Competition Between Humans and Large Carnivores: Case studies from the Late Middle and Upper Palaeolithic of the Central Balkans 9781407323770, 9781407355856

Citation preview

Competition Between Humans and Large Carnivores Case studies from the Late Middle and Upper Palaeolithic of the Central Balkans

Stefan Milošević B A R I N T E R NAT I O NA L S E R I E S 2 9 6 3

2020

210mm WIDTH

Competition Between Humans and Large Carnivores Case studies from the Late Middle and Upper Palaeolithic of the Central Balkans

Stefan Milošević B A R I N T E R NAT I O NA L S E R I E S 2 9 6 3

2020 297mm HIGH

210 x 297mm_BAR Milosevic TITLE ARTWORK.indd 1

17/1/20 6:42 PM

Published in 2020 by BAR Publishing, Oxford BAR International Series 2963 Competition Between Humans and Large Carnivores

978 1 4073 2377 0 paperback 978 1 4073 5585 6 e-format DOI https://doi.org/10.30861/9781407323770 A catalogue record for this book is available from the British Library isbn isbn

© Stefan Milošević 2020 cover image

Neanderthals hunting horses by Predrag Milošević

The Author’s moral rights under the 1988 UK Copyright, Designs and Patents Act, are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. Links to third party websites are provided by BAR Publishing in good faith and for information only. BAR Publishing disclaims any responsibility for the materials contained in any third party website referenced in this work.

BAR titles are available from: BAR Publishing 122 Banbury Rd, Oxford, OX2 7BP, UK email [email protected] phone +44 (0)1865 310431 fax +44 (0)1865 316916 www.barpublishing.com

Milosevic.indd 2

20/01/2020 13:24

Of Related Interest The Palaeolithic of the Balkans Proceedings of the XV IUPPS World Congress (Lisbon, 4-9 September 2006) Vol. 17 Session C33 Edited by Andreas Darlas and Dušan Mihailovic BAR International Series: 1819

Oxford, BAR Publishing, 2008

Sub-series: Proceedings of the XV World Congress UISPP (Lisbon, 4-9 September 2006) In Search of Total Animal Exploitation Case Studies from the Upper Palaeolithic and Mesolithic Edited by Laure Fontana, François-Xavier Chauvière and Anne Bridault BAR International Series: 2040

Oxford, BAR Publishing, 2009

Sub-series: Proceedings of the XV World Congress UISPP (Lisbon, 4-9 September 2006), 42 Relational Cohesion in Palaeolithic Europe Hominin-Cave Bear Interactions in Moravia and Silesia, Czech Republic, During OIS3 Patrick J. Skinner BAR International Series: 2379

Oxford, BAR Publishing, 2012

A Composite View to the Past A Methodological Integration of Zooarchaeology and Archaeological Geophysics at the Magdalenian Site of Verberie le Buis Jason Thompson BAR International Series: 2623

Oxford, BAR Publishing, 2014

Recent Discoveries and Perspectives in Human Evolution Papers arising from ‘Exploring Human Origins: Exciting Discoveries at the Start of the 21st Century’ Manchester 2013 Edited by Anek R. Sankhyan BAR International Series: 2719

Oxford, BAR Publishing, 2015

For more information, or to purchase these titles, please visit www.barpublishing.com iii

Milosevic.indd 3

02/01/2020 12:47

Abstract This research studies paleoecology and subsitence strategies of Palaeolithic hunter-gatherers in the Central Balkans during the Late Middle and Upper Palaeolithic (120-20 kya BP) through archaeozoological study of large mammal and avian remains from the cave sites of Pešturina, Hadži-Prodanova, and Smolućka. The aim of the is to understand ecological thresholds that influenced subsistence strategies and human ecological demands in the Central Balkans over the course of Last Interglacial and Glacial periods that are characterized by climatic instability affecting the distribution of various ecological niches and their biodiversity. Numerous remains of large carnivores in archaeozoological assemblages studied here show that they largely influenced the accumulation of ungulate remains in caves used by humans, reglardless of the intensity of human occupation. Important topics are: ecological continuity/discontinuity, relatively sparse traces of human occupation of palimpsests, similarities in prey choice between the humans and large carnivores. The results clearly point to ecological continuity of mosaic landscapes at least up to the Last Glacial Maximum (LGM), with impoverishment but not disintegration of ecological niches following the end of the Last Interglacial. Neanderthal occupations during Marine Isoptope Stage (MIS) 5 are characterized by diversified logistic subsistence, in which two ungulate species were preferred from the open grasslands, while several others were additionally hunted in different habitats (mostly forest). However, this strategy proved to be less successful during the Early Glacial, bringing higher mobility in settlement and subsistence patterns. Modern humans of the Gravettian culture employed a different subsistence strategy, with a constant focus on red deer or ibex, an intermediate ungulate species that can be encountered in a variety of habitats, imposing high mobility in that pursuit. Key words: archaeozoology, paleoecology, taphonomy, subsistence strategies, competition, Middle Palaeolithic, Gravettian, Central Balkans

iv

Milosevic.indd 4

02/01/2020 12:47

Contents List of Figures..................................................................................................................................................................... vii List of Tables........................................................................................................................................................................ ix Foreword............................................................................................................................................................................... x 1. Introduction..................................................................................................................................................................... 1 1.1 Aims of the study....................................................................................................................................................... 1 1.2 Chronological and biological framing of Upper Pleistocene humans, and their socio-economic behaviour ........... 3 1.3 Spatial framework of the study.................................................................................................................................. 7 1.4 History of research and archaeological data on the Palaeolithic in Serbia................................................................ 7 2. Theoretical Framework and Hypotheses.................................................................................................................... 12 2.1 Ecological niche concept......................................................................................................................................... 12 2.2 Foraging theories...................................................................................................................................................... 13 2.3 Hypotheses............................................................................................................................................................... 14 3. Materials and Methods................................................................................................................................................. 16 3.1 The sites................................................................................................................................................................... 16 3.2 Methodology of zooarchaeological analysis............................................................................................................ 19 3.2.1 Taxonomic determination, skeletal element determination and quantification of remains.............................. 20 3.2.2 Taphonomic analysis of the remains................................................................................................................ 21 3.2.3 Ageing of different mammalian species and age structure .............................................................................. 23 3.3 Spatial distributions and catchments........................................................................................................................ 24 3.4 Behaviour of wolves, spotted hyenas, lions, and bears............................................................................................ 24 4. Results............................................................................................................................................................................ 28 4.1 Pešturina................................................................................................................................................................... 28 4.1.1 Taxonomic composition................................................................................................................................... 28 4.1.2 Representation of skeletal elements................................................................................................................. 30 4.1.3 Taphonomic analysis........................................................................................................................................ 30 4.1.4 Age structures................................................................................................................................................... 45 4.1.5 Spatial distribution........................................................................................................................................... 45 4.2 The Hadži-Prodanova cave...................................................................................................................................... 45 4.2.1 Taxonomic composition................................................................................................................................... 47 4.2.2 Representation of skeletal elements................................................................................................................. 49 4.2.3 Taphonomic analysis........................................................................................................................................ 50 4.2.4 Age structure.................................................................................................................................................... 54 4.2.5 Spatial distribution........................................................................................................................................... 54 4.3 The Smolućka cave.................................................................................................................................................. 54 4.3.1 Taxonomic representation................................................................................................................................ 54 4.3.2 Representation of skeletal elements................................................................................................................. 55 4.3.3 Taphonomic analysis........................................................................................................................................ 55 4.3.4 Age structures................................................................................................................................................... 56 5. Discussion...................................................................................................................................................................... 66 5.1 Ecology of humans and carnivores during the Late Middle Palaeolithic and Gravettian in the Central Balkans... 66 5.2 Human and carnivore subsistence strategies of the Middle Palaeolithic and Gravettian of the Central Balkans.... 72 5.3 Late Neanderthal and Gravettian settlement strategies in the Central Balkans....................................................... 77 6. Conclusion..................................................................................................................................................................... 84 Appendix 1.......................................................................................................................................................................... 88 Appendix 2.......................................................................................................................................................................... 94 Appendix 3.......................................................................................................................................................................... 95 v

Milosevic.indd 5

02/01/2020 12:47

Competition Between Humans and Large Carnivores Appendix 4........................................................................................................................................................................ 100 Appendix 5........................................................................................................................................................................ 102 Appendix 6........................................................................................................................................................................ 104 Appendix 7........................................................................................................................................................................ 106 Appendix 8........................................................................................................................................................................ 112 Appendix 9a...................................................................................................................................................................... 113 Appendix 9b...................................................................................................................................................................... 114 Appendix 9c...................................................................................................................................................................... 115 Appendix 10...................................................................................................................................................................... 116 Appendix 11...................................................................................................................................................................... 117 Appendix 12...................................................................................................................................................................... 119 Appendix 13...................................................................................................................................................................... 121 Appendix 14...................................................................................................................................................................... 122 Appendix 15...................................................................................................................................................................... 123 Appendix 16...................................................................................................................................................................... 126 Appendix 17...................................................................................................................................................................... 128 Appendix 18...................................................................................................................................................................... 129 Appendix 19...................................................................................................................................................................... 132 Appendix 20...................................................................................................................................................................... 134 Bibliography..................................................................................................................................................................... 136

vi

Milosevic.indd 6

02/01/2020 12:47

List of Figures Figure 1. Distribution of various Transitional type industries and Proto-Aurignacian in Europe (45-38 kya BP).............. 2 Figure 2. Late Middle and Upper Palaeolithic sites on the Balkans studied in detail, and some sites in surrounding regions which are of importance for this study..................................................................................................................... 8 Figure 3. Map of systematically explored Palaeolithic sites in Serbia, and paleontological sites...................................... 10 Figure 4. Summary of optimal foraging. A decision was dependant upon the processing function between search/ processing time, distance from central place, energy spent by the hunters, and weight of anatomical unit/its caloric yield versus time required to process it. The point presents the start of gaining back spent caloric yield......................... 14 Figure 5. Geographic location, Pešturina cave plan, section and microregional position through DEM model from NE direction............................................................................................................................................................... 17 Figure 6. Geographic position, section and plan of the entrance part of the Hadži-Prodanova cave................................. 18 Figure 7. Geographic position, section and plan of the Smolućka cave............................................................................. 19 Figure 8. Graphic representation of connecting the specimens through their taxonomic and element part identification with taphonomic traces on them................................................................................................................... 21 Figure 9. NISP of different mammalian taxa and size classes in Pešturina cave................................................................ 32 Figure 10. Reciprocal Simpsonʼs index (1/D) given as total between the layers, herbivores only, and carnivores only in Pešturina cave....................................................................................................................................... 32 Figure 11. Representation of habitat sensitive avian species in Pešturina cave................................................................. 33 Figure 12. The least area in km2 necessary to cover in order to encounter different carnivore (left) and herbivore (right) species from Pešturina, if area of 1 km2 is taken as a length of its diagonal, counted as a distance travelled, in relation to the population density of those species in given ecosystems obtained on the basis of their body mass and ecological preferences.................................................................................................................................................. 34 Figure 13. Minimum of elapsed time (min.) until the first encounter with different carnivore and herbivore species from Pešturina in different ecological niches, if moving by diagonal of 1 km2 (1410 m), with walking speed of 5 km/h................................................................................................................................................................... 35 Figure 14. Skeletal profiles of a wild horse, European ass, and remains that could not be determined beyond the horse genus (Equus sp.) in Pešturina cave.................................................................................................................... 36 Figure 15. Skeletal profiles of size III cervids (Cervus elaphus and Dama dama) from Pešturina.................................... 38 Figure 16. Skeletal profiles of bison, roe deer, ibex, and hare in layer 4 of Pešturina....................................................... 39 Figure 17. Skeletal profiles of the cave bear and hyena in layer 4 of Pešturina................................................................. 40 Figure 18. Average specimen length according to layer with variations............................................................................ 40 Figure 19. Shaft circumference preservation in percentages according to layers in Pešturina: a) size II mammals, b) size III mammals, c) size IV mammals.......................................................................................................................... 41 Figure 20. Representation of specimens from Mammalia indet. category from Pešturina, according to anatomical regions they belong............................................................................................................................................................. 42 Figure 21. Circumference of pits and bitten through carnassial holes (mm) according to layers and way the measures were taken........................................................................................................................................................... 43 Figure 22. Cross section of randomly selected butchery traces and the way the outline was measured............................ 43 Figure 23. Relationship between MAU and FUI for a wild horse and the horse genus together with distribution of processing marks and hyena tooth scores................................................................................................... 44 Figure 24. Relationship between MAU and BMD for: size II bovids, wild horses, size III cervids.................................. 46 Figure 25. Age structures of best represented carnivore and herbivore species in layer 4 of Pešturina............................. 47 vii

Milosevic.indd 7

02/01/2020 12:47

Competition Between Humans and Large Carnivores Figure 26. Mammal NISP, beside cave bear, from the Hadži-Prodanova cave. ................................................................ 49 Figure 27. Reciprocal Simpsonʼs index (1/D) given as total between the layers, herbivores only, and carnivores only.................................................................................................................................................................... 49 Figure 28. The least area in km2 necessary to cover in order to encounter different carnivore and herbivore species from Hadži-Prodanova........................................................................................................................................... 50 Figure 29. Skeletal profiles of ibex and wolves from the Hadži-Prodanova cave according to layers.............................. 51 Figure 30. Skeletal profiles of cave bears from Hadži-Prodanova according to layers...................................................... 52 Figure 31. Presence of different fragments size from the Hadži-Prodanova cave.............................................................. 53 Figure 32. Traces of carnivore scores on animal remains from Hadži-Prodanova............................................................. 55 Figure 33. Relationship between the percentage of MAU and FUI for ibex in different layers of Hadži-Prodanova....... 57 Figure 34. Age structures of ibex and cave bears in different layers of Hadži-Prodanova. Ibex structure shown with some modern examples of wolf predation structure. ................................................................................................. 57 Figure 35. Mammal NISP, beside cave bear, from the Smolućka cave.............................................................................. 58 Figure 36. Reciprocal Simpsonʼs index (1/D) given as total between the layers, herbivores only, and carnivores only.................................................................................................................................................................... 59 Figure 37. The least area in km2 necessary to cover in order to encounter different carnivore (left) and herbivore (right) species from Smolućka, if an area of 1 km2 is taken as a length of its diagonal, counted as a distance travelled, in relation to the population density of those species in given ecosystems obtained on the basis of their body mass and ecological preferences................................................................................................................................ 60 Figure 38. Minimum of elapsed time (min.) until the first encounter with different carnivore and herbivore species from Smolućka in different ecological niches, if moving by diagonal of 1 km2 (1410 m), with walking speed of 5 km/h................................................................................................................................................................... 61 Figure 39. Skeletal profiles of ibex and cave bears from the Smolućka cave according to layers..................................... 62 Figure 40. Representation of size II specimens and size III/IV specimens from Mammalia indet. category from Smolućka, according to anatomical regions they belong. ......................................................................................... 63 Figure 41. Shaft circumference preservation in percents according to layers in Smolućka............................................... 64 Figure 42. Relationship between the percentage of MAU and FUI for ibex from Smolućka............................................ 63 Figure 43. Age structures of ibex from layer 5 and cave bears from different layers of Smolućka. ................................. 65 Figure 44. Relationship between mammal NISP from the Last Interglacial deposits of Kozarnika cave and layer 4 of Pešturina....................................................................................................................................................... 70 Figure 45. Relationship between mammal NISP from OIS 4 context TDVI of Temnata dupka and layer 3 of Pešturina............................................................................................................................................................. 70 Figure 46. Relationship between mammal NISP from OIS 3 cave deposits...................................................................... 71 Figure 47. Combined dismemberment traces and tooth scores on Proboscidean astragalus from layer 4 in Pešturina..... 71 Figure 48. Hypothetical interrelationship of faunal species used in subsistence and time necessary to elapse in search for prey, processing, and transport........................................................................................................................... 78

viii

Milosevic.indd 8

02/01/2020 12:47

List of Tables Table 1. Published C14 and ESR dates for different layers of Pešturina.............................................................................. 5 Table 2. Published C14 dates for different layers of the Hadži-Prodanova cave, according to Alex et al., (2019)........... 18 Table 3. Taxonomic composition of mammals from different layers of Pešturina ranked according to size classes......... 29 Table 4. Taxonomic composition of avian remains from different layers of Pešturina (Zlatozar Boev, pers. comm.)...... 31 Table 5.1. Population density model for large carnivores appearing in the studied materials in different ecosystems...... 33 Table 5.2. Population density model for herbivores appearing in the studied materials in different ecosystems.............. 33 Table 6. Representation of long bone shaft fragment breakage types (according to Villa and Mahieu 1991) for herbivores and indet. mammals from different layers of Pešturina.................................................................................... 42 Table 7. Specimens burnt with different intensity from layers 4 and 3 of Pešturina, classified on the basis of visual thermodynamical changes that affected the specimen........................................................................................................ 42 Table 8. Number of specimens with mechanical-chemical damage in different layers of Pešturina................................. 45 Table 9. Taxonomic composition of mammals from different layers of the Hadži-Prodanova cave ranked according to size classes..................................................................................................................................................... 48 Table 10. Taxonomic composition of avian remains from different layers of the Hadži-Prodanova cave......................... 48 Table 11. Specimens with processing marks from layer 2 of the Hadži-Prodanova cave according to size class/taxa...... 49 Table 12. Specimens with processing marks from layers 4+5 of the Hadži-Prodanova cave according to size class/taxa............................................................................................................................................................................. 53 Table 13. Specimens with gnawing marks from different layers of the Hadži-Prodanova cave according to size class/taxa...................................................................................................................................................................... 54 Table 14. Number of specimens with mechanical-chemical damage in different layers of the Hadži-Prodanova cave.... 54 Table 15. Taxonomic composition of mammals (Dimitrijević 1991) from different layers of the Smolućka cave ranked according to size classes................................................................................................................................. 56 Table 16. Taxonomic composition of avian remains from different layers of the Smolućka cave, according to Malez, Dimitrijević 1990.................................................................................................................................................... 59 Table 17. Number of specimens with trampling and weathering marks in different layers of the Smolućka cave............ 65

ix

Milosevic.indd 9

02/01/2020 12:47

Foreword Contrary to most of the Palaeolithic projects done through significant side grants and direct PhD funds, this one is a real praise of good will. Fundings received per annum for the entire project „Cultural changes and population dynamics in the Palaeolithic and Mesolithic of the Central Balkans“, funded by Serbian Ministry of Education and Science (ON 177023) would not suffice to even pay a rent for archaeological team of 10 people to spend a month of archaeological excavations in most of the countries of European Union. This research is an ultimate praise to good will and scientific devotion to all people who were engaged in it. In the shaping of my scientific comprehension of Palaeolithic archaeology I owe the most to dr Dušan Mihailović, professor of Palaeolithic archaeology at Faculty of Philosophy at Belgrade University and leader of the project under which this research was conducted. His devotion to pursue research by all means improved Palaeolithic research in Serbia in last 15 years. Regarding specialistation in archaeozoology nothing of this would be possible without the engagement of my mentor during my MA and PhD studies, dr Vesna Dimitrijević, to whom I am indebted for transferring her knowledge, ideas, and discussion, but also for meeting me with fellow archaeozoologists, some of which greatly helped this study. One of the most important colleagues I received my training from is dr Ana Belen MarínArroyo from Cantabria University in Santander, and her instructions in taphonomic perspectives greatly influenced this study. I am greatly indebted to dr Zlatozar Boev from Natural History Museum in Sofia, Bulgaria, and dr Sheila Hamilton-Dyer from Bournemouth University for their immense help with avian remains from the Pešturina and Hadži-Prodanova caves. Good looking images in this work are thanks to John Meenely from Belfast Univeristy for his work on 3D scan at Pešturina, dr Nikola Vuković working as technician at SEM-EDS laboratory of Faculty of Mining and Geology, Belgrade University, and archaeologist Irina Kajtez for GIS data. I need not mention that this was done from their side to support me

and for the love of science, without a single coin taken. The shaping of the thesis, control of data and flow of work owes thanks to dr Sonja Bogdanović, dr Marko Porčić and dr Boban Tripković who were immediate members of my PhD defense committee, and dr Bridget Alex who read it independently as a researcher also engaged on this project. During the PhD I was engaged in research on other projects than the one resulting in this book. Some of experiences acquired there greatly influenced this study. For those experiences I would like to specially thank dr Aleksandar Kapuran, dr Aleksandar Bulatović, dr Dragan Milanović, dr Petar Milojević, dr Bridget Alex, dr Ana Majkić, dr Jelena Bulatović, dr Aitor Ruiz-Redondo, Gordan Janjić, dr Bojana Mihailović, Danica Mihailović, and dr Dušan Borić. I thank all colleagues from Laboratory of Bioarchaeology at Faculty of Philosophy, Belgrade University for all their help about uncertainties with the material. Lectorship and language support for the first edition of PhD profited immensely from suggestions by Danica Vukićević-Milošević and Jasmina Živković. I owe thanks to the BAR proofreader for this volume. The most important credits however must go to my immediate family, especially parents Olivera and Predrag Milošević. Their support throughout my entire life and the liberty I had in choices while being raised enbled me to find a life-long interest in archaeology. Not being angry with me in turning over family attics and drawers, and experimenting with daring, dangerous, and stupid childhood ideas undoubtedly influenced the choice of this profession. I credit my father for the cover of this book and beyond-imagination he had in capturing the point of this study in one illustration. Immense thanks to my fiancee Jelena for all her care and support during my long mindabsent hours at home while working to shape this study. Same goes for my closest and best friends, who were sometimes forced to listen about Palaeolithic perhaps against their will, I know they will recognize themselves in this sentence.

