The Iron Gates in Prehistory: New perspectives 9781407303734, 9781407334028

Table of contents :
Front Cover
Title Page
Copyright
Table of Contents
Acknowledgements
A Note on Dates
List of Contributors
Introduction
Section 1: REGIONAL STUDIES
Lithic technology and settlement systems of the Final Palaeolithic and Early Mesolithic in the Iron Gates
The development of the ground stone industry in the Serbian part of the Iron Gates
Sturgeon fishing in the middle and lower Danube region
The Mesolithic–Neolithic transition in the Ðerdap as evidenced by non-metric anatomical variants
Demography of the Đerdap Mesolithic–Neolithic transition
Approaches to Starčevo culture chronology
Faunal assemblages from the Early Neolithic of the central Balkans: methodological issues in the reconstruction of subsistence and land use
Section 2: SITE STUDIES
Lepenski Vir animal bones: what was left in the houses?
New-born infant burials underneath house floors at Lepenski Vir: in pursuit of contextual meanings
DNA-based sex identification of the infant remains from Lepenski Vir
Dating burial practices and architecture at Lepenski Vir
The vertebrate fauna from Hajdučka Vodenica in the Danubian Iron Gates: subsistence and taphonomy from the Early Neolithic and Mesolithic
Velesnica and the Lepenski Vir culture
The human osteological material from Velesnica
The Mesolithic–Neolithic transition in the Trieste Karst (north-eastern Italy) as seen from the excavations at the Edera Cave

Citation preview

BAR S1893 2008

The Iron Gates in Prehistory New perspectives

BONSALL, BORONEANğ & RADOVANOVIû (Eds)

Edited by

Clive Bonsall Vasile BoroneanĠ Ivana Radovanoviü

THE IRON GATES IN PREHISTORY

BAR International Series 1893 2008 B A R

The Iron Gates in Prehistory New perspectives Edited by

Clive Bonsall Vasile Boroneanţ Ivana Radovanović

BAR International Series 1893 2008

ISBN 9781407303734 paperback ISBN 9781407334028 e-format DOI https://doi.org/10.30861/9781407303734 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

Contents Acknowledgements

ii

A Note on Dating

ii

List of Contributors

iii

Introduction

1

Regional studies Dušan Mihailović

Lithic technology and settlement systems of the Final Palaeolithic and Early Mesolithic in the Iron Gates.

11

Dragana Antonović

The development of the ground stone industry in the Serbian part of the Iron Gates.

19

László Bartosiewicz, Clive Bonsall & Vasile Şişu

Sturgeon fishing along the Middle and Lower Danube.

39

Mirjana Roksandic

The Mesolithic–Neolithic in the Ðerdap as evidenced by non-metric anatomical variants.

55

Mary Jackes, Mirjana Roksandic & Christopher Meiklejohn

Demography of the Ðerdap Mesolithic–Neolithic transition.

77

Joni L. Manson

Approaches to Starčevo culture chronology.

89

Haskel Greenfield

Faunal assemblages from the Early Neolithic of the central Balkans: methodological issues in the reconstruction of subsistence and land use.

103

Vesna Dimitrijević

Lepenski Vir animal bones: what was left in the houses?

117

Sofija Stefanović & Dušan Borić

New-born infant burials underneath house floors at Lepenski Vir: in pursuit of contextual meanings.

131

Biljana Čuljković, Sofija Stefanović & Stanka Romac

DNA-based sex identification of the infant remains from Lepenski Vir.

171

Clive Bonsall, Ivana Radovanović, Mirjana Roksandic, Gordon Cook, Thomas Higham & Catriona Pickard

Dating burials and architecture at Lepenski Vir.

175

Haskel Greenfield

Reanalysis of the vertebrate fauna from Hajdučka Vodenica in the Danubian Iron Gates: subsistence and taphonomy from the Early Neolithic and Mesolithic.

205

Rastko Vasić

Velesnica and the Lepenski Vir culture.

227

Mirjana Roksandic

The human osteological material from Velesnica.

243

Paolo Biagi, Elisabetta Starnini & Barbara Voytek

The Mesolithic–Neolithic transition in the Trieste Karst (north-eastern Italy) as seen from the excavations at the Edera Cave.

251

Site studies

i

Acknowledgements Many individuals and organizations have helped directly or indirectly in the preparation of this volume. We thank all the contributors for their cooperation and forbearance throughout the long production process. Each of the papers has been reviewed by at least two independent referees, and we have many colleagues to thank for their expert opinions on the manuscripts. The camera-ready copy was prepared using David Pilling’s excellent desktop publishing software, Ovation Pro for Windows. Financial support for the Edinburgh symposium was provided by the British Academy and The University of Edinburgh Development Fund, while practical help at the meeting was provided by Sarah Davison, Brian Dean, Elizabeth Hargreaves, Ian Morrison, Mary Jane Nicholls, Ishbel Ogilvie, Pat Storey, Diana Şişu, and Kirsten Thompson.

A Note on Dates Throughout this volume dates followed by ‘BP’ are in radiocarbon years before present. Dates followed by ‘cal BC’ are calibrated radiocarbon ages (calendar ages). Dates followed simply by ‘BC’ are calendar ages derived by some other method, e.g. luminescence dating. The following table shows the approximate correspondence between uncalibrated and calibrated 14C ages over the time-range from 2000 to 13,900 BP. The calibrations were performed with CALIB 5.0.2 (Stuiver & Reimer 1993; Stuiver et al. 2005) using the IntCal04 curve (Reimer et al. 2004) and rounded to the nearest five years.

Radiocarbon date calibration table BP

cal BC

BP

cal BC

BP

cal BC

BP

cal BC

2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000 3100 3200 3300 3400 3500 3600 3700 3800 3900 4000 4100 4200 4300 4400 4500 4600 4700 4800 4900

0 125 295 390 460 625 795 835 955 1085 1255 1395 1470 1570 1715 1820 1950 2085 2250 2405 2535 2645 2795 2905 3015 3215 3365 3420 3550 3675

5000 5100 5200 5300 5400 5500 5600 5700 5800 5900 6000 6100 6200 6300 6400 6500 6600 6700 6800 6900 7000 7100 7200 7300 7400 7500 7600 7700 7800 7900

3775 3850 3995 4130 4295 4345 4410 4525 4660 4760 4890 5020 5130 5265 5375 5475 5540 5625 5690 5765 5895 5995 6050 6155 6295 6400 6450 6530 6630 6735

8000 8100 8200 8300 8400 8500 8600 8700 8800 8900 9000 9100 9200 9300 9400 9500 9600 9700 9800 9900 10,000 10,100 10,200 10,300 10,400 10,500 10,600 10,700 10,800 10,900

6930 7065 7230 7400 7500 7560 7595 7685 7875 8075 8250 8290 8395 8565 8680 8790 8965 9220 9270 9320 9515 9765 9950 10,130 10,330 10,570 10,740 10,835 10,880 10,920

11,000 11,100 11,200 11,300 11,400 11,500 11,600 11,700 11,800 11,900 12,000 12,100 12,200 12,300 12,400 12,500 12,600 12,700 12,800 12,900 13,000 13,100 13,200 13,300 13,400 13,500 13,600 13,700 13,800 13,900

10,965 11,060 11,160 11,240 11,310 11,390 11,475 11,600 11,735 11,815 11,900 12,005 12,110 12,210 12,430 12,705 12,915 13,055 13,160 13,280 13,395 13,535 13,680 13,830 13,970 14,105 14,240 14,370 14,490 14,735

References Reimer P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., et al., 2004: IntCal04 terrestrial radiocarbon age calibration, 0–26 cal kyr BP. Radiocarbon 46: 1029–1058. Stuiver, M. & Reimer, P.J. 1993: Extended 14C data base and revised CALIB 3.0 14C age calibration program. Radiocarbon 35: 215–230. Stuiver, M., Reimer, P.J. & Reimer, R. 2005: CALIB Radiocarbon Calibration (rev. 5.0.2): Online Manual. http://radiocarbon.pa.qub.ac.uk/calib/manual/

ii

List of Contributors Dragana Antonović

Institute of Archaeology, Knez Mihailova 35/IV, 11000 Belgrade, Serbia. Email: [email protected]

László Bartosiewicz

Institute of Archaeological Sciences, Loránd Eötvös University, Múseum körút 4/B, H-1088 Budapest, Hungary. Email: [email protected] Dipartimento di Scienze dell’Antichità e del Vicino Oriente, Università Ca’ Foscari Venezia, Palazzo Malcanton Marcorà, Dorsoduro 3484/D, I-30123 Venezia, Italy. Email: [email protected]

Paolo Biagi

Clive Bonsall Dušan Borić Vasile Boroneanţ Biljana Čuljković Gordon Cook Vesna Dimitrijević Haskel Greenfield Thomas Higham

Mary Jackes Joni Manson Christopher Meiklejohn Dušan Mihailović Catriona Pickard Ivana Radovanović Mirjana Roksandic Stanka Romac Vasile Şişu Elisabetta Starnini

Sofija Stefanović Rastko Vasić Barbara Voytek

School of History, Classics and Archaeology, University of Edinburgh, Old High School, Infirmary Street, Edinburgh, EH1 1LT, U.K. Email: [email protected] Department of Archaeology, University of Cambridge, Downing Street, Cambridge, CB2 3DZ, U.K. Email: [email protected] Bucharest Museum of History and Art, Bd I.C. Bratianu no. 2, sector 3, 70058 Bucharest, Romania. Email: [email protected] Institute for Research in Immunology and Cancer, P.O. Box 6128, Station Centre-Ville, Montreal, Quebec, H3C 3J7, Canada. Email: [email protected] Radiocarbon Laboratory, S.U.E.R.C., Scottish Enterprise Technology Park, Rankine Avenue, East Kilbride, G75 0QF, U.K. Email: [email protected] Department of Archaeology, Faculty of Philosophy, University of Belgrade, Čika Ljubina 18–20, 11000 Belgrade, Serbia. Email: [email protected] Department of Anthropology, University of Manitoba, Fletcher Augue 435, Winnipeg, Manitoba, R3T 5V5, Canada. Email: [email protected] Oxford Radiocarbon Accelerator Unit, Research Laboratory for Archaeology and the History of Art, Dyson Perrins Building, University of Oxford, Oxford, OX1 3QY, U.K. Email: [email protected] Department of Anthropology, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada. Email: [email protected] Heritage Education and Research Services, 13269 Bevelheimer Road, Westerville, OH 43081, U.S.A. Email: [email protected] Department of Anthropology, University of Winnipeg, Winnipeg, Manitoba, R3B 2E9, Canada. Email: [email protected] Department of Archaeology, Faculty of Philosophy, University of Belgrade, Čika Ljubina 18-20, 11000 Belgrade, Serbia. Email: [email protected] School of History, Classics and Archaeology, University of Edinburgh, Old High School, Infirmary Street, Edinburgh, EH1 1LT, U.K. Email: [email protected] Department of Archaeology, University of Kansas, 616 Fraser Hall, 1415 Jayhawk Boulevard, Lawrence, KS 66045, U.S.A. Email: [email protected] Department of Anthropology, University of Winnipeg, Winnipeg, Manitoba, R3B 2E9, Canada. Email: [email protected] Faculty of Biology, University of Belgrade, Studentski trg 16, P.O. Box 52, Belgrade 11000, Serbia. Email: [email protected] Muzeul Regiunii Porţilor de fier, Strada Independentei Nr 2, Drobeta Turnu-Severin 1500, Mehedinţi, Romania. Email: [email protected] Dipartimento di Scienze dell’Antichità e del Vicino Oriente, Università Ca’ Foscari Venezia, Palazzo Malcanton Marcorà, Dorsoduro 3484/D, I-30123 Venezia, Italy. Email: [email protected] Department of Archaeology, Faculty of Philosophy, University of Belgrade, Čika Ljubina 18-20, 11000 Belgrade, Serbia. Email: [email protected] Institute of Archaeology, Knez Mihailova 35/IV, 11000 Belgrade, Serbia. Email: [email protected] Archaeological Research Facility, 2251 College Building, University of California, Berkeley, CA 94720-1076, U.S.A. Email: [email protected] iii

Introduction Clive Bonsall, Vasile Boroneanţ & Ivana Radovanović

This book had its origins in a symposium held at the University of Edinburgh from 30 March to 2 April 2000, which was attended by 54 archaeologists with a shared interest in the prehistory of the small but distinctive region of Southeast Europe known as the Iron Gates (Fig. 1). It is by no means a complete compilation of the papers delivered at the symposium. Some contributions dealt with aspects of the Romanian–British excavations at Schela Cladovei, and will form the basis of a separate publication. Moreover, of the 15 articles included in the present volume, only ten are based on papers presented at the Edinburgh meeting, while five — those by Dragana Antonović, Vesna Dimitrijević, Joni Manson, László Bartosiewicz et al., and Clive Bonsall et al. — are later additions. As a geographical label, the term ‘Iron Gates’ has acquired several different meanings.1 In the broad sense it refers to the section of the Danube valley where the river forms the modern political border between Serbia and Romania, and it is this definition we have adopted for the present volume. The 230 km long Iron Gates section marks the beginning of the Lower Danube. It divides naturally into two linear zones —

the Iron Gates ‘gorge’ (Fig. 2) and the ‘downstream area’ (Fig. 3). The gorge, really a series of narrow, steep-sided canyons separated by sections where the valley is wider and the sides less steep, cuts a winding course for 134 km between the southern Carpathian Mountains and the northwestern foothills of the Balkan Mountains. In Serbia the gorge is known as Đerdap (from the Turkish ‘girdap’ meaning whirlpool — literally, ‘gorge of whirlpools’) reflecting the former presence of several sets of rapids and the generally turbulent flow of the river along this stretch of the Danube. In contrast, the downstream area is a zone of much more moderate relief, marking the beginning of the Danube’s journey across the Wallachian Plain toward the Black Sea. Here the river is flanked by a broad alluvial plain comprising a series of terraces. Notwithstanding the differences in topography, microclimates and natural vegetation between the two zones, the ‘gorge’ and the ‘downstream area’ have many features in common archaeologically. The Danube is both an important route way and a natural barrier, and historically the Iron Gates has had great strategic importance. The region is exceptionally rich in sites dating to

Figure 1. Principal Mesolithic and Early Neolithic sites of the Iron Gates.

1

The Iron Gates in Prehistory

the Roman and later periods, especially military installations. Roman achievements in the Iron Gates include forts, a canal built to bypass the rapids at the lower end of the gorge, and Trajan’s Bridge constructed in AD 103–105 to supply Roman legions during the Second Dacian War. While there is a long tradition of Roman archaeology in the Iron Gates extending back more than a century, the prehistory of the region was virtually unknown until the second half of the twentieth century when the first systematic archaeological surveys were undertaken in advance of the impounding of the Danube by the Iron Gates I and II dams. The survey work and rescue excavations were concentrated in the parts of the valley that would eventually be submerged beneath the reservoirs created by the two dams — essentially a narrow zone along both banks of the Danube as well as some low-lying islands in the river. The fieldwork progressed in two phases. The first between 1960 and 1971 focused on the areas affected by the Iron Gates I dam; these were very largely within the gorge, although some work was undertaken in the immediately downstream area at Ostrovul Banului and Schela Cladovei. In the second phase between 1977 and 1984 fieldwork was concentrated in the downstream area between the Iron Gates I and II dams. Altogether, over two hundred archaeological sites of various periods were discovered. Among them were a number of Mesolithic sites, the first to be identified in the entire central Balkan region (Fig. 1). Excavations were undertaken at many of these sites — seven on the Serbian bank (Padina, Stubica, Lepenski Vir, Vlasac, Hajdučka Vodenica, Velesnica, and Kula) and 16 on the Romanian side (Privod, Alibeg, Vodneac, Ilişova, Cuina Turcului, Climente I and II, Veterani Cave, Veterani Terrace, Răzvrata, Icoana, Ostrovul Banului, Schela Cladovei, Ostrovul Corbului, and Ostrovul Mare km 875 and km 873). Some were major projects extending over several field seasons, but many were small, exploratory excavations lasting just a few days or weeks, and some of these were never completed, being abandoned when the level of the Danube became too high. The published record of these excavations is variable. Comprehensive reports are available for just two sites — Vlasac (Srejović & Letica 1978) and Cuina Turcului (Păunescu 1970, 1978). But there are many other publications that offer either interim excavation reports or studies of particular bodies of material, especially from Lepenski Vir, Padina, Hajdučka Vodenica, and Schela Cladovei. There have also been several attempts to provide regional syntheses, most notably by Radovanović (1996), Boroneanţ (2000), and Bonsall (2008). Although a number of important Bronze and Iron Age sites were also discovered, it is the Mesolithic that has tended to dominate the literature on the Iron Gates in Prehistory — and it is a recurrent theme throughout the present volume. The reasons are not hard to identify; the range and quality of the evidence from the Iron Gates relating to Mesolithic architecture, art, burial practices, bone and stone technology, and subsistence patterns, far exceeds that from any other region of Southeast or Central Europe west of the Black Sea. Moreover, some Mesolithic sites continued to be used into the Early Neolithic, making the Iron Gates an area of critical importance for understanding the processes involved in the transition from hunting and gathering to farming in Europe.

Since the completion of the dam projects (and the consequent ‘loss’ of many of the valley floor sites recorded in the 1960s to 1980s) the emphasis in research has shifted, although the focus has remained firmly on the Mesolithic and the Neolithic transition. While some fieldwork has continued along the banks of the Danube, notably at Schela Cladovei in Romania (Boroneanţ et al. 1999; Bonsall 2008) and Vlasac in Serbia (Borić 2007), many publications have been based on re-analyses of existing archaeological collections, taking advantage of advances in archeological science such as AMS 14C dating (e.g. Cook et al. 2002; Bonsall et al. 2004; Borić & Miracle 2004; Borić & Dimitrijević 2007), stable isotope analysis (e.g. Bonsall et al. 1997, 2000; Grupe et al. 2003), and techniques for the recovery of DNA from animal and human bones (e.g. Larson et al. 2007). The data generated by these new analyses in turn have permitted archaeologists to reassess the models of prehistoric settlement that were proposed on the basis of initial field observations. The papers that make up this volume exemplify both strands of this new phase of research.

Objectives and organization of the volume The objectives of this volume are broadly those of the original Edinburgh symposium. First and foremost the volume is intended to illustrate the immense research potential of the Iron Gates region, even though those areas along the Danube’s banks that proved attractive for prehistoric settlement are now largely flooded and there has been comparatively little additional fieldwork undertaken since the mid1980s. A second objective is to provide case studies that illustrate the nature of current research and the rich possibilities offered by the growing range of scientific techniques available to archaeologists and their application to existing archaeological collections. When editing the volume it soon became evident that there was considerable overlap between the papers in terms of chronological and/or geographical coverage, or theoretical and/or methodological emphasis. Therefore, we decided on a basic division between those papers that include an element of regional synthesis, and those that deal largely or exclusively with the evidence from an individual site: Regional studies This section begins with the paper by Dušan Mihailović, which focuses on the Late Glacial (Final Palaeolithic) and the Early Holocene (Early Mesolithic) in the Iron Gates. Mihailović identifies clear trends in settlement pattern, subsistence and technology during this time range, including a reduction in the size of annual territories, increased use of local resources, and a decline in the range and quality of chipped stone artefacts. He argues that these perceived trends reflect a reduction in residential mobility in response to changing environmental conditions and resource availability. While the technological emphasis in Dusan Mihailović’s paper is on chipped stone artefacts, Dragana Antonović reviews the evidence for a ground stone industry in the Iron Gates and its development during the Mesolithic and Early Neolithic. Assigned to the category of ‘ground stone arte2

Clive Bonsall, Vasile Boroneanţ & Ivana Radovanović: Introduction

Figure 2. Two views of the Iron Gates gorge. Top: The gorge near the sites of Hajdučka Vodenica (Serbia) and Icoana (Romania); the photograph was taken from the Serbian bank looking upriver (© Mirjana Roksandic, 2000). Bottom: The gorge photographed from Lepenski Vir, looking upriver (© Ivana Radovanović, 2008).

facts’ are “all stone implements worked by grinding/polishing, as well as unfinished examples with traces of flaking or pecking … [and] … implements that are naturally polished or polished through use” (Antonović, this volume: 19). Her analysis is restricted to sites on the Serbian bank of the Danube — proceeding downstream, Padina, Lepenski Vir, Vlasac, Hajdučka Vodenica, Ajmana,

Velesnica, Knjepište, and Ušće Kameničkog Potoka. Within the ground stone industry she recognizes two major components, local and ‘imported’. These are distinguishable not only in terms of typology and technique, but also raw material choices and ascribed function. The local component is seen to have developed independently within the Iron Gates region and comprises a wide range of tool forms including 3

The Iron Gates in Prehistory

Figure 3. Two views of the ‘downstream area’ of the Iron Gates. Top: The Danube floodplain at Schela Cladovei (Romania), 8 km downstream from the Iron Gates I dam. The Schela Cladovei archaeological site extends for about a kilometre along the riverbank. The town of Kladovo in Serbia can be seen on the opposite bank of the Danube (© Clive Bonsall, 2008). Bottom: The confluence of the Zamna river with the Danube on the Serbian side, 8 km upstream from the Iron Gates II dam. Opposite the confluence lies Ostrovul Mare (The Big Island). Several Mesolithic sites were found in the vicinity — cf. Figure 3 (© Ivana Radovanović, 2008).

4

Clive Bonsall, Vasile Boroneanţ & Ivana Radovanović: Introduction

various kinds of grinding tools, percussive instruments, weights, and ornamental artefacts, invariably made from locally available raw materials. The ‘imported’ component comprises mainly ground edge implements (axes, adzes, and chisels), which have clear parallels in the Starčevo and Vinča cultures of the central Balkans, and the materials are both local and imported. This, and the fact that both the local and ‘imported’ elements are found together in some sites, leads Antonović to the conclusion that the ‘imported’ element was an adopted technology. Her paper is both an important contribution to knowledge of the Iron Gates in Prehistory, and a valuable demonstration that ground stone tools (in the broad sense of non-chipped stone tools) were an important, and often abundant, component of Mesolithic and Neolithic sites throughout the Iron Gates. This is equally true of the Romanian bank of the Danube, although there they appear not to have been collected so systematically. Somewhat broader in its chronological and geographical scope is the paper by László Bartosiewicz, Clive Bonsall & Vasile Şişu, which provides an overview of the zooarchaeological and ethnohistorical evidence for sturgeon fishing along the middle and lower Danube as represented by two contrasting sections of the river, in the Carpathian Basin and the Iron Gates. The exploitation of migratory sturgeon is documented from the Late Glacial onwards, although interpretation of the osteoarchaeological evidence is rarely straightforward. The authors attempt to provide a framework for future investigations, highlighting the many variables that archaeologists need to consider when seeking to interpret the role of sturgeon fishing in the lives of prehistoric peoples living along the Danube. These include sturgeon biology and behaviour, taphonomy, river conditions, fishing techniques, as well as local beliefs and customs. Two papers focus on the effects of the Mesolithic–Neolithic transition in the Iron Gates region, and specifically the issue of population change. Both use data gathered by Mirjana Roksandic for her PhD study of the skeletal populations from four sites in the Serbian part of the Iron Gates — Hajdučka Vodenica, Lepenski Vir, Padina, and Vlasac (Roksandic 1999). Mirjana Roksandic’s first paper in this volume examines the incidence of selected skeletal non-metric traits among the populations from different time periods and, based on the assumption that the traits are linked to ancestry, uses the data as a measure of the degree of interaction between local foragers and non-local farmers.2 The 259 adult skeletons examined are divided into three chronological groups (Mesolithic, Mesolithic–Neolithic ‘Contact’ period, and Neolithic), and the groups compared. The results are interpreted as reflecting a substantial degree of population continuity between Mesolithic and Neolithic. It is suggested there may have been some ‘seeping in’ of immigrants during the contact period before the establishment of a farming economy in the Iron Gates, but evidence of population replacement at the transition is lacking. From this the author concludes that the Iron Gates Neolithic was primarily the result of local adoption. The paper by Mary Jackes, Mirjana Roksandic & Christopher Meiklejohn applies standardized palaeodemographic methods of analysis to Roksandic’s age/sex data for skeletal populations from the Iron Gates, in order to test for changes in fertility across the

Mesolithic–Neolithic transition. Whilst stressing that their interpretations are severely constrained by inadequate sample sizes and uncertainty over the dating of many skeletons, the authors offer the tentative conclusion that population levels in the Iron Gates were stable during the Mesolithic with an increase in fertility in the Neolithic, which is consistent with findings elsewhere in Europe. The last two papers in this section are concerned with specific aspects of the Early Neolithic in the wider region of the central–northern Balkans. Joni Manson discusses the chronology of the Starčevo culture. Several chronological frameworks are in current use, all of which rely on pottery typology. But none of these has been adequately tested by independent dating procedures. Most attempts to establish the typological sequence of stylistic change in Starčevo ceramics have relied on stratigraphy and/or 14C dating of associated organic materials. A complementary approach would be to use the pottery itself to obtain direct age measurements, since there are more assumptions involved in 14C dating of material found near pottery than in trying to date the pottery directly (Bonsall et al. 2002). In a pioneering study based on her PhD research (Manson 1990), Dr Manson describes the results of archaeomagnetic intensity dating of potsherds from 12 Starčevo sites. By combining these data with Aranđelović-Garašanin’s (1954) typological sequence and the available radiocarbon dates for Starčevo sites, she is able to propose a revised chronology of the Starčevo culture. In the following paper, Haskel Greenfield reviews the state of archaeozoological research in the central Balkans pertaining to the Early Neolithic. He points out that there are relatively few Early Neolithic sites with analyzed faunal assemblages, and the quality of the information from those sites is highly variable. Greenfield’s discussion ranges over issues such as the effects of site location, sample size and taxonomic diversity, taphonomy, recovery methodology, quantification methods, curation, and state of publication, leading to the conclusion that the majority of faunal assemblages from the region have only limited potential for reconstructing animal exploitation strategies and land use patterns in the Early Neolithic. Greenfield’s paper serves as both a warning of the fragility of conclusions based on existing archaeozoological reports, and a plea for a more rigorous approach to the analysis of Early Neolithic animal bone assemblages from Southeast Europe. Site studies Of the six papers that deal with individual sites, four focus on Lepenski Vir, arguably the most famous archaeological site in the Iron Gates region. All four papers present the results of new analyses of the finds from Dragoslav Srejović’s excavation between 1965 and 1971. Not all of the finds survive from those excavations and this applies particularly to the animal remains, which in any case were mainly hand-collected (relatively little sieving was undertaken at Lepenski Vir). Vesna Dimitrijević focuses on the small surviving collection of animal bones stored in the National Museum in Belgrade, and specifically those recovered from within the famous trapezoidal buildings. In a thought-provoking analysis she distinguishes between remains deposited while a building was in use, and those deposited when it was abandoned. In so 5

The Iron Gates in Prehistory

doing she identifies episodes of food preparation, bone manufacturing, and ritual deposition. Consideration is also given to the question of seasonality of building abandonment, although the evidence is limited. The other papers focus on Lepenski Vir are concerned primarily with the human remains from the site. An unusual feature of Lepenski Vir was the occurrence of the remains of neonates underneath the floors of the trapezoidal buildings or immediately to the rear of the buildings. The papers by Sofija Stefanović & Dusan Borić and Biljana Čjulković, Sofija Stefanović & Stanka Romac are both directly concerned with these remains. Stefanović & Borić provide a detailed osteoarchaeological evaluation of the burials, including a discussion of taphonomy, supported by plans and photographs many of which have not been published previously. This leads on to a wide-ranging and thought-provoking discussion of the possible chronological and social significance of the neonate burials, which draws inspiration from archaeological, ethnographic, and ethnohistorical sources. Knowledge of the sex of the infants is fundamental to the interpretation of the burial evidence. However, determining the sex of neonate skeletons using standard osteometric procedures is notoriously unreliable. An alternative approach is to use ancient (aDNA) techniques. Čuljković et al. have done just this. Of the 41 infant skeletons excavated from under buildings, 30 individuals were tested and successful sex identification is reported in nearly all cases. These results have important implications for the related study by Stefanović & Borić, and thus the two papers should be considered together. Although by no means the first successful amplification of aDNA from ancient human remains, Čuljković et al.’s paper stands as a pioneering study for the Iron Gates and demonstrates the potential of the Iron Gates sites for archaeogenetic research. The burial record from Lepenski Vir is not, of course, confined to neonates; there were also many burials of older children, adolescents, and adults, and these are the subject of the paper by Clive Bonsall, Ivana Radovanović, Mirjana Roksandic, Gordon Cook, Thomas Higham & Catriona Pickard. AMS 14 C dates and stable isotope values are presented for 24 burials. The data are then used to establish the chronological contexts of different forms of burial represented at Lepenski Vir, and to refine the dating of the trapezoidal buildings where these occur in a clear stratigraphic relationship with directly dated burials. The authors suggest that their data also have implications for dating the appearance of farming in the Lepenski Vir catchment, although they acknowledge that this is a controversial subject, which is further complicated by curatorial problems and a lack precision in the 14C dates. Three other papers in this section are concerned with sites downriver of Lepenski Vir. Haskel Greenfield provides an account of the (hitherto unpublished) animal bone assemblages from Late Mesolithic and Early Neolithic contexts at the site of Hajdučka Vodenica in the gorge, excavated by Borislav Jovanović between 1966 and 1969. In interpreting the results, Greenfield takes into account excavation methodology, uncertainties over stratigraphy, and the extent of taphonomic loss, in keeping with the approach advocated in his other contribution to this volume (pp. 103–114).

Comparisons are made with other sites in the Iron Gates gorge. A key finding is that bones of apparently domestic cattle and pig occur in both the Mesolithic and Neolithic samples which begs the important question, is this evidence of Mesolithic animal domestication, exchange with farmers, or ‘stratigraphic mixing’? The important site of Velesnica in the downstream area is the focus of separate but related papers by Rastko Vasić and Mirjana Roksandic. Excavated by Rastko Vasić between 1980 and 1984, Velesnica is a multi-period open-air site on the Serbian bank of the Danube (Fig. 1). Vasić reviews the evidence from the main settlement phases. The most abundant remains belong to the Early Neolithic (Starčevo culture) and include three graves, one of which contained seven skeletons; the burials are discussed in more detail in the companion paper by Mirjana Roksandic. Vasić draws attention to several features of the Starčevo settlement at Velesnica that he believes demonstrate a cultural connection with Lepenski Vir and other Late Mesolithic/Early Neolithic sites in the gorge, including the presence of carved stone ‘altars’. An earlier, Mesolithic occupation is suspected at Velesnica but could not be distinguished stratigraphically and has yet to be confirmed by radiocarbon dating. The final paper in the volume, by Paolo Biagi, Elisabetta Starnini & Barbara Voytek, stands apart in that it does not deal with the Iron Gates per se. Their article reviews the evidence from Edera Cave on the karst plateau at the head of the Adriatic Sea, which has a long occupation sequence including Mesolithic and Neolithic deposits. It is included here because it touches on the question of Mesolithic ‘survival’ and contact between hunter-gatherers and farmers — still a major point of controversy among prehistorians working in the Iron Gates, with differing views aired by, e.g., Voytek & Tringham (1989), Radovanović & Voytek (1997), Tringham (2000), Borić (2002), Radovanović (2006), and Bonsall (2007, 2008). The co-occurrence of pottery, wild and domestic fauna, and chipped stone artefacts of Mesolithic character in ‘layer 3a’ at Edera is reminiscent of the situation described for some sites in the Iron Gates and evokes similar questions regarding the nature and timing of the transition from Mesolithic to Neolithic. This volume presents new information and new perspectives on the prehistoric settlement of the Iron Gates. We also hope that it points out directions for future research. Many areas of uncertainty and controversy remain, and we look forward to further investments in AMS radiocarbon dating, isotopic analysis, and aDNA research in the future to help resolve some of these issues. In the years ahead we would also hope to see an extension of fieldwork into the hinterland on both sides of the Danube, beyond the areas affected by dam construction, in order to achieve a more representative picture of the human use of the Iron Gates in Prehistory. Notes 1. The original ‘Iron Gate’ of the Danube was the cliffs on either side of the rapids at the lower end of the gorge, where subsequently the Iron Gates I dam was built. In the Danube Convention of 1948 the term ‘Iron Gates’ was formally applied to “the section between Vince and Kostol on the right bank and between Moldova-Veche and Turnu-Severin on the left bank of the Danube”, which in-

6

Clive Bonsall, Vasile Boroneanţ & Ivana Radovanović: Introduction cludes most of the gorge and the first 8 km of the downstream area. A second dam was constructed 80 km downstream from Iron Gates I and given the official name ‘Iron Gates II’, thus effectively extending the term ‘Iron Gates’ to more or less the whole of the border region between Serbia and Romania. 2. The term ‘non-metric trait’, or ‘discrete trait’, refers to any minor (non-pathological) anomaly observed in bones and teeth that is not normally recorded by measurement. Non-metric traits may be recorded as being either present or absent or, less often, scored according to their degree of development (Mays 1998: 102).

In Thévenin, A. (ed.) L’Europe des Derniers Chasseurs: Épipaléolithique et Mésolithique. Paris: Éditions du Comité des Travaux Historiques et Scientifiques, 385–390. Cook, G.T., Bonsall, C., Hedges, R.E.M., McSweeney, K., Boroneanţ, V., Bartosiewicz, L. & Pettitt, P.B. 2002: Problems of dating human bones from the Iron Gates. Antiquity 76: 77–85. Grupe, G., Mikić, Ž., Peters, J. & Manhart, H. 2003: Vertebrate food webs and subsistence strategies of Meso- and Neolithic populations of central Europe. In Grupe, G. & Peters, J. (eds) Decyphering Ancient Bones: The Research Potential of Bioarchaeological Collections. Rahden (Westf): Marie Leidorf, 193–214. Larson, G., Albarella, U., Dobney, K., Rowley-Conwy, P., Schibler, J., Tresset, A., Vigne, J-D., Edwards, C.J., Schlumbaum, A., Dinu, A., Balacsescu, A., Dolman, G., Tagliacozzo, A., Manaseryan, N., Miracle, P., Van Wijngaarden-Bakker, L., Masseti, M., Bradley, D.G. & Cooper, A. 2007: Ancient DNA, pig domestication, and the spread of the Neolithic into Europe. Proceedings of the National Academy of Sciences of the United States of America 104(39): 15276–15281. Manson, J.L. 1990: A Reanalysis of Starčevo Culture Ceramics: Implications for Neolithic Development in the Balkans. Unpublished PhD dissertation, Southern Illinois University, Carbondale (University Microfilms, Ann Arbor). Mays, S. 1998: The Archaeology of Human Bones. London: Routledge. Păunescu, A., 1970. Epipaleoliticul de la Cuina Turcului-Dubova. Studii şi Cercetări de Istorie Veche 21(1): 3–47. — 1978: Cercetările archeologice de la Cuina Turcului-Dubova (Jud. Mehedinţi). Tibiscus Istorie, Volum Închinat Celei de-a 60 Aniversari a Unirii: 11–56. Radovanović, I. 1996: The Iron Gates Mesolithic. Ann Arbor: International Monographs in Prehistory. — 2006: Further notes on Mesolithic–Neolithic contacts in the Iron Gates region and the central Balkans. Documenta Praehistorica 33: 107–124. Radovanović, I. & Voytek, B. 1997: Hunters, fishers and farmers: sedentism, subsistence and social complexity in the Iron Gates Mesolithic. Analecta Praehistorica Leidensia 29: 19–31. Roksandic, M. 1999: Transition from Mesolithic to Neolithic in the Iron Gates Gorge: Physical Anthropology Perspective. Unpublished PhD thesis, Department of Anthropology, Simon Fraser University, Burnaby, B.C. Srejović, D. & Letica, Z. 1978: Vlasac: A Mesolithic Settlement in the Iron Gates. Vol. 1, Archaeology. Belgrade: Serbian Academy of Arts and Sciences. Tringham, R. 2000: Southeastern Europe in the transition to agriculture in Europe: bridge, buffer, or mosaic. In Price, T.D. (ed.) Europe’s First Farmers. Cambridge: Cambridge University Press, 19–56. Voytek, B. & Tringham, R. 1989: Rethinking the Mesolithic: the case of South-east Europe. In Bonsall, C. (ed.) The Mesolithic in Europe. Edinburgh: John Donald, 492–500.

References Aranđelović-Garašanin, D. 1954: Starčevačka kultura. Ljubljana. Bonsall, C. 2007: When was the Neolithic transition in the Iron Gates? In Spataro, M. & Biagi, P. (eds), A Short Walk through the Balkans: the First Farmers of the Carpathian Basin and Adjacent Regions. Trieste: Società per la Preistoria e Protostoria della Regione Friuli-Venezia Giulia, 53–65. — 2008: The Mesolithic of the Iron Gates. In Bailey, G. & Spikins, P. (eds), Mesolithic Europe. Cambridge: Cambridge University Press, 238–279. Bonsall, C., Lennon, R., McSweeney, K., Stewart, C., Harkness, D., Boroneanţ, V., Bartosiewicz, L., Payton, R. & Chapman, J. 1997: Mesolithic and Early Neolithic in the Iron Gates: a palaeodietary perspective. Journal of European Archaeology 5: 50–92. Bonsall, C., Cook, G., Lennon, R., Harkness, D., Scott, M., Bartosiewicz, L. & McSweeney, K. 2000: Stable isotopes, radiocarbon and the Mesolithic–Neolithic transition in the Iron Gates. Documenta Praehistorica 27: 119–132. Bonsall, C., Cook, G., Manson, J. & Sanderson, D. 2002: Direct dating of Neolithic pottery: progress and prospects. Documenta Praehistorica 29: 47–59. Bonsall, C., Cook, G., Hedges, R., Higham, T., Pickard, C. & Radovanović, I. 2004: Radiocarbon and stable isotope evidence of dietary change from the Mesolithic to the Middle Ages in the Iron Gates: new results from Lepenski Vir. Radiocarbon 46(1): 293–300. Borić, D. 2002: The Lepenski Vir conundrum: reinterpretation of the Mesolithic and Neolithic sequences in the Danube Gorges. Antiquity 76: 1026–1039. — 2007: New discoveries at the Mesolithic–Early Neolithic site of Vlasac: preliminary notes. Mesolithic Miscellany 18(1): 7–14. Borić, D. & Miracle, P. 2004: Mesolithic and Neolithic (dis)continuities in the Danube Gorges: new AMS dates from Padina and Hajdučka Vodenica (Serbia). Oxford Journal of Archaeology 23: 341–371. Borić, D. & Dimitrijević, V. 2007: When did the ‘Neolithic package’ reach Lepenski Vir? Radiometric and faunal evidence. Documenta Praehistorica 34: 53–72. Boroneanţ, V. 2000: Paléolithique Supérieur et Épipaléolithique dans la Zone des Portes de Fer. Bucureşti: Silex. Boroneanţ, V., Bonsall, C., McSweeney, K., Payton, R. & Macklin, M. 1999: A Mesolithic burial area at Schela Cladovei, Romania. 1

7

Section 1

REGIONAL STUDIES

Lithic technology and settlement systems of the Final Palaeolithic and Early Mesolithic in the Iron Gates Dušan Mihailović

Abstract: This paper deals with the changes that took place in the settlement system, economy and lithic technology during the Final Palaeolithic and Early Mesolithic of the Iron Gates. An attempt is made to explain the correlations, frequency and variability of various types of material remains in terms of the level of specialization in their acquisition, exploitation and usage. It is argued that the Final Palaeolithic communities of the Iron Gates were characterized by a high level of specialization in the production of artefacts, an organized system of settlements, and an organized system of supply of all resources at the regional level. Changes occurred at the beginning of the Holocene, reflected in: a) significantly narrower site catchments, b) intensification in the exploitation of local resources, and c) technological decline in stone artefact production. The technological decline was a consequence of cultural adaptation to the environmental changes during that period. Indirectly, it is manifested in reduced mobility of the communities and different activities conducted within the settlements. Key words: lithic technology, settlement systems, Palaeolithic, Mesolithic, Iron Gates

Introduction

Gates Mesolithic, especially in its earlier phase, were caused primarily by internal cultural, social and ecological factors. For this reason, I shall try to explain the correlations, frequency and variability of various types of material remains in terms of the level of specialization in their acquisition, exploitation and usage. The degree of specialization in technology and the mode of resource acquisition reflect to a considerable extent the social, economic, and even ideological demands of the community in the exploitation of the environment. Therefore, analysis of the various forms of specialization may lead to a better understanding of the changes in culture and way of life of Mesolithic hunter-gatherers in the Iron Gates.

Changes in the settlement system are fundamental to the interpretation of the cultural, social and ideological transformation of Mesolithic communities in the Iron Gates. However, in spite of the many features that point to a semi-sedentary or even sedentary way of life (Radovanović & Voytek 1997), the problem of establishing the degree of mobility of the Iron Gates Mesolithic communities is far from solved. The reasons are only partially theoretical in nature, related to the impossibility of the precise definition of the process of sedentarization (Voytek & Tringham 1989; Tchernov 1991; Bar-Yosef & Meadow 1995; Radovanović & Voytek 1997). Much greater difficulties result from fieldwork methods that were constrained by the ‘rescue’ character of the excavations and the uneven scope of investigation of certain sites; the perplexities of the stratigraphy and chronology of some of the key sites; and the lack of specialist analyses that are necessary for reliable conclusions regarding seasonality, mobility and functional differences among the settlements. However, since investigations of the Iron Gates Mesolithic have intensified in all respects over the past decade, it is reasonable to hope that in the near future most of the data necessary for reconstructing the settlement system will be at our disposal. This paper is an attempt to reconstruct the activities that resulted in the specific repertoire and patterning of the material remains in the Iron Gates sites from the perspective of lithic technology. The specific character of the Iron Gates Mesolithic necessitates a different approach to the relationship between lithic technology and settlement than that usually applied in ethnoarchaeological studies (Binford 1979; Binford & O’Connell 1984; Gamble 1993) or studies of other Mesolithic assemblages from Southeast Europe (Perlès 1987; Kozłowski 1999; Sinclair 1999). Specifically, exotic raw materials, curated technology and other features indicative of greater mobility of human groups and direct acquisition or exchange of goods over a wider area, in fact are not decisive in understanding the Iron Gates Mesolithic. Most of the available data lead to the conclusion that changes in the Iron

Final Palaeolithic The chipped stone industries from level I of the Cuina Turcului rockshelter and Climente II cave, attributed to the very end of the Pleistocene, are comparable with the Late Glacial industries of the Balkan and Apennine peninsulas, as discussed by several previous authors (Păunescu 1970, 1989; Boroneanţ 1999). However, from the point of view presented here, it is important to stress that over the wider region of the eastern and southern Mediterranean in this period, there is an increase in intensity of habitations in caves and rockshelters in mountainous areas. This phenomenon has been explained by stressing the importance of palaeoecological, palaeogeographical, demographical or cultural factors (Bailey et al. 1983; Bar-Yosef & Meadow 1995; Miracle 1996). It may be suggested, however, that the quantity and diversity of the finds in the Final Palaeolithic horizons of sites in Southeast Europe testify above all to an organized settlement system based on intensive exploitation of resources at the regional level (Mihailović 1999b). The technology and economy of the period are characterized primarily by specialization (economizing on resources and maximizing their exploitation) on the one hand, and diversification on the other. This is manifested in technology by the presence of a wider repertoire of 11

The Iron Gates in Prehistory

The frequency of the faunal remains from different biomes (forest, steppe and montane) in the lower levels of Cuina Turcului indirectly suggest that the rockshelter during this period was a base camp. The organization of hunting activities, the mode of hunting and, most probably, the weapons used in hunting the species from these environments differ significantly (for example, hunting ibex and chamois requires a high level of specific skills — Gamble 1999: 230–240). Therefore, in spite of the relatively short distances between different ecological zones in the area of Veliki Kazan (‘the Great Cauldron’) and their mutual connections (Bolomey 1973: 43–45), it may be assumed that the faunal remains in the lower levels of the site are the result of a number of forays with more-or-less specialized intentions. The remains from level I indicate intensive hunting of wild boar, ibex and chamois, while those from level II show an emphasis on ibex and chamois (and the important role played by steppe animals) with a very low ratio (c. 10%) of forest animals. The quantity and character of the lithic industries from the Final Palaeolithic sites of the Iron Gates offer additional information on the degree of mobility, and nature and duration of human occupations of natural shelters. Mobility and contacts at the regional level are attested by a significant presence of high quality raw materials at Cuina Turcului, especially in level II (Dinan 1996a, 1996b). Artefacts of local grey flint are prevalent in both levels, while quartz is used in only small percentages in both occupation phases. Concerning technological indicators of the character of the settlement, Dinan (1996b) suggested that cores had been brought to the site in a partially prepared state and that most of the artefacts (including tools and finished blanks) had been produced outside the habitation. She based this proposition on the fact that her analysis of a small sample of the artefacts from Cuina Turcului (representing only 5–10% of the total assemblage) revealed a small percentage of cortical pieces (23% in level I and only 13% in level II). However, the huge amount of chipping debris reported by Păunescu (1978) undoubtedly indicates that manufacturing of the majority of the artefacts was executed inside the rockshelter. Compared to the other Iron Gates sites, Cuina Turcului is also characterized by a lower percentage of cores and retouched artefacts (Fig. 3), as well as a high percentage of unretouched and retouched blades and bladelets. This pattern is surely influenced by the intensive exploitation of high quality cores, as well as by the fact that the bipolar flaking technique was not widely implemented during this phase. The typological composition of the retouched artefacts from levels I and II of Cuina Turcului is presented in detail in the works of Păunescu (1970, 1978). On the basis of the data presented, the tool assemblages from these levels are characterized by very high percentages of end-scrapers (the highest of all the Iron Gates sites) — up to 60%. Retouched blades and bladelets are well represented in the assemblage (over 20%) while backed tools make up 13% in level I, falling to 6% in level II. These data are well known, and it is the intention here only to emphasize the very small percentage of typologically undifferentiated (informal) tools. According to Păunescu’s classification (Păunescu 1978: 19–20), only the notched and denticulated tools may be listed as such, while side-scrapers do not appear at all. However, the tendency to-

Figure 1. Cuina Turcului, level I: frequency of ungulate species identified.

microliths, and in economy by the addition of new food resources (Miracle 1996). The most comprehensive data on the palaeoecology of the Iron Gates and the mode of exploitation by human groups have been gathered by pollen and faunal analyses of levels I and II at Cuina Turcului (Bolomey 1970, 1973; Pop et al. 1970; Păunescu 1978). In the lower levels of this rockshelter species from various biomes have been recorded, reflecting well the contrasts in the natural environment within the Iron Gates gorge and its immediate hinterland during this period. For example, in Cuina Turcului level I, along with the bones of ibex, chamois, bison and horse (assumed to be species adapted to an environment lacking forested habitats), a very significant percentage of woodland fauna also appears (Fig. 1). On the other hand, a lower incidence of woodland fauna and an increase in steppe and montane elements in level II (Fig. 2) indicate that this level formed under colder climatic conditions, most probably during the Younger Dryas. However, it has to be acknowledged that the avifauna and pollen analyses point to the opposite trend toward climatic amelioration (Pop et al. 1970; Păunescu 1978).

Figure 2. Cuina Turcului, level II: frequency of ungulate species identified.

12

Dušan Mihailović: Lithic technology and settlement systems

Figure 3. Tripolar graph representing a seriation of the cores, blades/flakes and tools at Iron Gates sites: A–Alibeg, P–Padina, VL–Vlasac, VT–Veterani Terrace, IC–Icoana, R–Răzvrata, OB–Ostrovul Banului, SC–Schela Cladovei.

wards a highly selective use of blanks in the production of tools is very marked, especially in level II. End-scrapers dominate among the tools made from flakes, while blades and bladelets were used in the production of backed and geometric tools, end-scrapers (11.4% in level I and 16.5% in level II), notched (>50% of notched pieces in both levels) and denticulated tools (25.3% in level I and 13.0% in level II).

finds from these layers are not dated, but they surely belong to the period before the beginning of the seventh millennium cal BC — the date of layer III. This site is not very well known, as is the case with Veterani Terrace, having elements of material culture of both this and the subsequent phase of the Iron Gates Mesolithic (Boroneanţ 1973, 1989, 1999; Radovanović 1996: 8–12). The remains of the earliest settlement (A1) at Padina are represented only by the finds from the central part of sector II (Radovanović 1981). The majority of the finds assigned to phase A2 also come from sector II and most of them were found above the construction of flat stones and pebbles in block 2a. In sector I a small number of artefacts were recorded below the hearth attributed to the hypothetical phase AB. There are indications, however, that the original settlement in this sector was situated on the very edge of the river and was probably destroyed by river erosion (Radovanović 1996: 62). Therefore, it is possible that excavation of this part of the site uncovered only a peripheral part of the settlement, and this may explain the small number of artefacts belonging to phase A. Artefacts from phase AB have been recorded in the hearth in block 1b of sector II, and have been compared to the industry from Alibeg owing to the presence of typical microliths and by-products of blade débitage (Radovanović 1981: 63). In any case, there is no certain archaeological evidence of settlement continuity at Padina between phases A and B.

Early Mesolithic The settlement of Padina, by Gospođin Vir (‘the Lady’s Whirlpool’), has been dated by a series of radiometric 14C ages on human bone (not corrected for the freshwater reservoir effect, cf. Cook et al. 2002) to the second half of the ninth millennium cal BC (Burleigh & Živanović 1980). It belongs to the earliest phase of habitation on the Danube bank, following level II of Cuina Turcului (Jovanović 1971, 1974, 1987; Radovanović 1981, 1996). One should however bear in mind that most of these dates have been established for skeletal remains from the so-called ‘stone construction of the necropolis’ in sector III, assigned to phase A2, so the finds from phase A1 could be somewhat older (Radovanović 1996: 173, 359). Layers I and II of the site of Ostrovul Banului may also belong to the initial phase of the establishment of openair settlements in the Iron Gates area (Boroneanţ 1999). The 13

The Iron Gates in Prehistory

Padina A is different from the other sites of the Iron Gates owing to the very poorly preserved architectural features and the large quantity of faunal remains and chipped stone artefacts concentrated in a small area (Radovanović 1981: 26). For example, in block 2a of sector II, as many as 75 lithic artefacts per square metre were recovered — more than in level I of Cuina Turcului, where the average density of finds was 72.9 per square metre (in level II as high as 143.9 per square metre). However, the composition of the finds from phases A1 and A2 in sector II does not meet the criteria for interpretation as a lithic workshop (Kozłowski 1980) — i.e. the percentage of retouched tools in these levels (23.7% in phase A1 and 15.1% in phase A2) is among the highest in the Iron Gates and is only matched by levels I and II of Ostrovul Banului (Radovanović 1996: 233). The high percentage of retouched tools on Palaeolithic and Mesolithic sites is usually an indication of short-term habitations, but in the case of Padina A it seems more probable that it reflects the nature and intensity of the activities performed in that part of the site. This question, just as with the issue of seasonality, will hopefully be solved after analysis of the faunal remains has been completed. The industry of Padina A differs from that in the lower levels of Cuina Turcului in having a lower frequency of high quality raw materials, a much higher percentage of flakes and tools on flakes, and a much smaller proportion of backed pieces in phase A2. A somewhat smaller percentage of unretouched artefacts in the overall inventory of Padina A1 and A2 may be explained by the higher percentage of bipolar cores (Fig. 5) and, on the other hand, by the less intensive use of this type of core (i.e. the smaller number of flakes removed from them). A similar pattern appears in layer II of Ostrovul Banului, but the frequency of cores (100% bipolar) is somewhat lower, retouched tools are more frequent, and the percentage of unretouched artefacts is lower, compared to Padina. A preliminary analysis of the raw material composition indicates that at Padina the same raw materials were used as on the neighbouring sites of Vlasac and Lepenski Vir (Fig. 6). Artefacts made from grey and grey-brown flint predominate. Low quality siliceous brown and greenish stones are much less frequent (including siltstone and mudstone). A very small percentage of the artefacts are of chalcedony, and red and green jasper, while finds of obsidian are very few. Quartz is poorly represented in the form of unretouched flakes, but not cores or retouched tools. All this points to a decrease in mobility and longer periods of occupation at the settlements. It may be argued that the differences between Cuina Turcului and Padina, noted in all three categories (raw materials, flaking technology, tools inventory), do not result simply from the fact that Padina is an open-air settlement while Cuina Turcului is a rockshelter. Rather, the differences probably reflect primarily the change in the mode of habitation and exploitation of resources taking place in this period. These changes are clearly recorded in the evolution of the industries at numerous Palaeolithic and Mesolithic sites in Montenegro and Greece (Mihailović 1999a, 2001) — especially at Crvena Stijena (Mihailović 1998) and Franchthi Cave on the Peloponnese (Perlès 1987, 1990, 1999). Further analyses of the material from Padina have shown

Figure 4. Padina A: frequency of ungulate species identified.

