.4 0.5 0.6
e)
45 40 35 r = 0.4938
a) b) c)
Aurochs Cattle Roe deer Caprines Red deer
Corrected NISP
30 25 20 15 10 5 0 0 0.1 0.2 0.3 Density 0.4 0.5 0.6
d) e)
Figure 7.2: median density against corrected NISP for bovids and cervids at Gomolava. Density values from Symmons 2002. Whiskers show interquartile range, r values are Spearmans rho.
232
Taphonomic analysis
a) All Sus
19 17 15 Corrected NISP 13 11 9 7 5 3 1 -1 0.5 1 Density 1.5 2 r = 0.3359
b) Domestic pig
13 r = 0.3085 11 9 Corrected NISP 7 5 3 1 -1 0.5 1 Density 1.5 2
c) Wild pig
13 r = 0.3029 11 9 Corrected NISP 7 5 3 1 -1 0.5 1 Density 1.5 2
Figure 7.3: mean density against corrected NISP for pigs at Gomolava. Density values from Puglsey 2002. Whiskers show 50% confidence interval, r values are Spearmans rho.
233
Taphonomic analysis
Density is plotted against abundance for pig elements in anatomical order in Figure 7.4. Bars represent MAU as a percentage of the highest observed value while points show average density values, again as an index. While the overall representation of elements and portions only follows density values loosely, the abundance of proximal and distal ends typically matches density very closely. Long-bone shafts are often underrepresented here, but fit well with density values when measured by NISP. The correspondence between MAU and density for individual portions of scapulae and ulnae is particularly striking given the close proximity of scan sites. Density-mediated destruction has thus played a detectable role in the formation of the assemblages, but is only the main factor in part representation at the intra-element level. Other factors must be at least as important between elements.
110 100 90 80
Density index
MAU Mean density index
110 100 90
%MAU (max = 17)
80 70 60 50 40 30 20 10 0
Dn-1 Dn-4 Dn-6 Dn-8 At-3 Ax-1 Sp-1 Sp-2 Hu-2 Hu-3 Hu-5 Ra-1 Ra-3 Ra-4 Ul-1 Ul-2 Mc-1 Mc-3 Mc-5 Ac-1 Il-2 Pu-1 Fe-2 Fe-4 Fe-6 Ti-2 Ti-3 Ti-4 As-1 Ca-3 Ca-2 Mt-1 Mt-3 Mt-5 P1-1 P1-3 P2-1 P2-2 P3-1
Portion
70 60 50 40 30 20 10 0
Figure 7.4: density vs. % MAU for pigs from Gomolava.
Of course, the same factors may not have shaped element representation in all contexts. Symmons (2002, 248-253), for example, demonstrates a greater correlation between abundance and density in external than internal contexts at atalhyk. Before presenting contextual comparisons here, it is worth noting that at least two effects may underlie differences in strength of correlation: 1. Varying degrees of density-mediated attrition. 2. Selection prior to primary deposition. In general, the closer the deposited assemblage to complete skeletons the more readily density-mediated effects will be detected. Figure 7.5 and Figure 7.6 show the relationship between density and abundance for cattle, red deer and pigs from each of the two main context types. For cattle, the higher correlation seen in the cultural layer probably indicates slightly greater attrition, while for red deer very similar coefficients
234
Taphonomic analysis
are seen. Sample sizes for pigs are problematic, but less so than they appear since n refers to total corrected NISP. Here, density is a poor predictor of abundance, and the anatomical distribution graphs suggest that this reflects prior selection rather than necessarily less severe attrition. The density-abundance relationship is not assessed by phase at Gomolava since most Ia faunal remains derive from pits, while most from Iab were found in the cultural layer. It is thus impossible to separate the effects of phase and context without producing unworkably small samples.
a)
100
% Corrected NISP
0.7 0.6 0.5

Corrected NISP Density
40 35 30 25 20 15 10 5 0 0 0.1 0.2 0.3 Density Cultural layer: r = 0.5052 0.4 0.5 0.6 Cultural layer: r = 0.6630 Pits: r = 0.4893
80 60 40
0.4 0.3 0.2
20 0 P Scap P Hum S Hum D Hum P Rad S Rad D Rad P MetaC S MetaC D MetaC P Pelv P Fem S Fem D Fem P Tib S Tib D Tib P MetaT S MetaT D MetaT Phal 1
0.1 0
b)
100
Corrected NISP
0.7 0.6 0.5 0.4 0.3

Corrected NISP
17 15 13 11 9 7 5 3 1 -1 0
Pits: r = 0.5203
80 60 40
0.2 20 0 P Scap P Hum S Hum D Hum P Rad S Rad D Rad P MetaC S MetaC D MetaC P Pelv P Fem S Fem D Fem P Tib S Tib D Tib P MetaT S MetaT D MetaT Phal 1 0.1 0
Density
0.1
0.2
0.3 Density
0.4
0.5
0.6
Figure 7.5: relationship between NISP and density by feature type for (a) cattle and (b) red deer at Gomolava. Corrected NISP values given as percentage of maximum, r values are Spearmans rho.
Analysis at Petnica follows the same pattern as Gomolava. Table 7.4 show basic results by MAU; Table 7.5 by corrected NISP. Density-mediated attrition has affected MAU profiles much more severely at Petnica, with correlations of around 0.5 for the smaller species, but its impact is still barely visible graphically (Figure 7.7). Wild pigs appear immune, and size may be behind a higher if still very modest correlation for domestic pigs. The latter sample is very small, but with a fixed number of data points this actually makes correlation less likely. No dramatic differences are seen between the cultural layer and houses (including all house-related contexts).
235
a) 110
Cultural layer (n = 27) 110 100 90 80 70
7 5 3 1 -1 0.5 1 Density 1.5 2 9 11 Pits: r = 0.3793 Cultural layer: r = 0.0528
100
Pits (n = 100)
90
80 60 50 40 30 20 10 0
70
60
50
Density index
% Corrected NISP
30
20
10
Dn-1 Dn-4 Dn-6 Dn-8 At-3 Ax-1 Sp-1 Hu-2 Hu-3 Hu-5 Ra-1 Ra-3 Ra-4 Ul-1 Mc-1 Mc-3 Mc-5 Ac-1 Il-2 Pu-1 Fe-2 Fe-4 Fe-6 Ti-2 Ti-3 Ti-4 As-1 Ca-3 Ca-2 Mt-1 Mt-3 Mt-5 P1-1 P1-3 P2-1 P2-2 P3-1
Portion
b) 110
Cultural layer (n = 18) 110 100 90 80 70 60 50 40 30 20 10 0 Pits (n = 26)
Corrected NISP
40
Cultural layer: r = 0.2525 7 Pits: r = 0.2906
100
90
80
70
60
50
Density index
% Corrected NISP
30
Corrected NISP
40
20
10
Dn-1 Dn-4 Dn-6 Dn-8 At-3 Ax-1 Sp-1 Hu-2 Hu-3 Hu-5 Ra-1 Ra-3 Ra-4 Ul-1 Mc-1 Mc-3 Mc-5 Ac-1 Il-2 Pu-1 Fe-2 Fe-4 Fe-6 Ti-2 Ti-3 Ti-4 As-1 Ca-3 Ca-2 Mt-1 Mt-3 Mt-5 P1-1 P1-3 P2-1 P2-2 P3-1
-1 0.5 1 Density 1.5 2
Portion
Figure 7.6: relationship between NISP and density by feature type for (a) wild pig and (b) domestic pig at Gomolava. Corrected NISP values given as percentage of maximum, r values are Spearmans rho.
236
Taphonomic analysis
Context/phase All
Species Bos taurus Cervus elaphus Capreolus capreolus Ovis/Capra Sus (all) Sus (wild) Sus (domestic) Bos taurus Cervus elaphus Capreolus capreolus Ovis/Capra Sus (all) Sus (wild) Sus (domestic) Bos taurus Cervus elaphus Capreolus capreolus Ovis/Capra Sus (all) Sus (wild) Sus (domestic)
Spearman's 0.3042 0.2369 0.4710 0.5990 0.2787 0.0745 0.2831 0.2277 0.1698 0.4994 0.4727 0.2904 0.0698 0.3685 0.3797 0.0742 0.2539 0.5990 0.1949 0.0745 0.2478
Pearson's 0.3559 0.2530 0.5439 0.5604 0.2868 0.0733 0.3400 0.3005 0.2343 0.5433 0.5177 0.3164 0.0938 0.3557 0.4108 0.1094 0.3401 0.5604 0.1752 0.0733 0.2476
Total MAU* 144 178 42 43 120 79 21 100 120 30 22 88 59 14 35 44 12 43 26 79 6
Data points 21 21 21 21 39 39 39 21 21 21 21 39 39 39 21 21 21 21 39 39 39
Cultural layer
House
Table 7.4: correlation coefficients and sample sizes for MAU vs. density at Petnica. *for portions with density measurements
Context/phase All
Species Bos taurus Cervus elaphus Capreolus capreolus Caprines Sus (all) Sus (wild) Sus (domestic) Bos taurus Cervus elaphus Capreolus capreolus Caprines Sus (all) Sus (wild) Sus (domestic) Bos taurus Cervus elaphus Capreolus capreolus Caprines Sus (all) Sus (wild) Sus (domestic)
Spearman's 0.7321 0.7048 0.7161 0.7324 0.2446 0.2177 0.3124 0.6873 0.6736 0.7708 0.6746 0.3199 0.2792 0.2579 0.6832 0.7038 0.4612 0.7776 0.0889 0.0677 0.0929
Pearson's 0.7013 0.7017 0.7687 0.7614 0.3222 0.3039 0.2767 0.6727 0.6857 0.7716 0.7161 0.3810 0.3511 0.2866 0.7044 0.6787 0.6666 0.7873 0.1119 0.1351 0.0732
Total NISP* Data points 342 516 163 96 337 181 98 223 368 108 54 234 132 63 99 118 45 29 85 44 25 21 21 21 21 31 31 31 21 21 21 21 31 31 31 21 21 21 21 31 31 31
Cultural layer
House
Table 7.5: correlation coefficients and sample sizes for corrected NISP vs. density at Petnica. *for portions with density measurements. .
237
Taphonomic analysis
a) Cattle
12
b) Red deer
18 16
Spearman's r = 0.3042
10 14 8 12 10 8 6 4 2 2 0 0.1 0.2 0.3 0.4 0.5 0.6 0 0.1 0.2 0.3 0.4 0.5 0.6
MAU
Density
MAU
Density
c) Roe deer
d) Caprines
6
14
12
10 4
MAU
MAU
2 0 0.1 0.2 0.3 0.4 0.5 0.6
0 0.1 0.2 0.3 0.4 0.5 0.6
Density
Density
Figure 7.7: median density against MAU for bovids and cervids at Petnica. Density values from Symmons 2002. Whiskers show interquartile range, r values are Spearmans rho.
As at Gomolava, correlations are higher when abundance is measured by corrected NISP: all major bovid/cervid species have Spearmans r greater than 0.7. The relationship is clearly visible from scatter plots (Figure 7.8). Pigs are much less affected than other taxa, although the difference between wild and domestic specimens observed for MAU disappears for NISP. Both wild and domestic pigs show stronger but still weak correlation between NISP and density in the cultural layer than among house remains (Figure 7.9), although samples of domestic specimens are small. This is true to a lesser extent of roe deer but not of any other species, as illustrated by Figure 7.10. The scatter-plots within Figure 7.10 show that while correlation coefficients for bovids and cervids vary little between context types, the slope of the relationship is steeper in the cultural layer. Attrition was apparently more severe in the cultural layer but the difference had little or no additional impact on element ranking. At this juncture one cannot say why the difference in degree
238
Taphonomic analysis
of attrition should be expressed more clearly for pigs, but the fact that this applies only to NISP and not to MAU suggests that differential fragmentation may be involved.
a) Cattle
45 40 35 30 25 20 15 10 5 0 0.1 0.2 0.3 0.4 0.5 0.6
b) Red deer
80 75 70 65 60 55
Corrected NISP
Corrected NISP
50 45 40 35 30 25 20 15 10 5 0 0.1 0.2 0.3 0.4 0.5 0.6
Density
Density
c) Roe deer
32 30 28 26 24 22
d) Caprines
22 20 18 16
Corrected NISP
20 18 16 14 12 10 8 6 4 2 0 0.1 0.2 0.3 0.4 0.5 0.6
Corrected NISP
14 12 10 8 6 4 2 0 0.1 0.2 0.3 0.4 0.5 0.6
Density
Density
Figure 7.8: median density against corrected NISP for bovids and cervids at Petnica. Density values from Symmons 2002. Whiskers show interquartile range, r values are Spearmans rho.
239
b)
Cultural layer (n = 132) House remains (n = 44)
16 Cultural layer: r = 0.2892 House remains: r = 0.0677 14 12 10 8 6 4 2
2.5 2 1.5 1 0.5 0

0.5 1 Density 0 1.5 2 Density
18
100
80
60
% Corrected NISP
20
c)
12
Dn-4 At-3 Ax-1 Sp-1 Hu-2 Hu-3 Hu-5 Ra-1 Ra-3 Ra-4 Ul-1 Mc-1 Mc-3 Mc-5 Ac-1 Fe-2 Fe-4 Fe-6 Ti-2 Ti-3 Ti-4 As-1 Ca-3 Mt-1 Mt-3 Mt-5 P1-1 P1-3 P2-1 P2-2 P3-1
Corrected NISP 14 Cultural layer: r = 0.2579 House remains: r = 0.0929 10 8 6 4 Corrected NISP 2 0 0.5
40
2.5 2 1.5 1 0.5 0

Density
100
80
60
% Corrected NISP
40
20
1 Density
Dn-4 At-3 Ax-1 Sp-1 Hu-2 Hu-3 Hu-5 Ra-1 Ra-3 Ra-4 Ul-1 Mc-1 Mc-3 Mc-5 Ac-1 Fe-2 Fe-4 Fe-6 Ti-2 Ti-3 Ti-4 As-1 Ca-3 Mt-1 Mt-3 Mt-5 P1-1 P1-3 P2-1 P2-2 P3-1
1.5
Figure 7.9: relationship between NISP and density by feature type for (a) wild pig and (b) domestic pig at Petnica. Corrected NISP values given as percentage of maximum, r values are Spearmans rho.
240
Taphonomic analysis
a)
100
% Corrected NISP
0.7 0.6 0.5

Density
35 Cultural layer: r = 0.6873 30 25 Corrected NISP 20 15 10 House remains: r = 0.6832
80 60 40
0.4 0.3 0.2
0.1
5
0
0 0 0.1 0.2 0.3 Density 0.4 0.5 0.6
b)
100
% Corrected NISP
0.7 0.6 0.5

Density
70 Cultural layer: r = 0.6737 60 50 Corrected NISP 40 30 20 House remains: r = 0.7038
80 60 40
0.4 0.3 0.2
0.1
10
0
0 0 0.1 0.2 0.3 Density 0.4 0.5 0.6
c)
100
% Corrected NISP
0.7 0.6
25 Cultural layer: r = 0.7708 20 House remains: r = 0.4612
80 60 40
0.5
Density
0.4 0.3 0.2
Corrected NISP
15
10
0.1 0
0 0 0.1 0.2 0.3 Density 0.4 0.5 0.6
d)
100
% Corrected NISP
0.7 0.6 0.5

Density
12 Cultural layer: r = 0.6746 10 8 6 4 2 0 0 0.1 0.2 0.3 Density 0.4 0.5 0.6 House remains: r = 0.7776
80 60 40
0.4 0.3 0.2
0.1 0
Figure 7.10: relationship between NISP and density by feature type for (a) cattle, (b) red deer, (c) roe deer and (d) caprines at Petnica. Corrected NISP values given as percentage of maximum, r values are Spearmans rho.
241
Corrected NISP
Taphonomic analysis
7.1.3 stage 1 summary

Density-mediated attrition has an appreciable but not overwhelming impact on observed skeletal part profiles at Gomolava. Correlations between NISP and density are frequently in the range 0.30.6, and occasionally as high as 0.7, using parametric or non-parametric coefficients. Values vary considerably but not consistently between taxa and context types. Attrition appears more severe at Petnica. In the overall sample, NISP-density correlations are consistently above 0.7 for taxa other than pigs. Attrition is most severe in the cultural layer, although the difference is only clearly expressed for pigs. Using MAU instead of NISP reduces correlations considerably at both sites, and removes the difference between context types at Petnica. It may thus be a more robust measure of predepositional skeletal part representation, although low correlations relate in part to the well-known tendency of minimum number counts to inflate proportions of rarer entities in small samples. Correlations between MAU and density remain above 0.5 for smaller taxa at Petnica, however. The degree of small bone completeness, especially at Petnica where reasonable standards of recovery applied, suggests that post-depositional destruction was fairly limited. This raises the possibility that peri-depositional processes such as weathering and carnivore/pig ravaging were important in assemblage formation.
7.2 Evidence for peri-depositional damage

7.2.1 overall frequency and impact on fragmentation
Overall frequencies of weathering and gnawing in Vina levels at Gomolava and Petnica are given in Table 7.6. Over 25% of specimens at Gomolava showed some cracking or flaking attributed to weathering, indicating that burial of specimens was not always very rapid. The near absence of markedly weathered specimens (< 4%), however, suggests that prolonged periods of surface exposure were rare. Weathering is more frequent and severe at Petnica, with nearly 40% showing some signs and over 10% markedly weathered. Gnawing is recorded on 15-20% of bones at each site, the vast majority being attributed to carnivores or pigs. Slightly higher frequencies at Gomolava may simply reflect the larger average size of specimens since gnawing is usually localized, smaller fragments are less likely to show traces regardless of original frequency. As with weathering, pig or carnivore access to bone material indicates that a substantial portion was exposed for some time between primary deposition and eventual burial. Almost all units at each site show some sign of weathering and gnawing, but rates
242
Taphonomic analysis
range widely, especially for weathering. The coincidence of higher weathering frequency and more severe density-mediated attrition at Petnica suggests that the role of the former in observed assemblage formation was considerable.
Weathering None Slight Moderate Heavy Total Gnawing None Possible Light Moderate Severe Rodent Digestion Total
Gomolava Frequency Percent 3278 72.47 1083 23.94 143 3.16 19 0.42 4523 100 Frequency 3789 121 496 228 66 8 2 4523 Percent 83.77 2.68 10.97 5.04 1.46 0.18 0.04 100
Petnica Frequency Percent 3747 61.78 1656 27.30 495 8.16 167 2.75 6065 100 Frequency 5255 159 632 5 14 6065 Percent 86.64 2.62 10.42 0.08 0.23 100
Table 7.6: frequencies and percentages of weathering and gnawing at Gomolava and Petnica. Teeth excluded.
Bar-Oz and Munro (2004, 206) advocate comparing frequencies of visible taphonomic variables with element abundance in order to test their impact, but the expected relationship is unclear. Different considerations could be invoked to predict either positive or negative correlation with abundance depending on exactly how density is assumed to affect susceptibility to weathering/gnawing. Comparing gnawing and weathering with specimen length to assess their impact on fragmentation is also problematic: there will obviously be a background positive correlation between gnawing and fragment length, since larger specimens simply have more surface area on which traces of gnawing might be preserved. This is also true for weathering, which was recorded according to the most severely affected part of the specimen. Gnawing shows no relationship with fragment length at Gomolava (Kruskal-Wallis: p = 0.3002). At Petnica, an almost imperceptible but weakly significant increase in average fragment length is seen amongst gnawed specimens (t-test: p = 0.0454; MannWhitney: p = 0.0028). A similarly tiny yet highly significant increase in length occurs with severity of weathering at both sites (Kruskal-Wallis: p value = 0.0000 at each site). Neither weathering nor gnawing seems to have increased fragmentation sufficiently to overcome the background correlation with size.
243
Taphonomic analysis
a)
60
A A A A AA
75
Percentag e weathered
Percentage gnawed
A A
40
20
A A A A A AA A A A A A A A A A A A A AA A A A A A A A A A A A A A AA A A A A AAA A A A A AA A AA A A A A AA A A A A A A A A A A AA A A A A A A AA A A A A A A AA A A A A A A A A A A A A A A
50
A A A A A A A A A A A A A AA A AAA A A A A AA A A A AAA A AAA A A A A A A A A A A A A A AA AA A A A AA A A A A A A AA A A A AA A A AA AA AAA A A A A A A A AA A A AA A A AA A A A AA A A A A A
25
A
0 0 50 100 150
0 0 50
100
150
Median length (mm )
Median le ngth (mm)
b)
100
A A A
40
Percentag e weathered
75
50
25
A A A A A A A A A A A A A AA A AA A A A AAA A A A A A AA A AA A A A A A A A A AA A A A A A AA A A AAA A A AA A A A A A AAA A A A A A A A A A A A A A AA A A A A A AA A A A A AA A A A A A A A A A A
Percentage gnawed
30
A A A A A A A AA A A A A A A AA A A A A A AA A AA A A A A A A A A A A A A A AA A AA A A A A AA A A A AAA A AA A A A A A A A A AA A AA A A A AA A A A A A A AA A A A A A A A A A
20
10
A A
0 0 25 50 75 100
0 0 25
A A A AA A A A A AA AA A AAAA
50
75
100
Median le ngth (mm )
Median le ngth (mm )
Figure 7.11: median long bone fragment length versus weathering/gnawing rates in individual units (NSP > 9) at (a) Gomolava and (b) Petnica.
More convincingly, no relationship between median fragment length and weathering/gnawing rates is observed at the unit level (Figure 7.11). Although poor recovery at Gomolava and the undiagnostic category at Petnica will have removed the smallest specimens, presumably obscuring some inter-unit variation, a considerable range of average lengths remains and is not correlated with either variable. Weathering and gnawing surely played an appreciable part in assemblage formation, but their specific impact on element profiles cannot be discerned at this point.
7.2.2 comparison between contexts, taxa and phases

Contexts at Gomolava can reasonably be split into pits, house remains and the cultural layer. The latter two have rough equivalents at Petnica, although the poor preservation of houses means that the distinction between them is less clear. The pits at Petnica are in no way comparable with those at Gomolava, and in any case contained very little bone.
244
Taphonomic analysis
Starting with Gomolava, the assemblages from the pits and the cultural layer show no significant difference in frequency of gnawing (2: p = 0.4304; see 4.4.3.3 for justification of statistical treatment), while that from the houses has a significantly higher frequency than either (2: p = 0.0000 in both cases). Unsurprisingly, weathering is significantly more common in the cultural layer than in pits (2: p = 0.0000) or houses (2: p = 0.0014), no significant difference being observed between the latter two (2: p = 0.5065). At Petnica, no significant difference is observed between the two main context types either for weathering or gnawing (2: p = 0.7942; p = 0.7355, respectively). Plotting frequency of weathering and gnawing at the unit level confirms these results (Figure 7.12), but reveals that overall differences in weathering rates between context types mask considerable overlap.
a)
10
Pit
15
Pit
Frequency
Frequency
8 6 4 2
10 5 0 15
Cultural layer
10
Cultural layer
Frequency
8 6 4 2
Frequency
10 5 0 15
Hous e
10
House
Frequency
8 6 4 2 0 25 50 Percentage weathered 75 100
Frequency
10 5 0
25
50 Percentage gnawed
75
100
b)
Frequency
Cult ural layer
Cult ural layer
Frequency
12 8 4 0
House
20 15 10 5 0
House
Frequency
Frequency
12 8 4 0 0 25 50 Percentage weath ered 75 100
20 15 10 5 0 0 25 50 Percentage gnawed 75 100
Figure 7.12: percentage of specimens gnawed in individual units (NISP > 9) at (a) Gomolava and (b) Petnica, split by context type.
The absence of any difference in gnawing rate between pits and the cultural layer at Gomolava suggests that bones deposited in the former either (a) became buried no faster than that elsewhere
245
Taphonomic analysis
i.e. pit-filling was very gradual or (b) included much secondarily deposited material. The higher rate of weathering in the cultural layer is hard to reconcile with this. It might favour option (b) if one argues that gnawing levels off with prolonged exposure while weathering becomes more marked: secondarily deposited material may have been exposed long enough to reach a plateau in the former but not the latter. Given the subtle difference observed, however, this is merely a suggestion. The house sample at Gomolava is small, but seems to show more evidence for gnawing compared to weathering than other context types. House remains at Petnica in general are indistinguishable from the cultural layer in this respect, but house floor units taken alone are more gnawed and less weathered than either the cultural layer or other house-related contexts. The difference is statistically insignificant, but would fit well with Gomolava. Considerable mixing is likely between context types at Petnica, which are relatively poorly-defined stratigraphically in any case. Remains of different taxa may have been disposed of in different ways, either as a direct result of cultural prescriptions or due to variations in modes of consumption. Differences in disposal may be reflected by the incidence of weathering and gnawing, shown by taxon as Table 7.7.
Taxon Bos taurus Bos primigenius Cervus elaphus Capreolus capreolus Caprines Canis Sus (all) Sus (domestic) Sus (wild)
n 1351 84 766 98 56 57 573 139 268
Gomolava % gnawed % weathered 19.69 30.22 16.67 33.33 15.27 22.88 14.29 12.09 8.93 11.32 7.02 10.91 24.96 26.01 25.18 32.61 22.39 23.13
n 885 1088 264 164 23 609 143 287
Petnica % gnawed 18.64 13.42 8.71 10.98 8.70 21.84 30.07 21.25
% weathered 50.06 44.49 23.48 33.54 17.39 44.99 48.95 47.39
Table 7.7: gnawing and weathering rates by taxon at Gomolava and Petnica.
Leaving aside pigs, both gnawing and weathering rates generally increase with body size. This probably reflects a number of factors, from recovery and average fragment size to scavenger preferences, but would be hard to interpret in terms of human behaviour. Pigs stand out as especially subject to gnawing and to a lesser extent weathering, but in this case the smaller domestic specimens are actually more affected. Plotting weathering and gnawing by both taxon and context type at Gomolava (Figure 7.13) reveals that the contextual difference in the former is driven almost entirely by pigs, which show a dramatic reduction in weathering rate when found in pits, suggesting that burial was much more rapid than for specimens deposited elsewhere. Since the difference between context types is much less
246
Taphonomic analysis
pronounced for cattle and red deer and the opposite seen for the relatively small aurochs sample one might suggest that deposition of pigs in pits followed a different pattern, perhaps taking place at intervals on a larger scale. That the same trend is not seen amongst pigs for gnawing implies that burial was not immediate, although the difference between wild and domestic specimens in this respect is hard to explain. High overall rates of gnawing and weathering on pig bones probably reflect physical properties rather than depositional practices.
35 Percentage gnawed 30 25 20 15 10 5 0
) ic ) C ap rin es (a ll) (w ild ph u ni ap re o au es t ig e Su s C an is Bo s ru s us s la lu s
Pits Cultural layer 45 32
B. t
C .e
rim
Su s
B. p
C .c
50 Percentage weathered 40 30 20 10 0
32
45
ru s
Su s
(d o
(w ild ) (d om es tic ) C ap rin e Su s
Bo s
(a ll)
us
lu s
ph u
ni
ap re o
au
ig e
B. t
Su s
la
C .e
rim
Figure 7.13: rates of gnawing and weathering by taxon in pits and the cultural layer at Gomolava. Figures for caprines, dogs and roe deer are given only for pits since these taxa are very rare in the cultural layer. Sample sizes < 50 marked on bars. Lighter shades represent wild and domestic sub-categories within Bos and Sus.
At Petnica, weathering and gnawing barely differ between context types for most taxa (Figure 7.14). Only roe deer are considerably less weathered when recovered from house remains, although they are joined by wild pig in terms of gnawing. An explanation cannot be offered at this stage, but it is notable that both roe deer and wild pig also have appreciably weaker NISP-density correlations
B. p
C .c
247
Su s
an i
Taphonomic analysis
in house-related contexts (Table 7.5), underlining the role of gnawing, at least, in density-mediated attrition.
35 Percentage gnawed 30 25 20 15 10 5 0
ru s (a ll) hu s ca pr eo lu ap rin e (w ild el ap Su s m es (d o Su s ta u tic ) s s )
Houses Cultural layer
43
Bo s
er vu s
Percentage weathered
70 60 50 40 30 20 10 0
ru s (a ll) hu s ca pr eo lu ap rin e (w ild el ap Su s m es (d o Su s ta u tic ) s s )
ap re o
lu s
Su s
43
Bo s
er vu s
Figure 7.14: rates of gnawing and weathering by taxon in house remains and the cultural layer at Petnica. Sample sizes < 50 marked on bars. Lighter shades represent wild and domestic sub-categories within Bos and Sus.
Trends across the three Vina phases at Gomolava are hard to assess given the change in feature types (6.3.5.3). Only the three most common taxa have sufficient samples for context type to be held constant while phases are compared (Table 7.8). The only notable change in weathering rate applies to pigs: specimens in Iab pits are considerably more likely to be weathered than those in Ia pits, while little difference exists in the cultural layer. Weathering amongst the other taxa is fairly constant. Pigs thus appear to come into line to some extent; whatever depositional practices underlie their relative lack of weathering in the Ia pits would seem to be less active by Phase Iab. If the small samples for definitively wild and domestic pigs are trusted, the discrepancy between cultural layer and pits is greater for wild specimens in both phases, and virtually disappears for domestic pigs by Iab. It is worth recalling the apparent selection for wild pigs in pits (6.3.5.7).
C ap
re o
lu s
248
Su s
Taphonomic analysis
Phase Species Weathering Ia Bos taurus Cervus elaphus Sus (all) Sus (wild) Sus (domestic) Iab Bos taurus Cervus elaphus Sus (all) Sus (wild) Sus (domestic) Gnawing Ia Bos taurus Cervus elaphus Sus (all) Sus (wild) Sus (domestic) Iab Bos taurus Cervus elaphus Sus (all) Sus (wild) Sus (domestic)
Cultural layer Percent n 31.46 21.36 41.67 55.56 41.18 32.01 25.93 38.54 41.03 44.00 21.35 17.48 31.25 22.22 29.41 18.60 13.33 18.75 15.38 24.00 178 103 48 9 17 328 135 96 39 25 178 103 48 9 17 328 135 96 39 25
Pits Percent 28.40 26.94 15.14 13.73 20.34 26.26 29.49 24.62 23.08 41.67 21.84 13.47 25.70 24.18 25.42 22.22 17.95 18.46 17.95 0.00 n 412 386 284 153 59 99 78 65 39 12 412 386 284 153 59 99 78 65 39 11
Table 7.8: weathering and gnawing rates by phase and feature type for three main taxa at Gomolava. Wild and domestic pigs are greyed-out to represent unreliably small samples.
Frequency of gnawing marks decreases across the board between phases in the cultural layer, but in pits it increases for red deer, stays constant for cattle and decreases markedly for pigs. Since the overall decline may well relate to a reduction in the numbers of live dogs and/or pigs within the settlement a factor which cannot be controlled for given the poor recovery and consequent absence of digested bone the increase seen for red deer in pits is probably the anomaly here. The situation is much simpler at Petnica: weathering and gnawing are extremely consistent across the three phases for both main context types (Table 7.9). Even allowing for some inter-phase mixing, the lack of change is remarkable.
Petnica phase 2 1144 1169.8 755 729.2 1899 1623 1610.4 273 285.6 1896
Unweathered Weathered 2 = 0.9887 Not gnawed Gnawed 2 = 0.9909
1 282 257.5 136 160.5 418 369 353.3 47 62.7 416
3 1760 1758.7 1095 1096.3 2855 1899 1927.2 370 341.8 2269
Total 3186 1986 5172 3891 690 4581
Table 7.9: observed and expected frequencies of weathering and gnawing by phase at Petnica.
249
Taphonomic analysis

Evidence for weathering and gnawing is common and ubiquitous in both assemblages. Weathering is particularly widespread at Petnica, affecting nearly 40% of specimens, and may be a factor in the higher attrition rate at the site. The visibility of both processes appears to be speciesspecific partly following a body-size trend and this can probably be attributed to differing physical properties. Differences between context types, however, are suggestive of depositional practices. In particular, pigs deposited in pits at Gomolava appear to have been buried more rapidly than other taxa, on the basis of reduced weathering rate. Such differences are not noticeably reflected in NISP-density correlations at Gomolava, suggesting that peri-depositional processes were a relatively minor aspect of density-mediated attrition. At Petnica, clear context-related differences are restricted to wild pigs (gnawing) and roe deer (gnawing and weathering) but do seem to be expressed in NISP-density correlations, suggesting that gnawing at least was an important agent of differential destruction.
7.3 Breakage and fragmentation

7.3.1 overall levels
Fragmentation of bones is caused by a variety of processes and agents, both pre- and postdepositional. In situ attrition, weathering and gnawing have already been mentioned, but much breakage is likely to have been caused by direct human action, from initial butchery to processing for marrow or grease (Outram 2002, 51). Insofar as subsequent processes can be controlled for, the degree of fragmentation in an assemblage serves as a measure of the intensity of carcass use (Halstead 2007, 30; Serjeantson 2006, 122-123). The following analyses are thus aimed at: Establishing the contribution of direct human action to assemblage fragmentation at Gomolava and Petnica (and hence improving our understanding of the context types and their formation processes). Assessing the intensity with which carcasses of the various taxa were processed and consumed at the sites. Identifying any differences in modes of consumption between taxa and/or context types.
Fragmentation is assessed through identification rate, average fragment length, long-bone completeness, percentage of fragments with fresh breaks (Petnica) and a freshness index (FI; Gomolava). Details are given in 4.4.3.4. Table 7.10 shows overall values of these variables at each
250
Taphonomic analysis
site. The Gomolava data must be treated with caution since poor recovery will certainly have inflated both average length and mean long-bone completeness. Only diagnostic specimens are included from Petnica.
Length (mm) Median Mean 87 88 87 51 66 63 69 60 91.6 92.0 91.7 77.6 71.5 68.7 74.1 67.2 36.7 32.4 31.1 33.8 33.7 30.4 29.7 32.3 0.10 0.10 0.09 0.16 0.08 0.09 0.07 0.07 0.07 0.07 0.05 20.0 14.2 27.6 10.8 7.7 27.4 0.0 29.6 20.6 29.0 19.7 13.2 Long-bone completeness* 1 2 3 0.31 0.28 0.26 0.22 0.23 0.16 Breakage 'freshness'** 1 2 3 1.24 1.18 1.26 0.85 1.22 1.18 1.39
n Gomolava Pits Cultural layer Houses Pits Cultural layer Petnica Houses (all) House base House 'floor' House rubble 2350 2103 183 51 4822 1632 551 472 560
ID rate 71.1 65 68.3 49.0 55.5 57.8 55.7 58.7 58.9
60.5 64.6 28.7 0.10 1 = sheep-sized; 2 = pig-sized; 3 = cattle-sized
Table 7.10: overall fragmentation levels at Gomolava and Petnica. Very small samples are excluded. *mean values, by body size. **Petnica: % of fragments having fresh breaks; Gomolava: FI.
Both sites show higher levels of fragmentation in the large category. This would be expected from peri- and post-depositional processes alone the impact of which has been documented above but at Petnica higher levels of fresh breakage among large taxa suggest that they were broken down further by humans prior to deposition. The FI at Gomolava is steady between large and small taxa, but if the completeness differential was entirely due to peri-/post-depositional processes one would expect lower FI values for large species. These results suggest that there was no gross difference in the manner of processing larger and smaller animals, i.e. that both were used fairly intensively. Lack of evidence for more wasteful consumption of large carcasses implies that they must have been shared amongst several households, at least if one discounts large-scale preservation (Halstead 2007).
7.3.2 pot-sizing and marrow extraction