x

Milosevic.indd 10

02/01/2020 12:47

1 Introduction Since it studies the human past, archaeology strives to create a relevant image of past human societies. As we experience archaeology through different materials from the past, which are not directly related to the phenomenon we study, we build up interpretations of research results to access various aspects of past human lives, but limited to what we are able, or what we think we are able, to observe. For that reason it is crucial to observe past remains from different perspectives – societies, various aspects of natural environment and landscapes which they inhabited, and, from the point of individuals and various social constructs, so that we can understand and interpret them more comprehensively, while still respecting the interpretation limits imposed by the archaeological data and materials we study. This is especially complicated for the Palaeolithic period, because it is not possible to draw any direct analogies between Palaeolithic and contemporary hunter-gatherers, not only with respect to aspects of material culture, but also behavioural and cognitive execution of different life tasks. With the emergence of processual archaeology there was a positive way of looking into the analogies with contemporary huntergatherers, and various anthropological and archaeological data created from that period onwards helped in selecting criteria that are measurable and comparable. Nonetheless, the boundary between interpretation and generalizing the Palaeolithic into a universal narrative remains very thin when accessing these analogies with any human being and society in recent history and present-day world.

contact and long isolation between differen palaeodemes at certain times. According to the traditional archaeological periods it can be divided, in Europe, into the Late Middle Palaeolithic corresponding to the material culture of Neanderthals, and Upper Palaeolithic, corresponding to the material culture of modern humans which emerged as they spread out of Africa. The biggest interest, in terms of both physical and cultural anthropology, surrounds “the big transition”, the last known cohabitation of different forms of humans on Earth, which, in Europe, happened roughly between 45 and 30 kya BP and ended with the extinction of Neanderthals. Although there are many hypotheses about the possible causes for Neanderthal extinction – climate change, rigid subsistence and settlement strategies, weak ties between the individuals and societies leading to weaker technological exchanges and social learning, violence, diseases, etc. (D’Errico and Sánchez-Goñi 2003; Horan et al., 2005; Jiménez-Espejo et al., 2007; Banks et al., 2008; Underdown 2008), it is still inconclusive what happened over the course of the transitional period. Earlier interpretation that early modern humans (EMH) colonized Europe in sort of a “big transition”, by bringing in lamellar technology, complex worked bone technology, manufacture of personal ornaments and other forms of symbolic behaviour such as the appearance of rock art, is now contested (Mellars 1998a, 1999; Gamble 1999; Marean and Henshilwood 2003). Aurignacian, the first techno-cultural complex, which emerged in Europe or in the Near East, is primarily thought to have EMH origins, but various factors that led to its genesis have not yet been essentially understood. As a lithic industry, it is partly preceded by, and is partly contemporaneous with a number of different stone tool industries, so called “Transitional type industries”, which differ from the Mousterian Middle Palaeolithic by having a larger proportion of various Upper Palaeolithic techno-typological indices and an elevated lamellar index. These industries are highly variable spatially and temporally, and have been found in the Near East, lower course of Don river, southern European peninsulas (Iberian, Apenine, Balkans), Moravia, southwest and west-central France (Djindjan et al., 2003; Kuhn 2003; Otte and Kozłowski 2003; Svoboda 2003; Bon 2006). Pre 40 kya (kilo-year age) BP Aurignacian in Europe, or Protoaurignacian, is defined by the presence of carinated core technology, the production of twisted and Dufour bladelets, and split-base antler points with complex worked bone technology (Gaudzinski 1999; d’Errico et al., 2003, 2011; Soressi et al., 2013), and is contemporaneous with several Transitional industries: Châtelperonnian, Uluzzian, Bachokirian and Bohunician (Moncel and Voisin 2006; Zilhão and d’Errico 2006) (Fig. 1). It is still debated whether these industries represent a Neanderthal

1.1 Aims of the study This research studies human subsistence strategies and lifestyles tied with food procurement in various past ecosystems, and mechanisms that influenced them in the Last Interglacial and Glacial environment, within Late Middle and Upper Palaeolithic of the Central Balkans, primarily, but not exclusively, based upon archaeozoological analysis, within the broader context of the Balkans and European Palaeolithic. In order to understand the faunal record it is necessary to examine Neanderthals and modern humans more closely – their lifestyles in Europe from the beginning of Last Interglacial to the Last Glacial Maximum (MIS 5-2), or until the beginning of Epipalaeolithic. Corresponding to time range geological units from the sites presented here, Last Interglacial to the Last Glacial Maximum (approx. 130.000 – 22.000 years BP) are periods that witnessed the evolution of archaic humans into anatomically distinct modern humans in Africa and Neanderthals in Eurasia, among others. The term evolution refers to series of changes in human behaviour, and a greater degree of diversity in human skeletal morphology due to nature of diathonic 1

Milosevic.indd 1

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 1. Distribution of various Transitional type industries and Proto-Aurignacian in Europe (45-38 kya BP). Sites: 0. Üçağızlı, 1. Vindija, 2. Senftenberg, 3. Geißenklösterle, 4. Fumane, 5. Riparo Mocchi, 6. Grotta della Fabrica, 7. EsquichoGrapau, 8. Isturitz, 9. Gatzarria, 10. Abric Romani/L’Arbreda. Dotted routes represent paths of initial spread of EMH.

acculturation to certain Protoaurignacian elements, EMH traditions parallel with the Protoaurignacian, or an exclusively Neanderthal innovation, since it is now widely recognized that archaic humans were able to produce lamellar technology much before the transitional period (Bordes 2003; Chabai 2003; Svoboda 2003; Fernández et al., 2004; Kozłowski 2006; Rigaud and Lucas 2006; Tsanova 2012).

et al. 2003; Soficaru et al. 2006, 2007; Alexandrescu et al. 2010) in Romania, and Mladeč (31 kya BP) (Wild et al. 2005), Vogelherd (31 kya BP) (Churchill and Smith 2000) in Central Europe (Fig. 1). Except the Oase individuals, these EMH specimens, correspond to the period when the Aurignacian already became widespread, so they are not typical anthropological representatives of the initial EMH groups in Europe.

Timing the arrival of EMH in Europe is hard to define, both on the basis of archaeological material and anthropological finds. It is still not possible to define a material culture which can be unequivocally attributed to the Neanderthals or EMH during the initial period of their arrival at some point between 50-45 kya BP. Moreover, because of the high skeletal diversity of both types of humans at that time, differences in skeletal morphology can only be studied on better preserved skeletal elements, making highly fragmented remains difficult to discern between late Neanderthals and EMH without a welldefined context (Harvati et al. 2004). The oldest EMH anthropological remains in Europe have been recovered at the sites of Peştera cu Oase (40 kya BP), Peştera Muierii (30 kya BP), Peştera Cioclovina (29 kya BP) (Trinkhaus

Already towards the end of Transitional industries and when Aurignacian became widespread, we witness appearance of the Gravettian, the genesis and origins of which are still being sought in Eurasia. It appears over the vast area from Europe to Central and Northern Asia. What distinguishes the Gravettian from other Palaeolithic periods is the fact that for the first time humans continually settled all climatic belts and landscapes, except the parts covered by the ice sheets. The origin and rapid spread of Gravettian over a vast territory of Eurasia remains an unsufficiently understood phenomenon (Kozłowski 2014). According to traditional Perigordian periodization, it was assumed that origin of Gravettian is local and based on cultural traits which evolved from the Aurignacian, but accumulated data suggest that the the Gravettian was 2

Milosevic.indd 2

02/01/2020 12:47

Introduction more likely part of a second, maybe larger, migration of EMH (Svoboda 2004). Early Gravettian (30-25 kya BP) appeared over the area of Central and Eastern Europe and could coincide with appearance of variant I of Y chromosome halpogroup in European populations, the appearance of which is estimated at 24±7,1 kya BP (Rootsi et al., 2004). Results of MtDNA studies also presented a significant mix of local European and new gene alleles at 20 kya BP (Pala et al., 2012).

consequence of the emergence and spread of their common ancestor Homo erectus out of Africa, followed by long periods of isolation between the populations. The emergence of skeletal morphologies typical for both human forms is subject to debate. According to palaeoanthropological data acquired so far, there is now growing evidence that they are the consequence of multiregional local evolution of the common ancestor (Wolpoff et al., 2000), but with occasional gene flow between some of these populations (de Castro and Martignón-Torres 2013). PreNeanderthal skeletal forms are defined on the basis of palaeoanthropological material recovered at the sites of Sima de Los Huesos, Arago-Tautavel, Petralona, Apidima, Steinheim, and Mauer, to which the Mala Balanica hominin from Central Balkans temporally corresponds, while the archaic forms of anatomically modern humans are defined on the basis of the remains recovered at the sites of Kabwe, Bodo, Omo 2, Ndutu and Elandsfontein (Rightmire 2007, 2008). These sites roughly fall into the period between 600 and 300 kya BP, and human remains recovered at them clearly show derived morphology of Homo erectus, and are classified globally as Homo heidelbergensis. The assumption is strengthened when the individuals that lived temporarily close, but spatially very distant from one another are compared, such is, for example, a striking geomorphometric similarity between Kabwe and Petralona skulls (Van Vark 1995; Rightmire 2008; Harvati 2009).

Local origins are also less probable because Gravettian chipped stone technology, based on blade production from narrow faced core, is temporarily overlapping with Aurignacian chipped stone assemblages. Gravettian chipped stone industry is most similar to Levantine Ahmarian (Skrdla 1997), in which the production sequence is charaterized by unipolar exploitation of blades from narrow faced core. Some authors are of the opinion that chipped stone industries appearing in Eastern Europe between 34-31 kya BP, notably in layer VII of Kozarnika cave and layer 3g of Temnata dupka in Bulgaria, layer III of Buran Kaya on the Crimean peninsula, and Kostienki 8 in the Don river valley contain these basic elements in chipped stone technology (Tsanova 2006; Prat et al., 2011) that define Gravettian. Others (Pessese 2008, 2010) see a link between final Aurignacian and initial Gravettian, but also agree that it is perhaps too simplistic to readily assign some of final Aurignacian techno-types, such as Font-Yves bladelets and points, to either Aurgnacian or Gravettian.

Besides skeletal morphology, the multiregional evolution model is supported by DNA studies. According to the molecular clock and ancient DNA analyses combined, it is possible to establish that the split between African and European human lineages happened between 700 and 500 kya BP (Prüfer et al., 2014, 2017). From that time and the next large-scale contact between these two populations around 100 kya BP, ones that lived out of Africa acquired around 3% of new gene alleles through gene flow and mutation (Green et al., 2006: 334). The point of these data is not to show that there was a considerable genetic difference between African and outside African populations, but that rather occasional gene flow happened within both of them, which led to the existence of the variety of temporal and spatial allotaxons, or palaeodemic variations. This is best seen on the example of ancient DNA fragments obtained from Denisova hominins, an archaic human population that lived in Central Asia 130-40 kyA BP, showing that it diverged from the population of Homo erectus even earlier than European population, between 1000 and 500 kya BP (Krause et al., 2010; Sawyer et al., 2015). Comparison of the ancient DNA fragments with human genomes worldwide made a big step in understanding of these variations. It is established that the maximum DNA difference, at the level of alleles, between Neanderthals and modern humans is always between zero and three per cent, and that the difference is always bigger outside of Africa (up to 3.7%) than within it (up to 1.7%). Such a result points that, at the level of DNA alleles, Neandertals are more closely related to contemporary Eurasian than African populations (Green et al., 2010: 713). Finally,

Research questions of this study are aligned in several different, but contextually integrated paths divided between the Late Middle Palaeolithic and Gravettian, corresponding to results of paleoecological, taxonomic, taphonomic, and spatial analyses of different levels at the studied sites: • Did humans and large carnivors target the same prey? What is the difference in ecological patch choice in the Central Balkans between Neanderthals, modern humans, and carnivores when targeting large herbivores? What is the level of biodiversity across MIS 5-2 span and availability of herbivores to Neanderthals, modern humans, and large carnivores within the same site? • What is the difference in skeletal parts representation and processing strategies of humans and carnivores? Is it possible to identify and discern activity zones of humans and carnivores on the basis of spatial distributions of different sets of finds? • Was there a competition between humans and carnivores for large herbivores preferring different ecological patches, or intra-site competition? 1.2 Chronological and biological framing of Upper Pleistocene humans, and their socio-economic behaviour Separation of two distinctive forms of Homo sapiens sub-species – Neanderthals and modern humans, is a 3

Milosevic.indd 3

02/01/2020 12:47

Competition Between Humans and Large Carnivores the analysis of ancient DNA recovered from the oldest known skeletal remains of modern humans in Europe, the Oase 1 individual from Peştera cu Oase, showed that it holds 10% of DNA allele similarity with the Neanderthals, more than twice the value of contemporary populations, which strongly suggests that Oase 1 had a Neanderthal ancestor not more than four to six generations before (Fu et al., 2015). This discovery seals the debate about the Neandertals being regarded as a different species from modern humans, as the oldest known modern human in Europe had itself Neanderthal admixture (Harvati and Roksandic 2016). In terms of biological similarity, their genomes were compatible and could produce fertile hybrids that could carry them on, at least in some instances as shown in Oase 1 individual.

subsistence shows that they hunted large mammals in open, lowland habitats (such as horses, bison, reindeer), while often exploiting medium-sized mammals in mountainous habitats (deer, ibex and chamois), as they were adaptive to the local game availability within an ecosystem. Subsistence is rarely tied to one species, but in those cases it can be more readily defined as animal foraging of the most productive species from the set of available habitats comprising the ecosystem (Stiner 1991, 1992, 2004b; Boyle 2000; Hoffecker and Cleghorn 2000; Mussi 2001: 152–154; Fiore et al., 2004; Valensi and Psathi 2004; Miracle 2007; Daujeard and Moncel 2010; Daujeard et al., 2012). Studies that synthetize Neanderthal subsistence in different regions always emphasize the fact that large game hunting is nearly always focused on adult animals, which is confirmed at least at 31 Middle Palaeolithic sites from La Cotte de Saint Brelade in Bretagne up to Teshik Tash and Aman Kutan in Uzbekistan (Gaudzinski 2006). These sites dominated by a single species are most probably just one of the faunal exploitation strategies within a settlement strategy in a wider area, rather than hunting specialization. Such assumption is strengthened by the fact these locations were not visited repetitively and seasonally and represent remains of one or several random herd hunting events, based on lithic assemblages encountered there. In other words, these sites were stations where prey was mass hunted, butchered, and brought out (Fairzy et al., 1994; Stiner 1994; Boyle 2000: 343; Gaudzinski 2000, 2004). Recent studies have shown that, besides being well-organized foragers (Adler and Bar-Oz 2009), Neanderthals were also top ecological opportunists, since their subsistence included small mammals, such are rabbit/hare (Cochard et al., 2012; Fa et al., 2013), birds (Blasco and Fernandez 2009) tortoises (Stiner et al., 2000; Blasco 2008; Starkovich 2012), marine mammals and shellfish (Cortés Sánchez et al., 2011; Stringer et al., 2008), as well as a narrow variety of plants (Henry et al., 2010). Stable isotope analyses, which reveal the origin of proteins in diet on one hand, and position in the food chain on the other, suggests that Neanderthals were predators at the top of the food chain, so that almost all of their diet was based upon protein originating from terrestrial mammals (Drucker and Bocherens 2004; Richards and Trinkhaus 2009; Dobrovolskaya and Tiunov 2011).

In comparison to the earlier periods of the Pleistocene, the Upper Pleistocene is characterized by abrupt and frequent climate changes on a global scale with two peaks – the Last Interglacial or marine isotopic stage (MIS) 5e around 130 kya BP, and the Last Glacial Maximum at the boundary of MIS 3 and 2 around 22 kya BP. The Last Interglacial, encompassing MIS 5e-5a, and the Last Glacial, encompassing MIS 4, 3 and 2, do not represent the periods of constant warm or cold climate. They both have shorter warmer and colder oscillations, with shorter cold oscillations during the interglacial climate and shorter warm oscillations during the glacial climate (Boroughs 2005; Sánchez Goni 2007). Futhermore, climatic conditions in Europe should not be assessed in general since they differ between Western Europe, where climate is influenced by the Atlantic Ocean, Northern Europe, where it is influenced by Scandinavian and Alpine ice sheets, Eastern Europe, with climate influenced by the “mammoth steppe” – the largest continuous steppe that existed on Earth in the Upper Pleistocene, and southern European peninsulas with climate influenced by the Mediterranean (Kukla et al., 2002; van Kolfschoten 2002). Because of that, expansion and contraction of various ecosystem types happened at different paces within given European regions. According to distribution of Late Middle Palaeolithic sites in Europe it has been established that Neanderthals preferred to settle in mosaic ecosystems – that is to have availability to choose between diferent ecological patches both in mountainous and lowland landscapes, in order to have better access to various economic resources (Wenzel 2007). As such, Neanderthals had their own histories of population bottlenecks and isolations in Western Europe, Mediterranean and sub-Mediterranean Europe through glacial periods, population expansions to Central and Eastern Europe (and the Middle East/Central Asia) and migrations, as ecolological patchiness increased or decreased between these regions (Pathou-Matis 2000; Dennell et al., 2011). This means that the cold climate would not necessarily lead to a demographic decline of European Neanderthals, but rather to their redistribution with higher concentration in the regions of Europe with milder climate conditions that enable a wider choice of ecological patches. Archaeological data about Neanderthal

Several independent lines of evidence from various sites, such as occupation stress and trauma markers at Krapina and Moula Gercy, stone tool technology (universal over the vast area of Eurasia) and wooden spears (Lehringen), indicate that Neanderthals hunted large mammals from close quarters. At the sites of Krapina and Moula Gercy, which contained the largest numbers of Neanderthal individuals found at one place and as such present the largest available samples, the outline of humerus cortical bone at 35% of its length from the distal end shows that the diaphysis was remodelled mainly by the pressure of flexor muscles (Churchill et al., 1996). Antero-posterior depth of trochlea humeri in relation to longitudinal humeral axis, which marks the maximum extension angle of the forearm (Hambücken 2012), as well as the 4

Milosevic.indd 4

02/01/2020 12:47

Introduction optimal angle of reaction force transferred from the forearm to humerus, is positioned at an average optimal angle of 101º. Schmitt et al., 2003: 104, Table 1, shown that this is the optimal position of the arm for reaction force transfer when thrusting a spear. The asymmetry of these parameters between left and right humerus is around 16.5%, which is generally low and also suggests a thrusting arm movement. Stress markers characteristic of throwing would be opposite to these, because throwing engages arm extensor muscles which cause lateral torsion of the distal end, and the alignment of trochlea axis with the longitudinal axis of humerus (observed in lateral norm), since the longer movement enhances the javelin speed at the moment of release and gives higher propulsion force to the projectile (Rhodes and Churchill 2009). In the period of the initial human settlement in Europe, during Aurignacian, we do not encounter any larger sample of complete human remains. But, remains of hunting technology encountered at Aurignacian sites such as antler split base points (Tartar and White 2013; Tejero 2016) might suggest the appearance of projectile technology. Gravettian populations in Europe, represented by a larger sample of well-preserved human remains (such as Sungir for e.g. Trinkhaus et al., 2014), show twice more stress marker asymmetry between left and right humerus (around 31.7%), but the evidence for thrown weighted atlatls, javelins and harpoons made of bone and antler is also abundant (Goutas 2016).

can be reached by closer examination of their hunting tool kit. Neanderthal chipped stone spear points were of standardized production, designed and conceptualized to be curated and resharpened several times before refusal, which is also a sign of a well-structured behaviour tied to the manufacture and use of hunting tool kit (Lazuen 2012). Techniques of tipping the spears and other chipped stone tools with adhesives such as bitumen (Böeda et al. 1996; Koller et al., 2001; Cărcuimaru et al., 2012) and tar obtained from birch bark by oxygen reduction (Mazza et al., 2006; Pawlik and Thissen 2011) are observed through the chemical analysis of their residues on the artefacts themselves, but were not implied to be used as collated lithic segments for a potential sort of projectile. In rare cases, remains of wooden hunting equipment have been found. At the site of Lehringen in NE Germany, dated to the Last Interglacial, a wooden spear split in several pieces but preserved in almost its entire length has been recovered, with an estimated length around 2,20-2,40 m, and was found together with elephant bones (Schmitt et al., 2003). The considerable length also means that Neanderthals could keep themselves at a safer distance when hunting large mammals from close quarters (Ruff et al., 1997). Direct evidence of Neanderthal close quarters hunting is also sometimes observed on fauna. A mesial fragment of a Levallois point was found embedded in a cervical vertebra of a wild donkey at the site of Umm el Tlel in Syria (Böeda et al., 1999), this suggests that Neandertals had exquisite knowledge of critical points when hunting, and aimed to bring their prey with one fatal blow, since wounded prey could easily flee.