The fauna of Padina A is represented by the remains of a large number of animal species and more than a thousand individual bone specimens (Clason 1980). Red deer remains are the most frequent, while wild boar, roe deer and aurochs (Bos primigenius) are represented by less than 15% (Fig. 4). The quantity of fish bones is very large, up to about 25% of the total faunal assemblage from this phase (in phase B fish bones constitute 87%). This fact, along with the kind of fish (a substantial number of large specimens), the context of the majority of bones, and the very position of the settlement (near Gospođin Vir), points to the importance of fishing in the economy of the early inhabitants of Padina. However, the quantity of red deer remains is a clear indication that Padina, especially in its early phase, was not a specialized fishing settlement. Unfortunately, the seasonality and the true nature of the habitation at Padina cannot be reconstructed at present, since the data indicating that the site was inhabited all through the year apply to Padina A as a whole (Clason 1980: 171), not its individual phases. Phase A1 may therefore be judged only on the basis of the chipped stone artefacts.

Figure 5. Percentage of bipolar cores in relation to the total number of cores: A–Alibeg, P–Padina, VL–Vlasac, LV–Lepenski Vir, VT–Veterani Terrace, IC–Icoana, R–Răzvrata, OB–Ostrovul Banului, SC–Schela Cladovei.

14

Dušan Mihailović: Lithic technology and settlement systems

the Privod hill, in the close vicinity of the site, is used (Radovanović 1996: 316). All this indicates that mobility and, most probably, the need to acquire high quality raw materials decreases markedly in this period. Underlying this approach to the exploitation of lithic resources must have been the achievement of a certain level of technological specialization, necessary for successful knapping of quartz cores.

Conclusion Changes in settlement and subsistence practices may be observed in the Iron Gates from the Final Palaeolithic to the beginning of the Neolithic (Tables 1 & 2). In the Final Palaeolithic of Cuina Turcului the ratio of faunal remains and the lithic raw materials used point above all to the greater exploitation of resources in the immediate vicinity of settlements, although there are elements that point to regional mobility and, even, inter-regional contacts (obsidian). According to the available data, it may be assumed that during this period the rockshelter was a base camp, and that the activities of the human groups were directed toward the Iron Gates hinterland. The differences in the span and nature of the habitations point to functional differences among Cuina Turcului and the caves of Climente II and Hoţilor Băile Herculane (Bolomey 1973; Dinan 1996a), although their contemporaneity and role in the settlement system cannot be established with certainty. As far as subsistence is concerned, there is no evidence of specialization on particular animal species. Economy in the use of resources and the level of elaboration of technological process are attested by microlithization of artefacts, standardization of tools, and selective use of particular raw materials and especially blanks in the production of certain categories of artefacts. The possibility cannot be ruled out that as early as the Younger Dryas, which over the last decade has been established as an abrupt, cold oscillation of global character (Sherratt 1997; Adams & Otte 1999), the Iron Gates region was a natural refugium that was intensively inhabited. There is, however, no evidence to back up this assumption. The settlements on the river bank, even if they existed, were washed away by river erosion, either by the beginning of Holocene, after the increase in the Danube level, or in the subsequent periods. However, in level II of Cuina Turcului, no matter whether it belongs to the Younger Dryas (as indicated by several lines of evidence) or to the very beginning of Holocene, there are features that point to the exploitation of a poor environment (orientation toward hunting ibex and chamois, the additional supply of resources from the steppe biome). Also the total quantity and composition of the artefacts and faunal remains in this level indicate that Cuina Turcului in this period had a clearer role as a base camp, but at the same time its economic territory was somewhat larger, embracing some elements of a logistical system of settlement. With regard to the Holocene and the establishment of open-air settlements, changes in the settlement system, acquisition of food and mineral resources, and flaking technology are gradual but radical — as seen clearly in the finds from Padina A, Ostovul Banului I and II, and Veterani Terrace. The continuity with the Final Palaeolithic is marked

Figure 6. Flint artefacts as a percentage of the total number of artefacts of flint, quartz/quartzite, and other siliceous rocks in Iron Gates sites: A–Alibeg, P–Padina, VL–Vlasac, LV–Lepenski Vir, VT–Veterani Terrace, IC–Icoana, R–Răzvrata, OB–Ostrovul Banului, SC–Schela Cladovei.

that selective use of raw materials in the production of certain types of blanks and retouched tools is very poorly expressed, but is present nonetheless. Grey flint is the basic material in bipolar débitage — attested by the relatively high number of bipolar cores of this material and a large number of flakes with a linear, almost unrecognizable striking platform. It is worth noting that the implementation of this technology seems to be more evident from the characteristics of the striking platform than the scars on the dorsal surface. A kind of specialization and economization of raw material resources is attested by the characteristic products of flaking: flakes with two ventral sides and blades resembling burin spalls, detached from so-called plate cores. All this clearly indicates that all the available supports were used for knapping (as cores), including very small thick flakes. An identical situation has been recorded at Vlasac (Kozłowski & Kozłowski 1982: 33–36, 43–45). Also, as at Cuina Turcului (Dinan 1996b: 19), many cores are secondarily processed — in fact, turned into retouched implements. It needs to be stressed that bipolar technology and microlithization represent the major traits of the chipped stone industry, not only of Padina, but also the other Iron Gates sites (Boroneanţ 1989, 1999). This is important because these two elements are not so well expressed on the sites in Montenegro and Greece, where the proportion of bipolar cores is lower and the tendency toward microlithization of tools made on flakes is not so pronounced (Perlès 1987; Mihailović 1996, 1998, 1999a). In the next phase of the Iron Gates Mesolithic, on the sites of the lower Gorge, the industries acquire a ‘quartzose’ character; bipolar technology is used on quartz too, and quartz is also used in the production of tools. As the technology of quartz flaking is mastered and tools of this material appear, the need for systematic acquisition of flint decreases. The composition of the raw material of the industries from the Veterani Terrace and Ostrovul Banului clearly shows a gradual decrease in acquisition of flint from the hill of Sviniţa in the upper Gorges (Boroneanţ 1989: 477; Radovanović 1996: 228–234), while in the case of Alibeg quartz represents the dominant material, although flint from 15

The Iron Gates in Prehistory Table 1. Comparison of Padina A and Cuina Turcului I–II in terms of flaking technology. Lithics

FINAL PALAEOLITHIC (Cuina Turcului)

EARLY MESOLITHIC (Padina)

Variability

Degree of operationalization

Intensification

- various raw materials, from different sources, collected and transported in different ways - different types of blanks (including microbladelets) - pronounced variability and standardisation of tools (especially in microlithic group)

- high degree of raw material selectivity in blade/flake production (decreasing) - high degree of blank selectivity in tool production

- intensive exploitation of blade/bladelet cores - rejuvenation of cores during exploitation - microlithisation - secondary modification of cores and tools (repairing and retooling)

- local, bad quality raw materials prevail - flakes predominate in artefact assemblages - low variability of tools; low level of tool standardisation

- high – but only in bipolar core exploitation - low level of raw material selectivity (with exception of grey flint and mudstone) - low level of blank selectivity in tool production

- intensive exploitation of bipolar and other (irregular flake) cores - characteristic by-products - microlithisation

Table 2. Comparison of Padina A and Cuina Turcului I–II in terms of subsistence strategy. Fauna

FINAL PALAEOLITHIC (Cuina Turcului)

EARLY MESOLITHIC (Padina)

Diversity in environment and archaeological record

Degree of operationalization of hunting technique, game selection and prey processing

Intensity in exploitation of environment, some faunal species and prey

- relatively poor environment ? - species from different biomes (Cuina Turcului I) - orientation towards exploitation of species from mountainous environment (Cuina Turcului II) - site location near or at moderate distance from area of exploitation - activities directed to the hinterland

- different hunting strategies - hunting equipment: projectiles with inserted backed points and/or geometric microliths

- unknown – confirmed in other sites in Balkans (Badanj, Klithi)

- relatively rich environment - orientation towards fishing and deer hunting - location of settlement near zone of exploitation

* selective hunting (size, age, sex groups) confirmed for some other Late Upper Palaeolithic sites in the Balkans (Badanj, Klithi); no data available for Iron Gates sites. - new hunting/fishing techniques – probably highly developed - hunting by bone points ? * no data available for selective hunting in this period; in fishing – orientation to bulky specimens

solely in the repertoire of tools — by the presence of backed pieces, geometric artefacts, and circular and thumbnail scrapers. On the other hand, in terms of the composition of raw materials there is a sudden decrease in the presence of high-quality stones, while in technological terms flakes and tools on flakes show a growing trend. The changes in subsistence are even more marked. Compared to the preceding period, when survival of the communities, living in a relatively poor environment, depended on maximal exploitation of diverse resources, in the Early Mesolithic there was an intensification of the exploitation of selected resources (deer

- no available data * Intensification is visible only in the number of hunted specimens (i.e. in the quantity of finds)

hunting and fishing), in a (probably) much richer environment. This appears to have been accompanied by a similar trend in technology — the narrowing of the repertoire of the artefacts and intensification in the production of implements of low-quality flint. This was a consequence of several factors: (i) the disintegration of cultural identity caused by the abandonment of the settlement system practised in the preceding period, (ii) changes in hunting techniques, and (iii) a reduction in mobility of human communities, with longer stays at habitation sites and, probably, different activities being performed there.

1

16

Dušan Mihailović: Lithic technology and settlement systems

References

Approaches and Goals in Polish Archaeology. Wrocław: Polska Akademia Nauk Instytut Historii Kultury Materialnej, 33–55. — 1999: Gravettian/Epigravettian sequences in the Balkans: environments, technologies, hunting strategies and raw material procurement. In Bailey, G.N., Adam, E., Perlès, C., Panagopoulou, E. & Zachos, K. (eds) The Palaeolithic Archaeology of Greece and Adjacent Areas. London: British School at Athens, 319–329. Kozłowski, J.K. & Kozłowski, S.K. 1982: Lithic industries from the multi-layer site Vlasac in Yugoslavia. In Kozłowski, J.K. (ed.) Origin of the Chipped Stone Industries of the Early Farming Cultures in Balkans. Warszawa–Kraków: Państwowe Wydawnictwo Naukowe, 11–109. Mihailović, D. 1996: Upper Palaeolithic and Mesolithic chipped stone industries from the rockshelter of Medena Stijena. In Srejović, D. (ed.) Prehistoric Settlements in Caves and Rockshelters of Serbia and Montenegro — Fascicule I. Belgrade: Centre for Archaeological Research, 9–60. — 1998: Gornji paleolit i mezolit Crne Gore. Unpublished PhD dissertation, Faculty of Philosophy, University of Belgrade. — 1999a: The Upper Palaeolithic and Mesolithic stone industries of Montenegro. In Bailey, G.N., Adam, E., Perlès, C., Panagopoulou, E. & Zachos, K. (eds) The Palaeolithic Archaeology of Greece and Adjacent Areas. London: British School at Athens, 343–356. — 1999b: Intensification of settlement in the Late Glacial of southwestern Balkans. In Kobusiewicz, M. & Kozłowski, J.K. (eds) Post-pleniglacial Re-colonisation of the Great European Lowland. Folia Quaternaria 70. Kraków: Polska Akademia – Komisja Paleogeografii Czwartorzędu, 385–392. — 2001: Technological decline of the early Holocene chipped stone industries in South-east Europe. In Kertész, R. & Makkay, J. (eds) From the Mesolithic to the Neolithic. Proceedings of the International Archaeological Conference held in the Damjanich Museum of Szolnok, September 22–27, 1996. Budapest: Archaeolingua, 339–347. Miracle, P.T. 1996: Diversification in Epipalaeolithic subsistence strategy along the eastern Adriatic coast: a simulation approach applied to zooarchaeological assemblages. Atti della Società per la Preistoria e Protostoria della Regione Friuli-Venezia Giulia 9 (1994–1995): 33–62. Păunescu, A. 1970: Epipaleoliticul de la Cuina Turcului-Dubova. Studii şi Cercetari de Istorie Veche 21(1): 3–47. — 1978: Cercetarile arheologice de la Cuina Turcului-Dubova (Jud. Mehedinţi). Tibiscus Istorie, Volum Închinat Celei de-a 60 Aniversari a Unirii: 11–56. — 1989, Le Paléolithique et le Mésolithique de Roumanie (un bref aperçu). L’Anthropologie 93(1): 123–158. Perlès, C. 1987: Les Industries Lithiques Taillées de Franchthi (Argolide, Grèce). Tome I: Présentation Générale et Industries Paléolithiques. Bloomington, Indianopolis: Indiana University Press. — 1990: Les Industries Lithiques Taillées de Franchthi (Argolide, Grèce). Tome II: Les Industries du Mésolithique et du Néolithique Initial. Bloomington, Indianopolis: Indiana University Press. — 1999: Long term perspectives on the occupation of the Franchthi Cave. In Bailey, G.N., Adam, E., Perlès, C., Panagopoulou, E. & Zachos, K. (eds) The Palaeolithic Archaeology of Greece and Adjacent Areas. London: British School at Athens, 311–318. Pop, E., Boscaiu, N. & Lupsa, V. 1970: Anexa nr. 1 — Analiza sporo-polenica a sedimentelor de la Cuina Turcului-Dubova. Studii şi Cercetari de Istorie Veche 21(1): 31–34. Radovanović, I. 1981: Ranoholocenska kremena industrija sa lokaliteta Padina u Ðerdapu. Belgrade: Arheološki Institut.

Adams, J. & Otte, M. 1999: Did Indo-European languages spread before farming? Current Anthropology 40: 73–76. Bailey, G., Carter, P., Gamble, C. & Higgs, H. 1983: Epirus revisited: seasonality and inter-site variation in the Upper Palaeolithic of north-west Greece. In Bailey, G. (ed.) Hunter-Gatherer Economy in Prehistory. Cambridge: Cambridge University Press, 64–78. Bar-Yosef, O. & Meadow, R.H. 1995: The origins of agriculture in the Near East. In Price, T.D. & Gebauer, A.B. (eds) Last Hunters–First Farmers: New Perspectives on the Prehistoric Transition to Agriculture. Santa Fe: School of American Research Press, 39–94. Binford, L.R. 1979: Organization and formation processes: looking at curated technologies. Journal of Anthropological Research 35: 255–273. Binford, L.R. & O’Connell, J. 1984: An Alawyara day: the stone quarry. Journal of Anthropological Research 40: 406–432. Bolomey, A. 1970: Anexa nr. III — Cîteva observaţii asupra faunei de mamifere din straturile romanello-aziliene de la Cuina Turcului. Studii şi Cercetări de Istorie Veche 21(1): 37–39. — 1973: An outline of the Late Epipalaeolithic economy at the ‘Iron Gates’: the evidence on bones. Dacia n.s. 27: 41–52. Boroneanţ, V. 1973: Recherches archéologiques sur la culture Schela Cladovei de la zone des ‘Portes de Fer’, Dacia n.s. 27: 5–39. — 1989: Thoughts on the chronological relations between the EpiPalaeolithic and the Neolithic of the Low Danube. In Bonsall, C. (ed.) The Mesolithic in Europe. Edinburgh: John Donald, 475–480. — 1999: The Mesolithic habitation complexes in the Balkans and Danube Basin. Living Past 1. http://www.cimec.ro/livingpast/ mesolithic.htm Burleigh, R. & Živanović, S. 1980: Radiocarbon dating of a CroMagnon population from Padina, Yugoslavia, with some general recommendations for dating human skeletons. Zeitschrift für Morphologie und Anthropologie 70(3): 269–274. Clason, A.T. 1980: Padina and Starčevo: game, fish and cattle. Palaeohistoria 22: 141–173. Cook, G.T., Bonsall, C., Hedges, R.E.M. , McSweeney, K., Boroneanţ, V., Bartosiewicz, L. & Pettitt, P. 2002: Problems of dating human bones from the Iron Gates. Antiquity 76: 77–85. Dinan, E.H. 1996a: A preliminary report on the lithic assemblage from the early Holocene level at the Iron Gates site of Băile Herculane. Mesolithic Miscellany 17(2): 15–24. — 1996b: Preliminary lithic analysis of the Epigravettian levels from the Iron Gates site of Cuina Turcului. Mesolithic Miscellany 17(2): 25–40. Gamble, C. 1993: Exchange, foraging and local hominid networks. In Scarre, C. & Healy, F. (eds) Trade and Exchange in Prehistoric Europe. Oxford: Oxbow Books, 35–44. — 1999: The animal bones from Klithi. In Bailey, G.N., Adam, E., Perlès, C., Panagopoulou, E. & Zachos, K. (eds) The Palaeolithic Archaeology of Greece and Adjacent Areas. London: British School at Athens, 207–244. Jovanović, B. 1971: Elements of the Early Neolithic architecture in the Iron Gate gorge and their functions. Archaeologica Iugoslavica 9: 19. — 1974: Praistorija gornjeg Ðerdapa, Starinar n.s. 22: 122. — 1987: Die Architektur und Keramik der Siedlung Padina B am Eisernen Tor, Jugoslawien. Germania 65: 116. Kozłowski, J.K. 1980: Technological and typological differentiation of lithic assemblages in the Upper Palaeolithic: an interpretation attempt. In Schild, R. (ed.) Unconventional Archaeology: New 1

17

The Iron Gates in Prehistory — 1996: The Iron Gates Mesolithic. Ann Arbor, International Monographs in Prehistory. Radovanović, I. & Voytek, B. 1997: Hunters, fishers and farmers: sedentism, subsistence and social complexity in the Iron Gates Mesolithic. Analecta Praehistorica Leidensia 29: 19–31. Sherratt, A. 1997: Climatic cycles and behavioural revolutions: the emergence of modern humans and the beginning of farming. Antiquity 71: 271–287. Sinclair, A. 1999: Technological decision making: the case of Megalakkos. In Bailey, G.N., Adam, E., Perlès, C., Panagopoulou, E. & Zachos, K. (eds) The Palaeolithic Archaeology of

Greece and Adjacent Areas. London: British School at Athens, 188–196. Tchernov, E. 1991: Biological evidence for human sedentism in Southwest Asia during the Natufian. In Bar-Yosef, O. & Valla, F. (eds) The Natufian Culture in the Levant. Ann Arbor, MI: International Monographs in Prehistory, Archaeological Series 1, 315–340. Voytek, B. & Tringham, R. 1989: Rethinking the Mesolithic: the case of South-east Europe. In Bonsall, C. (ed.) The Mesolithic in Europe. Edinburgh: John Donald, 492–499.

18

The development of the ground stone industry in the Serbian part of the Iron Gates Dragana Antonović

Abstract: The early Holocene ground stone industry of the Iron Gates comprises two components – local and ‘imported’. The local component was developed independently against a background of Mesolithic sedentary settlement, according to the basic economic needs (hunting and fishing) of the inhabitants of this area. The tools were manufactured by limited modification of natural pebbles. Alongside the artefacts typical of the Iron Gates ground stone industry, occur ground edge tools (axes, adzes and chisels) characteristic of the Neolithic of central Serbia that were used exclusively for woodworking – these are referred to as the ‘imported’ component. The tools of the ‘imported’ component have traits that are very different from the tools of the local component. The parallel appearance of two ground stone industries of differing character can be regarded as another line of evidence for the parallel development of the Iron Gates Mesolithic and the Neolithic of the neighbouring areas which, after all, is indicated by absolute dating. Key words: ground stone industry, Mesolithic, Neolithic, Iron Gates, Serbia

Introduction

1981, 1983). Other stones from Lepenski Vir, as well as from Padina and Velesnica, were analysed by the present author. The field documentation from some Iron Gates sites contains descriptive rock identifications, but these are not sufficiently detailed or accurate for present purposes. Functional–typological analysis. The method of manufacture of the artefacts, their shape and use (determined by microwear analysis) were considered. Striations on some of the tools from Lepenski Vir and Velesnica were examined under a binocular microscope at 50X magnification, and interpreted on the basis of previous experience and examination (cf. Semenov 1976: 126–142). Typological analysis of the ground edge tools (axes, adzes and chisels) was done according to the classification of the material from Vinča (Antonović 1992: 7–16). Analysis of other artefact types made use of the observations of previous researchers in the Iron Gates.

The early Holocene ground stone industry of the Iron Gates is an exceptional phenomenon in Serbia. Nothing resembling it has been discovered elsewhere in the country. On present evidence, there would appear to have been an autochthonous and continuous development of the industry from the Early Mesolithic, independent of the origin and development of the Neolithic in the other parts of Serbia and, it seems, from the Neolithic in Romania and Bulgaria. In the course of archaeological research in the Iron Gates, the ground stone tools have been analysed from Mesolithic and Early Neolithic (VIII–VI millennium cal BC) horizons at the following sites: Padina, Lepenski Vir, Vlasac, Hajdučka Vodenica, Ajmana, Velesnica, Knjepište, and Ušće Kameničkog Potoka (Fig. 1). Included in the category ‘ground stone artefacts’ are all stone implements worked by grinding/polishing, as well as unfinished examples with traces of flaking or pecking. Also included under this heading are implements that are naturally polished or polished through use. This group includes various polishers, hammerstones and working plates,1 made from unmodified pebbles that were suitable for use. All implements were studied from two main aspects: petrographic and functional–typological. Petrographic analysis. The raw materials from which the ground stone implements are made were examined only macroscopically by non-destructive methods. They were tested for hardness and compactness — hardness is given in terms of the Mohs’ scale, tested by scratching with a steel needle; compactness is estimated by simple visual observation of the specimen. Then the presence of calcium carbonate was determined by testing the specimen with dilute hydrochloric acid. The mineral composition, texture and structure of the rock were determined by inspection with a hand lens. Some of the stones recovered in the excavations at Lepenski Vir and Vlasac were identified by Miomir Babović of the Mining-Geology Faculty in Belgrade (Srejović & Babović

The raw materials The raw materials are extremely varied, which is not surprising since most of the ground stone implements are made from rounded, river pebbles. Magmatic, sedimentary and metamorphic rocks were used equally for their manufacture. The physical-technical characteristics of the stone were decisive in the selection of materials. Thus, predominantly finegrained, blow-resistant rocks without cracks and natural flaws were chosen for the manufacture of ground stone implements. Magmatic rocks Granite Granites are the most widespread plutonic rocks, most commonly exhibiting various shades of grey and reddish colour. Only hard rocks, resistant to abrasion, were selected. In the Lepenski Vir locality fine-grained, very compact, greycoloured varieties, are represented. They were used for the 19

The Iron Gates in Prehistory

Figure 1. Early Holocene sites in the Serbian part of the Iron Gates where ground stone implements have been recorded.

manufacture of hammers, which were found in all horizons of the site. At Velesnica fine- and medium-grained massive pebble varieties are found, which were used either as hammerstones or shaped as mallet-weights. Larger granite blocks served as querns or stationary grindstones/polishers for working hard materials. At Vlasac granite was used for the manufacture of polishers for working bone needles (Srejović & Letica 1978: 99). At Padina granite pebbles were utilized as anvils.

Diabase Diabases are rocks characterized by ophitic texture. They are hard, compact and abrasion resistant. At Lepenski Vir a finegrained, dark-grey variety is represented, which was used for the manufacture of axes and chisels for woodworking. At Velesnica dark-grey fine-grained diabase was used as pebble-hammerstones. Andesite Andesites are compact and tough rocks exhibiting clearly visible porphyritic texture. At Lepenski Vir there are finegrained, compact, dark-grey varieties that were used for the manufacture of woodworking axes. At Vlasac small pebbles of this material were utilized as sling-balls (Srejović & Babović 1981: 76).

Gabbro Like granites, gabbros are tough rocks — i.e. resistant to breakage. Artefacts of gabbro are reported from Vlasac and Velesnica. At Vlasac fine-grained gabbro was used for the manufacture of grooved polishers (Srejović & Babović 1981: 76). At Velesnica fine-grained, compact and hard varieties (gabbro-diabases?) are represented; these were used for the manufacture of axes for woodworking, pounders, and small fishing weights. At Padina uralitized and saussirized gabbro was used for making axes.

Aplite Aplites are fine-grained rocks (related to granites) composed mainly of quartz and feldspar, and therefore white in colour. At Velesnica small fishing weights were made from aplites. 20

Dragana Antonović: The ground stone industry in the Serbian Iron Gates

Tuff These are compact rocks that are easily working. At Padina slightly silicified tuff (hardness 6) exhibiting an irregular conchoidal fracture and various shades of grey and ochreous colour, was used for the manufacture of tranchets and hammerstones. It is likely that the tranchets from Vlasac were also made of tuff, and not of some undefined siliceous rock (cf. Srejović & Letica 1978: 98).

pletely silicified fine-grained types were only occasionally recorded. At this site sandstones were used for the manufacture of pounders, whetstones, polishers, millstones, working plates, mallets, pebble-axes, weights, line winders, and axes; all of these artefacts, except the last two types, were manufactured from partially modified pebbles. At Vlasac various coarse- and fine-grained varieties were used for the manufacture of sceptres, pebble-axes, hammerstones, and polishers for awls (Srejović & Letica 1978: 99–100; Srejović & Babović 1981: 76).

Other magmatic rocks It is possible that in the Iron Gates other magmatic rocks were also utilized, but precise identifications have not been made. At Lepenski Vir various kinds of hard, very fine-grained, basic rocks were used, which could not be characterized macroscopically. These were used for the manufacture of axes, both blunt forms (Srejović 1972: 133) and ground edge forms. At Velesnica poorly defined magmatic rocks were also noted. These comprise fine-grained, dense, compact and hard rocks, of uniform ochreous colour, which occurred as pebbles, as well as fine-grained, atypical, compact and hard rocks, ochre-grey in colour, whose constituents are not recognizable owing to the presence of a deposit on the surface of the pebble. Such pebbles were used, slightly reworked, as small fishing weights and as anvils for stone tool working. At Vlasac various magmatic rocks (not identified more precisely — Srejović & Letica 1978: 99–100) were used for the manufacture of sceptres, hammerstones, working plates and sling-balls, which occurred in all horizons of the Vlasac Mesolithic. From the site of Ušće Kameničkog Potoka a ‘green stone of volcanic origin’ is mentioned, which was used for the manufacture of axes and chisels (Stanković 1986a: 468). Most probably, this is a gabbroic rock.

Limestones These are compact, fine-grained rocks, of hardness 4 (except silicified limestone with hardness grading up to 6.5). However, they are brittle and, therefore, not suitable for the manufacture of hammering tools. At Lepenski Vir several varieties of these calcareous rocks were in use. Weakly silicified, compact limestone with hardness 5, exhibiting various shades of grey and light greyish-green colour, was utilized for the manufacture of axes, adzes and hammers. Dense, hard limestones, as well as marly varieties, were used for the manufacture of sceptres (Srejović & Babović 1983: 186–189); whereas ornamental artefacts — beads and amulets — were made from whitish and reddish crystalline limestones (Srejović & Babović 1981: 72, 92; Srejović & Babović 1983: 191–197). Limestones are present in all horizons of Lepenski Vir. At Velesnica fine-grained, silicified, compact and hard, grey-coloured varieties were utilized for the manufacture of small fishing weights that retained the form of the pebbles. At Vlasac artefacts of limestone are also represented (Srejović & Letica 1978: 98). Other sedimentary rocks In addition to the sandstones and limestones mentioned above, which were abundantly used for the manufacture of ground stone artefacts, other sedimentary rocks are present in the Iron Gates sites in much smaller quantities. Marlstone and mudstone without calcite were used at Lepenski Vir for the manufacture of sceptres (Srejović & Babović 1983: 188). Silicified sedimentary rocks were also used to a lesser extent; for example; at Velesnica, where silicified pelite is recorded. This rock closely resembles chert; it is off-white in colour with alternating thin darker and lighter bands, compact, finegrained, has a hardness around 6, a conchoidal fracture, and contains calcium carbonate. Axes were made from this rock, on account of its hardness and toughness. At Vlasac there were numerous sceptres made from sedimentary rocks the character of which has not been clearly defined (Srejović & Letica 1978: 99).

Sedimentary rocks Sandstones These rocks may be used, especially when quartzose, as an abrasive material. Medium- and fine-grained varieties can be distinguished on the basis of grain size, and on the basis of the cement, they can be classified as calcareous, marly, ferruginous or siliceous sandstones. These rocks can readily be accepted as the leading raw material used during the Mesolithic and Early Neolithic in the Iron Gates area. All art objects (sculptures, altars) are made from sandstone pebbles (Srejović & Babović 1983: 107–181). Sandstones were largely used in the production of ground stone artefacts. In all levels at Lepenski Vir there are sandstones with siliceous and calcareous cement, as well as very fine-grained, weakly metamorphosed varieties. Pebbles of quartzose sandstones are very abundant. They served for the manufacture of altars and sculptures. From the fine-grained varieties were made axes, mallet-weights, hammerstones and trapezoidal grindstones/polishers embedded in the floors of houses, as well as pounders, which were found lying on the house floors beside the grindstones (Srejović 1972: 133; Srejović & Babović 1981: 76–77; Srejović & Babović 1983: 186–189). At Velesnica sandstones were the most common raw material. Varieties with calcareous, marly and siliceous cement were used — most commonly these were fine-grained, occasionally they were nearly fine-grained, compact and hard varieties. Medium- to coarse-grained sandstones and com-

Metamorphic rocks Gneiss and granite gneiss The gneisses are distinguished from granites by parallel (schistose) structure. They are mostly medium- to coarsegrained, compact, hard rocks. When fine grained and with the poorly developed schistosity, especially granite gneisses, they differ slightly from granite macroscopically. Granite gneiss is a rock from which at Lepenski Vir sceptres were made (Srejović & Babović 1983: 185). Gneiss is recorded at Velesnica, where pebbles of it were used as hammerstones. 21

The Iron Gates in Prehistory

Phyllite and phyllite–mica schist These are fine-grained (slightly crystalline) schistose rocks. Phyllite–mica schist is a phyllite grading to mica schist. At Lepenski Vir axes and chisels were made from these rocks, these being found in all levels of the site.

axes, adzes and chisels are recorded in the Lepenski Vir IIIb levels, chisels are recorded at Velesnica, and grindstones/polishers for awls were found in the Vlasac II levels (Srejović & Babović 1981: 77). Other rocks Fine-grained sedimentary and thermal-contact metamorphic rocks These terms cover the main rock types used for ground stone artefacts in the Neolithic of Serbia. They are fine-grained, very dense rocks, resistant to percussion, of hardness 6–6.5, and having an irregular conchoidal fracture, which makes them suitable for knapping. They are variable in colour — light to dark greyish-green, grey, brown, green, dark grey and black, commonly with lighter or darker bands and light spots. They comprise mainly hornfelses, skarnoids, alevrolithic metasandstones, and silicified pelites. The hornfelses and skarnoids are contact metamorphic rocks, whereas the alevrolithic metasandstones and pelites are sedimentary rocks that have undergone slighter changes. Fine-grained variants of these rocks are recorded in Lepenski Vir I and III, where they were used for axe, adze and chisel manufacture. They are also present at Velesnica, where they were used for the manufacture of axes.

Amphibolite These rocks may exhibit both schistose and massive structure. Most are fine-grained and compact; when massive or with reduced schistosity, they could represent a high quality raw material for implement production. Amphibolite artefacts are present at Vlasac (pebble-axes; Srejović & Babović 1981: 75) and Lepenski Vir (sceptres; Srejović & Babović 1983: 185). Other types of schist These are medium- to coarse-grained rocks of medium to high crystallinity, with perfect schistosity, which makes them easily recognizable. The various schists were preferred for the manufacture of ground stone implements, especially sceptres. Compact rocks were always used, but it should be stressed that some of these are not exceptionally hard rocks, as is the case with chlorite schists. Amphibole schist was used at Padina for sceptre manufacture. Some sceptres from Vlasac are made of albite-chlorite-muscovite schists (Srejović & Babović 1981: 74), but also of other varieties of schists. The sceptres from Lepenski Vir are most commonly made from muscovite-chlorite schist (Srejović & Babović 1983, 185–187). At Velesnica amphibole-muscovite schist is recorded, which was utilized for the production of axes.

Jadeite/nephrite Jadeite/nephrite are an interesting occurrence on Neolithic sites in the Iron Gates. These are actually two different minerals — jadeite is pyroxene and nephrite is amphibole — but cannot be distinguished macroscopically (both are known under the common name, ‘jade’). For that reason, it is not possible to say which of the minerals occurs at the Iron Gates sites. They are very dense and tough minerals, 6–6.5 in hardness, translucent, and green to olive green in colour. They are hard to work because they are hard to break, and thus are suitable for the manufacture of high quality tools. Jade is a rare mineral in the early Holocene Iron Gates sites. It is recorded in Lepenski Vir Ib–c, where it was used for manufacturing chisels. At Velesnica a fragment of a jadeite/nephrite ring was found.

Eclogite This is a massive, metamorphic rock of high crystallinity produced by metamorphosis of basic rocks. It is hard and tough, suitable for the manufacture of implements, but was not common in the early Holocene Iron Gates settlements. It is recorded at Lepenski Vir, where it was used for the production of chisels. Marbles These are white, granular rocks, exhibiting a wide range in grain size. They are of relatively low hardness (3–4 on Mohs scale), and therefore easy to work. When polished they exhibit a high lustre. Because of these characteristics they were a popular raw material for the manufacture of decorative and ceremonial artefacts — pendants and amulets, which are recorded only in Lepenski Vir I–II (Srejović & Babović 1983: 190–194).

Malachite and azurite These minerals are different hydrocarbonates of copper, produced by weathering of primary copper ores, and thus are very widespread. Malachite is green, sometimes blackish green, and azurite is dark blue in colour. Both minerals are of low hardness (3.5–4), and are easy to work. A rare occurrence of these minerals is recorded from Lepenski Vir IIIa, in the form of amulets (Srejović & Babović 1981: 92, 94). Malachite also occurs at Padina.

Serpentinites These are generally loose, highly fractured rocks of lesser hardness (about 4). Although being of low hardness, compact varieties of this rock could be, due to its toughness, suitable for tool manufacture, especially when partially silicified. Most common are green or dark green varieties, but brown and nearly black varieties also occur. Serpentinites were a preferred material for the manufacture of small woodworking tools, such as chisels and small adzes, during the Starčevo and late Vinča culture periods in Serbia (Antonović 1998: 36, 144). In the early Holocene Iron Gates localities serpentinite

Occurrence of raw materials at selected sites Tables 1–4 summarize the occurrence of raw materials used for manufacturing ground stone artefacts at four key sites. Lepenski Vir The analysis of the material from Lepenski Vir (Table 1) is based on direct inspection of material in the National 22

Dragana Antonović: The ground stone industry in the Serbian Iron Gates

Figure 2. Geological map of the Iron Gates area (after: Geological Map of Yugoslavia 1970; Pavelescu 1961; Kovačević et al. 1997; Vasković & Matović 1997). A. Holocene, B. Miocene–Pliocene clastic, claystones, limestones, C. Upper and Lower Cretaceous limestones, dolomites, clastics, D. Cretaceous andesites and pyroclastics, E. Upper and Lower Jurassic limestones, dolomites, clastics, F. gabbro, diabase, G. serpentinite, H. Permo-Carboniferous grey sandstones with coal, I. Permo-Triassic red sandstones, J. hornfels, K. Hercynian granites (massifs in eastern Serbia — 1. Brnjica, 2. Neresnica, 3. Gornjane, 4. Plavna), L. Lower Palaeozoic phyllites, greenschists, amphibolites, clastics, M. Riphean-Cambrian + Proterozoic schists, gneisses, amphibolites (Romania), N. Riphean-Cambrian greenschists, O. amphibolites, P. RipheanCambrian gneisses, micaceous schists, amphibolites, Q. Proterozoic gneisses, R. corundum, S. quartzites.

Museum in Belgrade, as well as on the field documentation and publications concerning Lepenski Vir. The sample of 119 specimens relates to the implements; decorative and ceremonial–magic artefacts are not included in the total.

of the material (Padina A and Padina B) was not attempted, since the final detailed analysis of excavations at Padina is in progress, which should shed new light on the stratigraphy of the site.

Vlasac The analyses performed and published by the investigators of this site (Srejović & Letica 1978: 98–100) have been used to compile the list in Table 2.

Velesnica In the analysis of the ground stone industry from Velesnica, the material from the Early Neolithic layers of block A, excavated in 1982, was used (see Vasić, this volume). A sample of 109 specimens was examined (Table 4).

Padina Only part of the stone material from the early Holocene strata of Sectors I, II and III has been analysed (Table 3), so that the picture from this site is incomplete. The main source of the data was the field documentation for Padina; the actual raw materials were examined in a small sample of artefacts from all three sectors. A more precise chronological classification

Origin of the raw materials The choice of rocks for the ground stone implements found in the Iron Gates sites exhibits considerable variety. The most common are sandstones, used equally for both art objects 23

The Iron Gates in Prehistory Table 1. Lepenski Vir. Rock

Implements

Horizons

1. Granites

Hammers

L. Vir Ie, II

2. Diabase

Chisels, axes

L. Vir II, IIIa

3. Andesite

Axes

L. Vir IIIa

4. Basic magmatic rocks

Axes, chisels

L. Vir Ib–e, IIIa

5. Sandstones

Axes, sceptres, weights, grinders/polishers, pounders

L. Vir Ic–e, II, III

6. Limestones

Axes, adzes, hammers, sceptres, beads, amulets

L. Vir Ib–e, II, IIIa

7. Marlstone

Sceptres

L. Vir I, II

8. Mudstones without calcite

Sceptres

L. Vir II

9. Schists (slates)

Sceptres

L. Vir Ia–b, II

10. Phyllite

Chisels

L. Vir Ic

11. Phyllite–mica schist

Axes

L. Vir IIIa

12. Granite gneiss

Sceptres

L. Vir Ib–d

13. Eclogite

Chisels

L. Vir I

14. Marbles

Amulets

L. Vir Ib–c, II

15. Serpentinite

Axes, adzes, chisels

L. Vir IIIb

16. Fine-grained rocks

Axes, adzes, chisels

L. Vir Ib–e, IIIa

17. Nephrite

Chisels

L. Vir Ib–c

18. Malachite, azurite

Amulets

L. Vir IIIa

Table 2. Vlasac. Rock

Implements

1. Granite

Grinding/polishing stones

2. Gabbro

Grinding/polishing stones

3. Andesite

Sling-balls

4. Tuff

Tranchets

5. Undefined magmatic rocks

Sceptres, hammerstones, working plates, sling balls

6. Sandstones

Sceptres, pebble-axes, hammerstones

7. Limestones

?

8. Undefined sedimentary rocks

Sceptres

9. Amphibolite

Sceptres

10. Schists

Sceptres

Table 3. Padina. Rock

Implements

1. Granite

Anvils

2. Gabbro

Axes

3. Silicified tuff

Tranchets

4. Sandstone

Grinding/polishing stones, anvils, pounders, mallets, sceptres, querns–altars

5. Schist

Sceptres

6. Malachite

Pendant?

24

Dragana Antonović: The ground stone industry in the Serbian Iron Gates Table 4. Velesnica. Rock

%

Implements

1. Granite

12.84

2. Aplite

1.84

Fishing weights

3. Diabase

1.84

Hammerstones

6.42

Axes, pounders, fishing weights

4. Gabbro, gabbro-diabase

Hammerstones, querns, mallets

5. Ill-defined magmatic rocks

10.09

Pebble-axes, fishing weights, working plates

6. Sandstones

42.20

Grinding/polishing stones, whetstones, pounders, querns, hammerstones, pebble-axes, mallets, fishing weights, working plates, spool

7. Silicified limestones

0.92

Axes

8. Silicified pelite

0.92

Axes

9. Granite gneiss

8.26

Hammerstones, pounders, grinding/polishing stones, mallets, fishing weights

10. Gneiss

0.92

Pounders

11. Amphibole–muscovite schist

0.92

Axes

12. Silicified serpentinite

0.92

Chisels

13. Fine-grained contact-metamorphic and sedsedimentary rocks

8.26

Axes, adzes

14. Jadeite

0.92

Ring

Typological–functional analysis

(sculptures) and various kinds of implements. Next are crystalline schists, most commonly used for sceptres, but also for other implements. This representation of raw materials corresponds to the distribution of these rocks in the wider Iron Gates region (Fig. 2). In this region there are extensive formations bearing sandstones and metamorphic rocks, mostly crystalline schists, on both sides of Danube (Fig. 2: H, F, L–Q). Magmatic rocks are also abundant. Granites are the most widespread, found equally on both the Romanian and Serbian sides of the river (in the Serbian part of the Iron Gates there are four massifs: Brnjica, Neresnica, Gornjane and Plavna — Fig. 2: K). Andesites, pyroclastic rocks (including tuffs which are commonly used at Padina for the manufacture of tranchets and axes), gabbro and diabase occur mainly in the hilly hinterland of the Padina, Lepenski Vir and Vlasac localities. All of the rocks mentioned above, as well as limestones, marlstones and marbles, have certainly been found in the form of pebbles in the numerous rivers and streams that form the tributaries of the Danube in this area. This cannot be said for hornfels, nephrite/jadeite and malachite. Most probably, the acquisition of these rocks and minerals was organized. Only in the case of hornfelses, which occur in the western part of the Iron Gates region, where the Danube makes contact with the Brnjica massif, is there a possibility that they could be collected from the surface as pebbles (Fig. 2: J). Much more numerous deposits of hornfels are found in central Serbia, and it may be assumed that some of this material was imported.2 Jadeite/nephrite deposits have not been recorded in the Iron Gates or elsewhere in Serbia; the nearest source is in Macedonia, to the south of Skopje (Solunska Glava — Marić 1945: 225). They are also present on the Kikládhes Islands (Cyclades) in Greece, as well as in the Piemonte region of western Italy. Therefore, it may be assumed that jade was imported from distant regions. On the other hand, malachite almost certainly derives from eastern Serbia, where extensive copper deposits occur.

There are two basic assemblages of ground stone artefacts from the Iron Gates sites that date to the early part of the Holocene. The first comprises implements of the so-called ‘local’ ground stone industry: mallet-sceptres, malletweights, hammerstones, anvils, pounders, grinding/polishing stones, whetstones, working plates, querns, mortars, slingballs, fishing weights, pebble-axes (blunt axes), amulets, and various ornamental artefacts. To these can be added tranchets. In the second group are artefacts of the ‘imported’ ground stone industry, which includes ground edge tools (axes, adzes and chisels), used for woodworking, and hammers made from broken and blunted ground edge tools. The term ‘imported’ should not be taken literally. Judging by the raw materials used for their manufacture, these artefacts are not imported as such, although they are present in smaller numbers. It is the technology that was imported — developed elsewhere and adopted in its entirety by the Iron Gates population. This is the case at Ajmana, where there are workshops for the manufacture of ground edge axes (Stalio 1986: 29). Artefacts of the ‘local’ ground stone industry are produced from pebbles, naturally shaped for use, which are also transformed into tools with minimal working. The pebbles are of many different rocks, but sandstones are the most numerous. These are very abundant in the region and thus are available in the watercourses. Implements of the ‘imported’ variant are made most commonly from both fine-grained sedimentary and thermal contact metamorphic rocks, as well as from rocks that closely resemble them in their physical-technical features — gabbro, diabase, andesite, fine-grained basic magmatic rocks, very fine-grained sandstones, limestones, eclogite, serpentinite, and jadeite/nephrite. These sorts of rocks can be worked by chipping because of their finegrained texture, and using this technique the implements were manufactured from large flakes and from the cores themselves. 25

The Iron Gates in Prehistory

Figure 3. Mallet-sceptres: A. Lepenski Vir I–II (Srejović & Babović 1983: 187); B. Lepenski Vir I–II (Srejović & Babović 1983: 189); C. Padina, sector III, trench 5, block 1, F.I. 374/70; D. Padina, sector II, F.I. 291/70; E. Vlasac III (Srejović & Letica 1978: pl. CXVIII); F. Vlasac II (Srejović & Letica 1978: pl. CXVIII).

26

Dragana Antonović: The ground stone industry in the Serbian Iron Gates

Figure 4. Mallet-weights: A. Lepenski Vir II, block d/II, F.I. 805; B. Velesnica, block A/7, F.I. 417; C. Padina, sector I, profile 1, F.I. 17/68; D. Padina, sector I, profile 1, F.I. 8/68; E. Padina, sector III, profile III.

27

The Iron Gates in Prehistory

Figure 5. Anvils: A. Padina, sector I, profile I, F.I. 31/68; F. Vlasac II (Srejović & Letica 1978: pl. CXIX); I. Vlasac II (Srejović & Letica 1978: pl. CXIX). Hammerstones: B. Velesnica, block A/7, layer 26; C. Padina, F.I. 370; D. Velesnica, block A/9, F.I. 454. Grindstones/polishers: E. Velesnica, block A/3; G. Vlasac I (Srejović & Letica 1978: pl. CXIX); H. Velesnica, block A/3, F.I. 451.

28

Dragana Antonović: The ground stone industry in the Serbian Iron Gates

Figure 6. Working plates: A. Velesnica, block A/1, layer 30. Querns: D. Velesnica, block A/1, layer 31; C. Padina, sector III, profile III, F.I. 360/70. Mortars: B. Lepenski Vir Ib–c (Srejović & Babović 1983: 163). Fishing weights: E. Velesnica, block A/2-2a, F.I. 469; F. Velesnica, block A/2-2a, F.I. 420; G. Velesnica, block A/3, F.I. 402. Sling-balls: H. Vlasac III (Srejović & Letica 1978: pl. CXVIII); I–J. Padina, sector III, trench 6, block 1, house floor.

29

The Iron Gates in Prehistory

Figure 7. Pebble-axes: A. Velesnica, block A/8, F.I. 416; B. Velesnica, block A/7, F.I. 455; C. Vlasac III (Srejović & Letica 1978: pl. CXIX); D. Vlasac II (Srejović & Letica 1978: pl. CXIX).

30

Dragana Antonović: The ground stone industry in the Serbian Iron Gates

Figure 8. Ornamental and ceremonial artefacts: A. Labret of limestone — Lepenski Vir IIIb (Srejović & Babović 1981: 94); B. Amulet of marble — Lepenski Vir Ib–c (Srejović & Babović 1983: 190); C. Ring of jadeite — Velesnica, block A/7, F.I. 498; D. Winder of fine-grained sandstone — Velesnica, block A/8, F.I. 422. Tranchets: E. Kula I (Sladić 1986: Fig. 6/25); F. Padina, sector II, trench 2, block 2a, F.I. 96/70; G. Padina, F.I. 165; H. Padina, sector II, trench 2, block 2a, F.I. 145/70; I. Vlasac I (Srejović & Letica 1978: pl. CXVII); J. Lepenski Vir III (Kozłowski & Kozłowski 1984: 289). Cutting-edge tools: K. Padina, sector I, profile I, F.I. 27/68; L. Velesnica, block A/8, F.I. 418; M. Padina, sector I, profile I, F.I. 7/68; N. Velesnica, block A/3, F.I. 393; O. Padina, sector III, trench 5, block 1, F.I. 281/70; P. Velesnica, block A/7, F.I. 401.

31

The Iron Gates in Prehistory

Figure 9. Cutting-edge tools: A. Lepenski Vir IIIa, no. 227; 3 B. Lepenski Vir IIIa, no. 230; C. Lepenski Vir III, no. 249; D. Lepenski Vir Ia, house 20, no. 254; E. Lepenski Vir Id–e, house 26, no. 262; F. Lepenski Vir Ie, house 8, no. 306; G. Lepenski Vir Ie, house 27, no. 244; H. Lepenski Vir Ib-c, house 7, no. 242, I. Lepenski Vir IIIb, depot 2, no. 250; J. Lepenski Vir IIIb, depot 2, no. 251; K. Lepenski Vir Ib–c, house 35, no. 253; L. Lepenski Vir Ie, house 21, No. 256.

32

Dragana Antonović: The ground stone industry in the Serbian Iron Gates

The ‘local’ ground stone industry Mallet-sceptres (Fig. 3) These are massive stone artefacts, up to 50 cm long, made from elongated pebbles of elliptical or circular cross-section. Final shaping was done by grinding, and some specimens have ornamentation made by shallow incision. It is likely that they were used for killing game and large fish, although there is an opinion that they are sceptres used in sympathetic magic and fertility rituals (Srejović & Babović 1983: 183). On the Serbian bank of the Danube, mallet-sceptres are present at Lepenski Vir (Srejović & Babović 1983: 185–189), Hajdučka Vodenica (Radovanović 1996: 323), Vlasac (Srejović & Letica 1978: 99) and Padina. Usually they are made of crystalline schists, and less commonly from finegrained sandstones, marlstones or limestones.

dish sandstone, embedded in the floors of the houses (Srejović 1972: 133), as well as at Ajmana (Stalio 1986: fig. 5) and Velesnica. Whetstones are usually linked with the sharpening of metal objects, but these implements do occur, albeit rarely, in the Iron Gates area long before the appearance of metal. They were found at Velesnica, where they were used for the fine sharpening of bone awls and needles (Fig. 5: H). Working plates (Fig. 6: A) In Vlasac II and III, fragmented plates made of magmatic rocks were found. These were worked by hammering and grinding, with numerous circular recesses on the sides. How these plates were used is not clear; it is possible that they were utilized as working tables or as artefacts of ceremonial character (‘altars’; Srejović & Letica 1978: 100). At Velesnica the plates are made from sandstone and basic plutonic rocks. They are of irregular shape and have a ground working surface with circular recesses in it, and are interpreted as operating tables for working stone tools (Fig. 6: A).

Mallet-weights (Fig. 4) 4 As the name implies, these are massive pebbles used to shatter hard materials such as concreted soil or stone. The natural shape is altered only in the middle where a groove was made by hammering. They are recorded at Lepenski Vir and dated to LV Ie–II (Srejović 1972: 133). They were also found at Velesnica and Ajmana (Stalio 1986: 29), and in all the early Holocene strata at Padina and Hajdučka Vodenica (Radovanović 1996: 323). They are usually interpreted as fishing weights or as a kind of primitive mallet. The latter is more plausible given their size, weight and the used end. They are mainly made from large sandstone pebbles, and occasionally from certain other rocks (granite, granite gneiss).