Pot-sizing the reduction of large elements to a size suitable for boiling is often invoked as evidence for everyday consumption of stews or soups (e.g. Marciniak 2005a, 130; Serjeantson 2006, 123). Arguments are typically anecdotal, resting on observation of similar-sized long-bone specimens during recording. Since eventual fragment lengths depend on anatomy as well as on human action, element-specific modal lengths probably just represent the portions into which bones break most readily. Common modes across multiple elements and taxa would provide convincing evidence for deliberate sizing. Figure 7.15 and Figure 7.16 show long-bone length distributions for major species at Petnica and Gomolava respectively, excluding shaft fragments with no epiphysis
251
Taphonomic analysis
present. Pot-sizing is conspicuous by its absence at Petnica, although it could be obscured by subsequent fragmentation. At Gomolava there is some sign of a common peak for cattle at around 100 mm but this is partly a product of poor recovery, which truncates the lower tails of the distributions. No clearer peaks emerge when pits and the cultural layer are plotted separately. Serjeantson (2006, 123-124) shows that around one third of pig long-bones and a smaller
proportion of cattle specimens at Neolithic Runnymede are roughly half complete. These specimens, often chopped once through the shaft, might represent breakage for marrow without subsequent processing. A couple of dozen such cattle specimens showed evidence that breakage was preceded by heating of the shaft (Serjeantson 2006, see also Marciniak 2005a, 131). Having been noted on nine specimens at Petnica, across various elements, taxa and context types, this phenomenon was recorded systematically at Gomolava (Table 7.11), proving most common on cattle metapodials
Humerus Bos primigenius Bos sp. Bos taurus Caprines Cervus elaphus Sus (wild) Sus (domestic) Total
Radius
Femur
Tibia 1 2
1 1 1 1 1 2 3 1
Metacarpal Metatarsal 1 1 1 3(2) 4(4) 1
11
Total 3 1 16 1 4 1 1 27
Table 7.11: evidence for post-burning breakage of long-bones at Gomolava. Numbers in brackets represent possible cases.
Figure 7.17 shows long-bone completeness by element at Petnica and Gomolava. Following Serjeantson, elements with no diagnostic zones present are excluded from the graphs. While the majority of specimens at Petnica are reduced to less than a quarter of the original bone, presumably by a combination of breakage for boiling or grease extraction and post-depositional attrition, a substantial number approach half complete, indicating less intensive use. Poor recovery at Gomolava has inflated proportions of relatively complete specimens, especially for smaller taxa. Nonetheless, roughly half-complete specimens are much rarer amongst medium and especially large species at both Vina sites than for pigs and cattle respectively at Runnymede, suggesting that bone processing usually went beyond marrow extraction.
252
Taphonomic analysis
Bos
Frequency
15 10 5 0
Ce rvus
Sus
Humerus
Frequency
15 10 5 0
Radius
Frequency
15 10 5 0
Femur
Frequency
15 10 5 0
Tibia
Frequency Frequency
15 10 5 0 15 10 5 0
Metacarpal
Metatarsal
50
100
150
200
250
50
100
150
200
250
50
100
150
200
250
Length (mm)
Length (mm)
Length (mm)
Figure 7.15: length distributions for long bones at Petnica.
The exceptions to this are large species metapodials: at both sites, more than 30% of those with any zones present are 40-60% complete. This phenomenon is largely driven by the cattle, tying in very well with the direct evidence of post-burning breakage for marrow. Peaks at around half-length are also visible for cattle metapodials in Figure 7.15. Serjeantson (2006, 128-129) argues that freshlyextracted warmed marrow was a delicacy associated with feasting at Runnymede. Marciniak makes the same suggestion for Kujavian LBK sites, where post-burning breakage is restricted to pits between longhouses (Marciniak 2005b, 152). Interestingly, of the 20 Bos long-bones at Gomolava showing probable or possible post-burning breakage, 18 derive from the cultural layer. The number of roughly half-complete metapodials is steady across context types, however: it is the pattern of burning rather than of breakage which is restricted. One might suggest that heating and then breaking bones for immediate marrow consumption was an expedient practice, while remains in pits represent more systematic processing.
253
Taphonomic analysis
Bos
Frequency
12 8 4 0
Ce rvus
Sus
Humerus
Frequency
12 8 4 0
Radius
Frequency
12 8 4 0
Femur
Frequency
12 8 4 0
Tibia
Frequency
12 8 4 0
Metacarpal
Frequency
12 8 4 0
Metatarsal
50
100
150
200
250
50
100
150
200
250
50
100
150
200
250
Length (mm)
Length (mm)
Length (mm)
Figure 7.16: length distributions for long bones at Gomolava.
7.3.3 contextual differences

Other evidence for context-related differences in fragmentation is extremely limited. Various measures are shown by species and context type in Figure 7.18 and Figure 7.19, for Gomolava and Petnica respectively, with full data in Appendix 6. Bars are omitted where sample sizes are intolerably small (< 11 for lengths, < 16 for completeness/FI). Mean lengths for pits and the cultural layer at Gomolava are extremely similar, while few taxa can be compared in terms of completeness or FI. Wild pigs may be more complete in the pits but this is very subtle. FI values are generally slightly lower in the cultural layer but this can probably be attributed to greater weathering which both obscures fresh breakage surfaces and introduces new breaks especially since the largest differences are for pigs.
254
Taphonomic analysis
a)
60%
Hum erus
Radius
Fem ur
Tibia
Metacarpal
Metatarsal
Percent Percent Percent
Sm all
40%
20%
0%
60%
Medium
40%
20%
0%
60%
Large
40%
20%
0%
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
b)
50%
Hum erus
Radius
Fem ur
Tibia
Metacarpal
Metatarsal
Percent Percent Percent
40% 30% 20% 10%
Sm all
50% 40% 30% 20% 10%
Medium
50% 40% 30% 20% 10% 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0
Large
Figure 7.17: long-bone completeness by body size and element at (a) Petnica and (b) Gomolava.
Bones at Petnica are generally smaller, presumably due to improved recovery. The relative lengths of different taxa are very similar at both sites, however, and are probably determined to a great extent by the physical/mechanical properties of the bones. Context types again differ very little. Material from house-related contexts generally appears slightly more fragmented, but since it also features a lower percentage of fresh breaks this is probably a peri-/post-depositional effect. Caprines buck this trend, being less fragmented in house-related contexts but with fresh breaks more evident.
255
Taphonomic analysis
Splitting the house sample into its rather ill-defined constituent context types (6.4.3) results in small samples, but fragmentation is generally slightly higher than the cultural layer in rubble contexts, and slightly lower in house base units. Pig long-bones, however, are on average twice as complete in rubble/floor contexts as in house base units. House floor units always tentatively identified lie between the two in some respects, perhaps due to mixing, but are generally closer to the rubble. The precise formation of these contexts remains unclear, but it seems likely that at least some of the base units are just that: the make-up layer on which houses were constructed. Their composition and condition is very similar to the cultural layer, with slightly reduced fragmentation perhaps reflecting protection from subsequent disturbance. Bones from house rubble contexts probably represent a combination of material within the house at the point of destruction and that deposited on and mixed into its remains subsequently.

Comparison between phases reveals virtually no change at either site. In general, the evidence from fragmentation is of remarkable uniformity, with very little evidence emerging for differential modes of consumption, either between taxa or contexts. Processing seems to have been fairly intensive across the board, with specific techniques only identified for cattle metapodials. None of this means that differing social contexts of consumption and/or disposal did not exist, only that they are not reflected in the processing of carcasses at a coarse level, and probably that conspicuous display with its resultant wastage was rare.
7.4 Visible human modification

7.4.1 burning
Burnt zooarchaeological specimens may sometimes represent food preparation practices, but typically relate to more intense heating as caused by, for example, in situ burning of deposits. Burning patterns may thus be informative about depositional processes and the nature of specific contexts as well as constituting direct evidence for a potentially destructive taphonomic factor. This section aims to: Assess the overall extent of burning at Gomolava and Petnica, its distribution across contexts, and the likely processes which caused it. Evaluate the impact of burning on fragmentation and preservation. Identify any interspecific patterns in burning which might reveal different treatment in terms of processing and/or deposition.
256
10
0
20
30
40
50
60
70
80
90
0.1
0.2
0.3
0.4
100
120
140
0.1
0.2
0.3
20
40
60
80
0
0
Bos taurus
Bos taurus
Bos primigenius Bos primigenius Bos taurus Canis familiaris Capreolus capreolus Caprine Cervus elaphus Sus (wild) Bos taurus Canis familiaris Capreolus capreolus Caprine Cervus elaphus Sus (wild) Sus (domestic) Sus (all)
Canis familiaris Capreolus capreolus Caprine Castor fiber Cervus elaphus Sus (wild)
Houses (all) Cultural layer
Canis familiaris
Capreolus capreolus
Caprine
Castor fiber
a) mean length - all bones
a) mean length - all bones
c) mean completeness
c) mean completeness
Cervus elaphus
Sus (wild) Sus (domestic) Sus (all)
Pits
Sus (domestic)
Sus (domestic) Sus (all)
Cultural layer
Sus (all)
Taphonomic analysis
Figure 7.18: measures of fragmentation by taxon and context type at Gomolava.
Figure 7.19: measures of fragmentation by taxon and context type at Petnica.
257
40 45
0.2 0.4 0.6 0
10
15
20
25
30
35
0.8
1.2
1.4
1.6
1.8
100 90 80 70 60 50 40 30 20 10 0
100
120
140
20
40
60
80
Bos taurus Bos taurus Canis familiaris Capreolus capreolus Caprine Castor fiber Cervus elaphus Sus (wild) Sus (domestic) Sus (all)
Bos primigenius Bos taurus Canis familiaris Capreolus capreolus Caprine Cervus elaphus Sus (wild) Sus (domestic) Sus (all)
Bos primigenius Bos taurus Canis familiaris Capreolus capreolus
Canis familiaris
Capreolus capreolus
Caprine
d) mean FI
Castor fiber
Caprine Cervus elaphus Sus (wild) Sus (domestic) Sus (all)
b) mean length - long bones
b) mean length - long bones
d) % with 'fresh' breaks
Cervus elaphus
Sus (wild)
Sus (domestic)
Sus (all)
Taphonomic analysis
Burning can only be definitively identified where bones have begun to carbonize or calcine. Nonetheless, subtler variations in colour and/or texture can be attributed to less intensive burning with some confidence (4.4.3.5). Since a variety of other factors influence bone colour and texture, there will inevitably be some misidentifications. Conversely, many bones that were subjected to heating will not bear visible signs. At a statistical level, however, frequencies of recorded burning are indicative of its actual extent. Table 7.12 gives overall frequencies of burning at Gomolava and Petnica. Superficial burning is fairly common, especially at Petnica, but carbonized and/or calcined bones are very rare at both sites. At Gomolava, burning is considerably less common in the cultural layer than in the pits, and rarer still amongst house remains. Carbonized bones are more common in houses, as might be expected given destruction by fire, although this hinges on only two specimens. The overall lack of burning implies that many bones from house contexts represent either later (probably still Vina) intrusions or inclusions from building substructures. The 75 oven specimens are from a single pe and seem to have been built into the substructure, hence their high burning rate.
Cultural layer Pit House Oven Total Cultural layer Pit House House base House 'floor' House rubble Total
Unburnt # % 1778 84.31 1902 80.22 158 86.34 53 70.67 3891 82.12 3537 73.05 44 86.27 1272 77.18 426 77.17 341 71.94 466 81.33 4853 74.19
Burnt # % 327 15.50 466 19.65 23 12.57 22 29.33 838 17.69 1209 24.97 5 9.80 339 20.57 113 20.47 123 25.95 95 16.58 1553 23.74
Carbonized # % 3 0.14 3 0.13 2 1.09 0 0.00 8 0.17 63 1.30 2 3.92 20 1.21 6 1.09 5 1.05 7 1.22 85 1.30
Calcined # % 1 0.05 0 0.00 0 0.00 0 0.00 1 0.00 33 0.68 0 0.00 17 1.03 7 1.27 5 1.05 5 0.87 50 0.76
Total 2109 2371 183 75 4738 4842 51 1648 552 474 573 6541
Petnica
Gomolava
Table 7.12: overall frequencies of burning at Gomolava and Petnica.
Burning is also less common in house remains at Petnica, but the tentative separation of context types within these provides further insights. The highest frequency is seen for floor units, lending some credence to the idea that these represent house contents at the point of destruction. House base units are rather less burnt than the cultural layer, although they probably consist of material derived from it. This perhaps suggests that in situ burning in the cultural layer affected previously deposited material, while that incorporated into make-up levels was protected. Finally, a very low burning rate in rubble units could also be explained by much burning in the cultural layer having taken place post-depositionally. Alternatively, bones mixed into house rubble may have a different origin to those in the cultural layer, whether a matter of depositional practices or of later intrusion.
258
Taphonomic analysis
Superficial burning does not have a detectable impact on mean fragment length at Petnica, with affected specimens actually slightly larger on average (t-test: p = 0.0314; Mann-Whitney: p = 0.0043). Recovery standards at Gomolava skew size distributions, but no correlation is seen between burning rate and either mean fragment length or identification rate at the unit level (Pearsons r = 0.0317 and 0.0984 respectively, where NSP > 14). Carbonized and calcined specimens are clearly smaller at both sites, although samples are too small for formal testing. Severe burning is also much more common amongst undiagnostic than diagnostic fractions at Petnica: it seems that burning only becomes a serious destructive factor at the point of carbonization. Higher rates of burning in pits at Gomolava are interesting given (admittedly limited) evidence for in situ burning from pit sections and occasional excavation diary entries, plus the analogy with Opovo (6.3.4.2). Burnt patches and layers in pits could alternatively relate to burning of large quantities of waste prior to deposition. The difference between pits and the cultural layer is relatively slight and burning rates in individual units fall within the same range for both main context types (Figure 7.20a). Values for the cultural layer are clearly concentrated at 0-20%, however, while those from pits are more evenly distributed, perhaps with peaks at 5-10% and 3035%. It seems that only some units are affected by whatever processes underlie increased burning, and also that certain features are particularly implicated (Figure 7.20b). The distribution is bimodal, with one mode at 15-20% corresponding to that for the cultural layer and another standing out clearly at 30-35%. Higher levels of burning in certain pits may represent (a) more frequent in situ burning or (b) differential treatment of material prior to deposition. Burning rates also vary between taxa, even within particular context types (Table 7.13). They are probably influenced considerably by differing physical properties: some bones will simply have accumulated visible traces of burning more readily. This seems to be true for (a) smaller species and (b) the typically denser bones of wild animals: roe deer, wild pig and red deer/aurochs exhibit more burning than caprines, domestic pig and cattle respectively, but the rate also decreases with bodysize. Nonetheless, the existence of second-order differences between species and context types demonstrates that taxa were treated differently to some extent. Most taxa at Gomolava exhibit more burning in pits than in the cultural layer, but this is particularly pronounced for wild pigs, whose remains are almost five times as likely to be burnt in the former context. The opposite is seen for aurochsen, although samples are relatively small. At face value, it appears that wild pigs were preferentially involved in the practices underlying increased burning in (some) pits. However, given the size and wild/domestic effects noted above, it is actually the specimens in the cultural layer which appear less burnt than expected, and the elevated weathering rates amongst pigs in this context can be recalled (7.2.2.2). In any case, differential effects across
259
Taphonomic analysis
taxa indicate that burning in pits cannot exclusively be attributed to a purely functional burningdown of organic deposits.
a)
15
Frequency
Cultural layer
b)
6
Frequency
10 5 0 15
Pit
Frequency
10 5 0
0
0 25 50 Percentage burnt 75 100
25
50 Percentage burnt
75
100
Figure 7.20: burning rates at Gomolava (a) by unit (NSP > 9) and (b) by pit (NSP > 19).
At Petnica, differences between context types vary widely from species to species. Pig and caprine specimens in floor units are less likely to be burnt than in the cultural layer, while the opposite is true for deer and cattle. The low rate of burning in house base units, noted above, applies to red deer and cattle but not pig, while that amongst house rubble holds for domestic pig and cattle but not for either deer species, and hardly for wild pig. These results are hard to explain at present. High percentages of burning on long-bone articular ends relative to shaft fragments are sometimes taken as evidence for roasting (Miracle 2006, 76; Russell 1999), a relatively wasteful form of cooking. The relevant data for Gomolava and Petnica are given in Table 7.14 larger numbers represent a higher ends:shafts ratio, and arguably therefore more frequent roasting. At Gomolava, values are higher for pigs than for cattle and red deer. At Petnica, only the deer species stand out with low ratios, while caprines and domestic pigs have particularly high values. Relative values between species may be affected by physical properties of their bones as much as by human activities, but the difference between sites for cattle is interesting, suggesting more intensive consumption at Gomolava. Comparing between context types, the only dramatic difference at Gomolava is for domestic pigs, which have a very low end:shaft burning ratio in pits. Wild pigs have a much higher figure in the pits but are rarely found in the cultural layer for comparison.
260
a) Gomolava
Bos taurus Bos primigenius Cervus elaphus Capreolus capreolus Caprines Canis familiaris Sus (all) Sus (domestic) Sus (wild) b) Petnica 78 0 87 23 22 2 1 55 14 23
Bos taurus Bos primigenius Cervus elaphus Capreolus capreolus Caprines Canis familiaris Castor fiber Sus (all) Sus (domestic) Sus (wild)
Cultural layer Pits Houses %A %B n %A %B n %A %B n 12.96 0.27 733 16.46 0.35 571 11.84 1.32 76 23.40 47 6.25 32 2 20.98 0.35 286 23.55 467 13.79 29 16.67 6 35.48 93 33.33 3 12.50 8 24.49 49 100.00 1 50.00 4 16.98 53 50.00 2 9.22 206 23.61 360 14.29 14.29 7 12.50 80 18.06 72 2 5.00 80 23.96 217 25.00 25.00 4 Cultural layer Houses (all) House base House 'floor' %A %B %A %B %A %B %A %B n n n n 15.70 1.80 777 13.26 3.03 264 9.09 4.55 88 21.79 2.56 25.00 8 16.67 6 33.33 3 27.53 2.90 861 26.38 1.63 307 18.42 2.63 114 35.63 1.15 36.92 1.03 195 41.67 5.56 72 22.22 5.56 18 60.87 8.70 24.03 2.60 154 8.70 7.25 69 11.11 11.11 18 9.09 4.55 29.41 17 33.33 9 100.00 3 27.50 5.00 40 20.00 5 1 27.52 1.55 516 17.92 2.89 173 32.00 50 14.55 1.82 22.61 2.61 115 11.54 1.92 52 18.18 11 14.29 7.14 31.30 0.41 246 21.13 2.82 71 33.33 24 13.04 % A: percentage of specimens burnt; % B: percentage of specimens carbonized and/or calcined. House rubble %A %B n 9.78 2.17 92 2 28.26 1.09 92 38.71 3.23 31 7.41 27 3 33.33 3 10.94 6.25 64 7.69 26 17.39 8.70 23
Table 7.13: burning rates by taxon and context type at Gomolava and Petnica. Small samples are greyed out; very small samples omitted.
261
Taphonomic analysis
Bos taurus Cervus elaphus Sus (all) Sus (domestic) Sus (wild)
% end fragments burnt / % shaft fragments burnt* All Cultural layer Pits Houses 0.71 0.63 0.78 0.81 0.69 0.92 1.17 1.41 1.10 1.07 1.67 0.48 1.07 1.11 0.63 0.65 1.02 0.95 1.55 0.84
Gomolava
Bos taurus 1.15 1.37 Cervus elaphus 0.67 0.69 Capreolus capreolus 0.86 0.86 Caprines 1.39 1.91 Sus (all) 1.19 1.21 Sus (domestic) 1.48 1.46 Sus (wild) 1.10 1.11 * humerus, radius, femur, tibia
Table 7.14: burning on long-bone shafts versus articular ends at Gomolava and Petnica. Small samples are greyed out; very small samples omitted.
Samples are too small for cross-phase comparisons at Gomolava without pooling pits and the cultural layer. At Petnica, the ratio for red deer is almost constant between phases 2 and 3. That for cattle decreases dramatically from 1.81 to 0.21, but this relies on precariously small samples of burnt long-bone ends. If trusted, it implies a move away from roasting cattle in the final phase. Comparison of burning by element is only viable for the three main species at the site level (Figure 7.21 and Figure 7.22). Concentrated burning on the feet taken as evidence for roasting limbs at Pupiina (Miracle 2006, 78) is not observed here, although there is some indication of it in the unreliably small samples of roe deer and caprines at Petnica. The distribution of burning is similar for cattle and red deer, especially at Petnica. Between sites it differs primarily in degree, although burning on skulls is notably rare at Gomolava. For pigs, the overall pattern at Petnica is broadly similar to cattle and red deer. At Gomolava, burning on the hindlimb matches deer at the site and pigs at Petnica but there is a remarkable absence of burning on the lower forelimb, the significance of which is unclear.
7.4.2 butchery marks

Butchery marks are fairly common at Gomolava and Petnica. The vast majority are cut-marks, with chop surfaces extremely rare. Whether representing disarticulation or filleting, such cuts are unlikely to relate to bone fragmentation and destruction, hence the analytical phase is skipped for this variable. Two aims remain: To assess the overall intensity of butchery at Gomolava and Petnica and its distribution across body sizes and contexts.
Petnica
262
Taphonomic analysis
To identify any interspecific patterns indicative of differential processing.
Halstead (2007, 29-30) links cut-marking to butchery intensity, arguing that large carcasses must be broken down further to produce similar-sized parcels of meat, and hence that larger species should have more frequent cut-marks unless consumed wastefully. He notes, however, that knife butchery often leaves no trace, making cut-marks a crude measure. Table 7.15 shows the distribution of cut-marks at each site. Unsurprisingly, they are most common in the other category including carnivores, beaver and hare where they often clearly derive from skinning. Otherwise, frequency of marks at Gomolava increases with body size as predicted. At Petnica, however, cattle-sized specimens have surprisingly few cuts, perhaps indicating more wasteful consumption. Within this category, cattle and red deer have similar levels of butchery. Sheep-sized species differ, however, with more cut-marking on roe deer than on caprines. Wild pigs are consistently more cut-marked than domestic pigs.
Bos primigenius Bos taurus Cervus elaphus Capreolus capreolus Caprines Sus (all) Sus (domestic) Sus (wild) Cattle-sized Pig-sized Sheep-sized Other* Total Bos primigenius Bos taurus Cervus elaphus Capreolus capreolus Caprine Canis familiaris Castor fiber Sus (all) Sus (domestic) Sus (wild) Cattle-sized Pig-sized Sheep-sized Other* Total
Cultural layer Marked All 4 47 35 733 14 286 1 6 0 8 7 206 1 80 4 80 66 1632 9 356 2 109 0 6 77 2103 0 38 30 5 0 0 7 27 3 21 106 39 23 13 202 8 787 867 196 151 16 40 520 115 248 2997 890 793 116 4796
% Marked 8.51% 0 4.77% 10 4.90% 3 0 0 3.40% 0 1.25% 0 5.00% 0 4.04% 15 2.53% 1 1.83% 0 0 3.66% 16 4.83% 3.46% 2.55% 0.00% 1 12 14 1 2 0 0 16 3 9 49 23 6 3 83
Houses All 2 76 29 3 1 7 2 4 159 15 7 2 183 6 267 309 72 70 7 5 176 53 72 1009 301 288 24 1622
% 13.16% 10.34%
9.43% 6.67%
8.74% 4.49% 4.53% 1.39% 2.86%
Marked 1 33 25 3 0 22 3 17 96 29 6 7 138 1 0 0
Pits All 32 571 467 93 49 360 72 217 1571 487 251 60 2369 9 8 1
% 3.13% 5.78% 5.35% 3.23% 0.00% 6.11% 4.17% 7.83% 6.11% 5.95% 2.39% 11.67% 5.83%
Petnica
Gomolava
17.50% 5.19% 2.61% 8.47% 3.54% 4.38% 2.90% 11.21% 4.21%
9.09% 5.66% 12.50% 4.86% 7.64% 2.08% 12.50% 5.12%
0 0 0 0 1 0 1
7 1 2 4 34 12 50
2.94% 2.00%
Table 7.15: distribution of cut-marks at Gomolava and Petnica. Small samples are greyed out; very small samples omitted.
263
Taphonomic analysis
Figure 7.21: anatomical distributions of burning at Gomolava. Templates courtesy of INRAP Archozoo: Bos M. Coutureau, after Pales and Garcia (1981) ; Sus - M. Coutureau, after Barone (1976); Cervus J.G. Ferri / C. Beauval.
264
Taphonomic analysis
Figure 7.22: anatomical distributions of burning at Petnica. Templates as 7.22.
265
Taphonomic analysis
Frequencies also differ between context types. At Gomolava, pits feature more cut-marking than the cultural layer, although this must be at least partly due to reduced weathering. Certain pits have particularly high numbers. Frequencies in house remains are higher still, but this is largely down to Structure 11. At Petnica there is no clear difference between houses and the cultural layer. This is presumably partly down to the conflation of three context types in the former case, although cattlesized specimens the only category large enough for reliable subdivision have similar levels of cut-marking in base, rubble and floor units. Cut marks are compared by element at the body size level, permitting inclusion of ribs and vertebrae. Figure 7.23 shows results for cattle-sized specimens at Gomolava and Petnica, divided by context type in the former case. Disarticulation is most evident, with cuts concentrated around the major limb joints. Cuts on the mandibles are mostly to the articular process, those on the scapula are primarily to the collum scapulae or glenoid process, and those on the pelvis cluster around the acetabulum. A paucity of cuts at the knee joint is probably due to severe fragmentation of distal femora and proximal tibiae. There is also considerable evidence of defleshing, however, with numerous cut-marks on vertebral processes and rib shafts. Filleting marks on long-bone shafts are more common than they appear since many affected fragments could not be identified to element: just over 5% of unidentified large mammal shaft fragments have cuts. If roasting of substantial joints was the norm for large animals one might expect less evidence for disarticulation below the shoulder/hip and perhaps also less sign of meat removal, which leaves clearer traces when undertaken prior to cooking (Halstead 2007, 29-30). On the other hand roasting followed by carving might still cause more cuts than everyday stewing (Russell 1993, 359). Nonetheless, it seems that large carcasses were typically dismembered thoroughly and much meat removed prior to cooking. The only evidence for wastefulness is a surprising lack of cuts on feet, suggesting that phalanges were typically discarded unprocessed. This is corroborated by their overwhelming completeness even at Petnica plus occasional refits within units.
266
Figure 7.23: distribution of cut-marks on cattle-sized mammals at Gomolava and Petnica. Elements excluded (shown transparent) where total NISP < 15. Template courtesy of INRAP Archozoo: M. Coutureau, after Pales & Garcia (1981).
267
Taphonomic analysis
Limb disarticulation differs little between context types at Gomolava, but cuts on the ribs and axial skeleton are much more common in pits. Since more-or-less the same elements and portions are represented in each context this is not a matter of different stages in a single reduction process primary versus secondary butchery waste but rather of different pathways. The nature of these is hard to assess. Butchery around rib-heads and vertebral corpuses might represent reduction of the carcass into smaller packets, but this is hard to argue for cuts rather than chops on rib shafts and vertebral processes, which predominate here. Rather, defleshing is implicated, particularly in the pits where most rib cuts are on the distal or midshaft, compared to less than one third in the cultural layer. The question thus becomes why this is not seen in the cultural layer. One possibility is that removal of meat from ribs and vertebrae eventually deposited in the cultural layer mostly took place after cooking, leaving fewer marks. The corresponding remains in pits might then represent waste from more systematic filleting in preparation for cooking. Alternatively, the paucity of cuts in the cultural layer could be attributed to stewing meat on the bone, in which case those from the pits could represent either preparatory filleting or perhaps carving of roasted meat. In either case, the pattern in the pits at least for ribs and axial elements represents a departure from the stewing commonly assumed to be the norm on Neolithic sites (e.g. Russell 1993, 359; Serjeantson 2006, 123). The same trend towards greater cut-marking in the pits is seen on ribs from the pig- and sheepsized categories. Samples within these size classes are too small for anatomical comparison between contexts, but overall distributions of cuts in the pig-sized class are given in Figure 7.24. The evidence is more ambiguous in this case, perhaps due to smaller samples, but again a combination of disarticulation and defleshing marks is present. Cut-marks are too few to separate the effects of context type and of phase at Gomolava, so differences in butchery may be temporal as much as contextual. The frequency of butchery marks at Petnica rises slightly between phases 2 and 3 overall from 3.95% to 4.74% but more dramatically for red deer and wild pig (Table 7.16). Greenfield (1999) demonstrates increasing use of copper knives for butchery at Petnica, rising from 5.88% of cuts (a single case) in Petnica 1 to 16.28% (seven cases) in Petnica 3. He argues that this has implications for the assumption that early copper tools were not utilitarian (1999, 806), speculating that some form of hardening may have been employed. Conversely, one might suggest that his results have implications for the assumption that butchery choices were purely utilitarian: it is easy to imagine particular social contexts of butchery in which the prestige of copper tools might be favoured over the practicality of lithics. Unfortunately, Greenfield does not specify contextual, taxonomic or anatomical provenances.
268
Taphonomic analysis
Figure 7.24: distribution of cut-marks on pig-sized mammals at Gomolava and Petnica. Elements excluded (shown transparent) where total NISP < 15. Template courtesy of INRAP Archozoo: M. Coutureau, after Barone (1976).
Marked Bos primigenius Bos taurus Cervus elaphus Capreolus capreolus Castor fiber Caprine Sus (all) Sus (domestic) Sus (wild) 18 11 4 3 0 9 1 7
Petnica 2 All 353 387 93 17 63 226 47 110
% 5.10% 2.84% 4.30% 17.65% 0.00% 3.98% 2.13% 6.36%
Marked 1 22 34 7 3 2 23 3 16
Petnica 3 All % 14 7.14% 484 4.55% 534 6.37% 122 5.74% 17 17.65% 127 1.57% 341 6.74% 86 3.49% 147 10.88%
Table 7.16: frequency of cut-marks by taxon and phase at Petnica.
269
Taphonomic analysis

Burning and butchery show distinct differences in processing and/or disposal between taxa at both sites, and between context types at Gomolava. The interpretation of these is far from clear. Elevated levels of burning in certain pits at Gomolava might represent an in situ process, or alternatively the disposal of remains processed in different ways and perhaps reflecting different consumption contexts. A combination seems likely, and in situ burning probably also took place within the cultural layer to a lesser extent. In either case, second-order differences between taxa and context types indicate selectivity in the remains which were burnt. Wild pigs, for example, are preferentially burnt in pits and remarkably unburnt elsewhere. Evidence for roasting is as ever questionable, but shows differences between wild and domestic pigs recovered from pits. There is also some suggestion that cattle were roasted more at Petnica. Frequency of butchery marks follows the expected size trend at Gomolava, but there is evidence that large animals were butchered less intensively at Petnica. Within the smaller size categories, wild pigs and roe deer are more frequently affected than domestic pigs and caprines respectively. The location of marks suggests that carcasses were usually thoroughly dismembered and probably often filleted prior to cooking. At Gomolava, the cultural layer and pits feature similar patterns on the appendicular skeleton, but dramatic differences for the ribs and axial skeleton once again indicate that the two context types received remains from different consumption pathways.
7.5 Body-part representation

The interpretation of anatomical element profiles in terms of human action is split into two parts: correlation between utility indices and element abundance; and visual inspection of body-part profiles. The impact of density-mediated attrition on abundance especially when measured by NISP (7.1.2) obviously complicates interpretation. Such attrition does appear to have reduced numbers of the most vulnerable bones, but considerable variation remains amongst the denser elements, demonstrating that pre-depositional selection (a) took place and (b) is detectable.
7.5.1 correlation between utility and abundance

The uses, shortcomings and technical considerations of utility indices are discussed in 4.4.3.6. In the present context, correlating abundance with the FUI has two aims: To detect any interspecific differences in body-parts deposited on-site that might be related to practical transport decisions. This primarily relates to hunted species, but would also apply to domestic animals slaughtered off-site.
270
Taphonomic analysis
To investigate the nature of deposition in different context types, especially at Gomolava. Strong over-representation of high-utility elements might characterize the classic feasting deposit, for example. Likewise, if certain context types preferentially received material from particular stages in the butchery process one might expect either positive or negative correlations with utility.
Since utility is a culturally specific concept, the analysis should be seen as exploratory: while correlations might reveal patterns worthy of further considerations, their absence certainly does not mean that selection for particular body parts has not contributed to observed anatomical profiles. In fact, selectivity which cannot be explained either by density-mediated attrition or by practical butchery/transport concerns may be of particular interest. Very few appreciable correlations are seen, of which only three have reasonable underlying samples. Even these are of highly questionable significance when assessed visually (Figure 7.25). No species stand out as having high-utility parts preferentially deposited on-site; nor do the major context types differ in the utility of deposited elements. FUI is a very poor predictor of abundance at each site.
7 6 5
Red deer in cultural layer at Gomolava
Pig-sized mammals in cultural layer at Gomolava
4 3.5 3
Cattle in house remains at Petnica
2.5 MAU 0 2000 4000 6000 FUI (Metcalfe & Jones 1988) 8000 2
MAU
3
2
MAU
1.5 1
2 1 0 0 1000 2000 3000 4000 5000 FUI (Metcalfe & Jones 1988) 6000
1
0.5 0 0 1000 2000 3000 4000 5000 6000
FUI (Metcalfe & Jones 1988)
Figure 7.25: scatter plots showing strongest observed correlations between MAU and FUI at Gomolava and Petnica.
7.5.2 body-part profiles