Traces of physical trauma on more complete Neanderthal skeletons show that almost every individual had at least one sprained joint or a broken bone, and that most of these injuries affected hands, feet, head and neck (Berger and Trinkhaus 1995). These accumulated traumas on Neanderthals skeletons which point to repetitive bruises caused by running and falling under a heavy physical activity or on rough terrain add to the hypothesis of close quarters hunting, as hunting injuries may have been perceived as routine (Pettitt 2000). A similar conclusion

Data represented by archaeozoological research, stress markers on Neanderthal skeletons, and crafting technology and use of chipped stone implements reveals a well-structured behaviour, but also tremendous versatility concerning hunting activities. This pattern implies manufacture of specific tool kits and applying different

Table 1. Published C14 and ESR dates for different layers of Pešturina. layer

laboratory sign

RTD7148

/

RTK6446

ESR SD

LU γ

SD

16271±58

/

/

/

/

/

30888±622

/

/

/

/

RTD7231

AT23

33129±176

34300

± 1900

38900

± 2500

RTK6449

/

47608±3597

/

/

/

/

layer 3/4

RTK6450

/

42921±2171

/

/

/

/

layer 4

RTD7149

AT22

44599±591

86100

± 4500

92900

± 5200

/

AT32

/

95200

± 3500

101900

± 3800

layer 3

ESR

C14 years BP

EU γ

layer 2

C14

dates

C14 cal. BP ultrafiltration at 68% confidence; SD – standard deviation; EU – early uptake, LU – linear uptake. C14 dates from Alex and Boaretto (2014), Alex et al., (2019), ESR dates from Blackwell et al., (2014).

5

Milosevic.indd 5

02/01/2020 12:47

Competition Between Humans and Large Carnivores tactics and handling of prey, depending on animal species, surrounding landscapes and ecosystems, as they were able to hunt ungulates of quite different sizes, agilities, aggressiveness, living solitary or in herd, and in different ecological settings.

1996, 1998b, 2004). As shown on the example of southwestern France and northern Spain, EMH specialized in reindeer hunting, the remains of which are usually the most numerous at Proto- and Early Aurignacian sites. They specialized not only by species selection but also through an original hunting gear (Teyssandier 2002; Liolios 2006; Soulier 2013). Holding that true, it should be stated that increase in reindeer subsistence has also been observed in late Neanderthals and was already a dominant prey type before, even during MIS 4 at some of the sites in Western Europe such as Combe Grenal and Jonzac (Chase 1989; Niven et al., 2012). More recently, it was demonstrated that species selection at early EMH sites in the region was mostly due to ecologial conditions (Grayson et al., 2001; Grayson and Delpech 2002, 2006). Besides, butchery and transport strategies were not essentially different between Neanderthals and EMH (Chase 1989; Stiner 1994; Burke 2000; Costamagno et al., 2006; Grayson and Delpech 2008; Niven et al., 2012) not only in SW Europe but elsewhere, and use of small prey by Neanderthals has also been observed (Blasco and Fernández Peris 2009, 2012; Blasco et al., 2016). Contrary to Mellars’ theory, it is now accepted that main response of EMH in ecological competition was not species specialization, but broadening of diet (O’Connel 2006; Lloveras et al., 2016) to more often include species from a wider array of ecological sets than the Neanderthals.

Although there is no evidence that Neanderthals made early art parallel to Chauvet, Altxerri B, Coliboaia, and bone and ivory figurines found at the sites of Swabian Jura (Conrad and Bolus 2006; Clottes et al., 2012; González-Sainz et al., 2013), they were able to express symbollic behaviour through the use of pendants and colourants much earlier than the Transitional period. Mineral processing to obtain colourants is evidenced already in early Neanderthals, as observed at the site of Bečov in Moravia (MIS 7), where porcelanite chunks, fired to obtain different colours, were found beside a quern with traces of the pigment both on the grindstone and the quern itself (Šajnerová et al., 2006). Two caves in eastern Spain – cova Aviones and cova Anton contained shell caps perforated in the umbo region, while one contained traces of pigment and solder, and the layers in which they were found are older than 45kya BP (Zilhão et al., 2009). Neanderthal personal ornaments from the transitional period are rich in their variety of forms, although they appear as such only in regions where the presence of transitional type industries is strong: Châtelperronian, Bachokirian, Uluzzuan (d’Errico et al., 1998; Arnaud et al., 2016; Fabbri et al., 2016), so that some of the authors also contest their Neanderthal origin (Caron et al., 2011). Besides, it has been recently observed that Neanderthals butchered birds for feathers, probably used to garment or decorate themselves (Peresani et al., 2011; Finlayson et al., 2012).

In comparison to the Aurignacian, larger differences are observed in subsistence and settlement strategies in the Early Gravettian, since it is, above all, characterized by large mammoth bone hut settlements at the bands of large rivers – Morava, Danube, and Don. They subsist on large mammals living in herds: before all mammoths, but also horses, bison and reindeer (Svoboda et al., 2005; Bosch 2012; Nikolskiy and Pitulko 2013; Brugère 2014), and hunt small carnivores for fur (Wojtal et al., 2012). Existence of composite chipped stone armatures which minimize damage to the fur when hunting/trapping is also evidenced, and bone eyeneedles point to the existance of tailored clothes (Brühl 2005), which would have been thermodynamically much more efficient than interconnected patches of fur (Gilligan 2007). Gravettian societies traded raw materials and shells over longer distances than in previous periods, and show cohesion in ritual and burial behaviours (HenryGambier 2008; Trinkhaus et al. 2014). Some of these open-air sites were settled year round (Fišáková 2013). The use of symbollic objects in the Early Gravettian is quite complex, with a great variety of forms, from personal ornaments to antropomorphic and zoomorphic figurines carved in bone, teeth, ivory, stone, and for the first time from baked clay (Svoboda 1995; Trinkhaus and Svoboda 2006). Late Gravettian, after 25 kya BP, is characterized by a certain degree of cultural and population dispersal towards valleys between the Dniepr and Don rivers, which represents the period of the so called Eastern Gravettian (Sinitsyn 2007), with culmination in numbers and size of open air settlements such as Kostienki and Avdeevo. In other parts of Europe, in the Mediterranean and SubMediterranean climatic belt, cave and rockshelter sites

The timing of EMH arrival in Europe and demise of Neanderthals also coincides with climatic deterioration. Although mild glacial climatic conditions prevailed in MIS 3, glacial maximums, or Heinrich events, were too brief to severely change biotopes in Europe. However, between 46 and 38 kya BP, Heinrich events 5 and 4 occured temporarily close, and led to a much drier climate (Boroughs 2005; Sánchez Goni 2007). A similar, yet more pronounced climatic occurence happened with Heinrich events 3 and 2, leading to the LGM. Thus, Heinrich events 5 and 4 must had an impact on the demise of distribution of animal and human populations in Europe during the transitional period. Modelling demographic advancement of EMH in Europe and the replacement of Neanderthals predicts Neanderthal population compression and an increase in population density in milder regions of Europe, rather than population decline, which might have enabled EMH to settle in. Even if EMH had greater demographic rise, it would not suffice, without Neanderthal population admixiture itself, to demographically outcompete them (Morin 2008; Excoffier and Currat 2011; French 2011; Mellars and Dogandžić, McPherron 2013). It was earlier stated that the Middle to Upper Palaeolithic transition was marked by a shift in human subsistence from opportunistic to specialized hunting (Mellars 1992, 6

Milosevic.indd 6

02/01/2020 12:47

Introduction are more numerous, with subsistence orientated towards large mammals represented in milder climate and mixed steppe-forest habitats, notably bison, horses, red deer, and ibex, and small mammals and birds as well (Boscato 2007; Starkovich 2012; Stiner et et al., 2012; Tagliacozzo et al., 2012; Vacca et al., 2012; Starkovich 2017), but also small mammals (Condrad et al. 2013). Chipped stone (steeply gravette retouched bladelets, blades and their segments) and osseous technology is used for production of composite and projectile implements. Rare remains of organic matter document the existence of nets and traps (Adovasio et al., 1996) used in fishing and fowling (Bochenski et al., 2009). Gravettian was the start of the ultimate adaptation of human hunter-gathereres to every environment type of Ice Age of Europe. Their societies are socially, culturally and technologically highly organized and geared for settling different ecosystems year-round. The settlement pattern is mainly residential-logistic or residential-cyclic, which is observed, beside the existence of year-round settled locations, on the numerous sites showing pronounced seasonality of human subsistence activities at some locations and regions.

Such a geographic position made this area relatively open to climatic influences from the north and east of Europe along the Sava, Danube, and upper course of Morava river basins, but relatively closed to the climatic influences of sub-Mediterranean and Atlantic climates, and separated from Mediterranean and Atlantic with high mountain chains with glaciers: the Dinaric Alps, Prokletije, and Šar-planina. Contemporary climatic factors on the Central Balkans are complex, but are characterized by mild continental climate type, and continental climate at the heights above 1000 m a.s.l., while its easternmost lowland parts are under the influence of a sub-tropical climatic type from the Black Sea shores (Savić and Obuljen 1979). However, it is not possible to discuss what key climatic factors were represented through the Upper Pleistocene in the Central Balkans, largely because it lacks such a holistic and interdisciplinary study. For that reason, it is unknown at what pace biomass changed in regard to a warm/cold and wet/dry climate. Although it contains lowland corridors along larger rivers, with altitudes below 200 m, the Central Balkans is mostly a hilly-montane region with altitudes between 300-500 m, but as high as 2.600 m. Considering the relief and human settlement, it is worth asking what was the physical boundary in altitude as well. In the southern part of the Central Balkans, Palaeolithic sites were discovered at higher altitudes, such as Crvena Stijena (700 m), and Smolućka pećina (945 m), and it is possible that smaller mountain glaciers never dropped below 700 to 1000 m during the Upper Pleistocene, and even during the LGM (Đurović 2012). Variations in oxygen isotope ratios, oxygen isotope in ostracod shells from bental sediments of lake Ohrid in North Macedonia show that during the last 140 kya the climate was subtropical-Mediterranean and corresponds to oxygen isotope ratios from ostracods in Ioannina (Epirus, Greece), Monticchio (Campania, Italy), Corchia (Liguria, Italy), and core MD01-2444 in the Atlantic Ocean, approximately 100 km from the shores of Alentego (Portugal) (Belmecheri et al., 2012). Across the east-west axis, the Central Balkans is less accessible, with fewer natural communication routes. It is important to establish whether there were connections between lowland ecosystems as well on this axis, between the river basins of Timok and the Nišava, and Velika and Zapadna Morava rivers, which would have made possible the movement of large migrating herbiovore species between them.

There are the first indications of humans intentionally changing the landscape such as killing off large predators in resident territories to lessen their pressure on the common ungulate prey, competition for shelter, fur, or even food (Kitagawa et al., 2012; Bocherens et al., 2014; Demay et al., 2015; Wilczyński et al., 2015; Wojtal et al., 2015; 2018). Various human deposited contexts containing almost exclusively carnivore remains, and remains of carnivores containing cutmarks are more numerous than in previous periods. The evidence of possible early animal taming (not to be confused with domestication process) comes from the Early Gravettian site of Předmosti, where geometric morphometric studies of canine skulls has shown, although on a small sample, reduction of the teeth row lenghth, and difference in skull morphology between wolves and wolf-dog hybrids, which were, besides, uncovered in a „domestic space“ or inside of the dwelling structures (Germonpré et al., 2012; Germonpré et al., 2015). This suggests that if interaction of humans and wolves existed, wolves and bastard-dogs were not kept separate by humans in order to domesticate them. 1.3 Spatial framework of the study

1.4 History of research and archaeological data on the Palaeolithic in Serbia

The area of this study is the Central Balkans, the geographical boundaries of which are to the north Kupa, Sava, and Danube rivers, and Mediterranean and Black Sea from other cardinal points. The Central Balkans covers about one sixth of the Balkan peninsula. It is defined mostly by larger geographic barriers – river valleys of the Sava and Danube to the north, Drina river basin to the west, Rhodopian and Balkan mountain chains to the east, and the Dinaric and Šar-planina mountains to the south, south-west (fig. 2). The main feature of the Central Balkans, in terms of physical geography, is the existence of series of mostly lowland basins, but sometimes highland plateaus, separated by canyons, gorges, or mountains.

Systematic research of the Palaeolithic period in Serbia has a long tradition, but is characterized by large pauses in exploration. A brief review is presented here, since detailed history of Palaeolithic research in Serbia and Central Balkans was given already in several publications (Mihailović 2009; Mihailović et al., 2011; Mihailović 2014). The earliest Palaeolithic research conducted was the field and test trench surveys of Đoka Jovanović (Јовановић 1892, 1893) and Jovan Cvijić, who first excavated in the Prekonoška cave near Svrljig according to methodological standards at that time and published the 7

Milosevic.indd 7

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 2. Late Middle and Upper Palaeolithic sites on the Balkans studied in detail, and some sites in surrounding regions which are of importance for this study: 1. Temnata dupka, 2. Bacho Kiro, 3. Kozarnika, 4. Baranica, 5. Pešturina, 6. Tabula Traiana cave, 7. Peştera cu Oase, 8. Risovača, 9. Golema pesht, 10. Kastritza, 11. Asprochaliko, 12. Theopetra, 13. Klisooura, 14. Franchti, 15. Lakonis, 16. Bioče, 17. Crvena stijena, 18. Hadži-Prodanova cave, 19. Šalitrena cave, 20. Kadar, 21. Mujina cave, 22. Šandalja, 23. Vindija, 24. Grotta de Cavallo. Regression line shows gradual replacement and demise of the Neanderthal societies in Balkans between 38-33 kya BP.

results (Цвијић 1891). The first systematic, archaeological and paleontological excavations were conducted in the 1950s at the cave of Risovača near Aranđelovac (Гавела 1988). These were followed by excavations in the 1980s – in the Smolućka cave near Novi Pazar (Калуђеровић 1985), At an open air Aurignacian site near Vršac (Radovanović 1986), Pećurski kamen cave near Sokobanja (Malez and Salković 1988), and systematic surveys in gorges of Resava and Suvaja rivers (Вучковић 1985; Ђуричић 1990). During the first half of the 1990s, several surveys with test trenches were conducted: the Prekonoška cave near Svrljig (Калуђеровић 1992), the Markova and Pećurski kamen caves near Sokobanja (Kaluđerović

1993), the Baranica cave near Knjaževac (Сладић and Јовановић 1995; Михаиловић et al., 1997), Mirilovska cave near Paraćin (Ђуричић 1996), Drenaićka cave near Valjevo (Калуђеровић and Јеж 1996), and Kremenac, a flint raw material source near Niš (Калуђеровић 1996). In the last decade, systematic archaeological excavations of Palaeolithic sites occurred at the Hadži-Prodanova cave near Ivanjica (Михаиловић and Михаиловић 2006), the Šalitrena cave near Mionica (Mihailović 2008), Velika and Mala Balanica in the Sićevo gorges (Михаиловић 2009; Roksandic et al., 2011) and the Pešturina cave in Jelašnica near Niš (Михаиловић and Милошевић 2013). Sites studied in more detail are presented in fig. 3. Several 8

Milosevic.indd 8

02/01/2020 12:47

Introduction surveys and systematic excavations of Palaeolithic sites are currently underway in eastern Serbia and the Morava river basin.

European (Fig. 4) and Near Eastern regions may suggest population continuity and continuity of techno-cultural ideas, or population isolation.

A hominin mandible dated to around 500 kya BP recently discovered at the Mala Balanica cave represents the earliest evidence of archaic human settlement in the Central Balkans and, as already said, belongs to a small group of human fossils from pre-Neanderthal period in Europe. Some of them are characterized by derived skull morphology features that are absent in so called „neanderthalization“ (gradually acquiring typical Neanderthal skull landmark features). The mandible of this individual does not posses typical anatomical landmarks for pre-Neanderthals (Homo heidelbergensis sensu lato in Western/Central Europe) (Roksandic et al., 2011; Roksandic 2016), similar to more recent interpretation of Apidima 1 cranium from Greece (Harvati et al., 2019). This opens the question of timing of that process across various palaeo populations in Europe, and might point to multiple dispersals of modern humans from Africa at least in this part of Europe (Ivanova 2016; Spasov 2016).

The beginning of the Middle to Upper Palaeolithic transitional period in the Balkans temporally corresponds to a natural catastrophe – a volcanic eruption which happened in Campi Flegreii in Campania, Italy. Layers of volcanic ash and pyroclastic ignimbrites were identified in stratigraphic sequences at Crvena stijena, Franchthi, Klissoura, Golema pešt, the cave above Tabula Traiana and Temnata dupka (Lowe et al., 2012), and provide an indication of the scale of this eruption. Such a natural catastrophe could be one of the triggers influencing the demise of Neandethal populations in the Central Balkans, but also one of the reasons why we currently lack Early Upper Palaeolithic sites in this region, as well as in western Adriatic (Hoffecker et al., 2008; Fitzsimmons et al., 2013). The evidence for Early Upper Palaeolithic settlement in the Central Balkans is sparse, and documented just with several chipped stone artefacts in the cave above Tabula Traiana (34 kya BP, Borić et al., 2012) and Baranica (36 kya BP, Mihailović et al., 2011), which is relatively late compared to dates of EUP layers at the sites of Bačo Kiro and Klissoura, and human remains from Peştera cu Oase, in surrounding areas. The context and technology of Late Aurignacian finds from At may be connected with openair Aurignacian sites in Romania – Tinçova, Koşava, and Romaneşti-Dumbraviţa (Mihailović et al., 2011: 94), and probably to Šalitrena cave in western Serbia, where it is dated to around 32 kya BP, while the Middle Palaeolithic layer there, dated to 38 kya BP, has among the latest Neanderthal occupation ages in the Central Balkans (Mihailović and Mihailović 2012). The apparent lack of the Early Upper Palaeolithic seems to indicate that a large part of the Balkans, from the Mujina cave (Rink et al., 2002) on the Adriatic to the west, to the lower Morava river basin to the east, and Thessaly to the south, was not inhabited by EMH between 41-39 and 34-28 kya BP (fig. 2), as indicated by Alex et al., (2019). Contrary to the north-western parts of the Balkans and Danube corridor, which withnessed Early Aurignacian and Transitional industries periods, this area starts to be colonized by modern humans in the Early Gravettian, as observed at the sites of the Šalitrena cave, Kozarnika, and Temnata dupka (Mihailović 2008; Tsanova et al., 2012). Although it is apparent that Gravettian assemblages from the Danube corridor belong to the same technological group as the Central European Gravettian, the nature of Gravettian in the interior of the Balkans remains unresolved. The most recently published research on the Central Balkans Paleolithic will be presented in the discussion of results, especially in context of this study.