Querns and mortars (Fig. 6: B–D) The querns are massive stone objects either with a flat working surface or with shallow hollows, used for grinding seeds. Mortars are, like querns, massive stone implements, but with deeper concave hollows resembling a recipient (vessel) on one side (see Radovanović 1996: 278). Following Radovanović (1996: 277–278) some of the ‘altars’ from Lepenski Vir and Velesnica (Srejović & Babović 1983: 158–172; Vasić 1986, 277, fig. 7) may be interpreted as querns and mortars. The querns/altars from Padina may also be included in this category (Fig. 6: B–C). Simple querns were found in Padina, Sector IV–Babice (Radovanović 1996: 341) and at Velesnica (Fig. 6: D). The querns are made of sandstone, seldom of granite and other magmatic rocks.

Hammerstones, anvils, pounders (Fig. 5: A–D, F, I) These terms refer to artefacts with the same pebble shape and partially reworked surfaces, but which had different uses. The hammerstones are ball-shaped, egg-shaped or discshaped, suitable for holding in the hand, and were used for stone working, as can be seen from scars or minor recesses (one or more) on their working surfaces. When working stone and bone implements, the tools could have been held in the hand (hammerstones; Fig. 5: B–D) or used as a support (anvil — Fig. 5: F, I). Some of these pebbles were used for seed grinding (pounders). They have been found in all Mesolithic horizons at Kula (Kula I–III; Sladić 1986: 432–433), and in Vlasac I–III (Srejović & Letica 1978: 99–100), Hajdučka Vodenica (Radovanović 1996: 323), Lepenski Vir (Srejović 1972: 133), Padina and Velesnica. They are made from sandstones, magmatic rocks and tuffs.

Sling-balls (Fig. 6: H–J) These artefacts are rare, for the reason that they were hunting weapons. Sling-balls are not genuine ground stone artefacts, because it is not clear whether they are naturally ground pebbles or shaped by hand. They were recorded in the Mesolithic strata of Vlasac (Srejović & Letica 1978: 100) and at Padina. Relatively small pebbles (2.5–5 cm diameter) of magmatic rocks were used. Fishing weights (Fig. 6: E–G) These are small tabular pebbles with minor shallow notches in the lateral edges at the narrower end, made by hammering and knapping, which were probably used for fixing a line around the implement. There is an opinion that these are amulets (Vasić 1986: 269). Such weights are the commonest ground stone artefacts at Velesnica. On the basis of their abundance, we are inclined to think that they are fishing weights (for both fishnets and fishhooks). They were also recorded at Knjepište and Ušće Kameničkog Potoka (Stanković 1988: 97).

Grindstones/polishers and whetstones (Fig. 5: E, G, H) The grindstones/polishers are implements used for grinding and polishing objects of hard materials (stone, bone or antler). Therefore, they are manufactured only of rocks with abrasive characteristics, such as sandstones, and very occasionally magmatic rocks with a high quartz content. Grinders/polishers for working awls and needles, with one or more grooves (Fig. 5: G), were found in Lepenski Vir I (Srejović 1972: 133), Vlasac I–III (Srejović & Letica 1978: 99) and Padina. They are manufactured from sandstones and magmatic rocks. Immobile grinders for working artefacts of hard materials (Fig 5: E), were found in Lepenski Vir I, in the form of large trapezoidal tables made from fine-grained red-

Pebble-axes (blunt axes; Fig. 7) These are massive implements of elongated form made of magmatic rocks, one end of which has a worked cutting edge 33

The Iron Gates in Prehistory Table 5. Lepenski Vir. % Artefacts

LV I

LV II

LV III

?5

Total

1. Ground edge tools

17.80

6.78

11.86

30.51

66.95

2. Hammers

1.70

0.85

0

0

2.54

3. Undefined fragments

2.54

0.85

0

3.39

6.78

4. Sceptres

16.95

2.54

0.85

0

20.34

5. Mallets

0.85

0.85

0

1.70

3.39

6. Pebble-axes

+6

+

7. Grinders/polishers

+

+

8. Necklaces, pendants

+

+

+

+

Table 6. Vlasac. % Artefacts

Vlasac I

Vlasac II

Vlasac III

Total

1. Sceptres

6.11

12.21

19.85

38.17

0

1.53

0.76

2.29

2. Pebble-axes 3. Grinders/polishers

0.76

1.53

1.53

3.82

4. Anvils, hammerstones

8.40

25.19

17.56

51.15

0

1.53

3.05

4.58

5. Sling-balls

Table 7. Padina. % Artefacts

Sector I

Sector II

Sector III

Total

1. Ground edge tools

10.66

1.64

13.93

26.23

2. Mallets

11.48

3.28

13.93

28.69

3. Anvils, pounders, hammerstones

7.38

17.21

5.74

30.33

4. Querns, altars

1.64

0

1.64

3.28

5. Grinders/polishers

0

0.82

4.92

5.74

6. Sceptres

0

0

3.28

3.28

7. Sling-balls

0

0

2.46

2.46

Table 8. Velesnica. Artefacts

%

1. Ground edge tools

8.26

2. Hammerstones, pounders

22.02

3. Grinders/polishers, whetstones

7.34

4. Querns

4.59

5. Mallets

2.75

6. Pebble-axes

4.59

7. Fishing weights

41.28

8. Working plates

0.92

9. Ring

0.92

10. Spool

1.83

11. Undefined fragments

5.51

34

Dragana Antonović: The ground stone industry in the Serbian Iron Gates

or has been rounded. Such tools have been found at Lepenski Vir on the floors of the houses in settlements Id-e (houses 1, 21, 27, 32 and 37; Srejović 1972: 133) and in Vlasac II–III (Srejović & Letica 1978: 99), Hajdučka Vodenica Ib (Radovanović 1996: 323) and in the earliest strata of Velesnica. The most developed shape of the pebble-axes occurs at Velesnica (Fig. 7: A–B). These are ellipsoidal pebbles that, besides having one end transformed into a cutting edge or rounded, are characterized by hammered regular grooves at the other end, which was used for fixing the tool on a handle. Such implements could be used as axes when one end was worked into a blade, or as hammers when the end had been rounded. They are most commonly made from sandstone.

the blade is not in the plane of symmetry of the tool (Fig. 8: N; Fig. 9: H, I). They were used only for woodworking (hollowing out and splitting) as visible from striations. Chisels are put into a special category on the basis of their smaller dimensions (Fig. 9: D, E, J–L). They were used in the same manner as adzes, but due to their small size, were also used for finer woodworking. Ground edge tools occur in Lepenski Vir I–III, Velesnica, Padina, Ajmana (Stalio 1986: 29), as well as in the Neolithic horizons of Vlasac (Vlasac IV; Srejović & Letica 1978: 135), Ušće Kameničkog Potoka (Stanković 1986a: 468) and Knjepište (Stanković 1986b: 448). They are manufactured from very fine-grained, compact and tough rocks, such as hornfelses, silicified pelites, limestones, fine-grained sandstones, diabases, andesites, gabbros, and other poorly defined fine-grained basic magmatic rocks, as well as phyllites, phyllite-mica schists, chlorite schists, eclogites and serpentinites.

Ornamental artefacts (Fig. 8: A–D) In the early Holocene Iron Gates sites small numbers of stone ornamental artefacts were found. They are made of both ornamental stones and gemstones, such as marble, crystalline limestone, ordinary limestones of attractive colours, as well as of malachite, azurite and jade. The ornamental objects were mostly found at Lepenski Vir, in all horizons of the site. They comprise pendants of various shapes, made of marble and crystalline limestones, and necklaces with beads manufactured of pink and white limestone, all of them being found in Lepenski Vir I–II (Srejović & Babović 1983: 190–197). From the Neolithic horizons of Lepenski Vir (IIIa–b) there are amulets in form of a labret, made of limestone, as well as pendants of malachite and azurite (Srejović & Babović 1981: 92, 94). A tabular piece of malachite, found at Padina, could be interpreted as a half-finished pendant. At Velesnica was found a jadeite/nephrite ring fragment. Here also can be mentioned some delicately worked flat winders, from Velesnica and Knjepište (Stanković 1986b: 448), which may represent some kind of ornament, but not an object for use in everyday life.

Hammers (Fig. 9: F) In Lepenski Vir I and II damaged and blunted ground edge tools were subsequently used as hammers. So far, these tools have been recorded only at this site, where they are made of limestone and granite. *

*

*

*

*

*

Tables 5–8 summarize the occurrence of the different types of ground stone artefacts at the sites of Lepenski Vir, Vlasac, Padina and Velesnica.

Conclusions The Iron Gates early Holocene ground stone industry is composed of two elements: local and ‘imported’. The local variant developed quite independently, against a background of sedentary settlement, satisfying the essential needs of the inhabitants of this region. This led to the production of mallet-sceptres, mallet-weights with grooves, pebble-axes, grindstones/polishers and hammerstones. Judging by their shape and considerable dimensions, these implements were probably used in game hunting and for killing large fish. On the other hand, because of the blunt cutting edges of the pebble axes and the massiveness of the tools themselves, it may be assumed that they were not used in woodworking or agriculture. In parallel with this autochthonous ground stone industry, there occur in the Iron Gates early Holocene sites tools typical of the Starčevo and Vinča cultures — the ‘imported’ variant. Thus, on the floors of the houses of Lepenski Vir I were found axes and adzes of Starčevo–Vinča type, along with mallet-sceptres typical of the Lepenski Vir culture. These two variants of the ground stone industry are of quite different character. In essence, the local variant is based on the pebble, its natural shape being only finished. On the other hand, the ground edge implements of the ‘imported’ variant are produced by working large flakes of fine-grained rocks with a characteristic grey and grey-greenish colour, thus having a distinctive appearance reminiscent of the Starčevo–Vinča ground stone industry. It is this raw material that makes the ground stone industry of the entire Serbian

Tranchets (Fig. 8: F–J) Although these artefacts can be regarded as products of the chipped stone industry, and have been studied and published as such (Radovanović 1981: 50; Kozłowski & Kozłowski 1984: 274; Srejović & Letica 1978: 98) they are mentioned here, because they are identical in shape to the unfinished ground edge tools from Neolithic localities in other parts of Serbia. The tranchets are made only from tuffs that, in their physical-technical features, resemble hornfelses (which is the basic material for the manufacture of ground edge tools in other parts of Serbia). They were used for the extraction of roots and other ‘agricultural’ activities, and have been found in Vlasac I, Padina A, Lepenski Vir III and Kula I (Sladić 1986: 432). The ‘imported’ ground stone industry Ground edge implements These comprise axes, adzes and chisels (Fig. 8: K–P; Fig. 9: A–E, G–L). In terms of shape, axes are implements with a symmetrical profile, the cutting edge occurring in the plane of symmetry (Fig. 8: L, P; 9: A–C, G). These are tools whose basic traits indicate that they were used for cutting trees. Adzes differ from axes by the asymmetry of the blade — i.e. 35

?

Vir III nsk i Le pe

Le pe

Le pe

nsk i

nsk i

Vir II

Vir I

The Iron Gates in Prehistory

Vir III nsk i Le pe

Le pe

nsk i

Vir II

Vir I nsk i Le pe

a–S ect

or III

or I Pa din

Pa din

a–S ect

ca sni Vel e

or II a–S ect

Pa din

Ku la

Vla sac

Figure 10. The frequencies of particular tool types in the major horizons of Lepenski Vir.

Figure 11. The frequencies of local and ‘imported’ ground stone industries at selected Iron Gates sites.

36

Dragana Antonović: The ground stone industry in the Serbian Iron Gates

Neolithic definitely recognizable. From these observations, the occurrence of ground edge implements, made of greygreenish rocks (whether hornfelses — the main Neolithic raw material in Serbia — or local rocks with the same physicaltechnical properties) indicates the import of this technology and, occasionally, the import of the artefacts themselves into the Iron Gates region. The common appearance of two variants of the ground stone industry, different in character, is perhaps further evidence of the parallelism of the Iron Gates Late Mesolithic with the developed Neolithic of the neighbouring areas, which is otherwise indicated by absolute dates (Radovanović 1996: 291–292). The implements of the ‘imported’ ground stone industry (ground edge tools — axes, adzes and chisels) were used only for woodworking, and occur in greater numbers in settlements showing more intensive construction activity. Thus, they are well represented at Lepenski Vir (where the largest number of houses is recorded) and Padina–sector III; although their complete absence from Vlasac is surprising. Both the local and ‘imported’ variants occur in parallel in the Iron Gates sites. In some sites artefacts of the local industry are more abundant, while in others there are more ‘imported’ artefacts. In yet other settlements the ‘imported’ ground stone variant is absent (Kula I–III, Vlasac I–III). On the basis of the presence of both local and ‘imported’ ground stone variants, the following sequence of the Iron Gates early Holocene settlements can be made: in Vlasac I–III and Kula I–III only tools of local variant occur; in Velesnica and Padina the ground stone implements are largely restricted to the local variant; whereas in Lepenski Vir the local variant is of greater importance only in the deepest levels, but it disappears in layers II and III (Figs 10 & 11). This is not a chronological sequence, but merely an example illustrating the extent to which some Iron Gates early Holocene settlements accepted Neolithic products from outside. In this respect, Lepenski Vir appears to a higher degree as a neolithized settlement based on hunting and gathering supported by intensive fishing. In contrast, Velesnica is far behind Lepenski Vir and, according to the ground stone material, was a highly developed fisher’s settlement. Here, it should be mentioned that the local variant continued its development during the Late Neolithic of the Iron Gates. A considerable number of ‘local variant’ tools (mallet-weights, sling-balls, hammerstones, pounders, querns, grindstones) were found, along with ground edge implements of Starčevo–Vinča type in the Vinča settlement of Zbradila near Korbovo (Babović 1984: 95–96, 1986: 96). When attempting to establish analogies and parallels with the ground stone industries of other regions, it has to be said that there is a remarkable similarity between the Iron Gates local variant and the ground stone products from very distant areas, with which the Iron Gates inhabitants were not in direct contact. The Mesolithic and Early Neolithic stone implements of the northwest regions of the European part of Russia show a close resemblance, in both form and raw materials, to the material from Lepenski Vir, Padina and Velesnica (Gurina 1961: 190–201, 1973: 46–49). There is a basic similarity also with the Late Mesolithic stone industries of the Dnestr–Don area (Levenok 1966: 93; Telegin 1982: 110), and the coastal

areas of eastern Germany (Gramsch 1973: 19–30) and Pomerania in Poland (Galinski 1992: 103, 156–157, 162). All of these cultures developed near large bodies of fresh- or salt water and fishing was an important occupation, although none of these cultures in terms of their character or territory is close to the early Holocene Iron Gates. Thus it is not a matter of cultural influence, but of a similar pattern of development of sedentary living near to a body of water, which provided a reliable source of food all year round. In all of the regions mentioned the development of stone implements was initiated by the need to exploit large fish. This shows once more that culture is a response to the natural conditions to which humans were trying to adapt. Therefore, it may be stated that in the Iron Gates too there was an autochthonous development of the ground stone industry (only in the form of the local variant), and this development was directed by the natural surroundings and living conditions of the Iron Gates inhabitants. The ‘local variant’ implements were adapted for the successful fishing and hunting economy, supported by various kinds of percussive [hammering] implements, whereas the ‘imported variant’ artefacts can be seen as a later addition to the quality of life of the Iron Gates population. Notes 1. The term working plate refers to a large stone with a flat or slightly concave working surface; usually they are thin tabular stones, but not exclusively. 2. More detailed and precise analyses of the raw materials used for the manufacture of the ground stone implements would give a clearer picture of the possible circulation of raw materials in these districts. 3. Inventory number in the National Museum, Belgrade. 4. These artefacts are most commonly called weights. 5. Lacking contextual and stratigraphic details. For the majority of ground edge tools, the field documentation does not give details of the find circumstances or stratigraphic position, but they were found along with pottery. This is not an automatic indication that they are restricted to the Lepenski Vir III level, because a large number of tools (of both local and imported variants) derive from the Lepenski Vir I and II levels, and the field documentation records the presence of advanced Starčevo pottery in those levels (Starčevo IIa–b in Garašanin’s classification). Such is the case with the ground edge tools that are registered in the Field Inventory (F.I.) under numbers 122, 123, 124 and 129, found on the floor of house 5 (LV I), the axe F.I. no. 369 from house 1 (LV Id–e), axe F.I. no. 459, found in house 30 (LV Ic), chisel F.I. no. 476 from the floor of house 9 (LV Id), axe F.I. no. 667 from the floor of house 32 (LV Id–e), chisel F.I. no. 738 from house 47 (LV Ia). Pottery is found on the floor of houses 1, 32 and 47, but was not recorded in houses 5, 9 and 30 (Srejović 1972: 133). 6. For some of the artefacts from Lepenski Vir it is known only that they existed at this site, but precise data on their frequency are not available.

Acknowledgments Thanks are due to colleagues who made it possible for me to study the field documentation from the Iron Gates sites, which has been an important source of data for this paper — Borislav Jovanović and Rastko Vasić of the Archaeological Institute in Belgrade, as well as Ljubinka Babović of the National Museum in Belgrade. Special thanks are due to the late Professor Dragoslav Srejović, who gave me access to the important material from Lepenski Vir.

1

37

The Iron Gates in Prehistory

References

phylliennes dans la partie centrale et orientale de l’autochtone danubien. In Association géologique Carpato–Balkanique, Ve Congrès, 4–19 septembre 1961, Bucharest. Vol. II: Communications scientifiques. I-ère section: minéralogiepétrographie. Bucharest: Association géologique Carpato–Balkanique, 167–179. Radovanović, I. 1981. Ranoholocenska kremena industrija sa lokaliteta Padina u Đerdapu. Belgrade: Arheološki institut. — 1996: The Iron Gates Mesolithic. Ann Arbor: International Monographs in Prehistory. Semenov, S.A. 1976: Prehistoric Technology. London: Moonraker Press. Sladić, M. 1986: Kula près de Mihajlovac — un site préhistorique. Đerdapske sveske 3: 432–442. Srejović, D. 1972: Europe’s First Monumental Sculpture: New Discoveries at Lepenski Vir. London: Thames & Hudson. Srejović, D. & Babović, Lj. 1981: Lepenski Vir: Menschenbilder einer frühen europäischen Kultur. Mainz: von Zabern. Srejović, D. & Babović, Lj. 1983: Umetnost Lepenskog Vira. Belgrade: Jugoslavija. Srejović, D. & Letica, Z. 1978: Vlasac: A Mesolithic Settlement in the Iron Gates. Vol. 1, Archaeology. Belgrade: Serbian Academy of Arts and Sciences. Stalio, B. 1986: Le site préhistorique Ajmana à Mala Vrbica. Đerdapske sveske 3: 27–50. Stanković, S. 1986a: Emboucher du ruissen Kamenički potok site du groupe de Starčevo: compte-rendu des fouilles de 1981. Đerdapske sveske 3: 467–471. — 1986b. Localité Knjepište — une station du groupe de Starčevo; fouilles de 1982–1983. Đerdapske sveske 3: 447–452. — 1988: Šljivik — Stragari, Trstenik. In Srejović, D. (ed.) The Neolithic of Serbia. Belgrade: Faculty of Philosophy — Centre for Archaeological Research, 95–97. Telegin, D.Ja. 1982: Mezolitični pam’jatki Ukraini (IX–VI tisjačolittja do n. e.). Kiev: Naukova dumka. Vasić, R. 1986: Compte-rendu des fouilles du site préhistorique à Velesnica. Đerdapske sveske 3: 264–277. Vasković, N. & Matović, V. 1997: The Hercynian granitoides of Djerdap. In Grubić, A. & Berza, T. (eds) Geology of Djerdap Area. International Symposium, 23–26.IX.1997, Donji Milanovac–Orsova. Belgrade: Geoinstitute / Bucharest: Geological Institute of Romania, 129–141.

Antonović, D. 1992: Predmeti od glačanog kamena iz Vinče. Belgrade: Filozofski fakultet — Centar za arheološka istraživanja. — 1998: Nastanak i razvoj industrije glačanog kamena u neolitu Srbije. Unpublished PhD dissertation, Department of Archaeology, Faculty of Philosophy, University of Belgrade. Babović, Lj. 1984: Zbradila, Korbovo: izveštaj o arheološkim istraživanjima u 1980. godini. Đerdapske sveske 2: 93–100. — 1986: Zbradila — Korbovo: compte-rendu des fouilles en 1981. Đerdapske sveske 3: 95–115. Galinski, T. 1992: Mezolit Pomorza. Szczecin: Muzeum narodowe. Geological Map of Yugoslavia 1970: 1:500,000. Belgrade: Federal Geological Institute. Gramsch, B. 1973: Das Mesolithikum im Flachland zwischen Elbe und Oder, vol. 1. Veröffentlichungen des Museums für Ur- und Frühgeschichte Potsdam 7. Berlin: Deutscher Verlag der Wissenschaften. Gurina, N.N. 1961: Drevnjaja istorija severo-zapada evropeiskoi časti SSSR. Materialy i issledovanija po arheologii 87. Moscow–Leningrad: Akademija nauk SSSR. — 1973: Drevnie pamjatniki Kol’skogo poluostrova. In Gurina, N.N. (ed.) Etnokul’turnije obštnosti lesnoj i lesostepnoj zony evropejskoj časti SSSR v epohu neolita. Materialy i issledovanija po arheologii 172. Leningrad: Akademija nauk SSSR, 45–53. Kovačević, J., Jovanović, M. & Radošević, B. 1997: Nonmetallic mineral raw materials in Djerdap area. In Grubić, A. & Berza, T. (eds) Geology of Djerdap Area. International Symposium. Belgrade: Geoinstitute / Bucharest: Geological Institute of Romania, 169–174. Kozłowski, J.K. & Kozłowski, S.K. 1984: Chipped stone industries from Lepenski Vir, Yugoslavia. Preistoria Alpina 19: 259–293. Levenok, V.P. 1966: Mezolit srednorusskogo DneprovskoDonskogo meždureč’ja i ego rol’ v složenii metnoj neolitičeskoj kul’turi. In Gurina, N.N. (ed.) U istokov drevnih kul’turi (epoha mezolita). Materialy i issledovanija po arheologii 126. Moscow–Leningrad: Akademija nauk SSSR, 88–98. Marić, L. 1945: Sistematska petrografija. Zagreb: Nakladni zavod Hrvatske. Pavelescu, L. 1963: Contribution à l’étude du soubassement cristallin et l’aire de distribution des différentes formations cristallo-

38

Sturgeon fishing in the middle and lower Danube region László Bartosiewicz, Clive Bonsall & Vasile Şişu

Abstract: Migrating sturgeons were the largest fish in the middle and lower Danube region. Most of these species, however, have been brought to the brink of extinction by habitat loss and overfishing. This review is a synthesis of sporadic archaeological evidence, zoological and environmental data as well as ethnohistorical information in two regions: the Iron Gates at the southeast edge of the Carpathian Basin and the Hungarian section of the Danube within the basin. In addition to ichthyological and taphonomic questions, fishing techniques as well as the varying perceptions of these large fish are summarized in an attempt to draft a multidisciplinary interpretive framework for the archaeological evaluation of future finds. Key words: sturgeon, Acipenseridae, fishing, weirs, Danube, Iron Gates, seasonality

Introduction

Although the disciplinary boundaries between these areas of research often overlap, this list is intended to provide an interpretive framework for archaeologists interested in any period, touching upon the complex interactions between nature and society as reflected in sturgeon fishing. Our research hypothesis is that it should be possible to outline geographical locations as well as seasons when the probability of catching various species of sturgeon was increased. These parameters should correspond to and complement the scanty archaeological evidence for these important fish. This overview of zooarchaeological, documentary and ethnohistorical evidence of sturgeon fishing in the middle Danube region represented by two contrasting sections of the river, in Hungary and the Iron Gates (Fig. 1), is also aimed at interpreting ichthyoarchaeological data within a broader, culture–historical context.

Sturgeon has become one of the most elusive animals in modern day fishing. Its osseous remains, decimated by taphonomic loss in archaeological deposits, yield relatively scarce evidence of their dietary role in past times. Their dwindling stocks, brought to the brink of extinction in the 20th century, are only a pale shadow of their economic importance until the recent past. Over-exploitation, habitat loss and pollution have severely hit all 27 species in the Acipenserid family worldwide. Sturgeons, the largest fish in the Danube, were relatively common until dams were built in the Iron Gates section where the river forms the border between Romania and Serbia. The first dam (Iron Gates I) which became operational in 1971, was built where the Danube leaves the Iron Gates gorge, and effectively marks the divide between the middle and lower Danube. The dam was designed in part to improve navigation through the 130 km long gorge section, where the Danube (prior to impounding) was characterized by strong currents, rapids and exposed rocks that were hazardous to shipping. The second dam (Iron Gates II), operational in 1984, is located 80 km downriver, 875 km from the river’s mouth. The dams effectively cut off the migration route of endangered beluga sturgeon and other anadromous fish in all major sections of the Danube, as reviewed at the 1994 International Conference on Sturgeon Biodiversity and Conservation in New York (Bacalbaşa-Dobrovici 1997; Hensel & Holčík 1997). Against this background, reconstruction of ancient sturgeon fishing becomes a truly multidisciplinary task: 1. Ichthyoarchaeological finds offer evidence of which species were targeted. 2. The palaeohydrological reconstruction of riverine habitats helps in identifying locations where sturgeon fishing may have taken place. 3. Familiarity with fish behaviour points to seasons when migrating sturgeons were most easily caught. 4. Historical accounts describe techniques by which sturgeons were caught. 5. Ethnohistorical records reveal attitudes toward these great fish as food as well as symbols.

Figure 1. The Danube Valley in Europe and areas discussed in this study (see also Figures 7–8).

39

The Iron Gates in Prehistory

mains represented in the Schela Cladovei assemblage, dorsal and lateral scutes as well as pectoral fin rays could, to some extent, be identified to species. Osseous elements from small but mature individuals could be assigned to sterlet on the basis of size. The first (i.e. most cranially located) pectoral fin ray is well developed and has been used in osteometric analyses of common sturgeon by Desse-Berset (1994: 84). The mediolateral width of this bone in beluga sturgeon (Bartosiewicz & Takács 1997: 12, fig. 8/1) has also been used in size estimations in the present paper. Greatest lengths of the Schela Cladovei sturgeons were estimated as shown in Table 2. When plotted together with modern, historical data (Khin 1957), the distribution of these estimated lengths largely corresponds to that of the largest specimens from historical periods (Fig. 2). The latter display a slight positive skew, and the distribution of the prehistoric specimens is within the same range, suggesting that randomly caught prehistoric sturgeons were as large at Schela Cladovei as the largest modern specimens in Hungary. While, owing to the small number of cases no significant difference can be observed between the two groups, the Schela Cladovei sturgeons were indubitably large. Sporadic records of the amount of meat some record animals represented (Table 3) indicate that approximately two-thirds of the live weight estimated for large prehistoric sturgeons represented edible protein, and other lines of evidence show that fish were the major source of animal protein in the diet of the Mesolithic inhabitants of Schela Cladovei (Bonsall et al. 1997, 2000). Ichthyorachaeological research has shown early signs of overfishing common sturgeon (Acipenser sturio Linné 1758) in the southern Baltic region (Benecke 1986: 16, fig. 1) at archaeological sites in Gdańsk, Poland (10th to 13th century AD) and Ralswiek, Germany (8th to 12th century AD). Osteometric data show that these sturgeons were at least larger than the 1.4 m long modern reference specimen available to that author (Benecke 1986: 17, fig. 2). Recent sturgeon sizes in the Danube are also worth considering. Aside from a random element (fishermens’ luck), the frequency and actual size of the largest individuals landed

Figure 2. Lengths of Danubian sturgeon in the historical record (Hungary) and by prehistoric estimates (Iron Gates).

Ichthyoarchaeological data This study is built around the osteological evidence for sturgeon from the Danube. Sturgeon remains recovered from archaeological sites are discussed in terms of taxonomic identifiability and taphonomic bias, as well as possibilities of size reconstruction. The Romanian–British excavations (1992–1996) at the Late Mesolithic and Early Neolithic settlement of Schela Cladovei (Romania) 7 km downriver from the Iron Gates I dam brought to light 139 sturgeon bones, mostly identifiable only to family level (Table 1). These remains have been recovered by water sieving from several features, located on one of the river’s terraces, near the present day riverbank. Although these numbers of sturgeon bones may seem modest, given the sample size of hand-collected bones at this site (Bartosiewicz et al. 2001: 16, table 1) and others in the region they may be considered relatively high. Of the skeletal re-

Table 1. Sturgeon bones from Schela Cladovei. Mesolithic Acipenserid frontale parashenoideum praemamillo-maxillare dentale praeoperculare operculare cleithrum dorsal scute lateral scute pectoral fin ray dorsal/anal fin ray flat bone Total number Find weight (g)

Russian Sturgeon

Early Neolithic Sterlet

Acipenserid

Sterlet

3 5 1 1 1 1 10

1

5 7 2 75

2

108 287.5

2 2.4

40

1 1 3 2

2

2

2 1 6

10 8.2

14 63.6

1 1 3

5 2.2

László Bartosiewicz et al.: Sturgeon fishing in the middle and lower Danube region Table 2. Length and weight estimates for prehistoric sturgeons from Schela Cladovei. Period Mesolithic Mesolithic Mesolithic Mesolithic Neolithic Neolithic Neolithic

GW fin ray (mm)

Estimated length (m)

Estimated live weight (kg)

24.4 36.4 43.3 50.0 37.5 42.0 44.1

1.44 2.15 2.56 2.95 2.21 2.48 2.60

28.2 72.8 110.1 154.0 77.7 102.1 114.2

Table 3. The meat output of some record sturgeons from Hungary. Year

Location

Live weight (kg)

Carcass weight (kg)

Live weight (kg)

1922 1957 1987

Gemenc Paks Paks

90 138 181

70 100 100*

70.0 72.0 55.2*

* Pure meat with spine but without skin

Figure 3. Diachronic temporal decline in gross sturgeon catch in Romania and in the size of record sturgeon in Hungary.

Sturgeon taxonomy

depends on the reproductive capacity and growth characteristics of fish stocks (Miranda et al. 1987: 219). Since 1800, an estimated 1.84 kg average annual decrease in record sturgeon body weights was found to be statistically significant (Bartosiewicz & Takács 1997: 9). This decrease is paralleled by a decline of the overall weight of sturgeons landed in Romania (Fig. 3). The unusually large, 181 kg specimen caught at Paks (Hungary) in 1987 (Fig. 4), nearly two decades after the closure of the Iron Gates 1 dam (Pintér 1989: 24), may have attained this respectable size after having been trapped upstream, behind the dam.

In spite of the difficulties of identification at a species level in this family of fish, the body dimensions as well as skeletal variability among, and habitat preferences of, Acipenserids are worth considering. Appraising these parameters is of help in defining ranges within which archaeological finds can be better understood. Sturgeons (Acipenseriformes) are large (80–600 cm), late maturing (almost 25 years for some females) fish of concomitant longevity. They inhabit coastal sea waters, rivers and lakes, mostly within a latitudinal range from 30–70° N 41

The Iron Gates in Prehistory

Figure 4. Sturgeon, weighing 181 kg, caught at Paks (Hungary) in 1987. Photo: István Takács.

Table 4. Characteristic dimensions of Danubian sturgeons. Species Beluga sturgeon Common sturgeon Russian sturgeon Ship sturgeon Stellate sturgeon Sterlet

Huso huso Linné 1758 Acipenser sturio Linné 1758 Acipenser gueldenstaedti Brandt 1833 Acipenser nudiventris Lovetzky 1828 Acipenser stellatus Pallas 1771 Acipenser ruthenus Linné 1758

(Hankó 1931: 9). They feed on small animals, including molluscs, crustaceans and small fishes (anchovies, sprats, gobies) as well as plants, and are anadromous, i.e. migrate up-river to breed and spawn. Because of this migratory behaviour, Luigi Ferdinando Marsigli, an 18th century military engineer surveying the lower Danube valley, classified beluga sturgeons as fluviatiles marini, marine fish that live in rivers (Marsigli 1726). Systematic work on the complex taxonomy of sturgeons began relatively recently (Berg 1904). As with the archaeological record, most historic references to sturgeon are vague as to species identification. This is related to the fact that some sources not only predate Linnæan nomenclature, but the living fish species are also often difficult to tell apart. This is, to some extent, a consequence of natural hybridization. The main dimensions of species relevant to this study are summarized in Table 4. The large species central to this study include beluga sturgeon, Russian sturgeon, ship sturgeon and stellate sturgeon (Fig. 5). Occurrences of common sturgeon have been reported from the Danube only relatively recently. According to Antipa (1905) these fish spawn on the sandy sea bottom before the Danube estuary. Some distribution maps (Muus & Dahlström 1965: 63, fig. 26; Terofal 1971; Vuković & Ivanović 1971: 111; Wheeler 1978: 56; Müller 1983: 123, fig. 28) suggest that it does live in the Danube. Other sources, 42

Total length (m)

Live weight (kg)

2–3 (max. 10) 1.5–2.5 (max. 3.5) 2–2.5 2 1.5–2 1–1.2

80–100 (max. 1000) 70–100 (max. 320) 80–100 40–50 30–50 15–16 (max.)

however, assert that common sturgeon is a fish of the Atlantic/Baltic region (Maitland & Linsell 1978: 78; CurryLindahl 1985: 230). This species certainly has not been recorded in the Hungarian section of the Danube (Berinkey 1966: 17; Pintér 1989: 24), and a reference to its occurrence in Transylvania has also been questioned by Hankó (1931: 9). Although common sturgeon was mentioned in the discussion of ‘Neolithic’ Padina (Clason 1980: 167), that faunal list only contains the item ‘Acipenser sp./Huso huso’ (Clason 1980: 149). Sterlet is the smallest Acipenserid in the region. It is not anadromous, having adapted to freshwater conditions. Owing to its smaller size and behaviour it falls beyond the focus of this paper. Osteology and taphonomy The selective survival of different animal remains has a direct bearing on archaeological interpretations. Studying the post mortem history of excavated bone is indispensable for the critical understanding of archaeozoological assemblages. Taphonomic analyses must be based on familiarity with the original, complete skeleton. Interspecific comparisons between morphologically similar Acipenserid species are made difficult by at least two skeletal characteristics of these fish:

László Bartosiewicz et al.: Sturgeon fishing in the middle and lower Danube region

1m Figure 5. Sturgeon species of major importance in the Danube, drawn to scale on the basis of mean lengths in Table 4 (compiled and redrawn to scale after Berinkey 1967 and Pintér 1989).

1. The sheets of dermal bone covering the head of Acipenserids tend to be numerous and irregular in shape to such an extent that they may even be strongly asymmetric within the same individual. 2. While bones from large fish would more likely be recovered, their survival is poorest in old Acipenserids (especially beluga sturgeon), whose skeleton reabsorbs minerals with the advancement of age. Thus, the largest bones tend to be most easily destroyed or eroded beyond recognition in archaeological deposits. Brinkhuizen (1986) reviewed differences between the usually resistant and morphologically most characteristic dermal scutes of beluga and Russian sturgeons from the Iron Gates. The dorsal scutes of beluga sturgeon are oval in shape with an elongated, horn-like process. In mature individuals, these scutes are covered by skin. The lateral scutes of beluga sturgeon are toothed. In old individuals, they are partially reabsorbed and develop a spongy, eroded look. The dorsal scutes of common and stellate sturgeon display a more marked morphological difference. These bones are almond-shaped with roof-like cross-sections. The medial

edge of the ‘roof’ is largely symmetric in common sturgeon, while it is slightly skewed in a cranial direction in stellate sturgeon giving it a rose-thorn profile line. A strong radial pattern is also characteristic of stellate sturgeon. In other Danubian Acipenserids, even this character looks transitional between the two morphological extremes. It is chiefly large and compact dermal scutes with characteristic surface patterning that show up even in handcollected assemblages, which otherwise contain few fish remains. These scutes are arranged in dorsal, lateral and ventral rows along the body of sturgeons and differ both in size and shape by anatomical location (Casteel 1976: 38, fig. 19). According to Berinkey (1966: 18–22), Vuković & Ivanović (1971: 106–112), Pintér (1989: 24–31) and a review of nine authors by Brinkhuizen (1989: 41), the number of dermal scutes and fin rays varies between the discussed species as shown in Table 5 and Figure 6. As is shown by these data, dermal scutes not only form continuously changing rows on individuals of varying sizes, but also their numbers differ by species which makes their quantitative analysis a nightmarish enterprise. 43

The Iron Gates in Prehistory Table 5. The number of some skeletal elements in Danubian Acipenserids. Species

Fin rays

Beluga sturgeon Common sturgeon Russian sturgeon Ship sturgeon Stellate sturgeon Sterlet

dorsal 62–73 31–43 33–51 45–57 40–46 37–54

Huso huso Linné 1758 Acipenser sturio Linné 1758 Acipenser gueldenstaedti Brandt 1833 Acipenser nudiventris Lovetzky 1828 Acipenser stellatus Pallas 1771 Acipenser ruthenus Linné 1758

anal 28–41 22–27 21–33 23–37 24–29 19–31

Scutes dorsal 9–17 9–16 5–19 11–17 9–16 12–17

lateral 37–53 24–39 24–50 50–74 26–43 57–71

Branchiospinae ventral 9–12 9–14 14–16 11–17 9–14 10–19

17–31 18–25 15–31 24–36 24–26 16–21

Figure 6. Variability in the number of dermal scutes in Danubian acipenserid species.

Habitat reconstruction

and variation in discharge rate affect the abundance and distribution of anadromous fish populations (Schalk 1977). The quantity of oxygen dissolved in water (y mg/l) is a function of current speed (x1 m/s) and water temperature

The presence/absence of sturgeon bones at archaeological sites is not only a matter of preservation. It may be presumed that, at least in prehistoric times, primary butchery of large sturgeon took place near where the fish was landed. The Schela Cladovei finds on a low-lying terrace of the Danube are indicative of this tendency. In later periods, carcass dismemberment may have taken place away from where the fish was pulled ashore, at markets or high status sites of consumption. Prehistoric findspots (Fig. 7) as well as historically recorded catch sites (Fig. 8) suggest that understanding habitat preferences of sturgeon may point to locations where they could be caught most efficiently — information whose intimate knowledge was essential to fisherfolk throughout the millennia.

x1 = speed of current: x2 = temperature:

y = 1.953x + 1.984 y = -0.221x + 13.669

r = 0.943 r = -0.979

(x2 ºC). On the basis of empirical data published by Pénzes and Tölg (1977: 327, table 4) as well as Harka (1993) this relationship was expressed using the following regression equations by Bartosiewicz & Bonsall (2004: 263): The regression coefficients express the idea that every additional 1 m/s in water velocity increases dissolved oxygen content by almost 2 mg/l in rivers, while a 1 °C rise in temperature would result in a 0.2 mg/l loss. This is why oxygenloving sturgeons seek rapid rivers during the spring for sterlet (Acipenser ruthenus Linné 1758) pikeperch (Stizostedion lucioperca Linné 1758) tench (Tinca tinca Linné 1758)

Water properties Anadromous Acipenserids migrate into rivers for spawning: for this purpose they need deep, well aerated waters with a hard, preferably sandy or rocky substrate. Geographical latitude/climate, distance upstream from the river’s mouth,

3.0–3.5 mg/l 2.0–3.0 mg/l 0.7 mg/l

spawning, and why water temperature is of decisive importance in the timing of their migration. The minimum requirements of dissolved oxygen by three characteristic freshwater 44

László Bartosiewicz et al.: Sturgeon fishing in the middle and lower Danube region

Figure 7. Archaeological sites and sturgeon bone finds in the Iron Gates gorge section of the Danube.

Figure 8. Archaeological sturgeon finds and the occurrence of 19–20th century record specimens (full circles) in presentday Hungary. Archaeological site codes: 1= Tiszaföldvár (prehistoric), 2=Tiszaug (prehistoric), 3=Ács (Roman period), 4=Szentendre, 5=Esztergom, 6=Pilisszentkereszt, 7=Visegrád, 8=Vác, 9=Szentendre, 10–12=Buda Castle, 13=Sárszentlőrinc.

45

The Iron Gates in Prehistory

Figure 9. The relative frequency of sturgeon bones in hand-collected fish assemblages along the Iron Gates section of the Danube (after Bartosiewicz & Bonsall 2004: 266).

fish species are as follows (Pénzes & Tölg 1977: 327): Although sterlet is not an anadromous species, it clearly illustrates the highest requirement of dissolved oxygen of Acipenserids in this comparison. As the speed of the river is greater towards its source (and its temperature tends to decrease with increasing altitude), the further upstream sturgeons move, the better the circumstances for spawning. Aeration parameters were summarized by Harka (1993). From the tabulated summary of his data it is evident that foothill and lower foothill river sections with 3.0 to 4.0 mg/l of dissolved oxygen would provide ideal spawning grounds for sturgeons (Bartosiewicz & Bonsall 2004: 263, table 4). Even during the 20th century, beluga sturgeons as heavy as 900 kg used to be caught by the Donji Milanovac fishermen inside the Iron Gates gorges. Naturally, the distribution of such spots varies with the discharge along a river as determined by climate and topography. The c. 2300 m 3 /s average discharge of the Danube at Budapest more than doubles to over 5600 m3/s in the Iron Gates gorge. Before the construction of the Iron Gates 1 dam, this c. 130 km long section of the Danube was characterized by extreme changes in water levels. Minimum discharge was 1400 m3 /s, while 16,000 m 3/s values were also measured (Bǎncilǎ et al. 1972: 9). Moreover, prior to dam closure, the riverine environment of the Iron Gates gorges was characterized by strong currents, hard substrates, and was rich in nutrients, aquatic plants, insects and invertebrates (e.g. Gammaridae and Corophiidae sp.) that sustained rich and varied fish resources. Thus, this section of the Danube provided an ideal habitat for large sturgeons as well as sterlet.

Thus the net gradient is three orders of magnitude different (40% vs 0.037%) not to mention the considerable differences in topography and relief between these two sections of the Danube valley. Lake sturgeon in Canada spawn in rivers at depths of c. 0.5–5 m, in areas of swift water or rapids at the foot of low falls that slow down further migration (Needs-Howarth 1996: 149). Given the mass movement during the spawn run, such places must have been packed with sturgeons of all sorts in the Iron Gates gorge as well, making them increasingly vulnerable to human predation at these natural traps. This possibility seems to be supported by the differential proportion of sturgeon bones to those of other large fish in hand-collected archaeozoological assemblages from sites at different locations within the Iron Gates (Fig. 9; Bartosiewicz 1996, 1997). The Late Mesolithic/Early Neolithic site of Schela Cladovei, located downstream from the rapids, prior to dam construction marked the exit from the Iron Gates gorge, and must have been one of the ideal fishing spots where great numbers of migrating fish could be targeted. The reported absence of sturgeon bones at the prehistoric sites of Lepenski Vir and Vlasac raises the question of author-related bias: the bones were identified by Sándor Bökönyi, a leading expert in mammalian osteology. Did some of the large ‘catfish’ bones identified by Bökönyi (1969, 1978) originate from sturgeon? However, if nothing else, the ornate head bones and unmistakable dermal scutes of sturgeons would surely have attracted his attention since they can be most easily recognized in excavated assemblages (Desse & Desse-Berset 1992); Bökönyi himself identified sturgeon remains at the Neolithic site of Mihajlovac–Knjepiste (Bökönyi 1992: 79). More recent work suggests that Acipenserid bones did occur on sites upstream from Schela Cladovei, including sites within the Iron Gates gorges. Borić & Dimitrijević (2005) have reported sturgeon (beluga) bones from the floors of some of the trapezoidal buildings at Lepenski Vir. In addition to sterlet, commonly represented at prehistoric sites along the Danube in Romania, remains of great sturgeon (beluga) were also reported from the Epipalaeolithic sites of Ostrovul Banului and Icoana. Bones of Epipalaeolithic Russian sturgeon were also identified at Cuina Turcului (Pǎunescu 2000: 342). In the time of the

Topography and changes in riverbed gradient The river was confined to a width of only 170 m in the Khazan gorge. As is shown by the pre-regulation measurements of the river from 1872 (Bartosiewicz & Bonsall 2004: 265, fig. 6), depths varied tremendously, between 0.5 and 50 m before the river exited to the plain. In the lower part of the Iron Gates gorge, upstream from Schela Cladovei (near Turnu-Severin) in Romania, the riverbed has a very steep gradient, falling 8 m in only 20 km. By contrast, over the 935 km between Schela Cladovei and the Black Sea, the riverbed declines overall by only 34 m (Giurescu 1964: 101). 46

László Bartosiewicz et al.: Sturgeon fishing in the middle and lower Danube region

Figure 10. Schela Cladovei and major sturgeon catch sites along the Danube.

Boian and Gumelniţa cultures, anadromous sturgeons were also caught at sites along the lower Danube, notably Borduşani, Hârsova and Isaccea (Radu 1997, 2003). When some historically renowned sturgeon fishing grounds are plotted along the entire section of the Danube (Fig. 10), many of them pinpoint locations downstream from reaches with a steep gradient in the riverbed (large sturgeon were regularly caught at Tulln, upstream from Vienna, and a 1692 record is known from Bavaria). In addition to the river’s gradient, changes along its course also created opportunities for sturgeon fishing. Until the mid-19th century, the Danube meandered through floodplains and wetlands, shifting its course with each spring flood. Fish turned such marshes, e.g. around Szentendre, north of Budapest, into rich hatcheries. In the plains, the variable course of the river resulted in underwater shoals, fords, smaller and larger islands that all influenced currents and created spots where sturgeons could be caught more easily. In 1690, “50–100 sturgeons were caught and butchered daily” at the island of Ada-Kaleh, downstream from Orşova in the Iron Gates (Marsigli 1726). Plotting the catch sites of 19th–20th century record specimens on the map of present-day Hungary (cf. Fig. 8), the horizontal patterning appears more homogeneous, although confluences with tributaries and major river bends seem to be indicative of good sturgeon fishing grounds. Sturgeon were regularly caught in many of the Danube’s tributaries, including the Váh, Maros and Tisza rivers (Hankó 1931: 9). In the latter, sporadic prehistoric bone finds as well as modern records (Bartosiewicz 1999) confirm Hankó’s statement (cf. Fig. 8). In 1518, following a long medieval tradition, the city of Komárom in northern Hungary was given the rank of Royal Sturgeon Fishing Grounds (Herman 1980: 267). This strategically important point is located at the confluence of two branches of the Danube, downstream from

Europe’s largest inland river delta where sturgeon were already caught in Roman times (Bartosiewicz 1989: 611). The monopoly of Komárom fisherfolk can still be detected in the 18th century, when they let sturgeon fishing rights from as far as the Tisza river between Tiszacsege and Tiszafüred (Bencsik 1970: 98). The mouth of a small left bank tributary of the Danube, across from the northern tip of Margaret Island in Budapest, is called ‘Sturgeon Catcher’, another indication that such natural topographic features were exploited in sturgeon fishing. Sturgeon shoals thinned out as the river was regulated and walled off from the floodplain (Woodard 2000). Large-scale commercial navigation along the Danube also had a negative impact as it became necessary to keep the riverbed clean of sand and gravel deposits. By the mid-20th century, beluga sturgeon seldom swam upstream beyond the Iron Gates.