Body-part profiles by NISP and MAU are shown in Figure 7.26-Figure 7.30, split by taxon and context type. Percentages are used where sample sizes permit to aid comparison. Definitions of anatomical regions are given in 4.4.3.6, along with the methodology used to derive MAU scores. For individual taxa, the axial category is limited to pelvis and sacrum since ribs and most vertebrae were not identified to species. Likewise, neck here includes only atlas and axis.
271
Taphonomic analysis
Graphs using the two quantification methods should be read in different ways. NISP values are not corrected for symmetry or for the number of elements in each region, and would in any case be biased by differential fragmentation between elements. NISP-based profiles are thus uninformative taken alone, only becoming useful for comparisons between taxa and between contexts. In principle, MAU-based profiles show which body-parts are missing as much as which are present: given complete carcasses and perfect preservation and recovery all bars would be equal. Where body-parts are underrepresented, the missing bones must be accounted for by one of the following: 1. Differential destruction (density-mediated and/or linked to processing techniques). 2. Incomplete recovery due to limited excavation area or poor collection techniques. 3. Selectivity in elements brought on-site. 4. Removal of bones from the site by humans and/or carnivores. The effects of point (1) were assessed in 7.1 and are not negligible. Combining elements into larger groupings mitigates this somewhat since MAU values typically represent the denser elements in each (Stiner 1991; 2002), but differences remain (Pickering et al. 2003) The upper hindlimb category, consisting solely of the femur, is a case in point: the femurs generally porous articular ends are particularly vulnerable to attrition, and MAU values for this region are unsurprisingly typically low. Body-part groupings also reduce recovery biases, since most contain at least one large element. Feet are the exception, and their increasing under-representation with smaller body sizes is no coincidence. Incomplete recovery due to limited excavation is harder to control for. It is extremely unlikely that all carcasses represented in an excavation area will be represented only within that area, but biases should not exist unless (a) there are systematic differences in bone deposition across the site, or (b) the area is small enough for random variations to become important. The latter is not an issue for the 400 m2 studied at Gomolava, while sheer sample size at Petnica should even out random differences. Spatial patterning on a gross scale also seems unlikely given the rarity of spatial differentiation on Vina sites, but the discrepancies noted at Gomolava between Clasons sample and mine are worrying (6.3.5.4). Selectivity in body-parts brought to site may apply to hunted species, and also potentially to domestic animals if they are slaughtered off-site. Selective transport off-site is equally possible. Preservation and exchange of meat seem unlikely, especially on the bone, but certain classes of
272
Taphonomic analysis
material or material from certain contexts or events may have been deposited off-site. This was one possibility suggested to account for lack of change in taxonomic frequencies within the cultural layer at Gomolava after deposition in pits almost ceases (6.3.5.4). Most MAU graphs based on large samples are fairly flat. Cattle at Gomolava are typical: upper hindlimb and feet are underrepresented for the reasons mentioned above, but there is no reason to believe that any bones are genuinely missing. The MAU profile for cattle at Petnica is consistent with this interpretation given the greater impact of density-mediated attrition at the site (7.1.2.2). The profiles for red deer are not dissimilar to cattle especially by NISP but heads and necks are relatively underrepresented, reflecting either selective transport or special treatment of skulls, perhaps including some off-site deposition. Otherwise, even representation of body parts indicates that field butchery was not regularly undertaken. Mandibles are also underrepresented amongst aurochsen at Gomolava (Figure 7.27), although numerous cranial specimens were recovered. This may be a matter of identification: unmeasurable cattle skull fragments are easier to mistake for aurochsen than frequently measurable mandibular pieces. If heads are genuinely
underrepresented, as seems likely, the same two explanations can be offered as for red deer. In any case, the absence of wild cattle horn cores is interesting, especially since they are the most common aurochs element at Petnica (Figure 7.28). Discussion of smaller taxa dogs, caprines and roe deer is complicated by poor recovery. The rarity of all three outside the pits is certainly suspicious, but two factors suggest it may be genuine. Firstly, overall average fragment size does not differ between context types (6.3.5.4). Secondly, many of the specimens recovered from the pits near-complete dog skulls and tibiae, complete or near-complete mandibles from all three taxa would surely sometimes have been noticed in the cultural layer even if excavation was less careful outside of pits. Vina dog burials both whole and partial are known from pits at Opovo and in Gomolava Block I (6.3.4.2), and disturbed examples may explain the relatively intact specimens in Block VII. I would suggest that all three small species are indeed subject to a recovery bias across the whole site, but that their overrepresentation in the pits relates to a genuine depositional difference more than to improved collection. In the case of dogs this was caused by burials and/or deliberate head placements.
273
Taphonomic analysis
35 30 25 %NISP 20 15 10 5 0 0
Pits (total MAU = 56.75)
Cattle
Pits (n = 565) Cultural layer (n = 688)
20 15 %MAU 10 5
Cultural layer (total MAU = 68.5)
30 25 20 %NISP 15 10 5 0 %MAU
25
Red deer
Pits (n = 464) Cultural layer (n = 274) 20 15 10 5 0
Pits (total MAU = 44.625) Cultural layer (total MAU = 31.25)
90 80 70 60 %NISP 50 40 30 20 10 0 %MAU
40
Pig (domestic)
35 30 25 20 15 10 5 0
Pits (total MAU = 16) Cultural layer (total MAU = 21)
50
25
Pits (total MAU = 33) Cultural layer (total MAU = 14.125)
Pig (wild)
40 %NISP 30 20 10 0
Pits (n = 201) Cultural layer (n = 68) %MAU Pits Cultural layer MAU 20 15 10 5 0
25 20 NISP 15 10 5 0
4.5
Caprines
4 3.5 3 2.5 2 1.5 1 0.5 0
Pits Cultural layer
40 35 30 NISP 25 20 15 10 5 0
nd li m b fo re l im b fo re l im b H ea d dl im b et nc or e k Ax ia l N ec Fe hi n
Roe deer
Pits Cultural layer MAU
6 5 4 3 2 1 0
fo re l im b
Pits Cultural layer
nd li m b
ib le
Pe lv is
fo re l im b
dl im b
N ec
an d
le r /h or
Lo w er
U pp er
U pp er
U pp er
U pp er
An t
Figure 7.26: body-part profiles for the main species at Gomolava.
274
Lo w er
Lo w er
Lo w er
hi n
hi
hi
Fe
et
Taphonomic analysis
30 25 20 %NISP 15 10 5 0 MAU
3.5
Aurochs
3 2.5 2 1.5 1 0.5 0
Pits Cultural layer
40 35 30 25 20 15 10 5 0
Dog
Pits Cultural layer MAU
10 8 6 4 2 0
Pits Cultural layer
NISP
pp er hi nd li m Lo b w er hi nd li m b
el im b
ad
nc or e
or el im b
or el im b
vi s
ial
rf or el im
dl im b
ec k
ec
ib l
He
Ax
M an d
pp er fo r
in dli
Pe l
An t le r/h or
Lo we
pp er h
er f
pe rf
Figure 7.27: body-part profiles for aurochs and dog at Gomolava.
20
Beaver
15 NISP
Houses Cultural layer MAU
3.5 3 2.5 2 1.5 1 0.5 0 Mandible
Beaver
10 5 0
Neck
Upper Low er forelimb forelimb
Pelvis
Upper Low er hindlimb hindlimb
Lo we
Lo w
Up
rh in
Fe
4 3
NISP
Aurochs
Houses Cultural layer NISP
2 1 0
Up pe rf or el im Lo b we rf or el im Up b pe rh in dl Lo im w b er hi nd li m b ea d Fe et ec k or nc or e Ax ia l H N
20 18 16 14 12 10 8 6 4 2 0
Dog
Ax ia U l pp er fo re l im Lo b w er fo re l im U b pp er hi nd li m Lo b w er hi nd li m b
ea d
An t
le r/h
Figure 7.28: body-part profiles for beaver, aurochs and dog and Petnica.
Wild pig MAU profiles are fairly flat at both sites, apart from the expected paucity of femora and phalanges. Heads are perhaps underrepresented at Petnica by MAU, but the same is true for domestic pigs. Profiles for domestic pigs appear less complete in general, as might be expected given the high percentage of immature specimens (6.3.5.5; 6.4.4.5), which are probably more vulnerable to attrition. Low numbers of atlases, axes and lower hindlimbs (specifically metatarsals) may be an identification issue, since almost all measured specimens of these elements were assigned to the wild group from log-ratio scores (6.2.1.3). Differences in domestic pig profiles between sites cannot be explained in terms of attrition or identification, but given relatively small samples only the discrepancy for mandibles is convincing: these are clearly overrepresented at Gomolava and underrepresented at Petnica.
275
Fe et
ec k
et
Feet
Taphonomic analysis
Frequencies of recovered small carnivore bones are shown in Figure 7.29. Mandibles dominate, followed by lower limbs and feet, with only five other specimens recovered, suggesting that primary butchery usually took place elsewhere. Skinning marks, especially on mandibles and ulnae, indicate that most specimens were subsequently processed further. By comparison, domestic dogs at the site have all body-parts apart from feet represented by at least three specimens (Figure 7.28). Bear should probably be treated separately: while extremely rare overall, a butchered humerus indicates that at least one relatively complete carcass was brought on-site. Interest in beavers was also more than skin-deep, with a single pelvis and several butchered humeri and femora recovered alongside tibiae, metapodials and cranial elements (Figure 7.28).
25
Wild carnivores
20
NISP
Carnivora Lynx Otter Wolf Badger Marten Wildcat Fox Bear
15 10 5 0
Ax U ia pp l er fo re lim Lo b w er fo re lim U pp b er hi nd lim Lo b w er hi nd lim b H ea d N ec k Fe et
Figure 7.29: body-part profile for wild carnivores at Petnica.
No complementarity exists between context types at either site, and in fact clear contextual differences in element profiles are few and mostly head-related. Red deer mandibles, crania and antlers are all more common in pits than in the cultural layer at Gomolava, as are aurochs cranial elements. The same is probably true for roe deer, dogs and caprines. Dogs in particular are represented in the pits primarily by intact mandibles and several near-complete skulls, suggesting a rather unusual context of deposition. Pig cranial elements, meanwhile, are actually less common in pits. An additional difference between pits and the layer only becomes apparent at the body size level, with the inclusion of ribs and vertebrae: axial elements are considerably more common in the cultural layer for all size groups especially sheep-sized although, as noted above, they exhibit more butchery traces when found in pits. At this level samples are large enough to separate the otherwise conflated effects of phase and context type at Gomolava (Table 7.17). For cattle-sized specimens, the axial:appendicular ratio is a factor of phase, while the cranial:postcranial ratio depends primarily on context type. Data for pig-sized animals are less clear, but both ratios appear to differ over time and between pits and the cultural layer. The significance in changing axial:appendicular ratios over time is unclear, but is
276
Taphonomic analysis
presumably butchery-related. It therefore seems likely that differences in cut-marking between context types (7.4.2) are also primarily temporal.
45 40 35 %NISP 30 25 20 15 10 5 0 0 %MAU House (total MAU = 20.625) Cultural layer (total MAU = 53) 20 15 10 5
Cattle
25 Houses (n = 270) Cultural layer (n = 786)
30
Red deer
25 20 %NISP 15 10 5 0
25 Houses (n = 307) Cultural layer (n = 841) %MAU 20 15 10 5 0
Houses (total MAU = 25.25) Cultural layer (total MAU = 57.125)
50
Pig (domestic)
40 % NISP 30 20 10 0
Houses (n = 53) Cultural layer (n = 115)
40 30 %MAU 20 10 0
30 25 % NISP 20
Houses (n = 75)
Pig (wild)
25 20 %MAU
Cultural layer (n = 248)
15 10 5 0
15 10 5 0
40 50 40 % NISP 30 20 10 0
Caprines
Houses (n = 71) Cultural layer (n = 152) %MAU
35 30 25 20 15 10 5 0
45 40 35 30 25 20 15 10 5 0
r /h or nc or e
Roe deer
Houses (n = 71) Cultural layer (n = 194)

%MAU
30 25 20 15 10 5 0
%NISP
hi nd li m Lo w b er hi nd li m b
lim b
Fe et
ea d
hi nd li m b
ib le
im b
im b
Pe lv is
Ax ia l
pp er
pp er
pp er
er fo r
An t le
Lo w
pp er
Figure 7.30: body-part profiles for the main species at Petnica.
277
Lo w
Lo w
er hi nd li m b
M an d
fo re
er
fo r
fo r
el im b
el
el
Fe et
ec
ec
Taphonomic analysis
Only cattle and red deer were sufficiently numerous for consideration in house contexts at Gomolava (Figure 7.31). Red deer are represented similarly to elsewhere, but cattle heads specifically mandibles are surprisingly rare. Differences between the cultural layer and house remains at Petnica are of dubious significance given the conflation of context types in the latter category, and are rarely clear, although it may be significant that cattle and caprine heads are rarer in houses. Cattle and red deer are actually remarkably consistent between base, floor and rubble contexts, while erratic figures for pigs are attributable to small sample sizes.
Cattle-sized
Pig-sized
Ia Iab Ib Ia Iab Ib
%NISP - cultural layer Head Axial* Appendicular 17.6 21.2 61.2 18.4 29.9 51.7 15.4 37.8 46.8 25.0 32.1 42.9 29.2 42.1 28.7 18.8 41.7 39.6
n 170 1038 188 28 216 48
Head 24.9 23.5 24.9 38.3
%NISP - pits Axial* Appendicular 22.9 52.2 25.8 50.7 29.0 23.3 46.0 38.3
n 1217 221 389 60
Table 7.17: head, axial and appendicular specimens at Gomolava by phase and context type. Small samples shaded. *includes all vertebrae, ribs and pelvis.
35 30 25
%NISP
Cattle (n = 76) Red deer (n = 29)

MAU
20 15 10 5 0
Ax ia Up l pe rf or el Lo im w b er fo re l im U b pp er hi nd li m Lo we b rh in dl im b He ad Fe et ec k co re N
5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

an di bl e Pe lv is dl im b b pp er fo re or el im b pp er hi n in dl im b er h ec k N l im
Cattle Red deer
An t le r/h or n
Lo w
er f
Figure 7.31: body-part profiles for cattle and red deer in houses at Gomolava.

MAU profiles are typically relatively flat, with missing elements mostly accounted for in terms of preservation and recovery. There is some evidence for selective transport of red deer and aurochs in the form of underrepresentation of cranial elements, but this could also be explained by special treatment of skulls. Profiles for caprines and domestic pigs diverge further from complete representation, and while much of this may reflect their vulnerability to post-depositional attrition, differences between the two sites imply that other factors are at work. Underrepresentation of both wild and domestic pig heads at Petnica is difficult to explain other than by special treatment, perhaps including transport off-site.
278
Lo w
Fe et
Taphonomic analysis
Differences between context types are very limited apart from higher cranial:postcranial ratios in pits than the cultural layer for red deer and probably other taxa at Gomolava. Both wild and domestic pig heads are actually more common in the cultural layer than in the pits. Cattle skulls and mandibles are roughly equally common in pits and the cultural layer but are rarely found in houses. Higher appendicular:axial ratios in the pits represent changing butchery practices over time.
7.6 Spatial distribution at Gomolava

The final stage, not strictly part of the five-fold taphonomic analysis structure, involves the distribution of taphonomic variables across a site. This analysis is restricted to pits at Gomolava due to the small excavated area and paucity of discrete features at Petnica and has two aims: To provide further insights into the nature of pit-filling. To reveal any spatial patterning in faunal assemblages which might relate either to structuring of the depositional practices involved or to issues of differential access to and/or sharing of meat.
7.6.1 distribution of taphonomic modifications

A variety of taphonomic differences has been detected between the pits and cultural layer at Gomolava, and is interpreted as evidence for specific depositional practices. In some cases, particular pits appear to be implicated. Various variables are shown by pit in Figure 7.32 with the aim of determining (a) whether pits can be divided into those involved in non-standard deposition and those not, and (b) whether any taphonomic modifications show spatial patterning above the feature level. Weathering is relatively even across the pits while the other variables vary widely. There is little sign of any relationship between them, however, nor does any spatial clustering emerge. The sheer range of variation indicates that pit-fill formation was anything but uniform, but one cannot single out individual pits responsible for the overall differences seen between them and the cultural layer.
7.6.2 distribution of taxa and body parts

The manner in which the distribution of taxa and body-parts across the pits should be interpreted rests on the nature of pit-filling, revisited here. The sheer number of pits and the volume of their contents representing 88% of Gomolava Ia faunal material in the NE half of Block VII
279
Taphonomic analysis
indicate that they were not reserved for special events, as confirmed by both the patchiness and the quantitative rather than qualitative nature of evidence for varying depositional practices presented above. Rather, I follow Russell (2000, 43) in arguing that pit fills represent a combination of everyday deposition (matching that in the cultural layer) and more restricted practices, although I steer clear of the word feasting for now. Proportions of material from everyday and special social contexts probably vary widely between features, but data quality is insufficient to pursue this. One might well ask why the cultural layer should be taken as a baseline and pits considered special, especially given that the latter account for the bulk of material in the earliest phase. One reason is the ubiquity of the cultural layer during Vina occupation at Gomolava, with little change in taxonomic frequencies throughout. Another requires a return to Chapmans (2000b; 2000c, see 6.3.4) arguments regarding the significance of deposition in each context type. While both are subject to his concentration principle, pits are argued to have particular significance through exchange of material with the natural world and/or the ancestors, making them likely recipients of deposits carrying particular meaning. Acceptance of this concentration principle linking deposition to houses is crucial to my interpretation of the pits. As set out in 6.3.4.5, the working hypothesis is the Phase Ia examples represent some variant of household clusters, despite the lack of visible foci of dwelling, and that their contents are therefore potentially informative regarding sharing of and differential access to meat. It is recognised, however, that this is a speculative interpretation. Prizer (1994) argues for inter-household sharing of red deer meat at Opovo from refits between widely-spaced pits. No dedicated refitting study was undertaken at Gomolava, but it is notable that the seven refits recorded from various species were all within single units despite prior sorting of the material by species and element. Given the incomplete nature of the collection, however, this carries little weight. Some flow of meat on or off the bone between households, arguably represented by pit clusters, seem very likely a priori, especially given the social tensions posited for the period. Referring to similar issues in the Aegean Neolithic and Early Bronze Age, Halstead argues that:
settlements made a substantial material investment in projecting solidarity at the level both of the local community and of its constituent domestic groups. Crosscultural ethnography suggests that commensality is likely to have played a vital role in affirming and negotiating solidarity at both social scales Halstead 2007, 26
280
Taphonomic analysis
Figure 7.32: selected taphonomic variables by feature in Gomolava Block VII. Phase Ia on left, Phase Iab on right.
281
Taphonomic analysis
Sharing of meat between domestic groups may have involved distribution of joints, collective consumption, or indeed a combination (Halstead 2007, 30-31). The potential equifinality introduced by patterns of differential access to carcasses and/or subsequent sharing was discussed in 4.5.3.6, with hypothetical outcomes of a number of scenarios set out in Figure 4.3. In general, sharing acts to smooth differences in taxonomic availability, but may or may not introduce variation in body-part frequencies depending on the manner in which it is structured. Greater anatomical than taxonomic variation thus provides good evidence for sharing, while the opposite suggests that it was limited. Broadly similar levels are inconclusive. Of course, comparing the two is difficult, with no benchmark for what constitutes appreciable variation between features (or clusters) for a given sample size. Ideally, inter-specific comparison of variability in body-part representation would highlight those taxa most involved in sharing as a null hypothesis, the larger species but only the cattle-sized sample at Gomolava is large enough to plot profiles by pit. These are shown in Figure 7.33 along with taxonomic compositions. There is considerable variation in taxonomic representation, but certainly no clear evidence for differential access to particular species, either between individual pits or the three or four clusters that could be drawn amongst them. Unsurprisingly, the greatest variations from the overall proportions are seen in the smallest samples, and should be treated with caution. Nor is there any strong sign of particular reliance on wild species, although Pits 13, 7 and 11 have unusually few domestic specimens. In the latter two features this relates partly to high numbers of wild pigs, which are probably subject to different depositional practices from other taxa (see below). Bodypart profiles do vary considerably and there is even some sign of similarity between adjacent features (e.g. Pits 10 and 11), but much larger samples not to mention comparisons between taxa would be required to discuss this with any confidence. As it is, one can only say that the data are consistent with the a priori likelihood of routine sharing acting upon differing initial access to carcasses. Correlation matrices provide slightly more objective impressions. Table 7.18 presents correlations between taxa (%NISP) across the largest Gomolava pit-samples (NISP > 50), plus a reworking of Russells (2000) results for the main pits at Opovo (DZ > 30). Taxa are ordered by median frequency, with the most common at top left. Strong negative correlations are expected by default in this corner of the table, but become less constrained towards the bottom right, since even large differences amongst minor taxa are easily absorbed by more abundant species. Amongst common species, positive correlations are reliable and even neutral or weakly negative coefficients may indicate a positive relationship.
282
Phase Ia
Phase Iab
Figure 7.33: (a) taxonomic and (b) anatomical (cattle-sized only) profiles by pit at Gomolava, by %NISP. Minimum sample sizes 27 and 39 respectively. Pie areas represent sample size.
283
Taphonomic analysis
a) Gomolava Cattle Red deer Wild pig Roe deer Domestic pig Caprines Dog b) Opovo Red Roe Pig* Cattle Caprines Dog
Red deer -0.3103 1
Wild pig -0.7095 0.0335 1
Roe deer -0.4344 0.0276 0.2470 1
Domestic pig -0.4002 -0.1054 0.2922 -0.3031 1
Caprines -0.3611 -0.2358 -0.0812 0.1335 0.3899 1 Dog 0.2340 -0.8000 -0.8727 0.4697 0.6403 1
Roe -0.2437 1
Pig* -0.3404 0.5530 1
Cattle -0.5773 -0.6177 -0.2426 1
Caprines -0.0919 -0.5476 -0.4238 0.3581 1
Dog 0.1877 -0.1484 -0.2330 -0.5330 0.0284 -0.1540 1 Aurochs -0.0548 -0.4362 -0.6338 0.3665 0.5963 0.5423
Aurochs 0.2631 -0.5937 -0.0502 0.0444 -0.2419 0.1442 -0.2975
Table 7.18: correlations between frequencies of major taxa in individual pits, by (a) %NISP at Gomolava and (b) %DZ at Opovo. Opovo data from Russell 2000, figures 4.2-4.14. *mostly wild.
Cattle at Gomolava show the expected negative correlations with other major taxa, while neutral or weakly positive coefficients between red deer, roe deer and wild pig imply a positive relationship. The picture is similar at Opovo, although red deer are somewhat detached from pigs (mostly wild) and roe deer, the latter being very strongly related. Caprines appear positively correlated with cattle but not pigs or roe deer, although they are far enough down the table for this to be questionable. Overall, there is some suggestion that wild species cluster at each site, which could be taken as evidence for subtle differences in reliance on wild and domestic resources. Russell (2000, 47) goes on to interpret varying correlations between adjacent elements as evidence that meat circulated primarily in joints disarticulated at the shoulder and hip, at least among the larger taxa. Interspecific variations in correlations between right and left elements are interpreted as a size effect with larger species shared more and hence subject to random variations in right/left frequencies but in fact cattle stand out (Table 7.19). This may simply reflect the rarity and hence unreliability of cattle at Opovo, but at face value constitutes tentative evidence that they were shared more. Samples at Gomolava do not support such treatment, and even the Opovo data are worryingly sparse at this level.
Scap. Red deer Roe deer Pig Cattle 0.835 0.753 0.395 -0.468
Forelimb Hum. Rad./ulna 0.554 0.754 -0.078 -0.102 0.878 0.669 0.394 0.8
Average 0.756 0.725 0.237 0.077
Pelvis 0.782 0.905 0.933 -0.662
Hindlimb Femur Tibia -0.067 0.655 0.488 -0.218 0.393 0.218 0.876 0.237
Average 0.369 0.593 0.766 -0.214
Overall average 0.563 0.659 0.501 -0.069
Table 7.19: correlations between right and left elements in pits at Opovo. Data from Russell 2000.
284
Taphonomic analysis
7.7 Discussion
My discussion of the taphonomic data from Gomolava and Petnica starts from the unusual assertion that there is precisely no evidence for feasting in the classic sense of extravagant, wasteful mass consumption events, and that in all probability such events did not occur. There is evidence, however, that certain species notably pigs were sometimes treated in systematically different ways, and that these practices are preferentially represented in pits. Since many differences cannot be explained in mechanistic taphonomic terms they must be underlain by different patterns of slaughter, consumption and/or deposition of the animals involved, presumably occurring in differing social contexts. Feasting in its broadest sense, of communal consumption on an extra-quotidian basis, is therefore implicated. The evidence for differing treatment between taxa and context types is summarised in Table 7.20. Some of the clearest and most intriguing trends apply to pigs, especially the wild variety, when deposited in pits at Gomolava. Weathering rates suggest that burial was typically much quicker than for other taxa, while the selection for wild individuals in pits is striking (6.3.5.4). Most taxa are more commonly burnt when recovered from pits presumably at least partly due to in situ firing but this is clearest by far for wild pigs. Rates of burning on long-bone ends versus shafts, arguably evidence for burning during cooking i.e. roasting rather than in situ, are also generally higher in pits, but again wild pigs have the highest value of all. Interestingly, domestic pigs have the lowest. Discussion of the kinds of events or practices involved inevitably enters the realm of speculation, but some likely aspects can be suggested. Given the overwhelming lack of evidence for wasteful consumption or varying processing intensity, large animals must have been shared, at least if one discounts the possibility of preservation and storage. These include aurochsen, cattle, red deer and wild pigs at the very least, and the special treatment of the latter cannot be explained as a reaction to surplus meat alone. Sharing of meat was probably routine, involving the distribution of joints as argued for Opovo (Russell 2000, 47), but the practices culminating in rapid deposition of burnt pig remains are set apart. They need not even have centred on consumption, although it is very likely to have been important: if carcasses were deposited un-processed this would surely be visible in the fragmentation and butchery data.
285
Stage Larger fragment sizes, scavenger preferences, etc. Physical properties of pig bones? Some secondary deposition in pits Different mode of deposition for pigs in pits? Fewer dogs/pigs on site? ? Some mixing, but still remarkable Intensive processing of all categories? Or post-depositional?
Site
Pattern
Interpretation
Both
Larger species generally more weathered and gnawed Pigs slightly more weathered and much more gnawed than expected
Gom.
Weathering more common in cultural layer than pits or houses No difference in gnawing between pits and cultural layer; higher in houses
Pigs (especially wild) far more weathered in cultural layer than in pits - much smaller differences for other taxa.
Gnawing decreases between phases in cultural layer; varies in pits
Pet.
Roe deer less weathered and much less gnawed in house remains
Extreme consistency over time
Both
Fragmentation increases with size category
Pet. Gom. Intensive processing No sign of 'pot-sizing'
Freshness increases with body size FI steady between small and large taxa
Human role in increased fragmentation: would expect lower values for large species if primarily post-depositional
Both
Most cattle/pig-sized elements reduced to < 25% Modal long-bone end-and-shaft lengths differ between elements and species
Gom.
Post-burning breakage on (mostly) cattle metapodials Almost always in cultural layer
Expedient practice of eating freshly-extracted, warmed marrow? Artefact of weathering ?
FI lower in cultural layer
Pet.
Higher fragmentation but lower 'freshness' in house contexts (especially rubble units) Opposite for caprines
4a
Both
Superficial burning common (especially at Petnica); severe burning rare Mostly in situ burning; maybe also different processing Most 'house' material NOT contents at point of destruction Interpretation of 'floor' units as house contents valid? Surface density? Differential processing of wilds probably unsupportable Common in situ burning following deposition, or different processing Roasting more common in pit deposits? Wild and domestic pigs treated differently in pits (less so in cultural layer)
Gom.
Burning more common in SOME pits than in cultural layer Burning rarest in houses, but severe burning probably most common (small sample)
Pet.
Burning less common in houses overall, but 'floor' units have highest rate
Both
Burning more common on smaller animals, and on wild taxa within each size category
Gom.
Wild pigs MUCH (>x4) more frequently burnt in pits Aurochsen very rarely burnt in pits (fairly small sample)
Generally higher end:shaft burning ratio in pits Domestic pigs have v. low ratio in pits, wild pigs v. high. But similar values in cultural layer
Burning rare on crania
286
4b Less intensive processing of large carcasses? Or pigs treated differently? Skinning Possible butchery differences along wild/domestic lines (note burning trend above), but doesn't apply cattle and red deer Not processed Partly artefact of weathering? Different processing pathways (temporal rather than contextual trend?) More intensive processing? Link (non-utilitarian) to apparent copper use? No clear-cut differential transport, or high vs. low utility deposits Complete representation on-site Field butchery and differential transport? Depends on nature of house contexts. Avoidance of deposition? ? Brought to site partially processed Whole carcasses brought to site; probably eaten Changing butchery practices Selective deposition Disturbed dog burials / head placements
Gom.
Cut marks increase with body size category
As expected if all sizes processed intensively
Pet.
Surprisingly few cuts on largest taxa relative to pigs Many cut marks on carnivores/beaver
Both
Roe deer more cut-marked than caprines; wild pig more than domestic pig
Lack of cut-marks on large mammal feet
Gom.
Cut-marks more frequent in pits than cultural layer, except for aurochs More butchery marks on ribs and vertebrae in pits Cut-marks very frequent in houses - but mostly due to Structure 11
Pet.
Increasing evidence for butchery over time, especially on red deer and wild pig
Both
No correlation between abundance and FUI
MAU profiles fairly 'flat' for cattle, allowing for differential destruction
Heads and necks underrepresented for red deer (and probably aurochsen at Gomolava)
Cattle heads rare in houses (also caprine heads at Petnica)
Pet. Gom.
Pig heads underrepresented Pig heads overrepresented, especially in cultural layer
Pet.
Carnivores specimens other than bear mostly heads (mandibles) and lower limbs Cut-marked upper limb bones from bear and beaver
Gom.
Post-cranial axial:appendicular ratio increases over time
Cranial elements more common in pits for red deer (and probably roe deer and caprines) Opposite true for pig
Dogs represented by remarkably intact specimens (mostly cranial, also tibiae)
Table 7.20: summary of taphonomic evidence from Gomolava and Petnica.
287
Taphonomic analysis
The most obvious explanation is that the sharing of wild pig following successful hunts took the form of a social gathering, perhaps hosted by the successful hunter, and that specific patterns of processing and disposal applied. I favour a rather different scenario in which hunting was motivated by the social gathering rather than vice versa. Wild pigs are certainly favoured for deposition in pits, but there is also a weaker selection for male domestic pigs and possibly for older individuals (6.3.5.7). Certain social events called for the communal consumption of wild pig, but a large domestic male was apparently an acceptable substitute if necessary. The nature of the events involved is a moot point, but the usual suspects may be mentioned: rites of passage, visits from exchange partners and so-on. The involvement of both wild and domestic animals, however, indicates that the primary motivation was neither display of wealth nor celebration of hunting prowess. Further support for this theory derives from the fact that the decline of pit digging/filling is not matched by any appreciable increase in wild pig remains in the cultural layer. The relative representation of wild pigs declines overall, but holds steady in each context type (6.3.5.4). Unless practices were moved off-site in later phases, it seems that most pig hunting ceased along with deposition in pits. The putative communal consumption events were not a response to the availability of wild pig meat but rather a motivation for its procurement. Since they apparently ceased along with large-scale pit digging/filling in general, it seems likely that the underlying motivations were related, and the idea of reinforcing ties to place through inclusive activities around dwellings can be recalled (6.3.4.2). The pigs-in-pits phenomenon cannot be discussed at Petnica since no large external pits are represented in the sample. House remains are known from all three Vina phases, but given the limited excavation area it is quite possible that pits along the lines of those at Gomolava Iab or Opovo did exist alongside them. Indeed, Je (1985) reports a Petnica 1 zemunica (pithouse) in Trench 1. Apart from casting some doubt on the validity of overall taxonomic frequencies from incompletely excavated sites, this makes it impossible to determine whether the special treatment of pigs at Gomolava was site-specific, although the sheer number of wild pig specimens recovered from the cultural layer and house remains at Petnica suggests that different depositional principles applied. Russell (1993) notes no special treatment of pigs from the taphonomic evidence at Opovo, and the predominantly wild specimens are no more frequent in pits than elsewhere. The possibility that special treatment of pigs was specific to Gomolava cannot be excluded. Less dramatic evidence for differential processing, consumption and/or disposal exists for other taxa. The main evidence for differing contexts of consumption relates to the pits and the
288
Taphonomic analysis
cultural layer at Gomolava. These context types are both present in phases Ia and Iab, although the relative quantities of deposition in each shift considerably. Their contents differ in terms of species composition, burning, butchery marks, weathering and, for some taxa, anatomical representation. In most cases these reflect genuine contextual differences as well as trends across the phases, but the assemblages from each are not complementary: they represent different patterns of consumption/deposition rather than stages in a single pathway. It is a reasonable assumption that these patterns stem from different combinations of consumption contexts. I am not suggesting anything so simplistic as a pit mode and a surface mode not least because of the likelihood of secondary deposition but at a statistical level deposition in each context type must have been underlain by different practices. Apart from the peculiar treatment of wild pigs, the principal differences in post-mortem treatment appear to be broadly size-related. The smallest major taxa dogs, caprines and roe deer are almost entirely recovered from pits, and initial suspicions of taphonomic bias appear unfounded. Domestic pigs are equally common in each context and red deer are only slightly overrepresented in the phase Ia pits, although this becomes more evident in phase Iab. Cattle, the largest-bodied common species, are also the only one to be more frequent in the cultural layer than the pits. Roughly size-correlated differences also exist in fragmentation, burning and cut-marking at both Gomolava and Petnica. If one accepts the argument that the pits are associated with groups approximating to households even in Gomolava Ia when actual dwellings were very rarely detected then the presence of smaller taxa may reflect their consumption primarily at this scale. Surface deposition of larger species, especially cattle, would then indicate a more communal mode of consumption. Since there is little indication of particular foci of surface deposition on Vina sites apart from around houses as demonstrated at Divostin (McPherron & Gunn 1988) I envisage this to have involved the distribution of joints rather than feasting per se. As Halstead (2007, 30-31) notes, consumption events may often combine ceremonial slaughter and butchery with the distribution of parts for consumption elsewhere. One possibility is that the consumption of distributed portions at Gomolava typically took place visibly around dwellings, with deposition in situ, in order to reinforce the relationships played out in the distribution of meat. The practice of extracting and probably immediately consuming warmed marrow from the long-bones of large mammals, especially cattle, should probably be seen in this context given its strong association with the cultural layer. Where large animal remains were deposited in pits this could be the result of truly communal consumption hosted
289
Taphonomic analysis
by particular households or groups, or of secondary deposition, or simply an indication that consumption and deposition practices were not consistent. An alternative account starts from the opposite assumption: that pits were spatially bounded communal places of deposition, perhaps reserved for particular classes of event. In this case, small animals would paradoxically seem to have been consumed communally more than large ones, and cattle least of all. Since the taphonomy points to intensive consumption of large species they must nonetheless have been shared, pointing once again to the distribution of joints amongst households, with eventual deposition on the surface, probably around dwellings. In either case, modes of consumption varied with body size but also over time, since the proportion of material in each context type shifts dramatically. While size seems to be the primary determinant of post-mortem treatment, there are some tentative indications of a secondary wild/domestic distinction in processing within each size class. At both Gomolava and Petnica roe deer exhibit more cut-marking than caprines, and wild pigs more than domestic pigs, although consistent differences are not seen between cattle and red deer. Wild pigs and red deer are more associated with the Gomolava pits than domestic pigs and cattle respectively. Higher rates of burning amongst wild species at each site are somewhat conservatively attributed to physical properties of their bones, but taking the taphonomic variables together a division in processing along wild/domestic lines certainly cannot be excluded. Butchery evidence also changes over time. A decrease in cut-marking at Gomolava relates to the decline both of pit-filling and of the more cut-marked wild taxa. Meanwhile, there is a clear increase in cut-marking between phases at Petnica, driven mainly by wild pig and red deer. The exact significance of cut-marking frequency is unclear since some forms of food preparation are more likely to leave marks than others, but Greenfields (1999) identification of copper knife-marks at Petnica implies that butchery decisions were more than utilitarian. Forms of butchery likely to leave marks may be related to social contexts as much as to practicalities. The principal conclusion regarding house-related contexts is a confirmation of mixed origins. The majority of material was probably deposited after destruction, with a substantial portion deriving from floor make-ups and only a minority representing the actual contents of the house when fired. There is some control for these three origins at Petnica, but given the poor state of preservation considerable mixing is inevitable. Various taphonomic patterns amongst house remains, such as very frequent butchery at Gomolava and surprisingly intact caprine bones at Petnica, are therefore hard to explain. Dead houses must have maintained a visible presence in the landscape and probably considerable significance (Tringham 2005, 106-107) for neighbours and former occupants, and there may well have been principles structuring
290
Taphonomic analysis
deposition within their bounds. A marked absence of cattle heads in houses at Gomolava may be of significance, especially since both cattle and caprine cranial elements are also less common in house-related contexts than elsewhere at Petnica. These are the two taxa most associated with live houses through bucranium installations.
7.8 Conclusions
Consumption of all major taxa at Gomolava and Petnica appears typically to have been intensive. None of the taphonomic data support the proposition that larger species might sometimes have been consumed in an extravagant, wasteful manner the classic feasting argument whether for celebratory or competitive purposes. Sharing of meat almost certainly occurred within the sites, but probably involved the distribution of joints on a routine basis. Sharing is extremely hard to document directly even if one assumes that pits at Gomolava are related to specific households, but there is some evidence that reliance on wild versus domestic taxa varied somewhat and was reduced by sharing. This would in any case be expected a priori. There is, however, clear evidence for different practices of processing and consumption between taxa and between depositional contexts. Most notably, wild pigs appear to have been selectively deposited in pits at Gomolava. A distinctive taphonomic signature indicates that this took place in an unusual manner, presumably in relation to discrete events. Whatever the nature of these, there is reason to believe that they were a motivation for pig hunting as much as a response to it. Subtler differences between context types and over time exist for other taxa, primarily involving patterns of burning and butchery. The precise significance of these is unclear, but they are taken to indicate differing consumption pathways, and there is very tentative evidence for differential treatment of wild and domestic animals, at least among smaller taxa.
291
Synthesis
Chapter 8 Synthesis and interpretation