Mihailović gives, on multiple occasions, comprehensive studies of technological variability, similarities, and problems in connecting Middle and Upper Palaeolithic chipped stone industries of the Central Balkans with other parts of the Balkans and Europe (Mihailović 2009; Mihailović et al., 2001; Mihailović 2014; Mihailović and Bogićević 2016), so only the most important features of it will be presented here. From the techno-cultural point of view, the Central Balkans represents one of the areas which witnessed an early appearance of Charentian type Mousterian (Михаиловић 2009), and according to finds from the sites of Velika Balanica and Crvena stijena in Montenegro, they can be related to Protocharentian assemblage from the site of Karain in Anadolia, which came from layers dated to 330-280 kya BP (Kozłowski 2002), with such assemblages also appearing further in the Near East. There is a large temporal gap in Charentian technological tradition between the Balkans and the Near East, and Western and Mediterranean Europe where it appears around MIS 5/4 boundary (around 90-80 kya BP). On the other hand, the technological tradition in the Mediterranean zone of the Balkans is different, and is based on cores and flakes produced from smaller cores, or perhaps nodules, and intensive exploitation of them, which in the pre-refusal stage is done with centripetal knapping. This technology, named Micromousterian, appears at the sites of Bioče in Montenegro, Asprochaliko, Teopethra, and Klissoura in Greece, and the Mujina cave in Croatia (Dogandžić and Đuričić 2017; Karavanić and Bilich-Kamenjarin 1997; Darlas 2007; Mihailović 2014; Vujević et al., 2017), but tools of small size appear as well in Central Italy (Kuhn 1995). The authors who studied this technological phenomenon argued that it has to do with high moblity of late Neanderthal societieties within these regions, which prevented them from visiting the best raw material sources in the region frequently. Technological variability between the Central Balkans and other

The analysis of animal remains was an integral part of archaeological research from the first systematic Palaeolitic excavations in Serbia, at the cave of Risovača (Rakovec 1965). During the 1980s, the analysis of 9

Milosevic.indd 9

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 3. Map of systematically explored Palaeolithic sites in Serbia: 2. Baranica, 3. Pešturina, 5. Smolućka cave, 6. Hadži– Prodanova cave, 9. Risovača, 10. Šalitrena cave, 12. Tabula Traiana cave, 14. Petrovaradin, 15. At; and paleontological sites: 1. Vrelska cave, 4. Prekonoška cave, Pećurski kamen 7. Lazareva cave, 8 Ceremošnja, 11 Petnička cave, Visoka cave, 13 Janda hole.

faunal remains from the Smolućka cave near Novi Pazar (Dimitrijević 1985) and Pećurski kamen near Sokobanja (Malez and Salković 1988) was conducted. These faunal analyses were orientated towards species identification, relative stratigraphic position in absence of absolute dates, and paleoecological reconstruction. On the basis of fauna, Rakovec dates Risovača to MIS 4, while Malez and Salković (1988), and Dimitrijević (1997), assign Pećurski kamen and the Smolućka caves more broadly to Upper Pleistocene. A synthesis of micro and macrofaunal remains from the Upper Pleistocene archaeological and paleontological records in Serbia was conducted by Dimitrijević (1997). Upon those results, the basic picture of the paleoecology of the Central Balkans during the

Upper Pleistocene was established. It was concluded that taxa typical for steppe predominate, with lesser presence of taxa inhabiting forests, the smallest proportion being those inhabiting decidous forests. Presence of exclusively boreal species was not confirmed. In lowland parts of the Central Balkans, the landscape was dominated by steppe fauna, typically represented by mammoths, wooly rhino, horses, bison and hyenas (Dimitrijević 2011, 2013). Cave bear sites are quite numerous in this region, mostly in the hilly-montaine areas, but also in lowlands. Remains of cave bears from archaeological and paleontological sites in Serbia are well-studied paleontologically (Михајловић and Павловић 1988; Paunović 1991; Dimitrijević et al., 2002; Cvetković and Dimitrijević 2014). From the 10

Milosevic.indd 10

02/01/2020 12:47

Introduction paleontological work, the most important conclusions to this research explain their use of caves during winter hibernation. On the basis of the fauna from Upper Pleistocene sites in the Central Balkans the presence of mosaic ecosystems is apparent. Roe deer (C. capreolus), wild boar (S. scrofa), giant elk (M. giganteus), and narrow nosed rhino (S. hemiotoechus) are present in the Central Balkans Upper Pleistocene fauna and and are stenotypic faunal elements of a more temperate and humid climate. Contrary to that is the complete absence of cervids adapted to drier and colder climate – reindeer (R. tarandus) and elk (A. alces). Narrow nosed rhino went extinct in all regions of Europe during MIS 4, except in Italy and the Balkans, where it survived well into MIS 3, and most probably because of the milder climate (Bedetti et al., 2005; Pushkina 2007). For a more comprehensive picture about the Upper Pleistocene fauna in Central Balkans, it is important to chronologically integrate archaeological and paleontological (Fig. 3) sites. In this way, it is possible to observe more broadly whether and to what extent climate changes could influence the taxonomic distribution and composition, and whether the Central Balkans was a faunal refugia and if it was where these ecosystems could be found. In some faunal studies, taphonomic agents of bone accumulation were also presented and discussed on the basis of taphonomic traces (Dimitrijević 1993, 1996, 1997; Dimitrijević and Jovanović 2002; Kuhn et al., 2014). These studies suggested that all indicated sites contained faunal material deposited mostly by natural and carnivore agents, with little anthropogenic influence, but offered an ecosystem explanation in the form of food web reconstruction. Avifauna from Palaeolithic sites of Serbia are studied only for the Smolućka cave assemblage (Malez and Dimitrijević 1990). Over the area of Central Balkans, more precisely the territory of Serbia from which the material for this study comes from, there are just a few systematically excavated and published sites, out of which the majority, according to dates obtained, contain Late Pleistocene sediments. However, several Palaeolithic projects are currently undergoing and they shall produce more results which will allow further insight, and in greater detail, into human lifeways through the Late Middle and Upper Palaeolithic in Central Balkans.

11

Milosevic.indd 11

02/01/2020 12:47

2 Theoretical Framework and Hypotheses Ecological niche concept and optimal foraging are used as a theoretical framework in this study. They are closely connected, with optimal foraging actually being realized within the ecological niche concept, since it integrates the most important proxies in this study – ecological setting on one hand, and large carnivores, humans, their common herbivore prey, and trade-offs for processing it into a food web on the other. This is because both ecological niche concept and optimal foraging are firmly rooted in evolutionary ecology (Krebs and Davies 1978) and perceived through human behavioural ecology (Winterhalder and Smith 1981). Evolutionary theory suggests that the goal of foraging is to forage optimally – to maximize net profit of foraged foods (Smith 1979). As the topic of this study is competition, understanding foraging optimization is essential to interpret behavioural ecological pursuits of humans, and how they allocate their time within food procurement and processing towards available ecological patches, other competing agents appearing in those patches beside their prey such as large carnivores, but also other activities necessary to gear up for a certain foraging.

all species successfully thrive, after which it is possible to model different effects of ecosystem destabilization and decomposition, and the consequences it can have on animal, and other living communities. By using this theory in interpretation of zooarchaeological analysis it is possible to establish population fluctations of certain species, causes of their expansion, shrinkage, or even extinctions in certain regions, as well as their mosaicity, concentration or dispersal in the areas with several different ecosystem habitats. It identifies the ecological requirements of different species, including humans, from the ecosystem. Ecological niche theory used in contemporary archaeozoological analyses from Palaeolithic sites was defined by Stiner (1994). In her study of Late Middle Palaeolithic faunas in Central Italy she integrated the quantitative data on number of animal remains of different taxa while relying on taphonomy to study the interrelationship of different species, notably between carinivores and herbivores in comparison to humans and herbivores. Key points of the study were the creation of specific skeletal profiles of different species, notably herbivores, based on predation and processing intensity of prey by the most common large carnivore species in the landscape and humans, as well as age structures of herbivore and carnivore individuals characteristic for such assemblages, in different contexts. The study also stressed that humans could produce different patterns of processing marks in relation to their current settlement strategy and subsistence goals within it. The number of species which humans used in subsistence depended on the sustainability of the habitat, but more importantly on competition with other large carnivores. The study conclued that the population ecology of herbivores in the Upper Pleistocene of Central Italy influenced settlement strategies of late Neanderthals, and that humans and large carnivores often preferred the same type of prey, that they hunted equally successfully using cooperation, and alternately occupied the same sites in such a way that, from an ecological perspective, humans and large carinivores were part of the same predator guild (Stiner 1992). Because large predators and humans hold high rank in the food chain, this could lead to elevated exploitation pressure on certain herbivore species and their demise, and thus an imbalance of the ecosystem, requiring Neanderthal societies to move on to a new region. It could be also the main reason for the high mobility settlement pattern observed among late Neanderthals in Central Italy, without any base camp sites. In essence, Stiner combined Hutchinson’s definition of ecosystem as comprised of biologically ranked layers, and Elton’s based upon a holistically modeled branched ecosystem. In complex ecosystems characterized by the large number of interdependent species (large guilds),

2.1 Ecological niche concept Ecological niche concept can explain some causal relationships between hunter-gatherers and their animal prey on the basis of archaeozoological analysis, since Palaeolithic settlements represent a fragment – palimpsest of the past ecosystems in which humans and animals once coexisted. Ecological niche represents one of the approaches from which is possible to observe causal relationships between different animal species, their habitat preferences, common influences on the ecosystem or same interest in the food choice. The theory emerged in animal biology, and was systematised by Elton (1927) and Hutchinson (1965), based on Darwin’s theory of evolution by natural selection. A key factor in this theory is the correct identification of agents comprising the ecosystem. According to Elton, the most important component of reconstruction is the food web, because he holds that two or more species with the same food preferences cannot inhabit the same niche, one must take over eventually. Contrary to Elton, who views ecosystem as a branched system, Hutchinson conceives it as comprised of multiple layers – natural components of the landscape: sedimentological and pedological base, climate, biomass, etc. These enable certain species specialized for particular environments to inhabit them, while also inhibiting some other species from doing so, consequently creating the array of species suitable for their mutual cohabitation in certain environments. According to Hutchinson’s view, the starting point in niche construction is defining the ‘‘fundamental niche’’, the niche in which 12

Milosevic.indd 12

02/01/2020 12:47

Theoretical Framework and Hypotheses of the acquired carcass (Marean and Cleghorn 2003). Central place foraging and Prey choice models (Cannon 2003) are regarded as the most important for this study as they help to establish how far the foragers will travel to encounter, hunt, butcher and transport back different prey to the central place, as well as why that prey is specifically targeted compared to the total animal resource availability and when to swich habitat patch in search of food. The main premise of these models is that humans do not feed as they go but instead forage in landscape and then return food to camp. Thus, human settlement at the site (logistical or residential) is regarded as central place where they process animals acquired in the surrounding landscape and prepare/maintain/make toolkit necessary for such tasks, or prepare meals from acquired food. Butchery and transport, which are inversely proportional to the distance covered in search for the animal are presented as the sum of invested time and energy, and altoghether commonly determined as processing. Processing strategy, thus, is a set of decisions influencing which animals will be targeted and how anatomical parts of those animals will be ranked in butchery/transport to suffice according to the size of the carcass, number of hunters, carriers, distance over which it needs to be transported, as to not only return the spent energy, but to feed the remainder of the group. Although the exact sequence of these actions is not possible to establish and measure directly, it is possible to measure their relative pattern abundance through different skeletal elements. Processing strategies for different animal species is thus possible to identify by comparing the relative pattern abundance to establish processing intensity of different anatomical units in different species. For example, filleting during field processing makes transport easier and opens the possibility to extract marrow on the same spot, but for some anatomical units, such as axial skeleton or feet elements, this requires much more time. Dismemberment of the carcass takes less time, but dismembered anatomical units are bulkier for transport, especially in larger game. However, this model would present difficulties when interpreting transitional subsistence strategies that might often occurred within Upper Pleistocene human groups who could easily shift across vast uninhabited areas driven by food source depletion or other environmental causes. In these cases humans would actually feed as they go, without making distinction between where they live and where they process their food. Ethnographic research has provided valuable data to gain insight into ways in which hunter-gatherers balance between the effort and time spent, and the return rates when processing different herbivores in relation to distance from camp, which are all measures of their foraging efficiency. Butchery begins with skinning and dismemberment of animal parts which have the largest return rates for the least amount of processing time in regards to distance to the camp as a central place. It can end in marrow extraction on the spot even from marginal skeletal parts, for which a greater investment of effort is needed for a relatively small return (Fig. 4). We are interested in what the parameters which determine when hunters will discontinue processing different animal parts are. One of the main parameters is the ratio between the

this ecological theory is possible to amend with „Neutral theory“ (Chave 2004), which states that even highly competing species can adapt to cohabitation by making a compromise, based on avoidance in certain circumstances, or interdependence, based on cleptoparasitic behaviour of carnivore species from the top of food chain, where the behaviour of a represented carnivore species has an important role in understanding of the niche food web mechanisms. 2.2 Foraging theories Applying various foraging theories by using mathematical models that rely on net energy as a unit enable archaeozoologists to test hypotheses regarding human subsistence behaviours. In these models energy is established as a proxy for both reproductive fitness, reflected in size of human groups, and other activities that might enhance social aspects of human groups, since the less time spent in foraging the more time can be involved in other activities – exchange of ideas and socialization with others for e.g. since it is impossible to fully test the models because many parameters may be lacking or inconsistent, they may present the outcome of long-term effects for the same foraging behaviour. Models that are most commonly used in Palaeolithic archaeozoology are intended to study optimal foraging behaviours in order to produce theoretical explanations as to why certain foods, and patches that contain them, have been specifically selected, and why they are processed in the observed pattern. This has been previously used in biology to study foraging preferences and decision making in contemporary animal communities, in order to better understand their ecological demands against available resources. Soon after, foraging theories were adopted by archaeology and social anthropology (Smith and Winterhalder 1981; Winterhalder 1983; Smith et al., 1983; Mithen 1988; Winterhalder 1996) as an interpretative tool which can, besides human subsistence, assess human processing strategies tied to the efficiency and flexibility of human foragers when searching, foraging/hunting, and butchering their prey. It can explain how human foragers rank not only different prey species, but also different anatomical parts in different prey as well. To do so, it uses models related to prey choice and processing strategies in a previously reconstructed ecosystem to explain human decisions observed on animal remains in form of patterns (skeletal profiles, age selection, butchery marks, etc.) – Prey choice model, Patch choice model, Marginal value theorem, Central place foraging model (Emlen 1966; Macarthur and Pianka 1966; Charnov 1976; Stephens and Krebs 1986; Grayson 2001; Broughton 2002; Cannon 2003; Burger et al., 2005; Marín 2009; Bettinger and Grote 2016). Within them, the body of the prey is percieved as structured from anatomical parts of different nutritional values, each of which demands certain investment of time and effort during butchery and transport. Prey processing begins after the hunting episode or, if scavenging, when the carcass is accessed, and assumes butchery and transport 13

Milosevic.indd 13

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 4. Summary of optimal foraging. A decision was dependant upon the processing function between search/processing time, distance from central place, energy spent by the hunters, and weight of anatomical unit/its caloric yield versus time required to process it. The point presents the start of gaining back spent caloric yield.

distance to the camp and the size of the carcass, which if large mainly makes hunters to process the prey more intensely to easy the transport, especially if net gain in energy is high (Metcalfe and Barlow 1992; Cannon 2003; Nagaoka 2005, 2006; Lupo 2007; Niven 2007). Thus, an important notion is that the prey that has the highest net gain of energy per procesing time will always be pursued upon the encounter, whereas less rewarding prey can be ignored. However, low-ranked resources may predominate a diet if higher-ranked ones become sparse in the landscape. Processing steps are possible to observe on the basis of taxonomic composition of herbivore fauna, skeletal representation of various herbivore species, and animal parts containing butchery traces.

of medium and high ranked food utility elements and preferred age structure of prey. This strategy is oftenmost observed at the sites situated in a mosaic ecosystem. Specialized strategy is based upon the exploitation of a single species even if it occurs in multiple ecological patches, because of high energetic return rates in the way in which it is acquired, mostly through ambush mass hunting, but also trapping. Because of that, lower-return prey is always ignored, regradless of its abundance, since their opportunity cost is considered to be higher (Stephens and Krebs 1986). This strategy usually leaves almost complete skeletal profile of the hunted species, while age can differ. Butchery marks on hunting species are usually extenisve and appear repeatedly on same elements and same landmarks, pointing to standardized procedures of prey processing. The appearance of specialized strategy is dictated by the behaviour and population ecology of the herbivore species, since not all large herbivores can be mass hunted, as well as their population density in the region, or seasonal translocation of species moving in migratory herds.

In relation between ecological patches and ranking of prey species it is possible to conclude whether human subsistence was opportunistic, diverse or specialized (Phoca-Cosmetatou 2009). Opportunistic strategy is defined by presence of a variety of herbivores used in subsistence, with unbiased skeletal profiles and age structures, and it points to a random selection of ecological patches in search of prey species, rather than targeting specifically one species or one patch. Opportunism may include scavenging (Stiner 1991, 1994; Pobiner 2015). Diverse strategy assumes that hunters search for one to three species living in the closest proximity of the camp and that usually inhabit same ecological patch. It is characterized by skeletal pattern with good representation

2.3 Hypotheses Animal remains from the Pešturina, Hadži-Prodanova, and Smolućka caves are suitable for moving within given theories and hypotheses for several reasons. These cave sites are palimpsests which human societies used infrequently and shortly before moving on, and at which they conducted 14

Milosevic.indd 14

02/01/2020 12:47

Theoretical Framework and Hypotheses either specialized activities or activities necessary at the moment. Thus, they left much space for carnivore species to regularly settle in them. Such a site type is characteristic for the interior of the Balkans, and so the results obtained can be compared with other sites in the region. Each of the three cave sites has a different composition of carnivore remains, so it is instructive to observe which herbivores, and consequently ecological habitats were targeted by them. Although the archaeozoological material from palimpsest sites cannot reconstruct the complete hunting and processing strategies for every species targeted by humans, it is possible to assess the human ecological interaction with fauna living in the landscape through which they moved, since the remains of taxa deposited at such sites are mainly deposited independently of human prey choices, as a natural consequence of the food web.

to identify and discern activity zones of humans and carnivores on the basis of spatial distributions of different classes of zooarchaeological material and other sets of finds?

Concerning the first hypothesis which deals with ecological competition between humans and large carnivores, it is possible to observe which herbivore species met their demands, and whether it influenced human prey choices, processing and transport strategies, and more broadly, the settlement strategies. Did humans and large carnivors target the same prey? What is the difference in the age structure between the species, especially for herbivores? What is the level of overlap in the exploitation of various ecological patches and herbivores by humans and large carnivores within the same patch? This can be tested in regards to large mammal and bird taxonomic composition in main stratigraphic units, as to establish the evolution of ecological patches in local landscape. Within these, it is possible to compare the extent of human processing marks and large carnivore teeth marks appearing on different taxa. The second hypothesis can be tested by a comparison of taxonomic compositions between Middle and Upper Palaeolithic contexts, as within the sites in this study, but also with other sites. In such a way it is possible to observe whether ecological patches in the Central Balkans were uniformly or irregularly distributed, and the pace at which they changed during the Upper Pleistocene, compared to the Danube corridor and sub-Mediterranean and Mediterranean zones of the Balkans, and how this could be reflected on human societies. What is the difference in skeletal parts representation and processing strategies of between the Neanderthals, modern humans, and carnivores? Is there any inter-regional difference or difference in comparison to other Europen southern peninsulas? The third hypothesis is tested on the basis of spatial distribution at the site, and by studying debris zones created by human processing as described in ethnolographic studies of contemporary hunter-gatherers, and acutalistic large carnivore studies that describe debris left after their feeding episodes in controlled and natural conditions. Animals might occupy different parts of the cave than humans depending of reason behind its use – communal, natal or hibernation den, scavenging, etc. Is it possible 15

Milosevic.indd 15

02/01/2020 12:47

3 Materials and Methods Material analyzed in this study are large mammal and bird remains from the sites of the Pešturina, Hadži-Prodanova, and Smolućka caves. Microfauna, small rodents and lower vertebrates, for which methods of analysis differ, are not covered by this study. The faunal analysis is accompanied by spatial analysis for the Pešturina and Hadži-Prodanova caves which, beside temporal, may allow spatial segregation of the accumulated remains. Considerably more faunal remains come from Pešturina (NISP = 34766), than from the Hadži-Prodanova (6574) and Smolućka (1468) caves.

Some cuts from these periods into subsequent layer 2 led to material mixing in certain parts of cave. The boundary between layers 1 and 2 is quite clear, and occurs sharply, whereas the boundary between layers 2 and 3 is more gradual. Layers 2 and 3 have similar heterogenous structure in which alevrite predominates, with presence of clay and small rocks, and they cannot be clearly separated by colour but only by the difference in compactness, since layer 3 is far more compact than layer 2. Layer 3 ceases at the bedrock bench between the front and rear part of the cave, which is also seen in section (fig. 5). The boundary between layers 3 and 4 is less clear and diffused but can be defined by the approximate level of larger rocks, which marks the beginning of layer 4 (4a) and contains more densely packed smaller rocks with occasionally larger ones. After that layer 4 becomes quite loose, reddish in colour, and characterized by the presence of fine sand (4b). The lower part of the same deposit is characterized by a calcarous tufa level (4c), and a deposit extends below it (4d), but has only been recorded on a small surface, and remains from it are omitted in this study. Geochemical analysis, which can reveal possible more subtle distinction or groupings of sublayers, is in progress. For this study faunal remains were contextually grouped into layers 2, 3, and 4. Sediments and osseous material from Pešturina has undergone C14, and ESR, with OSL dating results still not published. Radicoarbon method, although more precise, covers the material deposited up to approx. 40 kya BP, while ESR can date deposits outreaching the temporal limits of radiocarbon dating. Overview of the dates obtained by C14 and ESR methods is shown in table 1. Deposits excavated up to now are of Upper Pleistocene age and were deposited during isotope stages 5c – 2. During the given period of research two trenches were opened. One was positioned around the central area in the front part of the cave, while the other was opened to the level of bedrock between the front and rear part, and encompassed a previous hole cut by treasure looters. Trenches covered a total area of 24 m2 (fig. 5), from which an equal volume of excavated sediments came from 9 m2 (squares L-M-N 9; L-M-N-O 10; M-N-O 11). Excavations were done in a square grid, with square sides of 1 m, split into four quadrants. Layer 2 contains a Gravettian chipped stone industry, while layers 3 and 4 contain a Mousterian industry (Михаиловић and Милошевић 2013).