Sturgeon behaviour Sturgeons may live for up to a 100 years. Male beluga sturgeons become sexually mature by the age of 12 years. Females usually start spawning at 18 years (Deckert 1967: 66). Prior to sexual maturity, Acipenserids approach the coastal areas of the Black Sea. Following a short period of adaptation, they start moving upstream into the rivers. From that point onwards, patterns of migration and seasonality directly influenced the success of sturgeon fishing inland. Some stocks already begin to move into the Danube delta in September–October. Others begin the spawning run in the sea in March–April. By late summer, young sturgeons move into the lower Danube. They reach sexual maturity by next spring (Table 6). The arrival of these immature individuals, as well as their prolonged presence in rivers seems to indicate 47

The Iron Gates in Prehistory Table 6. Spawning parameters of various sturgeon species in the Danube (Berinkey 1966). Species Beluga sturgeon Common sturgeon Russian sturgeon Ship sturgeon Stellate sturgeon Sterlet

Huso huso Linné 1758 Acipenser sturio Linné 1758 Acipenser gueldenstaedti Brandt 1833 Acipenser nudiventris Lovetzky 1828 Acipenser stellatus Pallas 1771 Acipenser ruthenus Linné 1758

Spawning time

Water temperature

March–May April–May April–May April–June, April–June April–June

9 ºC 8–18 ºC 10–17 ºC — 10–17 ºC 12–17 ºC

Figure 11. Seasonal variations in river discharge in the Iron Gates in ‘characteristic years’ prior to dam construction. Inset: mean annual values. (after Bǎncilǎ et al. 1972).

that they are keen on adapting to fresh water. Groups of migrating sturgeon, therefore, tend to include individuals of different ages and sizes, as was observed in North America (Roussow 1957). Beluga sturgeon is the first to move into fresh water, as soon as icy floods are over (Hankó 1931: 9). They are followed by Russian and stellate sturgeon swimming up to the Iron Gates. All three species are actively targeted by fishermen between January to June as well as October to December. These two periods largely correspond to the spring and autumn migrations in the Iron Gates gorge. Most notably, they also coincide with seasons of high discharge in this region (Fig. 11; Bǎncilǎ et al. 1972: 19), when low water temperatures and high water velocity favour spawning. Sturgeon fishing between June and September only took place opportunistically in the Iron Gates when water temperatures were high and discharge low. Mátyás Bél (1764: 39), an 18th century Hungarian naturalist, documented the same two seasons in Hungary: spring fishing began in March and continued uninterrupted until June (cf. Table 7). The autumn season lasted from August until December, unless winter began early. These data contradict somewhat 19th century records that in Hungary only 7–15% of the yearly sturgeon catch were landed between February and April, while

41–57% were caught between September to November. The rest were caught opportunistically throughout the year (Khin 1957). Identifying the season of death may be attempted by thin-sectioning pectoral fin rays from archaeological sites. The low precision of this method, however, makes the identification of fish caught during the spawn run contradictory (Desse-Berset 1994; Needs-Howarth 1996). Seasonal differences in fish size are also difficult to establish. Some 19th–20th century record specimens of known date are listed by season in Table 7. Owing to the random nature of the catch, no significant seasonal weight or length difference could be detected between record sturgeons. These data reconfirm, however, that equally large sturgeons could be caught both during the spring and in the late autumn/winter. Interspecific differences in spawning temperatures, as well as the presence of off-season stragglers meant that sturgeons were available year round. The use of their bones as seasonal indicators in archaeological assemblages, therefore, is limited to probabilistic interpretations. It would seem logical that sturgeons could be most successfully targeted when they were rushing upstream along the Danube in great numbers. However, this is contradicted by the 41–57% September to November catch statistics given previously. 48

László Bartosiewicz et al.: Sturgeon fishing in the middle and lower Danube region Table 7. Measurements of sturgeon caught at known seasonal dates in the Hungarian section of the Danube. Year

Weight (kg)

Length (cm)

138 90 132 63 86 181 134 50

273 220 300 215 235 300 280 203

109.3 44.2

253.3 39.5

102 117 250 100 117 63

252 263 365 200 263 213

Mean value, spring Standard deviation

109.3 44.2

253.3 39.5

P-value of Student’s t-test

0.600

0.819

1957 1922 1950 1936 1927 1987 1957 1954

Date March 2 March 3 March 8 March 18 ‘Good Friday’ May 16 May May 27

Location Paks Gemenc Százhalombatta Dunapataj Dunapentele Paks Baja Paks

Mean value, spring Standard deviation Winter 1911 1956 1870 1894 1955 1953

early September October 6 ‘winter catch’ ‘winter catch’ January 31 February 1

Orsova (Iron Gates) Ercsi Ásvány-Győr Pozsony Ercsi Ercsi

Historical accounts of sturgeon in the Danube

a maze and made from timber fencing which stretched across the Volga river. Sturgeons swimming upstream were trapped in the four, curvaceous ‘pockets’ of this complex structure (Khin 1957: 12, fig. 7). Nineteenth century sturgeon traps in the Iron Gates gorge were described by Jókai (1872: 7) as follows: “Between the islets, the narrow branches of the Danube are disrupted [by] double post structures made from robust timber, arranged in a V-shape, opening downstream... Once the sturgeons enter, it is not their habit to turn downstream. As they proceed in the ever-narrowing funnel, they wind up in the ‘death chamber’ at the end...”. Mixed-media weirs included the use of strong nets as well. Khin (1957: 14–15) refers to a 16th century description by Miklós Oláh. Wooden posts were staked across the Danube in November, before the waters grew icy. Fishermen in boats stretched nets between the evenly placed posts, then cannons were fired to scare sturgeons into the traps. Medieval documents suggest that full closure of rivers would have been considered an unfriendly or even illegal move. A 1528 court case in Hungary between the city of Vác and the royal capital, Buda (located downstream), clearly shows the competing interests of their sturgeon fishing communities (Szilágyi 1995: 114). Weirs, therefore, were often placed between the bank and smaller islets, thus, exploiting only one of the river’s branches. One of the most instructive documents on sturgeon fishing in the Iron Gates gorge is found in Volume IV of Marsigli’s 1726 work. The frontispiece (Fig. 12) portrays numerous technical details. It shows a tunnel of three A-shaped gates built between a small island and the right bank of the Danube (an arrow clearly indicates the downstream direction). In the lower half of the picture the narrow gorge is flanked by steep, misty cliffs. Dense ripples indicate a more rapid current here. Sturgeon are being caught at the downstream end of this sec-

While large Acipenserid bones commonly occur in some Mesolithic and Neolithic archaeological assemblages, they seem to become rare by the late Middle Ages. In part, this may be explained by a shift in the focus of archaeological research: catch sites were more likely to coincide with the sites of consumption during prehistory. By the Middle Ages, however, sturgeon remains are best known from high status settlements in the Danube Bend gorge in Hungary (Fig. 8). This valuable fish is known to have been traded over both short and long distances. From an archaeological point of view, however, most relevant historical information relates to methods of catching these animals. Given the large size of beluga sturgeon especially, landing it posed a special challenge to fisherfolk in all periods. According to Masen (1681: 898), “the beluga is similar to sturgeon and as strong as a tuna, frequent in the Danube”. While there is no direct evidence of prehistoric sturgeon fishing techniques along the Danube, elaborate systems for sturgeon fishing have been documented since the Middle Ages. Sturgeon weirs In accordance with the large size and strength of these animals, traps or enclosures placed in rivers to catch sturgeon, often included massive timber structures, sometimes in combination with large-holed netting. The placement of weirs was of strategic importance. Familiarity with sturgeon habitats, i.e. features of riverbed and adjacent topography, as well as the knowledge of sturgeon behaviour were, thus, instrumental in minimizing labour expenditure required for the construction of weirs. The German naturalist Johann Georg Gmelin became a professor of chemistry and natural sciences in St. Petersburg, Russia in 1731. He documented a sturgeon weir shaped like 49

The Iron Gates in Prehistory

Figure 12. Sturgeon fishing at the exit from the Iron Gates gorge (from Marsigli’s 1726 book).

tion, where the hills and the plain meet. Fishermen in small boats haul sturgeons onto the bank, where primary butchering takes place on a makeshift table. According to a 1702 description by Bél (1764: 39), willow withe fencing was combined with strong hemp nets at a slight elevation in the riverbed at Földvár (Tolna county, Hungary). The soft substrate mentioned in his text (onto which sturgeons were driven and caught) may be a reference to a submerged sandbank. The construction of weirs, even at the best loci, required tremendous investments of labour. In the 16th century along the Tisza river entire villages of serfs were enlisted to build weirs using oak logs, under the direction of the magister clausurae. Aside from having an allowance of fish during construction work, these serfs had the right to half of the catch in the new weir — with the exception of sturgeons

(Maksay 1959: 703). Building and maintaining sturgeon weirs not only required labour. In agricultural areas, shipments of appropriate oak logs from distant forests also had to be organized. According to a 1493–1495 price list, for example, 220 such ‘weir’ logs were bought by the Eger episcopate in Hungary (Kandra 1887: 378). Although the sporadic nature of the written data do not permit us to estimate the costs of sturgeon weirs, their construction must have represented a major investment, affordable only by well-organized estates in medieval Hungary. The lack of archaeological evidence for weirs does not exclude the possibility that such complex structures could have been erected in ancient times. The possibility of such cooperative efforts is clearly illustrated by the construction in AD 103–105 of Trajan’s bridge across the Danube at Drobeta (modern-day Turnu-Severin) a few kilometers downriver 50

László Bartosiewicz et al.: Sturgeon fishing in the middle and lower Danube region

from the prehistoric site of Schela Cladovei.

for beluga sturgeon in Medieval Hungary. Aside from the impressive quantity of food provided by a single specimen, the quality of sturgeon flesh was also appreciated, regardless of size. Of the species that resembled great sturgeon, Marsigli (1726) devoted most attention to the small sterlet whose meat was said to be simply ‘the best’. A quarter of a millennium later, this view was confirmed during ethnographic interviews conducted by Vasile Şişu among the fishermen of Dubova district (Iron Gates gorge, Romania). According to the Torah, only fish with scales and fins are kosher. Accordingly, the bones of sturgeon and similarly scaleless catfish occurred only in the Christian layers above deposits associated with the 13th century Jewish community in a well excavated in the Buda Royal Castle in Hungary. The latter contained only remains of kosher carp-like fish (Cyprinidae) and high status pike (Esox lucius L.; Bartosiewicz 2003). The ganoid type (Lagler et al. 1977: 108, fig. 4.2) scutes of sturgeon are not considered scales in this religious context, since the skin is often torn during descaling that renders the fish tref. Sturgeon meat and caviar therefore were avoided by Ashkenazi Jews who settled in Medieval Eastern Europe. ‘Legalizing’ sturgeon meat with reference to its ganoid scales, however, became an important halachic issue in the early 19th century Jewish religious reform in Hungary (Frojimovics et al. 1995: 293). Needs-Howarth (1996: 153) concluded that concentrations of lake sturgeon bones at certain archaeological sites in Canada indicate that these animals had mythical or religious meaning for Iroquoian people. In their artistic renditions of Lake Superior native American mythology, Longfellow (1855) and Kohl (1859 II: 143) indeed equate the ‘King of Fishes’ with sturgeon. It is not difficult to imagine that the much-awaited spring arrival of masses of sturgeon amounted to some sort of a feast during the Mesolithic and Early Neolithic of the Iron Gates. According to a structuralist interpretation by Radovanović (1997: 88–89), the upstream movement of beluga sturgeon may have symbolized life to the prehistoric inhabitants of the gorge, possibly counterweighted by the downstream orientation of the deceased in Mesolithic burials along the riverbank.1

Other equipment By the 19th–20th century hardly any references to sturgeon weirs occur in Hungary (Szilágyi 1995: 108). Large and strong, so-called pipola nets, were made especially for sturgeon fishing out of hemp. However, by the second half of the 19th century large sturgeons became scarce (cf. Fig. 3), so that these special nets were used but rarely (Herman 1887: 281). Sturgeons could also be caught with gill nets, suspended vertically in the water to trap fish by their gills, as the strong pectoral fins of sturgeons easily became hooked in its meshes (von Brandt 1964: 170). Sturgeons were also caught using sharp hooks strung on a strong rope and stretched across rivers. In contrast to similar methods of catfish (Silurus glanis Linné 1758) fishing, such hooks did not have to be baited, since curious sturgeons often ‘played’ with these glossy pieces of metal and were caught by chance (Khin 1957: 16). By the late 19th century, such hooks became the primary means of sturgeon fishing (Herman 1887: 368). Lake sturgeons in North America could be speared even from the shore in shallow waters (Needs-Howarth 1996: 149). Fish, injured and dazed, were then dragged near the bank. According to Marsigli (1726), fishermen enveloped such fish in a large net in the Danube and provoked it further onto the bank by ‘titillating’ it until it became stranded. Captured sturgeons were sometimes kept tethered to trees or strong poles, before being towed upstream by boat to the nearest market. The capitals of Buda and Vienna were supplied with live sturgeon in this way (Bél 1764: 41). Aside from hefty hooks and harpoons, grapnels must have been important tools in landing these large fish at all times. By the 18th century, the great sturgeons were even ‘hunted’ with firearms (Marsigli 1726, in Deák 2004: 74).

Cognitive aspects of sturgeon fishing Sturgeons were the largest fish in the Danube. Their sheer size and powerful movement must have impressed people throughout history. The renowned abundance of sturgeons during the spawning run must have further enhanced their awesome perception. Thus, it is likely that these great beasts had dualistic symbolic meanings related to both life and death in the spiritual life of peoples in the region.

Negative perceptions The perception of animals tends to be dualistic, ranging between extremes (Bartosiewicz 1998: 69). At the prehistoric site of Lepenski Vir in the Iron Gates gorge, the stone statue dubbed Danubius by the excavator (Srejović 1972: fig. 52) has been tentatively identified with sturgeon, having even a crest of ‘dermal scutes’ carved along its back (Radovanović (1997: 93, figs. 1–2). Although the exaggerated, even frightening, facial features (such as the large, bulging eyes) are not at all reminiscent of the modern perception of sturgeons — sturgeon have very small eyes (cf. Fig. 4), this hypothesis is not implausible if, as suggested by Srejović (1972: 111), the sculpture was conceived as an apotropaic representation with the power to avert evil or catastrophe. Bonsall et al. (2002) suggested that the Lepenski Vir sculptures were intended to protect against unpredictable and catastrophic floods. If so, the ability of sturgeon to swim against the flow even at times of high discharge, overcoming the power of the river, may

Positive perceptions Isotopic studies of human remains by Bonsall et al. (1997, 2000, 2004) have confirmed the results of previous faunal analyses (e.g. Bökönyi 1978, Bartosiewicz et al. 1995), indicating that Mesolithic populations in the Iron Gates consumed considerable amounts of fish (aquatic resources were still exploited during the Early Neolithic, but the dietary role of terrestrial resources increased). Even if sturgeon was available on a seasonal basis, owing to the quantities of meat even single individuals yielded, its contribution to the diet cannot be underestimated, and probably determined the prehistoric perception of these animals. Highly valued sturgeon meat won the name ‘Royal Fish’ 51

The Iron Gates in Prehistory

explain their use in imagery. Other, more negative associations must also be reckoned with. Although ancient fisherfolk were perfectly able to distinguish between species based on to their intimate first-hand knowledge, the image of powerful sturgeons may have been distorted by beliefs in water monsters. For example, in an Ojibway myth from North America, a giant fish threatens to swallow a young boy (Williams 1956: 168). However, in Lake Superior, the only fish that large is lake sturgeon, a harmless bottom feeder. A comparably large Danubian species is European catfish (also known as sheathfish or wels). Large specimens of this fish species reach the average size of sturgeons (c. 2.5 m, 120 kg; Pintér 1989: 135) and may grow up to 5 m (330 kg) in the Dniepr (Curry-Lindahl 1985: 259). In contrast to sturgeon, however, catfish are indeed ferocious carnivores with fearful reputations. One of their vernacular synonyms, parasztfaló, bluntly means ‘peasant gobbler ’ in Hungarian (Gozmány 1979: 957). The English synonym, sheathfish, seems to relate this species with sturgeon, whose IndoEuropean names were associated with ‘greave’ (a piece of armor worn to protect the shin) by von Sadovszky (1995: 17–19). Moreover, the Hungarian name for Acipenserids, tokhal, could be best translated as ‘case fish’ or, actually, ‘sheathfish’. Although the linguistic evaluation of apparent similarities is beyond the scope of this paper, it may indicate how the perceptions of these two large fish merged in popular mythology. Both in the Csallóköz region (the aforementioned inland delta, located upstream from Komárom), and in the Iron Gates, sturgeons following boats were considered a bad omen. As a 19th century fisherman put it: “See, how that infernal monster swims ceaselessly alongside our boat. It’s an old sturgeon, almost half a ton. It always means trouble, when evil beasts are racing with the boat like that” (Jókai 1872: 94). Danubian sturgeons of all species have four barbels in front of the mouth equipped with taste buds (Lagler et al. 1977: 350, fig. 11.9). According to Kriesch (1876: 68), rare specimens without such ‘whiskers’ were called ‘calamity fish’ in Hungary, an omen of disaster, threatening the region or even the entire country. They must have occurred among the young of stellate sturgeon especially. This means that the negative perception of sturgeon was not necessarily linked with large size and the concomitant challenge posed by sturgeon fishing.

reached the frenzy of the much-publicized, traditional Grindadrap on the Faeroe Islands, during which pilot whales (mostly the so-called Atlantic blackfish, Globicephala melaena Traill 1809), are driven ashore and massacred, mostly by men. A remarkable similarity between these small cetaceans and sturgeons in size and, to some extent, shape could be a functional basis of such hypotheses. The late 1st century BC Greek author, Strabo (Geographica Lib. VII.3.18, cited by Bél 1764: 35) actually compared sturgeons to dolphins in terms of size. While it is possible that some prehistoric sturgeon remains from the Iron Gates were the result of opportunistic fishing, historical data on the scale and intensity of sturgeon exploitation make the Grindadrap parallel sound reasonable. The vivid late Baroque description of sturgeon fishing by Bél (1764: 40–41) is as reminiscent of a gracious rite as of archetypal human predation.

Conclusions While it cannot be demonstrated that sturgeons always played a key role in the lives of all peoples living along the Danube, their significance should not be neglected at sites where their bones have been found. This multidisciplinary review of sturgeon fishing should provide archaeologists with insights into the possible ways to interpret sturgeon bones encountered in ichthyoarchaeological assemblages. 1. Owing to their poor resistance to taphonomic loss sturgeon bones tend to be underrepresented at most sites. On the other hand, their large size guarantees the recovery of the surviving remains even in hand-collected assemblages. Although species-level identification is limited to special elements, large Acipenserids must have played comparable roles in most cultures 2. Animals with such a large body mass must have been a desirable prey for subsistence fishers. Their capture, however, required special skills that included a thorough knowledge of both habitats and fish behaviour, which determined the probability of taking the best catch. 3. Aside from some degree of specialization, sturgeon fishing could only succeed as a fundamentally cooperative enterprise throughout history. It took planning and coordinated action by fishing communities, which only increased in significance as fishing techniques became ever more complex. As a group activity, carried out by men in historical times, sturgeon hunting may have had strong associations with male identity. 4. Developing infrastructure for large-scale sturgeon fishing required investments, which in turn increased the commercial value of these animals. Medieval trade in sturgeons shows the consumer’s end of this chain; some sturgeon remains were found at sites, spatially separated from catch sites, which were probably the places of primary butchering. 5. Sporadic archaeological data cannot be used in studying the depletion of stocks. Decline by the 19th century, however, is reflected both in fish sizes and in the ethnohistorical records. Over a century ago Kriesch (1876: 12–17) recognized that deforestation, river regulation

Overcoming sturgeons In spite of all these benign and fearsome connotations, it was only the landing of sturgeons that was indubitably trying and dangerous. Overcoming such a formidable adversary required fishing skills, strength and probably involved a major element of machismo at all times. According to 18th century Hungarian fishermen, the tail of a landed great sturgeon should not be approached because “even the most ablebodied lad can be killed by a single slap from it” (Bél 1764: 40). The hard work of sturgeon fishing probably also served as an outlet for social aggression. Taphonomic loss and problems of seasonal dating make it impossible to tell whether prehistoric sturgeon landings 52

László Bartosiewicz et al.: Sturgeon fishing in the middle and lower Danube region

works, damming, water pollution, steamboats and timber floating were directly responsible for habitat degradation. Owing to their low reproductive rate and high visibility, sturgeons were particularly hard hit by what he termed ‘human greed’, i.e. overfishing. 6. While the often-anecdotal ethnohistorical and literary examples cannot be used directly in the interpretation of archaeological sturgeon finds, they show the complexity of attitudes (mostly related to the immense size and particular appearance of sturgeons) that are unlikely to have developed only in recent centuries.

Guadalajara, Jalisco State, México. Bartosiewicz, L., Bonsall, C., Boroneanţ, V. & Stallibrass, S. 1995: Schela Cladovei: a preliminary review of the prehistoric fauna. Mesolithic Miscellany 16(2): 2–19. Bartosiewicz, L., Boroneanţ, V., Bonsall, C. & Stallibrass, S. 2001: New data on the prehistoric fauna of the Iron Gates: a case study fromSchela Cladovei, Romania. In Kertész, R. & Makkay, J. (eds) From the Mesolithic to the Neolithic. Budapest: Archaeolingua, 15–22. Bartosiewicz, L. & Bonsall, C. 2004. Prehistoric Fishing along the Danube. Antæus 27: 253–272. Bartosiewicz, L. & Takács, I. 1997: Osteomorphological studies on the great sturgeon (Huso huso Brandt). Archaeofauna 6: 9–16. Bél, M. 1764 [1984]: Tractatus de Rustica Hungarorum: A magyarországi halakról és azok halászatáról (Hungarian Country Life: The Fish of Hungary and their Fishing). Hungarian translation of the 1764 copy: Antal András Deák, 1984. Budapest: Vízügyi Történeti Füzetek. Bencsik, J. 1970: Egy jobbágyközség gazdasági, társadalmi élete az úrbérrendezéstől a jobbágyfelszabadításig (The economic and social life of a serfs’ community between the urbarium bill and emancipation). Acta Universitatis de Ludovico Kossuth Nominatae Series Historica 10: 49–91. Benecke, N. 1986: Some remarks on sturgeon fishing in the southern Baltic region in Medieval times. In Brinkhuizen, D.C. & Clason, A.T. (eds) Fish and Archaeology. Studies in Osteometry, Taphonomy, Seasonality and Fishing Methods. BAR International Series 294. Oxford: British Archaeological Reports, 9–17. Berg, L.S. 1904: Zur Systematik der Acipenseriden. Zoologischer Anzeiger 27: 665–667. Berinkey, L. 1966: Halak – Pisces. Budapest: Akadémiai Kiadó. Bonsall, C., Lennon, R., McSweeney, K., Stewart, C., Harkness, D., Boroneanţ, V., Bartosiewicz, L., Payton, R. & Chapman, J. 1997: Mesolithic and Early Neolithic in the Iron Gates: a palaeodietary perspective. Journal of European Archaeology 5: 50–92. Bonsall, C., Cook, G., Lennon, R., Harkness, D., Scott, M., Bartosiewicz, L. & McSweeney, K. 2000: Stable isotopes, radiocarbon and the Mesolithic–Neolithic transition in the Iron Gates. Documenta Praehistorica 27: 119–132. Bonsall, C., Macklin, M., Payton, R. & Boroneanţ, A. 2002: Climate, floods and river gods: environmental change and the Meso–Neolithic transition in southeast Europe. Before Farming: the archaeology of Old World hunter-gatherers 3_4(2): 1–15. Bonsall, C., Cook, G., Hedges, R., Higham, T., Pickard, C. & Radovanović, I. 2004. Radiocarbon and stable isotope evidence of dietary change from the Mesolithic to the Middle Ages in the Iron Gates: new results from Lepenski Vir. Radiocarbon 46: 293–300. Borić, D. & Dimitrijević, V. 2005: Continuity of foraging strategies in Mesolithic–Neolithic transformations: dating faunal patterns at Lepenski Vir (Serbia). Atti della Società per la Preistoria e Protostoria della Regione Friuli-Venezia Giulia 15 (2004–05): 33–80. Bökönyi, S. 1969: Kičmenjaci (prethodni izveštaj). In Srejović, D., Lepenski Vir: Nova praistorijska kultura u Podunavlju. Belgrade: Srpska književna zadruga, 224–228. — 1978: The vertebrate fauna of Vlasac. In Srejović, D. & Letica, Z. (eds) Vlasac: Mezolitsko naselje u Djerdapu, vol. 2. Belgrade: Serbian Academy of Science and Arts, 35–65. — 1992: Animal remains of Mihajlovac–Knjepište, an early Neolithic settlement of the Iron Gate gorge. (Hommage à Nikola Tasić à l’occasion de ses soixante ans). Balcanica 23: 77–87. Brandt, A. von 1964: Fish Catching Methods of the World. Farnham: Fishing News Books. Brinkhuizen, D.C. 1986: Features observed on the skeletons of

Note 1. Some Final Mesolithic burials at Lepenski Vir (cf. Bonsall et al. 2004, this volume) are oriented parallel to the Danube with their heads pointing downstream, but this is not a consistent feature of Mesolithic burials from the Iron Gates.

Acknowledgements A preliminary version of this paper was presented in September 2001 at the 12th Meeting of the Fish Remains Working Group (International Council for Archaeozoology) in Paihía Bay, New Zealand, organized by Drs Foss Leach and Janet Davidson. Figure 12, from Marsigli’s original work, was kindly provided by Andrea Kreutzer, Librarian in the Museum of Military History in Budapest. Dr Antal András Deák kindly gave permission to quote from his manuscript on Marsigli prior to publication. Thanks are due to Dr Alice M. Choyke for useful comments on the text as well as Dr Erika Gál and László Daróczi-Szabó, zooarchaeologists, who translated texts from Romanian.

References Antipa, G. 1905: Die Störe und ihre Wanderungen in den europäischen Gewässern mit besonderer Berücksichtigung der Störe der Donau und des Schwarzen Meeres. Wien: Sitzungberichte der internationale Fischerei Kongress. Bacalbaşa-Dobrovici, N. 1997: Endangered migratory sturgeons of the lower Danube River and its delta. Environmental Biology of Fishes 48: 201–207. Băncilă, I., Diakon, A., Djordjević, V., Hvoj, J., Jakovlević, P. Neimarević, K., Nouresku, A., Sandul, G., Vasiliju, A. & Verčon, M. (eds) 1972: Système hydro-énergétique et navigable Ðerdap (Les Portes de Fer). Belgrade: Export-Press. Bartosiewicz, L. 1989: Animal remains from the fort. In Gabler, D. (ed.) The Roman Fort of Ács-Vaspuszta (Hungary) on the Danubian Limes. Part ii. BAR International Series 531. Oxford: British Archaeological Reports, 600–623. — 1996: Early Medieval faunal remains from Pontes (Iron Gates gorge, eastern Serbia). Acta Archaeologica Academiae Scientiarum Hungaricae 48: 281–315. — 1997: Őskori vizahalászat a Duna vaskapui szakaszán (Prehistoric sturgeon fishing in the Danube). Halászatfejlesztés 20: 92–103. — 1998: Attitudes to pets in the ethnolinguistic record. In Anreiter, P., Bartosiewicz, L., Jerem, E. & Meid, W. (eds) Man and the Animal World. Studies in Memoriam Sándor Bökönyi. Budapest: Archaeolingua Kiadó, 65–78. — 1999: ‘Bronze Age fish remains in the Tisza River region, Hungary’. Paper delivered at the 10th meeting of the ICAZ Fish Remains Working Group, New York City. — 2003: ‘Eat not this fish — a matter of scaling’. Paper delivered at the 12th meeting of the ICAZ Fish Remains Working Group, 1

53

The Iron Gates in Prehistory some recent European Acipenseridae: their importance for the study of excavated remains of sturgeon. In Brinkhuizen, D.C. & Clason, A.T. (eds) Fish and Archaeology. Studies in Osteometry, Taphonomy, Seasonality and Fishing Methods. BAR International Series 294. Oxford: British Archaeological Reports, 18–33. — 1989: Ichthyo-archeologisch onderzoek: methoden en toepassing aan de hand van romeins vismateriaal uit Velsen (Nederland). Unpublished PhD thesis. Rijksuniversiteit Groningen, Groningen. Casteel, R.W. 1976: Fish Remains in Archaeology and Palaeoenvironmental Studies. London: Academic Press. Clason, A.T. 1980. Padina and Starčevo: game, fish and cattle. Palaeohistoria 22: 141–173. Curry-Lindahl, K. 1985. Våra fiskar. Havs- och sötvattenfiskar i Norden och övriga Europa. Stockholm: Norstedt & Söners. Deák, A.A. 2004: A Duna fölfedezése (The Discovery of the Danube). Budapest: Vízügyi Múzeum, Levéltár és Könyvgyűjtemény. Deckert, K. 1967: Urania-Tierreich: Band 4: Fische, Lurche, Kriechtiere. Leipzig: Urania Verlag. Desse, J. & Desse-Berset, N. 1992: Pêches locales, côtières ou lointaines: le poisson au menu des Parisiens du Grand Louvre, du 14e au 18e siècle. Anthropozoologica 16: 119–126. Desse-Berset, N. 1994: Sturgeons of the Rhône in Arles (6th–2nd century BC). In Van Neer, W. (ed.) Fish Exploitation in the Past. Tervuren: Koninklijk Museum voor Midden-Afrika, 81–90. Frojimovics, K., Komoróczy, G., Pusztai, V. & Strbik, A. 1995: A zsidó Budapest. Emlékek, szertartások, történelem I–II. (Jewish Budapest. Monuments, Rituals, History I–II). Hungaria Judaica 7. Budapest: MTA Judaisztikai Kutatócsoport. Giurescu, C.C. 1964: Istoria pescuitului şi a pisciculturii în Romania (The History of Fishing and Pisciculture in Romania). Bucureşti: Editoriul Academiei. Gozmány, L. 1979: Vocabularium nominum animalium Europae septem linguis redactum. Budapest: Akadémiai Kiadó. Hankó, B. 1931: Magyarország halainak eredete és elterjedése (The origins and distribution of fishes in Hungary). A Debreceni Tisza István Tudományegyetem Állattani Intézetéből 10: 3–34. Harka, Á. 1993: A folyóvizek halrégiói (The fish regions of rivers). A természet 44(5): 85–87. Hensel, K. & Holčík, J. 1997: Past and current status of sturgeons in the upper and middle Danube river. Environmental Biology of Fishes 48: 185–200. Herman, O. 1887: A magyar halászat könyve I–II (The Book of Hungarian Fishing). Budapest: Kir. Magyar Természettudományi Társulat. — 1980: Halászat és pásztorélet (Fishing and Pastoralism). Budapest: Gondolat Kiadó. Jókai, M. 1872 [2001]: Aranyember I (The Golden Man I). Budapest: Talentum Kiadó. Kandra, K. 1887: Bakócs-codex vagy Bakócs Tamás egri püspök udvartartási számadókönyve 1493–6. évekből (The Bakócs Codex, the court accounting books of Bishop Tamás Bakócs of Eger, 1493–6). Adatok az Egri Egyházmegye történelméhez 2: 333–452. Khin, A. 1957: A magyar vizák története (The History of Hungarian Sturgeons). Budapest: Mezőgazdasági Múzeum Füzetei 2. Kohl, J.G. 1859: Kitschi-Gami oder Erzählungen vom Oberen See I–II. Bremen: C. Schünermann. Kriesch, J. 1876: Hasznos és kártékony állatainkról. Kézikönyv II. rész. Halak (Handbook of Useful Animals and Pests. Vol. II. Fish). Budapest: Szent István Társulat. Lagler, K.F., Bardach, J.E., Miller, R.R. & May Passino, D.R. 1977: Ichthyology. New York: Wiley. Longfellow, H.W. 1855 [1964]: The Song of Hiawatha. New York:

Bounty Books. Maitland, P.S. & Linsell, K. 1978: Europas sötvattenfiskar. Stockholm: Albert Bonniers. Maksay, F. (ed.) 1959: Urbáriumok XVI–XVII. század (16th–17th Century Urbaria). Budapest: Akadémiai Kiadó. Marsigli, L.F. 1726: Danubius Pannonico-Mysicus, observationibus geographicis, astronomicis, hydrographicis, historicis, physicis perlustratus. Vol. VI. Amsterdam–The Hague. Masen, J.S.J. 1681: Speculum Imaginum Veritatis Occultae. Ed. tertia prioribus correctior. Köln: J.A. Kinckius. Miranda, L.E., Wingo, M.W.M., Muncy, R.J. & Bates, T.D. 1987: Bias in growth estimates derived from fish collected by anglers. In Summerfelt, R.C. & Hall, G.E. (eds) Age and Growth of Fish. Ames, IA: Iowa State University Press, 211–219. Müller, H. 1983: Fische Europas. Leipzig: Neumann. Muus, B. & Preben, D. 1965: Meeresfische der Ostsee, der Nordsee, des Atlantiks: Biologie, Fang, wirtschaftliche Bedeutung. München: BLV-Verlagsgesellschaft. Needs-Howarth, S. 1996: Lake sturgeon fishing at prehistoric Iroquoian sites near Lake Simcoe, Ontario. Archaeofauna 5: 147–154. Păunescu, A. 2000: Paleoliticul şi Mezoliticul din spaţial cuprins între Carpaţi şi Dunăre. Bucureşti: Agir. Pénzes, B. & Tölg, I. 1977: Halbiológia horgászoknak (Fish Biology for Anglers). Budapest: Natura–MOHOSZ. Pintér, K. 1989: Magyarország halai (The Fish of Hungary). Budapest: Akadémiai Kiadó. Radovanović, I. 1997: The culture of Lepenski Vir: a contribution to the interpretation of its ideological aspects. In Lazić, M. (ed.) Antidoron Dragoslavo Srejović completis LXV annis ad amicis, collegis, discipulis oblatum. Belgrade: Centre for Archaeological Research, University of Belgrade, 87–93. Radu, V. 1997: Archaeological researches at Borduşani-Popină (Ialomiţa County). Preliminary report 1993–1994. Pisces. Cercetări arheologice, 10: 96–105. — 2003: Exploitation des ressources aquatiques dans les cultures néolithiques et chalcolithiques de la Roumanie méridionale. Thèse de doctorat de l’Université Aix–Marseille I, Préhistoire, Archéologie, Histoire et Civilisation de l’Antiquité et du Moyen-Age, Aix-en-Provence, France. Roussow, G. 1957: Some considerations concerning sturgeon spawning periodicity. Journal of the Fisheries Research Board of Canada 14: 553–572. Sadovszky, O. von 1995: Fish, Symbol and Myth. Budapest: Akadémiai Kiadó. Schalk, R.F. 1977: The structure of an anadromous fish resource. In Binford, L.R. (ed.) For Theory Building in Archaeology: Essays on Faunal Remains, Aquatic Resources, Spatial Analysis, and Systemic Modelling. New York: Academic Press, 207–249. Srejović, D. 1972: Europe’s First Monumental Sculpture: New Discoveries at Lepenski Vir. London: Thames & Hudson. Szilágyi, M. 1995: A Tiszai halászat. Az eszközök és fogási módok történeti változásai (Fishing in the Tisza. Historical Changes in Equipment and Methods). Budapest: Akadémiai Kiadó. Terofal, F. 1971: Kwastsnoeken, steuren en verwanten. In Grzimek, B. (ed.) Het leven der dieren IV, Vissen I. Nijmegen: Thieme & Cie, 148–146. Vuković, T. & Ivanović, B. 1971: Slatkovodne ribe Jugoslavije (Freshwater Fish of Yugoslavia). Sarajevo: Zemaljski Muzej Sarajevo. Wheeler, A. 1978: Key to the Fishes of Northern Europe. London: Frederick Warne. Williams, M.L. (ed.) 1956: Schoolcraft’s Indian Legends. East Lansing: Michigan State University Press. Woodard, C. 2000: Singing the Danube blues. Christian Science Monitor, May 11, 2000.

1

54

The Mesolithic–Neolithic transition in the Ðerdap as evidenced by non-metric anatomical variants Mirjana Roksandic

Abstract: In order to examine the interactions of communities with different modes of subsistence — the foraging communities of Lepenski Vir and the farming groups of the central Balkans — the non-metric anatomical variants were examined in four Ðerdap sites. This paper argues for local continuity within the region, with high a degree of initial heterogeneity, and temporal ordering as the most likely explanation for the pattern of change. A more pronounced difference at the time of contact with Neolithic populations argues for a limited ‘seeping in’ of a non-local population that did not result in a change of either economic base or ideology. There is no evidence of an incoming population at the time of change to a farming economy and integration of the sites into the cultural circle of the Balkan Neolithic, but rather of a local population accepting the new way of life. Key words: Mesolithic, Neolithic, transition, Đerdap, non-metric anatomical traits

Introduction

ber of communities in the Holocene that continue with essentially the same mode of food procurement and mobility patterns as their Pleistocene counterparts as well as significantly different groups, the need to introduce ‘Mesolithic’ as a different term is not obvious. Introduction of economic parameters has, in a number of cases, led to economic determinism, with an open or subdued notion of evolutionism. Distinction is made between the Epipaleolithic — in which the Holocene adaptation does not produce any changes in way of life and lithic technology (Leroi-Gourhan 1965; Kozłowski & Kozłowski 1986) — and the Mesolithic with its substantial changes in economy, ecology, and material culture (Kozłowski & Kozłowski 1986). The latter would be found only in innovation zones leading to food production (Leroi-Gourhan 1965), or enabling change from food collection to food production (Clark 1980). This definition supposes an unidirectional evolution toward food production and contradicts the data from large areas of the world where substantial changes in economy, ecology and material culture did not lead to the introduction of agriculture (e.g. the Pacific Northwest Coast). In order to overcome chronological and typological ambiguity, as well as economic determinism characterized by an implicit evolutionist basis, Radovanović (1996a: 14) argues that a qualitatively different phenomenon can be found in the appearance of formal disposal areas for burial of the dead. Formal disposal areas need not be a phenomenon separate from the habitation site, as that would exclude all western and Central European sites (Meiklejohn, pers. comm.), with the exception of the newly excavated Mesolithic necropolae in France (Duday & Courtaud 1998) and Belgium (Cauwe 1998). They are defined as “...areas of continuous, ceremonial, mortuary disposal” (Radovanović 1996a: 14). Further, they are an archaeologically visible phenomenon that is interpreted as arising from the need to lay claim to territory by its ideological integration (Chapman 1981).1 The need to claim territory arises from a combination of linear rather than hexagonal arrangement of units within one hunter-gatherer group, or higher than usual population densities (Gamble

This article seeks to promote a better understanding of the population dynamics of the Lower Danube as reflected in the transition from the Mesolithic ‘Lepenski Vir culture’ to the Neolithic ‘Starčevo complex’ in the Đerdap (Serbia). Since a long period of coexistence of the Mesolithic and Neolithic ways of life in this region has been proposed (Radovanović 1995, 1996b), this paper will attempt to reconstruct the extent and mode of interactions between farmers and foragers through the examination of skeletal material from four of the most important sites excavated in the region: Lepenski Vir, Hajdučka Vodenica, Vlasac and Padina. The focus of this article being the Mesolithic population of the Đerdap and its presumed contact with Neolithic peoples, and since the debate over the meaning of this term has historically played an important role in discussions between principal investigators of the Đerdap sites (see Srejović 1971, 1979; Jovanović 1972; Srejović & Letica 1978; Boroneanţ et al. 1999), it is important to provide clear definitions of both Mesolithic and Neolithic as they are used here.

Mesolithic Current definitions can be classified into two major groups: one typological and the other chronological. For the proponents of a typological definition the characteristics of the flint industry (such as the appearance of microliths) have the most decisive value. For example, Orliac (1988: 686) proposes that those industries situated between Palaeolithic and Neolithic that possess “characters sufficiently different from those of the industries of the two periods” should be determined as Mesolithic. While possibly appropriate for Western Europe, typological definitions fall short in other areas of the world. For proponents of a chronological definition all hunter-gatherers of the Holocene are regarded as Mesolithic, regardless of whether they show differences from Palaeolithic hunter-gatherers. Since it would include a num55

The Iron Gates in Prehistory

1986: 52–53), based on intensive exploitation of a vital resource, or a greater variety of resources in the vicinity, with semi-sedentism or sedentism. This would result in a structural complexity of the social unit (Srejović 1979) usually expressed through developed ancestral and mortuary rituals, a clear sign of changing times in prehistory associated with changes in the social arena (Chapman 1993). Although archaeologically visible, the ambiguity in determining what constitutes a formal disposal area makes them less valid for defining the Mesolithic. On the other hand, economic parameters and mobility patterns combined, once evolutionist connotations are removed, have the potential to make the definition of Mesolithic more straightforward and appropriate even for regions where the typological distinction of Mesolithic is not as easy as in France (for example, Africa where microliths appear as early as 70,000 BP), or where ceramics (traditionally associated with the Neolithic) appear earlier than agriculture, as for example in the Jomon culture of Japan (Imamura 1996) or in Scandinavia (Werbart 1998: 37, and quoted literature). This change in adaptation is usually linked with intensification of exploitation of one or more abundant resources (r-selected resources) as opposed to the exploitation of Kselected resources (see Gamble 1986: 41) that characterizes mobile hunter gatherers. The availability of an abundant and stable resource that can be exploited in the relative vicinity of the camp has been linked to reduced mobility. The combination seems to be determinative of Mesolithic settlements. Structural complexity, seen as a segmentation or an increase in the parts that make up the whole (Kent 1989: 10), could arise from changes in mobility patterns and increased sedentism resulting in population increase and the need for an arbiter in settling disputes (Lee 1972a, 1972b). However, it cannot be taken for granted and must be demonstrated by independent data in many areas of the world where it has been assumed (Brinch-Petersen & Meiklejohn, forthcoming). Settlement distribution and mobility patterning, interacting with more intensive exploitation of r-selected resources (plants or fish), and aggregation necessary for tasks demanding cooperative effort, can be regarded as both determinative of the Mesolithic and as providing sufficient archaeological visibility. Within the Đerdap, the Mesolithic is best defined as primarily based on intensified exploitation of food resources on a limited territory with reduced mobility. Regardless of whether fish played a key role as staple (Bonsall et al. 1997) or as a vital resource (Radovanovic 1996a), its availability is the sine qua non of reduced mobility (sensu Kent 1989) and the change in its exploitation provides clear evidence of Mesolithic economy and social organization.

ally traceable. Preceding or subsequent changes in social structure, ideology or any other aspects of life need to be examined on a regional basis. The other two questions, that of how much evidence of plant manipulation and animal husbandry constitutes enough evidence (sensu Harris 1996, and quoted literature), and whether a horticultural stage of farming economy — characterized by lack of impact on the environment (Willis & Bennet 1994) — can be perceived as agriculture, also need to be regarded on a local scale. In Southeast Europe, the introduction of cultigens and domestic animals from the Near East seem to solve this problem, as these are not found in the wild. Their introduction indicates a shift in subsistence strategy and reliance (at least partial) on imported cultigens.

The Đerdap context Crucial to our understanding of the Đerdap Mesolithic/Neolithic transition is the period of coexistence of these two modes of life in the immediate vicinity. It is evidenced by 14C dates, Starčevo ceramics, and flint blades within Mesolithic strata of the Lepenski Vir culture. This period witnesses the coexistence, communication, trade and interaction between Lepenski Vir hunter-gatherers and Starčevo farmers. It is in this light that the semantic discussion of the two terms becomes increasingly important for understanding archaeological data. Interpretations of the Lepenski Vir–Schela Cladovei culture range from Epipaleolithic (Boroneanţ 1969, Letica 1971, Lazarovici 1979), Protoneolithic (Srejović 1968), Epipaleolithic in its early and Protoneolithic in its late phase (Srejović 1979), Mesolithic (Prinz 1987, Srejović 1989, Voytek & Tringham 1989), to Mesolithic and Early Neolithic (Jovanović 1972, 1974). This variety stems to a great degree from the above-mentioned definitions: one chronological, in which Holocene hunter-gatherers are differently viewed as Mesolithic or Epipaleolithic; and the other based on material culture where microliths are taken as a tell-tale sign of the Mesolithic, while ceramics, polished stone axes and adzes are used as markers of Neolithic (for discussion and appreciation of the theoretical positions in these different approaches, see Radovanović 1996a). The littoral distribution of the sites in the Đerdap that resulted in reduced mobility or possibly even sedentism, increased social complexity evidenced by specialization of ritual vs domestic activities (Srejović 1979; but see Chapman 1993) and increased population aggregation (Jackes et al., this volume) are all very prominent features of Mesolithic components in the studied sites. Neolithic in the region is characterized by the introduction and reliance, at least partial, on Near Eastern cultigens and livestock. Although hunting, fishing and gathering remain important, the proportion of domesticates vs wild fauna and flora is sufficient to argue for the introduction of a Neolithic economy. Therefore, if intensification of exploitation of food resources, on a limited territory with reduced mobility, characterizes the Lepenski Vir Mesolithic, then all strata within these sites that do not have evidence of food production and/or introduced domesticates (above 5% as suggested by

Neolithic An economic definition of Neolithic, in which “the shift in mode of subsistence to agro-pastoral farming remains the only process that is relatively clearly defined, geographically widespread and sufficiently archaeologically detectable” to act as a signature of the Neolithic (Zvelebil 1996b: 625) is widely accepted and both ethnologically and archaeologic56

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap Table 1. Synchronisation of the sequences identified at four key sites in the Iron Gates gorge: shaded areas represent the appearance of ceramics in the stratum: light shading — sporadic appearance, darker shading — ceramics are common as well as Pre-Balkan Plateau flint, ‘Montbani type’ chipped stone blades, along with geometric microliths (based on Radovanović 1996a: 289, 1996b, 1996c; Radovanović & Voytek 1997). Phase – millennium BC

Padina

L. Vir

Vlasac

H. Vodenica

6

- mid 6th

B(III)

II/IIIa

—

Ib

5

- 7th/6th

B(II)

I(3)

—

Ia

4

- 7th (2nd half)

B(I)

I(2)

III

Ia

3

- 7th (1st half)

A-B

I(1)

Ib-II

2

- 8th/7th

A/A-B

Proto-LV

Ia-b

Ia

1

- 8th

A

—

Ia

—

Zvelebil 1996a) should be regarded as Mesolithic. If we accept that the Neolithic in the region is characterized by introduced cultigens and livestock, then evidence for food production and appearance of these cultigens in the strata should be regarded as crucial for identifying them as Neolithic. In this perspective, the sporadic appearance of Starčevo-type ceramics and Starčevo blades in Mesolithic strata, i.e. Mesolithic-type houses with ceramics in situ (Jovanović 1984b), can be explained by a porous agricultural frontier with transfer of knowledge, material items, and individuals across the board (Zvelebil 1996a), whilst recognizing that this exchange could have involved both different direction and different form over this long period of coexistence. The presence of domestic animals should also be considered in this context. Dog was domesticated in the Đerdap without doubt and is found in the Mesolithic, Transitional and Neolithic strata of Lepenski Vir, Vlasac, Padina and Hajdučka Vodenica. At Icoana selective hunting of wild pig (very young and very old animals) and possible domestication of dog point in the same direction (Bolomey 1973). The analyses from Padina (Clason 1980) and Hajdučka Vodenica (Greenfield 1984, this volume) are not as conclusive as, in both cases, the material was studied initially without respect to chronological units. The scarcity of fish bones at Hajdučka Vodenica could account for the elevated percentage of domestic animals (c. 20% of identifiable specimens). Considering only the domestic species, the vast majority belong to incipient domestic pig and dog, and only c. 15% (c. 3% of the total sample) belong to Bos taurus (imported species). All of the other units, except for the Lepenski Vir III (Neolithic) strata present less than 5% of domestic animals which coincides with Zvelebil’s (1996a) explanation of a porous agricultural frontier, and serve as evidence of contact (trade or raiding) rather than of a Neolithic economy. A high proportion of wild animals and fish in Lepenski Vir III points to a strong local tradition among the Early and Middle Neolithic settlers of the site, or to the possibility that Lepenski Vir was an atypical Neolithic site, a non-sedentary

station for hunting and gathering. Either way, the knowledge of the region and the know-how of the Mesolithic hunters was already acquired. Again, it would be of great value to be able to distinguish between Early (LV IIIa) and Middle (LV IIIb) Neolithic settlements as the relative importance of Ovis/Capra and Bos taurus are expected to have changed in the region from one sub-period to the other. Bonsall et al. (1997: 56–57) argue that sampling bias could have played a role in the lack of domestic animals in Lepenski Vir I and II, which both Chapman (1993) and Whittle (1996) consider to be synchronous with Neolithic Starčevo. Although faunal analyses by Bökönyi (1969, 1972, 1974, 1978), performed within the framework of the 1960s and 1970s, are lacking in detail (see Lyman 1994, for new approaches to MNI and skeletal elements proportions), they are consistent throughout the Đerdap material. The very restricted numbers of identifiable specimens in LV I–II (less than 500) compared to LV III (over 2000) could account for some bias in species representation (cf. Bonsall et al. 1997), but not for the total lack of selected exploited species. If sieving can account for a greater share of fish bones in the recent excavations at Schela Cladovei (Bartosiewicz et al. 2001), there is no reason to suppose that the overall proportions of mammal skeletons would have been significantly altered. Pottery, long held as a tell-tale sign of the introduction of the Neolithic to Europe, appears in many of the sites within the Đerdap and has been a major source of discussion. On the right bank the stratigraphic position of ceramic bearing horizons is reasonably clear in the lower gorge, Ključ and downstream from Ključ (Hajdučka Vodenica, Kula, Ajmana, Velesnica). Most problematic is the situation of Lepenski Vir and Padina, both of which are located in the upper gorge (Radovanović 1996a: 282). Since both sites have complex vertical and horizontal stratigraphy and evidence of other imported material, but no evidence of change in the economic structure and ideological world of the local inhabitants (until the LV III stratum), the appearance of pottery (Table 1) is recognized as evidence of contact between local foragers 57

The Iron Gates in Prehistory

and pottery-bearing Neolithic farmers. Pottery appears in all horizons of Padina B at all sectors, while there is no evidence of its appearance in Padina A, or A/B. Similarly, ProtoLepenski Vir and Lepenski Vir Ia–b (Radovanović phase I-1) did not contain pottery, while it starts appearing in Lepenski Vir Ic-e (Srejović 1969), or Radovanović phase I-2 and I-3. Lepenski Vir II did not contain any pottery, and it appears again with the Neolithic economy of Lepenski Vir IIIa and IIIb. Vlasac I–III, akin to the lower gorge settlements, did not contain any pottery until the fully Neolithic Vlasac IV stratum. The appearance of pottery coincides with the distribution of Pre-Balkan Plateau flint, and argues for greater importance of trade. While assuming that pottery is necessarily a Neolithic invention throughout Europe is inherently problematic, there is no reasonable doubt that pottery was brought into the Đerdap Mesolithic communities by neighbouring Neolithic people since it fits well within the Gura Baciului and Starčevo tradition (Jovanović 1984a, 1987). In terms of the newly proposed periodization by Tasić (1997, 1998), the ceramics found at the sites of Lepenski Vir and Padina fall well within the Early and Middle Neolithic of the central Balkans with no particular developments that would suggest local invention. Therefore, the appearance of pottery on these sites can serve as a marker of contact between farmers and foragers, independent of absolute chronology and uncertainties of 14 C dates. The rationale behind the use of pottery as an independent marker of contact is found in its non-local origin that supposes either trade, transfer of knowledge, or transfer of people with this particular knowledge into the Đerdap communities. All of these imply the availability of contact, even where there is no firm evidence of contact itself. Conceptually, three periods of the Đerdap sequence are distinguished in the present study: 1. Mesolithic proper, based on a hunting-gathering economy with increased sedentism. 2. Mesolithic/Neolithic or Contact period, based on the same economic pattern but with either evidence or possibility of contact with Neolithic peoples — the ‘availability phase’ of Zvelebil. 3. Neolithic proper, marked by change from a predominantly foraging economy to a predominantly food producing one. No unidirectional evolution of economic pattern is assumed. Regardless of the economic pattern of a particular site, or phase within the site, once contact with Neolithic peoples in the region becomes possible, it is no longer regarded as purely Mesolithic but falls within the Mesolithic/Neolithic group, signifying the availability of the contact. The economic behaviour at any particular site will further determine whether it is Mesolithic/Neolithic (with little or no change in the economic domain) or Neolithic (implying increased importance of domesticates). This approach justifies the preference of relative over absolute chronology. Since our determination of any of the strata in the four settlements is based on economic patterns, if burials can be associated reasonably accurately with any of these occupations then absolute dates do not provide useful additional information. The synchronization for the four sites in question (Table 1) summarizes the chronology and outlines the period when

contact with ceramic-producing farming communities in the region becomes established. Although the appearance of ceramics and Pre-Balkan Plateau flint does not necessarily imply the ‘invasion’ or even ‘moving in’ of farming communities into the region, it is evidence of the availability of contact between Đerdap foragers and Balkan farmers.

Anthropological research Most of the previous anthropological research in the Iron Gates gorge was based on comparisons of metric data for the two sites that had yielded larger numbers of measurable cranial remains, namely Vlasac and Lepenski Vir. One of the first syntheses came from Nemeskéri and his colleagues (Nemeskéri 1969, 1978; Nemeskéri & Lengyel 1978; Nemeskéri & Szathmary 1978a, 1978b, 1978c, 1978d, 1978e). Nemeskéri’s research was very influential and remains one of the most comprehensive studies of the Vlasac material, albeit within the paradigm of ‘anthropotypology’. Given recent understanding of problems associated with ‘racial taxonomy’ (Jacobs et al. 1996), these conclusions could be either rejected or incorporated into a different understanding of population genetics and micro-evolutionary changes. If his findings are regarded without the taxonomic labels, several characteristics of the population can be identified: 1. Great heterogeneity within the local Mesolithic population. 2. Greater heterogeneity within the male group and greater homogeneity within the female group. 3. A temporal trend toward gracilization. 4. A trend toward homogenization of the population in later phases, which Nemeskéri attributes to greater inter-group gene flow. Concerning the Lepenski Vir material, in a preliminary report Nemeskéri (1969) concludes that the Mesolithic strata contain the finds of Cro-Magnon type, while the Neolithic strata show most probably three distinct types belonging to the Mediterranean taxon. It is important to note that the Mediterranean taxon does not imply Mediterranean origin, but only a skeletal ‘type’. The first synthesis of the Đerdap material by Mikić (1980, 1981a, 1981b) follows the general division into A and B types outlined by Nemeskéri and Szathmary. Mikić developed an explanatory scheme that accounts for possible processes that could have induced the change within the series (Mikić 1981a: 104, fig. 1). In his first synthesis of the material, strongly influenced by Srejović’s hypothesis of indigenous Lepenski Vir cultivation and domestication, Mikić proposed that micro-evolutionary trends at Lepenski Vir could account for gracilization as a consequence of neolithization. He introduced another set of ‘types’ all based on a generalized ‘Mediterranean’ morphology, all of which were derived through micro-evolutionary processes from the ‘CroMagnon’ type. This micro-evolution occurs within layer II of Lepenski Vir, and subsequent changes in both shape and size of the skeletons occur without interruption into the Neolithic, eventually producing ‘gracile’ and ‘generalized’ Mediterranean types. As with Nemeskéri, the ‘Mediter58

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap

ranean’ in the label does not imply the movement of people from the Mediterranean into the region. Abandoning the typological classification in his later works (Mikić 1988), Mikić argues for local continuity and isolation stressing the morphological similarities between the Palaeolithic Climente specimen and Late Mesolithic and even Neolithic individuals from Lepenski Vir (Mikić 1992: 40). Padina and Hajdučka Vodenica were very summarily treated by Živanović and most of the conclusions were based on archaeological interpretation of the sites. Thus, on the basis of morphological examination, he singled out seven skeletons belonging to the Lepenski Vir culture at the site of Hajdučka Vodenica, attributing all others to the Iron Age stratum (Živanović 1976). Although Živanović claimed tohave recognized a number of more robust skeletons belonging to the ‘Padina racial sub-group’ and substantially more gracile ones belonging to a much later population (Živanović 1976: 124), I was unable to make any such distinction once the sex was accounted for. This ‘Padina racial sub-group’ is viewed as autochthonous, different from all other known groups and best described as the ‘Obercassel type of the Dinaric race’ (Živanović 1975a, b). Živanović ascribes to Nemeskéri the conclusion that Lepenski Vir belonged to the same ‘Proto-Dinaric’ population. However, Nemeskéri disclaimed this quote (Nemeskéri & Szathmary 1978b: 180). All authors agreed on one important aspect of the material — essential continuity within the region. All of the changes were attributed to micro-evolution towards more gracile forms with intensification of contacts and admixture at the time of the Vlasac II/Lepenski Vir I phases. None of the authors perceived any abrupt change in the Neolithic populations of the region. Schwidetzky & Mikić (1988) argued that gracilization cannot be assumed to coincide with Neolithic adaptation. They reach the conclusion that the high rates of change support the micro-evolutionary processes in the Iron Gates rather than abrupt change in population (Schwidetzky & Mikić 1988: 117). Quite a different conclusion was reached by Menk (Menk & Nemeskéri 1989). While he also claimed a sharp decrease in robusticity between the Terminal Mesolithic and Early Neolithic of the series (see Roksandic 1999, for a critique) as well as considerable change in shape, the author concluded that the change could not be explained by local evolution, but rather by a progressive replacement of the population. Menk applied PCA to cranial and postcranial measurements provided by Nemeskéri. As Menk notes, the Lepenski Vir Neolithic sample “fractions into three parts with a remarkable gap in the central part of its area” (Menk & Nemeskéri 1989: 534). In itself, coupled with a small sample size and a problematic pooling of the sexes, this finding can invalidate the analysis, since the central value of the Neolithic Lepenski Vir population is derived from the strong dissimilarity of the individuals of which it is made. A look at the PCA graphic output for components 1, 2 and 3 offer by Menk & Nemeskéri (1989) shows strong differentiation along axis 1 (corresponding to size) for Lepenski Vir Mesolithic and Lepenski Vir Neolithic. But, the same is true for the distance between the Vlasac I and I? and Lepenski Vir Mesolithic. However, the distance between the two on the 2, 3 (measuring some form of shape differentiation) axis is small. It is, in effect, much smaller than between Vlasac I and

Vlasac I?. Furthermore, the Starčevo-Criş Neolithic sample seems to be less removed on the size axis from Mesolithic Lepenski Vir and practically identical with it on the shape axis (axis 2, 3). The only actual difference between different subsamples on the ‘shape’ axis is between Vlasac I and Vlasac II, III that show as much difference on the size axis as the Lepenski Vir Mesolithic and Neolithic samples. The conclusion, although not necessarily wrong, cannot be demonstrated on the basis of the published results; however, the article offers an interesting view on Mesolithic heterogeneity that is in accordance with all other published results.