8.1 Introduction
The previous chapters have considered both the overall economic roles of animal species in the Vina period and the taphonomic minutiae of post-mortem treatment and disposal at Gomolava and Petnica. It remains to bring these top-down and bottom-up analyses together in order to address (a) the character of human-animal relations in the Vina period, and (b) their contribution to wider questions of social change. Section 8.2 consists of a discussion of the evidence for human engagements with each taxon in turn. While perhaps reminiscent of the traditional archaeozoological report with its species-by-species considerations of morphology and ecology, the aim here is simply to respect the specificity of human-animal relationships. Other categories of archaeological evidence are also brought in at this point. In the next section (8.3) possible interpretations of the data in terms of change over time are considered. Finally, attention turns to the validity of the wild:domestic division in the Neolithic (8.4).
8.2 Human relationships with specific taxa

8.2.1 domestic pigs
The wide spectrum of possible human engagements with pigs is often noted (e.g. Albarella et al. 2007a), with many ethnographic examples blurring wild and domestic categories (e.g. Albarella et al. 2007b; Dwyer 1996a; Rosman & Rubel 1989). Domestic pigs in lowland New Guinea, for example, are very often the offspring of wild boars and domestic sows, or even of two wild parents (Dwyer & Minnegal 2005, 39). The degree of human control over the movement of domestic pigs also varies widely, from fencing and/or constant tending to nearcomplete freedom to roam. Amongst the Kubo, control of pigs is established through their familiarity with a particular human carer rather than through physical constraints, while for pigs kept by the Etoro the attachment is to people in general (Dwyer & Minnegal 2005, 3943). Similarly, Albarella et al. (2007b, 300-301) report entirely free-ranging pigs in Corsica that nonetheless recognise their owner on his occasional visits to their home range. Such examples constitute a good argument for the abandonment of strict notions of either physical control or reproductive isolation within our understanding of what it is to be domestic (3.2.2.5). Perhaps surprisingly, an altogether more clear-cut situation emerges from the Vina biometric data: the presence of distinct groups that cannot be attributed to sexual dimorphism (6.2.1.4) demonstrates effective reproductive isolation of wild and domestic populations. The extent of exploitation of these populations is uncorrelated; according to PCA, wild pigs tend
292
Synthesis
to be common where deer are frequent, while domestic pigs are grouped closely with cattle and caprines. Biometry cannot prove that interbreeding never occurred, especially since measurements overwhelmingly represent (sub)adult individuals. The situation described by Albarella et al. (2007b, 299) for Corsica and Sardinia in which interbreeding is considered inevitable but undesirable and the offspring slaughtered immediately would leave less biometric trace, although there are still wild sow/domestic boar unions to consider. Separation between groups is scarcely weaker for dP4 than for M2 and M3, however, suggesting that hybrid litters were rare at best. The frequency of wild pig remains typically 5-20% of NISP implies that they were fairly common in the vicinity of many Vina sites, so this isolation requires explanation. Given the probable Near-Eastern ancestry of the earliest European domestic pigs (Larson et al. 2007), some form of behavioural isolation from indigenous wild populations cannot be entirely excluded. Many examples of hybridization amongst subspecies of Sus scrofa exist (Groves 2007, 24), however, and I have already argued from biometry (admittedly hinging on a single site, Starevo) that most Vina domestic pigs probably had a local origin (5.2.5). Reproductive isolation must thus be explained in terms of human control, and three possible motivations can be suggested: 1. Deliberate prevention of interbreeding. Wild pigs may have been considered somehow dangerous or polluting to domestic herds, and indeed their consumption by humans, though frequent, does often seem to have been in special circumstances. The apparent substitution of domestic pigs in some cases, however, indicates that any conceptual boundary between wild and domestic populations was permeable. Practical beliefs about breeding might have the same effect, perhaps if relative docility was a valued trait. 2. Prevention of damage. More prosaically, pigs may have been tightly controlled to prevent crop-raiding and other damage, with isolation from wild populations an unintended consequence. Albarella et al. (2007b, 306) play down this factor, which obviously depends on the locations of forage vis--vis gardens or plots: if pigs are allowed to range fairly freely well away from cultivated areas it may not be a major concern. 3. Prevention of breeding within the domestic population. Individuals or households may have kept their own pigs separated from those of others to prevent uncontrolled breeding and concomitant ambiguities regarding ownership or stud
293
Synthesis
fees, with isolation from the wild population again incidental. However, breeding control could be maintained simply by tethering or enclosing intact males. Discussion of change over time is somewhat premature in the absence of sufficient data for a detailed biometric study similar to that recently conducted for the Italian Neolithic (Albarella et al. 2006b), but the existence of closer control during the Vina period would have different implications depending on which of the three possibilities above are taken to be important. Reproductive isolation could be taken to indicate the development of a conceptual wild:domestic boundary, at least as regards pigs. This is probably over-interpretation, however, and one might note that the Mafula of New Guinea maintain a clear ritual separation between wild and domestic pigs, which nonetheless form a single biological population (Williamson 1912, cited in Rosman & Rubel 1989, 31). Conceptual and genetic barriers cannot be assumed to coincide. Dwyer and Minnegal (2005) describe three forms of pig husbandry from the New Guinean ethnography in which pigs primary attachments are to individual people, to place (and people in general), and to other pigs, at low, medium and high altitudes respectively. At lower altitudes domestic males are typically all castrated, while some are kept intact for breeding in the highlands. They link these forms of husbandry to increasing population density and/or mobility, to the importance of exchanging live pigs difficult when bonding is to a specific carer and to labour requirements for gardening, noting that the domestic pigs which are most removed from contact with populations of wild animals are the ones whose contacts with other pigs are most like that of wild animals (Dwyer & Minnegal 2005, 45). Rappaports famous statement that the pig through its early socialization becomes a member of a Maring family (1968, 58-59) thus refers to one end of a spectrum in which increasing control of pig breeding is associated with decreasing intimacy between people and pigs. Extension of this New Guinean spectrum to the later Neolithic Balkans would be dangerous, to say the least, but it does give some idea of the potential variation in human-pig relationships. Superficially, the highland case seems the most appropriate analogy, with its relative sedentism and lack of interbreeding with wild pigs. However, the latter stems from their absence at high altitudes rather than from prevention by humans. Meggitt (1958) mentions frequent disputes over siring amongst one highland group, indicating that even in this case domestic boars were not always kept under tight control. Such analogies therefore shed no light on the reasons for reproductive isolation in the Vina case, but they do suggest some likely consequences. The attachment of pigs in many lowland groups to individual humans rather than other pigs is facilitated by isolation and hand-rearing of piglets, which are later encouraged to forage individually away from the village (Dwyer & Minnegal 2005, 39-
294
Synthesis
41). Access for wild boars is considered a benefit of this arrangement. Since appreciable numbers of pigs were kept by Vina communities it is hard to envisage a system in which they were both individually cared for and effectively prevented, intentionally or otherwise, from breeding with wild individuals, unless it was extremely labour-intensive. Perhaps more likely is a scenario similar to that described for the highlands:
though familiar with people, their primary associations are with an oftenchanging population of other pigs carers develop close attachment to them. It is unlikely, however, that this is reciprocated Dwyer & Minnegal 2005, 44
Such a system opens the way for regular exchange of live pigs, as hinted by the above quotation. This is extremely important in the New Guinean highlands, where transactions are typically structured by kin relations and specific social occasions (Sillitoe 2007, 331): the animals are made to stand for social relationships (1986, 16-17). These transactions drive the demand for pigs as much as vice versa. Efforts to prevent uncontrolled breeding amongst domestic pigs would have obvious social implications, indicating particular attitudes towards ownership of animals and more specifically their offspring. There is some evidence from age data that breeding was controlled within the domestic population. Specifically, the MWS profile from Gomolava shows remarkably clear peaks (6.3.5.8). If methodological problems can be discounted and I believe they can then this implies both seasonal culling and a suprisingly restricted birthing season. Since pigs are capable of breeding year-round (Lauwerier 1983), the very regular pattern observed here is suggestive of human manipulation, whether achieved deliberately or as a side-effect of intact males being controlled for social reasons. Even if isolation from wild pigs was maintained for other reasons, the very fact that domestic that is, owned boars were involved in reproduction introduces an extra dimension to pigs role in the political economy. Breeding rights are amongst the issues which make social transactions involving animals potentially much more complex than those mediated through other possessions; close attention to siring in this context would increase the overlap between human and pig socialities and kinships. It would also sit well with the arguments for enchained social relations within the later Neolithic (Chapman 2000a, see 2.3, 3.2.3), with animals constituting inalienable property, both they and their offspring potentially materialising exchanges between humans.
295
Synthesis
8.2.2 wild pigs

Wild pig remains are roughly as common as their domestic counterparts in the period as a whole, but there is a clear shift towards the latter over time. I will argue here that this represents the decline of pig hunting as a deliberate, targeted activity linked to specific social contexts. Evidence for the nature of pig hunting as a targeted rather than primarily opportunistic or protective activity comes from several sources. Firstly, there is an apparent focus on mature individuals, and especially males (5.3.1.4-5.3.1.5). Older age profiles amongst wild than domestic specimens are by no means inevitable, and indeed Higham (1968a) showed the opposite at Egolzwil, presumably indicating an equally focused hunting of piglets under six months. Deliberate hunting of adult wild pigs, especially males, implies a concern with the prestige involved in the activity, and probably only secondarily with the additional meat procured. Bringing down a mature boar they are usually solitary (Legge & Rowley-Conwy 1988, 18) would surely require an appreciable hunting party, and even with the aid of dogs would remain a dangerous pursuit. Secondly, very specific patterns of pig deposition at Gomolava were argued above to represent discrete events preferentially involving deposition of one or more wild pigs, presumably following consumption. The substitution of large domestic boars in some cases indicates that these events were not motivated by successful hunting, but rather that hunting may have been motivated by the circumstances calling for a consumption event (7.7). A late-spring/early-summer concentration of births is common in temperate populations of wild pig (e.g. Briedermann 1990; Mauget 1982). The MWS profile for wild pigs at Gomolava shows fairly clear peaks that as argued above for domestic pigs would be consistent with a seasonal kill (6.3.5.8). These are less regular in the wild case due perhaps to less restricted birthing or just smaller sample size but they generally line up rather well between the two populations. It is quite plausible that hunting of wild pigs was concentrated at the same times of year that the majority of domestic pigs were eaten. The decline in wild pig numbers in later phases of the site is clearly driven by the decrease in deposition in pits: background levels of wild pig deposition in the cultural layer remain fairly steady. When the practices underlying wild pigs specific mode of deposition cease, so does much of the hunting. Further corroboration of this hypothesis is provided by the age profiles (5.3.1.5; 6.3.5.5): the number of mature specimens decreases dramatically between Gomolava Ia and Iab, suggesting a much less targeted pattern of hunting in the later phase. Wild pigs were hunted throughout the Vina occupation of the site, probably on an
296
Synthesis
opportunistic or protective basis, but more active and dangerous pursuit became increasingly infrequent. Extrapolating this interpretation out to the rest of the region would be unwise. At Snandrei a similar decrease to Gomolava is seen between the Banat and Vina C levels, again associated with a reduction in average age-at-death. Wild pigs remain important throughout the occupation at Opovo, although the average age decreases between building horizons (Russell 1993, 318-319). The Opovo pigs mostly wild are no more common in pits than elsewhere on the site (Russell 1993, 187). Wild pig frequencies at Petnica actually increase over time. Unfortunately, the Petnica mandibular sample is uncomfortably small even when all phases are taken together, while a lack of large pits within the excavation area makes it impossible to say whether a similar situation to that at Gomolava ever obtained.
8.2.3 cattle
Although recovery bias has probably slightly exaggerated the predominance of cattle amongst Vina archaeofaunas, they are nonetheless the most common taxon by a wide margin, and their dominance increases substantially through the Vina period. This is particularly clear at several multi-phase sites, where I have already suggested (5.3.3) that it might reflect (a) settlement histories and trajectories of growth, or (b) an unusually strong expression of social changes taking place in the region as a whole, viz. intensification of production and increasing inter-household competition. Little evidence can be brought to bear on the manner in which cattle were herded in the Vina period. Bknyi (e.g. 1974a) has long argued for their local domestication in the region, with young aurochsen captured and tamed to supplement existing stock. This hypothesis has been comprehensively dismissed by Rowley-Conwy (2003), who argues that it is neither likely a priori nor supported by the biometry: while metrical separation implies effective genetic isolation, its absence in no way demonstrates genetic continuity. The local domestication hypothesis is further undermined by mtDNA evidence demonstrating distinct maternal lines for prehistoric European aurochsen and domestic cattle (Bollongino et al. 2006; Edwards et al. 2007), except possibly in Italy (Beja-Pereira et al. 2006). Interbreeding between wild males and domestic females whether unavoidable, tolerated or encouraged cannot be excluded, however, and appears to have contributed to modern Y-chromosome sequences in some parts of Europe (Gtherstrm et al. 2005). The clue to management forms provided by genetic isolation amongst pigs is thus not available for cattle. In the absence of stable isotope or dental microwear evidence, geographical herding patterns also remain obscure. Equally importantly, neither the absolute
297
Synthesis
nor the social scale at which herding was organized is known. While prehistoric herding should not necessarily be seen in terms of optimization for subsistence, the associated labour costs cannot be discounted. On this basis Halstead (1996) argues that large-scale herding in prehistory would require heavy specialization in a single domestic species, as well as a degree of mobility, of herds and herders if not of the whole human population. The parameters involved in the temperate climes of the Central Balkans would differ from his Greek example, but the basic reasoning holds. Accordingly, Greenfields (2008) argument for an essentially pastoral Starevo-Cri economy on the basis of high numbers of cattle and caprines at FoeniSala is unconvincing. By the second half of the Vina period, however, the dominance of cattle at many sites makes large-scale herding a distinct possibility. A lack of evidence for corrals or large pens on Vina sites suggests that cattle were either (a) largely kept off-site, where such evidence is unlikely to be found, or (b) kept on-site but dispersed and in small numbers. Certain buildings at Gomolava are suggested in the excavation diaries to have housed livestock. Ancillary buildings are associated with particular houses (e.g. House 2/73-House 3/73; House 1/73'-House 1/73; House 2/77-Structure 3/77; House 6'-House 6; see Figures 6.28 and 6.33), and the larger cases from Phase Iab could plausibly have housed cattle belonging to the occupants, although this is only one possible interpretation. For that matter, livestock could quite feasibly have dwelt alongside humans in the larger houses during this phase. Coupled with the zooarchaeological results, the reduction in size of houses and disappearance of large ancillary buildings in Phase Ib could represent a shift towards larger-scale herding, with cattle kept off-site and probably grazed at a greater distance, beyond cultivated plots. This might have involved the aggregation of herds above the household level for management purposes, which need not imply communal ownership. With their use for milking and/or traction also being debatable (5.3.2), it is hard to judge the closeness of human relations with cattle. That they will have been central to Vina peoples lives and increasingly so is undeniable; but the levels of intimacy and familiarity involved are open to question: if large herds were grazed well away from sites, live cattle would have been a less ubiquitous part of many peoples experience than less numerically important species such as pigs. At the same time, they would surely be central to the identity of those involved in the herding, as highlighted by Parker Pearsons Tandroy study (2000, 220-221). Such divisions are likely to have had a gendered aspect, but age and other factors may also have been important. Whittle (2003, 94) suggests a tension between the potential solitariness of herding and the commonality of periodic aggregations, characterized by large-scale consumption. The latter point is not supported by the Vina taphonomic evidence, but the
298
Synthesis
networks of social relations embodied by herds would in any case present a marked contrast with the solitary experience of mobile herding. Turning to the second aspect of domestication their role in property relations cattle are the only species in Neolithic Europe to meet the requirements set out by Ingold (1980, 224-226) for an animal to act as a store of wealth. Drawing on a range of ethnographic analogies, Russell (1998; 1999, 157-161) makes a strong argument for this having applied in the Vina case. Moderately high proportions of adult cattle are seen at Vina sites, and the presence of significant numbers of adult males in particular (5.3.2) implies that animals were sometimes kept alive for reasons other than efficient meat or milk production. Even if cattle were used for traction, a contentious issue (e.g. Chapman 1982; cf. Greenfield 1988; Sherratt 1981), this would presumably only increase their potential value as living property (see Bogucki 1993, 499). This value may have rendered cattle particularly suitable for the expression of growing tensions between household and community in the Vina period. This does not imply commodification, however. On the contrary, they are likely to have been inalienable possessions rather than wealth in a strict sense, as argued by Ray and Thomas (2003, 41) for Neolithic Britain. Exchanges of cattle, whether as bridewealth (Russell 1998), as loans for breeding or traction (Bogucki 1993, 499), or in other contexts will have been important in the creation and maintenance of social ties. As suggested above for pigs, individual cattle may have come to materialise these ties through biographies of involvement. With their relative longevity, slow growth and low fecundity resulting in low herd turnover cattle have particularly high potential as lasting records of human social relationships. Cattle and human socialities are likely to have been intricately intertwined (Pollard 2006, 139; Ray & Thomas 2003, 41). The proposed position of cattle as foci in the negotiation of social tensions may be reflected in the symbolic role they appear to have played in the Vina period. While zoomorphic figurines are typically either stylistic or non-specific, most of those judged to be representational resemble cattle (Bailey 2000, 184, referring to the Balkans as a whole), sometimes very clearly male (e.g. Milojkovi 1990, 416-417; ljivar & Jacanovi 2005). Of course, even assuming specific representational intent these could reference aurochsen rather than domestic bulls, but representations of wild species are otherwise rare. Referring to the earlier Neolithic of the region, Nanoglou (2008, 10) takes the representation of humans and domesticates but not wild species to indicate a focus on the relationship between humans and animals within the community, an argument which seems to hold also for the Vina period. At this level the evidence from figurines suggests some continuity from the earlier Neolithic in the place of domestic animals within the community, perhaps with an increasing focus on cattle. Small
299
Synthesis
clusters of apparent cattle figurines found in situ at some sites (Chapman 1981, 73; ljivar & Jacanovi 2005) have been taken as representations of herds (Marangou 1996). Bucrania are another line of evidence for the symbolic importance of cattle in the period. Found primarily within houses (Chapman 1998, 125), these have unsurprisingly been linked to ideas of cattle ownership and household wealth (Chapman 1981, 68; Russell 1998). Chapman expresses some caution in making this association (1981, 73; 2000a, 217), but given that the occurrence of bucrania increases along with the economic importance of cattle it is certainly not an unreasonable interpretation. The underrepresentation of cattle cranial elements in houses at Petnica and much more clearly Gomolava is interesting in the light of this probable association, and perhaps reflects principles structuring deposition in dead houses. The shrine at Para, where numbers of cattle remains increase dramatically in the earlier Vina period (5.2.5), included numerous real and modelled cattle crania (Lazarovici 1989; Lazarovici et al. 2001). Cattle as other large-bodied taxa appear generally to have been processed and consumed intensively, with very little wastage, indicating consumption on a level beyond the household. This may have involved communal consumption events without overtly competitive or extravagant elements, or distribution of portions between domestic groups, possibilities that are not mutually exclusive (Halstead 2007, 30-31). As argued in 7.7, the association of cattle with surface deposition suggests that their meat was commonly subject to dispersed consumption and deposition. At Opovo, meanwhile, there is some indication that cattle remains in the pits were more subject to dispersal than other taxa, even accounting for body size (7.6, see Russell 2000). Such distribution should probably be considered sharing in rather than sharing out in Ingolds terms (1980, 173-175; 1986, 233). In Chapmans (2000a, 5) language, the indissoluble links formed between an animal and all those involved in its siring, raising and subsequent life were respected in its butchery, distribution, consumption and deposition. Of course, this could apply to any owned animal, but as noted above the longevity of cattle make them particularly effective as vehicles and records of human relationships.
8.2.4 aurochsen
With its impressive size and reputed ferocity5, the aurochs has proven fascinating and enthralling, if not to Neolithic people then certainly to archaeologists. Interesting possibilities for exploring symbolism and classificatory systems are certainly presented by the existence of
A reputation partly owing to that most creative of sources, Caesars Gallic Wars.
300
Synthesis
a large and potentially dangerous animal, living away from human contact but nonetheless sharing unmistakable similarities with animals that were central to Neolithic life. This also describes wild pigs, however, and it is hard to avoid the conclusion that archaeological interest in aurochsen is fuelled partly by their present extinction. Unlike wild pigs, aurochsen in the Central Balkan Neolithic fall firmly within the domestic group derived from PCA, and an extremely high correlation between aurochs and domestic cattle NISPs (5.2.5) raises the spectre of widespread misidentification. The alternative
possibility that aurochsen were hunted more frequently at the sites where cattle were most important cannot be supported at this point, appealing as it may be. Distinguishing wild and domestic cattle in the southeast European Neolithic is certainly difficult but my own identifications are very conservative, considerably more specimens being assigned to Bos sp. than to Bos primigenius. While the relative numbers of cattle and aurochs may not be entirely accurate, and some of the latter may lurk amongst the former, I am fairly confident regarding specimens positively identified as aurochs. Even accepting all identifications, aurochs is a minor component of Vina faunal assemblages, especially by comparison to Middle and Late Neolithic groups further north on the Hungarian Plain (see e.g. Bartosiewicz 2005; Bknyi 1974a). Age data are therefore limited, but the eleven mandibles from Gomolava are mostly very young or senile, suggesting that aurochs hunting was largely opportunistic, in stark contrast to that of wild pigs. This would seem to be anything but hunting the mighty aurochs. Limited hunting does not imply that the occupants of Vina sites were unaware of, uninterested in, or unimpressed by aurochsen, however. Chapman (1981, 73) reports a complete aurochs skull, with horns, placed at the base of a ritual pit in Gomolava Block I, while Russell (1993, 84) mentions a horn core associated with a complete pot towards the bottom of a pit at Opovo. A large number of aurochs cranial fragments were found in the pits at Gomolava, along with many fragmentary domestic cattle horn cores. Given identification difficulties, it is plausible that both cores and crania include wild and domestic specimens. If so, aurochs and cattle skulls may have been interchangeable in this context. On the other hand several taphonomic variables at Gomolava give tantalising glimpses of marked differences in the post-mortem treatment of domestic cattle and aurochs, although samples are too small for detailed comment. Ray and Thomas (2003, 42) suggest that cattle and aurochsen were seen as parallel but contrasting communities in the Neolithic, and both aspects may be represented in their treatment at Gomolava. Aurochsen were probably a rare sight around most Vina sites and when encountered may have appeared as much an impressive archetype of the familiar domestic cattle as an opposed, contrasting form.
301
Synthesis
8.2.5 red deer

It has sometimes been suggested that deer were actively managed in prehistory (e.g. Jarman 1972), although convincing arguments for an intermediate position between wild and domestic typically involve transportation to islands (e.g. Sharples 2000). It is highly doubtful that red and roe deer herds could be maintained under human control without the aid either of geographical boundaries or of sophisticated fencing given the seasonally aggressive and territorial behaviour of males (Clutton-Brock et al. 1982; Legge & Rowley-Conwy 1988, 1518; Nowak & Paradiso 1983, 1225). While prey conservation may certainly have been a concern in some Neolithic cases, it is a far cry from sustainable hunting to herd management, let alone semi-domestication. Neither regular contact with humans nor a live role in property relations seems plausible for Vina period deer. Red deer are one of the most common species on Vina sites, and are likely to have been encountered relatively frequently. Given the probable wooded landscape around Vina sites, red deer hinds will have spent most of the year alone with their young, albeit with two or three family groups occasionally aggregating, while males will mostly have been solitary (Ahlen 1965; Legge & Rowley-Conwy 1988, 15-16): red deer will only have been encountered in any numbers during the September-October rut. A seasonal focus is only apparent at Selevac, where Legges (1990, 235) suggestion of hunting in the colder months of the year probably means after the rut. Varying sex-ratios between sites (5.3.1.4) may indicate differing hunting strategies. At an inter-site level red deer are associated statistically with the other main wild species. More tentatively, the frequencies of these taxa in individual pits at Gomolava and Opovo seem to be weakly correlated. Depending on ones interpretation of the pits, this might suggest variations in reliance on wild versus domestic species between households or other groups, albeit largely obscured by the distribution of meat. Evidence for symbolic importance of red deer is rare, but a bucranium featuring antlers was found at Gomolava (Na 1957, 404), and Russell (1993, 136) mentions a frontlet from Opovo. Higher proportions of red deer cranial elements especially antler and mandibles in the pits might represent a parallel with aurochs/cattle skulls and horn cores. Alternatively, deer hunted for particular events involving deposition in pits may have been brought to the site complete, while others were often butchered in the field to ease transport. The frequency of red deer decreases over time at Gomolava, as for other wild species. Unlike wild pig and roe deer, this is not clearly linked to the decline in pit digging/filling: red deer remains become scarcer even within the cultural layer. As with wild pig, however, the
302
Synthesis
reduction in frequency is accompanied by signs of less targeted hunting, and it seems likely that active pursuit declined over time in favour of opportunistic or garden hunting. A similar decrease in red deer frequencies is seen at Selevac, but demographic data are not available. At Opovo and Petnica little change is seen, and in the former case the average age of specimens actually increases over time. In general, a decline in red deer hunting is seen at sites where the importance of cattle increases amongst the domesticates. The Petnica mandible sample is too small for subdivision but is weighted much more towards adult specimens even than Gomolava Ia, suggesting that hunting was very focused throughout the sites occupation. Red deer cranial elements are underrepresented at Gomolava and Petnica, presumably due to field butchery and selective transport of a kind that would not be expected under a garden hunting model. The post-mortem treatment of red deer is broadly comparable to that of cattle, and their similar body size seems to have been important. No clear corroboration was found for Russells (1993; 1999) argument that red deer were particularly involved in feasting at Opovo: they appear to have been consumed as intensively as cattle at Gomolava and Petnica, probably primarily through the distribution of joints, perhaps structured by kin relations and/or contribution to hunting. Red deer are weakly associated with pits at Gomolava, however, and are more frequently burnt than cattle at both sites, regardless of context. They thus fit into a tentative overall pattern of differential processing between wild and domestic species, the precise significance of which is unclear.
8.2.6 roe deer

Roe deer are a minor part of the wild fauna at most sites. There is no sign that they replace the larger-bodied red deer and wild pig in times of resource stress, as predicted by diet-breadth models (e.g. Winterhalder 1981); rather, their frequency is broadly correlated with other wild taxa. Roe deer hunting patterns are hard to determine due to small mandible samples and early dental maturity, but the animals are again likely to have been encountered individually or in small groups (Legge & Rowley-Conwy 1988, 16-18). They are mostly male at Gomolava, Petnica and Opovo. Acquisition of raw materials, especially bone, may have been one motivation for hunting, but since roe deer antler is rarely worked at the sites this cannot account for the number of males (Russell 1993, 381, personal observation). Garden hunting might alternatively explain the sex ratio. Roe deer largely drop out of the fauna at Gomolava after phase Ia, being found almost exclusively in pits. Meat-bearing parts are underrepresented in both context types, so we may be seeing preferential deposition of crania and metapodials often bone-working waste in
303
Synthesis
the pits, with other elements removed by a combination of intensive fragmentation and poor recovery. Roe deer representation is steady between Petnica 1 and 2 but decreases
appreciably in Petnica 3. Anatomical representation is fairly even at Petnica, allowing for preservation and recovery biases. A marked decline is seen at Selevac, but the opposite at Snandrei. No overall regional pattern is visible. Roe deer are small enough that they could feasibly have been consumed by individual households. This may underlie their concentration in the pits at Gomolava, depending on how these are interpreted. Roe deer elements do seem to cluster in particular pits, while Russells (2000) left-right correlations indicate relatively low levels of sharing at Opovo. In terms of processing, roe deer fit in with both the size and possible wild:domestic trends noted in 7.7.
8.2.7 caprines
Caprines are the only unequivocally domestic species in the Balkan Neolithic in the sense that they have no wild counterparts. Sheep and goats have thus far been treated as a single category due to difficulty in discrimination, but both species are definitely present on Vina sites and might potentially have been managed in very different ways. Reported relative frequencies depend heavily on the identification sources used by each analyst, but sheep certainly seem to be more common. Regarding the husbandry of either species, many of the same questions can be asked as for cattle, including: where were caprines herded? Were they milked? At what social scale were flocks aggregated for management purposes? Each has a bearing on the presence of the animals in peoples lives and the intimacy of their relationships. Again, however, macro-osteological evidence alone cannot provide adequate answers. Frequency of caprines decreases over time at the regional level in favour of cattle and pigs, especially on the Hungarian Plain at the north of the study area. Whittle (2003, 113-114; 2005) argues that the importance of sheep and goats on Krs sites stemmed partly from their exotic character clearly contrasted to the indigenous fauna and from associations with the past. If so, their later decline may indicate that such associations faded over time. Caprines were probably more suited to the uplands, erdap and the terraces of the Morava, and indeed they remained economically important at sites such as Liubcova and Selevac. At Petnica caprine remains never exceed 10% of NISP, but their frequency actually increases considerably through time. The difficulty of interpreting age profiles was discussed above (5.3.2). Nonetheless, the limited available data for the Vina period certainly do not indicate the pastoral specialism proposed by the SPR model (Arnold & Greenfield 2004; 2006; Greenfield 2005). Sheep and
304
Synthesis
goats may well have been milked, but flocks were not managed intensively for this purpose. Following Halsteads logic (1996, see 8.2.3), large-scale, presumably mobile herding of caprines is plausible for Krs sites such as Ludo-Budak, but within the Vina period only for the Macedonian sites of Anza IV and Rug Bair. Average age-at-death increases from the Early Neolithic into the Vina period, possibly reflecting (a) smaller and/or less stable caprine flocks as cattle became the dominant domesticate, or (b) reduced co-ordination above the household level. In either case, relatively small flocks could quite feasibly have been pastured around settlements and perhaps sometimes penned on-site. If cattle were herded further afield on a larger scale, then caprines, along with pigs, may actually have been more visible in most peoples lives on a day-to-day basis, representing a more constant, domestic sociality (Whittle 2003, 94). It may also be significant that caprines are small enough to have been consumed entirely or primarily at a household level.
8.2.8 dogs
A vast literature exists on human-dog relationships across time (e.g. Crockford 2000; Serpell 1995a; Snyder & Moore 2006) and is not entered into here in detail. The potential for dogs to act as companions to humans is well established (e.g. Hart 1995), as is their common ambiguous position between human and non-human worlds, neither person nor beast (Serpell 1995b, 254). This ambiguity may be reflected in rules surrounding their treatment (e.g. Bulmer 1967). Neither point should be taken for granted, however. Dogs may be tolerated scavengers as much as valued companions (Serpell 1995b, 250), and the potential for genuine intimacy between humans and pigs, at least, may impinge upon the supposedly unique interstitial space occupied by dogs. Nor can we be sure that dogs in prehistoric cases had specific owners and did not, in Moreys (2006, 165) terms belong to nearly everybody. Nonetheless, humans and dogs/wolves certainly have a particular potential for interspecific sociality, with highly compatible social systems (Kubinyi et al. 2007). Referring to the Iron Gates Mesolithic, Radovanovi (1999) makes the cogent point that the status of individual dogs relies on their bonds with individual humans. Dogs are relatively rare on Vina sites, but do seem to have had a special status. Cranial elements dominate at Gomolava, Petnica and Opovo to an extent that cannot be accounted for by preservation and recovery alone, especially since mandibles and crania are usually remarkably complete. Russell (1993, 190) describes deliberate deposition of dog heads in pits and other discrete contexts at Opovo, and this fits very well with my own data. Of course, deer and possibly caprine heads also appear to have been deposited preferentially in pits, but the cranial:postcranial ratio is much higher for dogs. In addition, there is an articulated dog skeleton in Feature 41 at Opovo (Russell 1993, 84), and a double dog burial in Pit H at
305
Synthesis
Gomolava Ia (Brukner 1980a, 16). Articulated animals burials might be taken as parallels for human inhumations (Morey 2006, 164), perhaps indicating that the individual animals involved were attributed personhood in the sense usually reserved for humans. This argument is problematic for the Vina culture since human burials are extremely rare, with the only known on-site cemetery being from the final phase at Gomolava. The dogs at Opovo and in Pit H at Gomolava were not treated as pseudo-humans, but as something altogether unique. On the other hand, the number of dogs represented at Vina sites is surprisingly small given the frequency of gnawing, and numerous postcranial remains are certainly unaccounted for. The question of the missing dogs is perhaps equivalent to that of the missing people, and the answers may be related. The assumption of a special almost-human status for dogs may thus be valid in the Vina period, at least in death, but their roles in life have yet to be established. It seems likely a priori that dogs took part in hunting given the apparently targeted pursuit of dangerous game (i.e. adult boars), and their frequency is correlated with wild species. This is perhaps surprising given their apparently sporadic inclusion in on-site deposits, but there remains the intriguing possibility that dog remains were buried more often where their role in hunting was greater. Dogs were also skinned at Gomolava, with six mandibles showing the same pattern of light transverse cuts to the alveolar base seen on various wild species. In addition, two almost complete crania featured a mass of cuts to either side of the maxilla and across the nasals, which together with the mandibular marks would form a complete ring around the muzzle. It is possible that the anatomical profile represents the retention of heads and lower limbs onsite following primary skinning elsewhere as for wild carnivores but the intact crania nonetheless indicate unusually careful deposition. Russell (1993, 361) notes slightly more disarticulation/filleting than skinning marks on dog bones at Opovo, but the former are absent at Gomolava. Given the paucity of postcranial elements on which such cuts might be recorded, consumption of dogs cannot be ruled out. Interestingly, no butchery marks at all were recorded on dog specimens from Petnica, where wild carnivores were frequently present and skinned. Dog-skin might only have been used where the availability of other pelts was limited.
306
Synthesis
8.3 Explaining changes