3.1 The sites The Pešturina cave (or Jelašnička cave 1) is situated in low foothills of Suva mountain, between the Niš lowland and the entrance to the Jelašnica gorge, and between the village of Jelašnica and Jelašnica cementery (N43°17’58’’; E22°3’7’’, 305 m a.s.l.). Excavations were conducted by the department of Archaeology, Faculty of Philosophy in Belgrade, by Dušan Mihailović. Test trenches were laid in 2006 and 2008, and systematic excavations started in 2010, together with Winnipeg University from Canada and Mirijana Roksandić, and are still ongoing. The cave is situated below a karstic hilltop which encloses a funnel shaped stream valley. This valley stretches upslope in a north-west direction from the Niš lowlands, and represents a proluvial cone of a low slope. As it approaches the cave its slope rises and it gets narrower. Today this valley has a permanent running stream. The cave is situated almost at the highest point of this valley, below the top of the hill, from which the vantage point captures around half of the Niš lowland landscape (fig. 5).1 Such a dominant position was certainly of special importance to Palaeolithic hunters in observing the terrain for prey. The cave is relatively small, 22 m long, 15 m width of the enterance, and maximum height of around 3,5 m. The cave entrance faces west. The interior space has an approximate hourglass shape and can be divided into front, rockshelter-like part, and rear part, divided from the front by a narrowing, from which bedrock is much closer to the surface. The front part is characterized by thick deposits, from which almost 4 m is reached in some squares. These deposits are divided into 4 layers – one formed during the Holocene (layer 1), and three formed during the Upper Pleistocene (layers 2-4; fig. 5). The Holocene layer contained occasional finds from Neolithic/Eneolithic and perhaps other later Prehistory, Antiquity, and Medieval periods, and its use in the modern age was also apparent.

The Hadži-Prodanova cave is located around 3,5 km to the north of the town of Ivanjica, in the canyon of Raščićka river, which, before its confluence with Moravica river, joins with the Marina river. It is situated on the right side of the Raščićka river, at an altitude of 630 m, above the local road Ivanjica – Lisa, in the surroundings of the village Lisa and an area named Ocokoljići, across from a stone

 Based on topographic map, section for Bela Palanka 583/1/1 and geological map, section for Bela Palanka K34-33. 1

16

Milosevic.indd 16

02/01/2020 12:47

Materials and Methods

Figure 5. Geographic location (upper left), Pešturina cave plan (left), section (upper right) and microregional position through DEM model from NE direction. Photo: field documentation; DEM model in ArcGIS software, Irina Kajtez.

quarry.2 The cave entrance is narrow, around 3 m wide and comes to a narrow point of 1,75 m, after which cave expands into two vast chambers. In front of the cave there is a plateau covering a surface larger than 80 m2. Although the activity of the stone quarry changed the landscape, it is possible to conclude that the cave was situated in a smaller limestone canyon. The plateau in front of the cave does not offer much visibility of the surrounding landscape, and the cave is rather secluded. Archaeological excavations were conducted in 2003 and 2004,3and were salvaged, prior to restoration of a church on foundations from 1909, on the plateau in front of the cave, and cementing of pavement through the part of the cave for touristic visits.

layers 4 and 5 are Mousterian. Layer 3 contained no lithic material. Deposits of the site are dated and in preparation, but fall in a range between 44 to 25 kyA BP (table 2). It can be concluded that the excavated layers were deposited over a relatively short period of time during OIS 3. The excavated surface was 16 m2, with volume of excavated earth equal to over 10 m2 (squares C-D-E 8-9-10, and E11). The Smolućka cave is located in the valley of Smoluća creek, 2,5 km to the west of regional road Novi Pazar – Tutin, near the village Crkvine. It is situated in a karstic valley which forms a broken canyon several km long. Short and steep dolines of heavily eroded limestone cut the landscape which is gently sloped from the Pešter highlands towards the valleys of Sebečevska and Raška reka.4 The cave is at 945 m a.s.l. The entrance is south orientated, 4 m wide, 4,5 m high. It is 25 m long with a bedrock cascade in the back of the cave. Deposits on the cascade are shallow, while a larger lower area of the cave contains thicker deposits (Калуђеровић 1985). The potential of the Smolućka cave as a Paleolithic site was recognized during the „Archaeological survey with test excavations of Tutin municipality“ in 1982, and test excavations were conducted in 1984, and

Excavated deposits reached the depth of 2,5 m, and were divided into 5 layers (fig. 6). Boundaries between layers were clear, and the same deposits can be identified at the plateau as at the entrance part of the cave (Михаиловић and Михаиловић 2003: 13). Just 15 chipped stone artefacts were found during excavations. Finds from layer 2 have Epigravettian character, while those originating from   On the basis of topographic map, section Čačak 529/3/3, 529/3/4, and geologic map, section for Ivanjica K34-17. 3   Excavations were directed by dr Dušan Mihailović from Faculty of Philosophy in Belgrade, with Bojana Mihailović from National Museum in Belgrade, and Heritage protection office in Kraljevo. 2

 On the basis of topographic map, section for Novi Pazar 579/4/3, 579/4/4, and geological map, section for Sjenica K34-29. 4

17

Milosevic.indd 17

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 6. Geographic position (above), section (left) and plan of the entrance part of the Hadži-Prodanova cave (right, arrow points towards cave interior). Section drawings: field documentation.

from 1985 to 1987 systematic excavations continued. Excavations were conducted by the Archaeological Institute in Belgrade, under the direction of Zvonimir Kaluđerović. Archaeological trenches encompassed nearly half of the cave space below the bedrock cascade, and part in front of the entrance. Part of the field documentation is currently missing, so that the contexts were reconstructed by author on the basis of several preserved section drawings and reports from excavation seasons and earlier publications on fauna from the Smolućka cave (Dimitrijević 1985). Five layers are distingushed (1 to 5), one lens (layer 4z) which contextually and depositionally corresponds to layer 4. Layer 1 was deposited during the Holocene, layer 2 contains mixed Holocene and Pleistocene deposits of subsequent layer 3. Layers 3–5 are of Upper Pleistocene age (fig. 7). Radiocarbon dating was conducted on two occasions: a wood charcoal from layer 3 was dated by the AMS method (Hedges et al., 1990), and more recently several samples from layer 5 by Alex Bridget (personal comm.). The AMS age of layer 3 is close to 40 kya cal. BP, while the age of layer 5 surpasses the C14 decay period. Around 300 Mousterian lithic artefacts where collected in layers 3-5 (Kaluđerović 1993). Excavations were conducted using a square grid of 1 m sides, according to arbitrary excavation layers taking the levels of each surface. All deposits were

Table 2. Published C14 dates for different layers of the Hadži-Prodanova cave, according to Alex et al., (2019). layer

laboratory sign

C14 years BP

2

RTD-7274

25200±130

2

RTD-7280

18700±80

2

RTD-7271

25100±130

3

RTD-7277

31500±190

4

RTD-7276

35500±280

4

RTD-7281

24000±120

4

RTD-7273

33600±280

5

RTD-7275

29500±160

5

RTD-7282

47700±1650

5

RTD-7278

>51400

5

RTD-7270

39500±540

18

Milosevic.indd 18

02/01/2020 12:47

Materials and Methods

Figure 7. Geographic position (lower left), section (left) and plan of the Smolućka cave (right). Drawings of section and base: field documentation.

sieved with 4 mm mesh, while water sieving was done for microfaunal samples. From preserved section drawings, it can be concluded on the basis of excavation layers’ levels drawn projected on the section, that some contexts may contain finds mixed from more than one deposit, which posed contextual problems, except for the 4 and 4z mixing which originated around the same time. The appearance of osteological remains also pointed to this assumption, since part of the material packed together contained bones in different states of fossilization.

b) Identification and quantification of human and carnivore taphonomy: scored NISP per MNE, age structure. II Study of degree of destruction by taphonomic agents: a) Fragmentation analysis: repetition of diagnostic zones, breakage types, circumference preservation, teeth to long bone ratio; axial elements to extremities ratio; b) Per cent MAU to FUI ratio, carnivore ravage per element, weathering, trampling, bone dissolution, specific bone mineral density rankings.

3.2 Methodology of zooarchaeological analysis In order to explain ecological and economic pursuits of humans in the landscape they inhabited during different periods of time, the method of faunal analysis is conceptualized to interconnect taxonomical determination, taphonomic analysis, and quantification of taphonomic traces and skeletal remains. The advantage of such an approach (Bar-Oz and Munro 2004) compared to separate counts of taphonomic traces and skeletal elements lies in fact that it diminishes the equfinality of results (Lyman 2004) and their interpretation. The structure of the approach constructed for this study comprises of:

The results are partially obtained with inclusion of previous studies and with help of avifauna, GIS and SEM specialists. Taxonomic and age determination of large mammals from the Smolućka cave is based upon Dimitrijević (1991, 1993). Analysis of cave bear remains, and data for tripolar graphs presented in the study are based upon data from Banković (2006) for the Hadži-Prodanova cave; and Dimitrijević (1991) for Smolućka cave bears. Bird remains from Pešturina were identified thanks to Zlatozar Boev from National Natural history Museum in Sofia, Hadži-Prodanova bird assemblage thanks to Sheila Hamilton Dyer, Burnmouth University, while birds from Smolućka follow determination by Malez, Dimitrijević 1990. GIS models and spatial distributions were produced in collaboration with Irina Kajtez. SEM images and measurments were produced in collaboration with Nikola

I Sum of primary quantitative archaeozoological data: a) Identification and quantification of remains: NISP, MNE, MAU, MNI; 19

Milosevic.indd 19

02/01/2020 12:47

Competition Between Humans and Large Carnivores Vuković at SEM laboratory of the Faculty of Mining and Geology in Belgrade.

or underrepresentation of different anatomic regions. Based on MNE, taking into account the laterality of the elements, in this case mostly teeth, the MNI was calculated.

3.2.1 Taxonomic determination, skeletal element determination and quantification of remains

Bird remains from the Pešturina and Hadži-Prodanova caves were determined using comparative osteological collection from the Vertebrate Animals Department of the National Museum of Natural History (Bulgarian Academy of Sciences, Bulgaria), and comparative collection of Burnmouth University (United Kingdom). Avian taxonomy follows del Hoyo & Collar (2014, 2016). Species’ habitat preferences are after Harrison (1982).

Large mammal remains were first identified in different taxonomic categories. Taxonomic identification was done using comparative collection at the Department of Archaeology, Faculty of Philosophy, Belgrade Univeristy, and corresponding atlases in its library (Lavocat 1966; Pales and Lambert 1971; Eisenmann et al., 1988; Brugal 1983; Stampfli 1963). Because of the highly fragmentary state of the material all taxonomic categories are grouped, according to cortical thickness of bone, dentine and enamel, into classes I (largest) to V (smallest). Small fragments of horn/antler, bone, dentine and enamel for which it is only possible to identify as mammalian were counted together. A minority of specimens were readily identified as skeletal elements/element parts at the level of species or genus, while the majority were determined to the level of Class, Order or Family (and their subcategories) and were counted as belonging in general to: horn/antler, cranium, mandible, tooth, vertebra, rib, scapula/pelvis, short or long bone, when it was possible to identify them as such in terms of NSUTS. They were also ranked to one of the five size classes according to their body mass (Bunn 1982; Saarinen et al., 2016) and similar bone robusticity, having in mind species composition in the materials comprising this study. Class I includes megafauna – proboscideans and rhinos; Class II – large bovids, cervids, equids, and carnivores – bison, giant elk, horses, cave and brown bear, steppe lion; Class III – medium sized cervids, equids, bovids and carnivores – red and fallow deer, european ass, ibex, hyenas, leopards; Class IV – smaller medium sized ungulates and carnivores – roe deer, boar, chamois, wolves, lynx; Class V – lagomorphs, large rodents, and small carnivores – hare, beaver, porcupine, fox, badger, wild cat, mustelids.

Because of the fact that different skeletal elements have different preservation chances, both owing to their energetic payoff attractivity for humans to process them and carnivores to consume them, and their specific mineral density, MNE was choosen as the main analythical method for quantification of both taphonomic traces and number of individuals. It is conceptualized as an assemblage of diagnostic zones, meaning that the most numerous diagnostic zone of the element will account for the minimum number of that element. This approach is based on the premise that every skeletal element can be divided in several zones, or areas, containing characteristic morphological landmarks, which differ across the species/ genera. The system was developed by Dobney and Riely (1988), and has been accepted in quantification of elements of microfauna, avifauna, and ichtiofauna as well, which this study uses in simplified form because of the highly fragmentary state of the material. The overview of diagnostic zones designed for this study, following Dobney Riely (1988), is given in the appendix 1. Such a system is not only useful in quantifying skeletal elements, but also to merge quantification of taphonomic traces with them, specifically butchery or tooth marks, which can follow certain patterns, making them appear highly localized on some parts of the elements, depending on processing intentions of humans and carnivores. For that reason it is an optimal choice, consolidating quantification of skeletal elements and taphonomic traces, aiding in better understanding of human processing strategies and consumption by large carnivores (OtárolaCastillo 2010).

The quantification method of faunal remains was adapted to their highly fragmented state, as well as the fact that hunther-gatherers often transport their prey in parts (Fig. 8). The main quantification method was NISP, which counts every specimen. From it, MNE was calculated on the basis of NISP identified to a genus or species belonging to same skeletal elemtents. Then MNE was used to calculate minimum animal units (MAU; Binford 1978) for different elements of better represented species and genera. MAU is also useful for quantifying highly fragmented skeletal elements, where is often difficult to distinguish between left and right side of the skeleton. For the ungulate species most represented by number of elements, the utility index (Outram and Rowley Conwy; 2008; Madrigal and Holt 2002; Friesen 2001; Marín 2009) is given as the total Food Utility index (FUI) comprised of meat and marrow utility indices (Jones and Metcalfe 1988; Metcalfe and Jones 1988), as well as their specific mineral bone denisty (BD; Kreutzer 1992; Lam et al., 1998), so that we can better understand the reason of over-

Over the course of the skeletal element identification, special attention was given towards long bone shaft fragments (LBSF). A large number of authors (Stiner 1991, 1994; Marean and Kim 1998; Batram and Marean 1999; Rogers 2000a, 2000b; Marean and Cleghorn 2003; Bar-Oz and Munro 2004; Faith and Gordon 2007; Marín Arroyo 2009) have stressed that LBSF are very important agents in long bone element quantification, especially MNE and number of taphonomic traces per element, since the epiphyses of long bones are significantly more prone to destruction due to lower mineral density and feeding choice of predators. The studies of listed authors have shown that ignoring the LBSF in long bone MNE counts, when quantifying highly fragmented faunal 20

Milosevic.indd 20

02/01/2020 12:47

Materials and Methods

Figure 8. Graphic representation of connecting the specimens through their taxonomic and element part identification with taphonomic traces on them. Quantification method is conceived to entangle MNE in different taxa and taphonomic traces appearing on them. In such way it is possible to define human influence on accumulation of faunal remains more precisely, and processes tied to them at the site.

remains, will always give the same skeletal pattern, the socalled Klasies River pattern (Klein 1976; Bindord 1985; Chase 1986) or head-and-feet pattern. The pattern can be observed in the Balkans for the material from Crvena stijena in the study of Malez (Малез 1975), where long bone shaft fragments were ignored. Contrary to that, archaeozoological study from recent excavations brought different results. It is characterized by the low presence of upper limb and axial elements compared to the presence of head elements (mostly teeth) and lower limb elements (metapodials/carpal-tarsals/phalanges), which influenced the interpretation that, at least, archaic humans (Late Lower and Middle Palaeolithic) were obtaining animal protein mainly through scavenging. Since it is now established that there is a direct connection between the difference in bone mineral densities of different elements and their representation in zooarchaeological materials (Lam et al., 1998, 2003; Stiner 2004a) it is exactly the teeth and lower limb bones that have the highest mineral densities and thus the chances of preservation as well.

carnivore consumption. Utility indices help to understand the reason behind the presence of different butchery and teeth marks on different elements, and processing intensity of different elements with nutritional yield of the elements, enabling us to understand if the processing occurred as to their utility ranks.

Quantification of anatomical regions through MAU is instructive when interpreting butchery and transport strategies, or identifying the predator responsible for further processing or accumulation of bone material after humans. Humans tend to leave smaller bones that have little or no nutritional value unprocessed, such as carpals/ tarsals/phalanges, and they are often transported as a side (schlepp) effect. MAU can explain different skeletal patterns and tie them to human processing strategies or

3.2.2 Taphonomic analysis of the remains

MNI was used to indirectly quantify human and carnivore strategies in exploitation of different habitats that existed close to the site, compared to paleoecological demands of herbivores and a model of their population density in certain habitats. Population density is given in log scale, based on the model from Silva et al., 2001, which is calculated from the mass of an adult individual and whether it is specialized to a certain type of habitat or is a habitat generalist. For herbivore species the log of species density is 1.43-0.68 (log m of taxon), and for carnivores 1.41 (log m of taxon)2+0.28 (log m taxa)3. The index of -0.367 is subtracted if given habitat is not the preferred habitat of the species.

Since every archaeological site has its own taphonomic history, understanding of taphonomic traces on faunal remains is a key to identification of agents which affected the formation of zooarchaeological assemblages. The ultimate goal of taphonomic analysis of faunal remains from Palaeolithic cave sites is to explain human influence on deposition and modification of faunal material within 21

Milosevic.indd 21

02/01/2020 12:47

Competition Between Humans and Large Carnivores the site. As human influence is usually small compared to other agents, it is important to identify all taphonomic agents and quantify their contribution to accumulation of different taxa and skeletal elements. Different agents affect the material in different ways, but it is possible to identify their order of appearance on a relative time scale. The basic division of taphonomic traces is those incurred before, and those incurred after the deposition, and which could be from various biotic, abiotic, active or passive agents. For identification of taphonomic traces, specimens were observed under a light microscope with x80 magnification, and a small number of specimens containing unclear traces were run under SEM-EDS.

rather small per long bone – around 4.5% (Pickering and Egeland 2006), and is directly tied to the specific mineral bone density of that element. Thus, less impact is needed to crack open upper limb elements than lower ones, with ungulate metapodials being the toughest to crack (Lam et al., 1999). Since taphonomic traces accumulate over specimens as non-linear process, survivability of human processing marks also depends on other agents as well. On a relative taphonomic time scale human induced marks are among the first to be inflicted, and in order to remain visible in the present day, they have to „survive“ the influence of other taphonomic agents (DomínguezRodrigo and Yravedra 2009). Adding to that, the number of cuts and number of marks left on bones by cutting are not directly proportional to the intensity of processing, and that they are highly variable, and depend on carcass size, carcass stance, and toolkit used in processing. Randomly chosen cutmarked specimens were observed with SEM-EDS, where most prominent points across the cross section of traces were measured, and compared with criterions for cutmark identification from the studies of Walker and Long (1977), Bello and Soligo (2008). The goals of high magnification scan were twofold – first, to establish the existence of morphometric variations in cumark cross-sections on various specimens, and secondly, to observe how they differ from trampling marks (Domínguez-Rodrigo et al., 2009) on the material heavily affected by water dissolution. Trampling marks may look similar to cutmarks, but originate from the scratching against ground surface by the movement of cave dwellers, they often exhibit have multidirectional pattern, and can have rounded edges (Behrensmeyer et al., 1986). They are numerous in caves which were extensively used by carnivores, but especially when used for hibernation by cave bears. Because of their weight they can break most of the bones they stomp, or polish them by repeated rolling over in their hibernation nest. Since trampling take place on surface prior to deposition of the material, it can affect previous traces such as cutmarks and teeth scores, because it happens after them (Gaudzinski et al., 2010). Traces of weathering appear due to material exposure to the atmospheric conditions for a long period of time prior to burial under sediments, and they appear as vertical and horizontal cracks in bone structure or exfoliation of bone lamellae (Behrensmeyer 1978). Still, the intensity of weathering is not linear with the passage of time, so it cannot be assessed how much time the material spent on the surface before deposition (Lyman and Fox 1989). Presence of cracks augments the chances for further fragmentation of the material after deposition, while exfoliation leads to partial or complete loss of previous taphonomic traces due to disintegration of bone lamellae on which they were etched.