Material and methods Since basic demographic data are given elsewhere (Roksandic 1999; Jackes et al., this volume), only the period designation for each of the burials is presented here (Table 2). Padina According to the chronology of Padina and the division of the site strata into Mesolithic and Mesolithic/Neolithic contact, skeletal remains were assigned to either of the two periods according to the site documentation, superpositioning of certain features, and Radovanović’s (1996a) analysis of the burials. It was not possible to assign four individuals from the disturbed unidentified graves to either period with acertainty. Lepenski Vir According to the presented chronology of the site and division of strata into the Mesolithic, Mesolithic/Neolithic contact and Neolithic periods, the skeletal remains were assigned to particular phases according to Radovanović’s analysis of Mesolithic burials (Radovanović 1996a: 174–189) and site documentation (kindly provided by Prof. Srejović, for the Neolithic burials (Antunović 1987). Vlasac According to the building horizons, Srejović & Letica (1978) identified three phases of the Mesolithic settlement (Vlasac I, II and III) and one of the Neolithic (Vlasac IV). Since no human skeletal remains were associated with the latter, it will not be discussed in detail. Radovanović has observed significant changes in burial practices over time and has proposed a division of formal disposal areas into two chronological phases. Based on the published material, as well as field documentation (kindly provided by the principal investigator, the late Prof. D. Srejović) she was able to distinguish an important change in burial practices that began to occur at the end of Vlasac I but was certainly present in the Vlasac II phase (Radovanović 1996a: 187–218, appendix II). However, both Vlasac I and II would belong to the Mesolithic pre-contact period and only Vlasac III would chronologically belong to the period when contact with the Neolithic populations was possible. All of the cases where Radovanović could not distinguish between the Vlasac II and III burials are treated as Mesolithic/Neolithic contact. However, a separate test was run with these individuals included in the Mesolithic group since contacts between Lepenski Vir and the surrounding farming population(s) is least attested in Vlasac of 59

The Iron Gates in Prehistory Table 2: Classification of burials from Padina, Lepenski Vir, Vlasac, and Hajdučka Vodenica. PADINA Mesolithic: single burials:

1; 18b; 21; 22; 39.

double burials:

12; 12(1); 14; 14(1); 17; 17(1); 23; 23(1).

multiple burials:

15(15-16a); 15-16(15-16a); 16(15-16a); 16a(15-16a); 16(1)(15-16a).

Mesolithic/Neolithic Contact: single burials:

1a; 3; 6; 7; 8; 9; 10; 11; 13; 18; 19; 24; 26; 26a; 27; 28; 29; 30.

double burials:

2, 2(1); 6a, 6a(1); 25, 25(1).

multiple burials

4/4+5+5a/, 5/4+5+5a/, 5a/4+5+5a/, 5a(1)/4+5+5a/; 20, 20(1), 20(2).

LEPENSKI VIR Mesolithic: single burials:

3; 21; 22; 46; 60; 61; 67; 69; 110; 111; 112; 113; 117; 118; 119; 120; 121; 132; 133.

double burials:

50, 50(1); 64, 64(1); 99, 99(1); 102, 102(1); 109,109a.

multiple burials:

101, 101(1), 101(2).

Mesolithic/Neolithic Contact: single burials:

11; 12; 15; 16; 17; 23; 26; 28; 46; 68; 70; 90; 91; 92; 94; 95; 96; 97; 100; 103; 104; 105; 115; 116; 122; 126; 127; 128; 129; 130; 131; 134.

double burials:

7a,7b; 13, 13(1); 14, 14(1); 45a, 45b; 63, 63(1); 74, 74(1); 93, 93(1); 98, 98a; 99, 99(1); 106, 106(1); 107, 107(1); 108, 108(1); 114, 114(1); 123, 123(1); 124, 124(1); 125, 125(1).

multiple burials

54a, 54b, 54c, 54d, 54d(1), 54e.

Neolithic: single burials:

1; 4; 5; 6; 8; 9; 20; 31a; 33; 35; 37; 38; 39; 42a; 43; 44; 48; 51; 53; 56; 57; 59; 66; 71; 88.

double burials:

18, 18(1); 19, 19a; 32a, 32c; 52, 52a; 55a, 55b; 73, 73(1).

multiple burials:

83a, 83a(1), 83b.

Period uncertain: single burials:

2; 10; 24; 25; 29; 36; 40; 41; 49; 58; 62; 65; 72; 75; 76; 78; 86.

double burials:

77, 77(1); 80; 81; 82; 84, 84(1).

multiple burials:

27a (27a+e), 27(27b), 27 (27C), 27 (27d), 27(1), 27 (27f), 27(2); 34a, 34b, 34c; 79a, 79b, 79c; 85, 85a, 85b/85(1)/, 85b; 87, 87(1), 87(2), 87(3), 87(4); 89a, 89b, 89b(1).

VLASAC Mesolithic: single burials:

7; 8; 10, 11; 13, 20; 25; 28, 30; 31; 32; 33; 34; 37; 38; 39; 40; 41; 44; 59; 61; 63; 68; 72; 79; 81.

double burials:

9, 9(1); 12a, 12b; 19, 19a; 35, 35a; 42a, 42b; 47, 47a; 48, 48(1); 53, 53(1); 56, 56(1); 57, 57(1); 60, 60(1); 62, 62(1); 66a, 66a(1); 80, 80a; 84, 84(1).

multiple burials:

4a, 4b, 4b(1); 5, 5(1), 5(2); 6, 6a+6(1), 6(2); 18a, 18b, 18c; 21, 21(1), 21(2); 36, 36(1), 36(2); 45, 45(1), 45a+45(1); 49, 49(1), 49(2); 50, 50(1)+50a(1), 50a, 50a(2)+50b(1), 50a(3), 50b; 51, 51a, 51b, 51+51a+51b(1), 51+51a+51b(2); 52, 52(1), 52(2), 52(3); 54, 54(1), 54(2); 55, 55(1), 55(2), 55(3); 58, 58a, 58b; 64, 64a, 64b; 65, 65(1), 65a; 67, 67(1), 67(2), 67(3); 82, 82(1)+ 82b, 82(2)+82c, 82(3), 82(4)+82a; 83, 83a+83(1), 83(2), 83(3).

Mesolithic/Neolithic Contact: single burials:

2; 14; 16; 17; 22; 23; 24; 43; 46; 75; 76.

double burials:

15, 15(1); 27, 27(1); 70, 70(1); 71, 71(1); 73, 73(1); 77, 77(1); 78, 78a.

multiple burials

29, 29(1), 29a; 69, 69(2), 69a, 69(1)+69a(1); 74, 74(1), 74(2).

Period uncertain: single burials:

1; 3.

double burials:

26, 26(1).

HAJDUČKA VODENICA Mesolithic/Neolithic Contact: missing burials:

1; 2; 3; 4; 5; 6; 7; 9; 10; 12.

single burials:

8; 11; 16; 21; 22; 30; 31.

double burials:

14, 14(1); 29, 29(1); 33, 33(1).

multiple burials:

15 m, 15 s, 15s(1); 17/17-20(3)/, 18/17-20/, 19/17-20(2)/, 20/17-20(4)/, 17-20(1), 20, 20(1); 23+24+25(1), 23+24+25(2), 23+24+25(3), 23+24+25(4), 23+24+25(5); 26+27+28(1), 26+27+28(2), 26+27+28(3), 26+27+28(4); 32, 32(1), 32(2).

60

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap

all of the sites; no pottery was found in these layers, and all of the Pre-Balkan Plateau flint was explained as intrusive (Srejović & Letica 1978; Kozłowski & Kozłowski 1982).

and Buikstra & Ubelaker (1994) as well as some additional characters observed during the first field season on the Đerdap material itself (sutura squamo-mastoidea and tuberculum marginale). The original list comprised 29 paired (that could be recorded bilaterally) traits and eight axial (that had only sagittal expression) traits for the skull, and 21 paired and one sagittal character for the postcranial skeleton. Procedures for recording followed Standards (Buikstra & Ubelaker 1994) where applicable, and Hauser & De Stefano (1989) in all other cases. Most of the traits were recorded on a scale rather than present or absent in order to allow more flexibility in the final analysis. However, they are treated as discrete in the statistical analysis. Since Buikstra & Ubelaker (1994) offer very little in terms of postcranial non-metric traits, a list of postcranial traits adapted from Czarnetski (1972b), Czarnetski et al. (1985) and Saunders (1978) was added. Only adult skeletons from all of sites were taken into consideration since the occurrence of non-metric traits in subadults can be ambiguous. This has reduced the total number of individuals examined from 438 (MNI) from all four sites to 259 (MNI) adult individuals. Since lack of intercorrelation and correlation with age and sex cannot be assumed, trait frequencies were compared with regards to age, sex and each other. Since the chance of purely random significant correlation occurring on the tested samples becomes greater with the number of correlation tests performed, the first step in the procedure was to remove all of the variables that could not be observed (both as absent or present) on at least 10% of the examined adult sample. This resulted in the elimination of seven cranial and nine postcranial traits. As, depending on the sample size, a small absolute number of occurrences (less than 5) can produce biased results, six cranial and four postcranial variables were excluded before their frequencies within subpopulations were examined. Of the remaining variables another group of characters, those with low overall frequencies, were checked against chronological and spatial subpopulations in order to assess their overall variability. (Sjøvold 1977; Molto 1983: 113). Sjøvold (1977) recognizes two types of low variability traits: those that have reached fixation in every sample studied, and others that have very low uniform incidence in any set of population samples. Rather than using the χ2 or Fisher’s exact test between samples to exclude the variables for which the significant difference is not obtained in at least one pairwise comparison (as suggested by Sjøvold 1977), the empirical results of Molto (1983: 114) for an Ontario Iroquois sample were applied. In Molto’s study (1983: 115) the largest range of frequencies among the traits that had low variability was 7.1 (for example, 0.0% in one sample to 7.1% in another). Molto has excluded these traits from further consideration and kept those with minimum range in any of the samples equal or greater than 10% (e.g. 21% in one and 31% in another). By using this observation as a rule of thumb in the present study, rather than increasing the possibility of finding statistical significance (where there might be none) through a large number of tests performed, three paired and one sagittal traits were determined as having low overall variability and excluded. Three variables were excluded because of the high inter-

Hajdučka Vodenica Within the proposed framework, all of the burials from Hajdučka Vodenica would fit within the Mesolithic/Neolithic contact period. Unlike the other sites, the graves show remarkable uniformity in the burial position and even orientation. This uniformity, as well as the restricted and wellrespected space for the burials, is well in accord with the (comparative to other sites) short time-span of the necropolis and of the chamber tomb.

Analyses Several reasons have guided preference towards non-metric traits: 1) They have been understudied in the present series. 2) They are potentially more appropriate for material of unequal preservation, as we are discussing population frequencies and not individual characteristics. This would be increasingly important in case of differential treatment for different groups. 3) Although there is an ongoing discussion among anthropologists about the ability of different analyses to establish relationships among skeletal samples, a paired study of analyses of diverse non-metric traits and cranial measurements demonstrates that the former are more powerful in this respect (Jackes et al. 1997). Non-metric traits do not follow simple Mendelian patterns of inheritance. Hauser & De Stefano (1989: 5–10) accept the model of ‘threshold character’ proposed by Falconer (1965) in relation to pathological conditions, as the underlying theoretical basis for all of the characters. The liability to develop a trait is normally distributed, and depending on the position of the individual’s inherited tendency to develop the character relative to the threshold, the character may or may not be expressed. The genes involved are multiple genes with small additive frequencies. Threshold models permit a number of other environmental and developmental factors to be included in the determination of the trait’s expression and allow for the observed gradients in some of the traits. An individual situated just below the threshold in one environment may be pushed over it in another (Hauser & De Stefano 1989: 7) which reinforces the population specific character of the frequencies of trait expressions. The proportion of total variance attributed to the additive effects of genes, known as the heritability of the trait, was calculated from the study of the frequency of the condition in a series of related individuals of known sex and age (Sjøvold 1984) and was shown to be significant although low. However, any attempt to relate individuals within a series to one another failed to perform, because of this underlying complex genetic basis of the traits (e.g. Crubezy 1991).

Choice of characters The choice is based on a number of characters for which low environmental influences were suggested by Saunders (1989) 61

The Iron Gates in Prehistory Table 3. Traits used in various combinations in the final analyses. Sides pooled. Trait name – common

Latin

Code

Trait no.

Marginal tubercle

tuberculum marginale

(TZ)

1

Squamomastoid suture

sutura squamomastoidea

(SSM)

2

Supraorbital notch

incisura supraorbitalis

(SN)

3

Supraorbital foramen

supraorbital foramen

(SF)

4

Zygomatico facial foramen

foramen zygomatico-faciale

(ZFF)

5

Parietal foramen

foramen parietale presens

(PF)

6

Coronal ossicle

(COR)

7

Lambda ossicle

(LAM)

8

(AUDT)

10

Auditory torus

torus auditivus

Mental foramen

foramne mentale

(MEF)

11

Maxillary torus

torus maxilaris

(MAX)

12

Mylohyoid bridge

ponticulus mylohyoideus

(MHBD)

13

Septal aperture

perforatio fossae olecranii

(PFO)

15

Third trochanter

trochanter tertius

(TT)

17

(APIC)

23

(INCA)

24

(PAL)

26

Apical bone Inca bone

os inca

Palatine torus

torus palatinus

or intra-observer error calculated from the observations recorded in 1996 and those recorded in the 1998 field season on a randomly chosen subsample. The Fisher exact test of significance was performed on the rest of the traits in order to check for possible correlation with sex. Only one variable pair was found to be potentially correlated with sex: the mandibular torus. In the published literature, there is no definite pattern of preference according to sex, but the general trend of predominance in females is reported (Hauser & De Stefano 1989, and quoted literature). The trait was excluded from further consideration. Tests of side correlation were performed on all of the paired traits. The ones that showed correlation were excluded. In doing so the risk of increasing the probability of false correlation was ignored, as potential benefits in increasing the number of observations by pooling the sides outweighed the concerns. Traits that had less than 9 observations on left and right side combined in any of the subsamples (see further discussion on the sides recording of the traits) were also excluded. Only 17 traits were used in different combination in the analyses (Table 3). Since a more restricted number of appropriately chosen variables can distinguish better between populations (Molto 1983; Krenzer 1996) this reduction was deemed beneficial. Trait description and scoring procedures are detailed in Roksandic (1999). Three assumptions were made: 1) Non-metric traits have a strong genetic basis. 2) The sample is representative of the populations we are trying to compare. 3) The outlier is a substantially removed population studied by the same methods.

Assumption 1 Although the influence of environment (occupation, habitat, nutrition) cannot be excluded for most traits, assumption is reasonably well founded in the research on the genetics of non-metric traits. We are examining the population structure and not the genetic make-up of the individuals, and therefore even if the influence of environment on the expression of traits (due to their threshold character) cannot be disregarded, the validity of the population comparisons is not reduced. Assumption 2 The assumption of the representativeness of the sample that is examined can rarely be proven in archaeological samples. Burial samples are often (if not always) biased as not everyone gets a formal burial, and mode of deposition is strongly dependent on the social persona (Masset 1993; Roksandic 2002, and quoted literature). Furthermore, the direction of bias can be discerned only rarely, after a thorough study of all of the social, biological and taphonomic aspects of the skeleton. Although we can not assume the representativeness of the sample for the purpose of studying mortuary ritual and its social implications, there is a hope that the populational biology (or the genetic make-up) of a changing population will still be represented adequately. Only under the circumstances of a completely different burial ritual for the local and the supposed incoming population, that would obliterate one or the other from our sample, the assumption of the representativeness could not be sustained. Since burial ritual in the Mesolithic varies greatly and becomes more or less canonized only in the Lepenski Vir IIIb period which belongs to the Middle Neolithic (Antunović 1990), and since inhumation is a demonstrated pattern for both of the periods, there is no 62

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap

reason to suspect total obliteration of one group or another in the current sample.

for dendrograms were used. Monotonic scaling, Kruskal Stress (measuring how well the curve fits all the points), and two dimensions because of the small number of points plotted were found the most appropriate. In each of the MDS plot figures, captions include the scores for the two dimensions, the Kruskal stress of final configuration — which should be less than 0.1 in a ‘good fit’ (Wilkinson et al. 1996: 667) — and proportion of variance expressed. Of further methodlogical concern is the treatment of sides. I have opted for adding sides and treating the material by elements and not by individuals. This method is preferred by Ossenberg (1981). She argues that the observed correlation between the intensity of trait incidence and the proportion of bilateral expression reflects genetic factors since an individual expressing a trait bilaterally has a stronger dose of trait positive alleles than an individual with unilateral expression of the trait. Therefore, computing the frequency of a discrete trait on the basis of pooled sides quantifies the genetic potential in the population better than does the individual count. This way of recording has the benefit of expressing the underlying threshold character of the traits as it takes into account the total genetic potential for the trait expression within a population. It also increases the sample sizes in many cases without violating the biological bases of the trait expression. Accordingly, sides were added in the following manner: k/N L + k/N R = k/N or (2/5 + 3/8 = 5/13) where k is the number of instances in which the trait was recorded as present, while N is the total number of possible observations.

Assumption 3 Inclusion of Franzhausen I, a Bronze Age population from Austria (Wiltschke-Schrotta 1992), was based on a number of variables that were recorded in both studies, and the system of recording that followed the same general procedures (Czarnetzki 1972a, 1972b, 1972c; Hauser & De Stefano 1989; Buikstra & Ubelaker 1994). Also, the site is sufficiently removed chrono-spatially, but still within the same general geographical area. Importantly, Wiltschke-Schrotta has recorded sides separately and has presented the side data in a manner that made it possible to add them up without problems and obtain a methodologically comparable sample.

Statistical analysis of non-metric traits Following the publication of Berry & Berry’s (1967) article on population studies based on cranial non-metric traits, the Smith–Grewal statistic was commonly used for calculation of average distances between sample populations (Saunders 1989: 98). Smith’s Mean Measure of Divergence (MMD) has further been investigated and developed by Sjøvold (1977) and serves, with minor modifications, as the major statistic used for examining the inter-sample distance. Stated simply, the Mean Measure of Divergence (MMD) is the summed divergence between two samples, divided by the number of traits included in the analysis. For this type of analysis, proportions of the sample exhibiting a trait are given as theta (θ) values symmetrical around 0, such that the incidence of 50% equals a theta of zero (Jackes et al. 1997: 645). Sjøvold has determined that the Anscombe formula is the best modification for calculation of θ, most accurately transforming the incidences of traits and stabilizing the variance well, except in cases when incidences are extremely high or low (Sjøvold 1977). If the sample sizes are small and incidences are accordingly low, the Freeman–Tukey transformation is judged to provide somewhat better variance stabilization than Anscombe (Jackes et al. 1997: 645). The actual formulae used in this study were taken from Jackes et al. (1997). The programming as well as the running of some of the data sets was done by Mary Jackes on the QuattroPro spreadsheet program at the University of Alberta in Edmonton. Others were run by myself on the Microsoft Excel program provided by Jackes. In analyzing the distance between the populations, it is possible to use MMD, Z and ‘stand’, all of which are represented on the tables. The choice here is based on the fact that Z is a way of standardizing the MMD in case of unequal sample sizes and is more explicitly correlated with both MMD and DI, the latter being the measure of significance of the distance (Jackes & Gao, forthcoming). Therefore Z matrices are reproduced with the results. Since dendrograms can link the samples in only one direction, a spatial distance between different samples can be better appreciated through multidimensional scaling (MDS) plots. To produce MDS plots, the same distance matrices as

Subsamples based on sites with Franzhausen I The results of these analyses are presented in Tables 4–6 and Figure 1. A glance at Table 6 shows that Hajdučka Vodenica and Padina are virtually identical, while all other sites seem to be significantly different from each other. Franzhausen, as expected, is the most removed from other sites, while Padina and Lepenski Vir and Padina and Vlasac show the next most significant difference. The multidimensional scaling plot also shows the clear grouping of the Iron Gates gorge sites against the more removed Franzhausen I site. The general pattern is that of heterogeneity (Fig. 1).

Subsamples based on chronology The results of these analyses are presented in Tables 7–9 and Figure 2. As described above, the basis for distinguishing the units is provided by the evidence of economic behaviour and evidence of contact with peoples with different economic patterns. Mesolithic refers to the strata within any of the sites where there is no evidence of contact. The Mesolithic/Neolithic is the period when contact with farming communities in the region becomes possible, while Neolithic is, primarily, characterized by greater importance of domesticates in the economic base (>5%). Evidence of adoption of 63

The Iron Gates in Prehistory Table 4. ‘k’ and ‘N’ values used in the analysis of sites with Franzhausen I. Trait no.

1 k

HV N

2 k

L. Vir N

3 k

Padina N

2

4

6 10

4 k

Vlasac N

5 k

FR.I

9

14

46

13

26

4

62

8

588

7

10

31

60

23

28

44

77

211

425

7

22

5

48

17

33

27

76

130

638

11

1

21

2

60

3

40

6

82

46

530

12

1

9

6

41

1

10

15

59

8

365

13

2

14

12

18

4

25

16

32

28

446

17

0

10

12

29

4

13

13

35

116

318

Table 5. The output of the statistical analysis of sites with Franzhausen I. Site 1

Site 2

Site 1 name

1

2

HVmn

1

3

HVmn

1

4

1 2

Site 2 name

mmd FT

sd FT

Stand FT

LV

0.2954

0.0590

5.0063

P

0.0233

0.0724

0.3214

HVmn

V

0.2678

0.0547

5

HVmn

FRI

0.3548

3

LV

P

0.2900

2

4

LV

V

2

5

LV

3

4

P

3

5

4

5

Total N

Z FT

di FT

S FT

formula

57

3.4171

0.1774

24.6591

ft

39

0.2896 -0.1216

7.5861

ft

4.8985

74

3.0388

0.1585

22.0737

ft

0.0460

7.7127

486

4.2742

0.2628

31.0451

ft

0.0429

6.7659

68

5.3277

0.2043

39.9015

ft

0.0946

0.0263

3.5981

104

3.5686

0.0420

25.7342

ft

FRI

0.4219

0.0169 24.9836

516

9.5513

0.3881

86.5518

ft

V

0.2290

0.0385

5.9491

85

4.9147

0.1520

36.2973

ft

P

FRI

0.3769

0.0299 12.5989

498

8.7299

0.3171

76.0812

ft

V

FRI

0.2554

0.0118 21.6578

533

8.9086

0.2319

78.3023

ft

Table 6. Matrix of Z values for sites with Franzhausen I. Significant relationships are outlined in bold. Z(ft) matrix

1 H. Vodenica

1

0

2

3.4171

0

0.2896

5.3277

0

3.0388

3.5686

4.9147

0

4.2742

9.5513

8.7299

8.9086

3 4 5

2 L. Vir

3 Padina

4 Vlasac

5 FR.I

0

Dimensions

1

HV

0.16

2 -0.14

LV

1.00

0.39

PA

0.12

-1.00

VL

0.23

0.69

FR

-1.51

0.05

Kruskal Stress of final configuration: 0.0270 Proportion of variance: 0.9941

Figure 1. Multidimensional scaling for the sites with Franzhausen I. Based on dissimilarity matrix.

64

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap Table 7. ‘k’ and ‘N’ values for traits used in the analysis of chronological units. Trait no.

Mesolithic k

1 N

Meso/Neo k

2 N

Neolithic k

3 N

1

30

33

20

45

1

8

2

10

55

16

66

8

14

3

40

60

43

77

6

14

4

9

66

14

65

2

12

5

23

42

29

43

4

9

6

42

67

48

76

8

18

7

0

62

4

70

0

18

8

16

62

23

73

5

16

10

30

50

26

85

0

15

11

6

81

4

82

1

18

12

12

46

7

51

2

13

23

8

33

3

32

0

10

Table 8. The output of the analysis of chronological units. Site 1

Site 2

Site 1 name

Site 2 name

mmd FT

sd FT

Stand FT

1

2

Meso

MN

0.1298

0.0155

8.3884

1

3

Meso

LVn

0.4667

0.0398

11.7260

2

3

MN

LVn

0.1017

0.0388

2.6230

Total N

Z FT

di FT

S FT

formula

119

4.8968

0.0988

46.9740

ft

69

6.4826

0.3871

63.6011

ft

78

2.4289

0.0242

26.0986

ft

Table 9. Matrix of Z values for chronological units. Significant relationships are outlined in bold. Z(ft) matrix

1 Mesolithic

2 Meso/Neo

1

0

2

4.8968

0

3

6.4826

2.4289

3 Neolithic

0

Dimensions Meso

1 -1.34

2 0.00

Contact

0.27

0.00

Neo

1.07

0.00

Kruskal Stress of final configuration: 0.00 Proportion of variance: 1.00

Figure 2. a) Dendrogram showing internal relationship of chronological units.Derived from dissimilarity matrix, Euclidean distance, complete linkage. b) MDS plot of chronological units based on dissimilarity matrix.

65

The Iron Gates in Prehistory

Subsamples based on combined chronology and sites

cultural elements of the surrounding farmers of the Balkano–Anatolian and Balkano–Carpathian basin (ENCB and MNCB of Tasić 1998) although considered, was not taken as sufficient for determining the find as Neolithic. Clustering of the Mesolithic/Neolithic (Contact) period with the Neolithic period contradicts the wave of advance model for the neolithization of this region. In order to demonstrate the spread of Neolithic farmers themselves, and not only their domesticates and/or knowledge, the result should show a slight to non-existent change in the Contact period (as some exchange of genes could be expected) and an abrupt change with the advent of the Neolithic. This pattern would argue for an insurgence of people with different genetic make-up who brought about the change in the economic base (as proposed by Ammerman & Cavalli Sforza 1971; 1984; Cavalli Sforza 1996). All of the relationships are significant according to the ‘di’ value (Table 7). It is noteworthy that the distance between Mesolithic and Contact is more than twice the distance between Contact and Neolithic. This is even more suggestively shown by the multidimensional scaling plot. If there is, indeed, an exchange of genes as well as goods at the time of the first contact, it does not destabilize the Mesolithic society and ideology. Even more importantly, as the basis of subsistence remains hunting, gathering and fishing, this supposed exchange of genes does not bring about a fully developed farming economy. It is important to note that a certain amount of change in the genetic make-up, as evidenced in the non-metric traits, would be expected due to secular trends. However, for secular trends to be the only source of change, the distances between different periods would need to be approximately the same. The diagram in Figure 2b strongly suggests that, apart from the obvious secular trend reflected in the alignment of the units, a greater amount of change happens between Mesolithic and Contact periods. The introduction of an outlier in the next analysis is aimed at clarifying how important this difference was in the amount of genetic change.

Subsamples that respected both sites and chronological determination were analyzed in order to provide a more finegrained understanding of the relationship among them (Tables 13–18; Figs 4–5). The Padina sample, when divided into Mesolithic and Mesolithic/Neolithic, made comparisons almost impossible because of the small numbers of observations in almost all variables. Since Padina clusters consistently with Hajdučka Vodenica, and since most of the individuals from Padina belonged to the same chronological unit (Mesolithic/Neolithic) they were assigned to the Hajdučka Vodenica subsample and thus form the HVPmn (Hajdučka Vodenica–Padina Mesolithic/Neolithic). The remaining individuals from Padina that belong to the Mesolithic were assigned to the Lepenski Vir Mesolithic subsample (forming LVPm — Lepenski Vir–Padina Mesolithic). This was done in order to strengthen the Mesolithic sample of Lepenski Vir after a careful examination of frequencies. Although the frequencies do not show substantial differences, this should be kept in mind in the analyses and interpretation. According to the distance matrix produced here, the most similar are the Hajdučka Vodenica/Padina Contact group and the Lepenski Vir Neolithic. This points to a strong continuity between the two periods and is important because this continuity is not site specific. Along these lines is the similarity between Lepenski Vir Contact and Lepenski Vir Neolithic. But the Hajdučka Vodenica/Padina group also shows little difference from the Lepenski Vir Mesolithic subsample. At Lepenski Vir itself, the change is pronounced at the time of Contact and very restricted between Contact and Neolithic. It is interesting to note that Lepenski Vir Mesolithic is most different from Vlasac Mesolithic and Lepenski Vir Contact and less, but still significantly, different from Lepenski Vir Neolithic. Lepenski Vir Neolithic shows little difference from Vlasac Mesolithic and somewhat more from Vlasac Contact. Vlasac Mesolithic and Mesolithic/Neolithic appear to be virtually identical along the second dimension and different along the first dimension where they are pooled by similarity to Lepenski Vir Neolithic. The general outline argues for temporal trend and continuity within the sample with greater variability in the Contact period. According to the ‘di’ values (Table 14), several distances are non-significant: Lepenski Vir Contact shows little distance from the Neolithic period at the same site, and there is a clear continuity between Mesolithic and Contact at Vlasac. Such a strong association raises doubts that the chronological determination based on stratigraphy and stylistic analysis (Radovanović 1996a) could be incorrect. In order to check if this grouping is indeed evidence of continuity, and not a consequence of unreliable separation into different chronological groups, a different chronological division of the sample, based on Srejović’s (Srejović & Letica 1978) determinations was run through the same procedures. None of the significant relationships changed and, more importantly, Vlasac retains practically the same non-difference for its two chronological subsamples. Lepenski Vir Neolithic is equidistant from Vlasac Mesolithic and Vlasac Contact.

Subsamples based on chronology with Franzhausen I These results are presented in Tables 10–12 and Figure 3. As is evident from Table 10, the difference between Contact and Neolithic periods ceases to be significant when an outlier is introduced (shown by a negative ‘di’ value). Furthermore, distances between Mesolithic and Contact, and Contact and Neolithic, according to the Z statistic, become almost equal. The introduction of Franzhausen shows that the distances on the local scale become less obvious and that in general they follow the secular trend. While the Contact period is almost equidistant from the Mesolithic and Neolithic periods, the two are differently positioned with respect to Franzhausen (Bronze Age). In other words, the Contact, Neolithic and Franzhausen are to be found on the same axis, while the Mesolithic period forms a different pattern and is situated on a different axis together with the Contact period. 66

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap Table 10. ‘k’ and ‘N’ values for traits. Analysis of chronological units with Franzhausen I. Trait no.

Mesolithic k

1 N

Meso/Neo k

2 N

Neolithic k

3 N

Franz.I k

4 N

2

10

55

16

66

8

14

8

588

5

23

42

29

43

4

9

90

451

6

42

67

48

76

8

18

211

425

8

16

62

23

73

5

16

27

190

10

30

50

26

85

0

15

130

638

11

6

81

4

82

1

18

46

530

12

12

46

7

51

2

13

8

365

13

16

39

15

42

1

5

28

446

23

8

33

3

32

0

10

16

194

Table 11. The output of the analysis of chronological units with Franzhausen I. Site 1

Site 2

Site 1 name

1

2

Meso

1

3

1

4

2

Site 2 name

mmd FT

sdFT

Stand FT

MN

0.0411

0.0187

2.1905

Meso

N

0.3244

0.0530

6.1157

Meso

FrI

0.3561

0.0111 31.9687

3

MN

N

0.1024

0.0522

1.9597

2

4

MN

FrI

0.2874

4

3

FrI

N

0.3327

Z FT

di FT

S FT

formula

2.0405

0.0036

18.9949

ft

66

5.1131

0.2183

42.6539

ft

478

12.7567

0.3339

142.4636

ft

74

2.4522

-0.0021

21.6171

ft

0.0103 27.8134

486

12.0167

0.2668

130.2460

ft

0.0448

438

5.8392

0.2430

49.6241

ft

7.4201

Total N 114

Table 12. Matrix of Z values for chronological units with Franzhausen I. Significant relationships are outlined in bold. Z(ft) matrix

1 Mesolithic

2 Meso/Neo

3 Neolithic

1

0.0000

2

2.0405

0.0000

3

5.1131

2.4522

0.0000

4

12.7567

12.0167

5.8392

Dimensions Meso

4 Franz.I

0.0000

1

2

-0.88

0.48

Contact

0.75

-0.35

Neo

0.15

-0.27

Franz. I

1.48

0.14

Kruskal Stress of final configuration: 0.00 Proportion of variance: 1.00

Figure 3. Multidimensional scaling plots for chonological units with Franzhausen I. Based on dissimilarity matrix.

67

The Iron Gates in Prehistory Table 13. ‘k’ and ‘N’ values for traits used in the analysis of subsamples based on site/chronology combination. Trait no.

HVP(mn) k

1 N

LVP(m) k

2 N

LV(mn) k

3 N

LV(n) k

4 N

V(m) k

5 N

V(mn) k

6 N

1

6

11

12

15

2

11

1

8

18

28

12

23

2

11

24

7

19

4

15

8

14

3

36

1

26

3

16

26

20

22

8

19

6

14

20

48

19

32

4

4

19

2

20

5

18

2

12

7

46

5

28

5

7

12

10

12

5

9

4

9

13

30

17

22

6

17

24

19

22

12

22

8

18

23

45

19

30

7

0

17

0

20

2

18

0

18

0

42

2

25

8

4

17

2

18

4

18

5

16

14

44

15

28

11

4

34

0

27

0

20

1

18

6

54

0

28

13

5

24

4

19

6

6

1

5

12

20

4

26

23

1

9

2

9

0

9

0

10

6

24

2

14

24

1

11

1

10

2

11

0

9

0

28

0

15

Table 14. The output of the analysis for site/chronology combination. Site 1

Site 2

mmd FT

sd FT

Stand FT

1

2

1

3

HVPmn LVPm

0.0205

0.0520

0.3935

HVPmn LVmn

0.2698

0.0584

4.6233

1

4

HVPmn LVn

-0.0104

0.0630

1

5

HVPmn Vm

0.0959

1

6

HVPmn Vmn

0.0875

2

3

LVPm

LVmn

4

2

LVn

2

5

LVPm

2

6

3 3

Total N

Z FT

di FT

S FT

formula

37

1.0880

-0.0836

17.3097

ft

34

3.0129

0.1531

30.4878

ft

-0.1647

32

-0.1692

-0.1364

10.7029

ft

0.0381

2.5196

56

2.6350

0.0198

27.6085

ft

0.0437

2.0041

44

2.5068

0.0002

26.6643

ft

0.5018

0.0586

8.5607

32

5.3937

0.3846

51.9135

ft

LVPm

0.3109

0.0632

4.9183

30

4.0290

0.1845

38.9387

ft

Vm

0.2715

0.0382

7.1145

55

5.6937

0.1952

55.0150

ft

LVPm

Vmn

0.1599

0.0438

3.6504

43

3.5494

0.0723

34.8213

ft

4

LVmn

LVn

0.1905

0.0718

2.6525

27

1.2510

0.0469

18.2819

ft

5

LVmn

Vm

0.2242

0.0459

4.8892

52

3.3118

0.1325

32.8670

ft

3

6

LVmn

Vmn

0.3885

0.0505

7.6860

39

4.1027

0.2874

39.5918

ft

4

5

LVn

Vm

0.1661

0.0507

3.2744

50

2.4544

0.0647

26.2831

ft

4

6

LVn

Vmn

0.1712

0.0553

3.0968

37

3.0302

0.0606

30.6230

ft

6

5

Vmn

Vm

0.1217

0.0300

4.0608

62

3.2961

0.0617

32.7397

ft

Table 15. Matrix of Z values for site/chronology combination Significant relationships are outlined in bold. Z(ft) matrix

1 HVP(mn)

2 LVP(mn)

3 LV(mn)

1

0

2

1.0880

0

3

3.0129

5.3937

0

4

-0.1692

4.0290

1.2510

0

5

2.6350

5.6937

3.3118

2.4544

0

6

2.5068

3.5494

4.1027

3.0302

3.2961

68

4 LV(n)

5 V(m)

6 V(mn)

0

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap

Dimensions HVPmn

1

2

-0.39

0.20

LVPm

1.33

-0.24

LVmn

0.97

0.79

LVn

0.27

0.34

Vm

0.89

0.63

Vmn

-0.41

-0.95

Kruskal Stress of final configuration: 0.027 Proportion of variance: 0.993

Figure 4. Multidimensional scaling plot showing internal relationship between the site/chronology units in the Iron Gates gorge. Based on dissimilarity matrix.

This, somewhat unorthodox approach was suggested by M. Jackes (pers. comm.) based on a published analysis by Christensen (1997). Simply stated, ‘θ’ values are treated as ordinal values and submitted to the PCA in order to ascertain which of the traits contributed the most to the observed pattern. The site chronology analysis being the one on which the interpretation is mostly based, it was deemed unnecessary to subject the results of other analyses to the same procedures. The output in Table 19 shows that trait 1 (marginal tubercle 0.90), trait 3 (supraorbital notch 0.92), trait 6 (parietal foramen 0.89) and trait 23 (apical bone 0.80), contribute the most to the first dimension. In real ordinal data this dimension represents size; here it shows the traits with high frequencies. Trait 4 exhibits a strong negative association with dimension one (supraorbital foramen – 0.89), while trait no 13 (mylohyoid bridge – 0.59) shows negative association of a lesser extent. On the second dimension, trait 24 (Inca bone 0.89) has a strong positive association, while trait no 11 (double mental foramen – 0.69) has a strong but negative association. The two components explain 66.6% of the total variation within the sample. In Varimax rotation (Table 20), applied to reduce the number of variables on the ‘size’ axis, Variables 1 and 23 show even stronger positive association (0.97 and 0.98, respectively) while variable 4 shows strong negative association. On the second component, variable 11 shows even stronger negative association, while variable 5 shows the strongest positive association with this component. These two components explain 55.7% of the variation in the sample. All of the analyses (with or without Franzhausen) show a strong temporal trend. A significant amount of change within the examined population may be due to non-directional micro-evolution that is expected for a series covering a timespan of 1500 years. However, the position of Hajdučka Vodenica and Padina and the MDS of chronological units show significantly more change occurring between Mesolithic and Contact period (due to the availability of

The relationships between different site/chronological units become more complex. Lepenski Vir and Padina in the Mesolithic resemble the Contact period at Hajdučka Vodenica and Padina. This could be due to the fact that Padina is present in both components. However, while the Padina Contact sample is almost the same size as Hajdučka Vodenica in the same period, the Padina Mesolithic sample is very small and comparable in frequencies to Lepenski Vir Mesolithic. Therefore it is unlikely that the analysis would pool these two sites together were they different. As stated, a strong case of continuity is present between Vlasac Mesolithic and Vlasac Contact, as well as Lepenski Vir Contact and Neolithic groups. Some shifting and moving of population within the region could explain the similarities between the Contact subsamples at Lepenski Vir, Vlasac, Hajdučka Vodenica and Padina. This would coincide with Radovanović’s phase of greater territorial integrity and more ideological integration in the region (Radovanović 1995, 1996a, 1996b, 1996d). Availability of contact with Neolithic farmers in the region could have emphasized the need for ideological and conceptual unity among the foragers. Franzhausen, as expected, is far removed from the rest of the sample and is a definite outlier. The sites examined form a pattern similar to the ‘horseshoe’ shape typical of chronological series (Greenacre 1984). However, several features contradict an interpretation of the pattern as reflecting merely change over time. According to the temporal change explanation, the Mesolithic sites should be on one end, Mesolithic/Neolithic at the bottom and Neolithic on the other end of the ‘horseshoe’ diagram. Although the pattern observed reflects this situation, Hajdučka Vodenica and Padina Contact and Vlasac Mesolithic do not fit perfectly.

Contribution of traits Theta values obtained in the Site/Chronology analysis (above) were submitted to Principal Component Analysis. 69

The Iron Gates in Prehistory Table 16. ‘k’ and ‘N’ values for traits used in the analysis of subsamples based on site/chronology combination with Franzhausen I. Trait no.

HVP(mn) k

1 N

LVP(m) k

2 N

LV(mn) k

3 N

LV(n) k

4 N

V(m) k

5 N

V(mn) k

6 N

FR.I k

2

11

24

7

19

4

15

8

14

3

36

1

26

8

588

5

7

12

10

12

5

9

4

9

13

30

17

22

90

451

6

17

24

19

22

12

22

8

18

23

45

19

30

211

425

8

4

17

2

18

4

18

5

16

14

44

15

28

27

190

10

16

38

12

23

1

18

0

15

18

47

9

29

130

638

11

4

34

0

27

0

20

1

18

6

54

0

28

46

530

12

1

16

3

13

0

9

2

13

9

33

6

26

8

365

23

1

9

2

9

0

9

0

10

6

24

2

14

16

194

Table 17. The output of the analysis for site/chronology combination with Franzhausen I. Site 1

Site 2

mmd FT

sd FT

Stand FT

1

2

HVPmn LVPm

0.0236

0.0625

0.3779

1

3

HVPmn LVmn

0.0638

0.0683

1

4

HVPmn LVn

0.1346

1

5

HVPmn Vm

0.0891

1

6

HVPmn Vmn

1

7

2

3

2

Total N

Z FT

di FT

S FT

formula

40

0.8608

-0.1014

11.2045

ft

0.9335

37

1.8817

-0.0729

16.5584

ft

0.0668

2.0161

36

2.5822

0.0011

20.8349

ft

0.0439

2.0302

61

2.1528

0.0013

18.1553

ft

0.1679

0.0510

3.2912

47

3.3415

0.0659

26.0245

ft

HVPmn FRI

0.2940

0.0317

9.2615

444

7.1037

0.2305

60.2433

ft

LVPm

LVmn

0.2480

0.0708

3.5028

33

3.1803

0.1064

24.8741

ft

4

LVPm

LVn

0.4032

0.0692

5.8225

32

4.5975

0.2647

35.8746

ft

2

5

LVPm

Vm

0.1918

0.0462

4.1505

57

3.7504

0.0994

29.0578

ft

2

6

LVPm

Vmn

0.1510

0.0534

2.8303

43

2.9377

0.0443

23.1927

ft

2

7

LVPm

FRI

0.5434

0.0339 16.0297

441

8.7306

0.4756

79.4253

ft

3

4

LVmn

LVn

-0.0231

0.0754 -0.3070

29

-0.1041

-0.1740

7.1021

ft

3

5

LVmn

Vm

0.1954

0.0523

3.7372

54

2.9894

0.0908

23.5464

ft

3

6

LVmn

Vmn

0.1621

0.0593

2.7350

40

2.4154

0.0436

19.7722

ft

3

7

LVmn

FRI

0.1425

0.0400

3.5631

438

3.0453

0.0625

23.9311

ft

4

5

LVn

Vm

0.2797

0.0509

5.4949

53

4.0364

0.1779

31.2792

ft

4

6

LVn

Vmn

0.3378

0.0580

5.8255

40

4.2705

0.2218

33.1581

ft

4

7

LVn

FRI

0.3628

0.0385

9.4119

437

5.9376

0.2857

48.1238

ft

5

6

Vm

Vmn

0.0671

0.0349

1.9210

65

1.8855

-0.0028

16.5800

ft

5

7

Vm

FRI

0.1764

0.0153 11.5023

462

6.3152

0.1457

51.8994

ft

6

7

Vmn

FRI

0.3445

0.0226 15.2427

448

8.1575

0.2993

72.3665

ft

Table 18. Matrix based of Z values for site/chronology combination with Franzhausen I. Significant relationships are outlined in bold. Z(ft) matrix

1 HVP(mn)

2 LVP(m)

3 LV(mn)

4 LV(n)

1

0

2

0.8608

0

3

1.8817

3.1803

0

4

2.5822

4.5975

-0.1041

0

5

2.1528

3.7504

2.9894

4.0364

0

6

3.3415

2.9377

2.4154

4.2705

1.8855

0

7

7.1037

8.7306

3.0453

5.9376

6.3152

8.1575

70

5 V(m)

6 V(mn)

7 FR.I

0

7 N

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap

Dimensions

1

2

HVPmn

0.46

0.42

LVPm

1.10

0.31

LVmn

-0.25

0.14

LVn

-0.42

0.77

Vm

0.15

-0.77

Vmn Franz. I

0.67

-0.69

-1.70

-0.19

Kruskal Stress of final configuration: 0.064 Proportion of variance: 0.97

Figure 5. Multidimensional Scaling plot showing internal relationships between site/chronology units with the introduction of Franzhausen. Based on dissimilarity matrix.

contact with a different population?) than between Contact and Neolithic. This can be interpreted as showing a high degree of population heterogeneity during the Mesolithic and/or availability of contact with some influx (but not on a large scale) of new genes from a different population(s). The degree of difference (as well as economic behaviour) of these populations with the indigenous foragers remains, of course, impossible to assess from the current study.

greater isolation of Vlasac as a specific locality? In terms of anthropological change within the period, some regrouping of the population is evident. Vlasac seems to be very closed and little change occurs at the time of availability of contact. Similarities between Padina, Hajdučka Vodenica and Lepenski Vir seem to point towards greater mobility within the group as a result of possible pressure from the outside. This seems to confirm Radovanović’s interpretation of this period as a phase of consolidation of the Lepenski Vir culture, with greater ideological integration among the previously dispersed and distinct sites, reflected archaeologically in increased evidence of art and ritual (Radovanović 1996b, 1996c, 1996d; Radovanović & Voytek 1997). Can the observed difference between Mesolithic and Contact populations be attributed to admixture between the existing local subpopulations of the Mesolithic Iron Gates gorge, or does it provide evidence for the influx of other, more remote, genes? The observed differences between the chronological units examined seems to be largely due to a secular trend. At the time of first contact with the Neolithic population, a slightly more marked change in the genetic profile of the population occurs. This indicates higher levels of admixture with a nonlocal population. It could have been brought about by an influx of non-local foragers, by an influx of surrounding farmers, or both. In order to answer this question with certainty, a better understanding of local Neolithic populations as well as a wider base of the Mesolithic Iron Gates populations (both from the Romanian side of the Danube as well as from sites situated further away) would be needed. The Neolithic site of Velesnica contained only three female skeletons, while Ajmana (with 17 individuals) was not available for study at the time of this research. The published report by Radosavljević-Krunić (1986) does not give enough information for its inclusion in any of the analyses undertaken. On the Romanian side, only Schela Cladovei has yielded a signific-

Combining the lines of interpretation In view of the proposed porous frontier between Mesolithic and Neolithic cultures in the study region, can we presume interactions between bearers of these respective cultures? What forms did these interactions take? Was neolithization their ultimate consequence? Interactions, in the broadest sense, involving any amount of change within a population and resulting from the availability of contact or presence of another population, can be assumed even without any specific explanatory mechanism. It is improbable that two populations existing in a relatively restricted geographic area would never interfere or interact with one another. Beyond assumptions, the contact between groups with distinct material culture, which in the case of Đerdap archaeology correspond well to subsistence groups, is evidenced on many of the sites in the region through the exchange of trade items. The question is therefore centred more on the nature and consequences of this contact than on its existence. First it is important to stress that this contact need not be uniform and could have been site specific. For example, while there is no evidence for ceramics in the Contact period at Vlasac, Hajdučka Vodenica is rich in potsherds, and ceramics were found in situ in Padina houses. Is the close clustering of the two periods at Vlasac indicative of the 71

The Iron Gates in Prehistory Table 19. PCA output for Theta values of traits analyzed in the text. Latent Roots (Eigenvalues) 1

2

3

4

5

6

7

8

9

5.0401

2.9539

2.3817

1.1441

0.4803

0.0000

0.0000

0.0000

0.0000

Component loadings 1

2

3

4

V1

0.9061

-0.1902

0.0914

0.3483

V2

-0.0838

0.2467

-0.8662

-0.4264

V3

0.9282

0.3265

-0.0888

-0.1543

V4

-0.8026

0.4243

0.0552

0.0716

V5

0.7571

0.4489

0.4066

-0.2393

V6

0.8946

0.3982

-0.0725

0.0127

V7

-0.4109

0.6277

0.6539

-0.0315

V8

-0.3348

-0.4634

0.7356

-0.3070

V11

-0.2278

-0.6903

-0.5614

0.1728

V13

-0.5891

0.4175

0.0267

0.6585

V23

0.8072

-0.3803

0.1785

0.4010

V24

0.0366

0.8978

-0.3536

0.1874

Table 20. PCA output for Theta values of traits used in the text. Varimax rotation. Rotated Loading Matrix ( VARIMAX, Gamma = 1.0000) 1

2

3

4

0.9748

0.1010

0.0652

0.1492

-0.4468

-0.2620

0.7440

0.4220

V3

0.5848

0.4893

0.4623

0.4524

V4

-0.7235

0.0737

0.0293

-0.5501

V5

0.4276

0.8425

0.1079

0.3048

V6

0.6245

0.4965

0.5019

0.2757

V7

-0.4395

0.7166

-0.2264

-0.4832

V1 V2

V8

-0.2219

0.0959

-0.9464

0.0893

V11

0.0161

-0.9302

-0.0160

0.0872

V13

-0.2559

-0.0350

0.1576

-0.9295

V23

0.9802

-0.0225

-0.1318

0.1019

V24

-0.1509

0.3787

0.8303

-0.3345

‘Variance’ Explained by Rotated Components 1

2

3

4

3.8798

2.8133

2.7148

2.1118

Percent of Total Variance Explained 1

2

3

4

32.3321

23.4440

22.6232

17.5984

ant number of individuals (100+: C. Bonsall, pers. comm.) which are still under study, while a survey beyond the banks of the Danube on both the Romanian and Serbian sides is yet to be undertaken.