8.3.1 animals and the (political) economy
Chapter 5 closed with the observation that changes in taxonomic representation across the Vina period as a whole were driven to a great extent by marked trends at four large multiphase sites. Specifically, remains of wild taxa decline in numbers at these sites while cattle increasingly dominate amongst the domesticates. Meanwhile, a second group of sites maintains roughly steady taxonomic profiles, and one Liubcova shows the opposite trend. I suggested three possible explanations for the common shift away from wild taxa and towards cattle: 1. Resource depletion around larger sites over time. 2. Settlement histories, with founder groups trading-up to cattle from smaller stock and hunting less as herds became more stable. 3. Regional-scale social trends, with domestic animals increasingly important in the political economy as tensions between household and community increased. All may have played a part, but the latter two explanations were favoured (5.3.3). Having studied one site from each main group in some detail, we are now in a position to return to this issue. I shall start discussion from what is perhaps a surprising angle, namely the phenomenon of wild pigs in the pits at Gomolava. As argued above (7.7), there is good reason to believe that wild pigs were often hunted specifically for particular consumption events, not least because large domestic boars occasionally appear to have been used in their stead. This draws attention to the role of specific social contexts in motivating the killing of animals, a point with particular relevance for domestic species. The ethnographic literature on non-market societies indicates that domestic animals are typically killed for particular social events rather than purely for subsistence (Keswani 1994; e.g. Parker Pearson 2000, 223; Sillitoe 2007, 343346), and this was very probably the case in the Vina period. The contexts envisaged under this heading vary extremely widely in scale, purpose, regularity, predictability and standardization. Marriages, funerals, and other rites of passage; seasonal events and sacrifices in response to weather condition or disease; house closures; the sealing of exchanges or alliances; visits from neighbouring settlements; the list could go on, but all may have been causes for the consumption of domestic meat, be it exclusive or inclusive, with or without specific prescriptions and proscriptions. Of course, the social calendar might have been manipulated for reasons of subsistence just as the subsistence base was structured by social 307
Synthesis
imperatives, but slaughter purely for nutritional purposes was probably the exception rather than the rule. Other than the relative predictability of the availability of domestic meat, there is no reason why these arguments should not apply also to hunted species. Such events specific or generic were presumably one motivation for raising and maintaining animals in the first place, to which can be added their often crucial role in exchange while alive (see review in Russell 1998), their importance for identity (e.g. Parker Pearson 2000), and of course the pleasure and satisfaction potentially derived from their ownership and care (e.g. Campbell 2005). Indeed, the significance of slaughter and subsequent consumption is surely related to all these aspects of their role in society. If the keeping and killing of livestock are motivated by social factors, it follows that changes in the numbers and balance of animals kept, and in forms of management, will to a great extent be driven by social changes. Accordingly, explanations for the changes seen at certain Vina sites should be couched in terms of changing demand for both livestock and deadstock, the latter being the more directly visible from zooarchaeological evidence. Depositional practices at Gomolava change dramatically between periods, and this can reasonably be assumed to reflect, amongst other things, changes in modes of consumption. Given the nature of taphonomic data and the disparate range of different practices probably involved, it is not possible to flesh out possible forms taken by particular events; nor am I suggesting anything so simple as a shift from one mode of consumption/deposition to another. However, some suggestions can be made regarding changes in the overall character of the practices involved. Deposition at Gomolava was increasingly on the surface probably around dwellings, if one accepts the analogy with Divostin (McPherron & Gunn 1988) and Chapmans (2000b, 82) concentration principle where the deposited materials would have been visible for longer, perhaps in direct association with particular houses. Depending on how one interprets the pits, their decline represents a reduction in truly communal deposition, maybe related to events with a primarily celebratory character (sensu Hayden 1996), and/or in consumption at the household level. In either case, the emphasis appears to shift towards the distribution of joints of large animals and their dispersed deposition, although slaughter, butchery and some consumption may often have taken place on a more communal basis. Both wild and domestic species are implicated here cattle, aurochsen and red deer although their distribution was probably structured in different ways. Distribution of wild meat is likely to have been akin to Ingolds (1980, 173-174; 1986, 233) sharing out concept, while that of domestic carcasses was probably closer to sharing in. This distinction is unlikely to have been absolute, however, with kinship ties probably playing a major role in both cases.
308
Synthesis
This could potentially be cast in terms of Chapmans fragmentation theory, with carcasses constituting the archetypal fractal artefacts (Chapman 2000a, 40) and the distribution of parts confirming and solidifying social relationships. Given the nature of faunal materials and a certain reservation regarding Chapmans empirical argument for re-use after breakage I would suggest that (initially) visible surface deposition around dwellings served this purpose in the same way that curation of pot-sherds may have done. That said, the retention of certain elements as for example pig mandibles amongst the Trobriand Islanders (Malinowski 1922, 171) cannot be excluded, although no likely candidates were identified in the faunal study. In the case of domestic animals, the relationships referenced through distribution would be those established by the sequence of transactions making up the biography of the animal(s) in question, maybe even reflecting parentage (3.2.3). At a time when tensions between household and community were growing, and networks of enchained relations were increasingly important in negotiating these tensions, the enormous potential of domestic animals to establish, maintain and embody relationships between humans was surely recognized and exploited. As noted above, the long life-span and high individual value of cattle makes them particularly powerful in this respect. The importance of domestic animals, particularly cattle, for establishing and maintaining social ties is argued here to underlie their increasing representation at sites such as Gomolava and Selevac. I would suggest that this led to the first large-scale cattle herding in the region. I do not envisage this constituting pastoralism in its more specialized sense, not least because the actual contribution of animal products to the Vina economy is currently unquantified. Rather, as herds increased in size they probably came to be grazed away from sites for appreciable periods of time, perhaps on a seasonal basis. If this involved the aggregation of livestock belonging to various individuals or households then proxy management would represent another form of transaction. There are obvious implications of this form of management for the division of labour, and for the importance of cattle to identity on a gendered, age-related or possibly household or lineage basis. Changes in the manner in which animals were kept must thus have fed back into social tensions and cannot be seen as epiphenomenal. A tension emerges here between the accumulation of livestock and ideas of inalienability and enchainment, and I should stress that I do not see the intensification of cattle keeping and exchange in terms of commodification. Rather, I see this tension as being an expression of that inherent in the status of animals as sentient property. Widespread exchange of domestic animals serves to create overlap between human and animal kinship networks (Ray & Thomas 2003), the latter providing a powerful metaphor for human sociality. Herds would
309
Synthesis
have embodied relationships spanning the community and beyond, and the distribution and consumption of domestic meat thus has a particular capacity to be simultaneously competitive and cohesive. The proposed incentive to increase herds relates to the ability they confer to forge and manipulate social relationships, rather than the accumulation of wealth in any abstract sense. Whether or not one accepts these specific interpretations, the changing fauna at Gomolava clearly goes hand-in-hand with developing social practices. The same may be true at other sites where the same faunal trends apply, although I would not wish to imply identical processes in each case. The question thus returns to why these changes should only have occurred on a subset of Vina sites: is this a matter of settlement size, of duration of occupation or trajectories of growth, perhaps even of environment? It is worth briefly reviewing the two relatively faunally static sites with detailed data available. Occupation at Opovo is relatively short-lived, with no apparent change in depositional practices within the excavated area. Domestic animals are never very important, and the contributions of particular species change little between building horizons. Interestingly, some specific consumption practices differ from those at Gomolava. The pigs mostly wild are no more common in pits than in surface middens (Russell 1993, 187), while red deer are instead suggested to have had a particular role in feasting (Russell 1999). The small size and short occupation could be invoked to argue that social tensions are likely to have been weaker here than at Gomolava or Selevac indeed Tringham (1992) explains the foundation of the site in terms of fissioning from a larger settlement where they had become unsupportable. The case of Petnica is perhaps harder to reconcile. The site was occupied for a long period continuously, according to the excavators but again there is no apparent change in depositional practices. This may be an artefact of limited excavation, and reports from earlier campaigns suggest that pit digging was more common in Petnica 1 (Greenfield & Je n.d.; Je 1985). In this case, the fauna from the earlier phases may in fact be misleading. In any case, much deposition clearly took place on the surface from at least Petnica 2. There are suggestions that this differed somewhat from that at Gomolava appreciable numbers of dog remains were recovered from the cultural layer, for example but the main taphonomic variables largely follow the same taxonomic patterns at both sites. Taxonomic representations, however, change very little between phases at Petnica, except for a considerable relative increase in caprines. The current study clearly cannot provide all the answers, and one should stress a likely degree of site-specificity in social changes during the Vina period. Tentatively, however, I suspect
310
Synthesis
that a change in cattle management may be responsible, driven by imperatives to increase herds for reasons of social status but also feeding back into social organization in ways which may have been self-reinforcing. While the tension between household and community probably existed across the region, albeit to varying extents, new forms of herding were only adopted in certain cases. Environment as well as settlement size and history may be important here. The apparently abundant wild fauna around Opovo and Petnica may have allowed red deer, at least, to maintain a parallel status to cattle, with the latter never developing the social role that underlay their increasing importance elsewhere. At larger sites, however, a combination of greater social tensions, more complex networks of livestock-mediated relations, and limited availability of nearby grazing land stimulated the development of larger-scale cattle herding, enabling a form of accumulation which for the time being was reconcilable with the maintenance of social practices emphasizing horizontal, corporate kin relations (Chapman 2000a, 47) due to the peculiar status of domestic animals as sentient property. An important corollary of the argument that social impetuses drove an increase in the scale of cattle herding, and possibly also the number of pigs kept, is that the decreasing representation of wild fauna might not actually imply a dramatic reduction in hunting: more meat may have been consumed in total. It is notable that numbers of wild remains do not decrease appreciably on sites where the balance of domestic species is fairly constant. On the other hand, there is clear evidence for less targeted hunting in later phases of several sites (5.3.3). In either case, wild animals clearly retained a widespread importance to later Vina communities, some 1500 years after the first domesticates appeared in the region.
8.3.2 sedentism and settlement histories