Biotic agents that actively accumulate bone material are mostly carnivores and humans. The need for prey transport in carnivores arises for different reasons: moving to eat in a less stressful area, sharing or caching of food. During feeding, carnivores leave scores on bone surfaces in the form of pits, scars, and characteristic bone breakage pattern. Identification and comparison of these traces is done both through the results of actualistic studies during feeding episodes of contemporary animals living in captivity or controlled conditions, as well as through data acquired from fossil remains from archaeological and paleontological sites. Through the course of the analysis of carnivore modifications on bones it is especially important to single out species which created them, since different large carnivore species possess different jaw morphologies and teeth rows. The examples of carnivore traces from the material are shown in appendix 2. For taphonomic traces of large carnivores in the Upper Pleistocene of Europe it is crucial to understand the feeding behavior of large predators – steppe lion, hyenas, wolves, and cave bears. The relationship between the number of remains of these carnivore species and composition, number, taphonomy and age structure of ungulates can already give first indications about the use of the cave as a den. Traces of human carcass butchery may follow certain patterns, but in palimpsest sites, because of high degree of mobility, it can differ every time the location was visited. Osteological material from the sites in this study contains traces of lithic artifacts, the function of which through the course of carcass processing can be interpreted based upon their position on the skeletal element. The material also exhibits specimens with breakages characteristic of hammerstone percussion, using the surface as an anvil. Cutmarks were classified according to the method given by Binford (1981: 133, 141) to skinning, dismembering, and filleting marks, based upon the element part on which they appear, although it should be kept in mind that these traces are also highly variable during the Palaeolithic (Egeland 2014). Conchoidal bone breakage by hammerstone percussion is characterized by appearing only on fresh bones and depends on the presence of organic matter (Villa and Mahieu 1991). Experiments have shown that the number of percussion marks is

Traces of burning on osseous material in cave sites can be related almost exclusively to human activities since caves are well-protected from forest fires. Osteological material was mostly exposed to fire by chance, when in close proximity to a hearth. Long bones can be heated on purpose during marrow extraction, since fire turns marrow 22

Milosevic.indd 22

02/01/2020 12:47

Materials and Methods 3.2.3 Ageing of different mammalian species and age structure

into a liquid state, which eases the extraction. Long bones are also heated when used as fuel (Costamagno 2005; Morin 2010). During exposure of bone to high temperatures, the bone tissue recrystallizes. Depending on temperature and length of exposure, burning can affect the bone surface or entire specimen (Tompson et al., 2009). On the basis of burnt specimens it is possible to argue for the existence of hearths and their primary position, since the changes affect only specimens that were within 10 cm of the center of heat. These can be displaced out by trampling or sediment mixing. This study uses changes in colour and texture of bone due recrystallization of bone tissue imposed by Stiner et al., (1995), that is divided into 5 stages.

Carnivores often inhabited caves when they were not used by humans. Hyenas and wolves used caves as both communal and natal dens, while cave bears use them exclusively as natal and hibernation dens. Since carnivores are very competitive for cave space, both inter- and intra-species, it is not possible that two carnivore species will den at the same time. The age structure of carnivores can reveal their nature of cave use. When age structure is close to that in living populations, with progressive decline of number of individuals from youngest to oldest, it is likely that the cave was used as a communal den for a hyena clan or wolf pack (Voorhies 1969; Klein 1982). If a large carnivore community resided in the cave we can expect a considerable amount of ungulate remains brought in and modified by them, as a consequence of predation. When the age structure comprises largely neonate and juvenile (0-3 years) individuals, the cave was apparently used as a natal den to shelter pups.

After deposition, physical and chemical agents can affect the material by further destructive processes or redeposition of the material. For material presented here, the most important post-depositional agents are those leading to bone dissolution. Changes in osteological material that occurred through water dissolution are called corrosive changes. Corrosion occurs when sediment containing osseous material is soaked with water trickling from the cave walls and ceiling. Calcium-carbonate (CaCO3) is already present in surrounding bedrock, but reaches cave sediments with dripping water. If it does not start forming calcium-carbonate floatstones or stalagmites, it raises the level of pH in the sediment in which it precipitates (Hedges and Millard 1995; Hedges 2002). Water dissolution can affect whole bone or just the surface, while affecting more heavily the areas already containing previous taphonomic traces where bone structure is weakened. Thus, these traces are dissolved together with bone lamellae, and when the whole bone is affected, bone structure becomes brittle in situ under the pressure of the sediments above. Also, water dissolution can largely influence bone chemistry, so that the bones heavily affected by this process should not be considered for physical and chemical analyses. Weaker corrosive changes appear in the form of deformed first layer of bone lamellae and spongy look of the bone surface, while bone dissolution is observable under magnification. Weak corrosion does not affect teeth. Strong corrosive changes lead to the rounding of bone breakage, deformation of larger specimens, and are often followed by exfoliation of several layers of bone lamellae. Some of the previous taphonomic traces still retain their characteristics, but it is often impossible to make more subtle distinctions, such as between cut and trampling marks. Strong corrosion affects teeth as well, where it is especially destructive to the enamel, but also weakens dentine structure, which brittles. Such specimens cannot be used for ageing animals on the basis of tooth wear. Criterion for classifying corrosive changes from weaker to stronger is based on Pineda et al., (2014). Presence of mineral oxides in the form of black patches and dots, most probably of manganese origin, should also be mentioned. It could originate from either manganese/irong minerals or organic waste (Marín-Arroyo et al., 2014), but further geochemical and stain testing is needed at Pešturina to adress this issue.

When raising pups wolves tend to hide or mask their presence since they are helpless and blind at the time of birth and can be attacked by other carnivores. In first couple of months mother spends time alone with pups before introducing them to a pack. With such behaviour wolves do not accumulate remains of their prey in natal dens. In this case we can assume that most of the ungulate remains were not deposited as wolf prey. Hyenas raise pups in communal dens right away since their pups are born welldeveloped. Cave bears are unique here, because they give birth towards the end of hibernation, which is also the most critical period for its survival on an annual cycle. Thus, hibernation of the cave bear is marked by high mortality of juveniles up to three years old, dependant on their mother, and young pup-bearing females. Adult male and female bears rarely use same caves, with mortality again rising in old individuals, so the age structure is dominated by young and old individuals. High representation of cave bear remains is often associated with high presence of trampling marks, since the same individuals prefer to hibernate in the same „nest“ inside the cave. The age structure of ungulates adds considerably to taphonomic interpretation of their remains inside the caves, especially in distinguishing between humans and large carnivores as agents of their accumulation, since ungulates do not inhabit caves, although bovine species might use them for shelter during severe climate conditions or violent storms. For hyenas and wolves, which acquire their food both through hunting and scavenging, a natural age structure is excpected – with majority of young, fewer prime age, and fewest old individuals. Contrary to that, humans target prime adults in order to acquire as much bulk of an individual as possible for the same amount of effort, except when they are mass-hunting, where age structure is also as occuring in nature, or when hunting megafauna where they tend to hunt juveniles and subadults. Individual age stage is determined on the basis of tooth eruption and wear stages. Teeth were chosen before bones, 23

Milosevic.indd 23

02/01/2020 12:47

Competition Between Humans and Large Carnivores since they are better preserved. Bone ends, which are the key element parts for ageing, occur very rarely in the assemblages of this study, and are mostly gnawed off. The tooth age stages for a wolf was determined on the basis of lower D4-M1 and its wear stages, while lower D4-P3 were used for the hyena, from Stiner (2004), in which the age histogram is divided into nine age cohorts: the first three are juveniles, next four are adults, and last two are old individuals. The same age divisions are used for cave bear, by counting decidual canines in juveniles, and wear of upper P4, upper M1 and M2, and lower molars, according to Stiner (1998). For dentitions (preserved parts of maxilla or mandible with teeth row sequence), cohort I exhibits only deciduous teeth, cohort II mixed dentition of deciduous and permanent teeth, cohort III finished eruption of permanent teeth and start of wear, while other cohorts follow dental wear patterns. For ageing of ungulates, on the basis of eruption and wear of lower D3, D4, and P4, M1-3 for bovids and cervids, and incisors for horses. Cornevin, Lesbre (1894) was used for horses; for large bovids and cervids Klein and Cruz-Uribe (1984); for caprines Payne (1987); and for megafauna Loguet (2006).

the 2012 campaign in Pešturina. The rest of the material recovered from the Hadži-Prodanova and Pešturina caves in is also displayed according to square grids used at these two sites. The goal was to observe the distribution of human and carnivore activities within the site, so that specimens with butchery and percussion marks, and burnt specimens were plotted against the specimens with tooth scores, regrutiated specimens, and hyena coprolites. For modelling the Pešturina topographic environment, with data on elevation and slope, GRASS GIS software was used to model human movement speed across the surrounding landscape, as cumulative time required to cross raster cells from the starting point, which represents the site location. A model used by the software is conducted from Naismith’s rule, with Atiken-Langmuire’s correction. Essential to this formula is the dependence of movement speed on the slope of terrain in degrees. A shortcoming of this method, influencing the precision of the results, is that the only known factor is the slope. Other parameters which determine movement speed remain unknown, including layout/disposition of water surfaces or forests for e.g. This study used NASA’s STRM DEM models at a resolution of 3 arc/s, or around 90 m, which can be accessed freely at NASA’s website. With these data, a model of terrain slope was made, split into seven categories, and given as a percentage.

A tripolar graph is chosen for display of ageing results. Its main advantages are the ability to show results of all taxa together regardless of MNI differences between the taxa, and to bridge the disproportion in sample sizes between the sites (Stiner 1998). Although it basically balances between young, adult, and old individuals, it is divided into five areas. The two areas represent the expected living structure of the species that occurs in nature, and the „U“ profiled age structure characterized by lack of adults. Such an approach may appear too simplified. However, in a taphonomic scenario of studied material, in which ungulate remains are alternatively deposited by hyenas and humans, and where there are no true settlement horizons, it is not possible to attribute seasonality results exclusively to the human occupation. Moreover, because of the fact that human prey mostly comprises of adults in periods covered by this study, while the seasonality can be studied just on the basis of isolated teeth or dentitions of juveniles, it is misleading to discuss seasonality of human settlement, because the hyenas largely influenced their accumulation.

3.4 Behaviour of wolves, spotted hyenas, lions, and bears It was already mentioned that a large number of criteria defined by different authors, which are used in this study, rely on actualistic studies tied to the subsistance behaviour of contemporary carnivore populations in the wild or controlled conditions, and ethnoarchaeological studies of contemporary hunter-gatherers. Although contemporary ecosystems cannot be used as direct analogies in order to study past ones from the Upper Pleistocene in Europe, they still hold some general parameters which can used in interpretation, such as mosaic or monospecific ecosystems, dependence of mass-population density, herbivore to carnivore population density dependence, seasonal differences in the biomass of different ecosystems, etc. Studies of large carnivore behaviour are also significant for the interpretation of results, considering the large number of their remains and the presence of numerous taphonomic traces of their activity on the material. Social organization of large carnivores can be very structured and complex. As hunters-scavengers they are very versatile in the exploitation of the ecosystems they inhabit and as such, they present competition to humans. General data about the behaviour and lifecycles of large carnivores appearing in the assemblage studied here is based on Mills and Breader (2005) for hyenas (observed in Kenya), Макридин (1959) and Mech (1974) for wolves (observed in Siberia and Alaska), Haas et al., (2005) for lion (observed in Kenya), Pasitschniak-Arts (1993) for brown bears (observed Kamchatka peninsula, Russia).

3.3 Spatial distributions and catchments Spatial distributions are important since they show the interrelationship between finds and their contexts within the site, and on the other hand between the sites and their geographical and ecological landscape surroundings. Spatial distribution analyses were conducted for materials from Pešturina and Hadži-Prodanova, for which a subsquare grid of 50 x 50 cm, and square grid of 1 m, were implemented consistently in material provenience. The method used here is based upon the absolute positions of finds, or in situ, according to the drawings and level database for the Hadži-Prodanova cave, and according to provenience data collected by the total station used during 24

Milosevic.indd 24

02/01/2020 12:47

Materials and Methods The Wolf is a carnivore from the Canidae family which inhabits all continents in the northern hemisphere. As an exceptional ecological opportunist, wolves successfully inhabit a large number of different ecosystems, mostly owing to their complex social organization. Its basic social group is a pack formed around a parent pair and their offspring, which leave the pack temporarily or permanently upon reaching maturity and may establish a new pack or a relationship with another pack. Although packs are mostly mutually intolerant and tend to come into conflict in the case of territory violation, especially in times of year when prey is scattered, several packs can join into a larger one (Mech 1978; Павлов 1990). Cannibalism is rare in wolves, and it mostly happens after the clash of two packs, having the effect of intimidation upon the defeated pack within their territory as a gesture of dominance. Adult wolves require 3-5 kg of food per day, although when starved they can eat up to three times that amout per feeding. For that reason a wolf pack has high mobility and covers around 25 km per day (Павлов 1990; Макридин 1959). When hunting large ungulates a wolf pack shows a considerable level of cooperation which enables them to separate or guide the prey in the desired direction, and to stalk it 5 or more kilometers (Макридин 1959). Stalking tactics comprise grouped assault when the prey is large and slower, while faster prey is stalked by using relay strategy where tired individuals are replaced with fresh ones that keep pace with the pursuer at nearby distance (Peterson and Page 1988; Макридин 1959). Young individuals accompany the pack but do not participate in hunting. Since they reach maturity after two years, and litter 5-6 pups, the wolf population in a given region is fast to re-establish and grow, and for that reason they are highly competitive species with other large carnivores and humans, usually vastly outnumbering them in most of the ecosystems (Павлов 1990). Pups are born between May and July (Schmidt et al., 2008; Павлов 1990), which is important for understanding site seasonality on the basis of juvenile wolf remains. The natal den is positioned ususally in the core part of pack’s territory. With the tendency to transport parts of its prey to save the food for later or share it, a wolf is a considerable taphonomic factor because they actively deposits osteological material. The wolf cannot split open long bones of Class I and II mammals but only gnaw off the ends, so that the carcass elements consumed by wolves are better preserved, and long bone splinters are rarely below 1/4 of the circumference (Fosse et al., 2012).

the European Upper Pleistocene, a wolf is most often a marginal predator. The hyena family comprises specialized hunterscavengers, from which today only four species exist, with the spotted hyena being the only one that lives in a community, while the other three species live solitarily except in the breeding period. A spotted hyena is an opportunistic carnivore, which inhabits ecosystems with high diversity of herbivore species, especially large ungulates. Their societies are named clans, which can be very numerous with as many as 70 individuals, broken in several groups according to the most dominant female line. Contrary to a wolf pack which moves together, a hyena clan is rarely all in one place, with individuals often separating from rest of the clan up to a third of a day in order to hunt small animals or attempt kleptoparasitic action (Smith and Holekamp 2010). A hyena’s metabolism is fast and requires the continuous intake of food, which explains its kleptoparasitic nature of trying to obtain food through a broader array of means, including taking over from other predators whenever possible. Subgroups of a clan are usually formed spontaneously around a common feeding event such as hunting, poaching prey from other carnivores, and scavenging, but also pup raising. Since females are larger they lead clans as well as sub-groups, and their rank in a clan is determined upon female ancestral line – a mother and her siblings. Hyenas follow this social organization very rigidly, and fighting over position in a clan is quite rare. Their litter numbers no more than one or two pups, which are born very well developed. In contrast with other large carnivores they are not blind at birth and have all deciduous teeth developed. The lactation period is relatively long, although juveniles can chew and digest solid food already by three months, when they start to practice hunting on small prey, such as reptiles, small rodents and birds. Interestingly, mothers do not share solid food with their offspring, but other females from the same clan will gladly nurse other females’ offspring. The clan raises juveniles collectively for two years, after which they fall to their rank. Some authors think that this may have to do with the relatively long period needed to develop fully formed sagittal crest on the skull as a main surface of jaw muscle attachments and end of teeth replacement – when full strength of jaw is acquired (Binder and Van Valkenburgh 2000). After the jaws are fully formed, a hyena is capable of processing any large ungulate. Massive conical premolars and carnassials and great crushing force of jaw are adapted to their diet, of which bone marrow is an important component.

Actualistic studies have shown that the skeletal profile of large ungulates from wolf dens is characterized by a large number of limb elements (Stiner 2012), while smaller prey can be transported whole (Esteban-Nadal et al., 2010: 2968, table 5). When large mammals are considered, age structure of the prey almost excludes adult individuals. Since wolves are exceptional ecological opportunists, the presence of wolves in bone assemblage is not a reliable ecologial marker. However, wolf populations usually thrive in ecosystems where they are alone at the top of food chain, while in cohabitation with other large carnivores of

Hyena activities are closely related to high bone fragmentation, since they can break open even the bones of megafauna. Teeth scores of hyena can be expected over all skeletal elements, with high degree of variation in their localization and intensity, and usually appear over 15% of the carcass (Cruz-Uribe 1991; Domínguez-Rodrigo and Piqueras 2003). Cancellous bone ends are mostly gnawed off, while tooth scores on compact bone are densely distributed, and characterized by wide variety of sizes. The 25

Milosevic.indd 25

02/01/2020 12:47

Competition Between Humans and Large Carnivores presence of small regurgitated bone pieces is also indicative of hyena taphonomy since they often swallow them in the course of feeding. A wolf uses a similar approach when opening long bones to consume bone marrow, but has smaller teeth and weaker jaw strength, so that tooth scores are less intense than in hyenas, and span over 6-10% of the carcass. Hyenas can breed in any season, so juvenile and neonate remains do not contribute to site seasonality studies. They form two types of den: natal and communal (Boydston et al., 2006). For natal dens, a secluded space is chosen – if it is a karstic formation, low ceiling and the existance of several entrances/exits are preferred. Communal dens are positioned at the centre of a clan’s territory and its social activity, and placed where part of the clan is constantly present. Communal dens are usually located on the margin of a habitat densely populated with herbivores, and represent good ecological markers because they contain a large variety of herbivore species they brought in, which inhabited the surrounding landscape at a given time. The bones are heavily gnawn, and such sites are also characterized by the presence of hyena coprolites. Hyenas do not posess population flexibility like wolves due to small offspring litter and rely steadily on species richness and density in a given habitat, while wolves can inhabit ecological patches with a low density and low variety of prey. Also, hyenas rely heavily on the presence of other large carnivore species with which they overlap in the prey choice. Such circumstances enable them to use their most effective foraging strategy – kleptoparasitism (Honer et al., 2002), in which higher ranking females organize a group for taking over the carcass, after they scouted and felt that the predator that brought down and started processing and consuming the prey can be overwhelmed and chased off completely, or for a period of time enough to dismember a part of the carcass and get away with it. Social organization of hyena clans is thus adapted for inhabiting areas where competition with other carnivores is strong. Actualistic studies have shown that hyenas mostly choose prey weighing between 56-182 kg (Hayward 2006), corresponding to size classes III and IV in this study, for which hunting is not age selective. When targeting larger prey (size II and I), they select primarily young or weaker/disabled individuals); for example in a savanna ecosystem, prey overlap with lions is no larger than 50%, and since they are larger themselves, lion tend to hunt larger herbivores (Rapson and Bernard 2007).

different ones. Lionesses prefer team hunting and males are mostly lone hunters, although both prefer ambush tactics. The main task of male leaders, however, is to protect the pack and their offspring from other males. This is important since the males of a competitive pack wipe the entire offspring they encounter if males protecting the clan are killed or chased away, in order to produce their own (Haas et al., 2005). Lions can mate during any part of the year (Bertram 1975), so remains of juvenile individuals are not a seasonal marker. A litter usually numbers 1-4 cubs that can spend up to 6 weeks in isolation with the mother. During this period the offspring is most vulnerable, and as soon as they gain vision and start walking, the mother introduces them to the pack, where they are readily accepted and cared for by other lionesses. Contrary to wolves and hyenas, a lion pack spends much of its time randomly browsing within its territory. Subsistence activities of the pack can be divided into diurnal, which are more lethargic, and nocturnal when they actively hunt, especially in twilight periods. The fact that 76% of lion hunting episodes are successful (Hayward and Kerley 2005) makes lions impressive hunters. Weighting in average of 175 kg for males, and 120 kg for females, lions are capable of bringing down much larger prey than other carnivores. They ignore small and fast prey, since their cardiac capacity is too small compared to their body size for long chases characteristic for wolves and hyenas. When hunting, they rush ferociously at the closest prey, either from running close around a group of herbivores, tending to tighten them, pouncing one target with explosive sprint and a jump. This approach is more often in lionesses, while males prefer ambushing from taller vegetation. Lionesses tend to choose prey not much larger than themselves, average weight of around 126 kg, while males usually choose prey around 400 kg (Rapson and Bernard 2007), which would correspond to herbivores from class sizes II and III in this study. Lion jaws are characterized by massive canines for cutting the cervical vertebras and soft tissues of large prey. In general, teeth of large felids do not leave extensive damage on bones of animals they hunt. Exceptionally, traces of canines, in the form of large holes, may appear over spinal processi of vertebra, scapula blade, and at the humero/radiocubital and femuro/tibial joint, or deep and long, sharp grooves from carnassials (Haynes 1983; Domínguez-Rodrigo 1999; Stiner 2012). Only up to 5% of the complete carcass is scored by large felid teeth, and they mostly appear along the axial skeleton, and pelvis, with the partial destruction of the femural head and tibio-femoral joint including the patella on rear limbs, and greater humeral tuber and ulna anconeus. Variability from this pattern is very small. Smaller prey, as of size class III and IV of this study, can have more damaged bone ends, but they are never completely chewed off (DomínguezRodrigo 1999; Parkinson et al., 2015). A large lion pack can overhunt, which as a consequence results in dragging the prey off to the secluded place and guarding it, and consuming it on several occasions which may last a couple of days. Towards other predators, as well as towards the individuals from other packs, lions manifest aggressive, antagonistic behaviour. They would also take over the

Among large cats the lion is the only species with developed social organization. Like other cats, a lion is a specialized carnivore – its claws and teeth are adapted to laceration of soft tissues, and its digestive enzymes cannot cope with ingestion of bone fragments. The social life of lions is organized mostly according to sex. An important stage in life of males is a period of nomadism in early adolescence, and leadership of the lion pack a couple of years after reaching maturity. However, the core of a lion pack are lionesses from the same litter. A pack usually numbers six lionesses in sister sibling line and the offspring they raise, and two male leaders. Male leaders usually come from the same litter, although there are cases of joined males from 26

Milosevic.indd 26

02/01/2020 12:47

Materials and Methods carcass from other predators, sometimes just parading beside it as a trophy. In African savannas lions take over up to 50% of the large ungulates killed by hyenas. Lion remains from Upper Pleistocene deposits, found both in hyena and cave bear dens occasionally present traces of trauma inflicted in combat with these species, some of which have healed, resulting from such experiences in the past (Rotschild and Diedriech 2012). From everything listed above it can be concluded that the lion is dominant predator at the top of the food chain in the ecosystems it inhabited. However, young lions are prone to starvation to their third year of life in habitats where competition with other large predators is high (Lehmann et al., 2008). Also, tendency to inflict mass infanticide in moments when clan leadership changes and thus higher juvenile mortality makes lions less numerous in comparison to other large carnivores in the same ecological niche (Bertram 1975). Thus, a depleted lion population in a region renews very slowly.

the Upper Pleistocene, wolves seem to thrived exactly in such ecosystems. In mosaic ecosystems wolf populations are suppresed by larger carnivores, but at higher altitudes, away from lowland ecosystems, the wolf is a dominant predator. Their subsistence strategy, combining hunting and scavenging, as well as the higher reproductive rate compared to the lion and hyena, makes wolves the master of marginal Upper Pleistocene ecosystems in Europe.