Based on the data presented here, the distances between the Mesolithic components of Vlasac, Lepenski Vir and Padina seem to be important. The great heterogeneity of the population observed by other researchers also supports this 72

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap

finding. However, a simple trend towards homogenization in the Contact period, would have resulted in the Mesolithic/Neolithic subsamples being positioned close to one another and equidistant from the Mesolithic subsamples. This is not the case. As stated earlier, Vlasac seems to remain the most isolated while a marked degree of similarity is observed between the Lepenski Vir Mesolithic and the Hajdučka Vodenica and Padina Contact periods. Some introduction of new genes is possible. Ascertaining that they come either from surrounding Neolithic people, or other foragers moving as a consequence of neolithization of the surrounding region would be over-interpreting the scant evidence. If some influx of new genes is indicated in the Contact period, the Neolithic in the region (in terms of population biology) represents a continuation of the local Mesolithic. This is evident in both the non-metric traits where the Neolithic group helps make the ‘horseshoe-shaped’ curve typical of temporal ordering, and in metrics where there is practically no significant difference between the Contact and the Neolithic groups (Roksandic 1999).

theAcknowledgements My research was made possible by Wenner-Gren Foundation for Anthropological Research Grant no. 6250, an Ada and Isabelle Steele Memorial Scholarship (SFU), and a Graduate Travel Award (SFU). For making the material available for the present research, I am indebted to Dr Borislav Jovanović and late Professor Dragoslav Srejović, principal investigators of the Iron Gates gorge sites studied. My thanks also go to Chris Mieklejohn and Mary Jackes for their invaluable support. locality, in smaller groups, and adopted a different material culture and architecture, but retained the same burial practices, and to a large extent the hunting and fishing economic base. The greater percentage of domestic animals and definite use of domesticated varieties of cereals classifies them as a Neolithic group, but in many respects this population remained unchanged. Only within the fully developed Starčevo phase (Lepenski Vir IIIb) and with a change in burial ritual towards more canonized forms (Antunović 1990) did this population finally integrate itself into a larger Neolithic community.

Conclusions

Note 1. Although this monocausal explanation, which was applied to all the mortuary monuments in Neolithic Europe, is overly simplistic and reveals more about the preoccupations of modern-day western scholars than the prehistoric inhabitants of Europe (cf. Cullen 1995: 286), it remains one of the possible, and even plausible, reasons but cannot be perceived as the only cause of the growing importance of mortuary ritual in the period (Masset 1993).

In conclusion, no large-scale population admixture can be demonstrated from the above data. Some ‘seeping in’ of the population suggested by Menk (Menk & Nemeskéri 1989: 531), but without the successive replacement that he argued for, can be proposed on the basis of current research. This ‘seeping in’ happens more perceptibly at the beginning of the contact, rather than at the time of change in subsistence. Even when the change in subsistence finally occurs, fishing and hunting still account for the major portion of the animal assemblage in both Neolithic sub-phases at Lepenski Vir. Although the reasons for the change in subsistence are beyond the scope of this research, it can be stated on the basis of the anthropological information that it is not brought about by an incoming population. It must be regarded as a consequence of cultural and social factors operating within the Mesolithic of Lepenski Vir itself, which brought about its disintegration. The Mesolithic Lepenski Vir culture is based on the rich riverine environment. Such environments tend to support more affluent societies, and these are not ‘among the first to make the transition to food production. Rather they appear to be late lasting in historic terms’ (Brinch Petersen & Meiklejohn, in press). The Lepenski Vir Mesolithic successfully paralleled local Neolithic developments over a long period of time. Contacts with the Neolithic population in the region seem to have helped to form an ideological unity of sites and thus brought into full expression the artistic achievements of an already affluent society. Ideological integration evidenced at the time of possible Contact could have resulted from the growing wealth of the sites based on trade in salt-preserved foodstuffs (fish from Iron Gates gorge and wild game from Gura Baciului) as proposed by Tasić (1998). Internal conflicts, over-exploitation of the environment and innumerable other factors may have played a role in the disintegration of the Lepenski Vir tradition. The biological descendants of the Lepenski Vir culture remained in

Acknowledgements My research was made possible by Wenner-Gren Foundation for Anthropological Research Grant no. 6250, an Ada and Isabelle Steele Memorial Scholarship (SFU), and a Graduate Travel Award (SFU). For making the material available for the present research, I am indebted to Dr Borislav Jovanović and late Professor Dragoslav Srejović, principal investigators of the Iron Gates gorge sites studied. My thanks also go to Chris Mieklejohn and Mary Jackes for their invaluable support.

References Ammerman, A.J. & Cavalli-Sforza, L.L. 1971: Measuring the rate and spread of early farming in Europe. Man 6: 674–688. Ammerman, A.J. & Cavalli-Sforza, L.L. 1984: The Neolithic Transition and the Genetics of Populations in Europe. Princeton, New Jersey: Princeton University Press. Antunović, M. 1987: Populacija i način sahranjivanja na Centralnom Balkanu tokom ranog I srednjeg Neolita. Unpublished BA Honours paper, Department of Archaeology, University of Belgrade. — 1990: Anthropological and archaeological survey concerning mortuary practices in the central area of Balkan Peninsula during Early and Middle Neolithic. In Cahen, D. & Otte, M. (eds) Rubané et Cardial, Actes du colloque international de Liège (11–13 décembre 1988). E.R.A.U.L. 39. Liège: University of Liège, 39–50. Bartosiewicz, L., Boroneanţ, V., Bonsall, C. & Stallibrass, S. 2001: New data on the prehistoric fauna of the Iron Gates: a case study from Schela Cladovei, Romania. In Kertész, R. & Makkay, J. 1

73

The Iron Gates in Prehistory (eds) From the Mesolithic to the Neolithic. Budapest: Archaeolingua (Main Series), 15–22. Berry, A.C. & Berry, R.J. 1967: Epigenetic variation in the human cranium. Journal of Anatomy 101: 361–379. Bolomey, A. 1973: An outline of the late Epipalaeolithic economy at the ‘Iron Gates’: the evidence of bones. Dacia n.s. 27: 41–52. Bökönyi, S. 1969: Kičmenjaci (prethodni izveštaj). In Srejović, D., Lepenski Vir, Nova praistorijska kultura u Podunavlju. Belgrade: Srpska književna zadruga, 224–228. — 1972: The vertebrate fauna. In Srejović, D., Europe’s First Monumental Sculpture: New Discoveries at Lepenski Vir. London: Thames & Hudson. — 1974: History of Domestic Mammals in Central and Eastern Europe. Budapest: Akadémiai Kiadó. — 1978: The vertebrate fauna of Vlasac. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Science and Arts, 35–65. Bonsall, C., Lennon, R., MCSweeney, K., Stewart, C., Harkness, D., Boroneanţ, V., Bartosiewicz, L., Payton, R. & Chapman, J. 1997: Mesolithic and Early Neolithic in the Iron Gates: a palaeodietary perspective. Journal of European Archaeology 5: 50–92. Boroneanţ, V. 1969. Découverte d’objets d’art épipaleolithique dans la zone des Portes de Fer du Danube. Rivista de Science Prehistoriche 24: 283–298. Boroneanţ, V., Bonsall, C., MCSweeney, K., Payton, R. & Macklin, M. 1999: A Mesolithic burial area at Schela Cladovei, Romania. In Thévenin, A. (ed.) L’Europe des Derniers Chasseurs: Épipaléolithique et Mésolithique. (Actes du 5e colloque international UISPP, commission XII, Grenoble, 18–23 septembre 1995). Paris: Éditions du Comité des Travaux Historiques et Scientifiques, 385–390. Brinch Petersen, E. & Meiklejohn, C. Forthcoming. Historical context of the term ‘complexity’ in the Scandinavian Mesolithic. In Janik, L., Kaner, S., Matsui, A. & Rowley-Conwy, P. (eds) From the Jomon to Star Carr. Oxford: BAR Publishing. Buikstra, J.E. & Ubelaker, D.H. (eds) 1994: Standards for Data Collection from Human Skeletal Remains. Fayetteville: Arkansas Archaeological Survey Research Series 44. Cauwe, N. 1998: Sépultures collectives du mésolithique au néolithique. In Guilaine, J. (ed.) Sépultures d’Occident et génèse des mégalithismes (9000–3500 avant notre ère). Paris: Editions Errance, 11–25. Cavalli-Sforza, L.L. 1996: The spread of agriculture and nomadic pastoralism: insights from genetics, linguistics and archaeology. In Harris, D.R. (ed.) The Origins and Spread of Agriculture and Pastoralism in Eurasia. London: UCL Press, 51–69. Chapman, J. 1993: Social power in the Iron Gates Mesolithic. In Chapman, J. & Dolukhanov, P. (eds) Cultural Transformations and Interactions in Eastern Europe. Aldershot: Avebury, 71–121. Chapman, R. 1981: The emergence of formal disposal areas and the ‘problem’ of megalithic tombs in prehistoric Europe. In Chapman, R., Kinnes, I. & Randsborg, K. (eds) The Archaeology of Death. Cambridge: Cambridge University Press, 71–81. Christensen, A.F. 1997: Cranial non-metric variation in north and central Mexico. Anthropologischer Anzeiger 55:15–32. Clark, J.G.D. 1980: Mesolithic Prelude. The Paleolithic–Neolithic Transition in Old World Prehistory. Edinburgh: Edinburgh University Press. Clason, A.T. 1980: Padina and Starčevo: game, fish and cattle. Palaeohistoria 22:141–173. Crubézy, E. 1991: Caractères Discrets et Evolution. Example d’une Population Nubienne: Missiminia (Soudan). Thèse de Sciences. Université de Bordeaux I. Cullen, T. 1995: Mesolithic mortuary ritual at Franchthi Cave, Greece. Antiquity 69: 270–289. Czarnetzki, A. 1972a: Epigenetische Skelettmerkmale im 1

74

Populationsverleich. I. Rechts-links-Unterschiede bilateral angelegter Merkmale. Zeitschrift für Morphologie und Anthropologie 63: 238–254. — 1972b: Epigenetische Skelettmerkmale im Populationsverleich. II. Frequenzunterschiede zwischen den Geschlechtern. Zeitschrift für Morphologie und Anthropologie 63: 341–350. — 1972c: Epigenetische Skelettmerkmale im Populationsverleich. III. Zur frage der Korrelation zwischen den Große des epigenetischen Abstandes und dem Grad der Allopatrie. Zeitschrift für Morphologie und Anthropologie 64: 145–158. Duday, H. & Courtaud, P. 1998: La nécropole mésolithique de La Vergne (Charente-Maritime). In Guilaine, J. (ed.) Sépultures d’Occident et génèse des mégalithismes (9000–3500 avant notre ère). Paris: Editions Errance, 25–39. Falconer, D.S. 1965: The inheritance of liability to certain disease, estimated from the incidence among relatives. Annals of Human Genetics 29: 51–76. Gamble, C. 1986: The Paleolithic Settlement of Europe. Cambridge: Cambridge University Press. Greenacre, M. J. 1984. Theory and Application of Correspondence Analysis. London: Academic Press. Greenfield, H.J. 1984: Analysis of the vertebrate fauna from Hajdučka Vodenica: an Early Neolithic site in the Iron Gates, Yugoslavia. Unpublished manuscript submitted to Borislav Jovanović, Archaeological Institute, Belgrade. Harris, D.R. 1996. Introduction: themes and concepts in the study of early agriculture. In Harris, D.R. (ed.) The Origins and Spread of Agriculture and Pastoralism in Eurasia. London: UCL Press, 1–9. Hauser, G. & De Stefano, G.F. 1989: Epigenetic Variants of the Human Skull. Stuttgart: Schweizerbart. Imamura, K. 1996. Prehistoric Japan. New Perspectives on Insular East Asia. London: UCL Press. Jacobs, K., Wyman, J. & Meiklejohn, C. 1996. Pitfalls in the search for ethnic origins: a cautionary tale regarding the construction of ‘anthropological types’ in pre-Indo-European northeast Europe. In Jones-Bley, K. & Huld, M.E. (eds) The Indo-Europeanization of Northern Europe. Journal of Indo-European Studies 17. Washington D.C.: Institute for the Study of Man, 285–305. Jackes, M. & Gao, Q. Forthcoming. Jiangzhai and BanPo (Shaanxi, PRC): new ideas from old bones. In Janik, L., Kaner, S., Matsui, A. & Rowley-Conwy, P. (eds) From the Jomon to Star Carr. Oxford: BAR Publishing. Jackes, M., Lubell, D. & Meiklejohn, C. 1997: Healthy but mortal: human biology and the first farmers of Western Europe. Antiquity 71: 639–658 (supplementary material at http://intarch.ac.uk/antiquity). Jovanović, B. 1972: The autochthonous and the migrational components of the Early Neolithic in the Iron Gates. Balcanica 3: 49–58. — 1974: Praistorija Gornjeg Đerdapa. Starinar n.s. 22: 1–22. — 1984a: Hajdučka vodenica, praistorijska nekropola. Starinar n.s. 23–24:305–313. — 1984b: Padina, naselje mezolita i starijeg neolita. Starinar n.s. 23–24: 159–167. — 1987: Die Architektur und Keramik der Siedlung Padina B am Eisener Tor, Jugoslawien. Germania 65: 1–16. Kent, S. 1989: Cross-cultural perceptions of farmers as hunters and the value of meat. In Kent, S. (ed.) Farmers as Hunters: The Implications of Sedentism. Cambridge: Cambridge University Press, 2–17. Kozłowski, J.K. & Kozłowski, S.K. 1982: Lithic industries from the multi-layer Mesolithic site Vlasac in Yugoslavia. In Kozłowski, J.K. (ed.) Origin of the Chipped Stone Industries of the Early Farming Cultures in Balkans. Prace archeologiczne 33. Warszawa–Kraków: Państwowe Wydawnictwo Naukowe,

Mirjana Roksandic: The Mesolithic–Neolithic transition in the Ðerdap 11–109. Kozłowski, J.K. & Kozłowski, S.K. 1986: Foragers of Central Europe and their acculturation. In Zvelebil, M. (ed.) Hunters in Transition: Mesolithic Societies of Temperate Eurasia and their Transition to Agriculture. Cambridge: Cambridge University Press, 95–108. Krenzer, U. 1996: Anthropologischer Beitrag zur mesolitisch–neolithischen Transition. Homo 47: 223–247. Lazarovici, G. 1979: Neoliticul Banatului. Cluj-Napoca: Muzeul de Istorie al Transilvaniei. Lee, R.B. 1972a: The intensification of social life among the !Kung Bushmen. In Spooner, B. (ed.) Population Growth: Anthropological Implications. Cambridge, MA: M.I.T. Press, 343–350. — 1972b: Population growth and the beginnings of sedentary life among the !Kung Bushmen. In Spooner, B. (ed.) Population Growth: Anthropological Implications. Cambridge, MA: M.I.T. Press, 329–343. Leroi-Gourhan, A. (ed.) 1965: La Préhistoire. Paris: Press Universitaire Française. Letica, Z. 1971: Vlasac — nouvel habitat de la culture Lepenski Vir à Djerdap. Archaeologia Iugoslavica 10: 7–12. Lyman, R.L. 1994: Vertebrate Taphonomy. Cambridge: Cambridge University Press. Masset, C. 1993: Les dolmens. Sociétés néolithiques et pratiques funéraires. Paris: Editions Errance. Menk, R. & Nemeskéri, J. 1989: The transition from Mesolithic to Early Neolithic in Southeastern and Eastern Europe: an anthropological outline. In Hershkovitz, I. (ed.) People and Culture in Change. Proceedings of the Second Symposium on Upper Palaeolithic, Mesolithic and Neolithic Populations of Europe and the Mediterranean Basin. BAR International Series 508. Oxford: British Archaeological Reports, 531–539. Mikić, Ž. 1980: Anthropologische Typen der Đerdap (Eisernen-Tor) — Serie. In Kozłowski, J.K. & Machnik, J. (eds) Problèmes de la néolithisation dans certaines régions de l’Europe: actes du colloque international. Prace Komisiji Archeologicznej 21. Kraków: Polska Akademia Nauk, 151–162. — 1981a: Đerdapska serija kao antropološki model neolitizacije. Glasnik Etnografskog instituta SANU 30: 103–130. — 1981b: Stanje i problemi fizičke antropologije u Jugoslaviji — Praistorijski period. Posebna izdanja 53. Sarajevo: Centar za balkanološka ispitivanja, knjiga 9. — 1988: Die Anthropologischen Funde vom Eiserner Tor als Neolithisationsmodell. Berytus: Archaeological Studies 26: 45–52. — 1992: The Mesolithic population of the Iron Gates region. Balcanica 23: 33–45. Molto, J.E. 1983: Biological Relationships of Southern Ontario Woodland Peoples: The Evidence of Discontinuous Cranial Morphology. Ottawa: National Museums of Canada. Nemeskéri, J. 1969: Stanovnistvo Lepenskog Vira (Prethodni izvestaj). In Srejović, D. Lepenski Vir. Nova praistorijska kultura u Podunavlju. Belgrade: Srpska Književna Zadruga, 239–262. — 1978: Demographic structure of the Vlasac Epipaleolithic population. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Sciences and Arts, 97–134. Nemeskéri, J. & Lengyel, I. 1978: The results of paleopathological examinations. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Sciences and Arts, 231–260. Nemeskéri, J. & Szathmary, L. 1978a: Analysis of the variation of quantitative traits. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates. Vol. 2, Geology–Biology–Anthropology. Belgrade: Serbian Academy of

Sciences and Arts, 157–176. Nemeskéri, J. & Szathmary, L. 1978b: Anthroposcopic and epigenetic variation. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Sciences and Arts, 135–156. Nemeskéri, J. & Szathmary, L. 1978c: Individual data on the Vlasac anthropological series. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Sciences and Arts, 285–426. Nemeskéri, J. & Szathmary, L. 1978d: Sex and sexualisation. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Sciences and Arts, 77–96. Nemeskéri, J. & Szathmary, L. 1978e: Taxonomical structure of the Vlasac Mesolithic population. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Sciences and Arts, 177–229. Orliac, M. 1988: Mésolithique. In Leroi-Gourhan, A. (ed.) Dictionnaire de la Préhistoire. Paris: Presses Univérsitaires de France, 686. Ossenberg, N.S. 1981: An argument for the use of total side frequencies of bilateral nonmetric skeletal traits in population distance analysis: the regression of symmetry on incidence. American Journal of Physical Anthropology 54: 471–479. Prinz, B. 1987: Mesolithic Adaptations on the Lower Danube: Vlasac and the Iron Gates Gorge. BAR International Series 330. Oxford: B.A.R. Radosavljević-Krunić, S. 1986: Resultats de l’étude anthropologique des squélettes provenant du site Ajmana. Đerdapske sveske 3: 51–58. Radovanović, I. 1995: Mesolithic and Early Neolithic in the Iron Gates region — settlements, subsistence and chronology. In Thévenin, A. (ed.) L’Europe des Derniers Chasseurs: Épipaléolithique et Mésolithique. (Actes du 5e colloque international UISPP, commission XII, Grenoble, 18–23 septembre 1995). Paris: Éditions du Comité des Travaux Historiques et Scientifiques, 379–383. — 1996a: The Iron Gates Mesolithic. Ann Arbor: International Monographs in Prehistory. — 1996b: Kulturni identitet đerdapskog mezolita. Zbornik radova Narodnog muzeja 26(1): 39–47. — 1996c: The Lepenski Vir culture: a contribution to interpretation of its ideological aspects. In Lazić, M. (ed.) Antidoron completis LXV annis Dragoslavo Srejović ab amicis collegis discipulis oblatum. Belgrade: Center for Archaeological Research, Faculty of Philosophy, University of Belgrade, 87–93. — 1996d: Mesolithic/Neolithic contacts: a case of the Iron Gates region. Poročilo o raziskovanju paleolitika, neolitika in eneolitika v Sloveniji 23: 39–48. Radovanović, I. & Voytek, B. 1997: Hunters, fishers or farmers: sedentism, subsistence and social complexity in the Djerdap Mesolithic. Analecta Praehistorica Leidensia 29: 19–31. Roksandic, M. 1999: Transition from Mesolithic to Neolithic in the Iron Gates Gorge: Physical Anthropology Perspective. Unpublished PhD thesis, Simon Fraser University, Burnaby, B.C. — 2002: Position of Skeletal Remains as Key to Understanding Mortuary Behaviour. In Haglund, W.G. & Sorg, M.H. (eds) Advances in Forensic Taphonomy. Boca Raton, FL: CRC Press, 95–113. Saunders, S.R. 1978: The Development and Distribution of Discontinuous Morphological Variation of the Human Infracranial Skeleton. Archaeological Survey of Canada Paper no. 81.Ottawa: National Museums of Canada. Saunders, S.R. 1989: Nonmetric skeletal variation. In Iscan, M.Y. & Kennedy, K.A.R. (eds) Reconstruction of Life from the Skeleton. New York: Liss, 95–108.

1

75

The Iron Gates in Prehistory Schwidetzky, I. & Mikić, Ž. 1988: Lepenski Vir und das Grazilisationsproblem in der Anthropologie. Godišnjak Centra za Balkanološka ispitivanja (ANUBiH) 24: 113–123. Sjøvold, T. 1977: Non-metrical divergence between skeletal populations. Ossa 4 (supplement I): xii, 133. — 1984: A report on the heritability of some cranial measurements and non-metric traits. In van Vark, G.N. & Howells, W.W. (eds) Multivariate Statistical Methods in Physical Anthropology: a Review of Recent Advances and Current Developments. Dordrecht–Boston: Reidel, 223–246. Srejović, D. 1968: Lepenski vir (Boljetin) — predneolitska i neolitska naselja. Arheološki Pregled 10: 85–87. — 1969: Lepenski Vir, Nova praistorijska kultura u Podunavlju. Belgrade: Srpska književna zadruga. — 1971: Predneolitske i neolitske kulture u Đerdapu. Materijali 6 (VIII kongresa arheologa Jugoslavije): 19–25. — 1979: Protoneolit – kultura Lepenskog Vira. In Praistorija Jugoslovenskih, zemalja II: Neolitsko Doba. Sarajevo: Svjetlost, 33–76. — 1989: The Mesolithic of Serbia and Montenegro. In Bonsall, C. (ed.) The Mesolithic in Europe. Edinburgh: John Donald, 481–491. Srejović, D. & Letica, Z. 1978: Vlasac: A Mesolithic Settlement in the Iron Gates, 2 volumes. Belgrade: Serbian Academy of Sciences and Arts. Tasić, N. 1997: Apsolutna hronologija neolita centralnog Balkana. Unpublished PhD thesis, Department of Archaeology, University of Belgrade. — 1998: U potrazi za solju. IQ 3: 12–38. Voytek, B.A. & Trinham, R. 1989: Rethinking the Mesolithic: the case of southeast Europe. In Bonsall, C. (ed.) The Mesolithic in Europe. Edinburgh: John Donald, 492–499.

Werbart, B. 1998: Subneolithic: what is it? — ‘Subneolithic’ societies and the conservative economies of the circum-Baltic region. In Zvelebil, M., Dennell, R. & Dománska, L. (eds) Harvesting the Sea, Farming the Forest: The Emergence of Neolithic Societies in the Baltic Region. Sheffield: Sheffield Academic Press, 37–44. Whittle, A. 1996: Europe in the Neolithic: The Creation of New Worlds. Cambridge: Cambridge University Press. Wilkinson, L., Blank, G. & Gruber, C. 1996: Desktop Data Analysis with SYSTAT. Upper Saddle River, NJ: Prentice Hall. Willis, K.J. & Bennet, K.D. 1994: The Neolithic transition — fact or fiction? Palaeoecological evidence from the Balkans. The Holocene 4: 326–330. Wiltschke-Schrotta, K. 1992. The Early Bronze Age graveyard of Franzhausen I, lower Austria. 3. Study of epigenetic characteristics. Anthropologischer Anzeiger 50(1-2): 27–49. Zvelebil, M. 1996a: The agricultural frontier and the transition to farming in the circum-Baltic region. In Harris, D.R. (ed.) The Origins and Spread of Agriculture and Pastoralism in Eurasia. London: UCL Press, 323–345. — 1996b: What’s in the name: the Mesolithic, the Neolithic and social change at the Mesolithic–Neolithic transition. In Edmonds, M. & Richards, C. (eds) Understanding the Neolithic of North-Western Europe. Glasgow: Cruithne Press, 1–35. Živanović, S. 1975a: Mesolithic population in Đerdap region. Balcanica 6: 1–7. — 1975b: A note on the anthropological characteristics of the Padina population. Zeitschrift für Morphologie und Anthropologie 66: 161–175. — 1976: Ostaci ljudskih skeleta iz praistorijskog nalazišta na Hajdučkoj vodenici. Starinar n.s. 26: 124–129.

76

Demography of the Đerdap Mesolithic–Neolithic transition Mary Jackes, Mirjana Roksandic & Christopher Meiklejohn

Abstract: Models for the Mesolithic–Neolithic transition are critically dependent on demographic variables. A major issue has been whether Mesolithic populations were stationary or increasing in demographic terms. The development of methods allowing for palaeodemographic comparison among archaeological skeletal samples provides the opportunity to test assumptions about the adequacy of samples and about changes in fertility during the European Late Mesolithic and Early Neolithic. Previous work has suggested that the Mesolithic was a period of stationary population in the far west and north of Europe, while the Neolithic brought with it a significant increase in fertility. We apply our standardized palaeodemographic methods to data from Ðerdap skeletal samples to examine whether period assignments need reassessment, and whether the Neolithic brought increased fertility to the Iron Gates region. Our conclusions indicate that continuing reassessment of the assignment of Lepenski Vir skeletons to archaeological periods is needed, that Mesolithic populations were stationary, and that there is a high probability for increased fertility rates in the Neolithic at Lepenski Vir. Key words: Đerdap, demography, Mesolithic, Neolithic, transition, Lepenski Vir

Introduction

A key element of osteological work relates to the accurate determination of demographic variables. Problems involved in the determination of skeletal age, especially in adults, are important here (Jackes 1992, 1994, 2000). Whereas markers of sex in adult skeletons are relatively consistent throughout the life span, age markers are dependent on the process of ageing itself. Since this process is gradual and dependent upon factors such as health and physical condition, the determination is problematic. Individuals who are both healthy and in good physical condition show slower rates of change. As a result, ageing methods tend to provide ‘physiological’ age as opposed to ‘chronological’ results. The exception to this is in subadults, where the nature and rapidity of agedependent changes allows for quite accurate determination.

The shift to food production and the Neolithic is one of the great transitions in human history. In this paper we will examine the evidence provided by skeletons excavated from the Đerdap sites in order to test whether, prior to the introduction of agriculture, the Đerdap Mesolithic population was increasing or was stable and stationary. Demographic factors are central to how we view the transition. Arguments about the social or environmental basis of the shift return repeatedly to the issue of the relative size of Mesolithic and Neolithic populations in Europe. The analyses underlying the ‘demic diffusion’ model are dependent upon the relative population densities of ‘indigenous’ and ‘colonizing’ groups. For Greece and the Balkans other than in the Đerdap much is made of the invisibility of the Mesolithic population. As a result, a key question is whether the apparent peak and concentration of population in the early Holocene Đerdap is real, especially in the Mesolithic. This is a major problem both for archaeology and demography.

Parameters used in this analysis and the relationships among them Because of variability among adults in the expression of age-dependent characteristics, there is debate over the determination of prehistoric demographic variables that require the accurate estimation of skeletal age. There have been proposals that archaeological settings require an approach different from the full adult age profiles used in demographic analysis. Jackes (1986, 1992) has suggested the conjoint use of two values that may provide insight into overall population structure, the juvenile to adult ratio (J:A) and mean childhood mortality (MCM). The J:A (the Indice de Juvenilité, or IJ, introduced by Bocquet & Masset 1977) is the ratio of children 5 to 14 years of age to adults over 20 years, both readily determined from skeletal samples. MCM (Jackes 1986) is the mean mortality for ages 5 to 19: the mean of the life table mortality quotients is a summary value of the probability of death before reaching adulthood among those surviving to age 5. It is calculated from the q values across the three five-year age categories, 5–9, 10–14 and 15–19 years. J:A and MCM can be determined with some accuracy, and seem to bear a consistent relationship to each other unless

General palaeodemographical methodology: the questions that can be answered by demographic data Demography deals with variables such as birth rates, death rates, and their mutual relationship in figures such as growth rates. Demographers studying modern populations have the advantage of a relatively complete data base. Errors are minor and hidden within the statistical power afforded by large samples. Those who study prehistoric populations are less fortunate. Reconstructions must be based on a single set of connected variables, related to death, and extended with an unknowable degree of accuracy to breakdown by sex and age. As a result, studies of prehistoric demography focus on the distribution of ages at death. 77

The Iron Gates in Prehistory

Figure 1 Total fertility (TF) estimates derived from mean childhood mortality (MCM) and juvenile:adult ratio (J:A). Plot of 170 points: reference points 1. Dobe !Kung, 2. Hutterites dying between 1941 and 1950, 3. Coale’s Index of Marital Fertility; grey circles 132 archaeological samples (excludes the Đerdap material); solid triangles three archaeological samples from Dickson Mounds; open triangles 31 historical data sets.

perturbed by special circumstances, such as migration or sample bias. Model demographic data (Coale & Demeny 1983) provide the basis for estimates of population fertility. For each of a large number of pooled sex West model tables, increasing, decreasing and stationary, we calculate J:A, MCM and the total fertility rate (TF). These three variables are very highly correlated and allow us to estimate the TF values for archaeological populations by regression: J:A and MCM, which express the relative numbers of subadults among the dead, provide a reflection of population fertility. While there is a perception that human fertility without contraception would lead to an average woman who survives to menopause having 15 to 20 children, this does not seem to be true. This is reflected in the fertility values we use here, which derive from West model tables 1 to 8, at levels of increase up to 5% for West 1.1 We can demonstrate the validity of the approach by testing the values derived from the Coale and Demeny model data against modern populations that show high and low fertility. Since fertility has biological and behavioural constraints, it is clear that normal unbiased samples must have J:A and MCM values which have their own limits and a biologically determined relationship among the variables can be assumed. If

shown to be valid, the estimates will provide us with a method of examining past population demography. Is the use of West model tables valid? It may be argued that the West model tables are not appropriate bases for conjectures on the fertility of past populations; the fertility estimated from West model tables could be too low because the West tables specifically exclude data with high infant and early childhood mortality, and are derived from 129 sets of data from, in general, industrialized countries, mostly Northwest Europe or countries colonized from the British Isles. The West tables might therefore largely represent a part of the world with low fertility because of factors like late marriage and a high incidence of unmarried people (Hajnal 1982). Are the fertility estimates used here plausible? To test this, we compare the Coale and Demeny model fertility predictions (incorporating also United Nations model data: United Nations 1982) with data for American Hutterite women, considered the best example of ‘natural fertility’ under almost ideal conditions. The total fertility (TF) 2 value for Hutterite married women aged 15 to 49 from 1921 to 1930 was estimated as 12.44 live-born children. This figure is normally giv78

Mary Jackes et al: Demography of the Đerdap Mesolithic–Neolithic transition

en as the value of Coale’s Index of Marital Fertility, by which the Hutterites of this period are taken to represent the highest overall level of childbearing, exceeded only under rare conditions (e.g. Weeks 1996). Figure 1 demonstrates that the samples used for reference in this study have an estimated total fertility below this limit (marked by reference point 3). A TF of 12.44 is markedly above the TF of other historical data sets from Europe, Asia, North and South America covering the period 1650 to 1950. While reference point 3 marks maximum Hutterite fertility in the 1920’s, reference point 2 on Figure 1 marks the estimated level of fertility derived from the 309 Hutterite deaths between1941 to 1950 (age at death reordered from Eaton & Mayer 1953: table 16). The calculated J:A of 0.278 for this decade gives an estimated total fertility of 8.9 liveborn children for women aged 15 to 44; TF would be 7.5 when estimated from MCM. The higher estimated TF value for Hutterites dying between1941 and 1950 may be more correct, because the summed age-specific fertility rates of all Hutterite women of reproductive age between 1936 and 1940 gives a TF rate of 9.4 (see also Jackes & Meiklejohn 2004). A more conservative estimate would be influenced by the 8.1 value for Hutterite TF from 1946 to 1950 (Eaton & Mayer 1953). This was obviously a period of change, one disturbed by World War II, as is probably evidenced by the lack of accord between the J:A and MCM estimates of TF. Nevertheless, it is clear that the estimated values give a sense of Hutterite fertility even for a time of transition. Bocquet-Appel (pers. comm. 2000) has suggested that Hutterite fertility is lower than it might be, because infant mortality is low; when infant mortality is high, fertility is high. The relationship between infant mortality and fertility is, in fact, complex (see, e.g., Montgomery & Cohen 1998). Infant mortality may lead to shorter birth intervals, but this in turn is associated with an increased risk of adverse effects perinatally (Zhu et al. 1999) and maternal mortality (Andersson et al. 2000; Conde-Agudelo & Belizán 2000). If we move to the opposite extreme, low fertility, then excluding such recent contraceptive regimes as twentieth century USA, the Dobe !Kung lie at the bottom of the scale of known fertility levels as shown on Figure 1. The location of the Dobe !Kung (reference point 1) marks the total fertility as estimated from the age at death distribution. The Dobe !Kung fertility is estimated by the J:A fit as 3.9, a reasonable approximation of the Dobe !Kung TF for the 1963–1973 period of 4.3 (Howell 1979) given the small sample size and unstable conditions.

famine, marital separation, behavioural constraints, no sterility, and no age specific variation in fertility. In fact, foetal wastage due to chromosomal abnormalities is high, and thus the chance of a live birth resulting from intercourse is surprisingly low. Holman et al. (2000) show that 50% of a 20-year-old woman’s pregnancies result in foetal loss and that this increases with age. The general proposition is that over 70% of conceptions result in lost pregnancies. As a result, both this foetal loss (detected and undetected) and, of great importance for small archaeological populations, the effects of inbreeding (Dorsten et al. 1999; Ober et al. 1999) must be taken into consideration. While acknowledging that our fertility estimates for past populations could be conservative, we point out that they seem very realistic. Even in a situation encouraging rapid population growth, such as the southern region of North America during the first half of the eighteenth century, total marital fertility was 8.1. Yet here high fertility could be expected, especially because of the possibility of wet nurses reducing the period of lactational amenorrhoea (Houdaille 1995). Fertility in archaeological populations We cannot know details of the fertility of archaeological populations: the type of evidence available for the seventeenth century Huron (Jackes 1994) is surely almost unique. But that evidence does indicate low fertility. Even Colyton, the seventeenth century English parish studied for over 30 years, from written records, and determined to have ‘natural fertility’, is still under discussion (Vann 1999); the evidence suggests that age at marriage determined the period of highest fertility, and that there was family limitation towards the end of a woman’s reproductive period. Archaeological demography will, then, always present us with questions, but translating the J:A and MCM values into fertility estimates using quadratic regression provides a method of deriving a basic demographic parameter from skeletal remains, and acknowledges and circumvents the central problem of our inability to provide correct ages for adult skeletons. We may, in this way, recognize archaeological demographic trends. The database collected by Jackes, together with selected data from the database of Steckel et al. (2002; McCaa 2002), demonstrates that under normal conditions J:A and MCM values for historical and archaeological samples, of reasonable size and without obvious bias, will not fall beyond certain limits. We therefore consider that there are normal biological limits which are reached at slightly under J:A = 0.4 and MCM = 0.14, and that at this point TF must be below 14. Only seven problematical samples of the 142 archaeological samples in our database (which is taken to exclude the Đerdap material for the purpose of this discussion) fall beyond this limit on both axes. When the estimates derived from J:A and MCM are plotted against each other, we have a method of identifying samples which are probably flawed by errors and biases in such a way that an adequate TF estimate cannot be made. Some examples of problematic data appear on Figure 1, which illustrates the fertility estimations derived from archaeological skeletal age distributions. The two extreme outliers are one of the Dickson Mounds samples discussed in

Constraints on fertility Obviously there are behavioural restraints on high fertility, even in non-contracepting societies, and we must add biological factors to any discussion on fertility: higher risks of foetal loss or stillbirth with increasing age and parity; the adverse effects on mother and foetus of short birth intervals; the possibility of longer postpartum amenorrhoea with high parity or age (Larsen & Vaupel 1993). While 35 years of continuous exposure to sexual intercourse could theoretically lead to 26 children per woman, each newborn enjoying 6 months of breast feeding, this model is not plausible, even if there were no disease, death, 79

The Iron Gates in Prehistory

Figure 2 Total fertility (TF) estimates derived from mean childhood mortality (MCM) and juvenile:adult ratio (J:A). General archaeological samples (excluding the Đerdap material) and the three reference points provide the background context for estimates for Portuguese Mesolithic and Neolithic samples, suggesting fertility increase in the Neolithic.

Jackes (1993) and a sample of 170 individuals from Irene Mound (Steckel et al. 2002). The method demonstrates that the Irene Mound sample, representing c. 61% of the original excavated, must be considered inadequate. Thus, while we acknowledge that biases inherent in methods of adult age assessment may weaken the value of palaeodemography as an instrument of interpretation in bioarchaeology, we propose that the use of the J:A and MCM will provide a method of estimating population fertility and allow us to determine whether a sample is unsuitable for analysis (Jackes 1993), because of biases which may derive from incomplete excavation or reporting, taphonomic factors or selective burial.

Demography and the Mesolithic–Neolithic transition

What was not clear, however, was the actual relationship between the two transitions: agricultural and demographic. Deevey assumed the appearance of agriculture to come first, with demographic change being the dependent variable. However, by the late 1960s and 1970s, based on the demonstration that sedentary settlement preceded agriculture in the Near East, several scenarios (e.g. Binford 1968; Smith & Young 1972; Cohen 1977) argued that the prime mover was the demographic transition, and Meiklejohn argued forcibly over many years for a ‘population pressure’ driver to the agricultural transition (Meiklejohn 1978, 1979; Meiklejohn et al. 1984). It now seems that one of the key assumptions behind the population pressure model, the presence of high fertility levels prior to the transition, may be demonstrably false (Jackes 1988; Meiklejohn et al. 1997). This paper will discuss this possibility and the place of the Đerdap samples in the debate.

Previous assumptions It is forty years since Deevey (1960) outlined the concept of demographic transition and applied it to the appearance of agriculture, postulating that the scale of modern overall population size could be traced to the agricultural transition.

Analysis of Portuguese data — the Mesolithic is not a high growth population In the 1980s, Mesolithic skeletal samples excavated in the nineteenth and earlier twentieth centuries from the shell midden sites at Muge in central Portugal were re-examined 80

Mary Jackes et al: Demography of the Đerdap Mesolithic–Neolithic transition

Figure 3 Total fertility (TF) estimates derived from mean childhood mortality (MCM) and juvenile:adult ratio (J:A). General archaeological samples (excluding the Đerdap material) and the three reference points provide the background for estimates for northern European Mesolithic samples (Skateholm in Scania, Olenii Ostrov in Russian Karelia and combined data from sites on the Danish island of Sjælland) indicating a stationary population.

(Jackes 1988; Jackes & Lubell 1999a, 1999b; Lubell et al. 1994; Jackes 1992; Meiklejohn et al. 1997; Jackes et al. 1997a, 1997b). As part of that study, there was an attempt to understand the demographic structure of the samples from the Mesolithic sites Cabeço da Arruda and Moita do Sebastião and the Neolithic site Casa da Moura. While the analysis concentrated on mortality profiles, showing that the Mesolithic samples had lower mortality than the Neolithic sample, the clear implication was that mortality also reflects fertility (as with Figure 2). The Mesolithic samples did not show marked signs of population growth. 3 while the Neolithic sample did. Our work does not support ‘population pressure’ as the prime impetus for the agricultural transition in Portugal, nor do we see evidence that immigration fuelled the Neolithic population growth (see Jackes et al. 2001), leading to the conclusion that fertility increased during and/or after the period of transition.

a decade later, when Meiklejohn was asked to discuss his work on Danish Mesolithic samples (Meiklejohn et al. 1997). The Danish material was excavated, mostly under the Vedbæk Project, on the island of Sjælland. Because the Danish sample was small and therefore had to be pooled, other North European Mesolithic material (primarily Skateholm in Scania and Olenii Ostrov in Russian Karelia) was compared with the Portuguese results. The comparison demonstrated that, despite the problems of mixed samples and the absence of juveniles in both the Danish and Skateholm groups, the results were reasonably consistent with those found in Portugal. None of the Mesolithic samples was in the high mortality, high fertility grouping. Though often treated as a population whose ‘complexity’ prefigured the Neolithic (e.g. Price & Brown 1985; Tilley 1996), the Mesolithic population of southern Scandinavia could be interpreted as “...a stationary population, with both low mortality and fertility” (Meiklejohn et al. 1997: 320). This conclusion still seems valid (Fig. 3), indicating a stationary population in Mesolithic Northern Europe.

Analysis of north European Mesolithic data in accord with Portuguese results The ideas in Jackes’ (1988) paper were not tested until almost 81

The Iron Gates in Prehistory

The Đerdap samples — their importance to the question

While ‘extra individuals’ within any of the burials could be the result of the inclusion in the grave of earth from disturbed burials, these ‘extra bones’ are so common in Đerdap burials that their patterning requires more detailed analysis in order to allow their incorporation into the palaeodemographic reconstruction (Roksandic, in preparation). A further set of problems results from the loose human remains, bones and fragments of bone found in the archaeological deposits without any evidence of burial, a circumstance noted as a common occurrence in the Mesolithic elsewhere (Meiklejohn & Denston 1987). Theoretically such bones could belong to any of the buried individuals; verification of this would have required all individuals in the series to be checked (an impossible task given time constraints), so they were not included in the sample as separate individuals. The situation is especially complex for Lepenski Vir where the quantity of loose bones can sometimes exceed the quantity of bones present in a recognized grave. In order to see whether this situation created a bias, repeated analyses were run, both including and excluding these ‘individuals’ from the Lepenski Vir sample. Sex determination was based on the pelvic bones whenever possible and included standard procedures (Phenice 1969; Workshop of European Anthropologists 1980; Buikstra & Ubelaker 1994). It is noteworthy that the preauricular sulcus was present in almost all of the examined pelves that showed female morphology. As discussed by Roksandic (1999, 2000), the degree of sexual dimorphism is remarkable and the secondary skeletal markers of sex on postcranial bones could be used with great reliability where the pelvic remains were missing. A different pattern is observed with skull remains, which could account for discrepancies between present determinations and those of Nemeskéri (1978), Zoffmann (1983) and Živanović (1975). Age determination presented more problems. In order to avoid point age estimates in adults, since they are highly dependent on the reference population (Bocquet-Appel & Masset 1982) and unreliable in building mortality profiles (Müller & Love 1999), adult ages were assigned to two large categories, namely, ‘young’ and ‘old’. For the present study, all individuals older than 25 were grouped (cf. Jackes 1992). The problem created by individuals represented by a single bone or a bone fragment could be only partially circumvented in this way. Although for most of these partial skeletons it was possible to establish whether they were adults over 25 or not, some had to remain in the undetermined group comprising all from 15 to 80+ years of age. Age determination for children up to 12 years of age was based on observation of tooth formation and eruption and long bone epiphyseal union, when available. Age was assigned by reference to tables in Buikstra & Ubelaker (1994). In other cases, the general aspect of bones was used to establish that the skeleton belonged to any of the subadult groups. The precision with which the age in subadults was assessed depended greatly on preservation and the representation of different body parts, therefore some of the individuals were assigned to quite a wide age range (2–15 years, for example). All these problems had to be circumvented in the statistical analysis. Archaeological considerations are of greatest importance.

Questions arising from the above analyses Confirmation of the pattern of lower Mesolithic fertility and higher Neolithic fertility requires examination of further large Mesolithic samples. Work on material from four Đerdap sites (Roksandic 1999, 2000) provides the opportunity. Meiklejohn et al. (1997) concluded that the data from Nemeskéri’s (1978) study of the Vlasac sample were closely comparable to Olenii Ostrov, suggesting that the Đerdap samples would conform to the pattern. Furthermore, Meiklejohn & Zvelebil (1991), based on previously published data, suggested that the general health of the Đerdap population was similar to that found in southern Scandinavia. Roksandic (1999, 2000) studied four Đerdap samples: Hajdučka Vodenica, Lepenski Vir, Padina and Vlasac. She demonstrates that the data used by Nemeskéri and others are incomplete. We present a new set of demographic profiles here, and also examine whether the revised Đerdap data support our model of low mortality and fertility in Mesolithic populations and higher fertility in Neolithic populations. The basic demographic profiles: methods specific to the Đerdap site analysis The burial practices of Mesolithic peoples are characterized by great variability, and this is certainly true for the Đerdap Mesolithic. The burial practices include cremation, primary inhumation and secondary interment including removal and re-organization of body parts, with re-burial of skulls and fragmentary remains. The Đerdap sites were excavated as a rescue operation, and re-analysis of the excavation records continues. Further study of burial practices planned by Roksandic will include detailed consideration of taphonomic factors. This will no doubt alter the period assignments used here, especially in light of new analyzes already published (Bonsall et al. 2000; Radovanović 2000) and others in progress. In this paper we use the previously accepted archaeological assignment of the burials (Radovanović 1996a; Roksandic 2000). Our method of demographic analysis provides a contribution to the on-going discussion about the validity of the attribution of burials to archaeological units. Within any single burial, the assessment of the MNI followed the common procedures of establishing recognizable osteological elements that were doubled, as well as those that presented incompatibility of age and sex markers. Pairing of bones on the basis of age and general robusticity was accepted only in cases of good preservation and obvious similarities. Since the burials usually comprised one or a few individuals already recognized as separate entities in the field, and since there was some mixing of the smaller elements, all burials were treated as units. For Lepenski Vir, assessment was limited because field documentation was not available, but for Vlasac we could rely on published drawings (Srejović & Letica 1978) and for Padina and Hajdučka Vodenica on unpublished documentation provided by B. Jovanović. Since the MNI in any single burial unit did not exceed eight individuals, it was not necessary to use the procedures appropriate for ossuaries. 82

Mary Jackes et al: Demography of the Đerdap Mesolithic–Neolithic transition Table 1. Lepenski Vir. Age category

Mesolithic

Meso–Neo

M & M–N

Neolithic

M–N & N

All

0

20.6

34.4

55.0

5.3

39.7

60.3

5

2.2

4.2

6.4

6.0

10.2

12.4

10

2.8

2.7

5.5

3.9

6.6

9.4

15

0.0

1.5

1.5

1.5

3.0

3.0

20

1.3

2.5

3.8

1.3

3.8

5.1

25

13.6

45.5

59.1

40.9

86.4

100.0

Total

40.5

90.8

131.3

58.9

149.7

190.2

TF estimated from MCM

7.1

4.2

4.8

5.9

4.9

5.2

TF estimated from J:A

10.8

5.1

6.2

7.6

6.2

6.8

Table 2. Vlasac. Age category

Mesolithic

Meso–Neo

Neolithic

All

0

30.6

0.3

0

30.9

5

9.1

0.3

0

9.4

10

2.3

0.3

0

2.6

15

5.0

0

5.0

20

0.0

3.0

0

3.0

25

81.9

31.1

0

113.0

Total

128.9

35.0

0

163.9

TF estimated from MCM

4.6

4.2

3.8

TF estimated from J:A

4.9

5.1

4.1

Table 3. Padina. Age category

Mesolithic

Meso–Neo

Neolithic

All

0

1

2

0

3

5

1

1

0

2

10

0

1

0

1

15

0

0

0

0

20

0

1

0

1

25

15

26

0

41

Total

17

31

0

48

TF estimated from MCM

3.2

TF estimated from J:A

3.4

Table 4. Hajdučka Vodenica. Age category

Mesolithic

Meso–Neo

Neolithic

All

0

0

0

0

0

5

0

2

0

2

10

0

2

0

2

15

0

2

0

2

20

0

1

0

1

25

0

29

0

29

Total

0

36

0

36

83

The Iron Gates in Prehistory Table 5. Đerdap sites grouped. Age category

Mesolithic

Meso–Neo

Neolithic

All

0

52.3

36.7

5.3

94.2

5

12.3

7.5

6.0

25.8

10

5.1

6.0

3.9

14.9

15

5.0

3.5

1.5

10.0

20

1.3

7.5

1.3

10.1

25

110.5

131.6

40.9

283.0

Total

186.5

192.8

58.9

438.2

TF estimated from MCM

4.6

3.5

5.9

4.2

TF estimated from J:A

5.4

3.9

7.6

4.9

Table 6. Neolithic samples. Age category

Neo LV

Velesnica

Ajmana

All Neo

0

5.3

2

4

11.3

5

6.0

2

3

11.0

10

3.9

3

6.9

15

1.5

2

3.5

20

1.3

1.3

25

40.9

3

5

48.9

Total

58.9

7

17

82.9

TF estimated from MCM

5.9

9.7

TF estimated from J:A

7.6

11.6

Table 7. Site and period groupings. Age category

Vlasac, Padina & Hajdučka Vodenica

All

0

33.9

42.0

5

13.4

13.5

10

5.6

9.9

15

7.0

5.0

20

5.0

8.8

25

183.0

172.5

Total

247.9

251.7

TF estimated from MCM

3.7

3.9

TF estimated from J:A

4.0

4.7

For comparison of Mesolithic and Neolithic mortality and fertility patterns, it was necessary to assign individuals to different periods within the series. The details of period definition and chronological ordering of the skeletons can be found in Roksandic (1999, 2000), which is based on Radovanović (1996a) as well as re-analysis of the documentation from Padina (Jovanović, pers. comm. to MR, 1998). To strengthen the numbers of Neolithic individuals, Velesnica (Roksandic, this volume) and Ajmana (Radosavljević-Krunić 1986) were included in the analysis.