It is worth returning to the question of mobility and sedentism here. At the start of the thesis (2.3.4) I accepted the current consensus that central Balkan Early Neolithic communities were fairly mobile, but argued that this probably involved relatively frequent movement of residential sites without particularly extensive herding around those sites. Some logistical mobility around short-lived sites is also envisaged, and pigs and cattle probably ranged at some distance from sites, but their numbers were relatively few and long-distance seasonal movements away from the main human population are unlikely. In the Vina period, there is considerably more evidence for permanent foci of occupation, with increasing commitment to place expressed through investment in domestic architecture. This undoubtedly had an important symbolic aspect, but I see no reason not to take it also as evidence for an actual decrease in residential mobility. At the same time, I have suggested that the scale and extensiveness of cattle herding increased, probably entailing greater month-to-month
311
Synthesis
movement on the part of those involved in herding: there was a shift away from residential mobility and towards a pattern of radiating mobility in which a segment of the community probably defined along age/gender lines was actually increasingly mobile on a seasonal or otherwise short-term basis. I see this as being limited to a small portion of the community at any one time, with established Vina sites otherwise being essentially sedentary. The idea of large settlements as foci for periodic, probably seasonal, aggregations of dispersed populations argued by Bailey (1999b) for eastern Balkan tells and hinted at by Whittle (1996, 105) for Vina sites such as Selevac may however be useful in understanding their development. Pit-sites such as Gomolava Ia, Banjica V (the earliest horizon) and the first Vina occupation at Belo Brdo might represent brief aggregations of several mobile groups, with archaeologically-invisible temporary dwellings and activities involving deposition in pits. Investment in architecture at these sites increased over time along with the duration and possibly frequency of aggregations, some of the groups (households?) involved eventually coming to occupy the site more-or-less continuously. The development of Selevac from its earliest phases can probably be seen in the same way. Interestingly, this process occurs on a site-by-site basis, with the pit-site of Gomolava Ia, for example, founded long after substantial houses are known elsewhere. Sedentism appears to have rested on the establishment of commitment to specific places and particular agglomerations of smaller communities, rather than representing a region-wide settling-down of the population.
8.4 Wild:domestic?
The focus here has been on domestic animals due to the greater potential for involvement in the political economy that domestication by definition confers. Amongst the main points made in Chapter 3, however, was that the wild:domestic dichotomy, while a valid analytical construct, is not of universal conceptual importance. Likewise, Chapter 5 highlighted the continuing importance of hunting within later Neolithic economy and experience, in the central Balkans at least. One of the tentative hypotheses in 3.5 was that the distinction between owned and un-owned animals would become more resonant with Neolithic experience and perhaps more clearly expressed in post-mortem treatment or in symbolism given (a) an increase in the importance of animals as property, and (b) the possible adoption of forms of secondary products exploitation, entailing more direct human-animal contact. I have argued that the former condition probably did apply, while the degree of secondary products usage remains uncertain until more direct forms of evidence can be brought to bear.
312
Synthesis
The possibility of a Vina period wild:domestic distinction approximating our own is revisited briefly below. Quantitative analysis of taxonomic frequencies on central Balkan Neolithic sites actually reveals a remarkably clear-cut division between wild and domestic groups (5.2.6), indicating that these do, to some extent, form packages that go together. One must be careful not to over-interpret this, however: while reinforcing the analytical validity of these groupings, the PCA results can tell us nothing about perceptions or classifications. These are in fact extremely difficult to approach from (zoo)archaeological data, since there is no reason why categorizations and cosmologies should be reflected in archaeologically visible behaviour in any straightforward manner. Marciniaks (2005a, 240-241) tentative identifications of life forms and unique beginners (sensu Berlin et al. 1973) within the Polish Neolithic are rather optimistic. Nonetheless, differential processing, consumption and deposition of taxa may give some idea of the ways in which classifications were ordered, building on the a priori arguments in 3.2.4. From the taphonomic study of Gomolava and Petnica, the primary axis along which postmortem treatment of animals was structured would appear to be size (7.7). This no doubt represents the practicalities of butchery and distribution to a great extent, but since these processes go a long way towards defining the character of consumption events their importance cannot be avoided. Other variables, such as locus of deposition at Gomolava, certainly cannot be explained away in functional terms: whatever the significance of deposition in pits and it is debatable the selection of this context for the smallest taxa must indicate that they were treated comparably at some level. At the same time, there are fairly consistent differences between wild and domestic representatives of each size class in terms of both occurrence and pattern of burning and butchery, and at Gomolava also in terms of deposition. Again, the precise significance of these patterns is unclear, but in at least some cases they represent structured differences along wild:domestic lines in the ways in which animals were consumed and their remains disposed of. With their parallel wild and domestic forms in the European Neolithic, pigs and cattle may be informative here. Pigs in particular have the potential to blur the boundary between wild and domestic through a wide range of possible management systems. This does not seem to have been the case in the Vina period, however, where there are two effectively reproductively isolated populations, both having considerable importance at some sites. Of the explanations offered for this (8.2.1), however, only one involves deliberate separation. Wild individuals do seem to be implicated in certain consumption events, perhaps suggesting that they were attributed a special significance, but since domestic boars appear to have been an acceptable
313
Synthesis
substitute it may once again have been size as much as wild status that mattered. Aurochsen, meanwhile, are treated conspicuously differently than domestic cattle as far as can be told from the small samples. Finally, concentrations of predominantly wild faunal remains, including aurochs horn cores and red deer antlers, in the ditch terminals at Uivar perhaps functioned as apotropaic symbols marking the transition between the outside world and the realm of the settlement (Schier 2008, 61). This is a compelling case in broad support of Hodders (1990, 90) argument for the increasing definition of and emphasis on the agrios towards the end of the Neolithic. The presence of enclosure ditches per se is extremely unusual within the Vina period, however, and at Uivar we may be seeing a very idiosyncratic development. Publication of the faunal data is eagerly awaited. There are two rather different issues at stake here: the existence of a wild:domestic dichotomy as regards animals, and the validity of Hodders (1990) domus and agrios concepts. The latter is largely beyond the scope of this work, although I would note that the Uivar deposits and the figurines if one extends Nanoglous (2008) interpretation to the later Neolithic go some way towards supporting it. Turning to the former, wild and domestic status does seem to have played some part in structuring the treatment of animals at Gomolava and Petnica but appears subordinate to size, while the similarity of wild and domestic pigs seems to be recognized in consumption practices despite their segregation in life. I would argue that the commonality created between cattle, caprines and domestic pigs by their role in, and proximity to, human society was recognized again implicit in Nanoglous argument from figurines (see 8.2.3) but that this was not necessarily the most important dimension in their classification. It was argued in 3.2.4 that some form of protective animism, in Descolas (1996) terminology, is likely to have characterized human relations with domestic animals in the Neolithic, entailing a difference between wild and domestic animals in terms of mode of relation with humans but not in terms of mode of identification. Both groups of species may have been perceived to share an essential kinship with humans even if some were distinguished by their incorporation into human society and others by their place outside of it. It is quite possible that attitudes to wild animals in the Vina period were not radically different from those characteristically attributed to hunter-gatherers, entailing ideas of trust and generalized reciprocity. The suggestion of a sense of guilt on the part of humans for altering this relationship by incorporating some species into human society (Whittle 2003, 93) may thus have some merit. At the same time, ideas of affection, care, order (see Theodossopoulos 2005) and/or partnership within a system of balanced reciprocity may have been important in perceptions of domestic animals, and I would stress after Morris (1998, 4)
314
Synthesis
the complex, ambivalent and frequently contradictory nature of attitudes to animals amongst groups who keep domesticates. Although they can never prove or disprove them, the zooarchaeological results are consistent with the above suggestions: selection and treatment of animals certainly references the wild:domestic divide, but seems to do so in a fluid manner.
(Bori & Jovanovi in press; Breunig 1987; Burleigh et al. 1982; Burleigh & Nelson 1979; Chapman 1981; Draovean 1994; Draovean 2005; Gimbutas 1974; Gimbutas 1976; Glser 1996; Horvth & Hertelendi 1994; Kohl & Quitta 1970; Lszl 1997; Linick 1977; Luca 2003; Luca et al. 2006; McPherron et al. 1988; Quitta & Kohl 1969; Schier 1996; Schier 2008; Srdo et al. 1987; Srdo et al. 1977; Srdo et al. 1975; Tasi & Petrovi 1988; Todorovi & ermanovi 1961; Tringham & Krsti 1990a; Vogel & Waterbolk 1963; Waterbolk 1988; Whittle et al. 2002)
(Greenfield 2006; Noddle 1982)
315
Conclusions
Chapter 9 Conclusions
The conclusions of this thesis can be grouped into several sets. The first relates to the first aim stated in Chapter 2: to document changes in the frequencies of mammalian taxa represented on sites over the course of the Vina period. It was shown in Chapter 5 that wild species have an importance throughout the central Balkan Neolithic that cannot be reduced to a seasonal or riskbuffering role. Their frequency rises somewhat into the earlier part of the Vina period before declining slightly in the latter. At the same time, numbers of cattle and to a lesser extent pigs increase amongst the domesticates. Overall changes are subtle, however, and are driven to a great extent by clear trends at certain large multiphase sites: Selavac, Gomolava and Snandrei all show a marked decrease in wild taxa coupled with a relative increase in cattle amongst the domesticates, while fauna from several smaller sites including Petnica and Opovo change little over time. It is notable that the latter have particularly high proportions of hunted species throughout their occupations. Not all sites fit into these two groups. The second set of conclusions derives from the taphonomic study in Chapter 7. No evidence was found at Gomolava or Petnica to support the prediction that domestic animals would be involved in competitive, extravagant consumption. Even large species seem to have been consumed intensively, indicating that carcasses were shared without wastage, although it is likely that large species at least were mostly slaughtered in specific social contexts and consumed above the household level. This seems to be the case for wild pig in the earlier phases at Gomolava: male individuals were, I have argued, selectively hunted specifically for the purpose of events that resulted in their deposition in pits, with unusually frequent burning. Pig hunting declines along with pit-filling practices. Importantly, however, large male domestic pigs occasionally seem to have been treated in the same way. In general, the principal axis structuring post-mortem treatment of animals at the two sites appears to have been size, presumably testifying to the importance of sharing, but there are also consistent differences between wild and domestic taxa within each size class. It is not possible to detail the specific practices involved, but both size and domestication status appear to have influenced the ways in which animals were processed, consumed and/or deposited. The prediction that the latter would become more prominent over time could not be supported, although this may be partly due to the difficulty of splitting samples by phase at Gomolava These two datasets were brought together in Chapter 8 in order to address the second primary aim: relating changes in faunal assemblages to the models of social change proposed for the period. It is suggested that the rise in cattle remains reflects an increasing social importance, driven by their involvement in networks of enchained relations. Constituting both inalienable property and social beings tied into human kinship relations, domestic animals especially cattle had a particular potential for competition and status negotiation between households that did not overtly conflict with the egalitarian ethos restricting accumulation. Rising numbers of cattle may have stimulated
316
Conclusions
more extensive herding such that increasing commitment to place and reduced mobility for communities as a whole was accompanied by greater movement around the landscape on the part of a segment, with implications for the role of animals in identity on a gendered, age or (less likely) household/lineage basis. Hunting, meanwhile, appears to have become less targeted, probably due to increased social incentives to use domestic animals in exchange and communal consumption. Crucially, faunal (and architectural) changes do not apply to all sites and are displaced in time where they do occur, leading me to argue that the Tringham model of social change should be seen in terms of the establishment and varying development of individual communities rather than of a region-wide settling-down and transformation of social organization. This account is as much hypothesis as conclusion, and would benefit greatly from detailed faunal investigations at other sites with an observed increase in cattle Selevac, Snandrei and Para and at those with similar settlement histories to Gomolava, such as Banjica. Equally importantly, phytolith analysis and/or soil micromorphology from ancillary buildings on Vina sites might confirm or refute the suggestion that they housed animals, while dental microwear has the potential to elucidate geographical herding patterns, and lipid residue analysis can now reliably detect the use of dairy products (Mukherjee et al. 2005). To my knowledge, none of these analyses has yet been applied to Vina material. The potential of zooarchaeological data to address cosmology and classification is limited. Nonetheless, the results presented here are consistent with the a priori suggestion (Chapter 3) that Neolithic perceptions of animals may have constituted protective animism in Descolas terms: both wild and domestic species sharing an essence with humans but the latter distinguished by (a) hierarchical subordination and (b) the possibility of direct intersubjective relations with people. This is not intended as a blanket prescription for all Neolithic cases but as one likely possibility, and in any case leaves considerable scope for variation. In the Vina case, a distinction between wild and domestic certainly seems to have been recognised in post-mortem treatment and possibly in the breeding isolation of pigs but size was at least as important. Categorizations were in any case probably fluid, as suggested by the apparent substitution of domestic for wild individuals in the proposed pig-events. However wild and domestic animals were perceived in the Vina case, it is clear that both played a ubiquitous role both in the economy and in peoples lives. I would like to close with a comment regarding methodology. I hope that this thesis has demonstrated the benefits of approaching zooarchaeological interpretation from both a top-down and a bottom-up perspective simultaneously. From a strictly materialist tradition, we are in danger of moving too far in the opposite, idealist direction, the overwhelming focus on production and coarse-scale change being replaced by exclusive concentration on consumption, symbolism and contextual detail (see also Miracle 2006, 63). I believe that it is necessary to integrate both approaches and both types of datasets if we are to achieve a productive, balanced zooarchaeology that can provide a basis for realist interpretations of the place of animals in prehistoric communities. 317
References
References
Abe, Y., C.W. Marean, P.J. Nilssen, Z. Assefa & E.C. Stone. 2002. The analysis of cutmarks on archaeofauna: a review and critique of quantification procedures, and a new image-analysis GIS approach. American Antiquity, 67:643-663 Ahlen, I. 1965. Studies on the red deer, Cervus elaphus L. in Scandinavia. III. Ecological investigations. Viltrevy, 3:177-376 Aitchison, J. 1986. The Statistical Analysis of Compositional Data. London: Chapman and Hall. Aitken, R.J. 1975. Cementum layers and tooth wear as criteria for ageing Roe deer (Capreolus capreolus). Journal of Zoology, 175:15-28 Albarella, U., K. Dobney, A. Ervynck & P.A. Rowley-Conwy. 2007a. Introduction. In U. Albarella, K. Dobney, A. Ervynck & P.A. Rowley-Conwy (eds) Pigs and Humans: 10,000 years of interaction. pp. 1-12. Oxford: Oxford University Press. Albarella, U., K. Dobney & P. Rowley-Conwy. 2006a. The domestication of the pig (Sus scrofa): new challenges and approaches. In M.A. Zeder, D.G. Bradley, E. Emshwiller & B.D. Smith (eds) Documenting Domestication: new genetic and archaeological paradigms. pp. 209-227. Los Angeles: University of California Press. Albarella, U., F. Manconi, J.-D. Vigne & P.A. Rowley-Conwy. 2007b. Ethnoarchaeology of pig husbandry in Sardinia and Corsica. In U. Albarella, K. Dobney, A. Ervynck & P.A. Rowley-Conwy (eds) Pigs and Humans: 10,000 years of interaction. pp. 287307. Oxford: Oxford University Press. Albarella, U. & D. Serjeantson. 2002. A passion for pork: meat consumption at the British Late Neolithic site of Durrington Walls. In P. Miracle & N. Milner (eds) Consuming Passions and Patterns of Consumption. pp. 33-49. Cambridge: McDonald Institute. Albarella, U., A. Tagliacozzo, K. Dobney & P. Rowley-Conwy. 2006b. Pig hunting and husbandry in prehistoric Italy: a contribution to the domestication debate. Proceedings of the Prehistoric Society, 72:193-227 Armitage, P.L. & J. Clutton-Brock. 1976. A system for classification and description of the horn cores of cattle from archaeological sites. Journal of Archaeological Science, 3:329-348 Arnold, E.R. & H.J. Greenfield. 2004. A zooarchaeological perspective on the origins of vertical transhumant pastoralism and the colonization of marginal habitats in temperate southeastern Europe. In M. Mondini, S. Munoz & S. Wickler (eds) Colonisation, Migration and Marginal Areas. pp. 96-117. Oxford: Oxbow. Arnold, E.R. & H.J. Greenfield. 2006. The Origins of Transhumant Pastoralism in Temperate Southeastern Europe (British Archaeological Reports International Series 1538. Oxford: Archaeopress. Ascher, R. 1968. Times arrow and the archaeology of a contemporary community. In K.C. Chang (ed.) Settlement Archaeology. pp. 43-52. Palo Alto: National Press Books. Babovi, L. 1986. Zbradila Korbovo: compte-rendu des fouilles en 1981erdapske Sveske III. pp. 95-115. Belgrade: Archaeological Insititute. Bailey, D. 1999a. The built environment: pit-huts and houses in the Neolithic. Documenta Praehistorica, 26:153-162
318
References
Bailey, D.W. 1999b. What is a tell? Settlement in fifth millennium Bulgaria. In J. Brck & M. Goodman (eds) Making Places in the Prehistoric World: themes in settlement archaeology. pp. 94-111. London: UCL Press. Bailey, D.W. 2000. Balkan Prehistory: exclusion, incorporation and identity. London: Routledge. Bailey, D.W. 2005. Prehistoric Figurines: representation and corporeality in the Neolithic. London: Routledge. Bankoff, H.A. & H.J. Greenfield. 1984. Decision making and culture change in Yugoslav Bronze Age. Balcanica, 15:7-32 Bar-Oz, G. & N.D. Munro. 2004. Beyond cautionary tales: a multivariate taphonomic approach for resolving equifinality in zooarchaeological studies. Journal of Taphonomy, 2:201-221 Barker, G. 1975. Early Neolithic land use in Yugoslavia. Proceedings of the Prehistoric Society, 41:81-104 Barker, G. 2006. The Agricultural Revolution in Prehistory: why did foragers become farmers? Oxford: Oxford University Press. Barone, R. 1976. Anatomie Compare des Mammifres Domestiques. Paris: Vigot. Barrett, J.H. 2007. Environmentalism, materiality and paradigm shifts in archaeology: a zooarchaeological view. Paper presented to the Theoretical Archaeology Group, York, December 2007. Barth, F. 1975. Ritual and Knowledge among the Baktaman of New Guinea. New Haven: Yale University Press. Bartosiewicz, L. 1987. Cattle metapodials revisited: a brief review. Archaeozoologia, 1:47-51 Bartosiewicz, L. 2005. Plain talk: animals, environment and culture in the Neolithic of the Carpathian Basin and adjacent areas. In D. Bailey, A. Whittle & V. Cummings (eds) (Un)settling the Neolithic. pp. 51-63. Oxford: Oxbow. Bartosiewicz, L., V. Boronean, C. Bonsall & S. Stallibrass. 2001. New data on the prehistoric fauna of the Iron Gates: a case study from Schela Cladovei, Romania. In R. Kertsz & J. Makkay (eds) From the Mesolithic to the Neolithic. pp. 15-22. Budapest: Archaeolingua. Baxter, M.J. 1993. Comment on D. Tangri and R.V.S. Wright, Multivariate analysis of compositional data . Archaeometry, 35:112115 Baxter, M.J. 1994. Exploratory Multivariate Analysis in Archaeology. Edinburgh: Edinburgh University Press. Baxter, M.J. & I.C. Freestone. 2006. Log-ratio compositional data analysis in archaeometry. Archaeometry, 48:511-531 Bayliss, A., C. Bronk Ramsey, J. van der Plicht & A. Whittle. 2007. Bradshaw and Bayes: towards a timetable for the Neolithic. Cambridge Archaeological Journal, 17:1-28 Bayliss, A. & C. Orton. 1994. Strategic considerations in dating, or how many dates do I need? Bulletin of the Institute of Archaeology, 31:151-167 Behrensmeyer, A.K. 1975. The taphonomy and paleoecology of Plio-Pleistocene vertebrate assemblages of Lake Rudolf, Kenya. Bulletin of the Museum of Comparative Zoology, Harvard, 146:473-578.
319
References
Beigel, R. & J. Kuhn. 2005. Das tgliche Leben - Umwelt und Wirtschaft der Tellsiedlung von Uivar. In W. Schier (ed.) Masken Menschen Rituale: Alltag und Kult vor 7000 Jahren in der prhistorischen Siedlung von Uivar, Rumnien. pp. 41-47. Wrzburg: Universitt Wrzburg. Beja-Pereira, A., D. Caramelli, C. Lalueza-Fox, C. Vernesi, N. Ferrand, A. Casoli, F. Goyache, L.J. Royo, S. Conti, M. Lari, A. Martini, L. Ouragh, A. Magid, A. Atash, A. Zsolnai, P. Boscato, C. Triantaphylidis, K. Ploumi, L. Sineo, F. Mallegni, P. Taberlet, G. Erhardt, L. Sampietro, J. Bertranpetit, G. Barbujani, G. Luikart & G. Bertorelle. 2006. The origin of European cattle: Evidence from modern and ancient DNA. Proceedings of the National Academy of Sciences, 103:8113-8118 Bell, A.M. 2007. Animal personalities. Nature, 447:539-540 Bennett, J.L. 1999. Thermal alteration of buried bone. Journal of Archaeological Science, 26:1-8 Berlin, B., D.E. Breedlove & P.H. Raven. 1973. General principles of classification and nomenclature in folk biology. American Anthropologist, 75:214-242 Biddick, K.A. & J. Tommenchuck. 1975. Quantifying lesions and fractures on long bones. Journal of Field Archaeology, 2:239-249 Binford, L.R. 1978. Nunamiut Ethnoarchaeology. New York: Academic Press. Binford, L.R. 1981. Behavioural archaeology and the Pompeii premise. Journal of Anthropological Research, 37:195-208 Binford, L.R. & J.B. Bertram. 1977. Bone frequencies and attritional processes. In L.R. Binford (ed.) For Theory Building in Archaeology. pp. 77-156. New York: Academic Press. Bird-David, N. 1990. The giving environment: another perspective on the economic system of gatherer-hunters. Current Anthropology, 31:189-196 Bird-David, N. 1992. Beyond 'The Original Affluent Society': a culturalist reformulation. Current Anthropology, 33:25-47 Blai, S. 1985. Ostaci faune sa arheolokog nalazita kod Vizia. Rad Vojvoanskih Muzeja, 29:33-36 Blai, S. 1992. Fauna Donje Branjevine: preliminarni rezultati. Arheologija i Prirodne Nauke. pp. 65-67. Belgrade. Blai, S. 2005. The faunal assemblage. In S. Karmanski (ed.) Donja Branjevina: a Neolithic settlement near Deronje in the Vojvodina (Serbia). Trieste: Societa per la Preistoria e Protostoria della Regione Friuli-Venezia Giulia. Boaz, N.T. & A.K. Behrensmeyer. 1976. Hominid taphonomy - transport of human skeletal parts in an artificial fluviatile environment. American Journal of Physical Anthropology, 45:53-60 Boessneck, J. 1969. Osteological Differences between Sheep (Ovis aries Linn.) and Goat (Capra hircus Linn.). In D. Brothwell & E. Higgs (eds) Science in Archaeology. pp. 331-358. London: Thames & Hudson. Bogaard, A. 2004. The nature of early farming in central and south-east Europe. Documenta Praehistorica, 31:49-58 Bogaard, A. 2005. 'Garden agriculture' and the nature of early farming. World Archaeology, 37:177-196
320
References
Bogdanovi, M. 1988. Architectural and structural features at Divostin. In A. McPherron & D. Srejovi (eds) Divostin and the Neolithic of Central Serbia. pp. 35-91. Pittsburgh: University of Pittsburgh. Bogucki, P. 1993. Animal traction and household economies in Neolithic Europe. Antiquity, 67:492-503 Bknyi, S. 1969. Archaeological problems and methods of recognizing animal domestication. In P.J. Ucko & G.W. Dimbleby (eds) The Domestication and Exploitation of Plants and Animals. pp. 219-230. Chicago: Aldine. Bknyi, S. 1970. A new method for the determination of the number of individuals in animal bone material. American Journal of Archaeology, 74:291-292 Bknyi, S. 1972. Zoological evidence for seasonal or permanent occupation of prehistoric settlements. In P.J. Ucko, R.E. Tringham & G.W. Dimbleby (eds) Man, Settlement and Urbanism. pp. 121-126. London: Duckworth. Bknyi, S. 1974a. History of Domestic Mammals in Central and Eastern Europe. Budapest: Akadmiai. Bknyi, S. 1974b. The vertebrate fauna. In M. Gimbutas (ed.) Obre I and II: Neolithic sites in Bosnia. Wissenschaftliche Mitteilungen des Bosnisch-herzegowinischen Landesmuseums, Band 3, Heft B (Archologie). pp. 55-154. Sarajevo: Zemaljski Muzej. Bknyi, S. 1976. The vertebrate fauna from Anza. In M. Gimbutas (ed.) Neolithic Macedonia: as reflected by excavation at Anza, Southeast Yugoslavia. Monumenta archaeologica 1. pp. 313316. Los Angeles: University of California Institute of Archaeology. Bknyi, S. 1978. The vertebrate fauna of VlasacD. Srejovi & Z. Letica, Vlasac: a Mesolithic settlement in the Iron Gates. pp. 35-65. Belgrade: Serbian Academy of Science and Arts. Bknyi, S. 1984. Die frneolitische Wirbeltierfauna von Nosa. Acta Archaeologica Hungarica, 36:29-41 Bknyi, S. 1988. The Neolithic fauna of Divostin. In A. McPherron & D. Srejovi (eds) Divostin and the Neolithic of Central Serbia. pp. 419-445. Pittsburgh: University of Pittsburgh. Bknyi, S. 1989. Definitions of animal domestication. In J. Clutton-Brock (ed.) The Walking Larder. pp. 22-27. London: Unwin Hyman. Bknyi, S. 1992. Animal remains of Mihajlovac-Knjepite; an early Neolithic settlement of the Iron Gate Gorge. Balcanica, 23:77-87 Bollongino, R., C.J. Edwards, K.W. Alt, J. Burger & D.G. Bradley. 2006. Early history of European domestic cattle as revealed by ancient DNA. Biology Letters, 2:155-159 Bonsall, C., G. Cook, R. Lennon, D. Harkness, M. Scott, L. Bartosiewicz & K. McSweeney. 2000. Stable isotopes, radiocarbon and the Mesolithic-Neolithic transition in the Iron Gates. Documenta Praehistorica, 27:119-132 Bori, D. 2005. Deconstructing essentialisms: unsettling frontiers of the Mesolithic-Neolithic Balkans. In D. Bailey, A. Whittle & V. Cummings (eds) (Un)settling the Neolithic. pp. 16-31. Oxford: Oxbow. Bori, D. & V. Dimitrijevi. 2006. Continuity of foraging strategies in Mesolithic-Neolithic transformations: dating faunal patterns at Lepenski Vir (Serbia). Atti della Societ per la Preistoria e Protoistoria della Friuli-Venezia Giulia, 15:33-80 321
References
Bori, D. & B. Jovanovi. in press. Absolute dating of metallurgical innovations in the Vina culture of the Balkans. In T. Kienlin & B. Roberts (eds) Festschrift Barbara Ottaway. Bori, D. & P. Miracle. 2004. Mesolithic and Neolithic (dis)continuities in the Danube Gorges: new AMS dates from Padina and Hajduka Vodenica (Serbia). Oxford Journal of Archaeology, 23:341-371 Borojevi, K. 1988. Differences in plant macro remains from the Neolithic level at Gomolava and the Neolithic site of Opovo. In N. Tasi & J. Petrovi (eds) Gomolava: symposium. pp. 109-115. Novi Sad: Muzej Vojvodine. Borojevi, K. 1998. The Relations Among Farming Practices, Landownership, and Social Stratification in the Balkan Neolithic Period. Unpublished Ph.D. thesis, University of St. Louis. Borojevi, K. 2006. Terra and Silva in the Pannonian Plain: Opovo agro-gathering in the Late Neolithic (British Archaeological Reports International Series 1563. Oxford: Archaeopress. Bottema, S. & B.S. Ottaway. 1982. Botanical, malacological and archaeological zonation of settlement deposits of Gomolava. Journal of Archaeological Science, 9:221-246 Boyle, K.V. 2006. Neolithic wild game animals in Western Europe: the question of hunting. In D. Serjeantson & D. Field (eds) Animals in the Neolithic of Britain and Europe. pp. 10-23. Oxford: Oxbow. Bradley, R. 1984. The Social Foundations of Britain. London: Longman. Bradley, R. 2005. Ritual and Domestic Life in Prehistoric Europe. London: Routledge. Brain, C.K. 1969. The contribution of Namib Desert Hottentots to an understanding of australopithecine bone accumulations. Scientific Papers of the Namib Desert Research Station, 39:13-22 Breunig, P. 1987. 14C-Chronologie des vorderasiatischen, Sdost- und mitteleuropischen Neolithikums (Fundamenta A 13. Briedermann, L. 1990. Schwarzwild. Berlin: VEB Deutsher Landwirtschartverlag. Brody, H. 2000. The Other Side of Eden: hunters, farmers, and the shaping of the world. Vancouver: Douglas & McIntyre. Bronk Ramsey, C. 1995. Radiocarbon calibration and analysis of stratigraphy: the OxCal program. Radiocarbon, 37:425-430 Bronk Ramsey, C. 2001. Development of the radiocarbon program OxCal. Radiocarbon, 43:355-363 Bronk Ramsey, C. 2007. OxCal 4.0.5. Brown, W.A.B. & N.G. Chapman. 1991a. Age assessment of red deer (Cervus elaphus) from a scoring scheme based on radiographs of developing permanent molariform teeth. Journal of Zoology, 225:85-97 Brown, W.A.B. & N.G. Chapman. 1991b. The dentition of red deer (Cervus elaphus) - a scoring scheme to assess age from wear of the permanent molariform teeth. Journal of Zoology, 224:519-536 Brck, J. 1999. What's in a settlement? Domestic practice and residential mobility in early Bronze Age southern England. In J. Brck & M. Goodman (eds) Making Places in the Prehistoric World: themes in settlement archaeology. pp. 52-75. London: UCL Press.
322
References
Brukner, B. 1965. Gomolava, Hrtkovci - vieslojno nalazite, II Neolitski i stariji eneolitski horizont na Gomolavi 1965. g. Arheoloki Pregled, 7:20-23 Brukner, B. 1966. Neolitski i ranoeneolitski sloj na Gomolavi (1965-1966. godina). Rad Vojvoanskih Muzeja, 15:237-247 Brukner, B. 1967. Gomolava, Hrtkovci - praistorijsko vieslojno nalazite. Arheoloki Pregled, 9:45-47 Brukner, B. 1968. Gomolava, Hrtkovci - vieslojno nalazite. Arheoloki Pregled, 10:18-20 Brukner, B. 1971. Gomolava Hrtkovci: site prhistorique plusiers couches. In A. Benac, M. Garaanin & N. Tasi (eds) Epoque prhistorique et protohistorique en Yougoslavie - recherches et rsultats. pp. 175-176. Belgrade: Socit archologique de Yougoslavie. Brukner, B. 1972. Gomolava, Hrtkovci - vieslojno nalazite. Arheoloki Pregled, 14:25-31 Brukner, B. 1973. Gomolava, Hrtkovci - vieslojno nalazite. Arheoloki Pregled, 15:4-15 Brukner, B. 1974. Gomolava, Hrtkovci - vieslojno nalazite. Arheoloki Pregled, 16:24-27 Brukner, B. 1980a. Naselje Vinanske grupe na Gomolavi (neolitski i ranoeneolitski sloj). Izvetaj sa iskopavanja 1967-1976. g. Rad Vojvoanskih Muzeja, 26:5-55 Brukner, B. 1980b. Terenski Dnevnik 1980.godinu Lokalitet "Gomolava". In Museum of Vojvodina. Novi Sad. Brukner, B. 1981. Terenski Dnevnik 1981.godinu sa Lok. "Gomolava". In Museum of Vojvodina. Novi Sad. Brukner, B. 1982. Terenski Dnevnik za 1982.godinu sa Lokaliteta "Gomolava". In Museum of Vojvodina. Novi Sad. Brukner, B. 1984. Terenski Dnevnik za 1984.godinu sa Lokaliteta "Gomolava". In Museum of Vojvodina. Novi Sad. Brukner, B. 1988. Die siedlung der Vina-gruppe auf Gomolava (die wohnschicht des sptneolithikums und frhneolithikums (Gomolava Ia, Gomolava Ia-b und Gomolava Ib) und der wohn horizont des neolithischen humus (Gomolava II). In N. Tasi & J. Petrovi (eds) Gomolava: symposium. pp. 19-38. Novi Sad: Muzej Vojvodine. Brukner, B. 1990. Typen und siedlunges Modellen und Wohnobjekte der Vina-Group in der panonnischen Tiefebene. In D. Srejovi & N. Tasi (eds) Vina and its World. pp. 7983. Belgrade: Serbian Academy of Sciences and Arts. Brukner, B. 2003. Vinanska kultura u prostoru i vremenu. Rad Muzeja Vojvodine, 43-45:728 Brukner, B. & J. Petrovi. 1977. Gomolava, Hrtkovci - vieslojno nalazite. Arheoloki Pregled, 19:24-27 Buck, C.E., W.G. Cavanagh & C.D. Litton. 1996. Bayesian Approach to Interpreting Archaeological Data. Chichester: Wiley. Budiansky, S. 1992. The Covenant of the Wild: why animals chose domestication. London: Weidenfeld & Nicolson. Bull, G. & S. Payne. 1982. Tooth eruption and epiphysial fusion in pigs and wild boar. In B. Wilson, C. Grigson & S. Payne (eds) Ageing and Sexing Animal Bones from Archaeological Sites. British Archaeological Reports British Series 109. pp. 55-71. Oxford: BAR.
323
References
Bulmer, R.N.H. 1967. Selectivity in hunting and in disposal of animal bones by the Kalam of the New Guinea Highlands. In G. Sieveking, I. Longworth & K. Wilson (eds) Problems in Economic and Social Archaeology. pp. 169-186. London: Duckworth. Burkitt, M. 1921. Prehistory. London: Cambridge University Press. Burleigh, R., K. Matthews & J. Ambers. 1982. British Museum natural radiocarbon measurements XIV. Radiocarbon, 24:229-261 Burleigh, R. & A. Nelson. 1979. British Museum natural radiocarbon measurements XI. Radiocarbon, 21:339-352 Campbell, B. 2005. On 'loving your water buffalo more than your own mother': relationships of animal and human care in Nepal. In J. Knight (ed.) Animals in Person: cultural perspectives on human-animal intimacy. pp. 79-100. Oxford: Berg. Carlson, K.J. & T.R. Pickering. 2004. Shape-adjusted bone mineral density measurements in baboons: other factors explain primate skeletal element representation at Swartkrans. Journal of Archaeological Science, 31:577-583 Carter, H.H. 1975. A Guide to Rates of Tooth Wear in English Lowland Sheep. Journal of Archaeological Science, 2:231-233 Carter, R.J. 1997. Age estimation of the Roe Deer (Capreolus capreolus) mandibles from the Mesolithic site of Star Carr, Yorkshire, based on radiographs of mandibular tooth development. Journal of Zoology (London), 241:495-502 Carter, R.J. 1998. Reassessment of Seasonality at the Early Mesolithic Site of Star Carr, Yorkshire Based on Radiographs of Mandibular Tooth Development in Red Deer (Cervus elaphus). Journal of Archaeological Science, 25:851-856 Carter, R.J. 2001. New Evidence for Seasonal Human Presence at the Early Mesolithic Site of Thatcham, Berkshire, England. Journal of Archaeological Science, 28:1055-1060 Carter, R.J. 2006. A method to estimate the ages at death of red deer (Cervus elaphus) and roe deer (Capreolus capreolus) from developing mandibular dentition and its application to Mesolithic North West Europe. In D. Ruscillo (ed.) Recent Advances in Ageing and Sexing Animal Bones. pp. 40-61. Oxford: Oxbow. Cartmill, M. 1993. A View to a Death in the Morning: hunting and nature through history. Cambridge Mass.: Harvard University Press. Caulfield, S. 1978. Star Carr an alternative view. Irish Archaeological Research Forum, 5:15-22 Cauvin, J. 1972. Religions Nolithiques de Syro-Palestine. Saint-Andre-de-Cruzires: Centre des Recherches d'Ecologie et de Prhistoire. Cauvin, J. & M.-C. Cauvin. 1984. Origines de l'agriculture au Levant. Facteurs biologiques et socio-culturels. In T.C. Young, P. Smith & P. Mortensen (eds) The Hilly Flanks. pp. 43-55. Chicago: University of Chicago Press. Chaplin, R.E. 1971. The Study of Animal Bones from Archaeological Sites. New York: Seminar Press. Chapman, J. 1981. The Vina Culture of South-East Europe: studies in chronology, economy and society (British Archaeological Reports International Series 117. Oxford: BAR. Chapman, J. 1982. The secondary products revolution and the limitations of the Neolithic. University of London Institute of Archaeology Bulletin, 19:107-122
324
References
Chapman, J. 1988. Comment on H. J. Greenfield 'Origins of milk and wool production...' Current Anthropology, 29:587-588 Chapman, J. 1989. The early Balkan village. In S. Bknyi (ed.) Neolithic of Southeastern Europe and its Near Eastern connections. Varia Archaeologica Hungarica 2. pp. 3353. Budapest: Archaeolingua. Chapman, J. 1990. The Neolithic in the Morava-Danube confluence area: a regional assessment of settlement pattern. In R. Tringham & D. Krsti (eds) Selevac: a Neolithic village in Yugoslavia. pp. 13-43. Los Angeles: University of California Press. Chapman, J. 1992. Arenas of social power: the case of Serbian prehistory. Zbornik Narodnog Muzeja (Beograd), 14:305-317 Chapman, J. 1996. Enchainment, commodification, and gender in the Balkan Copper Age. Journal of European Archaeology, 4:203-242 Chapman, J. 1998. Objectification, embodiement and the value of places and things. In D. Bailey (ed.) The Archaeology of Value: essays on prestige and the processes of valuation. British Archaeological Reports International Series 730. pp. 106-130. Oxford: BAR. Chapman, J. 2000a. Fragmentation in Archaeology: people, places and broken objects in the prehistory of south-eastern Europe. London: Routledge. Chapman, J. 2000b. Pit-digging and structured deposition in the Neolithic and Copper Age of Central and Eastern Europe. Proceedings of the Prehistoric Society, 61:51-67 Chapman, J. 2000c. Rubbish-dumps or places of deposition? Neolithic and Copper Age settlements in Central and Eastern Europe. In A. Ritchie (ed.) Neolithic Orkney in its European context. pp. 347-362. Cambridge: McDonald Institute. Chapman, J. 2008. Meet the ancestors: settlement histories in the Neolithic. In D. Bailey, A. Whittle & D. Hofmann (eds) Living Well Together? Settlement and materiality in the Neolithic of South-East and Central Europe. pp. 54-67. Oxford: Oxbow. Chapman, J. & B. Gaydarska. 2007. Parts and Wholes: fragmentation in prehistoric context. Oxford: Oxbow. Childe, V.G. 1925. The Dawn of European Civilization. London: Kegan Paul. Childe, V.G. 1936. Man Makes Himself. London: Watts & Co. Choyke, A.M. 2001. Late Neolithic red deer canine beads and their imitations. In A.M. Choyke & L. Bartosiewicz (eds) Crafting bone skeletal technologies through time and space. British Archaeological Reports International Series 973. pp. 251-266. Oxford: Archaeopress. Clark, J. & K.K. Kietzke. 1967. Palaeoecology of the Lower Modular Zone, Brule Formation, in the Big Badlands of South Dakota. Fieldiana: Geology Memoirs, 5:111-137 Clark, J.G.D. 1969. World Prehistory: a new outline. Cambridge: Cambridge University Press. Clark, S.R.L. 1988. Is humanity a natural kind? In T. Ingold (ed.) What Is an Animal? pp. 1734. London: Unwin Hyman. Clason, A.T. 1979a. The farmers of Gomolava in the Vina and La Tene period. Palaeohistoria, 21:42-81
325
References
Clason, A.T. 1979b. The farmers of Gomolava in the Vina and La Tene period. Rad Vojvoanskih Muzeja, 25:60-114 Clason, A.T. 1980. Padina and Starevo: game, fish and cattle. Palaeohistoria, 22:142-173 Clason, A.T. 1988. The equids of Gomolava. In N. Tasi & J. Petrovi (eds) Die siedlung der Vina-gruppe auf Gomolava (die wohnschicht des sptneolithikums und frhneolithikums (Gomolava Ia, Gomolava Ia-b und Gomolava Ib) und der wohn horizont des neolithischen humus (Gomolava II). pp. 99-104. Novi Sad: Muzej Vojvodine. Clutton-Brock, J. 1989. Introduction to domestication. In J. Clutton-Brock (ed.) The Walking Larder: patterns of domestication, pastoralism and predation. pp. 7-9. London: Unwin Hyman. Clutton-Brock, T.H. & F.E. Guinness. 1975. Behaviour of red deer (Cervus elaphus L.) at calving time. Behaviour, 55:287-299 Clutton-Brock, T.H., F.E. Guinness & S.D. Albon. 1982. Red deer: behavior and ecology of two sexes. In G.B. Schaller (ed.) Wildlife Behaviour and Ecology. Chicago: University of Chicago Press. Collier, S. & J.P. White. 1976. Get them young? Age and sex inferences on animal domestication in archaeology. American Antiquity, 41:96-102. Collins, M.B. 1975. The sources of bias in archaeological data, an appraisal. In J.W. Mueller (ed.) Sampling in Archaeology. pp. 26-32. Tucson: University of Arizona Press. Cook, R.G. & A.J. Ranere. 1989. Hunting in Pre-Columbian Panama: a diachronic perspective. In J. Clutton-Brock (ed.) The Walking Larder: patterns of domestication, pastoralism and predation. London: Unwin Hyman. Counihan, C. & P.v. Esterik. 1997. Introduction. In C. Counihan & P.v. Esterik (eds) Food and Culture: a reader. pp. 1-8. London: Routledge. Cowgill, G.L. 1994. Unknown sampling bias is not a license to ignore statistical theory. In I. Johnson (ed.) Methods in the Mountains: proceedings of UISPP Commission IV meeting. pp. 7-11. Sydney: Sydney University Press. Craig, O.E. 2002. The development of dairying in Europe: potential evidence from food residues on ceramics. Documenta Praehistorica, 29:97-107 Crockford, S.J. (ed.) 2000. Dogs Through Time: an archaeological perspective (British Archaeological Reports International Series 889. Oxford: Archaeopress. Crockford, S.J. 2006. Rhythms of Life: thyroid hormone and the origin of species. Victoria: Trafford. Dahl, G. & A. Hjort. 1976. Having Herds: pastoral herd growth and household economy. Stockholm: University of Stockholm. Daly, P. 1969. Approaches to faunal analysis in archaeology. American Antiquity, 34:146-153 Darwin, C. 1859. On the Origin of Species. London: Murray. Davidson, D.A. & B. Thomas. 1986. Geomorphological studies. In C. Renfrew, M. Gimbutas & E. Elster (eds) Excavations at Sitagroi: a prehistoric village in northeast Greece. pp. 25-40. Los Angeles: University of California Press. Davis, S.J.M. 1987. The Archaeology of Animal Bones. New Haven: Yale University Press.
326
References
Degerbl, M. & B. Fredskild. 1970. The Urus (Bos primigenius Bojanus) and Neolithic domesticated cattle (Bos taurus domesticus Linne.) in Denmark. Det Kongelige Dansk Videnskabernes Selskab, Biologiski Skrifter, 17:1-177 Descola, P. 1996. Constructing natures: symbolic ecology and social practice. In P. Descola & G. Plsson (eds) Nature and Society: anthropological perspectives. pp. 82-102. London: Routledge. Descola, P. & G. Plsson. 1996. Introduction. In P. Descola & G. Plsson (eds) Nature and Society: anthropological perspectives. pp. 1-21. London: Routledge. Dietler, M. 2001. Theorizing the feast: rituals of consumption, commensal politics, and power in African contexts. In M. Dietler & B. Hayden (eds) Feasts: archaeological and ethnographic perspectives on food, politics and power. pp. 65-114. Washington: Smithsonian Institution Press. Dietler, M. & B. Hayden. 2001. Good to eat, good to drink, good to think. In M. Dietler & B. Hayden (eds) Feasts: archaeological and ethnographic perspectives on food, politics and power. pp. 1-20. Washington: Smithsonian Institution Press. Dimitrijevi, S. 1956. Prehistorijski nalazi sa Gomalavu u Hrtkovcima ( u Arheolokom Muzeju u Zagrebu). Zbornik Matice Srpske, 15:5-49 Dimitrijevi, V. forthcoming. Vina Belo Brdo, campaigns 1998-2003: vertebrate fauna. Dobney, K. & K. Reilly. 1988. A method for recording archaeological animal bones: the use of diagnostic zones. Circaea, 5:79-96 Dowling, J. 1968. Individual ownership and the sharing of game in hunting societies. American Anthropologist, 7:502-507 Draovean, F. 1994. Die Stufe Vina C im Banat. Germania, 72:409-425 Draovean, F. 2005. Zona thessalo-macedonean i Dunrea mijlocie la sfritul mileniului al VI-lea. Apulum, 42:11-26 Ducos, P. 1989. Defining domestication: a clarification. In J. Clutton-Brock (ed.) The Walking Larder: patterns of domestication, pastoralism, and predation. pp. 28-30. London: Unwin Hyman. Dwyer, P.D. 1996a. Boars, barrow and breeders: the reproductive status of domestic pig populations in mainland New Guinea. Journal of Anthropological Research, 52:481500 Dwyer, P.D. 1996b. The invention of nature. In R.F. Ellen & K. Fukui (eds) Redefining Nature: ecology, culture and domestication. pp. 157-186. Oxford: Berg. Dwyer, P.D. & M. Minnegal. 2005. Person, place or pig: animal attachments and human transactions in New Guinea. In J. Knight (ed.) Animals in Person: cultural perspectives on human-animal intimacy. pp. 37-60. Oxford: Berg. Edwards, C.J., R. Bollongino, A. Scheu, A. Chamberlain, A. Tresset, J.-D. Vigne, J.F. Baird, G. Larson, S.Y.W. Ho, T.H. Heupink, B. Shapiro, A.R. Freeman, M.G. Thomas, R.M. Arbogast, B. Arndt, L. Bartosiewicz, N. Benecke, M. Budja, L. Chaix, A.M. Choyke, E. Coqueugniot, H.-J. Dhle, H. Gldner, S. Hartz, D. Helmer, B. Herzig, H. Hongo, M. Mashkour, M. zdogan, E. Pucher, G. Roth, S. Schade-Lindig, U. Schmlcke, R.J. Schulting, E. Stephan, H.-P. Uerpmann, I. Vrs, B. Voytek, D.G. Bradley & J. Burger. 2007. Mitochondrial DNA analysis shows a Near Eastern Neolithic origin for domestic cattle and no indication of domestication of European aurochs. Proceedings of the Royal Society B: Biological Sciences, 274:1377-1385
327
References
Efremov, I.A. 1940. Taphonomy: new branch of paleontology. Pan-American Geologist, 74:81-93 Ehrich, R.W. 1977. Starevo revisited. In V. Markoti (ed.) Ancient Europe and the Mediterranean. pp. 59-67. Warminster: Aris & Phillips. Ehrich, R.W. & H.A. Bankoff. 1992. Geographical and chronological patterns in east central and southeastern Europe. In R.W. Ehrich (ed.) Chronologies in Old World Archaeology. pp. 375-392; 341-363. Chicago: University of Chicago Press. El Susi, G. 1995a. The animal husbandry during the Late Neolithic and Bronze Age in settlements of Danube Valley (southern Banat)Kulturraum Mittlere und Untere Donau: Traditionen und Perspectiven des Zusammenlebens. pp. 65-72. El Susi, G. 1995b. Economia animalier a comunitilor Neo-Eneolitice de la Para (jud. Timi). Banatica, 13:23-51 El Susi, G. 1996a. A general survery on the animal husbandry in the tell of Para by compareing with the Vina communities in the Danube valley. In F. Draovean (ed.) The Vina Culture, its Role and Cultural Connections. pp. 309-322. Timioara: Museum of Banat. El Susi, G. 1996b. A survey on animal husbandry of Turda community from Ortie Dealul Pemilor (X2) (Hunedoara County). Sargetia, 26:169-177 El Susi, G. 1996c. Vntori, Pescari i Cresctori de Animale n Banatul Milenilor VI .CH I D.CH: studiu arheozoologic. Timioara: Mirton. El Susi, G. 1998. Studiu preliminar al resturilor de faun din aezarea neolitic de la Para-tell II (jud. Timi). Analele Banatului, 6:129-151 El Susi, G. 2000. Studiul resturilor de faun din aezarea neolitic trzie de la Snandrei (jud. Timi). Analele Banatului, 7+8:193-204 El Susi, G. 2003. Analogii i diferente ntre economiile animaliere ale comunittilor Vina C i grupi foeni n aezri din Banat. Banatica, 16:135-151 Elkin, D.C. 1995. Volume density of South-American camelid skeletal parts. International Journal of Osteoarchaeology, 5:29-37 Enloe, J.G. & F. David. 1989. Le remontage des os par individus: le partage du renne chez les Magdalniens de Pincevent (La Grande Paroisse, Seine-et-Marne). Bulletin de la Socit Prhistorique Franaise, 86:275-281 Enloe, J.G. & E. Turner. 2006. Methodological problems and biases in age determinations: a view from the Magdelanian. In D. Ruscillo (ed.) Recent Advances in Ageing and Sexing Animal Bones. pp. 129-144. Oxford: Oxbow. Ervynck, A. 1997. Detailed recording of tooth wear (Grant, 1982) as an evaluation of the seasonal slaughtering of pigs? Examples from medieval sites in Belgium. Archaeofauna, 6:67-79 Ervynck, A. 2005. Detecting the seasonal slaughtering of domestic mammals: inferences from the detailed recording of tooth eruption and wear. Environmental Archaeology, 10:153-169 Fiddes, N. 1991. Meat: a natural symbol. London: Routledge. Fletcher, T.J. 1974. The timing of reproduction in red deer (Cervus elaphus) in relation to latitude. Journal of Zoology, 172:363-367
328
References
Garaanin, M. 1949. Naselja i stan prvobitnog oveka neolitskog doba u Srbiji. Istoriski Glasnik, 2:38-67 Garaanin, M. 1951. Hronologija Vinanske grupe. Ljubljana: Arheoloki Seminar. Garaanin, M. 1979. Centralnobalkanska zona. In A. Benac (ed.) Praistorija jugoslavenskih zemalja II: Neolit. pp. 33-79. Sarajevo: Akademija nauka i umetnosti Bosne i Hercegovine. Garaanin, M. 1984. Uvod. In Anon. (ed.) Vina u Praistoriji i Srednjem Veku. Belgrade: Galerija Srpska Akademija Nauka i Umetnosti. Garaanin, M. 1998. Vinanska kultura i jadranski uticajiArheoloko Blago Kosova i Metohije od Neolita do Ranog Srednjeg Veka. pp. 56-87. Belgrade: Serbian Academy of Arts and Sciences. Gardner, A. 1999. The ecology of Neolithic impacts - re-evaluation of existing theory using case studies from Hungary & Slovenia. Documenta Praehistorica, 26:163-183 Gell, A. 1998. Art and Agency: an anthopological theory. New York: Clarendon. Gimbutas, M. 1974. The gods and godesses of Old Europe 7000:3500 B.C.: myths, legends and cult images. Berkeley: University of California Press. Gimbutas, M. 1976. Chronology. In M. Gimbutas (ed.) Neolithic Macedonia: as reflected by excavation at Anza, Southeast Yugoslavia. Monumenta Archaeologica 1. pp. 29-77. Los Angeles: University of California Institute of Archaeology. Giri, M. 1960. Iskopavanja na Gomolavi 1957 godine. Rad Vojvoanskih Muzeja, 9:130-150 Giri, M. 1965. Osvrt na dosadanja iskopavanja na Gomolavi. Rad Vojvoanskih Muzeja, 14:109-112 Glser, R. 1996. Zur absoluten Datierung der Vina-Kultur anhand von 14C-Daten. In F. Draovean (ed.) The Vina Culture, its Role and Cultural Connections. pp. 175-212. Timioara: Museum of Banat. Goodman, F. 1992. Ecstasy, Ritual and Alternate Reality. Bloomington: Indiana University Press. Gtherstrm, A., C. Anderung, L. Hellborg, R. Elburg, C. Smith, D.G. Bradley & H. Ellegren. 2005. Cattle domestication in the Near East was followed by hybridization with aurochs bulls in Europe. Proceedings of the Royal Society B: Biological Sciences, 272:2345-2350 Grant, A. 1975. Appendix B: The use of tooth wear as a guide to the age of domestic animals a brief explanation. In B. Cunliffe (ed.) Excavations at Portchester Castle. pp. 437450. London: Society of Antiquaries. Grant, A. 1982. The use of tooth wear as a guide to the age of domestic ungulates. In B. Wilson, C. Grigson & S. Payne (eds) Ageing and sexing animal bones from archaeological sites. British Archaeological Reports British Series 109. pp. 91-108. Oxford: British Archaeological Reports. Grant, A. 2002. Food, status and social hierarchy. In P.T. Miracle & N. Milner (eds) Consuming Passions and Patterns of Consumption. pp. 17-23. Cambridge: McDonald Institute. Grayson, D.K. 1978. Minimum numbers and sample size in vertebrate faunal analysis. American Antiquity, 43:53-65
329
References
Grayson, D.K. 1979. On the quantification of vertebrate archaeofaunas. In M.B. Schiffer (ed.) Advances in Archaeological Method and Theory. pp. 199-237. New York: Academic Press. Grayson, D.K. 1981. The effects of sample size on some derived measures in vertebrate faunal analysis. Journal of Archaelogical Science, 8:77-88 Grayson, D.K. 1984. Quantitative Zooarchaeology. New York: Academic Press. Grayson, D.K. & C.J. Frey. 2004. Measuring skeletal part representation in archaeological faunas. Journal of Taphonomy, 2:27-42 Greenfield, H.J. 1986. The Paleoeconomy of the Central Balkans (Serbia) (British Archaeological Reports International Series 304. Oxford: BAR. Greenfield, H.J. 1988. The origins of milk and wool production in the Old World: a zooarchaeological perspective from the central Balkans. Current Anthropology, 29:573-593 Greenfield, H.J. 1989. Zooarchaeology and aspects of the secondary products revolution: a central Balkan perspective. Archaeozoologia, 3:191-200 Greenfield, H.J. 1991. Fauna from the Late Neolithic of the Central Balkans: issues in subsistence and land use. Journal of Field Archaeology, 18:161-186 Greenfield, H.J. 1993. Faunal remains from the Early Neolithic Starevo settlement at Bukovaka esma. Starinar, 43-44:103-113 Greenfield, H.J. 1999. The origins of metallurgy: distinguishing stone from metal cut-marks on bones from archaeological sites. Journal of Archaeological Science, 26:797-808 Greenfield, H.J. 2005. A reconsideration of the Secondary Products Revolution in southeastern Europe: on the origins and use of domestic animals for milk, wool and traction in the central Balkans. In J. Mulville & A.K. Outram (eds) The Zooarchaeology of Fats, Oils, Milk and Dairying. pp. 14-31. Oxford: Oxbow. Greenfield, H.J. 2006. Sexing fragmentary ungulate acetabulae. In D. Ruscillo (ed.) Recent Advances in Ageing and Sexing Animal Bones. pp. 68-86. Oxford: Oxbow. Greenfield, H.J. & F. Draovean. 1994. Preliminary report on the 1992 excavations at FoeniSla: an early Neolithic Starevo-Cri settlement in the Romanian Banat. Analele Banatului, 3:45-86 Greenfield, H.J. & . Je. n.d. Excavations at Petnica, Yugoslavia: preliminary report for 1982-1986. In Petnica Science Centre - archaeology department archives. Valjevo. Greenfield, H.J. & T. Jongsma. 2006. The intrasettlement spatial patterning of Early Neolithic settlements in temperate southeastern Europe: a view from Blagotin, Serbia. In E.C. Robertson, J.D. Seibert, D.C. Fernandez & M.U. Zender (eds) Space and Spatial Analysis in Archaeology. pp. 69-79. Calgary: University of Calgary Press. Greenfield, H.J. & T. Jongsma. 2008. Sedentary pastoral gatherers in the early Neolithic: architectural, botanical, and zoological evidence for mobile economies from FoeniSala, south-west Romania. In D. Bailey, A. Whittle & D. Hofmann (eds) Living Well Together? Settlement and materiality in the Neolithic of South-East and Central Europe. pp. 108-130. Oxford: Oxbow. Griffin, D. 1976. The Question of Animal Awareness. New York: Rockefeller University Press. Grigson, C. 1969. The uses and limitations of differences in absolute size in the distinction between the bones of aurochs (Bos primigenius) and domestic cattle (Bos taurus). In
330
References