To conclude, actualistic studies of presented species gave some important clues in interpreting results of archaeozoological analysis, and the impact of their presence in past ecosystems and competition that it created. A lion is the predator at the top of the food chain and can have large pressure on ungulates of class II size, but their population is the smallest in the guild of large carnivores, and most of their activities happen in the twilight period. Lions thrive in ecosystems where herbivore populations of size classes I and II are stable, and certain level of competition with humans can be expected to occur in all habitats in which human subsistence is also tied to herbivores of that size. The hunting activities of lions always attracts other carnivores and humans. Since they do not consume bone marrow and leave small strips of meat on the carcass, it leaves the carcass which can be further processed by others. Based on subsistence strategy, the hyena is the most flexible large carnivore of the Upper Pleistocene Europe. It is expected that hyena populations were especially numerous wherever mosaic ecosystems existed, where their abilitiy to subside on various herbivore species, and kleptoparasitic behaviour towards other large carnivores, comes to the fore. In such an environment hyenas conduct slightly more pressure towards size III herbivores, and nearly equal pressure to the herbivores of other sizes, except size I. It should be added that the processing level is most advanced in hyenas because of their jaw structure adapted for marrow consuption. Compared to wolves and lions, the degree of carcass nutritional exploitation is maximal and equals that of human processing. It can be concluded that the hyena is an overwhelming factor of competition in the Upper Pleistocene mosaic ecosystems, where mass aggregation of large herbivores occurs. In such conditions they can to develop quite numerous populations and adapt through the expanding their subsistance to the various prey species, for which parts of a clan forage almost constantly. In single-habitat Upper Pleistocene ecosystems, where low diversity of large herbivores is usually observed, the hyena was probably more marginal compared to the wolf. During 27

Milosevic.indd 27

02/01/2020 12:47

4 Results This chapter presents the results of zooarchaeological analysis – mammalian and avian remains from the Palaeolithic sites Pešturina, Hadži-Prodanova, and Smolućka pećina. The results are grouped in the context of climatic cycles – Last Interglacial MIS 5, and Last Glacial MIS 4-3, as the main mechanisms driving paleoecological changes in regions in which the sites are situated.

carnivore diversity is shown on fig. 10. Similar ecological tendencies are observed in avian remains, where diversity of species is also large. Layer 4 contains two avian species inhabiting rivers and marshes, several species inhabiting open landscapes with seasonal high grass vegetation, as well as those inhabiting the Mediterranean and subMediterranean climatic belts, and deciduous or mixed forests. Among them, long-distance migratory species are also present. In general, most of birds from Pešturina inhabit the steppe, while species inhabiting broad-leaf forest regularly occur in the Last Glacial contexts of layers 3 and 2. No boreal species were identified, although they largely differ ecologically between the layers – only two of 22 species appear in more than one layer. Also, three of 22 species: corn crake, tree pipit, and golden oriole are long-distance migrants, and today migrate to sub-Saharan Africa during the wintertime in the northern hemisphere.

4.1 Pešturina Presented archaeozoological material from Pešturina was collected during systematic archaeological excavations from 2010-2013. The total number counts 42652 mammalian and 43 avian remains from all contexts. Material was sorted according to layer it comes from, and according to arbitrary excavation layers varying between 5-10 cm, which were documented by drawing, proveniencing and photographing of surfaces. During excavations, the absolute position for in situ finds was documented for 524 remains, mostly those larger than 5 cm. The average length of remains is 27 mm. The total number of secure spatial-depositional contexts from which material comes is 1506. Beside hand collecting, material was collected by sieving the context deposits on 4 and 2 mm mesh, some of which was wet sieved. The material is largely fragmented as a consequence of different biotic and abiotic taphonomic agents affecting the material during and after the deposition, among which bone corrosion and dissolution largely prevail.

During the formation of the explored layers, based on composition of ecologically sensible taxa, the main types of habitats existing in the vicinity of the cave can be established. Of large mammals these are: 1) steppe and forest steppe: mammoth, wooly rhino, bison, horses, megaloceros, 2) montane-broken: ibex, chamois, 3) mixed forest: fallow deer, roe deer, boar, while remainder of species are ecological generalists. Number of ecologically sensible taxa for different ecological niches is shown according to layers on fig. 11a. Same habitat preference is done for avian remains (fig. 11b), with their representatives:

4.1.1 Taxonomic composition Taxonomic composition from the Pešturina cave is shown on tables 3 and 4. It is characterized by a broad spectre of species, especially in layer 4. Fauna from layer 3 comprises almost the same species, but a reduction of species characteristic of wet and warm climate is apparent, such as fallow and roe deer, boar, megaloceros, brown bears, and porcupines from layer 4. A considerable drop in herbivores depending on decidous forests in layer 3 can be observed, while almost all of them are completely absent in layer 2. A proportion of hyena remains towards the rest of large mammals is high in all layers, and its nature to accumulate osteological material is taphonomically well acknowledged. Remains of horses, bison, red deer, and ibex appear in all layers in similar proportions, but with their number of remains constantly dropping (fig. 9). Layer 4 contains 96%, layer 3 contains 69%, while layer 2 contains 57% of identified taxa. The ratio of carnivore to herbivore MNI is 23:46 (50%) in layer 4, 13:18 (72%) in layer 3, and 13:12, in layer 2 from which there are more carnivore than herbivore individuals. Reciprocal Simpson’s indices for inter-layer diversity, herbiovore, and

1) rivers and marshes: Eurasian teal, golden oriole, 2) montane-broken: griffon vulture, black grouse, chough, Eurasian crag martin, 3) forest: eurasian nuthatch, Eurasian woodcock, grey headed woodpecker, Eurasian bullfinch, 4) steppe and forest steppe: common quail, grey partridge, eurasian skylark, common chaffinch, great bustard, western jackdaw, Eurasian jay. Tables 5.1, 5.2, and fig. 12, contain mathematical functions of carnivore and herbivore population densities in relation to their body mass and habitat preference, and population densities related to encounter frequencies with them in identified habitats. They reflect basic parameters on which not only human prey choice depended, but large carnivores as well, in the exploitation of given habitats. These are presented as population density on 1 km2 and elapsed time and distance when moving through it. These data show that there are differences in economic potential 28

Milosevic.indd 28

02/01/2020 12:47

Results Table 3. Taxonomic composition of mammals from different layers of Pešturina ranked according to size classes. size class

Taxa

I

layer 2 NISP

MNE

layer 3 MNI

NISP

NISP

MNE

MNI

Mammuthus sp.

2

1

1

Proboscidea indet.

7 3

2

20

11

5

28

17

Coelodonta antiquitatis

2

Rhinocerotidae indet.

1

MNE

layer 4

2

MNI

2

1

Megafauna indet. II

6

Bison priscus

3

2

Bos s. Bison

8

3

Equus ferus germanicus

9

5

2

12

6

7

1

31

12

2

99

35

6

4

2

1

6

4

2

21

5

6

28

20

3

Ursus arctos

1

1

1

Ursus sp.

2

2

1

1

2

Megaloceros giganteus Urusus spelaeus

12

Panthera spelaea

1

Ungulata indet.

1

6

1

6

3

1 2

Carnivora indet. II/III

5

1

5

1

Equus sp.

13

Cervidae indet.

1

3

4

Ruminantia indet.

8

7

13

3

29

10

62

Ungulata indet.

27

2

Carnivora indet. III

Equus hydruntinus

4

2

2

20

6

2

50

18

6

Cervus elaphus

16

7

2

34

15

2

92

43

8

1

1

1

11

9

2

49

28

4

7

4

2

91

41

11

3

3

1

Dama dama Cervidae indet.

3

3

Capra ibex

10

6

1

Sus scrofa

2

1

1

Crocuta spelaea

15

5

3

14

42

Panthera pardus Ruminantia indet.

1

7

Ungulata indet.

III/ IV

11 9

15

2

4

13 3

Mammalia indet.

6

13

39

Caprinae indet.

3

1

5

4

11

Carnivora indet.

29

Milosevic.indd 29

02/01/2020 12:47

Competition Between Humans and Large Carnivores IV

Capreolus capreolus

4

1

41

30

3

Rupicapra rupicapra

2

2

1

1

1

1

4

2

1

Canis lupus

3

1

1

6

2

1

17

9

4

Ruminantia indet.

1

1

11

Ungulata indet.

1

1

2

1

1 22

21

2

Castor fiber

4

3

1

Hystrix cristata

2

2

1

20

18

2

348

69

Mammalia indet. V

6

Lepus europaeus

6

5

1

6

6

1

Vulpes vulpes

2

2

1

4

4

1

Felis silvestris

1

1

1

1

1

1

Meles meles

1

1

3

Carnivora indet.

8

8

12

Total ident. to genus/species (I to V)

108

52

25

242

104

31

670

Mammalia indet. indet.

7019

8958

17538

Mammalia I-V/Total NISP

108/7160 (2%)

242/9255 (3%)

670/18354 (4%)

MNE/NISP

48%

43%

42,6%

Grand total

1020/34766 (3%)

Ursus spelaeus and Crocuta spelaea. Skeletal profiles for herbivore taxa are given on figs. 14-16, and appendixes 3–5, and for carnivore taxa on fig. 17 and appendix 5. In equids, representation of skeletal elements/animal units in Equus ferus and Equus sp. appears to be relatively uniform in layer 4, while head and feet elements are most numerous in layers 3 and 2. Head and feet pattern characterizes Equus hydruntinus skeletal representation in all layers. In size II bovids (Bison priscus, Bos s. Bison), head and feet are the most numerous elements in all layers. In size III cervids skeletal element composition follows a similar pattern as Equus ferus and Equus sp. Ibex, roe deer, and rabbit are characterized by more unbiased representation of elements, and their remains appear numerous enough for skeletal profiles only in layer 4. Among carnivores, hyenas and bears are also characterized by rather unbiased representation of skeletal elements, with prevailance of teeth. Patterned representation of skeletal elements of herbivore taxa is typically identified as a consequence of human processing activities, and in carnivore remains it usually marks their deposition by natural processes (Stiner 1991, 1994). Incomplete representation of anatomic regions or high representation of head and feed elements of herbivores is more characteristic but not exclusive to carnivore activities and processing (Otárola-Castillo 2010).

between identified habitats, which are reflected in the largest concentration of large mammals in steppe/forest steppe, and dispersal of animals in montane-broken, and forest habitats. Also, it can be observed that carnivore diversity is the largest in forests, but that their population density there is rather low. Wolf population density is larger in forest and montanebroken terrain, while the hyena prevails in open habitats. Large cats have lower population densities, with lions tied to steppe and forest steppe, leopards being quite versatile with unpredictable populations between the habitats, and with lynxes probably finding more suitable habitat in montane forests. Fig. 13 shows minimum time before which the probability to spot a certain species in the given habitat is null, if one moves with walking speed of 5 km/h, and it was calculated from indices given in fig. 12. Most of the species can be spotted in steppe or forest steppe for the shortest amount of elapsed time, others more often in forest or montane broken terrain, but large mammals of size class II can be encountered considerably more often in steppe or forest steppe rather than in any other identified habitats. 4.1.2 Representation of skeletal elements The presence of various skeletal elements is not numerous to establish a large sample skeletal profiles across all layers, but with results still meaningful for the following herbivore taxa: in all layers for Equus ferus germanicus, Equus sp., Equus hydruntinus, Cervus elaphus, and in layer 4 only Capra ibex, Capreolus capreolus, and Lepus europaeus, and for carnivore taxa: for layer 4 only

4.1.3 Taphonomic analysis Average specimen length according to layers is shown on fig. 18. The material is very fragmented in all of 30

Milosevic.indd 30

02/01/2020 12:47

Results Table 4. Taxonomic composition of avian remains from different layers of Pešturina (Zlatozar Boev, pers. comm.). layer 2

layer 3

layer 4

NISP

NISP

NISP

class

taxa

Anseriformes

Anas crecca

1

Galliformes

Tetrao tetrix

2

Perdix perdix

2

6

Coturnix coturnix

2

Perdicinae indet.

1

Falconiformes

Falco tinnunculus

Accipitriformes

Gyps fulvus

1

Gruiformes

Crex crex

1

Otis tarda

1

1

Charadriiformes

Scolopax rusticola

Strigiformes

Aegolius funereus

Piciiformes

Picus canus

1

Passeriformes

Alauda arvensis

1

Anthus trivialis

1

1

Ptionoprogne rupestris

1

Petronia petronia

1

Sitta europaea

1

Fringilla coelebs

1

Pyrrhula pyrrhula

1

Oriolus oriolus

1

Garrulus glandarius

2

Pyrrhocorax graculus

1

Corvus monedula

2

Aves indet. Total

1

9

the layers, with largest variations in length in layer 4, and dropping gradually towards layer 2. Among long bone shaft fragments of all size classes, longitudinal splinters prevail, which encompass up to one eight of full shaft circumference (fig. 19a – c). Presence of various breakage patterns on long bone shaft fragments are shown on table 6. Numerous shaft fragments with ragged breakage from layer 4 points that carnivore gnawing considerably influenced their fragmentation. Stepped breakage, originating from post-depostion pressure of the sediments is also numerous, even more than carnivore breakage in layer 3, while carnivore breakage is dominant in layer 4. In comparison to other breakage types, conchoidal fractures typical for hammerstone percussion are present equally in layers 4 and 3, and of somewhat higher proportion but generally lower in number in layer

2

3

1

2

5

25

2. Layer 2 exhibits almost equal long bone destruction by carnivores, humans, and post-deposition agents. Small number of shaft fragments from layer 4 contain breakages characteristic for both humans and carnivores. Basic anatomical units of the mammal remains that cannot be identified to the level of genus/species are shown on fig. 20. According to it long bones may have been originally better represented than they are identified to genus/species level. This is probably same for axial elements. Fragments of horn/antler and teeth confirm the initial high presence of heads as in all layers. Further, overall low representation of pelvis and scapula regions is striking, and in layers 3 and 2 these elements are the least represented in the anatomical units of front and rear limbs, maybe due to their difficult identification when highly fragmented. Also, such pattern confirms that bone preservation is most probably sorted 31

Milosevic.indd 31

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 9. NISP of different mammalian taxa and size classes.

Figure 10. Reciprocal Simpsonʼs index (1/D) given as total between the layers, herbivores only, and carnivores only. The closer the number to total number of taxa, the greater the diversity.

according to element mineral density, since teeth and long bone fragments are better preserved than the rest of skeletal portions.

traces of filleting appear almost equally on size II and III herbivores. Elements of horses contain filleting traces that appear to be concentrated on elements with the most meat – humerus, femur and tibia, while in size III cervids (red and fallow deer) appear on lower parts of the limbs. Among the remains from Mammalia indet. indet. category there are 16 more specimens with processing traces, 4 on long bone shaft fragment, and 28 shaft fragments with conchoidal breakage. Burning appears on 416 specimens (2%). The number of specimens burnt with different intensity is shown in table 7. Their presence points to the existance of one or several fireplaces in layer 4 which were not preserved in situ.

The relationship between the number of processing marks and MNE of large ungulates from layer 4 is shown in appendix 6. Processing marks in layer 4 appear on 51 of 816 specimens determined to various taxa that can be put in a size class category (6%), but in terms of total NISP are rather small. Out of that, traces of dismemberment appear on 9, filleting on 22, skinning on 1, and percussion on 26 specimens. It can be observed that traces of dismemberment occur more often on remains of size II herbivores, and that 32

Milosevic.indd 32

02/01/2020 12:47

Results

Figure 11. Representation of habitat sensitive avian species (a) from Pešturina, according to MNI from layer 4 (Zlatozar Boev, pers. comm.). Representation of habitat sensitive mammalian species (b) from Pešturina, according to MNI from layer 4. The most numerous in all layers are the individuals inhabiting steppe/forest steppe. MNI of mammalian species inhabiting forest gradually drops towards layer 2, while bird species inhabiting forest are present in layer 2. Species inhabiting hillymountainous ecosystem is constant. Table 5.1. Population density model for large carnivores appearing in the studied materials in different ecosystems. Average weights from: bear (Pasitschniak-Arts 2003), lion (Haas et al., 2005), hyena (Mills and Breader 2006), leopard (Kingdon et al., 2013: 159), wolf (Mech 1974). carnivores

ind. per km2 = 1.43 – 1.83(LogM taxa) – 0.34(LogM taxa)2 + 0.28(LogM taxa)3

species (weight)

hilly/ mountainous

forest

steppe/forest steppe

bear (300 kg)

-2.706

-3.073

-3.073

lion (170 kg)

-2.622

-2.622

-2.255

hyena (80 kg)

-1.748

-2.155

-1.748

leopard (60 kg)

-1.539

-1.539

-1.539

wolf (40 kg)

-1.245

-1.245

-1.612

Table 5.2. Population density model for herbivores appearing in the studied materials in different ecosystems. Average weights from: mammoth (van Geel 2008), rhino (Boeskorov 2012), horses (Bennet and Hoffman 1999), bison (Boeskorov et al., 2015), megaloceros (Moen et al., 1999), red deer (Janiszewski and Kolasa 2006), fallow deer (Feldhamer et al., 1988), ibex (Parrini et al., 2009), wild boar (Heptner et al., 1988), chamois (Garel et al., 2009), roe deer (Sempéré et al., 1996), brown hare (Puig et al., 2007).

Traces of carnivore teeth on animal remains from layer 4 are presented in similar a manner as human processing in appendix 7. They appear over 221 out of 816 (27%) specimens identified to sizeable taxonomic category. Material from Mammalia indet. indet. category contains 760 more specimens with gnawing traces and 114 digested specimens, or 6% of total NISP, compared to human processing marks which are meagre on the scale of total NISP (1%). Beside being more numerous than human traces, carnivore traces appear on a larger number of species, such are giant elk, European ass, ibex, and small animals of size V, but also on other carnivores, mostly bears. Besides, they appear over all anatomical units present, while human processing is mostly observed on long bones. Among elements, short bones are underrepresented, while among diagnostic zones of long bones the absence of epiphyses regions is apparent, which suggests they were prone to destruction by means of various taphonomic factors, in the first place carnivores. It should be pointed out that 15 specimens contain both human and carnivore processing marks, one short bone of a Proboscidean, and other mostly long bone shafts (10 out of 16) specimens of size II herbivores.

herbivores

ind. per km2 = 1.43 – 0.68(LogM taxa)

species (weight)

hilly/ mountainous

forest

steppe/forest steppe

mammoth (5 t)

-1.936

-1.936

-1.569

rhino (2 t)

-1.181

-1.181

-0.814

bison (800 kg)

-0.911

-0.911

-0.544

wild horse (800 kg)

-0.911

-0.911

-0.544

megaloceros (600 kg)

-0.826

-0.826

-0.459

hydruntinus (300 kg)

-0.621

-0.621

-0.254

red deer (200 kg)

-0.134

-0.134

-0.501

fallow deer (100 kg)

-0.297

+0.07

-0.297

ibex (80 kg)

+0.175

-0.191

-0.191

wild boar (80 kg)

-0.191

+0.175

-0.191

chamois (45 kg)

+0.305

-0.061

-0.061

roe deer (30 kg)

+0.156

+0.637

+0.156

brown hare (7 kg)

+0.488

+0.488

+0.855

33

Milosevic.indd 33

02/01/2020 12:47

Competition Between Humans and Large Carnivores

a

b

c

Figure 12. The least area in km2 necessary to cover in order to encounter different carnivore (left) and herbivore (right) species from Pešturina, if area of 1 km2 is taken as a length of its diagonal, counted as a distance travelled, in relation to the population density of those species in given ecosystems obtained on the basis of their body mass and ecological preferences. Species for which the values are positioned along the area where regression line crosses y axis are those that are most common to encounter in given ecological niche. Ecological niches: a) broken hilly-mountainous, b) mixed forest, c) steppe/forest steppe. In course of every function there is a relationship between the randomness and variation (R2) of results that model carnivore and herbivore population densities. The smaller the R2 value between the habitats, the model has sounder statistical base.