Analysis of the data Because sample sizes are small, the data were reworked so that any unknown age or unknown period individuals were included in the analysis. The unknowns were redistributed proportionately into the known cells.4 It was immediately evident that no simple conclusions were to be derived from the Đerdap samples. The Lepenski Vir and Vlasac samples are summarized in Tables 1 and 2. The unreliability of the Lepenski Vir Mesolithic subsample can be judged from the fact that TF 84

Mary Jackes et al: Demography of the Đerdap Mesolithic–Neolithic transition

Figure 4 Total fertility (TF) estimates derived from mean childhood mortality (MCM) and juvenile:adult ratio (J:A). General archaeological samples (excluding the Đerdap material) and the three reference points provide the background for estimates derived from pooled Đerdap data: Padina (Table 3); Vlasac (Table 2); ALL (Table 5 plus Velesnica, Ajmana from Table 7); Lepenski Vir (Table 1). The distribution shown indicates low Mesolithic fertility and increased fertility in the Neolithic.

estimates range from 7 to 11 – such a broad range is a good indication of a problematic sample. To say that the rich riverine environment of the Lepenski Vir Mesolithic resulted in population increase would push the evidence beyond acceptable limits, because the Vlasac Mesolithic estimate, which is clearly robust, suggests that a Mesolithic woman’s completed family size would be around 4 children (Table 2). The Vlasac sample is interpreted on other evidence as representing a closed and conservative population (Roksandic 2000). Initial arguments for continuity in subsistence practices at Lepenski Vir (Radovanović 1996a, 1996b) are being reconsidered (Bonsall et al. 2000; Radovanović 2000), and the Mesolithic– Neolithic transition at Lepenski Vir is dated to a period of unstable climate (von Grafenstein et al. 1998; Barber et al. 1999). While it might be argued that there is a dramatic drop in fertility during the transition, so that the Lepenski Vir Mesolithic–Neolithic sample came to have a TF of between 4 and 5, it is more likely that the Mesolithic estimate is unacceptable, either because of the small sample size, or because of incorrect period assignment. Thus, the

suggestion of a stable and generally stationary Mesolithic TF of 4 to 5 children seems most reasonable. The sample from Padina (Table 3) is too small to give a reasonable assessment of its demography; we can simply indicate probable low fertility. The paucity of children under age five suggests sample bias, and this is even more obvious for Hajdučka Vodenica (Table 4). While the method of estimating fertility used here is specifically designed to circumvent the frequent problem of infant under-representation in archaeological sites, it is important to note instances in which childhood under-representation, beyond five years of age, may also occur as at Hajdučka Vodenica (Table 4). Given inadequate samples, grouping sites might provide a method to get better demographic estimates by period (Table 5). The Neolithic sample is clearly the weakest, and data from two other sites are added in Table 6. Two further groupings may be made in order to attempt to provide appropriate sample sizes. In Table 7 we examine firstly all sites and all periods pooled, but exclude Lepenski Vir because the Mesolithic data for that site suggest special 85

The Iron Gates in Prehistory

Conclusions

circumstances. We then group all the data for the Mesolithic–Neolithic with those for the Neolithic, in order to see whether high fertility could be considered a characteristic of contact with and development of agricultural practices. Both these groupings of samples again indicate a TF of around 4. Overall, whatever method of pooling is used, the data suggest a rather low fertility, one unlikely to lead beyond a stationary population unless circumstances were exceptional. There is, however, one set of data that is unexpected, the Lepenski Vir Mesolithic. The sample is small and the data give an indication of some bias. A special use of Lepenski Vir for the preferential burial of subadults is one possibility. On the other hand, the anomalous nature of the Mesolithic sample would be diminished by the addition of further adults, as suggested by the stable C and N isotopic values (Bonsall et al. 1997: table 5). One other factor should be considered in Đerdap demography: the transition period has a slight over-representation of adults among the dead. This might occur, not because of low fertility, preservation bias against the young, or selective burial of adults, but as a result of immigration of young adults. The Mesolithic–Neolithic sample could indicate a fall in fertility consequent upon a period of instability associated with cultural change and an influx of adults from outside leading to an apparent over-representation of adults. Such an influx could result in a drop in fertility: the drop could be actual, as a result of the changing and unstable conditions, or perceived, resulting from an unbalanced sex ratio among the migrants (an excess of males). However, it is worth pointing out that several of the Mesolithic–Neolithic sample adults might be considered Mesolithic on stable isotope values for the 33 Mesolithic and Neolithic skeletons from Lepenski Vir first analyzed (Bonsall et al. 1997), and that, until full details are published on the entire sample of 68 Lepenski Vir individuals analyzed for stable isotopes by Bonsall et al. (2000), we cannot speculate on the transition period demography. Nevertheless, a calibration of 14C dates for the Black Sea 5 suggests there may have been a change in Danube aquatic resources just before 8000 years ago. Thus, a drop in fertility might result from altered circumstances. Figure 4 provides a summary restatement of what we can derive from the Đerdap demographic data. Firstly, the Vlasac and Padina data indicate a stationary population for the Mesolithic, just as for the North European Mesolithic. Thus, if the contact period did have a fall in fertility, the result of resource instability, immigration or climate change, there would have been little buffering in the event of crises. Thirdly, while data for Neolithic samples are shown to be inadequate, nevertheless there is an indication of a marked increase in fertility such that population increase would occur. This can be said on the basis of Lepenski Vir data. The presence at Lepenski Vir of Neolithic material has clearly raised the TF above the previous levels. In fact, the Neolithic TF must be 7 as a minimum estimate. Overall, the 462 individuals represented in this analysis (the unlabelled triangle in Figure 4) provide a reasonable sample that will be of great value to palaeodemography once questions of period assignments are clarified.

The evidence is imperfect, because of inadequate sample sizes and the complication of apparent differences among sites, perhaps resulting from the difficulty of differentiating among periods. Our method has allowed us to identify those samples that must be considered as problematic. Despite uncertainties, the evidence suggests a stable and stationary Mesolithic population in the Đerdap. While a possible ‘seeping in’ of immigrants at the Mesolithic–Neolithic contact period, indicated by the non-metric traits in Đerdap samples (Roksandic 1999, 2000, this volume), could be indicated by a drop in fertility and/or in-migration of adults, such low fertility could not be maintained for any length of time. The Neolithic appears to be a time of population growth. In general, then, the pattern of stationary Mesolithic and increasing Neolithic population structures discerned in the far west and north of Europe is repeated and confirmed by the analysis of samples from the Đerdap. Notes 1. West model tables used in the regression analysis comprise 15 decreasing tables with a mean total fertility (TF) of 3.5 (SD = 0.614); 35 increasing tables with a mean TF of 8.9 (SD = 5.967); 11 stationary tables, mean TF 3.9 and SD of 0.782. TF was calculated for 30 years of childbearing from the Cx column of the life tables (the sex ratio for these 30 years is derived from the appropriate model tables and the excess of males over females varies between 0.01 and 0.04). TF estimates are predicted values derived from quadratic model curve estimation regression (SPSS v.12). 2. The total fertility rate is the number of live births a woman might have were she to live to menopause and bear children according to the age-specific fertility rates for the population. Note that live births, not pregnancies, are counted. As Terrisse (1986) points out, this theoretical figure may be shown to overestimate fertility when tested against actual historical data. 3. The demography was based on mandibular dentition MNI estimates from collections in Lisbon and Porto. For discussion on methodology and sites, see Jackes & Meiklejohn (2004) and Jackes & Alvim (2006). The present paper was written before a complete assessment of Arruda history and MNI: new information is provided in Jackes & Meiklejohn (2004). We have also now done the same for Moita, during which process the technique of estimating fertility has been tested and confirmed as reliable, in conjunction with a complete reassessment of the demography (Jackes & Meiklejohn in press). 4. This explains why the figures for age distribution of individuals in Tables 1–2 and 5–7 are not whole numbers. 5. Data from Ballard et al. (2000) calibrated with OxCal 3.5 (Bronk Ramsay 1998) using R = 67±26 (Siani et al. 2000) and the marine curve (Stuiver & Braziunas 1993). While there are serious questions regarding the interpretation of the Black Sea data (e.g. Aksu et al. 2002), the reality of a period of reduced temperature and precipitation over several hundred years around 8200 BP is now well established (see, e.g., Barber et al. 1999; Mayewski et al. 2004).

Acknowledgements MR thanks Dr B. Jovanović, Prof. Ž. Mikić, and the late Prof. D. Srejović for permission to study the Đerdap material and is grateful to the Wenner-Gren Foundation (Grant No. 6250) and Simon Fraser University for providing funding. Research on Portuguese materials was funded by operating grants 410-84-0030 and 410-86-2017 from

86

Mary Jackes et al: Demography of the Đerdap Mesolithic–Neolithic transition the Social Sciences and Humanities Research Council of Canada to D. Lubell, MJ and CM who thank Dr. M.M. Ramalho, Serviços Geológicos de Portugal, for permission to study collections in his care. MJ thanks D. Lubell for his help, Dr Robert McCaa for obtaining permission for the use of the databases collected by Steckel and Rose, and Dr J-P. Bocquet-Appel for helpful comments. CM thanks Dr. D.C. Merrett for a very close reading of the manuscript prior to final submission. Two anonymous reviewers provided useful suggestions.

Hajnal, J. 1982: Two kinds of preindustrial household formation system. Population and Development Review 8: 449–494. Holman, D.J., Wood, J.W. & Campbell, K.L. 2000: Age-dependent decline of female fecundity is caused by early fetal loss. In te Velde, E.R., Broekmans, F. & Pearson, P. (eds) Female Reproductive Ageing. Studies in Profertility Series, vol 9. Carnforth UK: Parthenon Publishing Group, 123–136. Howell, N. 1979: Demography of the Dobe !Kung, Population and Social Structure. New York: Academic Press. Houdaille, J. 1995: Puritans and Virginians. Population: an English Selection 7: 204–210. Jackes, M. 1986: The mortality of Ontario archaeological populations. Canadian Journal of Anthropology 5: 33–48. — 1988: Demographic change at the Mesolithic–Neolithic transition: evidence from Portugal. Rivista di Antropologia (supplement) 66: 141–158. — 1992: Paleodemography: problems and techniques. In Saunders, S.R. & Katzenberg, M.A. (eds) Skeletal Biology of Past Peoples: Research Methods. New York: Wiley-Liss, 189–224. — 1993: On paradox and osteology. Current Anthropology 34: 434–439. — 1994: Birth rates and bones. In Herring, A. & Chan, L. (eds) Strength in Diversity: a Reader in Physical Anthropology. Toronto: Canadian Scholar’s Press, 155–185. — 2000: Building the bases for paleodemographic analysis: adult age determination. In Katzenberg, M.A. & Saunders, S.R. (eds) Biological Anthropology of the Human Skeleton. New York: John Wiley and Sons, 407–456. Jackes, M. & Alvim, P. 2006: Reconstructing Moita do Sebastião, the first step. In Bicho, N. & Veríssimo, N.H. (eds) Do Epipaleolítico ao Calcolítico na Península Ibérica. Faro: Centro de Estudos de Património, 13–25. Jackes, M. & Lubell, D. 1999a: Human skeletal biology and the Mesolithic–Neolithic transition in Portugal. In Thévenin, A. (ed.) L’Europe des derniers chasseurs: Épipaléolithique et Mésolithique. Actes du 5e colloque international UISPP, commission XII, Grenoble, 18–23 septembre 1995. Paris: Éditions du CTHS, 59–64. Jackes, M. & Lubell, D. 1999b: Human biological variability in the Portuguese Mesolithic. Arqueologia 24: 25–42. Jackes, M., Lubell, D. & Meiklejohn, C. 1997a: Healthy but mortal: human biology and the first farmers of western Europe. Antiquity 71: 639–658 (supplementary material at http://intarch.ac.uk/ antiquity). Jackes, M., Lubell, D. & Meiklejohn, C. 1997b: On physical anthropological aspects of the Mesolithic–Neolithic transition in the Iberian Peninsula. Current Anthropology 35: 839–846. Jackes, M. & Meiklejohn, C. 2004: Building a method for the study of the Mesolithic–Neolithic transition in Portugal. Documenta Praehistorica 31: 89–111. Jackes, M. & Meiklejohn, C. In press: The palaeodemography of central Portugal and the Mesolithic–Neolithic transition. In Bouquet-Appel, J.-P. (ed.) Recent Advances in Paleodemography: Data, Techniques, Patterns. New York: Springer Verlag. Jackes, M., Silva, A.M. & Irish, J. 2001: Dental morphology – a valuable contribution to our understanding of prehistory. Journal of Iberian Archaeology 3: 97–119. Larsen, U. & Vaupel, J.W. 1993: Hutterite fecundability by age and parity: strategies for frailty modeling of event histories. Demography 30: 81–102. Lubell, D., Jackes, M., Schwarcz, H., Knyf, M. & Meiklejohn, C. 1994: The Mesolithic–Neolithic transition in Portugal: isotopic and dental evidence of diet. Journal of Archaeological Science 21: 201–216. McCaa, R. 2002: Paleodemography of the Americas from ancient times to colonialism and beyond. In Steckel, R.H. & Rose, J.C.

References Aksu, A.E., Hiscott, R.N., Kaminski, M.A., Mudie, P.J., Gillespie, H., Abrajano, T. & Yaşar, D. 2002: Last Glacial-Holocene paleoceanography of the Black Sea and Marmara Sea: stable isotopic, foraminiferal and coccolith evidence Marine Geology 190: 119–149 Andersson, T., Bergstrom, S., Hogberg, S. & Hogberg, U. 2000: Swedish maternal mortality in the nineteenth century by different definitions: previous stillbirths but not multiparity risk factor for maternal death. Acta Obstetrica Gynecologica Scandinavica 79: 679–686. Ballard, R., Coleman, D. & Rosenberg, G. 2000: Further evidence of abrupt Holocene drowning of the Black Sea shelf. Marine Geology 170: 253–261. Barber, D.C., Dyke, A., Hillaire, M.C., Jennings, A.E., Andrews, J.T., Kerwin, M.W., Bilodeau, G., McNeely, R., Southon, J., Morehead, M.D. & Gagnon, J.M. 1999: Forcing of the cold event of 8,200 years ago by catastrophic drainage of Laurentide lakes. Nature 400: 344–348. Binford, L.R. 1968: Post-Pleistocene adaptations. In Binford, S. & Binford, L.R. (eds) New Perspectives in Archaeology. Chicago: Aldine, 313–341. Bonsall, C., Lennon, R., McSweeney, K., Stewart, C., Harkness, D., Boroneanţ, V., Bartosiewicz, L., Payton, R. & Chapman, J. 1997: Mesolithic and Early Neolithic in the Iron Gates: a palaeodietary perspective. Journal of European Archaeology 5: 50–92. Bonsall, C., Cook, G., Lennon, R., Harkness, D., Scott, M., Bartosiewicz, L. & McSweeney, K. 2000: Stable isotopes, radiocarbon and the Mesolithic–Neolithic transition in the Iron Gates. Documenta Praehistorica 27: 119–132. Bocquet, J-P. & Masset, C. 1977: Estimateurs en paléodémographie. L’Homme 17: 65–90. Bocquet, J-P. & Masset, C. 1982: Farewell to palaeodemography. Journal of Human Evolution 11: 321–333. Bronk Ramsey, C. 1998: Probability and dating. Radiocarbon 40: 461–474. Buikstra, J.E. & Ubelaker, D.H. (eds) 1994: Standards for Data Collection from Human Skeletal Remains. Fayetteville: Arkansas Archaeological Survey Research Series 44. Coale, A.J. & Demeny, P., with B. Vaughan 1983: Regional Model Life Tables and Stable Populations, 2nd edition. New York: Academic Press. Cohen, M.N. 1977: The Food Crisis in Prehistory. New Haven: Yale. Conde-Agudelo, A & Belizán, J.M. 2000: Maternal morbidity and mortality associated with interpregnancy interval: cross sectional study. British Medical Journal 321: 1255–1259. Deevey, E.S. Jr. 1960: The human population. Scientific American 203: 195–204. Dorsten, L.E., Hotchkiss, L. & King, T.M. 1999: The effect of inbreeding on early childhood mortality: twelve generations of an Amish settlement. Demography 36: 263–271. Eaton, J.W. & Mayer, A.J. 1953: The social biology of very high fertility among the Hutterites: the demography of a unique population. Human Biology 25: 206–264.

87

The Iron Gates in Prehistory (eds) The Backbone of History: Health and Nutrition in the Western Hemisphere. New York: Cambridge University Press, 94–124. Mayewski, P.A., Rohling, E. E., Stager, J.C., Karlén, W., Maasch, K.A., Meeker, L. D., Meyerson, E.A., Gasse, F., van Kreveld, S., Holmgren, K., Lee-Thorp, J., Rosqvist, G., Rack, F., Staubwasser, M., Schneider, R.R. & Steig, E. J. 2004: Holocene climate variability. Quaternary Research 62: 243–255. Meiklejohn, C. 1978: Review of M.N. Cohen, ‘The Food Crisis in Prehistory: Overpopulation and the Origins of Agriculture’. Yale University Press [1977]. American Journal of Physical Anthropology 49: 286–287. — 1979: Ecological aspects of population size and growth in LateGlacial and early Postglacial north-western Europe. In Mellars, P.A. (ed.) The Early Postglacial Settlement of Northern Europe: an Ecological Perspective. London: Duckworth, 65–79. Meiklejohn, C. & Denston, B. 1987: The human skeletal material: inventory and initial interpretation. In Mellars, P.A. (ed.) Excavations on Oronsay: Prehistoric Human Ecology on a Small Island. Edinburgh: Edinburgh University Press, 290–300. Meiklejohn, C. & Zvelebil, M. 1991: Health status of European populations at the agricultural transition and the implication for the adoption of farming. In Bush, H. & Zvelebil, M. (eds) Health in Past Societies: Biocultural Interpretations of Human Skeletal Remains in Archaeological Contexts. BAR International Series 567. Oxford: BAR Publishing, 129–145. Meiklejohn, C., Schentag, C.T., Venema, A. & Key, P. 1984: Socioeconomic change and patterns of pathology and variation in the Mesolithic and Neolithic of western Europe: some suggestions. In Cohen, M.N. & Armelagos, G.J. (eds) Paleopathology at the Origins of Agriculture. Orlando: Academic Press, 75–100. Meiklejohn, C., Wyman, J.M., Jacobs, K. & Jackes, M.K. 1997: Issues in the archaeological demography of the agricultural transition in western and northern Europe: a view from the Mesolithic. In Paine, R.R. (ed.) Integrating Archaeological Demography: Multidisciplinary Approaches to Prehistoric Population. Southern Illinois University, Center for Archaeological Investigations, Occasional Paper no. 24. Carbondale: Southern Illinois University, 311–326. Montgomery, M.R & Cohen, B. (eds) 1998: From Death to Birth: Mortality Decline and Reproductive Change. Washington, D.C.: National Academy Press. Müller, H-G. & Love, B. 1999: ‘A semi-parametric model for estimating a lifetime distribution from paleodemographic data’. Paper presented at the Workshop Mathematic Modeling in Paleodemography: Coming to Consensus, Rostock, Germany, 9–12 June 1999. Nemeskéri, J. 1978: Demographic structure of the Vlasac Epipaleolithic population. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Sciences and Arts, 97–134. Ober, C, Hyslop, T. & Hauck, W.W. 1999: Inbreeding effects on fertility in humans: evidence for reproductive compensation. American Journal of Human Genetics 64: 225–231. Phenice, T.W. 1969: A newly developed visual method of sexing the Os pubis. American Journal of Physical Anthropology 30: 297–302. Price, T.D. & Brown, J.A. (eds) 1985: Prehistoric HunterGatherers: the Emergence of Cultural Complexity. Orlando: Academic Press. Radosavljević-Krunić, S. 1986: Resultats de l’étude anthropologique des squélettes provenant du site Ajmana. Ðerdapske sveske 3: 51–58.

88

Radovanović, I. 1996a: The Iron Gates Mesolithic. Ann Arbor: International Monographs in Prehistory. — 1996b: Mesolithic–Neolithic contacts: a case of the Iron Gates region. Poročilo o raziskovanju paleolitika, neolitika i eneolitika v Sloveniji 23: 39–48. — 2000: Houses and burials at Lepenski Vir. European Journal of Archaeology 3: 330–349. Roksandic, M. 1999: Transition from Mesolithic to Neolithic in the Iron Gates Gorge: Physical Anthropology Perspective. Unpublished PhD thesis, Department of Anthropology, Simon Fraser University, Burnaby, B.C. — 2000: Transition from Mesolithic to Neolithic in the Iron Gates gorge: physical anthropology perspective. Documenta Praehistorica 27: 1–100. — In preparation. ‘Extra individuals at Lepenski Vir: burial infill or grave inclusions?’. Siani, G., Paterne, M., Arnold, M., Bard, E., Métivier, B., Tisnerat, N. & Bassinot, F. 2000: Radiocarbon reservoir ages in the Mediterranean Sea and Black Sea. Radiocarbon 42: 271–280. Smith, P.E.L. & Young, T.C. Jr. 1972: The evolution of agriculture and early culture in greater Mesopotamia: a trial model. In Spooner, B. (ed.) Population Growth: Anthropological Implications. Cambridge, MA: M.I.T. Press, 1–59. Srejović, D. & Letica, Z. 1978: Vlasac: A Mesolithic Settlement in the Iron Gates, 2 volumes. Belgrade: Serbian Academy of Sciences and Arts. Steckel, R.H, Sciulli, P.W. & Rose, J.C. 2002: A health index from skeletal remains. In Steckel, R.H. & Rose, J.C. (eds) The Backbone of History: Health and Nutrition in the Western Hemisphere. New York: Cambridge University Press, 61–93. Stuiver, M. & Braziunas, T. 1993: Modelling atmospheric 14C influences and 14 C ages of marine samples to 10,000 BC. Radiocarbon 35: 137–189. Terrisse, M. 1986: Reconstitution des familles en Scandinavie. Annales de démographie historique: 325–352. Tilley, C.Y. 1996: An Ethnography of the Neolithic. Cambridge: Cambridge University Press. United Nations 1982: Model Life Tables for Developing Countries. Department of International Economic and Social Affairs, Population Studies no. 77. New York: United Nations. Vann, R.T. 1999: Unnatural infertility, or, whatever happened in Colyton? Some reflections on ‘English population history from family reconstitution 1580–1837’. Continuity and Change 14: 91–104. Von Grafenstein, U., Erlenkeuser, H., Müller, J., Jouzel, J. & Johnsen, S. 1998: The cold event 8200 years ago documented in oxygen isotope records of precipitation in Europe and Greenland. Climate Dynamics 14: 73–81. Weeks, J.R. 1996: Population: an Introduction to Concepts and Issues. 6th edition. Belmont (California): Wadsworth Publishing Company. Workshop of European Anthropologists 1980: Recommendations for age and sex diagnosis of skeletons. Journal of Human Evolution 9: 517–549. Zoffmann, Zs. 1983: Prehistoric skeletal remains from Lepenski Vir (Iron Gates, Yugoslavia). Homo 34: 129–148. Zhu, B.P., Rolfs, R.T., Nangle, B.E. & Horan, J.M. 1999: Effect of the interval between pregnancies on perinatal outcomes. New England Journal of Medicine 340: 589–594. Živanović, S. 1975: A note on the anthropological characteristics of the Padina population. Zeitschrift für Morphologie und Anthropologie 66: 161–175.

Approaches to Starčevo culture chronology Joni L. Manson

Abstract: The Early Neolithic period in the central and northern Balkans is represented by the cultures of the Starčevo-Körös-Criş complex. Pottery is considered one of the hallmarks of this culture complex, and the dating of Early Neolithic sites in the region has typically relied on typological sequences based on analyses of pottery decoration. Radiocarbon dating got off to a slow start in the Balkans partly because of the long-held emphasis on seriation and partly because of the poor preservation of organic material at many sites. A third dating technique, based on archaeomagnetic intensity analyses, holds great promise for dating these pottery-rich sites. The results of archaeomagnetic intensity analyses on potsherds from twelve Starčevo sites are described. By combining the typological sequences, radiocarbon dates, and archaeomagnetic intensity data it is possible to propose a more refined chronology of the Starčevo culture. Key words: Starčevo culture, chronology, Neolithic, pottery, radiocarbon, archaeomagnetic intensity

Introduction

nomic and ecological analyses, this has meant that the interpretation of many sites as ‘Neolithic’ has depended almost entirely on the presence of a single artefact type — pottery.

The Starčevo culture was named for the archaeological site Starčevo-Grad, located just west of the small village of Starčevo, some 20 km east-northeast of Beograd, Serbia. Starčevo-Grad lies on the northern (left) bank of the Danube, practically opposite the later Neolithic site, Vinča. M. Grbić conducted brief test excavations at Starčevo-Grad in 1928. A joint Yugoslav–American team carried out larger-scale excavations in 1931 and 1932. The concept of an early Neolithic ‘Starčevo culture’ was defined as a result of this work (Fewkes et al. 1933; Ehrich 1977; Benac et al. 1979a). Starčevo sites are concentrated in the central Balkan and southern Pannonian region, particularly Serbia, including Vojvodina and Kosovo, as well as portions of Croatia, Bosnia-Hercegovina, and Macedonia (Fig. 1). The Starčevo culture shares many similarities with roughly contemporaneous neighbouring cultures, especially the AnzabegovoVršnik culture (Macedonia), the Karanovo culture (southern Bulgaria), the Kremikovci culture (western Bulgaria), the Criş culture (Romania), and the Körös culture (Hungary). For detailed information on these culture groups, see Tringham (1971), Girić (1974), Brukner (1979), Dimitrijević (1979), Garašanin (1979), Radišić (1984), Manson (1990), Whittle (1996), and Bailey (2000). Although the territory of many of these groups roughly corresponds to modern political boundaries, it is possible that they may also reflect slight variations in adapting to the many diverse microenvironments of the Balkan peninsula (Kaiser 1984). Typically, these cultures are considered part of the Early Neolithic period in this area (Fig. 2). Many Iron Gates sites (e.g. Lepenski Vir, Schela Cladovei, Padina, and Hajdučka Vodenica) have Early Neolithic Starčevo and Starčevo-Criş occupations. The term ‘Neolithic’ (as it is used in the central Balkans) implies an agricultural and stockraising economy as well as the production of polished stone tools and pottery (e.g. Benac et al. 1979a). However, organic materials tend to be poorly preserved at many Starčevo sites. Combined with a slow acceptance of the practices required for the systematic retrieval of the floral and faunal remains necessary for detailed eco-

Starčevo pottery and typological sequences Potsherds are by far the most commonly recovered artefacts from Starčevo sites, accounting for almost 95% of the total assemblage from some sites (Dimitrijević 1979). The pottery is generally described in terms of its fabric, form, and decoration. Strong similarities exist in the pottery found at Starčevo, Körös, Anzabegovo-Vršnik, and Veluško-Porodin. Starčevo ceramics include coarse, medium, and fine fabrics, or wares. Paste texture and surface decoration are the major factors involved in the designation of wares. Coarse ware is predominant at all Starčevo sites. In coarse ware ceramics, the clay often appears to have been poorly cleaned. At Starčevo-Grad, the native loess underlying the site was used for pottery production (Fewkes et al. 1933; see also Horton 1938). The clay itself was relatively fine-textured, but the pottery also contained chaff, mica, sand, and occasionally small pebbles. It is assumed that the chaff would have been purposely added as a tempering material; some grain impressions on pottery have been identified as cultivated cereals. The sand, mica, and pebbles may have been natural accessories in the raw clay or they may have been intentionally added. Coarse ware vessels were often large and thickwalled. This pottery usually appears to be low-fired, with a dark central core. Exteriors may be light brown to red in colour; interior surfaces are usually darker and may be finely polished, possibly to reduce vessel wall permeability (Garašanin 1984; Manson 1990). Medium ware closely resembles the coarser ware in clay texture and tempering material. Vessel walls may be thinner, but they have the same dark core. Both surfaces are usually light brown, although hints of red or yellow may be seen. Fine ware contains little or no organic tempering, but fine sand or mica is usually present. The vessel walls are thin and usually even-coloured throughout. Surface colours cover a 89

The Iron Gates in Prehistory

Figure 1. Geopolitical map of the central and northern Balkans.

wide range from yellow to buff to red to brown or grey. Starčevo pottery occurs in a limited range of forms. The basic shape is a spherical or hemispherical pot, regardless of fabric or size. Vessels with flaring walls, and shouldered pots with cylindrical necks of various heights, are found in smaller numbers. Straight or slightly thickened bases appear on vessels of every fabric but are more common on coarse ware pots. Low ring bases are also associated with all fabric types, but higher ring bases and true pedestal bases are more common on fine ware vessels. Some larger vessels had handles, usually in the form of protuberances (some with holes to facilitate hanging or carrying), but band or strap handles are also found. Most Starčevo ceramic analyses have directed their attention to pottery decoration. In very broad terms, coarse ware is associated with barbotine surfaces, medium ware with impressed or incised decoration, and fine ware with painted decoration. The term ‘barbotine’ is credited by Fewkes (1938) to M. Vasić, who first used the term to describe the coarse pottery found in the lowest levels at Vinča. Fewkes considered coarse barbotine ceramics to be the ‘chief domestic ware’ of the Starčevo culture. Two major variants of barbotine decoration have been defined. In the first, small clumps of clay appear to have been spattered in an irregular or regular manner on the exterior surface of a vessel. In the second, a rough coating of clay was applied and then streaked over with sticks or fingers, giving the appearance of a barbotine coating that was arranged or organized into strips or ridges. In cross-section, the barbotine

coating may appear to be a different colour than the vessel wall, but the coating consists of the same material as the wall. It has the same fabric texture and contains the same inclusions (Horton 1938). Although barbotine is generally considered a decorative technique, the roughened surfaces would also have made the vessel less slippery to handle (Kutzián 1947) and may have increased the vessel’s resistance to thermal shock by increasing the exterior surface area (Kaiser 1984). Incised and impressed ornamentation appears on both coarse and (more commonly) on medium ware pots. Incising was less common, and generally took the form of a net-like pattern across the body of the vessel. Impressed designs were most often executed by means of fingertips or fingernails, either in definite patterns or unevenly spaced. Finger and fingernail impressions also appear on appliqué bands or rosettes. Occasionally impressions were made with an instrument (such as a reed) which left circular designs in the clay. A number of medium ware pots were simply smoothed, or sometimes burnished, and left undecorated. Although painted pottery is considered a diagnostic trait of the Early Neolithic Balkan cultures, painted ceramics make up a very small fraction of the total ceramic inventory at most of these sites. Pots were covered with a wash (interior and exterior), painted, burnished, and then fired (Fewkes et al. 1933). Three types of painting occur: white painting on a reddish background, dark painting on a reddish background, and polychrome (usually white outlining dark) painting on a reddish background. The dark paint colours were probably 90

Joni Manson: Approaches to Starčevo culture chronology

Figure 2. Neolithic cultures of the Balkans.

derived from iron oxides, while the white paint is thought to have an organic origin (Dimitrijević 1979). Painting is associated with fine wares, although some fine ceramics were burnished and left unpainted (‘monochrome’ ceramics). Common motifs of white painted ceramics include garlands, zigzags, and crosshatch designs near the rim. Groups of vertical lines, joined on the lower portion of the pot’s body, are found, as well as some curvilinear designs and (rarely) the spiral motif. Dark painted motifs include the above, plus triangles, groups of vertical bands filled with a crosshatch design, several types of spirals (including ones that end in ‘claws’), and meanders. Sometimes a rim will have a dark band painted around the top or slightly inside. Polychrome painting is relatively rare. In most cases, dark designs are outlined with white. Most motifs are curvilinear (including spirals) but some linear designs also occur in polychrome. The earliest published chronological sequence of pottery from Starčevo sites was devised by Vladimir Milojčić (1950). By comparing pottery from various sites, he established a four-part ceramic sequence and relative chronology, Starčevo I–IV (Fig. 3). The next serious work on a Starčevo seriation was carried out by Draga Aranđelović-Garašanin (1954). Her sequence closely followed that of Milojčić but recognized closer similarities between the middle two periods. Thus, the stages were labelled Starčevo I, IIa, IIb, and III. Aranđelović-Garašanin based much of her analysis on about 50,000 sherds recovered during the early excavations at Starčevo-Grad, particularly material from pit 5A, which she considered a closed, stratified feature.

According to Aranđelović-Garašanin’s (1954) sequence, painted ceramics are absent from Starčevo I; coarse wares, particularly barbotine ceramics, predominate but fine burnished monochrome ceramics are also fairly common. Incised, impressed, and appliquéd pottery is present but not common. Garašanin (1979) includes materials from Lepenski Vir III, Donja Branjevina, Gura Baciului, Cîrcea, Grivac, and Divostin I in the Starčevo I period (Fig. 4). Starčevo II is considered the beginning of ‘classical’ Starčevo. In Starčevo IIa, coarse barbotine ceramics predominate, but fine ware makes up a slightly higher percentage of the total ceramic assemblage. The proportion of appliquéd, impressed, and incised sherds also increases slightly. White and dark painted linear and curvilinear designs (including spirals) make an appearance. In Starčevo IIb, white painted designs disappear, except for an occasional polychrome piece, and the frequency of all styles of painted ceramics declined. During Starčevo III, the percentage of fine ware increased again. Coarse barbotine is still predominant, but it was now more often applied in a regular, organized pattern. There is a somewhat higher frequency of appliquéd, impressed, and incised ceramics compared to the preceding period. The same dark painted designs continued, with spirals (especially in polychrome) becoming more common. The frequency of biconical forms, and high pedestals also increased. Aranđelović-Garašanin based the Starčevo III period on finds from other sites because material from this period was not collected at Starčevo-Grad until the 1969–70 excavations. 91

The Iron Gates in Prehistory

Figure 3. Typological sequences of Starčevo ceramics.

Criticism of the Aranđelović-Garašanin (1954) typological sequence has centred on the reliability of the stratigraphy of pit 5A at Starčevo-Grad and questions regarding the use of painted pottery as a major distinguishing trait of the sequence. Typically, painted sherds make up less than 2% of the total ceramic assemblage from Starčevo sites. See Korošec (1973) and Ehrich (1977) for discussions on the stratigraphy of Starčevo-Grad and a statistical re-examination of Aranđelović-Garašanin’s typology. Srejović (1972) proposed combining Aranđelović-Garašanin’s Starčevo IIa, IIb, and III into a Classical Starčevo period. He saw little evidence of a Starčevo I period, and preferred to call materials from Lepenski Vir III, ‘Proto-Starčevo’ (see also Vetnić 1986). An alternative ceramic typology was offered by Stojan Dimitrijević (1974, 1979). His sequence begins with a twostage ‘Preclassic’ Starčevo, characterized by the absence of barbotine ceramics. The first part of this Preclassic is called the Monochrome Phase, the second is the White Linear (Linear A) Phase. This is followed by a two-stage ‘Classic’ Starčevo, which includes the Dark Linear (Linear B) Phase and the Garlandoid Phase. The major feature of Dimitrijević’s Classic Starčevo is the very large proportion of barbotine pottery. The Dark Linear Phase is characterized by such motifs as vertical bands, zigzags, triangles, and crosshatching executed in dark paint. In the Garlandoid Phase the major motif is a garland, executed in either dark or white paint. Dimitrijević’s Classic Starčevo corresponds to Aranđelović-Garašanin’s Starčevo II period. Dimitrijević’s ‘Late Classic’ Starčevo also consists of two phases: Spiraloid A and Spiraloid B. The spiral motif first appears in Spiraloid A and increases in frequency in Spiraloid B. A ‘Final’ Starčevo period follows the Late Classic. This period appears to be restricted to peripheral sites and the pottery exhibits much more similarity to Körös forms than to Starčevo ones. A majority of the pottery consists of fingerstreaked or spattered barbotine and appliquéd specimens. Dimitrijević (1979) suggested that this Final Starčevo period may have served as the basis of the Linear Banded Pottery

culture (LBK) of Central Europe (cf. Milisauskas & Kruk 1989; Bogucki & Grygiel 1993; Bogucki 1995; Gronenborn 1999). The Dimitrijević typology seems to have more relevance for the northern and western Starčevo sites (in Croatia) than for the majority of the Starčevo sites, and has not been readily accepted by archaeologists working in the region. In spite of the criticisms levelled against it, the ceramic typology of Aranđelović-Garašanin (1954) is now deeply embedded in the literature of this region and is still relied on for ceramic sequencing. The system does appear to be supported at the few Starčevo sites that have vertical stratigraphy, especially Rudnik at Metohija (Garašanin 1979). In fact, this pottery typology often serves as the only source of chronological control for the Starčevo culture because vertically stratified sites and radiocarbon dates are both scarce (e.g. Brukner 1979; Garašanin 1979). The chronological relationship between Körös and Starčevo sites also remains somewhat problematic. In many cases, the internal chronology of the Körös culture was based on a correlation with Starčevo material. Most of this research associated Early Körös with Starčevo II and Late Körös with Starčevo III (Brukner 1968, 1979; Brukner et al. 1974; Radišic 1984). Brukner (1979) added that the Late Körös period may have continued up to the beginning of the VinčaTordoš (B1) period.

Radiocarbon dates Many of the archaeologists specializing in Starčevo studies were slow to accept the radiocarbon dating technique (Benac et al. 1979b; see also Evans & Rasson 1984), and relatively few 14C dates have been published for Starčevo sites although the situation is improving (e.g. Whittle et al. 2002). Throughout this article, radiocarbon dates are given as years BP; calibrated dates are cited as cal BC. As some referenced sources gave radiocarbon dates and some gave calibrated 92

Joni Manson: Approaches to Starčevo culture chronology

Figure 4. Selected Neolithic sites in the Balkans.

6900 BP (c. 6050–5750 cal BC). The Starčevo occupation at Grivac dates to about 7250 BP (c. 6100 cal BC), that at Banja to about 7050 BP (c. 5950 cal BC) (McPherron 1988). Garašanin (1979) believed Divostin I could be correlated with Starčevo I, while Gimbutas (1974) fitted Divostin I into Starčevo IIa. Radiocarbon dates from Padina indicate a Starčevo occupation from about 7100 to 6550 BP (c. 5950–5500 cal BC) (Clason 1980). Similar dates have been obtained from charcoal samples from the houses at Lepenski Vir that contained Starčevo pottery. A series of 14C dates from these houses indicates a Starčevo occupation around 7150 to 6750 BP (c. 6000–5650 cal BC) (Budja 1999; Bonsall et al. 2000). Recently two studies focusing on the origins of the LBK culture of Central Europe looked at 14C dates from Starčevo and Körös sites (Gläser 1991; Lenneis & Stadler 1995). Gläser proposed dates of 5950 to 5470 cal BC for the duration of the Starčevo culture and 5700 to 5260 cal BC for the Körös culture. Lenneis & Stadler’s study dated the Starčevo culture to 6050 to 5450 cal BC, with dates of 5760 to 5280 cal BC for the Körös culture. Research reported by Hertelendi et al. (1995) dated Early Körös sites to 5950–5400 cal BC, and Late Körös sites to 5770–5230 cal BC.

dates, conversions were performed with the CALIB 5 program distributed by the Quaternary Isotope Laboratory of the University of Washington (Stuiver & Reimer 1993; Stuiver et al. 2005) using the IntCal04 curve (Reimer et al. 2004). In Macedonia, the eleven 14C dates for Anzabegovo I indicate a temporal span from about 7250 to 7100 BP (c. 6100–6000 cal BC). Eight dates for Anzabegovo II span the period of about 7050 to 6850 BP (c. 5950–5750 cal BC). And four dates for Anzabegovo III indicate a span from about 6850 to 6500 BP (c. 5750–5500 cal BC) (Gimbutas 1976). Gimbutas correlates Anzabegovo I with the Gura Baciului culture, Anzabegovo II with Starčevo IIa, and Anzabegovo III with Starčevo IIb and III. A radiocarbon date for the earliest level of the stratified site of Veluška Tumba, also in Macedonia, shows that Veluška Tumba I dates to about 6950 BP (c. 5800 cal BC), while Veluška Tumba II was dated to about 6700 BP (5600 cal BC) (Todorović et al. 1977). Garašanin (1979) correlated Veluška Tumba I and II to the Gura Baciului culture, Veluška Tumba III to Anzabegovo II (therefore Starčevo IIa), and Veluška Tumba IV to Anzabegovo III (therefore Starčevo IIb and probably III). A single 14C date for the Starčevo settlement at Gornja Tuzla in Bosnia-Hercegovina indicates an occupation around 6650 BP (c. 5600 cal BC). The site was correlated with Starčevo III (Tasić & Tomić 1969). Bone samples from the later excavations at StarčevoGrad, Serbia, yielded five radiocarbon dates between 6850 and 6650 BP (c. 5750–5600 cal BC) (Ehrich 1977). Seven 14C dates from Divostin I span the period from about 7200 to

Archaeomagnetic intensity dating The plasticity of clay and the nature of pottery-making result in products well suited to stylistic analyses and seriation. However, the physical characteristics of the clay also allow 93

The Iron Gates in Prehistory

us to date pottery by means of archaeomagnetic intensity analysis. Clay often contains magnetic minerals such as magnetite and haematite as impurities. When clay is heated to a temperature higher than the ‘blocking temperature’ of these minerals, the magnetic particles can record the direction and strength of the earth’s magnetic field in that location at that particular time. Blocking temperatures fall just below a magnetic mineral’s Curie point (the temperature above which a substance loses its ferromagnetic properties). The blocking temperatures of magnetite and haematite lie in the range of 500° C to 700° C. The primary remanent magnetization acquired by heating clay above the blocking temperature of its magnetic minerals is called thermoremanent magnetization (TRM). Since the geomagnetic field changes direction and strength over time and space, it is possible to determine when an artefact of baked clay was last subject to temperatures high enough to allow acquisition of TRM by comparing its magnetic parameters with the known geomagnetic record for a particular region. Although the directional data is not particularly useful when dealing with portable objects such as pottery or figurines, these artefacts can still provide important information in the form of intensity data. Most of the techniques now being used to determine magnetic intensities of baked clay artefacts are variations of the method devised by Thellier & Thellier (1959). This method of determining ancient geomagnetic intensity involves the use of double heatings at a series of temperature steps. The Thellier–Thellier technique depends on the principle that magnetic grains which have acquired a remanent magnetization after cooling through the blocking temperature will also lose that remanent magnetization at the same temperature if reheated. A sample of baked clay usually has such a variety of magnetic grains that the blocking temperatures are continuously distributed from the Curie temperature downward (Aitken 1983: 19). Therefore, a series of laboratory reheatings to increasingly higher temperatures up to the Curie point will yield a series of partial TRMs which can be plotted against the natural remanent magnetization (NRM) being removed from the sample. The NRM is generally considered to be the magnetism a sample has before it is subjected to experimental procedures in the laboratory. The progressive decrease of NRM plotted against the progressive increase of TRM should result in a straight line with a slope proportional to the ancient field intensity. This relationship can be expressed as: FA = FL x Δ NRM / Δ TRM, where FA is the ancient magnetic field intensity, FL is the laboratory field intensity, and Δ NRM / Δ TRM is the absolute value of the slope of the line. Pottery, which has been described as pyrophilic in nature, is generally well suited to the multiple heatings of the Thellier–Thellier method (Carmichael 1977). The first step in this procedure is to measure the NRM of the sample. Then the sample is heated to a temperature, T1, and allowed to cool in a known magnetic field, FL, with a known orientation. The remanent magnetism, J(T1, 0), resulting from this step is measured. The sample is rotated 180° about a horizontal axis, reheated to T 1 , cooled in F L again, and the remanent magnetism, J(T1, 180), is measured. The vector sum of J(T1,

0) and J(T1, 180) equals twice the PTRM (partial thermoremanent magnetization) that the sample acquired in the ancient field when it cooled from above its Curie point, Tc , to T1 . The vector difference of J(T1, 0) and J(T1, 180) equals twice the PTRM acquired in field FL between T1 and room temperature (Collinson 1983: 427–428). This double heating process is repeated at progressively higher temperatures up to the Curie point. One curve can then be generated from the sums of the measurements obtained from the initial and reversed positions for each temperature, while a second curve can be derived from the differences between these measurements (Michels 1973: 144). A third curve, the Arai diagram (Nagata et al. 1963), or NRM–TRM diagram, generally depicts NRM lost on a y-axis and TRM gained on an x-axis. The Thellier–Thellier method has features built into it that allow researchers to check the reliability of the results and to discard those which cannot pass these checks. In fact, it is not unusual for the rate of failure to be relatively high — often more than one-third of the samples must be rejected on the basis of a lack of internal consistency (e.g. Sasajima 1965; Boyd 1986; Wei et al. 1987). A number of explanations have been offered to account for deviations from ‘ideal’ behaviour, which often show up in the form of non-linear curves on the Arai plots. Many researchers have suggested guidelines to follow when determining which results can be considered reliable (Table 1). Manson (1990) subjected 101 samples of Starčevo pottery to archaeomagnetic intensity analysis (Table 2). The standard Thellier–Thellier method was used throughout the laboratory procedure, which was carried out at the Paleomagnetism Lab at the University of Pittsburgh, under the direction of Victor Schmidt. The NRM (at 20° C) of each sample was first established using a cryogenic magnetometer from Superconducting Technology, Inc. (SCT). The first four samples (from Divostin, Vinča, and Starčevo) were processed using a spinner magnetometer from Princeton Applied Research. All of the other specimens were processed with the cryogenic, or SQUID (Superconducting Quantum Interference Device), magnetometer. The heating and cooling of the specimens was carried out in a Schonstedt thermal specimen demagnetizer. Of the 101 samples processed, 65 yielded results considered to be acceptable. Samples that lacked at least three points on a straight-line segment of the NRM–TRM plots were immediately rejected. The remaining samples were then checked for consistency on Zijderveld diagrams, which combine intensity and directional changes on the same plot (Fig. 5). Regularity on these diagrams ensures that loss of TRM and gain of NRM are both univectoral, indicating that there are no additional components complicating the picture. Samples with irregular Zijderveld diagrams were also rejected from the study. Of the samples considered acceptable, the number of points on the straight-line segment of the NRM–TRM plot averaged 5 or 6. Ordinarily, the mid-range of temperature points formed the linear part of the plot. Furthermore, the linear portion of the plots covered an average span of 60% of the total extrapolated NRM. The slopes of the lines were calculated using SAS programs. In order to facilitate comparisons with the published master curves derived from other archaeomagnetic intensity 94

Joni Manson: Approaches to Starčevo culture chronology Table 1. Determining the reliability of NRM–TRM diagrams. Source

Minimum # of Temperature Points

Temperature (°C) Range to Use

Dubois & Watanabe (1965)

——

100–400

Sasajima (1965)

——

——

Weaver (1966)

——

150–450

Bucha (1967)

4

——

Points should cover 2/3 of the distance on the line between the Yand X-axes.

Bucha (1967)

3

——

Standard deviation from the line should not exceed ± 10%.

Coe (1967)

6

——

Standard deviation from the line should not exceed ± 5%.

Bucha et al. (1970)

4

——

——

Rusakov & Zagniy (1973)

6

——

——

Barbetti et al. (1977)

——

——

Don’t use high temperature points of concave-up lines.

Levi (1977)

——

——

Linear plot is a necessary but insufficient condition for judging reliability.

4

——

Points must span at least 15% of the total extrapolated NRM.

Kovacheva (1980)

——

mid-range

Tanaka (1982)

——

——

Need at least 3 samples per site.

Arbour & Schwartz (1982)

3

——

——

Kovacheva (1983)

6

mid-range

——

Thomas (1983)

——

——

Walton (1983)

——

mid-range

Aitken (1983)

——

——

Need regular NRM demagnetization as well as linear plots.

Aitken et al. (1983)

——

——

Linear portion should cover 200°C and 60% of the demagnetization range (for Class A); or 150°C and 50% of the range (for Class B+); TRM capacity must not change more than 10% (Class A) or 20% (Class B+) between 200°C and the upper point of the linear portion; unless VRM is present, the 150°C point should be no more than 10% (Class A) or 20% (Class B+) away from an extension of the line; the 550°C point should be no more than 10% (Class A) or 40% (Class B+) away from an extension of the line.

Sakai & Hirooka (1985)

——

——

The more points used, the better the reliability (in their sample, 172 out of 300 specimens gave ‘good’ results).

Burakov & Nachasova (1985)

——

up to and incl. 450°C

Kovacheva & Zagniy (1985)

——

mid-range

Wei et al. (1987)

——

——

Standard deviation should not exceed ± 6%.

4

——

Remanence direction must remain stable; allow 40% of his sample). ——

——

Portion of plots used is highly subjective.

95

Linearity is no guarantee that mineral alteration hasn’t occurred.

—— Use small temperature steps (25°C) to insure having enough points on a line.

The Iron Gates in Prehistory Table 2. Sites (with geographical coordinates), numbers of samples, and modern magnetic intensity (F0 = intensity at time of collection). Site Name

# Samples Tried

# Samples Accepted

Latitude (°N)

Longitude (°E)

F0 (oe.)

Ludaš-Budžak

26

15

46° 10´

19° 50´

0.46910

At

7

2

45° 08´

21° 15´

0.47160

Nosa

4

3

46° 10´

19° 50´

0.46690

Kaonik-Gradina

5

2

43° 30´

21° 30´

0.46730

Pančevo-Nadela

33

22

44° 52´

20° 45´

0.47000

Tečić

7

5

43° 50´

21° 10´

0.46210

Kozluk

2

2

45° 07´

21° 18´

0.46390

Vrtište

6

5

43° 20´

21° 45´

0.46080

Aradac-Leje

7

5

45° 22´

20° 18´

0.46970

Divostin

1

1

44° 01´

20° 52´

0.47470

Vinča

2

2

44° 46´

20° 37´

0.46720

Starčevo

1

1

44° 48´

20° 42´

0.46740

Table 3. Intensity values of Starčevo sites. Sample #

% of Extr. NRM

# of Temp. Points

Temp. (°C) Range Used

FA

FA / F0 ± S.E. (S.E. = Standard Error of Estimate)

Site Average FA / F0 ± S.E.

AT A001 A008

44 67

4 5

100–350 350–500

0.2900 0.3718

0.61 ± 0.00 0.79 ± 0.07

0.70 ± 0.04

LUDAŠ–BUDŽAK B001 B003 B004 B005 B006 B007 B014 B016 B019 B021 B025 B026 B027 B028 B029

41 90 44 50 37 47 38 42 90 67 65 99 77 57 77

3 6 3 8 5 4 4 5 6 6 4 9 9 9 5

200–400 450–575 200–350 300–550 350–500 200–400 200–400 350–500 350–550 200–475 300–450 350–600 100–525 200–550 300–475

0.3443 0.3170 0.2486 0.3751 0.2590 0.2342 0.2973 0.3890 0.2897 0.2296 0.4188 0.3655 0.3844 0.4052 0.3171

0.73 ± 0.05 0.68 ± 0.04 0.53 ± 0.03 0.80 ± 0.04 0.55 ± 0.01 0.50 ± 0.09 0.63 ± 0.08 0.83 ± 0.02 0.62 ± 0.02 0.49 ± 0.02 0.89 ± 0.02 0.78 ± 0.05 0.82 ± 0.04 0.86 ± 0.04 0.68 ± 0.03

KOZLUK K001 K002

53 41

5 5

200–450 300–475

0.4714 0.3885

1.02 ± 0.09 0.82 ± 0.05

0.93 ± 0.07

KAONIK–GRADINA KG001 KG003

42 54

6 5

300–500 200–450

0.3997 0.3973

0.86 ± 0.09 0.85 ± 0.09

0.86 ± 0.09

ARADAC–LEJE L001 L002 L003 L004 L007

69 32 47 66 51

5 3 4 7 7

100–400 200–350 200–400 200–500 200–500

0.4049 0.4430 0.4754 0.5322 0.5397

0.86 ± 0.04 0.94 ± 0.02 1.01 ± 0.04 1.13 ± 0.06 1.15 ± 0.06

96

0.69 ± 0.04

1.02 ± 0.04

Joni Manson: Approaches to Starčevo culture chronology Table 3 (cont). Intensity values of Starčevo sites. Sample #

% of Extr. NRM

# of Temp. Points

Temp. (°C) Range Used

FA

FA / F0 ± S.E. (S.E. = Standard Error of Estimate)

NOSA N001 N003 N004

68 47 67

6 4 5

200–475 200–450 100–400

0.4097 0.3984 0.3889

0.88 ± 0.06 0.85 ± 0.07 0.83 ± 0.08

PANČEVO–NADELA PN001 PN003 PN004 PN005 PN006 PN007 PN014 PN015 PN016 PN018 PN021 PN023 PN024 PN026 PN027 PN029 PN030 PN031 PN033 PN035 PN035b PN035c

32 58 65 65 61 64 64 47 32 92 58 50 48 92 59 68 74 79 65 51 49 57

5 5 7 5 6 8 4 5 3 6 5 5 5 5 6 6 9 9 7 5 5 5

300–475 300–475 200–500 300–475 350–525 300–550 100–350 300–475 300–400 200–475 200–450 200–450 300–475 100–400 300–500 100–450 350–625 200–550 300–575 350–500 350–500 350–500

0.1994 0.3241 0.4512 0.5117 0.3351 0.4723 0.6396 0.5139 0.5380 0.1388 0.3635 0.2943 0.3876 0.4767 0.4318 0.5401 0.4152 0.4189 0.3978 0.4191 0.3449 0.4001

0.42 ± 0.02 0.69 ± 0.03 0.96 ± 0.07 1.09 ± 0.06 0.71 ± 0.02 1.00 ± 0.07 1.36 ± 0.05 1.09 ± 0.04 1.14 ± 0.09 0.30 ± 0.02 0.77 ± 0.06 0.63 ± 0.05 0.82 ± 0.06 1.01 ± 0.08 0.92 ± 0.05 1.15 ± 0.06 0.88 ± 0.01 0.89 ± 0.05 0.85 ± 0.05 0.89 ± 0.08 0.73 ± 0.05 0.85 ± 0.04

TEČIĆ T002 T003 T004 T005 T006

87 27 55 69 66

7 4 4 9 9

200–500 300–450 200–400 200–575 350–600

0.5309 0.4315 0.3777 0.4318 0.3817

1.15 ± 0.06 0.93 ± 0.11 0.82 ± 0.02 0.93 ± 0.07 0.83 ± 0.04

0.93 ± 0.06

VRTIŠTE V002 V004 V005 V006 V007

84 67 76 59 68

5 6 9 8 9

200–450 350–525 350–600 200–525 300–575

0.4380 0.2591 0.2971 0.3807 0.3303

0.95 ± 0.04 0.56 ± 0.02 0.64 ± 0.02 0.83 ± 0.05 0.72 ± 0.04

0.74 ± 0.03

DIVOSTIN YU.4

64

5

200–500

0.265

0.57

0.57

VINČA YU.1 YU.3

63 82

4 6

400–550 200–550

0.285 0.465

0.61 1.00

0.81

STARČEVO YU.2

46

3

300–450

0.500

1.07

1.07

experiments in Southeast Europe (Kovacheva & Veljovich 1977, 1985; Kovacheva 1977, 1980, 1983, 1995), the results were reported as the ratio of ancient field intensity to modern field intensity at the site (FA / F0). ‘Modern’ field strength actually refers to the field intensity at the site at the time the sample was collected. Modern field strengths were determined through the use of the program GEOMAG (U.S. Department of Commerce n.d.). The program uses a site’s latitude, longitude, elevation, and collection date to calculate the appropriate geomagnetic data for a particular location and time (Manson & Schmidt 1991).