P.J. Ucko & G.W. Dimbleby (eds) The Domestication and Exploitation of Plants and Animals. pp. 277-294. London: Duckworth. Grigson, C. 1981. Fauna. In I. Simmons & M. Tooley (eds) The Environment in British Prehistory. pp. 110-124. London: Duckworth. Grigson, C. 1982. Sex and age determination of some bones and teeth of domestic cattle: a review of the literature. In B. Wilson, C. Grigson & S. Payne (eds) Ageing and Sexing Animal Bones from Archaeological Sites. British Archaeological Reports British Series 109. pp. 7-23. Oxford: BAR. Gromova, V. 1950. Opredelitelj mlekopitajuih SSSR po kostjam skeleta, v.1: opredelitelj po krupnim trubastim kostjamTrudi Komisii po izueniju etvertinovo perioda. pp. 1105. Moscow: Akademija nauk SSSR. Gromova, V. 1960. Opredelitelj mlekopitajuih SSSR po kostjam skeleta, v.2: opredelitelj po krupnim kostjam zapljusniTrudi Komisii po izueniju etvertinovo perioda. pp. 1116. Moscow: Akademija nauk SSSR. Groves, C. 2007. Current views on taxonomy and zoogeography of the genus Sus. In U. Albarella, K. Dobney, A. Ervynck & P.A. Rowley-Conwy (eds) Pigs and Humans: 10,000 years of interaction. pp. 15-29. Oxford: Oxford University Press. Habermehl, K.-H. 1985. Altersbestimmung bei Wild- und Pelztieren. Hamburg: Paul Parey. Hall, S.J.G. 1989. Chillingham cattle: social and maintenance behaviour in an ungulate that breeds all year round. Animal Behaviour, 38:215-225 Hallowell, I. 1960. Ojibwa ontology, behavior, and worldview. In S. Diamond (ed.) Culture in History: essays in honor of Paul Radin. pp. 19-52. New York: Columbia University Press. Halstead, P. 1985. A study of mandibular teeth from Romano-British contexts at Maxey. In F. Pryor, C. French, D. Crowther, D. Gurney, G. Simpson & M. Taylor (eds) The Fenland Project No. 1: archaeology and the environment in the Lower Welland Valley. East Anglian Archaeological Report 27. pp. 219-224. Cambridge: Cambridgeshire Archaeological Committee. Halstead, P. 1989. The economy has a normal surplus: economic stability and social change among early farming communities of Thessaly, Greece. In P. Halstead & J.M. OShea (eds) Bad Year Economics. pp. 68-80. Cambridge: Cambridge University Press. Halstead, P. 1996. Pastoralism or household herding? Problems of scale and specialization in early Greek animal husbandry. World Archaeology, 28:20-42 Halstead, P. 1999. Neighbours from hell? The household in Neolithic Greece. In P. Halstead (ed.) Neolithic Society in Greece. pp. 77-95. Sheffield: Sheffield Academic Press. Halstead, P. 2005. Resettling the Neolithic: faunal evidence for seasons of consumption and residence at Neolithic sites in Greece. In D. Bailey, A. Whittle & V. Cummings (eds) (Un)settling the Neolithic. pp. 38-50. Oxford: Oxbow. Halstead, P. 2006. Sheep in the garden: the integration of crop and livestock husbandry in early farming regimes of Greece and southern Europe. In D. Serjeantson & D. Field (eds) Animals in the Neolithic of Britain and Europe. pp. 42-55. Oxford: Oxbow. Halstead, P. 2007. Carcasses and commensality: investigating the social context of meat consumption in Neolithic and Early Bronze Age Greece. In C. Mee & J. Renard (eds) Cooking up the Past: food and culinary practices in the Neolithic and Bronze Age Aegean. pp. 25-48. Oxford: Oxbow.
331
References
Halstead, P., P. Collins & V. Isaakidou. 2002. Sorting the Sheep from the Goats: Morphological Distinctions between the Mandibles and Mandibular Teeth of Adult Ovis and Capra. Journal of Archaeological Science, 29:545-553 Hamayon, R. 1990. La Chasse a lme: esquisse dune thorie du chamanisme Sibrien. Nanterre: Socit dEthnologie. Hambleton, E. 1999. Animal Husbandry Regimes in Iron Age Britain (British Archaeological Reports British Series 282. Oxford: Archaeopress. Hamilakis, Y. 2001. Re-inventing environmental archaeology In U. Albarella (ed.) Environmental Archaeology: meaning and purpose. pp. 29-38. New York: Kluwer. Hamilakis, Y. 2003. The sacred geography of hunting: wild animals, social power and gender in early farming societies. In E. Kotjabopoulou, Y. Hamilakis, P. Halstead, C. Gamble & P. Elefanti (eds) Zooarchaeology in Greece: Recent Advances. pp. 239247. London: British School at Athens. Hammond, J., Jr., I.L. Mason & T.J. Robinson. 1971. Hammonds Farm Animals. London: Edward Arnold. Harker, A.B. 1968. Sheep husbandry. In W.P. Blount (ed.) Intensive Livestock Farming. pp. 64-86. London: Heinemann. Harris, D.R. 1996. Domesticatory relationships of people, plants and animals. In R.F. Ellen & K. Fukui (eds) Redefining Nature: ecology, culture and domestication. pp. 437-463. Oxford: Berg. Hart, L.A. 1995. Dogs as human companions: a review of the relationship. In J. Serpell (ed.) The Domestic Dog, its Evolution, Behaviour and Interactions with People. pp. 161178. Cambridge: Cambridge University Press. Hayden, B. 1996. Feasting in prehistoric and traditional societies. In P. Wiessner & W. Schiefenhovel (eds) Food and the Status Quest. pp. 127-147. Providence: Berghahn. Hecker, H.M. 1982. Domestication revisited: its implications for faunal analysis. Journal of Field Archaeology, 9:217-236. Helmer, D. & M. Rocheteau. 1994. Atlas du Squelette Appendiculaire des Principaux Genres Holocenes de Petits Rumiant du Nord de la Mditerrane et du Proche-Orient (Capra, Ovis, Rupicapra, Capreolus, Gazella). Juan-les-Pins: Editions APDCA. Hesse, B. 1988. Comment on H. J. Greenfield 'Origins of milk and wool production...' Current Anthropology, 29:590-591 Hesse, B. & P. Wapnish. 1985. Animal Bone Archaeology. Washington: Taraxacum. Higgs, E.S. & M.R. Jarman. 1972. The origins of animal and plant husbandry. In E.S. Higgs (ed.) Papers in Economic Prehistory: studies by members and associates of the British Academy major research project in the early history of agriculture. pp. 3-13. Cambridge: Cambridge University Press. Higgs, E.S. & M.R. Jarman. 1975. Palaeoeconomy. In E.S. Higgs (ed.) Palaeoeconomy. pp. 1-7. Cambridge: Cambridge University Press. Higham, C. 1968a. Patterns of prehistoric economic exploitation on the Alpine foreland. Vierteljahrsschrift der Naturforschenden Gesellschaft, 113:41-92 Higham, C. 1968b. Size trends in prehistoric European domestic fauna and the problem of local domestication. Acta Zoologica Fennica, 120:1-21
332
References
Hill, J.D. 1995. Ritual and Rubbish in the Iron Age of Wessex: a study on the formation of a specific archaeological record (British Archaeological Reports British Series 242. Oxford: Archaeopress. Hillson, S.M. 1992. Mammal Bones and Teeth: an introductory guide to methods of identification. London: UCL Institute of Archaeology. Hillson, S.M. 2005. Teeth. Cambridge: Cambridge University Press. Hodder, I. 1986. Reading the Past. Cambridge: Cambridge University Press. Hodder, I. 1987. The meaning of discard: ash and domestic space in Baringo. In S. Kent (ed.) Method and Theory for Activity Area Research. pp. 424-448. New York: Colombia University Press. Hodder, I. 1990. The Domestication of Europe: structure and contingency in Neolithic Europe. Oxford: Blackwell. Hodder, I. 2006. atalhyk: The Leopard's Tale. Revealing the mysteries of Turkey's ancient 'town'. London: Thames & Hudson. Holste, F. 1939. Zur chronologische Stellung der Vina-Keramik. Wiener Prhistorische Zeitschrift, 26:1-21 Hopf, M. 1974. Pflanzen reste aus Siedlungen der Vina-Kultur in Jugoslawien. Jahrbuch des Rmisch-Germanischen Zentralmuseums, Mainz, 21:1-11 Horvth, F. & E. Hertelendi. 1994. Contribution to the 14C based absolute chronology of the Early and Middle neolithic Tisza region. Jsa Andrs Mzeum vknyve 36:11-133 Hster-Plogmann, H., J. Schibler & K. Steppan. 1999. The relationship between wild mammal exploitation, climatic fluctuations and economic adaptations. A transdisciplinary study on Neolithic sites from the Lake Zurich region, Southwest Germany and Bavaria. In C. Becker, H. Manhart, J. Peters & J. Schibler (eds) Historia Animalium ex Ossibus: festschrift fr Angela von den Driesch. pp. 189-200. Rahden: Leidorf. Ingold, T. 1980. Hunters, Pastoralists and Ranchers. Cambridge: Cambridge University Press. Ingold, T. 1984. Time, social relations and the exploitation of animals: anthropological reflections on prehistory. In J. Clutton-Brock & C. Grigson (eds) Animals and Archaeology 3: early herders and their flocks. British Archaeological Reports International Series 202. pp. 3-12. Oxford: BAR. Ingold, T. 1986. The Appropriation of Nature: essays on human ecology and social relations. Manchester: Manchester University Press. Ingold, T. 1988. Introduction. In T. Ingold (ed.) What Is an Animal? pp. 1-16. London: Unwin Hyman. Ingold, T. 1994. From trust to domination: an alternative history of human-animal relations. In A. Manning & J. Serpell (eds) Animals and Human Societies: changing perspectives. pp. 1-22. London: Routledge. Ingold, T. 1996a. Hunting and gathering as ways of perceiving the environment. In R.F. Ellen & K. Fukui (eds) Redefining Nature: ecology, culture and domestication. pp. 117155. Oxford: Berg. Ingold, T. 1996b. The optimal forager and economic man. In P. Descola & G. Plsson (eds) Nature and Society: anthropological perspectives. pp. 25-44. London: Routledge.
333
References
Ingold, T. 1998. Nature, culture, environment: steps to an ecology of life. In B. Cartledge (ed.) Mind, Brain and Environment: the Linacre Lectures 1995-96. pp. 158-180. Oxford: Oxford University Press. Ingold, T. 2000. The Perception of the Environment: essays in livelihood, dwelling and skill. London: Routledge. Ioannidou, E. 2003. Taphonomy of animal bones: species, sex, age and breed variability of sheep, cattle and pig bone density. Journal of Archaeological Science, 30:355-365 Isaakidou, V. 2006. Ploughing with cows: Knossos and the Secondary Products Revolution. In D. Serjeantson & D. Field (eds) Animals in the Neolithic of Britain and Europe. pp. 95-112. Oxford: Oxbow. Jarman, M.R. 1972. European deer economies and the advent of the Neolithic. In E.S. Higgs (ed.) Papers in Economic Prehistory: studies by members and associates of the British Academy major research project in the early history of agriculture. pp. 125147. Cambridge: Cambridge University Press. Jarman, M.R. & P.F. Wilkinson. 1972. Criteria of animal domestication. In E.S. Higgs (ed.) Papers in Economic Prehistory: studies by members and associates of the British Academy major research project in the early history of agriculture. pp. 83-96. Cambridge: Cambridge University Press. Je, . 1985. Pregled neolitskih i eneolitskih kultura Gornje Kolubare. Istraivanja, 2 Johnson, E. 1985. Current developments in bone technology. In M.B. Schiffer (ed.) Advances in Archaeological Method and Theory. pp. 157-235. New York: Academic Press. Jones, G.G. 2006. Tooth eruption and wear observed in live sheep from Butser Hill, the Cotswold Farm Park and five farms in the Pentland Hills, UK. In D. Ruscillo (ed.) Recent Advances in Ageing and Sexing Animal Bones. pp. 155-178. Oxford: Oxbow. Jongsma, T. 1997. Distinguishing pits from pit houses through daub analysis: the nature and location of Early Neolithic Starevo-Cri houses at Foeni-Sala, Romania. Unpublished M.A. thesis, University of Manitoba. Jongsma, T. & H.J. Greenfield. 1996. The vertebrate fauna from Middle and Late Neolithic Snandrei, SW Romania, 1992. In F. Draovean (ed.) The Vina Culture, its Role and Cultural Connections. pp. 295-308. Timioara: Museum of Banat. Jovanovi, B. 1965a. Gomolava Hrtkovci - vieslojno nalazite, I. Neka pitanja o stratigrafiji Gomolave. Arheoloki Pregled, 7:16-20 Jovanovi, B. 1965b. Opta stratigrafija Gomolave. Rad Vojvoanskih Muzeja, 14:113-137 Jovanovi, B. 1969. Gomolava, Hrtkovci, Srem. Arheoloki Pregled, 11:24-28 Jovanovi, B. 1971a. Metalurgija Eneolitskog Perioda Jugoslavije. Belgrade: Archaeological Institute. Jovanovi, B. 1971b. Stratigrafija Gomolave u iskopavanjima 1967-1971. godine. Rad Vojvoanskih Muzeja, 20:95-102 Jovanovi, B. 1990. Die Vina-Kultur und der Beginn der Metallnutzung auf dem Balkan. In D. Srejovi & N. Tasi (eds) Vina and its World. pp. 55-60. Belgrade: Serbian Academy of Sciences and Arts. Jovanovi, B. & J. Glisi. 1960. Eneolitsko naselje na Kormadinu kod Jakova. Starinar (new series), 11:113-142
334
References
Jovanovi, B. & B.S. Ottaway. 1976. Copper mining and metallurgy in the Vina group. Antiquity, 50:104-113 Jovanovi, M. 2004. itarice u praistoriji u podunavlju i na balkanskom poluostrvu. Rad Muzeja Vojvodine, 46:101-127 Jovanovi, S., M. Savi, R. Trailovi, Z. Jankovi & D. ljivar. 2003. Evaluations of the domestication process in Serbia: palaeozoological remnants at Neolithic settlement of Belovode. Acta Veterinaria (Beograd), 53:427-434 Jovanovi, S., M. Savi, R. Trailovi, . Jankovi & D. ljivar. 2004. Evaluation of the domestication process in Serbia - domestication of neolithic cattle. Acta Veterinaria (Beograd), 54:467-473 Kaiser, T. & B. Voytek. 1983. Sedentism and economic change in the Balkan Neolithic. Journal of Anthropological Archaeology, 2:323-353 Kelly, R.L. 1992. Mobility/sedentism, concepts, archaeological measures, and effects. Annual Review of Anthropology, 21:43-66 Kent, S. 1989. Cross-cultural perceptions of farmers as hunters and the value of meat. In S. Kent (ed.) Farmers as Hunters. pp. 1-17. Cambridge: Cambridge University Press. Kent, S. 1993. Variability in faunal assemblages: the influence of hunting skill, sharing, dogs, and mode of cooking on the faunal remains at a sedentary Kalahari community. Journal of Anthropological Archaeology, 12:323-385 Keswani, P.S. 1994. The social context of animal husbandry in early agricultural societies: ethnographic insights and an archaeological example from Cyprus. Journal of Anthropological Archaeology, 13:255-277 Klein, R.G. & K. Cruz-Uribe. 1983. The computation of ungulate age (mortality) profiles from dental crown heights. Paleobiology, 9:70-78 Klein, R.G. & K. Cruz-Uribe. 1984. The Analysis of Animal Bones from Archaeological Sites. Chicago: University of Chicago Press. Knight, J. (ed.) 2005a. Animals in Person: cultural perspectives on human-animal intimacy. Oxford: Berg. Knight, J. 2005b. Introduction. In J. Knight (ed.) Animals in Person: cultural perspectives on human-animal intimacy. pp. 1-13. Oxford: Berg. Kohl, G. & H. Quitta. 1970. Berlin radiocarbon measurements IV. Radiocarbon, 12:400-420 Konigsberg, L.W. & S.R. Frankenberg. 1992. Estimation of age structure in anthropological demography. American Journal of Physical Anthropology, 89:235-256 Kotsakis, K. 2005. Across the border: unstable dwellings and fluid landscapes in the earliest Neolithic of Greece. In D. Bailey, A. Whittle & V. Cummings (eds) (Un)settling the Neolithic. pp. 8-15. Oxford: Oxbow. Kozdrowski, R., M. Dzieciol & A. Dubiel. 2005. Reproductive biology of roe deer. Medycyna Weterynaryjna, 61:1242-1244 Kreutzer, L.A. 1992. Bison and deer bone-mineral densities - comparisons and implications for the interpretation of archaeological faunas. Journal of Archaeological Science, 19:271-294 Kubinyi, E., Z. Virnyi & A. Miklsi. 2007. Comparative social cognition: from wolf to dog to humans. Comparative Cognition and Behaviour Reviews, 2:26-46
335
References
Lam, Y.M., X.B. Chen & O.M. Pearson. 1999. Intertaxonomic variability in patterns of bone density and the differential representation of bovid, cervid, and equid elements in the archaeological record. American Antiquity, 64:343-362 Lam, Y.M. & O.M. Pearson. 2005. Bone density studies and the interpretation of the faunal record. Evolutionary Anthropology, 14:99-108 Lam, Y.M., O.M. Pearson, C.W. Marean & X. Chen. 2003. Bone density studies in zooarchaeology. Journal of Archaeological Science, 30:1701-1708 Lamming, G.E. 1975. The control of oestrus in cattle. In J.C. Taylor (ed.) The Early Calving of Heifers and its Impact on Beef Production. pp. 47-57. Brussels: Commission of the European Communities. Larson, G., U. Albarella, K. Dobney, P.A. Rowley-Conwy, J. Schibler, A. Tresset, J.-D. Vigne, C.J. Edwards, A. Schlumbaum, A. Dinu, A. Blsescu, G. Dolman, A. Tagliacozzo, N. Manaseryan, P. Miracle, L.V. Wijngaarden-Bakker, M. Masseti, D.G. Bradley & A. Cooper. 2007. Ancient DNA, pig domestication, and the spread of the Neolithic into Europe. Proceedings of the National Academy of Sciences, 104:15276-15281 Lszl, A. 1997. Datarea prin radiocarbon n arheologie. Bucureti Lathrap, D.W. 1968. The hunting economies of the tropical forest zone of South America: an attempt at historical perspective. In R.B. Lee & I. DeVore (eds) Man the Hunter. pp. 23-30. Chicago: Aldine. Lauwerier, R.C.G.M. 1983. Pigs, piglets and determining the season of slaughtering. Journal of Archaeological Science, 10:483-488 Lawrence, M. & R. Brown. 1967. Mammals of Britain: their tracks, trails and signs. London: Blandford. Lazarovici, G. 1981. Die Periodisierung der Vina-Kultur in Rumnien. Prhistorische Zeitschrift, 56:169-196 Lazarovici, G. 1989. Das neolitische Heiligtum von Para. In S. Bknyi (ed.) Neolithic of Southeastern Europe and its Near Eastern Connections. pp. 149-174. Budapest: Archaeolingua. Lazarovici, G., F. Draovean & Z. Maxim. 2001. Para. Timioara: Museum Banaticum Temesiense. Lazi, M. 1988. Fauna Kimenjaka sa Boljevaca i njeno Mesto u Istoriji Faune Neolitskih i Eneolitskih Lokaliteta Centralnog Balkana. Unpublished Magisterium. Lazi, M. 1993. Eksploatacija domaih i divljih ivotinja centralnobalkanskom podruju. Glasnik SAD, 9:93-98 tokom neolita na
Leach, H.M. 2007. Selection and the unforseen consequences of domestication. In R. Cassidy & M.H. Mullin (eds) Where the Wild Things are Now: domestication reconsidered. pp. 71-99. Oxford: Berg. Lederman, R. 1986. What Gifts Engender: social relations and politics in Mendi, Highland Papua New Guinea. Cambridge: Cambridge University Press. Legge, A.J. 1990. Animals, economy and environment. In R. Tringham & D. Krsti (eds) Selevac: a Neolithic village in Yugoslavia. pp. 215-242. Los Angeles: University of California Press. Legge, A.J. & P.A. Rowley-Conwy. 1988. Star Carr Revisited. London: Birkbeck College Centre for Extra-Mural Studies.
336
References
Lekovi, V. 1990. The vinanization of Starevo culture. In D. Srejovi & N. Tasi (eds) Vina and its World. pp. 67-74. Belgrade: Serbian Academy of Sciences and Arts. Lenneis, E. & P. Stadler. 1995. Zur Absolutchronologie der Linearbandkeramik aufgrund von 14 C-Daten. Archologie sterreichs, 6:4-12 Levi-Strauss, C. 1962. Le Totmisme Aujourd'hui Paris: Presses Universitaires de France. Lichter, C. 1993. Untersuchungen zu den Bauten des sdosteuropischen Neolithikums und Chalkolithikums. Buch am Erlbach: Verlag Marie Leidorf. Linares, O. 1976. Garden hunting in the American tropics. Human Ecology, 4:331-349. Linick, T.W. 1977. La Jolla natural radiocarbon measurements VII. Radiocarbon, 19:19-48 Lubbock, J. 1865. Pre-historic Times. London: Williams & Norgate. Luca, S.A. 2003. Date noi cu privire la cronologia absolut a eneoliticului timpuriu din Transilvania - Rezultatele prelucrrii probelor radiocarbon de la OrtieDealul Pemilor, punct X2, jud. Hunedoara. Banatica, 16:77-102 Luca, S.A., D. Diaconescu, A. Georgescu & C. Suciu. 2006. Cercetrile arheologice de la Miercurea Sibiului Petri, (jud. Sibiu) campaniile anilor 1997-2005. Stratigrafie i cronologie. Brukenthal. Acta Musei, 1:9-20 Lyman, R.L. 1984. Bone density and differential survivorship of fossil classes. Journal of Anthropological Archaeology, 3:259-299 Lyman, R.L. 1985. Bone frequencies: differential transport, in situ destruction and the MGUI. Journal of Archaelogical Science, 12:221-236 Lyman, R.L. 1993. Density-mediated attrition of bone assemblages: new insights. In J. Hudson (ed.) From Bones to Behaviour: ethnoarchaeological and experimental contributions to the interpretation of faunal remains. pp. 324-341. Carbondale: Southern Illinois University. Lyman, R.L. 1994. Vertebrate Taphonomy. Cambridge: Cambridge University Press. MacCormack, C.P. & M. Strathern. 1980. Nature, Culture & Gender. Cambridge: Cambridge University Press. Mace, R. & A. Houston. 1989. Pastoralist strategies for survival in unpredictable environments: a model of herd composition that maximises household variability. Agricultural Systems, 31:185-204 Madrigal, T.C. & J.Z. Holt. 2002. White-tailed deer meat and marrow return rates and their application to Eastern Woodlands archaeology. American Antiquity, 67:745-759 Magnell, O. 2006. Tooth wear in wild boar (Sus scrofa). In D. Ruscillo (ed.) Recent Advances in Ageing and Sexing Animal Bones. pp. 189-203. Oxford: Oxbow. Makkay, J. 1990. The Protovina problem - as seen from the northernmost frontier. In D. Srejovi & N. Tasi (eds) Vina and its World. pp. 113-122. Belgrade: Serbian Academy of Sciences and Arts. Malinowski, B. 1922. Argonauts of the Western Pacific. Marangou, C. 1996. Assembling, displaying and dissembling Neolithic and Eneolithic figurines and models. Journal of European Archaeology, 4:177-202 Marciniak, A. 1999. Faunal materials and interpretive archaeology epistemology reconsidered. Journal of Archaeological Theory and Method, 6:293-320
337
References
Marciniak, A. 2001. Scientific and interpretive components in social zooarchaeology. The case of early farming communities in Kujavia. Archaeologia Polona, 39:87-110 Marciniak, A. 2005a. Placing Animals in the Neolithic: social zooarchaeology of prehistoric farming communities. London: UCL Press. Marciniak, A. 2005b. Social changes in the early European Neolithic. A taphonomy perspective. In T.P. OConnor (ed.) Biosphere to Lithosphere: new studies in vertebrate taphonomy. pp. 148-154. Oxford: Oxbow. Marean, C.W. 1991. Measuring the postdepositional destruction of bone in archaeological assemblages. Journal of Archaeological Science, 18:677-694 Marean, C.W., Y. Abe, P.J. Nilssen & E.C. Stone. 2001. Estimating the minimum number of skeletal elements (MNE) in zooarchaeology: a review and a new image-analysis GIS approach. American Antiquity, 66:333-348 Markoti, V. 1984. The Vina Culture. Calgary: Western. Marshall, F. 1993. Food sharing and the faunal record. In J. Hudson (ed.) From Bones to Behaviour: ethnoarchaeological and experimental contributions to the interpretation of faunal remains. pp. 228-246. Carbondale: Southern Illinois University. Marshall, F. & 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 Martin, L. & N. Russell. 2000. Trashing rubbish. In I. Hodder (ed.) Towards Reflexive Method in Archaeology: the example at atalhyk. pp. 57-69. Cambridge: McDonald Institute. Marx, K. 1964. The Economic and Political Manuscripts of 1884 (trans.) M. Milligan. New York: International Publishers. Mauget, R. 1982. Seasonality of reproduction in the wild boar. In D.J.A. Cole & G.R. Foxcroft (eds) Control of Pig Reproduction. pp. 509-526. London: Butterworth Scientific. Mauss, M. 1954. The Gift: forms and functions of exchange in archaic societies (trans.) I. Cunnison. London: Cohen & West. McPherron, A. 1988. Introduction. In A. McPherron & D. Srejovi (eds) Divostin and the Neolithic of Central Serbia. pp. 1-9. Pittsburgh: University of Pittsburgh. McPherron, A., V. Bucha & M.J. Aitken. 1988. Absolute dating of Divostin, Grivac-Barice and Banja. In A. McPherron & D. Srejovi (eds) Divostin and the Neolithic of Central Serbia. pp. 379-387. Pittsburgh: University of Pittsburgh. McPherron, A. & K.C. Christopher. 1988. The Balkan Neolithic and the Divostin project in perspective. In A. McPherron & D. Srejovi (eds) Divostin and the Neolithic of Central Serbia. pp. 463-489. Pittsburgh: University of Pittsburgh. McPherron, A. & J. Gunn. 1988. Quantitative analysis of excavated materials at Divostin. In A. McPherron & D. Srejovi (eds) Divostin and the Neolithic of Central Serbia. pp. 359-377. Pittsburgh: University of Pittsburgh. Meadow, R.H. 1978. Bonecode system of numerical coding for faunal data from Middle Eastern sites. In R.H. Meadow & M.A. Zeder (eds) Approaches to Faunal Analysis in the Middle East. pp. 169-186. Cambridge Mass.: Peabody Museum. Meadow, R.H. 1984. Animal domestication in the Middle East: a view from the eastern margin. . In J. Clutton-Brock & C. Grigson (eds) Animals and archaeology 3: early
338
References
herders and their flocks. British Archaeological Reports International Series 202. pp. 309-337. Oxford: BAR. Meadow, R.H. 1989. Osteological evidence for the process of animal domestication. In J. Clutton-Brock (ed.) The Walking Larder. pp. 80-90. London: Unwin Hyman. Meggitt, M. 1958. The Enga of the New Guinean Highlands: some preliminary observations. Oceania, 28:253-330 Metcalfe, D. & K.T. Jones. 1988. A reconsideration of animal body-part utility indexes. American Antiquity, 53:486-504 Midgley, M. 1983. Animals and Why They Matter. Athens: University of Georgia Press. Milner, N. 2005. Can seasonality studies be used to identify sedentism in the past? In D. Bailey, A. Whittle & V. Cummings (eds) (Un)settling the Neolithic. pp. 32-37. Oxford: Oxbow. Miloji, V. 1949. Chronologie der Jngeren Steinzeit Mittel- und Sdosteuropas. Berlin: Gebr. Mann. Milojkovi, J. 1990. The anthropomorphic and zoomorphic figurines. In R. Tringham & D. Krsti (eds) Selevac: a Neolithic village in Yugoslavia. pp. 397-436. Los Angeles: University of California Press. Milton, K. 2005. Anthropomorphism or egomorphism? The perception of non-human persons by human ones. In J. Knight (ed.) Animals in Person: cultural perspectives on human-animal intimacy. pp. 79-100. Oxford: Berg. Miracle, P. & L. Pugsley. 2006. Vertebrate faunal remains from Pupiina Cave. In P. Miracle & S. Forenbaher (eds) Prehistoric Herders of Northern Istria: the archaeology of Pupicina Cave. pp. 260-399. Pula: Arheoloki Muzej Istre. Miracle, P.T. 2002. Mesolithic meals from Mesolithic middens. In P.T. Miracle & N. Milner (eds) Consuming Passions and Patterns of Consumption. pp. 65-88. Cambridge: McDonald Institute. Miracle, P.T. 2006. Neolithic sheperds and their herds in the Northern Adriatic Basin. In D. Serjeantson & D. Field (eds) Animals in the Neolithic of Britain and Europe. pp. 6394. Oxford: Oxbow. Miracle, P.T. & N. Milner (eds) 2002. Consuming Patterns and Patterns of Consumption. Cambridge: McDonald Institute. Mitchell, B., W. Grant & J. Cubby. 1981. Notes on the performance of red deer, Cervus elaphus, in a woodland habitat. Journal of Zoology, 194:279-284 Mleku, D. 2007. 'Sheep are your mother': rhyta and the interspecies politics in the Neolithic of the eastern Adriatic. Documenta Praehistorica, 34 Monchot, H., M. Mashkour & J.-D. Vigne. 2005. Kernel smoothing and mixture analyses for the determination of the sex ratios at death, at the beginning of the domestication of ungulates. In J.-D. Vigne, J. Peters & D. Helmer (eds) First Steps of Animal Domestication: new archaeozoological approaches. pp. 55-60. Oxford: Oxbow. Monks, G.G. 1981. Seasonality Studies. In M.B. Schiffer (ed.) Advances in Archaeological Method and Theory, vol. 4. pp. 177-240. New York: Academic Press. Moran, N.C. & T.P. O'Connor. 1994. Age attribution in domestic sheep by skeletal and dental maturation: a pilot study of available sources. International Journal of Osteoarchaeology, 4:267-285
339
References
Morey, D.F. 2006. Burying key evidence: the social bond between dogs and people. Journal of Archaeological Science, 33:158-175 Morlan, R.E. 1994. Bison bone fragmentation and survivorship: a comparative method. Journal of Archaeological Science, 21:797-807 Morris, B. 1998. The Power of Animals: an ethnography. Oxford: Berg. Moskalewska, A.L. & V. Sanev. 1989. Preliminary analysis of bone remnants of animals from the Neolithic archaeological site Tumba Madari near Skopje (Yugoslavia). Macedoniae Acta Archaeologica, 10:55-75 Mukherjee, A.J., M.S. Copley, R. Berstan, K.A. Clark & R.P. Evershed. 2005. Interpretation of 13C values of fatty acids in relation to animal husbandry: food processing and consumption in prehistory. In J. Mulville & A.K. Outram (eds) The Zooarchaeology of Fats, Oils, Milk and Dairying. pp. 77-93. Oxford: Oxbow. Mullin, M.H. 1999. Mirrors and windows: sociocultural studies of human-animal relationships. Annual Review of Anthropology, 28:201-224 Munson, P.J. & R.C. Garniewicz. 2003. Age-mediated survivorship of ungulate mandibles and teeth in canid-ravaged faunal assemblages. Journal of Archaeological Science, 30:405-416 Na, . 1957. Gomolava kod Hrkovaca. Preistorijako, rimsko i srednjevekovno nalazite. Starinar, 7-8:403-404 Na, . 1960. Zatito iskopavanje na Gomolavi kod Hrtkovca - prethodni izvetaj za 19551958. godini. Rad Vojvoanskih Muzeja, 9 Nadasdy, P. 2007. The gift in the animal: the ontology of hunting and human-animal sociality. American Ethnologist, 34:25-43 Nanoglou, S. 2008. Representation of Humans and Animals in Greece and the Balkans during the Earlier Neolithic. Cambridge Archaeological Journal, 18:1-13 Nicholson, R.A. 1993. A morphological investigation of burnt animal bone and an evaluation of its utility in archaeology. Journal of Archaelogical Science, 20:411-428 Nicolescu-Plopor, C.S. 1980. The osteological remains from Rast. In V. Dumitrescu (ed.) The Neolithic Settlement at Rast. British Archaeological Reports International Series 72. pp. 124-125. Oxford: BAR. Noddle, B.A. 1982. The size of red deer in Britain - past and present, with some reference to fallow deer. In M. Bell & S. Limbrey (eds) Archaeological Aspects of Woodland Ecology. British Archaeological Reports, International Series 146. Oxford: BAR. Nowak, R.M. & J.L. Paradiso. 1983. Walker's Mammals of the World2). Baltimore: The John Hopkins University Press. OConnor, T.P. 1997. Working at relationships: another look at animal domestication. Antiquity, 71:149-156 OConnor, T.P. 1998. On the difficulty of detecting seasonal slaughtering of sheep. Environmental Archaeology, 3:5-12 OConnor, T.P. 2000. The Archaeology of Animal Bones. Stroud: Sutton. OConnor, T.P. 2003. The Analysis of Urban Animal Bone Assemblages. York: Council for British Archaeology.
340
References
OShea, J.M. 1989. The role of wild resources in small-scale agricultural systems: tales from the Lakes and the Plains. In P. Halstead & J.M. OShea (eds) Bad Year Economics. pp. 57-67. Cambridge: Cambridge University Press. Obeli, B., M.K. krivanko, B. Marijan & I.K. Broni. 2004. Radiocarbon dating of Sopot culture sites (Late Neolithic) in eastern Croatia. Radiocarbon, 46:245-258 Oma, K. 2006. Human-Animal Relationships: mutual becomings in the household of Scandinavia and Sicily 900-500 BC. Unpublished Ph.D. thesis, University of Southampton. Ortner, S.B. 1974. Is female to male as nature is to culture? In M.Z. Rosaldo & L. Lamphere (eds) Woman, Culture and Society. pp. 67-88. Stanford: Stanford University Press. Orton, C.R. 2000. Sampling in Archaeology. Cambridge: Cambridge University Press. Orton, D.C. 2003. The Construction of Seasonal Mortality Profiles for Domestic Sheep Based on Radiographic Assessment of Tooth Development. Unpublished B.A. thesis, University of Cambridge. Orton, D.C. 2004. Mortality Profiling for the Cattle Mandibles from the West Yorkshire Chariot Burial. Unpublished M.Sc. thesis, Univeristy of York. Outram, A.K. 2001. A new approach to identifying bone marrow and grease exploitation: why the indeterminate fragments should not be ignored. Journal of Archaeological Science, 28:401-410 Outram, A.K. 2002. Bone fracture and within-bone nutrients: an experimentally based method for investigating levels of marrow extraction. In P. Miracle & N. Milner (eds) Consuming Patterns and Patterns of Consumption. pp. 51-63. Cambridge: McDonald Institute. zdoan, M. 1993. Vina and Anatolia: a new look at a very old problem (or redefining Vina culture from the perspective of Near Eastern tradition). Anatolica, 19:173-193 zdoan, M. 1997. The Beginnings of Neolithic Economies in Southeastern Europe: an Anatolian perspective. Journal of European Archaeology, 5:1-33 Pales, L. & M.A. Garcia. 1981. Mammifres de Quaternaire. Paris: CNRS. Plsson, G. 1996. Human-environmental relations: orientalism, paternalism and communalism. In P. Descola & G. Plsson (eds) Nature and Society: anthropological perspectives. pp. 63-81. London: Routledge. Pappa, M., P. Halstead, K. Kotsakis & D. Urem-Kotsou. 2004. Evidence for large-scale feasting at Late Neolithic Makriyalos, N Greece. In P. Halstead & J.C. Barrett (eds) Food, Cuisine and Society in Prehistoric Greece. pp. 16-44. Oxford: Oxbow. Parker Pearson, M. 2000. Eating money: a study in the ethnoarchaeology of food. Archaeological Dialogues, 7:217-232 Payne, S. 1972. On the interpretation of bone samples from archaeological sites. In E.S. Higgs (ed.) Papers in Economic Prehistory. pp. 65-81. Cambridge: Cambridge University Press. Payne, S. 1973. Kill-off patterns in sheep and goats: the mandibles from Asvan Kale. Anatolian Studies, 23:281-303 Payne, S. 1985. Morphological distinctions between the mandibular teeth of young sheep, Ovis, and goats, Capra. Journal of Archaeological Science, 12:139-147
341
References
Payne, S. 1987. Reference codes for wear states in the mandibular cheek teeth of sheep and goats. Journal of Archaelogical Science, 14:609-614 Payne, S. & G. Bull. 1988. Components of variation in measurements of pig bones and teeth, and the use of measurements to distinguish wild from domestic pig remains. Archaeozoologia, 2:27-66 Pearce, J. & R. Luff. 1994. The taphonomy of cooked bone. In R. Luff & P.A. RowleyConwy (eds) Whither Environmental Archaeology. pp. 51-56. Oxford: Oxbow. Peri, S. 1998. Vieslojna neolitska naselja i problem kulturne stratigrafije neolita na teritoriji Srbije. Starinar, 49:11-38 Perls, C. 2001. The Early Neolithic in Greece. Cambridge: Cambridge University Press. Petrovi, J. 1978. Gomolava, Hrtkovci, vieslojno nalazite. Arheoloki Pregled, 20:32-34 Petrovi, J. 1979. Gomolava, Hrtkovci, vieslojno nalazite. Arheoloki Pregled, 21:20-22 Petrovi, J. 1982. Gomolava, Hrtkovci, vieslojno nalazite. Arheoloki Pregled, 23:15-20 Petrovi, J. 1984. Gomolava. Novi Sad: Museum of Vojvodina. Petrovi, J. 1992. Arhitektura kuce 4 na Gomolavi. Rad Vojvoanskih Muzeja, 34:19-32 Petrovi, J. 1993. Keramika i alatke is Kue 4 na Gomolavi, naselje mlae vinanske kulture. Rad Vojvoanskih Muzeja, 35:7-26 Pickering, T.R. & K.J. Carlson. 2002. Baboon bone mineral densities: implications for the taphonomy of primate skeletons in South African cave sites. Journal of Archaeological Science, 29:883-896 Pickering, T.R., C.W. Marean & M. Dominguez-Rodrigo. 2003. Importance of limb bone shaft fragments in zooarchaeology: a response to On in situ attrition and vertebrate body part profiles (2002), by M.C. Stiner. Journal of Archaeological Science, 30:1469-1482 Pike-Tay, A. (ed.) 2001. Innovations in assessing season of capture, age and sex of archaeofaunas. Grenoble: La Pense Sauvage. Pipe, A. n.d. The animal bones from a Neolithic settlement at Bardhosh, Kosova - excavation 2002: Museum of London Specialist Services. Pitts, M. 1979. Hides and antlers: a new look at the hunter-gatherer site at Star Carr, North Yorkshire, England. World Archaeology, 11:32-42 Pluciennik, M. 1998. Deconstructing the Neolithic' in the Mesolithic-Neolithic transition. In M. Edmonds & C. Richards (eds) Understanding the Neolithic of North-Western Europe. pp. 61-83. Glasgow: Cruithne. Pollard, J. 2006. A community of beings: animals and people in the Neolithic of southern Britain. In D. Serjeantson & D. Field (eds) Animals in the Neolithic of Britain and Europe. Neolithic Studies Group Seminar Papers 7. pp. 135-148. Oxford: Oxbow. Pond, W.G. 1983. Modern pork production. Scientific American, 5:78-87 Prizer, K.M. 1994. Food Sharing in the Neolithic: refitting red deer skeletons from Opovo, Serbia. Unpublished Senior Honors thesis, University of California at Berkeley. Prummel, W. 1988. Distinguishing features of postcranial skeletal elements of cattle, Bos primigenius f. taurus, and red deer, Cervus elaphus. Schriften aus der ArchaologischZoologischen Arbeitsgruppe Schleswig-Kiel, 12:1-52.
342
References
Prummel, W. & H.J. Frisch. 1986. A guide for the distinction of species, sex and body side in bones of sheep and goats. Journal of Archaeological Science, 13:567-577 Pugsley, L. 2002. Exploitation Patterns: food utility and bone mineral density indices for wild and domestic pigs. Unpublished Ph.D. thesis, University of Cambridge. Quitta, H. & G. Kohl. 1969. Neue Radiocarbondaten zum Neolithicum und zur frhen Bronzezeit Sdosteuropas und der Sowjetunion. Zeitschrift fur Arch, Radovanovi, I. 1999. "Neither person nor beast" - dogs in the burial practice of the Iron Gates Mesolithic. Documenta Praehistorica, 26:71-87 Rafferty, J.E. 1985. The archaeological record on sedentariness: recognition, development and implications. Advances in Archaeological Method and Theory, 8:113-56 Rappaport, R. 1968. Pigs for the Ancestors: ritual in the ecology of a New Guinea people. New Haven: Yale University Press. Raajski, R. 1954. Gomolava kod Hrtkovaca. Rad Vojvoanskih Muzeja, 3:187-219 Ray, K. & J. Thomas. 2003. In the kinship of cows: the social centrality of cattle in the earlier Neolithic of southern Britain. In M.P. Pearson (ed.) Food, Culture and Identity in the Neolithic and Early Bronze Age. British Archaeological Reports International Series 1117. pp. 37-51. Oxford: Archaeopress. Reid, A. 1996. Cattle herd and the redistribution of cattle resources. World Archaeology, 28:43-57 Reimer, P.J., M.G.L. Baillie, E. Bard, A. Bayliss, J.W. Beck, C. Bertrand, P.G. Blackwell, C.E. Buck, G. Burr, K.B. Cutler, P.E. Damon, R.L. Edwards, R.G. Fairbanks, M. Friedrich, T.P. Guilderson, K.A. Hughen, B. Kromer, F.G. McCormac, S. Manning, C. Bronk Ramsey, R.W. Reimer, S. Remmele, J.R. Southon, M. Stuiver, S. Talamo, F.W. Taylor, J.v.d. Plicht & C.E. Weyhenmeyer. 2004. Radiocarbon, 46:1029-1058. Reitz, E.J. & E.S. Wing. 1999. Zooarchaeology. Cambridge: Cambridge University Press. Richards, C. & J. Thomas. 1984. Ritual activity and structured deposition in later Neolithic Wessex. In R. Bradley & J. Gardiner (eds) Neolithic Studies: a review of some current research. British Archaeological Reports British Series 133. Oxford: BAR. Ringrose, T.J. 1993. Bone counts and statistics: a critique. Journal of Archaeological Science, 20:121-157 Robb, J.E. & M.A. Dobres. 2005. Doing agency introductory remarks on methodology. Journal of Archaeological Method and Theory, 12:159-166 Rogers, A.R. 2000a. Analysis of bone counts by maximum likelihood. Journal of Archaeological Science, 27:111-125 Rogers, A.R. 2000b. On equifinality in faunal analysis. American Antiquity, 65:709-723 Rogers, A.R. & J.M. Broughton. 2001. Selective transport of animal parts by ancient hunters: a new statistical method and an application to the Emeryville Shellmound fauna. Journal of Archaeological Science, 28:763-773 Rosman, A. & P.G. Rubel. 1989. Stalking the wild pig: hunting and horticulture in Papua New Guinea. In S. Kent (ed.) Farmers as Hunters. pp. 27-36. Cambridge: Cambridge University Press. Rowley-Conwy, P.A. 2003. Early domestic animals in Europe: imported or locally domesticated? In A. Ammerman & P. Biagi (eds) The Widening Harvest. The
343
References
Neolithic transition in Europe: looking back, looking forward. pp. 99-117. Boston: Archaeological Institute of America. Ruscillo, D. (ed.) 2006. Recent Advances in Ageing and Sexing Animal Bones. Oxford: Oxbow. Russell, N. 1990. The bone tools. In R. Tringham & D. Krsti (eds) Selevac: a Neolithic village in Yugoslavia. pp. 521-548. Los Angeles: University of California Press. Russell, N. 1993. Hunting, Herding and Feasting: human use of animals in Neolithic Southeast Europe. Unpublished Ph.D. thesis, University of California at Berkeley. Russell, N. 1998. Cattle as wealth in Neolithic Europe: where's the beef? In D.W. Bailey (ed.) The Archaeology of Value: essays on prestige and the processes of valuation. British Archaeological Reports International Series 730. pp. 42-54. Oxford: Archaeopress. Russell, N. 1999. Symbolic dimensions of animals and meat at Opovo, Yugoslavia. In J.E. Robb (ed.) Material Symbols: culture and economy in prehistory. pp. 153-172. Carbondale: Center for Archaeological Investigations. Russell, N. 2000. Household variation and meat sharing in Neolithic (spatial dimensions of the faunal remains from Opovo, Yugoslavia). In L. Nikolova (ed.) Technology, Style and Society: contributions to the innovations between the Alps and the Black Sea in prehistory. British Archaeological Reports International Series 854. pp. 41-50. Oxford: Archaeopress. Russell, N. 2002. The wild side of animal domestication. Society and Animals, 10:285-302 Russell, N. 2007. The domestication of anthropology. In R. Cassidy & M.H. Mullin (eds) Where the Wild Things are Now: domestication reconsidered. pp. 27-48. Oxford: Berg. Ryder, M.L. 1982. Sheep - Hilzheimer 45 years on. Antiquity, 56:15-23 Sahlins, M.D. 1968. Tribesmen. Englewood Cliffs: Prentice Hall. Sahlins, M.D. 1972. Stone Age Economics. London: Tavistock. Sahlins, M.D. 1976. Culture and Practical Reason. Chicago: University of Chicago Press. Schier, W. 1996. The relative and absolute chronology of Vina: new evidence from the type site. In F. Draovean (ed.) The Vina Culture, its Role and Cultural Connections. pp. 141-162. Timioara: Museum of Banat. Schier, W. 2000. Measuring change: the Neolithic pottery sequence of Vina-Belo Brdo. Documenta Praehistorica, 27:187-197 Schier, W. 2005. Masken Menschen Rituale: Alltage und Kult vor 7000 Jahren in der prhistorischen Siedlung von Uivar, Rumnien. Wrzburg: Universitt Wrzburg. Schier, W. 2008. Uivar: a late Neolithic-early Eneolithic fortified tell site in western Romania. In D. Bailey, A. Whittle & D. Hofmann (eds) Living Well Together? Settlement and materiality in the Neolithic of South-East and Central Europe. pp. 5467. Oxford: Oxbow. Schiffer, M.B. 1976. Behavioural Archaeology. New York: Academic Press. Schiffer, M.B. 1983. Towards the identification of formation processes. American Antiquity, 48:675-706 Schmid, E. 1972. Atlas of Animal Bones: for prehistorians, archaeologists and quaternary geologists. New York: Elsevier.
344
References
Schwartz, C.A. 1976. The vertebrate fauna from Rug Bair. In M. Gimbutas (ed.) Neolithic Macedonia: as reflected by excavation at Anza, Southeast Yugoslavia. Monumenta archaeologica 1. pp. 364-374. Los Angeles: University of California Institute of Archaeology. Schwartz, C.A. 1992. arkovo Ledine: the vertebrate and molluscan remains 1948 season. Zbornik Naradnog Muzeja (Beograd), 14:123-133 Sekere, L. 1961. Iskopavanja na Gomolavi u Hrtkovcima 1957. godine - parcijalni izvetaj. Rad Vojvoanskih Muzeja, 10:79-88 Serjeantson, D. 2006. Food or feast at Neolithic Runnymede? In D. Serjeantson & D. Field (eds) Animals in the Neolithic of Britain and Europe. Neolithic Studies Group Seminar Papers 7. pp. 113-134. Oxford: Oxbow. Serpell, J. (ed.) 1995a. The Domestic Dog, its Evolution, Behaviour and Interactions with People. Cambridge: Cambridge University Press. Serpell, J. 1995b. From paragon to pariah: some reflections on human attitudes to dogs. In J. Serpell (ed.) The Domestic Dog, its Evolution, Behaviour and Interactions with People. pp. 245-256. Cambridge: Cambridge University Press. Serpell, J.A. 1986. In the Company of Animals. Oxford: Blackwell. Service, E.R. 1962. Primitive Social Organization. New York: Random House. Shanklin, E. 1985. Sustenance and symbol: anthropological studies of domesticated animals. Annual Review of Anthropology, 14 Shapland, A. 2006. Over the horizon: human-animal relations in Neopalatial Crete. Paper presented to the International Council for Archaeozoology, Mexico City, August 2006. Shapland, A. in press. Wild nature: human-animal relations on Neopalatial Crete. In J. Mulville & A. Powell (eds) A Walk on the Wild Side: hunting in farming societies. Oxford: Oxbow. Sharples, N. 2000. Antlers and Orcadian rituals: an ambiguous role for red deer in the Neolithic. In A. Ritchie (ed.) Neolithic Orkney in its European Context. pp. 107-116. Cambridge: Cambridge University Press. Shennan, S. 1997. Quantifying Archaeology. Edinburgh: Edinburgh University Press. Sherratt, A. 2005. Settling the Neolithic: a digestif. In D. Bailey, A. Whittle & V. Cummings (eds) (Un)settling the Neolithic. pp. 140-146. Oxford: Oxbow. Sherratt, A.G. 1981. Plough and Pastoralism: aspects of the Secondary Products Revolution In I. Hodder, G. Isaac & N. Hammond (eds) Patterns of the Past: studies in honour of David Clarke. pp. 261-306. Cambridge: Cambridge University Press. Sherratt, A.G. 1984. Social evolution: europe in the later Neolithic and Copper Ages. In J. Bintliff (ed.) European Social Evolution: archaeological perspectives. pp. 123-134. Bradford: University of Bradford Press. Shipman, P., G. Foster & M. Schoeninger. 1984. Burnt bones and teeth: an experimental study of color, morphology, crystal structure and shrinkage. Journal of Archaeological Science, 11:307-326 Shotwell, J.A. 1958. Inter-community relationships in Hemphillian (Mid-Pliocene) mammals. Ecology, 39:271-282
345
References
Sillitoe, P. 2007. Pigs in the New Guinea Highlands: an ethnographic example. In U. Albarella, K. Dobney, A. Ervynck & P.A. Rowley-Conwy (eds) Pigs and Humans: 10,000 years of interaction. pp. 330-356. Oxford: Oxford University Press. Silver, I.A. 1969. The ageing of domestic animals. In D. Brothwell & E. Higgs (eds) Science in Archaeology: a survey of progress and research. London: Thames & Hudson. Simpson, G.G., A. Roe & R.C. Lewontin. 1960. Quantitative Zoology. New York: Harcourt Brace. ljivar, D. 1996. The eastern settlement of the Vina culture at Plonik: a relationship of its stratigraphy to the hoards of copper objects. Starinar, 47:85-97 ljivar, D. & D. Jacanovi. 2005. Zoomorphic figurines from Belovode. Zbornik Narodnog Muzeja (Beograd), 18:69-78 Snyder, L.M. & E.A. Moore (eds) 2006. Dogs and People in Social, Working, Economic or Symbolic Interaction. Oxford: Oxbow. Souvatzi, S. 2008. Household dynamics and variability in the Neolithic of Greece: the case for a bottom-up approach. In D. Bailey, A. Whittle & D. Hofmann (eds) Living Well Together? Settlement and materiality in the Neolithic of South-East and Central Europe. pp. 17-27. Oxford: Oxbow. Speth, J.D. & S.L. Scott. 1989. Horticulture and large-mammal hunting: the role of resource depletions and constraints of time and labor. In S. Kent (ed.) Farmers as Hunters: the implications of sedentism. pp. 71-79. Cambridge: Cambridge University Press. Srdo, D., B. Obeli, A. Sliepevi, I.K. Broni & N. Horvatini. 1987. Rudjer Boskovi Insitute radiocarbon measurements X. Radiocarbon, 29:135-147 Srdo, D., A. Sliepevi, B. Obeli & N. Horvatini. 1977. Rudjer Boskovi Insitute radiocarbon measurements IV. Radiocarbon, 19:465-475 Srdo, D., A. Sliepevi & J. Planini. 1975. Rudjer Boskovi Insitute radiocarbon measurements III. Radiocarbon, 17:149-155 Srejovi, D. 1988. The Neolithic of Serbia: a review of research. In D. Srejovi (ed.) The Neolithic of Serbia: archaeological research 1948-1988. pp. 5-19. Belgrade: Faculty of Philosophy. Stahl, P.W. 1999. Structural density of domesticated South American camelid skeletal elements and the archaeological investigation of prehistoric Andean charki. Journal of Archaeological Science, 26:1347-1368 Starovi, A. 1997. Petnica '97 Dnevnik Iskopavanja. In Petnica Science Centre - archaeology department archives. Valjevo: Istraivaka Stanica Petnica. Stevanovi, J. & B. Mihaljevi. 1996. Stambena arhitektura naselja vinansko-plonike faze u valjevskom kraju. Petnike Sveske, 42:10-19 Stevanovi, M. 1996. The age of clay: the social dynamics of house destruction. Unpublished PhD thesis, University of California at Berkeley. Stevanovi, M. 1997. The age of clay: the social dynamics of house destruction. Journal of Anthropological Archaeology, 16:334-395 Stevanovi, M. & R. Tringham. 1998. The significance of Neolithic houses in the archaeological record of south-east Europe. In M. Lazi, M. Garaanin, N. Tasi, A. Cermanovi-Kuzmanovi, P. Petrovi, Z. Miki & M. Ruzi (eds) Zbornik posveen Dragoslavu Srejoviu. pp. 193-208. Belgrade: Balkanological Institute.
346
References
Stiner, M.C. 1991. Food procurement and transport by human and non-human predators. Journal of Archaeological Science, 18:455-482 Stiner, M.C. 2002. On in situ attrition and vertebrate body part profiles. Journal of Archaeological Science, 29:979-991 Strathern, M. 1980. No nature, no culture: the Hagen Case. In C.P. MacCormack & M. Strathern (eds) Nature, Culture & Gender. pp. 174-222. Cambridge: Cambridge University Press. Strathern, M. 1988. The Gender of the Gift. Berkeley: University of California Press. Studer, J. & D. Pillonel. 2007. Traditional pig butchery by the Yali people of West Papua (Irian Jaya): an ethnographic and archaeozoological example. In U. Albarella, K. Dobney, A. Ervynck & P.A. Rowley-Conwy (eds) Pigs and Humans: 10,000 years of interaction. pp. 308-329. Oxford: Oxford University Press. Svensson, A. & O. Wendel. 1965. The Techniques of Crime Scene Investigation, 2nd edn. New York: Elsevier. Sykes, N. & R.H. Symmons. 2007. Sexing cattle horn-cores: problems and progress. International Journal of Osteoarchaeology:514-523 Symmons, R.H. 2002. A Re-examination of Sheep Bone Density and its Role in Assessing Taphonomic Histories of Zooarchaeological Assemblages. Unpublished Ph.D. thesis, University College London. Symmons, R.H. 2004. Digital photodensitometry: a reliable and accessible method for measuring bone density. Journal of Archaeological Science, 31:711-719 Symmons, R.H. 2005. Bone density variation between similar animals and density variation in early life: implications for future taphonomic analysis. In T.P. O'Connor (ed.) Biosphere to Lithosphere: new studies in vertebrate taphonomy. pp. 86-93. Oxford: Oxbow. Tambiah, S.J. 1969. Animals are good to think and good to prohibit. Ethnology, 8:424-459 Tangri, D. & R.V.S. Wright. 1993. Multivariate analysis of compositional data: applied comparisons favour standard principal components analysis over Aitchisons loglinear contrast method. Archaeometry, 35:103112 Tanner, A. 1979. Bringing Home Animals: religious ideology and mode of production of the Mistassini Cree Hunters. New York: St Martin's Press. Tapper, R.L. 1988. Animality, humanity, morality, society. In T. Ingold (ed.) What Is an Animal? pp. 47-62. London: Unwin Hyman. Tasi, N. 1998. Eneolit. Arheoloko Blago Kosova i Metohije od Neolita do Ranog Srednjeg Veka. pp. 89-115. Belgrade: Serbian Academy of Arts and Sciences. Tasi, N. & B. Brukner. 1975. Gomolava, Hrtkovci - vieslojno nalazite. Arheoloki Pregled, 17:11-13 Tasi, N. & B. Brukner. 1976. Gomolava, Hrtkovci - vieslojno nalazite. Arheoloki Pregled, 18:12-14 Tasi, N. & J. Petrovi (eds) 1988. Gomolava: symposium. Novi Sad: Muzej Vojvodine. Theodossopoulos, D. 2005. Care, order and usefulness: the context of the human-animal relationship in a Greek island community. In J. Knight (ed.) Animals in Person: cultural perspectives on human-animal intimacy. pp. 15-35. Oxford: Berg. Thomas, J. 1991. Rethinking the Neolithic. Cambridge: Cambridge University Press.
347
References
Thomas, J. 1993. Discourse, totalization and 'The Neolithic''. In C. Tilley (ed.) Interpretative Archaeology. pp. 357-393. Oxford: Berg. Todorovi, J. & A. ermanovi. 1961. Banjica: naselje vinanske kulture. Belgrade: Muzej Grada. Trigger, B.G. 2006. A History of Archaeological Thought. Cambridge: Cambridge University Press. Tringham, R. 1971. Hunters, Fishers and Farmers of Eastern Europe: 6000-3000BC. London: Hutchinson. Tringham, R. 1991. Households with faces: the challenge of gender in prehistoric architectural remains. In J. Gero & M. Conkey (eds) Engendering Archaeology. pp. 93-131. Oxford: Blackwell. Tringham, R. 1992. Life after Selevac: why and how a Neolithic settlement is abandoned. Balcanica, 23:133-145 Tringham, R. 2000a. The continuous house: a view from the deep past. In R.A. Joyce & S.D. Gillespie (eds) Beyond Kinship: social and material reproduction in house societies. pp. 115134. Philadelphia: University of Pennsylvania Press. Tringham, R. 2000b. Southeastern Europe in the transition to agriculture in Europe: bridge, buffer or mosaic. In T.D. Price (ed.) Europes First Farmers. pp. 19-56. Cambridge: Cambridge University Press. Tringham, R. 2005. Weaving house life and death into places: a blueprint for a hypermedia narrative. In D. Bailey, A. Whittle & V. Cummings (eds) (Un)settling the Neolithic. pp. 98-111. Oxford: Oxbow. Tringham, R., B. Brukner, T. Kaiser, K. Borojevi, L. Buhvi, P. teli, N. Russell, M. Stevanovi & B. Voytek. 1992. Excavations at Opovo, 1985-1987: socioeconomic change in the Balkan Neolithic. Journal of Field Archaeology, 19:351-386 Tringham, R., B. Brukner & B. Voytek. 1985. The Opovo project: a study of socioeconomic change in the Balkan Neolithic. Journal of Field Archaeology, 12:425-444 Tringham, R. & D. Krsti. 1990a. Relative and absolute chronology. In R. Tringham & D. Krsti (eds) Selevac: a Neolithic village in Yugoslavia. pp. 45-55. Los Angeles: University of California Press. Tringham, R. & D. Krsti. 1990b. Selevac and the transformation of southeast European prehistory. In R. Tringham & D. Krsti (eds) Selevac: a Neolithic village in Yugoslavia. pp. 567-616. Los Angeles: University of California Press. Tringham, R. & M. Stevanovi. 1990. Field research. In R. Tringham & D. Krsti (eds) Selevac: a Neolithic village in Yugoslavia. pp. 57-213. Los Angeles: University of California Press. Tripkovi, B. 2003. A dialogue between the household and community: a case study of Banjica. In L. Nikolova (ed.) Early Symbolic Systems for Communication in Southeast Europe. British Archaeological Reports International Series 1139. Oxford: Archaeopress. Uerpmann, H.-P. 1979. Probleme der Neolithisierung des Mittelmeeraumes (Wiesbaden: Biehefte zum Tbinger Atlas des Vordern Orients, Reihe A 27. van Zeist, W. 2002. Plant husbandry of tell Gomolava , Vojvodina, Yugoslavia. Palaeohistoria, 43/44:87-115
348
References
Vasi, M.M. 1910. Die Hauptergebnies der prahistorischen Ausgrabung in Vina im Jahre 1908. Prhistorische Zeitschrift, 2:23-39 Vasi, M.M. 1936. Vina IV. Belgrade: Dravne tamparija. Viveiros de Castro, E. 1998. Cosmological deixis and Amerindian perspectivism. The Journal of the Royal Anthropological Institute, 4:469-488 Vlahovi, P. & N. Miloevi. 1971. Petnika Peina, Valjevo - paleolitska slanica. Arheoloki Pregled, 13:10-16 Vogel, J.C. & H.T. Waterbolk. 1963. Groningen radiocarbon dates IV. Radiocarbon, 5:163202 von den Driesch, A. 1976. A Guide to the Measurement of Animal Bones from Archaeological Sites. Cambridge, Mass.: Harvard University Peabody Museum. Vrs, I. 1980. Zoological and palaeoeconomical investigations of the Early Neolithic Krs culture. Folia Archaeologica, 31:3-64 Waterbolk, H.T. 1988. C14-datirungen von Gomolava. In N. Tasi & J. Petrovi (eds) Gomolava: symposium. pp. 117-121. Novi Sad: Muzej Vojvodine. Watson, J.P.N. 1979. The estimation of the relative frequencies of mammalian species: Khirokitia 1972. Journal of Archaeological Science, 6:127-137 Weiner, A. 1985. Inalienable wealth. American Ethnologist, 12:210-227 Weiner, A. 1992. Inalienable Possessions: the paradox of keeping-while-giving. Berkeley: University of California Press. Westropp, H. 1872. Prehistoric Phases. London: Bell & Daldy. White, T.E. 1953. A method of calculating the dietary percentage of various food animals utilized by aboriginal peoples. American Antiquity, 18:396-398 Whittle, A. 1996. Europe in the Neolithic: the creation of new worlds. Cambridge: Cambridge University Press. Whittle, A. 1997. Moving on and moving around: Neolithic settlement mobility. In P. Topping (ed.) Neolithic Landscapes. pp. 15-22. Oxford: Oxbow. Whittle, A. 2003. The Archaeology of People: dimensions of Neolithic life. London: Routledge. Whittle, A. 2005. Lived experience in the Early Neolithic of the Great Hungarian Plain. In D. Bailey, A. Whittle & V. Cummings (eds) (Un)settling the Neolithic. pp. 64-70. Oxford: Oxbow. Whittle, A., L. Bartosiewicz, D. Bori, P. Pettitt & M. Richards. 2002. In the beginning: new radiocarbon dates for the Early Neolithic in northern Serbia and south-east Hungary. Antaeus, 25:63-117 Whittle, A., L. Bartosiewicz, D. Bori, P. Pettitt & M. Richards. 2005. New radiocarbon dates for the Early Neolithic in northern Serbia and south-east Hungary: some omissions and corrections. Antaeus, 28:347-355 Willey, P., A. Galloway & L. Snyder. 1997. Bone mineral density and survival of elements and element portions in the bones of the Crow Creek massacre victims. American Journal of Physical Anthropology, 104:513-528 Williams, H.L. 1968. The onset and the modification of the breeding season in sheep. In W.P. Blount (ed.) Intensive Livestock Farming. London: Heinemann.
349
References
Williams, R. 1972. Ideas of nature. In J. Benthall (ed.) Ecology: the shaping enquiry. pp. 146164. London: Longman. Williamson, R.W. 1912. The Mafulu: mountain people of British New Guinea. London: Macmillan. Willis, K.J. & K.D. Bennet. 1994. The Neolithic transition - fact or fiction? Palaeoecological evidence from the Balkans. The Holocene, 4:326-330 Wilson, B., C. Grigson & S. Payne (eds) 1982. Ageing and sexing animal bones from archaeological sites. British Archaeological Reports British Series 109. Oxford: British Archaeological Reports. Winterhalder, B. 1981. Optimal foraging strategies and hunter-gatherer research in anthropology: theory and models. In B. Winterhalder & E.A. Smith (eds) HunterGatherer Foraging Strategies: ethnographic and archeological analyses. Chicago: University of Chicago Press. Woodburn, J. 1998. 'Sharing is not a form of exchange': an analysis of property sharing in immediate-return hunter-gatherer societies. In C.M. Hann (ed.) Property Relations: renewing the anthropological tradition. pp. 48-63. Cambridge: Cambridge University Press. Zeder, M. 2005. A view from the Zagros: new perspectives on livestock domestication in the Fertile Crescent. In J.-D. Vigne, J. Peters & D. Helmer (eds) First Steps of Animal Domestication. pp. 125-146. Oxford: Oxbow. Zeder, M. 2006a. Central questions in the domestication of plants and animals. Evolutionary Anthropology, 15:105-117 Zeder, M. 2006b. Reconciling rates of long bone fusion and tooth eruption and wear in sheep (Ovis) and goat (Capra). In D. Ruscillo (ed.) Recent Advances in Ageing and Sexing Animal Bones. pp. 87-118. Oxford: Oxbow. Zohary, D., E. Tchernov & L.K. Horwitz. 1998. The role of unconscious selection in the domestication of sheep and goats. Journal of Zoology, 245:129-135 Zvelebil, M. 1992. Hunting in Anthropozoologica, 16:7-18 farming societies: the prehistoric perspective.
Zvelebil, M. 1998. Whats in a name: the Mesolithic, the Neolithic, and social change at the Mesolithic-Neolithic transition. In M. Edmonds & C. Richards (eds) Understanding the Neolithic of North-Western Europe. pp. 1-36. Glasgow: Cruithne Press. Zvelebil, M. & P. Rowley-Conwy. 1984. Transition to farming in northern Europe: a huntergatherer perspective. Norwegian Archaeological Review, 17:104-128 Zvelebil, M. & P. Rowley-Conwy. 1986. Foragers and farmers in Atlantic Europe. In M. Zvelebil (ed.) Hunters in Transition. pp. 67-93.
350