34

Milosevic.indd 34

02/01/2020 12:47

Results

Figure 13. Minimum of elapsed time (min.) until the first encounter with different carnivore and herbivore species from Pešturina in different ecological niches, if moving by diagonal of 1 km2 (1410 m), with walking speed of 5 km/h.

35

Milosevic.indd 35

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 14. Skeletal profiles of a wild horse, European ass, and remains that could not be determined beyond the horse genus (Equus sp.). Almost complete skeletal profiles characterize Equus sp. and E. ferus, while profile of E. hydruntinus is characterized by higher presence of head and feet elements in relation to the rest of the skeleton. These differences may result from different accumulation agents - wild horses by the humans, European ass by the carnivores.

36

Milosevic.indd 36

02/01/2020 12:47

Results

Figure 14, continued

37

Milosevic.indd 37

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 15. Skeletal profiles of size III cervids (Cervus elaphus i Dama dama) from Pešturina. Layer 4 contained elements of all anatomical units, while layers 3 and 2 contained mostly head and feet elements.

38

Milosevic.indd 38

02/01/2020 12:47

Results

Figure 16. Skeletal profiles of bison, roe deer, ibex, and hare in layer 4 of Pešturina. In layers 3 and 2 assemblages of these species are too small for making of skeletal profiles. Hare skeleton is mostly complete, while roe deer and ibex are more characterized by head and feet pattern.

39

Milosevic.indd 39

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 17. Skeletal profiles of the cave bear and hyena in layer 4 of Pešturina. Skull elements, especially teeth are the most numerous.

anatomical unit is shown in appendix 8a. Among remains which cannot be taxonomically attributed there is further 2 specimens with butchery traces, and 11 specimens with conchoidal breakage. Traces of burning appear on 35 specimens (< 2%). Number of specimens burnt to a different intensity is shown on table 7. Their presence indicates the use of fire in this layer, which has not been preserved as a recognizeable archaeological structure. Traces of carnivore processing on remains from layer 3 are shown in appendix 7b. Material which was not possible to identify to taxa and size contains 287 gnawed and 182 digested specimens, which comprises 6% of total NISP. As in the layer 4, carnivore marks appear on a larger number of species, mostly as those from layer 4. Contrary to layer 4, however, human and carnivore processing traces are confined to more or less the same elements and anatomical units, mostly long bones. The epiphyses of long bones are the least preserved among the diagnostic zones of long bones. On 4 specimens it is possible to observe both butchery and gnawing traces.

Figure 18. Average specimen length according to layer with variations (CI>, CI – maximum, the confidence intervals at 95%.

Figure 22. Cross section of randomly selected butchery traces and the way the outline was measured. Lines represent the modelled relation of the angle between the bone surface and bottom of the cutmark section (according to Bello and Soglio 2008).

cutmarks can be related to dismemberment, while angled ones are perhaps related more to filleting, although more experiment is needed to support such claims. The relationship between percentage of different animal units (MAU) for size II bovids, equids, and size III cervids and their food utility index (FUI) is shown according to layers on fig. 23, together with taphonomic traces of humans and carnivores on them.

mineral oxide strains, or calcium carbonate coating. The relationship between those traces is shown on table 8. Traces of manganese oxide in the form of strains appears on all specimens without exclusion. While the number of specimens containing trampling marks and weathering is not large, on the other hand, almost one third of the material is affected by the corrosive process of water dissolution. The alkaline local sedimentation environment affected the weakening of bone structure, bone leaching, and is the main reason for degradation and destruction of organic and mineral matter in osteological material.

Traces of passive taphonomic agents in this case can be divided on those which appeared when material was still exposed on the surface – trampling and weathering, and those that appeared after the material was buried beneath the sediments – corrosive changes, subsidence of

Figure 24 shows the relationship between the presence of various anatomical regions and mineral bone density of 43

Milosevic.indd 43

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 23. Relationship between MAU and FUI for a wild horse and the horse genus (FUI-according to Friesen 2001; Outram and Rowley Conwy 2008) and size III cervids (FUI-according to Madrigal-Holt 2002) together with distribution of processing marks and hyena tooth scores.

44

Milosevic.indd 44

02/01/2020 12:47

Results Table 8. Number of specimens with mechanical-chemical damage in different layers of Pešturina. layer 2

layer 3

layer 4

trampling

12

53

92

weathering

22

54

129

corrosion

2601 (36%)

2382 (25.7%)

4419 (24%)

cemented in CaCO3

18

98

42

part of the trench, where distribution of processing marks closely closely correspond to burnt specimens. Gnawed specimens and coprolites appear randomly distributed. In the Gravettian layer the distribution of specimens modified by humans and carnivores is random, and zones of activity cannot be established.

Age structures of the most represented herbivore and carnivore species in each layer are shown on fig. 25 for the layer 4 of Pešturina. These results, hovewer, should be taken with a lot of caution due to small sample sizes. In layer 4 of Pešturina large ungulates of size II may exhibited some differences in age structure between the horse (E. ferus), in which adult individuals prevail, and in bison, where age structure is close to the livinig structure. Among medium sized herbivores (size III) individuals of European ass (E. hydruntinus) seem to group around living structure, cervids are characterized by the prevailence of young individuals, while old individuals predominate in ibex. In carnivores, the cave bear is characterized by the prevalence of young individuals, while hyenas have living clan age structure. Wolf remains come only from adult individuals. Layers 3 and 2 have samples too small to argue about age structure.

The map showing topographic properties of relief and its slope, on the basis of which the isochrone lines were created is shown in appendix 10. Maps are given with radius of 30 km from the site of Pešturina as a central place to ephermal logistical-foraging visits. In topographic microregion, almost 90% comprises hilly-montainous relief, with close proximity of the highest peak of mountain Trem at 1811 m a. s. l. Also, a terrain slope above 26% is highly present. It influenced main lines of movement, both in logistical and transitional movements, although it should be noted that disposition of certain topographic barriers may have been different during the Upper Pleistocene and also susceptible to dynamics of relief change. If one assumes that in logistical settlement the base camp was somewhere at the edge or more closely than 30 km radius, it is clear that lowland steppe, or steppe-forest ecosystem stands out because it makes it possible to cover the largest distance in search for food for the shortest period of time. On the basis of isochrone contours it can be observed that it takes around 50 minutes to walk a 15 km half-radius if walking in a straight line. When combined with previously modelled least amount of time necessary to encounter preferable prey of the Neanderthals (fig. 13) in this occasion – horses and red/fallow deer, and their modelled population density, it is not before around 30 minutes of forage and 4,6 km covered in forage that these species will be spotted. Such predictions appear relatively synchronous with the model, since the search for the food is rarely in a straight line, and suggest that hunting episodes could occur relatively close to the site. Compared to lowland ecosystems, movement across the hilly-montane terrain would require around 136 minutes for the same walking distance, while movement through the woodland ecosystem is at the present moment highly variable, since the basic representation of plant species recovered from various samples at the cave are still to be studied, and type of the forest influences the model and values of raster cells.

4.1.5 Spatial distribution

4.2 The Hadži-Prodanova cave

Distribution of human and carnivore activities according to layers is shown in appendices 9a-c. On the basis of spatial distribution of human activities in layer 4, zones of activity cannot be unequivocally defined, especially while still lacking distribution of chipped stone artefacts. However, when distribution of burnt specimens is overlapped with specimens with human processing traces, they may have formed two zones, one of which is at the boundary of the L and M lines of squares, and the other in the O line of squares. It appears that the distribution of carnivore traces is random, and somewhere mixed with human traces. In layer 3 of Pešturina there is one zone of human activities which can be distinguished, and encompasses the central

Mammalian and avifaunal remains from the HadžiProdanova cave number 5793 specimens, split into three major contextual assemblages: layer 2 (Gravettian/ Epigravettian), layer 3 (without human activities), and layers 4 and 5 (Mousterian). Because of the specific sedimentation, characterized by the presence of large boulders, which sometimes covered almost the entire excavation surface, finds were collected in 1 m2 resolution, and arbitrary excavation layers varying between 5-15 cm, which were documented with drawings, level measuring and photo documentation. During excavations all sediment was sieved by 4 mm mesh. Absolute positions were taken only exceptionally – for five larger and better preserved

the elements for the three most abundant ungulate species: large bovids, horses, and medium sized cervids, according to layers from which they originate. Since the chances of the preservation of different bone elements depends on their specific mineral density, it was compared with obtained skeletal profiles and FUI for given taxa. From such approach it is possible to learn more precisely whether the represented elements were accumulated as a cosequence of human and carnivore activities concentrated on bulk, or the elements are sorted according to mineral density under the influence of various physical and chemical agents influencing the weakening of bone structure. 4.1.4 Age structures

45

Milosevic.indd 45

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 24. Relationship between MAU and BMD for: a) size II bovids, b) wild horses, c) size III cervids. BMD is presented on the basis of studies: under a) and c) according to: Kreutzer 1992 (bovids and cervids), under b) according to Lam et al., 1999.

46

Milosevic.indd 46

02/01/2020 12:47

Results

Figure 25. Age structures of best represented carnivore and herbivore species in layer 4 of Pešturina.

remains of cave bears. Studied material comes from both excavation campaigns and 1134 spatial and stratigraphic contexts, and it is very fragmented, with more than 80% comprising of specimens between 2 and 5 cm long. Rounded and partially polished specimens are numerous and appear in zones, as a consequence of trampling by cave residents and/or chemical weathering. Animal remains from pleistocene deposits of the Hadži-Prodanova cave were taphonomically analysed in the Masters thesis of the author.1 During this study, a more detailed quantification of animal remains through MNE, MAU and MNI was done, with more insight into skeletal element patterns for the most numerous species. Also, population density for herbivore and carnivore species are important for the aim of the study were established.

and there are no important ecological differences in faunal composition between the layers. Such an ecological trend is also supported by avian remains (table 11) among which the most numerous is chough, inhabiting montane cliffs. Among large fauna, remains of cave bears are by far the most numerous, and this species oftenmost frequented the cave. Ecologically sensitive species from different habitats are: 1) hilly-montainous, broken: bear, lynx, ibex, chamois, chough, 2) forest: roe deer, capercaille, 3) steppe-forest: aurochs. Among mammalian and avian remains there are no boreal species, and ecologically sensitive species presented here depend mostly on forest cover – roe deer on deciduous/ mixed forest, and capercaille on coniferous/mixed forest. If we ignore remains of cave bear, it can be observed that the spectre of other species is the smallest in layer 3, in which there was no traces of human activity. In the same layer spectre of carnivore species is largest.

4.2.1 Taxonomic composition Faunal representation from the Hadži-Prodanova cave is shown in table 10 and fig. 26. Layers 4 and 5 hold 66.6% of faunal spectrum, layer 3 holds 50% of the spectrum, and layer 2 holds again 66,6%. The relationship between carnivores and herbivores MNI is 3:8 (38%) in layers 4 and 5, 4:6 (66.6%) in layer 3, and 3:6 (50%) in layer 2. Simpson’s indices for inter-layer diversity are shown in fig. 27. Large fauna is characterized by a high presence of species inhabiting hilly-montaneous and broken terrains,

Figure 28 shows population densities of herbivores and carnivores depending on their body mass and preferred habitats, or population densities in relation to encounter rates with them in identified habitats, with the same goal as for the site of Pešturina. On the basis of species represented and identified habitats it is possible to conclude that the largest pressure by carnivores was in the forest-steppe ecosystem, and that the cave bear is the carnivore which

  Milošević, S. 2010. Taphonomic properties and accumulation agents of animal remains from the Palaeolithic site Hadži-Prodanova cave near Ivanjica. MA thesis, Faculty of Philosophy, Belgrade University. 1

47

Milosevic.indd 47

02/01/2020 12:47

Competition Between Humans and Large Carnivores Table 9. Taxonomic composition of mammals from different layers of the Hadži-Prodanova cave ranked according to size classes. size class

Taxa

II

layer 2

layer 3

NISP

MNE

MNI

Bos primigenius

2

2

1

Ursus spelaeus

156

73

5

Ursus arctos

3

2

1

Ursus sp.

21

11

Panthera leo

NISP

168

MNE

86

27

22

1

1

layer 4+5 MNI

5

IV

V

MNI

4

3

1

783

289

25

3

1

1

1

Mammalia indet.

323

351

Cervus elaphus

2

2

1

4

3

1

3

2

1

Capra ibex

19

13

1

16

8

2

29

16

3

2

2

1

12

10

1

23

18

1

4

4

1

1

1

1

1

1

1

345

36

Crocuta spelaea

III/IV

MNE

1

Ruminantia indet.

III

NISP

873

Ruminantia indet.

6

1

1

Mammalia indet.

17

1

20

Caprinae indet.

2

Mammalia indet.

681

Rupicapra rupicapra

5

5

1

Capreolus capreolus

1

1

1

Canis lupus

14

9

1

Lynx lynx

1

1

1

1 520

8

1795

7

1

Castor fiber Lepus sp.

6

5

1

Vulpes vulpes

1

1

1

Mammalia indet.

40

Total ident. to genus/species (I to V)

231

1

1

1

13 125

15

1113

20 130

12

3580

Mammalia indet. indet.

234

210

137

Mammalia I-V/Total specimens

231/1534 (15%)

227/1323 (17.2%)

863/3717 (23.2%)

MNE/NISP

49%

48%

38%

Grand total

1321/6574 (20.1%)

Table 10. Taxonomic composition of avian remains from different layers of the Hadži-Prodanova cave. (Sheila HamiltonDyer, pers. comm.). layer 2

layer 3

layer 4+5

NISP

NISP

NISP

class

taxa

Galliformes

Tetrao urogallus

1

Accipitriformes

Accipitridae indet.

1

Passeriformes

Phyrrhocorax sp.

46

9

18

48

Milosevic.indd 48

02/01/2020 12:47

Results

Figure 26. Mammal NISP, beside cave bear, from the Hadži-Prodanova cave.

Figure 27. Reciprocal Simpsonʼs index (1/D) given as total between the layers, herbivores only, and carnivores only. The closer the number to total number of taxa, the greater the diversity.

is most frequently encountered in all identified habitats. The data shows that subsistence and exploitation of local broken hilly-montane habitat is to an absolute advantage because it guarantees the lowest search times for prey and transport of the carcass.

Table 11. Specimens with processing marks from layer 2 of the Hadži-Prodanova cave according to size class/taxa. size class/taxa

element

No of cutmarks

No of impact

Ungulata III/IV

long bone shaft frag.

16

8

rib body

1

ulna

filleting

Cervus elaphus Total Ungulata III/IV

18

4.2.2 Representation of skeletal elements Skeletal profiles were made for species from which the majority of remains came and have the highest MNI: ibex, chamois, and in carnivores cave bears and wolves. They are shown on figs. 29 and 30, and appendices 11 and 12. Representation of skeletal elements in caprines is

8

49

Milosevic.indd 49

02/01/2020 12:47

Competition Between Humans and Large Carnivores

a

b

c

Figure 28. The least area in km2 necessary to cover in order to encounter different carnivore (left) and herbivore (right) species from Hadži-Prodanova, if area of 1 km2 is taken as a length of its diagonal, counted as a distance travelled, in relation to the population density of those species in given ecosystems obtained on the basis of their body mass and ecological preferences. Species for which the values are positioned along the area where regression line crosses the y axis are those that are most common to encounter in given ecological niche. Ecological niches: a) broken hilly-mountainous, b) mixed forest, c) steppe/forest steppe. In course of every function there is a relationship between the randomness and variation (R2) of results that model carnivore and herbivore population densities. The smaller the R2 value between the habitats, the model has sounder statistical base.

4.2.3 Taphonomic analysis

characterized by high presence of head and feet elements in all of the layers. Similar can be said for remains of wolves, while the cave bear is characterized by the almost equal presence of all anatomical regions. These differences may be due to a large sample size of cave bear remains in contrast to other large mammals.

Length of the specimens classified into the 5 categories is shown on fig. 31a. Material is very fragmented in all of the layers, with most of the specimens in the category between 2 and 5 cm. Average length of specimens is: layer 2 – 2,31 50

Milosevic.indd 50

02/01/2020 12:47

Figure 29. Skeletal profiles of ibex and wolves from the Hadži-Prodanova cave according to layers. Most numerous are head and feet elements.

Results

51

Milosevic.indd 51

02/01/2020 12:47

Competition Between Humans and Large Carnivores

Figure 30. Skeletal profiles of cave bears from Hadži-Prodanova according to layers. This cave bear is characterized by an almost complete skeleton.

cm, layer 3 – 2,76 cm, layer 4+5 – 4,04 cm. Representation of the material identifiable only to size class is shown according to main anatomical units for mammals of size II and III/IV on figs. 31b and 31c. It shows that small long bone shaft fragments and rib fragments prevail in all layers and suggests that these anatomical regions may have been much better represented originally than they are preserved, and that the material is at least to some extent sorted according to specific bone element density. Similarly as for Pešturina, underrepresentation of pelvis and scapula fragments is apparent.

The number of human marks according to MNE and MAU for herbivores per layer is given in tables 12 and 13. In both Gravettian and Mousterian layers, butchery traces appear on just 1% of total NISP. It is interesting to note that only traces of filleting were observed. Also, the number of specimens with hammerstone percussion marks is larger than cutmarks in large bovids and cervids. Burnt specimens are not numerous: 15 specimens in layer 2, and 7 specimens in layers 4 and 5, but they do signal possible presence of fireplaces in both Gravettian/Epigravettian and Mousterian contexts. 52

Milosevic.indd 52

02/01/2020 12:47

Results

a

Figure 31. Presence of different fragments size from the Hadži-Prodanova cave; Representation of size II specimens from Mammalia indet. category from Hadži-Prodanova, according to anatomical regions they belong; Representation of size III/IV specimens from Mammalia indet. category from the Hadži-Prodanova cave, according to anatomical regions they belong.

Comparing total NISP per layer, gnawed specimens are in similar proportions between the layers (4% in layer 2, 3% in layer 3, 2% in layers 4 and 5). The number of specimens with tooth scores is the largest on remains of size III/IV mammals in the Mousterian context, while tooth scores on size II remains are most numerous in layer 3. Distribution of gnawing marks is mostly confined to limb extremities – metapodials and phanalges (table 14), and their appearance (fig. 32) and pattern is characteristic for the wolf. Tooth scores appear on remains of all herbivore species on which cutmarks were encountered, with the exception of aurochs, and are relatively numerous, on the other hand, on cave bear elements.

Table 12. Specimens with processing marks from layers 4+5 of the Hadži-Prodanova cave according to size class/taxa. size class/taxa

element

Mammalia II

long bone shaft

Bos primigenius

humerus

filleting

humerus

filleting

metatarsus

filleting

1

long bone shaft frag.

12

6

rib body

2

Mammalia III/IV

No of cutmarks

No of impact 6

Cervus elaphus

metatarsus

Rupicapra rupicapra

radius

filleting

Capra ibex

radius

filleting

1

The relationship between the percentage of MAU and FUI for ibex are shown per layer on fig. 42. Such a positively sorted relationship in a Mousterian context probably suggests that the elements with bulk meat and marrow are relatively well represented compared to other anatomic regions of medium and low FUI, while in layers 3 and 2 this relationship is rather unbiased.

2

Total Mammalia II

3

8

Total Mammalia III/IV

16

8

Grand total

19

16

Of abiotic taphonomic traces, rounding and polishing are most numerous as the result of trampling, followed by weathering, while a substantial amount of the material, 53

Milosevic.indd 53

02/01/2020 12:47

Competition Between Humans and Large Carnivores 4.2.5 Spatial distribution

Table 13. Specimens with gnawing marks from different layers of the Hadži-Prodanova cave according to size class/ taxa. N layer 2

N gnawed specimens per anatomical unit* o

1 Mammalia II

1

Mammalia III/ IV

63

Mammalia V

2

2

3

4

5

1 11

26

Areas of carnivore and human activities are presented in appendices 13 and 14. In layers 4 and 5 Neanderthal activity is confined to the plateau in front of the cave, and towards the interior up to the line of squares 12 and 13. In layer 2, activities of modern humans are concentrated by the cliff wall to the left of the cave enterance, but also no further than the line of squares 12 and 13, suggesting mostly outdoor use of cave for animal processing by humans. Distribution of specimens with taphonomic traces of the wolf is bimodal. A larger presence of gnawed and reguritated bones is observed on the plateau in front of the cave, then drops before the cave enterance, and rises again in squares further into the entrance. Hadži-Prodanova exhibits well sorted activity zones of its inhabitants from the plateau towards the cave entrance. It is evident that most of the cave bear remains from all layers are concentrated from the part of the trench around the enterance, and that their number drops towards the plateau in front of the enterance (appendix 15). It is just the opposite with specimens containing processing traces, which were found exclusively at the plateau, away from the enterance. The distribution of specimens with trampling marks generally overlaps with those where cave bear remains are densest.

% tooth scored