Beyond Hunting and Herding

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Beyond Hunting and Herding

Uploaded by

Copyright:

Available Formats

Beyond Hunting and Herding: Humans, animals, and the political economy of the Vina period.

David Clive Orton Darwin College

Faculty of Archaeology and Anthropology University of Cambridge

For Professor Clive Orton (dad)

Declaration of original research and content

2.4 Statement of aims __________________________________________________ 18

Chapter 3 Theoretical perspectives on animals in the Neolithic______________19

3.3 Animals as food ____________________________________________________ 42

3.4 Animals in the Vina period: a new model ______________________________ 46

3.5 Statement of hypotheses _____________________________________________ 49

Chapter 4 Outline of methodology _____________________________________50

4.4 Taphonomic and contextual analysis___________________________________ 60

4.5 Summary and structure of results _____________________________________ 77

Chapter 5 Regional trends in faunal data _______________________________78

5.3 Discussion ________________________________________________________ 108

5.4 Summary ________________________________________________________ 123

Chapter 6 Results from Gomolava and Petnica__________________________124

6.3 Gomolava ________________________________________________________ 150

6.4 Petnica __________________________________________________________ 204

6.5 Summary ________________________________________________________ 228

Chapter 7 Taphonomic analysis at Gomolava and Petnica ________________229

7.2 Evidence for peri-depositional damage ________________________________ 242

7.3 Breakage and fragmentation ________________________________________ 250

7.4 Visible human modification _________________________________________ 256

7.5 Body-part representation ___________________________________________ 270

7.6 Spatial distribution at Gomolava _____________________________________ 279

7.7 Discussion ________________________________________________________ 285 7.8 Conclusions ______________________________________________________ 291

Chapter 8 Synthesis and interpretation ________________________________292

8.3 Explaining changes ________________________________________________ 307

8.4 Wild:domestic? ___________________________________________________ 312

Chapter 9 Conclusions _____________________________________________316 References ________________________________________________________318

Chapter 2 Archaeological background and aims

2.2 Hunting versus herding

2.3 The Vina culture

Miloji 1949 Start

Chapman 1981, 18; b.c. End Range (BC)* 4260-4240 5450-5100

Tringham & Krsti 1990, 54; BC Start End ~5250 5000

Lenneis & Stadler 1995, 7; BC Start End 5300 Start** 5317-5267

New estimates see below; BC End** Range*** 5276-5214 5295-5250

Tordo I Tordo II Gradac Plonik I Plonik IIA

4260-4240 B2 C1 C2 D1 3950-3850 D2 4100

2.3.3 models of social change

2.4 Statement of aims

Theoretical perspectives on animals

Chapter 3 Theoretical perspectives on animals in the Neolithic

Theoretical perspectives on animals

Theoretical perspectives on animals

Theoretical perspectives on animals

3.2 Attitudes to animals

Theoretical perspectives on animals

Theoretical perspectives on animals

3.2.1 hunter-gather human-animal relations

Theoretical perspectives on animals

Theoretical perspectives on animals

Theoretical perspectives on animals

Viveiros de Castro 1998).

In this conception, hunting becomes a social rather than a

Theoretical perspectives on animals

Theoretical perspectives on animals

3.2.2.1 classical definitions

Theoretical perspectives on animals

3.2.2.2 ecological approaches

Theoretical perspectives on animals

7.7 Discussion __ 285 7.8 Conclusions 291

Chapter 9 Conclusions _316 References ____________318

New estimates see below; BC End Range* 5276-5214 5295-5250