Site Average FA / F0 ± S.E.

0.85 ± 0.07

0.87 ± 0.05

Acceptable sample intensity ratios were then averaged to obtain average site intensity values. The relatively short life span and single component quality of most Starčevo sites (Manson 1990) lent support to this decision. Table 3 lists the number of temperature points and the temperature range used, the percentage of total extrapolated NRM included, the experimentally derived field intensity (FA), and the ratio of ancient to modern field intensity (FA / F0) for each sample, as well as the average FA / F0 for each site. Since magnetic field intensity does not vary unidirectionally through time, any intensity value may correspond to 97

The Iron Gates in Prehistory

The future of Starčevo chronological studies The dates proposed in Table 4 for the Starčevo and Körös cultures of the central Balkans cover a slightly broader range than the dates suggested by Gläser (1991), Hertelendi et al. (1995) and Lenneis & Stadler (1995). The latter three studies all derived their chronologies from available radiocarbon dates. It is likely that a combination of dating techniques will offer the greatest potential for creating absolute chronologies and for assessing the validity of the ceramic typologies currently in place (Sternberg 1997; Bonsall et al. 2002). It is also obvious that clear communication between the archaeologists who recover the objects, and the geophysicists who often perform the chronometric analyses on the samples, is essential for achieving reliable results (Sternberg 1990; Kovacheva 1995). The archaeologist is responsible for providing the contextual data about site stratigraphy and stylistic seriations and the geophysicist must be able to explain the basic theory, methodology and limitations of the techniques, as well as the results of the analysis. Finally, the archaeologist must be able to interpret that information and incorporate it into new research designs. Many of the major research questions concerning the Starčevo culture require a well-grounded chronological framework. Questions surrounding the introduction and adoption of domesticated plant and animal species, relationships to neighbouring (contemporaneous?) groups, relationships to earlier and later populations in the same area, changes in technology (including pottery production), settlement strategies, and social complexity can best be addressed when a temporal framework has been established (e.g. Greenfield 1986; Tringham et al. 1992; Bailey et al. 1998; Manson 1990, 1995; Gronenborn 1999; Bonsall et al. 2000). The goal, after all, is not merely to create more refined chronological frameworks, but to use those frameworks to focus attention on other aspects of past cultures. The establishment of temporal frameworks is a vital step in archaeology, but it can no longer be considered the final objective of the discipline. A more-refined chronology for the Starčevo culture should be viewed as an important tool for understanding the life and times of these Neolithic Balkan peoples.

Figure 5. NRM–TRM (Arai) and Zijderveld diagrams.

more than one possible date. The Aranđelović-Garašanin (1954) typology was used to determine a relative temporal association for each site and its most plausible position on the archaeomagnetic intensity master curve. The master curve for the region (see especially Kovacheva & Veljovich 1985; Kovacheva 1995) was based on materials from radiocarbondated sites in Bulgaria and Serbia, so it has absolute dates assigned to portions of the curve. After determining the positions of the intensity values on the curve, it was possible to 1

Table 4. Starčevo and Starčevo-Körös chronology. Phase

14C

age range

Calibrated age range

Starčevo I

At least 7250–7050 BP

c. 6100–5900 cal BC

Starčevo IIa

c. 7050 to 6800 BP

c. 5900–5700 cal BC

Starčevo IIb

c. 6800 to 6450 BP

c. 5700–5400 cal BC

Starčevo III

c. 6450–?6150 BP

c. 5400–5100 cal BC

Acknowledgements I gratefully acknowledge the financial assistance of the Institute for International Education (Fulbright Scholarship for Dissertation Research in Yugoslavia), Southern Illinois University at Carbondale (Doctoral Research Award), and the United States Department of Education (Jacob K. Javits Fellowship). The analyses reported here could not have been completed without the help and cooperation of many people in Beograd, Pančevo, Zrenjanin, Vršac, Priština, Niš, Kikinda, Bitola, Subotica, Zagreb, and Svetozarevo, as well as in the United States. The archaeomagnetic analyses were performed at the Paleomagnetism Laboratory of the Department of Geology and Planetary Science at the University of Pittsburgh, under the direction of Victor Schmidt. Alan McPherron, Department of Anthropology, University of Pittsburgh, arranged for my first field visit to Yugoslavia (Kragujevac). Carroll Riley, Department of Anthropology, Southern Illinois University at Carbondale, cheerfully (for the most part) provided much-appreciated advice at many stages of the project. I also wish to thank my husband, David Snyder, for his encouragement and assistance, especially where computers are concerned.

Starčevo I = Gura Baciului Starčevo IIa + IIb = Classical Starčevo = Early Körös Starčevo III = Late Körös

assign probable absolute dates to each of the Starčevo culture sites in the study. Furthermore, by combining the typological sequences with the archaeomagnetic intensity data and the relatively few 14C dates available from Starčevo culture sites, it is possible to propose a correlation between the Aranđelović-Garašanin ceramic sequence and absolute dates (Table 4). 98

Joni Manson: Approaches to Starčevo culture chronology

References

Carmichael, C.M. 1977: Introduction to ‘Paleomagnetic field intensity, its measurement in theory and practice’. Special issue of Physics of the Earth and Planetary Interiors 13(4): v–vii. Clason, A.T. 1980: Padina and Starčevo: game, fish, and cattle. Palaeohistoria 22: 141–173. Coe, R.S. 1967: The determination of paleo-intensities of the earth’s magnetic field with emphasis on mechanisms which could cause non-ideal behavior in Thellier ’s method. Journal of Geomagnetism and Geoelectricity 19: 157–179. Coe, R.S., Grommé, S. & Mankinen, E.A. 1978: geomagnetic paleointensities from radiocarbon-dated lava flows on Hawaii and the question of the Pacific nondipole low. Journal of Geophysical Research 83: 1740–1756. Collinson, D.W. 1983: Methods in Rock Magnetism and Palaeomagnetism. Techniques and Instrumentation. New York: Chapman & Hall. Dimitrijević, S. 1974: Problem stupnjevanja Starčevačke kulture s posebnim obzirom na doprinos južnopanonskih nalazišta rješavanju ovog problema. In Tasić, N. (ed.) Počeci Ranih Zemljoradničkih Kultura u Vojvodini i Srpskom Podunavlju. Beograd: Srpsko Arheološko Društvo / Subotica: Gradski Muzej. Materijali 10: 59–83. — 1979: Sjeverna zona. Praistorija Jugoslavenskih zemalja II: 229–360. Sarajevo: Akademija Nauka i Umjetnosti Bosne i Hercegovine. Dubois, R.L. & Watanabe, N. 1965: Preliminary results of investigations made to study the use of Indian pottery to determine the paleointensity of the geomagnetic field for the United States 600–1400 AD. Journal of Geomagnetism and Geoelectricity 17: 417–423. Ehrich, R.W. 1977: Starčevo revisited. In Markotić, V. (ed.) Ancient Europe and the Mediterranean. Warminster: Aris & Phillips, 59–67. Evans, R.K. & Rasson, J.A. 1984: Ex Balcanus Lux? Recent developments in Neolithic and Chalcolithic research in southeast Europe. American Antiquity 49: 713–741. Fewkes, V.J. 1938: Neolithic sites in the Yugoslavian portion of the lower Danubian valley. Proceedings of the American Philosophical Society 78: 329–402. Fewkes, V., Goldman, H. & Ehrich, R. 1933: Excavations at Starčevo, Yugoslavia, seasons 1931 and 1932. Bulletin of the American School of Prehistoric Research 9: 33–54. Garašanin, D. 1984: Naselje Starčevačke Kulture. In Vinča u Praistoriji i Srednjem Veku. Beograd: Galerija Srpska Akademije Nauka i Umjetnosti, 13–21. Garašanin, M. 1979: Centralnobalkanska zona. Praistorija Jugoslavenskih zemalja II: 79–212. Sarajevo: Akademija Nauka i Umjetnosti Bosne i Hercegovine. Gimbutas, M. 1974: Anza, ca. 6500–5000 B.C.: a cultural yardstick for the study of Neolithic southeast Europe. Journal of Field Archaeology 1: 26–66. — 1976: Neolithic Macedonia as Reflected by Excavation at Anza, Southeast Yugoslavia. Los Angeles: Institute of Archaeology, University of California. Girić, M. 1974: Körös–Starčevo nalazišta u severnom Banatu. In Tasić, N. (ed.) Počeci Ranih Zemloradničkih Kultura u Vojvodini i Srpskom Podunavlju. Beograd: Srpsko Arheološko Društvo / Subotica: Gradski Muzej. Materijali 10: 169–187. Gläser, R. 1991: Bemerkungen zur absoluten datierung des beginns der westlichen Linienbandkeramik. Banatica 11: 53–64. Greenfield, H.J. 1986: The Paleoeconomy of the Central Balkans (Serbia): A Zooarchaeological Perspective on the Late Neolithic and Bronze Age (4500–1000 BC). BAR International Series 304. Oxford: British Archaeological Reports. Gronenborn, D. 1999: A variation on a basic theme: the transition to farming in south-central Europe. Journal of World Prehistory 13:

Aitken, M.J. 1983: Basic techniques for archaeointensity determination. In Creer, K.M. et al. (eds) Geomagnetism of Baked Clays and Recent Sediments. Amsterdam: Elsevier, 78–87. Aitken, M.J., Alcock, P.A., Bussell, G.D. & Shaw, C.J. 1983: Paleointensity studies on archaeological material from the Near East. In Creer, K.M. et al. (eds) Geomagnetism of Baked Clays and Recent Sediments. Amsterdam: Elsevier, 122–127. Aranđelović-Garašanin, D. 1954: Starčevačka kultura. Ljubljana. [see also Garašanin, D.] Arbour, G. & Schwartz, E.J. 1982: Archeomagnetic intensity of Indian potsherds from Quebec, Canada. Journal of Geomagnetism and Geoelectricity 34: 129–136. Bailey, D.W. 2000: Balkan Prehistory. Exclusion, Incorporation and Identity. London: Routledge. Bailey, D.W., Tringham, R., Bass, J., Stevanović, M., Hamilton, M., Neumann, H., Angelova, I. & Raduncheva, A. 1998: Expanding dimensions of early agricultural tells: The Podgoritsa Archaeological Project, Bulgaria. Journal of Field Archaeology 25: 373–396. Barbetti, M.F., MCElhinny, M.W., Edwards, D.J. & Schmidt, P.W. 1977: Weathering processes in baked sediments and their effects on archaeomagnetic field-intensity measurements. Physics of the Earth and Planetary Interiors 13: 346–354. Benac, A., Garašanin, M. & Srejović, D. 1979a: Uvod. Praistorija Jugoslavenskih zemalja II: 11–31. Sarajevo: Akademija Nauka i Umjetnosti Bosne i Hercegovine. Benac, A., Garašanin, M. & Srejović, D. 1979b: Zaključna razmatranja. Praistorija Jugoslavenskih zemalja II: 635–667. Sarajevo: Akademija Nauka i Umjetnosti Bosne i Hercegovine. Bogucki, P. 1995: The Linear Pottery culture of central Europe: conservative colonists? In Barnett, W.K. & Hoopes, J.W. (eds) The Emergence of Pottery. Washington D.C.: Smithsonian Institution Press, 89–97. Bogucki, P. & Grygiel, R. 1993: The first farmers of central Europe: a survey article. Journal of Field Archaeology 20: 399–426. Bonsall, C., Cook, G., Lennon, R., Harkness, D., Scott, M., Bartosiewicz, L. & MCSweeney, K. 2000: Stable isotopes, radiocarbon and the Mesolithic–Neolithic transition in the Iron Gates. Documenta Praehistorica 27: 119–132. Bonsall, C., Cook, G., Manson, J. & Sanderson, D. 2002: Direct dating of Neolithic pottery: progress and prospects. Documenta Praehistorica 29: 47–59. Boyd, M. 1986: A new method for measuring palaeomagnetic intensities. Nature 319: 208–209. Brukner, B. 1968: Neolit u Vojvodini. Belgrade: Arheološko Društvo Jugoslavije, Dissertationes V. — 1979: Körös-grupa. Praistorija Jugoslavenskih zemalja II: 213–226. Sarajevo: Akademija Nauka i Umjetnosti Bosne i Hercegovine. Brukner, B., Jovanović, B. & Tasić, N. 1974: Praistorija Vojvodine. Novi Sad: Institut za Izučavanje Istorije Vojvodine i Savez Arheoloških Društava Jugoslavije. Bucha, V. 1967: Intensity of the earth’s magnetic field during archaeological times in Czechoslovakia. Archaeometry 10: 12–22. Bucha, V., Taylor, R.E., Berger, R. & Haury, E.W. 1970: Geomagnetic intensity: changes during the past 3000 years in the western hemisphere. Science 168: 111–114. Budja, M. 1999: The transition to farming in Mediterranean Europe — an indigenous response. Documenta Praehistorica 26: 119–141. Burakov, K.S. & Nachasova, I.E. 1985: Correcting for chemical change during heating in archaeomagnetic determinations of the ancient geomagnetic field intensity. Physics of the Solid Earth 21(10): 801–803. Izvestiya, Akademiya Nauk, SSSR. 1

99

The Iron Gates in Prehistory 123–210. Hertelendi, E., Kalicz, N., Raczky, P., Horváth, F., Veres, M., Svingor, É., Futó, I. & Bartosiewicz, L. 1995: Re-evaluation of the Neolithic in eastern Hungary based on calibrated radiocarbon dates. Radiocarbon 37: 239–244. Horton, D. 1938: Note on a microscopic study of a sample group of barbotine sherds with positive appliqué from ‘Grad’, Starčevo. Proceedings of the American Philosophical Society 78: 397–402. Kaiser, T. 1984: Vinča Ceramics: Economic and Technological Aspects of Late Neolithic Pottery Production in Southeast Europe. Unpublished PhD Dissertation, University of California, Berkeley. (University Microfilms, Ann Arbor). Korošec, B. 1973: Application de méthodes d’analyse statistique au problème de la chronologie de Starčevo. Arheološki Vestnik – Acta Archaeologica 24: 271–302. Kovacheva, M. 1977: Archaeomagnetic investigations in Bulgaria: field intensity determinations. Physics of the Earth and Planetary Interiors 13: 355–359. — 1980: Summarized results of the archaeomagnetic investigation of the geomagnetic field variation for the last 8000 yr in southeastern Europe. Geophysical Journal of the Royal Astronomical Society 61: 57–64. — 1983: Archaeomagnetic data from Bulgaria and south eastern Yugoslavia. In Creer, K.M. et al. (eds) Geomagnetism of Baked Clays and Recent Sediments. Amsterdam: Elsevier, 106–110. — 1995: Bulgarian archaeomagnetic studies. In Bailey, D.W. & Panayotov, I. (eds) Prehistoric Bulgaria. Madison, WI: Prehistory Press, 209–224. Kovacheva, M. & Veljovich, D. 1977: Geomagnetic field variations in southeastern Europe between 6500 and 100 years ago. Earth and Planetary Science Letters 37: 131–138. Kovacheva, M. & Veljovich, D. 1985: Arheomagnitno izsledvane na praistoricheski obekti ot NRB i SFRYU. Arheologia 4(2): 21–26. Kovacheva, M. & Zagniy, G. 1985: Archaeomagnetic results from some prehistoric sites in Bulgaria. Archaeometry 27: 179–184. Kutzián, I. 1947: The Körös Culture. Budapest: Dissertationes Pannonicae. Lenneis, E. & Stadler, P. 1995: Zur absolutchronologie der Linearbandkeramik aufgrund von 14 C-Daten. Archäologie Österreichs 6: 4–13. Levi, S. 1977: The effect of magnetite particle size on paleointensity determinations of the geomagnetic field. Physics of the Earth and Planetary Interiors 13: 245–259. McPherron, A. 1988: Radiocarbon determinations. In McPherron, A. & Srejović, D. (eds) Divostin and the Neolithic of Central Serbia. Pittsburgh: University of Pittsburgh Ethnology Monograph 10, 379–383. Manson, J.L. 1990: A Reanalysis of Starčevo Culture Ceramics: Implications for Neolithic Development in the Balkans. Unpublished Ph.D. dissertation, Southern Illinois University, Carbondale (University Microfilms, Ann Arbor). — 1995: Starčevo pottery and Neolithic development in the central Balkans. In Barnett, W.K. & Hoopes, J.W. (eds) The Emergence of Pottery. Washington D.C.: Smithsonian Institution Press, 65–77. Manson, J.L. & Schmidt, V.A. 1991: Archaeomagnetic dating of Neolithic Yugoslavian potsherds. Paper presented at the 56th Annual Meeting of the Society for American Archaeology, New Orleans. Michels, J.W. 1973: Dating Methods in Archaeology. New York: Seminar Press. Milisauskas, S. & Kruk, J. 1989: Neolithic economy in central Europe. Journal of World Prehistory 3: 403–446. Milojčić, V. 1950: Körös–Starčevo–Vinča. Mainz: Reinecke Festschrift. Nagata, T., Arai, Y. & Momose, K. 1963: Secular variation of the

geomagnetic total force during the last 5000 years. Journal of Geophysical Research 68(18): 5277–5281. Radišić, R. 1984: Starčevačka kultura i Kereška kultura. In Praistorijske Kulture na Tlu Vojvodine. Novi Sad: Vojvoðanski Muzej, 9–13. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Bertrand, C., Blackwell, P.G., Buck, C.E., Burr, G., Cutler, K.B., Damon, P.E., Edwards, R.L., Fairbanks, R.G., Friedrich, M., Guilderson, T.P., Hughen, K.A., Kromer, B., MCCormac, F.G., Manning, S., Bronk Ramsey, C., Reimer, R.W., Remmele, S., Southon, J.R., Stuiver, M., Talamo, S., Taylor, F.W., van der Plicht, J. & Weyhenmeyer, C.E. 2004: IntCal04 terrestrial radiocarbon age calibration, 0–26 cal kyr BP. Radiocarbon 46: 1029–1058. Rusakov, O.M. & Zagniy, G.F. 1973: Intensity of the geomagnetic field in the Ukraine and Moldavia during the past 6000 years. Archaeometry 15: 275–285. Sakai, H. & Hirooka, K. 1985: Variation of the geomagnetic intensity deduced from archaeological objects in Japan. Rock Magnetism and Paleogeophysics 12: 12–19. Sasajima, S. 1965: Geomagnetic secular variation revealed in the baked earths in west Japan (part 2). Change of the field intensity. Journal of Geomagnetism and Geoelectricity 17: 413–416. Srejović, D. 1972: Europe’s First Monumental Sculpture: New Discoveries at Lepenski Vir. New York: Stein & Day. Sternberg, R.S. 1990: The geophysical basis of archaeomagnetic dating. In Eighmy, J.L. & Sternberg, R.S. (eds) Archaeomagnetic Dating. Tucson: University of Arizona Press, 5–28. — 1997: Archaeomagnetic dating. In Taylor, R.E & Aitken, M.J. (eds) Chronometric Dating in Archaeology. London: Plenum Press, 323–356. Stuiver, M. & Reimer, P.J. 1993: Extended 14C database and revised CALIB radiocarbon calibration program. Radiocarbon 35: 215–230. Stuiver, M., Reimer, P.J. & Reimer, R.W. 2005: CALIB 5.0. [WWW program and documentation]. Tanaka, H. 1982: Geomagnetic paleointensities for the period 6000 to 3000 years BP determined from lavas and pyroclastic flows in Japan. Journal of Geomagnetism and Geoelectricity 34: 601–617. Tasić, N. & Tomić, E. 1969: Crnolačka Bara: Naselje Starčevačke i Vinčanske Kulture. Kruševac: Narodni Muzej / Beograd: Arheološko Društvo Jugoslavije. Dissertationes 8. Thellier, E. & Thellier, O. 1959: The intensity of the earth’s magnetic field in the historical and geological past. Izvestiya. Geophysical Series 7: 929–949. Akademiya Nauk SSSR. Thomas, R.C. 1983: Mineralogical problems. In Creer, K.M. et al. (eds) Geomagnetism of Baked Clays and Recent Sediments. Amsterdam: Elsevier, 98–104. Todorović, J., Simoska, D. & Kitanoski, B. 1977: The settlement of Pešterica and the problem of the Early Neolithic in Pelagonia. Archaeologia Iugoslavica 18: 1–8. Tringham, R. 1971: Hunters, Fishers and Farmers of Eastern Europe 6000–3000 B.C. London: Hutchinson. Tringham, R., Brukner, B., Kaiser, T., Borojević, K., Bukvić, L., Šteli, P., Russell, N., Stevanović, M. & Voytek, B. 1992: Excavations at Opovo, 1985–1987. Socioeconomic change in the Balkan Neolithic. Journal of Field Archaeology 19: 351–386. U.S. Department of Commerce n.d.: Program #609 (GEOMAG). Environmental Science Services Administration, Coast and Geodetic Survey, Geomagnetism Div., Washington Science Center, Rockville, MD. Converted to VAX/VMS and MacIntosh by Victor A. Schmidt, University of Pittsburgh, Pittsburgh, PA. Vetnić, S. 1986: Elementi Protostarčeva-Starčeva I na neseljima tipa Rajkinac kod Svetozarevu u gornjem Pomoravlju. Starinar n.s. 37: 89–93.

1

100

Joni Manson: Approaches to Starčevo culture chronology Walton, D. 1983: The reliability of ancient intensities obtained by thermal demagnetization. In Creer, K.M. et al. (eds) Geomagnetism of Baked Clays and Recent Sediments. Amsterdam: Elsevier, 88–97. Weaver, G.H. 1966: Measurement of the past intensity of the earth’s magnetic field. Archaeometry 9: 174–186. Wei, Q., Zhang, W. & Li, D. 1987: Geomagnetic intensity as evalu-

ated from ancient Chinese pottery. Nature 328: 330–333. Whittle, A. 1996: Europe in the Neolithic. The Creation of New Worlds. Cambridge: Cambridge University Press. Whittle, A., Bartosiewicz, L., Borić, D., Pettitt, P. & Richards, M. 2002: In the beginning: new radiocarbon dates for the Early Neolithic in northern Serbia and south-east Hungary. Antæus 25: 63–117.

101

Faunal assemblages from the Early Neolithic of the central Balkans: methodological issues in the reconstruction of subsistence and land use Haskel J. Greenfield

Abstract: The earliest food producing societies in Europe north of the Aegean littoral appear in the central Balkans during the Early Neolithic. As a result, they are frequently included in discussions of the evolution of animal domestication in Europe. However, relatively little zooarchaeological research has been undertaken in the central Balkans. This paper has two goals. First, it introduces and discusses the nature, quality and problems pertaining to each of the major Early Neolithic zooarchaeological samples from the central Balkans to demonstrate the difficulty of inter-assemblage comparison. Second, it explores the potential of these assemblages for reconstructing animal exploitation strategies and land use patterns during the Early Neolithic, which are essential for understanding the economic processes involved in the evolution of early food producing societies in Southeast Europe. It is shown that it would be misleading to uncritically accept these assemblages as high quality sources of information for reconstructing the origins of animal domestication in Southeast Europe and animal exploitation strategies and land use patterns during the Early Neolithic. These assemblages have a very limited potential for understanding the economic processes involved in the evolution of early food producing societies in Southeast Europe. Key words: fauna, Neolithic, southeast Europe, subsistence, land use patterns

The earliest food producing societies in temperate Southeast Europe north of the Mediterranean littoral appear in the central-north part of the Balkan peninsula (hereafter referred to as the central Balkans — Fig. 1) during the earlier half of the Neolithic. The Early Neolithic (Starčevo culture) of the central Balkans, c. 7300–6400 BP (6150–5350 cal BC), witnessed the earliest spread of farming adaptations in Europe north of the Mediterranean littoral (Tringham 1971; Barker 1985). It is within this region that food production strategies based essentially upon an eastern Mediterranean complex of plants and animals are modified before spreading to the temperate climatic zones of Central and Northern Europe (Champion et al. 1984; Barker 1985). Many scholars have proposed explanatory models, often including economic or subsistence components (with a paucity of supporting data), for the spread of the Early Neolithic food producing cultures of temperate Southeast Europe (e.g. Ammerman & Cavalli-Sforza 1971; Garašanin 1973, 1983; Bökönyi 1974a; Barker 1985; Whittle 1985). While the Early Neolithic of the central Balkans is considered one of the most crucial periods in European prehistory, it nonetheless remains one of the most poorly investigated in terms of prehistoric economics. But, if economic processes are recognized as fundamental to the emergence of food producing society, economic data must be collected and used to test models for their evolution. The appearance of early food producing societies in temperate Southeast Europe is often assumed by archaeologists to be the result of the migration of previously established food producing communities from the south (e.g. Childe 1929, 1958; Ammerman & Cavalli-Sforza 1971; Tringham 1971: 70ff.; Champion et al. 1984: 100, 120; cf. Dennell 1985: 153ff.; Whittle 1985: 54). Two often unstated assumptions are used to support this view of the spread of food producing societies into temperate Southeast Europe:

1. If domestic fauna are found to be dominant in Early Neolithic sites in the region, these sites belong to intrusive colonists colonizing an essentially open nearly uninhabited environment; and 2. Since terminal Mesolithic sites are poorly represented in the region (with the exception of agriculturally marginal areas such as the Danubian Iron Gates or Montenegrin highlands), significant indigenous Mesolithic huntinggathering populations survived only in such areas, and slowly adopted domestic economies. Therefore, any sites with domesticated plants and animals found outside of these refugia were assumed to belong to intrusive agricultural populations. The first Early Neolithic zooarchaeological studies from the region demonstrated a preponderance of domestic fauna in Early Neolithic sites outside of the Iron Gates (e.g. Bökönyi 1974b, 1976, 1988; Clason 1980). These were used in support of this circular reasoning. Recent studies indicate a more complex picture. As will be shown in this paper, the frequency of domestic fauna in sites cannot be used to argue for the presence or absence of indigenous populations in and outside of agriculturally marginal areas. Rather, the frequencies of faunal remains in sites are a reflection of differential resource exploitation, recovery methodology, and bone assemblage attrition. The Early Neolithic of the central Balkans has been extensively investigated because of the interest in the origins of food production in Europe. But, little effort has been placed upon the systematic recovery of faunal remains from such sites. This has resulted in a dramatically biased zooarchaeological database, and one that has been uncritically accepted in the literature to explain the origins of food production in Southeast Europe (e.g. Murray 1970; Barker 1985; Whittle 1985; Halstead 1988). Neither the few systematic zooarchaeological studies nor the secondary discussions employing such data have paid attention to the taphonomic history of

103

The Iron Gates in Prehistory

Figure 1. Map of the central Balkans, showing Early Neolithic sites with analyzed faunal assemblages mentioned in text. Site name abbreviations used in all figures and tables are: A=Anza; B=Blagotin; BC=Bukovačka Česma; D=Divostin; F=Foeni-Salaş; G=Golukot; HV=Hajdučka Vodenica; LV=Lepenski Vir; LB=Ludoš Budžak; M=Madžari; N=Nosa; O=Obre; P=Padina; RB=Rug Bair; S=Starčevo; V=Vlasac.

samples or to the methodological problems connected with the faunal samples from this period and region (cf. Gifford 1981; Greenfield 1991, 1993). In the region, there is a dearth of published systematic analyses of Early Neolithic fauna. Most published accounts consist of species-frequency lists or preliminary reports, usually lacking detailed supporting data. These shortcomings in the database are significant influences during reconstruction of prehistoric economies on the basis of inter-site patterning in bone assemblages. It is only after methodological and taphonomic issues are addressed that animal exploitation strategies can be discussed. Some of the more important Early Neolithic faunal samples from the central Balkans are discussed here in order to demonstrate

some of the difficulties involved in inter-assemblage comparison and their potential for reconstructing animal exploitation strategies and land use patterns.

The region, archaeological ‘culture’ and sites The region The central Balkans encompasses the eastern half of the former Yugoslavia. It includes the countries presently known as Serbia, Bosnia, and Macedonia. The term Balkan was originally a Turkish word, meaning a chain of (forested) mountains (Naval Intelligence Division 1944; Klaić 1982). The

104

Haskel Greenfield: Early Neolithic faunal assemblages from the central Balkans

Balkan Peninsula is a topographically complex environment with several interconnected mountain systems coursing through the area. Highlands and lowlands are juxtaposed, with more than 37% of the land above 500m (Turnock 1989: 8). This variation in landforms within a small area has a strong influence on local climate. The regional climate is transitional between that of temperate Central Europe and the more arid Mediterranean basin. Climate and plant and animal communities take on different characteristics not only in a general north–south gradient, but also with increasing altitude. Neighbouring valleys often exhibit very different combinations of regional environmental variables, yet retain the general pattern of environmental diversity within the area as a whole (Pounds 1969). The Early Neolithic of the central Balkans The advent of the Neolithic in this region is generally connected to the appearance of food producing technology and adaptations, such as settled communities, ceramics, polished stone tools, domestic plants and animals. These are found in large quantities for the first time with the Starčevo culture. The Early Neolithic of the central Balkans coincides for the most part with the spatial extent of the Starčevo ‘culture’ c. 7300–6400 BP (6150–5350 cal BC — cf. Tringham 1971; Gimbutas 1976; Garašanin 1983).1 Geographical variants of the Starčevo culture are known from Romania (Criş) and Hungary (Körös). The Starčevo culture has been temporally subdivided into several sub-phases. Several competing chronological systems have been proposed on the basis of, primarily, the seriation of ceramic materials found at stratified and unstratified sites throughout the region (e.g. Milojčić 1949; Aranđelović-Garašanin 1954; Gimbutas 1976; Garašanin 1979, 1983; Srejović 1988). It is considered to be the archaeological manifestation of the earliest Neolithic populations in the region. Large-scale excavations have demonstrated that Starčevo settlements are not internally differentiated into functionally distinct areas (such as domestic houses, specialized work areas, cemeteries). There is little evidence for socioeconomic differentiation among houses or burials that would imply significant status distinctions within communities. Remains of structures include both semi-subterranean and ground-level dwellings, often with associated hearths and refuse/storage pits. Artefacts include coarse and fine ceramics, small ceramic altars, amulets, chipped and ground stone tools, bone and antler tools, and unmodified animal bones and shells. The ceramics are thick-walled and decorated with a variety of typical Starčevo decorative motifs, including mono- and polychrome painting, interior burnishing, pinching, impressing, channelling, and barbotine surface decorations (Tringham 1971; Garašanin 1979, 1983; Srejović 1979; Leković 1985; Bogdanović 1988; McPherron & Srejović 1988; Chapman 1989; Greenfield 2000; Jongsma & Greenfield 2001). The sites Most of our knowledge of Early Neolithic subsistence derives from four groups of sites. Each is located in a different environmental context. The nature of the subsistence data differs between each of these environments.

1. The first are found along the banks of the Danube in the gorge known as the Iron Gates (Lepenski Vir: Bökönyi 1971; Padina: Clason 1980; Hajdučka Vodenica: Greenfield, this volume). The sites in the gorge share access to a similar range of resources and have similar faunal proportions. 2. The second group is located in the northern end of the Mediterranean environmental zone. They are also at the southern end of the Starčevo range. Sites include Anza and Rug Bair in a small Macedonian upland basin (Bökönyi 1976; Schwartz 1976). 3. The third group includes those found in the rolling hills and river valleys of central Serbia and Bosnia, such as Divostin and Blagotin nestled among the rolling hills of central Serbia, Bukovačka Česma in the Morava river valley, and Obre I in a valley deep in the mountains of northern Bosnia (Bökönyi 1974b, 1988; Greenfield 1994, Greenfield & Jongsma n.d.). This is the transition zone between the Mediterranean and the Central European temperate climatic system. River valleys, such as the Morava, Bosna and Vardar were important routes for the spread of Early Neolithic adaptations from the Mediterranean to Central Europe and a centre for Starčevo settlement (cf. Barker 1975; Srejović 1979). As such, they represent an important link in understanding the spread of early farming adaptations out of the southern Balkans into the rest of Europe. 4. The fourth group is found in the Pannonian Plain (a nonpolitical term for the great lowlands encompassing Hungary, eastern Romania, and the northern part of the former Yugoslavia). These sites are at the northern end of the Starčevo range. Sites from this group include the type site of Starčevo on the Danube along the south edge of the plain and those further north and east, in the interior of the plain, such as Foeni-Salaş, Ludoš-Budžak, Nosa, and Golukot (Bökönyi 1974a; Clason 1980; Lazić 1988; Greenfield et al. n.d.). They are found close to the Hungarian or Romanian borders and are located in a more Central European climatic system (Pounds 1969).

Environmental variability in site location Based on vegetation, climate and topography, the region can be divided into four major zones: 1. A Mediterranean zone in the southern half of the Balkan Peninsula, with low annual precipitation, high summer temperatures, and semi-steppe vegetation (Anza and Rug Bair). 2. A temperate transition zone in the hills of Serbia and Bosnia, transitional in many of its environmental characteristics between the Mediterranean and Central European systems. This zone is characterized by high, year-round precipitation, deciduous forests and strong soil development. The sites in this zone are located on hills overlooking river floodplains and surrounded by agriculturally fertile soils (Blagotin, Obre I, Divostin, Bukovačka Česma). 3. A temperate highland zone in the Iron Gates gorge of the Danube, surrounded by the mountains of eastern Serbia, thick deciduous forests, soils with poor agricultural po-

105

The Iron Gates in Prehistory

tential, and overlooking the Danube and its rich aquatic resources (Hajdučka Vodenica, Lepenski Vir, Padina); and 4. A temperate Central European lowland zone in the plains of Pannonia (Starčevo, Golukot, Foeni-Salaş, Nosa, and Ludoš-Budžak). These sites overlook the floodplains of rivers, streams and marshes, with flora and fauna adapted to soils with high moisture levels.

Some methodological issues in inter-assemblage comparison Although consideration of the taphonomic history of each sample is a necessary precondition to the reconstruction of prehistoric subsistence and land use, most of the samples were not excavated, analyzed or published in a manner that allows a systematic reconsideration of their taphonomic history. Consideration of the methodological problems underlying the faunal studies can be used, however, to enhance our understanding of some of the inter-assemblage variability. Eventually, we may be able to differentiate patterning in the samples attributable to methodology from that of prehistoric human behaviour. Some of the more important methodological issues in inter-assemblage comparison are discussed next. The issues chosen for discussion are those that can be re-evaluated given the limited nature of most faunal publications. Recovery methodology It has long been a truism of faunal analysis that recovery methodology dramatically affects the outcome conclusions of the analysis (Payne 1972; Casteel 1973). As a result, collection methodology must be considered before any conclusions concerning human behaviour can be made. In general, bone collection from this region has been unsystematic. All of the samples were unsieved except for Anza, Blagotin, Foeni-Salaş and Rug Bair. Only Foeni-Salaş and Rug Bair’s fauna were entirely sieved. The sieved and unsieved samples from Anza and Blagotin were not separated during analysis. Hand recovery affects the results of all the analyses in the following manner: 1. The bones of larger animals will be more completely collected than smaller animals. This will result in overrepresentation of larger animals in assemblages. For example, cattle remains were collected more frequently than sheep remains, over-representing them in the assemblage relative to their original frequencies. Such is the situation at Obre (Bökönyi 1974b; Clason & Prummel 1977; Greenfield 1986, 1991). 2. The size of the bone is also an important variable. The larger the bone of the same species, the more likely it is that it will be collected. Large bones are more frequently collected than the smaller bones of the same taxon. Thus, reconstruction of preferential selection of body parts for use and/or disposal is compromised. If excavators preferentially collect large unbroken long bones (e.g. humerus) and ignore the smaller bones (e.g. phalanges, tarsals and carpals), then one may mistakenly assume that the prehistoric occupants disposed of the missing bones elsewhere. However, the reason that the smaller bones are missing from the assemblage is due more to the behaviour of the

archaeologists conducting the excavation than to the behaviour of the prehistoric occupants (Lyman 1994). While most faunal studies from the region do not list the type and frequency of various element categories, some studies do (Clason 1980; Greenfield 1994, n.d. a, b, this volume). There is a systematic recovery bias against small bone elements. When the distribution of element categories are compared in the unsieved samples, there is a total or near absence of small elements (phalanges, carpals and tarsals) for the medium- and small-sized mammals (e.g. ovicaprines, pigs, dogs, beaver). They are relatively more common among large mammals, both domestic and wild. In the sieved samples, the small element categories are more common. It is also possible to note the effect of recovery technology upon overall sample size by comparing the Iron Gates samples (Lepenski Vir, Padina, Vlasac, and Hajdučka Vodenica — Bökönyi 1971, 1978; Clason 1980; Greenfield, this volume). The excavators of Lepenski Vir, Padina, and Hajdučka Vodenica openly acknowledge that bone recovery was limited and biased toward what was considered to be tools or otherwise modified bones (Greenfield, this volume; B. Jovanović, pers. comm.). The result was extremely low recovery rates when compared to Vlasac (Srejović & Letica 1978) where efforts were made to avoid the mistakes of the earlier excavations in the gorge. The other assemblages are dominated by large mammal taxa (red deer, cattle). At Vlasac, contrary to the other excavations in the gorge, the remains of fish and various small- and medium-sized taxa dominate the assemblage in all phases. Quantification Only two methods have been used to quantify species representation from Early Neolithic faunal assemblages in the central Balkans: number of individual specimens (NISP) and minimum number of individuals (MNI) (see Bökönyi 1970; Greenfield 1986, 1991 — for reviews of procedures in the Balkans). In this study, only NISP counts are employed for several reasons. First, the problems with MNI seem to be more severe than with NISP (Grayson 1984). For example, both MNI and NISP appear to be equally predictive of species abundance in large samples (>10,000 fragments). Some of the samples, however, are relatively small, and NISP is more useful for the analysis of small samples, especially those with unequal species frequency ratios (Gilbert et al. 1982). Also, MNI counts are not given for all of the samples, being available only for the total remains of each species from six of the sites (Divostin, Lepenski Vir, Ludoš-Budžak, Nosa, Obre and Anza I–III). In each instance, the values were calculated in the same manner (Bökönyi 1970). Finally, NISP counts are available for all of the sites, making the samples more comparable. In the NISP counts from Blagotin, FoeniSalaş, Hajdučka Vodenica and Bukovačka Česma, whole and partially articulated skeletons and limbs were not doublecounted. Articulated specimens were counted only once. Antler and horn fragments were few and mostly attached to cranial elements (Greenfield 1986). It is not known if the NISP counts from the other sites were calculated in exactly the same way since there is no discussion of NISP methodology in their publication. Even though the question of individual groups of articulated elements was

106

Haskel Greenfield: Early Neolithic faunal assemblages from the central Balkans Table 1. Frequency distribution of burnt and unburnt bones from Early Neolithic levels of each site. Site

Context

Burnt N

Blagotin

Exterior Pit house - central Pit house - peripheral Plough zone above pithouse Subtotal

Bukovačka Česma

All

Foeni-Salaş

If grey bones are deducted

Hajdučka Vodenica

All

Unburnt %

N

Total N

%

20 319 482 2 823

2.27% 1.82% 3.58% 1.38% 2.57%

861 17228 12982 143 31214

97.73% 98.18% 96.42% 98.62% 97.43%

881 17547 13464 145 32037

1

0.18%

543

99.82%

544

497

6.06%

7700

93.94%

8197

2

1.79%

110

98.21%

112

Table 2. Frequency distribution of weathering stages from Early Neolithic levels of each site. Weathering Stages Site

Context type

Blagotin

Exterior in cultural horizon Exterior - in ploughzone above ZM02 Subtotal - Exterior Interior (Pit house - central) Interior (Pit house - peripheral) Subtotal - Interior Grand total

Bukovačka Česma

Mixed

Foeni-Salaş

Exterior in cultural horizon Interior (Pit house) Grand total

Hajdučka Vodenica

Mixed

Light

Medium

Heavy

Total

50 4 54 10761 2194 12955 13009

14.49% 2.17% 10.21% 61.27% 16.30% 41.77% 41.24%

133 135 268 5893 9994 15887 16155

38.55% 73.37% 50.66% 33.56% 74.27% 51.22% 51.21%

162 45 207 908 1268 2176 2383

46.96% 24.46% 39.13% 5.17% 9.42% 7.02% 7.55%

345 184 529 17562 13456 31018 31547

544

100.00%

0

0.00%

0

0.00%

544

485 632 1117

14.52% 12.20% 13.09%

2592 4375 6967

77.58% 84.46% 81.67%

264 183 447

7.90% 3.53% 5.24%

3341 5190 8531

27

24.11%

29

25.89%

56

50.00%

112

not directly addressed in the previously published analyses of any of the sites except those analyzed by the author, the articulated remains of whole skeletons were not found at Anza, Obre, Divostin, Hajdučka Vodenica and Bukovačka Česma. Only the analyses conducted by the author have recognized the importance of documenting the presence of articulated limbs in order to modify NISP or MNIs. Articulated limbs were found at Blagotin, Bukovačka Česma and Foeni-Salaş (Bökönyi 1974b, 1976, 1988; Greenfield 1994, this volume, n.d. a, b). Data publication Each of the published assemblages has been published in ways that limit reanalysis. Even though the samples from Anza, Divostin and Obre were excavated by both natural and artificial stratigraphic units, the Early and Late Neolithic samples were not distinguished for publication except for initial summary statistics. For example, while species proportions over time are noted, changes in age and sex proportions are not reported. As a result, it is impossible to identify shifts in culling strategies. Curation Saving and storing bone collections in safe and accessible facilities is fundamental to the pursuit of the discipline. Advances made in one generation can be improved upon if

the next generation can access the materials. However, the central Balkans is plagued by poor storage facilities and an inadequately developed ethos in relation to the storage of zooarchaeological materials. The faunal remains from most of the largest collections were discarded immediately after analysis (Obre, Lepenski Vir, Divostin — personal communications from Sandor Bökönyi, Dragoslav Srejović, Svetozar Stanković). Only the bone tools were kept since they were considered to be artefacts, in addition to small samples of other bones. The faunal remains from Anza and Rug Bair were kept for several years in a storage shed outside of the local museum. By the time I tried to re-examine the Anza remains in 1982, rats had gotten into the shed. All of the bags and tags were destroyed, and the faunal remains subsequently discarded by the curators. The remains from Hajdučka Vodenica and Padina are stored at the central storage facility for the Serbian Iron Gates sites at Karataš. It is extremely difficult to access material in this facility since crates of material are stored one on top of the other, without any central cataloguing system. The remains from Starčevo and Blagotin are stored in Belgrade (at the National Museum and the University of Belgrade, respectively), while those from Foeni-Salaş are in Timişoara, Romania (Muzeul Banatului). Selected samples from Hajdučka Vodenica, Blagotin and Foeni-Salaş are in Winnipeg, Canada (University of Manitoba). The fate of the other collections is

107

The Iron Gates in Prehistory Table 3. Faunal remains from central Balkan Early Neolithic sites. The ‘number of taxa’ column includes all bones identified to a class or finer taxonomic level. Unidentified ovicaprines (i.e. those that are not identified to either Ovis or Capra) are not counted if Ovis or Capra are present. If Ovis or Capra are not present, then unidentified ovicaprines would be counted only once. All invertebrates, bird, and fish remains, respectively, are each treated as a single taxon owing to the low level of analysis accorded their remains (Bökönyi 1970, 1974, 1976, 1988; Clason 1980; Lazić 1988; Moskalewska 1986; Schwartz 1976; Greenfield n.d. a, b). Site

Figure abbreviations

Anza I–III

Number of taxa

Wild

A

14

138

Domestic 4.25%

3112

NISP

Notes

95.75%

3250

a a

Blagotin

B

18

719

8.29%

7959

91.71%

8678

Bukovačka Česma

BC

11

166

61.48%

104

38.52%

270

Divostin I

D

14

203

8.45%

2198

91.55%

2401

Foeni-Salaş

F

17

416

19.87%

1678

80.13%

2094

Golukot

G

10

893

68.38%

413

31.62%

1306

Hajdučka Vodenica

H

11

211

84.40%

39

15.60%

250

Lepenski Vir III

L

21

1765

74.50%

640

27.02%

2369

Ludoš-Budžak

LB

18

572

20.91%

2163

79.09%

2735

Madžari

M

17

185

6.44%

2689

93.56%

2874

Nosa

N

16

736

71.88%

229

22.36%

1024

Obre I

O

19

1700

20.75%

6491

79.25%

8191

Padina B

P

19

2656

92.64%

211

7.36%

2867

Rug Bair

R

9

13

1.88%

680

98.12%

693

Starčevo

S

19

1029

25.91%

2943

74.09%

3972

b

a. Sieved and unsieved samples were not separated during analysis. b. Sieved fauna from the American excavations.

unknown at present, but I suspect that they have been discarded. The various analysts also had an influence on curation. In general, Sandor Bökönyi advised his Yugoslavian hosts that they could discard the remains after he finished his analysis. In contrast, all other analysts encouraged that their collections be curated. However, since the advent of the Balkan wars of the 1990s, the condition of any of the collections is uncertain.

Physico-chemical assemblage attrition Most analyses ignore the issue of assemblage attrition through weathering and other destructive forces. The controversy surrounding the ‘schlepp effect’ is a good example (Lyman 1994). Without a discussion of taphonomic variables, it is impossible to judge the value of assemblages for subsistence reconstructions. Burning Most Early Neolithic faunal reports do not mention burnt bones at all. Only a very small fraction of burnt bones were found at Blagotin (2.57%), Bukovačka Česma (0.18%), Foeni-Salaş (6%) and Hajdučka Vodenica (1.8% — Table 1). The highest values come from Foeni-Salaş, which is two to three times higher than the values found in all the rest. Similarly low frequencies of burnt bones were found in Late Neolithic and post-Neolithic assemblages from the region (Greenfield 1986, 1991). Why are so few bones burnt? There are several possible reasons. Most bones probably were still covered with meat

when exposed to fire, were boiled in a stew-like concoction, or had their meat removed prior to cooking. In each case, there would have been limited contact between the bone and the flame. Also, the burnt bone may have simply disintegrated in the soil, since burnt bone breaks up into small fragments more readily than uncooked bone (Bonfield & Li 1966). Another reason that few pieces of burnt bone are found is differential disposal patterns. Cooking areas (hearths) and nearby middens are the depositional contexts in which burnt bone most likely would have been discarded. Other than in the Iron Gates sites, few hearths have been found. At Bukovačka Česma, almost all of the bones around the two excavated hearths (trench 5, N=40; trench 8, N=11) were not burnt. At Anza, most debris was acknowledged to derive from exterior middens (Bökönyi 1976), where the burnt bone would have broken up and disintegrated relatively quickly. However, due to the nature of excavation methodology at most of the sites that gathered together disparate pieces of bone from large-scale excavation units, no spatial analyses of bone distributions can be conducted. Why were so many more burnt bones found at Foeni-Salaş than at the other sites? Given the absence of any evidence for large-scale burnt destruction of the settlement, and the fact that most of the burnt bone remains were from small unidentifiable fragments, it would appear most likely that the extensive dry and wet sieving operation would have been responsible for the differences. Bone weathering and depositional context The degree of weathering of samples can also profoundly affect the level of reconstruction. Analyses that avoid this issue

108

Haskel Greenfield: Early Neolithic faunal assemblages from the central Balkans

Figure 2. Semi-logarithmic scale scatter-plot showing relationship between number of taxa identified and identified taxa assemblage size in each assemblage.

are in danger of reconstructing the patterns of preservation rather than any prehistoric behaviour. There are profound differences in preservation between the samples analyzed by the author. One major source of bone weathering is soil pH. Soil pH levels, however, were not measured for most sites, since they were excavated prior to an interest in taphonomy in the region. However, pH was measured at Blagotin and FoeniSalaş. At both, the Early Neolithic deposits ranged from pH 5.0 to 7.0, depending upon the context — relatively normal values for the region. While pH may have had less of an influence, depositional context probably had a great influence. There appear to be regional variations in bone preservation. The material from the Iron Gates seems to be much more poorly preserved than the other regions. Most of the Hajdučka Vodenica sample was heavily eroded (50% — Table 2), with characteristic weathering cracks, pitted bone

surfaces, and the loss of delicate features. Based on a study of some of the bone tools from Lepenski Vir and Vlasac curated at the University of Belgrade, the state of preservation of the Hajdučka Vodenica sample is probably characteristic of the other Iron Gates samples. In the central Serbian sites, bones tend to be much better preserved. The Bukovačka Česma sample is remarkable for the preservation of even delicate morphological features (Greenfield 1994). All of the bones were very well preserved. At Divostin, bones were well preserved when buried in the ashy and alkaline soil of pits. Bones from the cultural horizon (which is composed of Smonica-type soils, with a heavy clay content and high water retention), however, were so fragile and soft that they fragmented or disintegrated during excavation (Lyneis 1988: 301).2 At nearby Blagotin, there were similar patterns of preservation to those found at Divostin (Greenfield 2000, n.d. a; Greenfield & Jongsma n.d.). Most of the bones were lightly (41%) or moderately (51%) weathered (Table 2). However, when the remains are divided between those found within features (such as pits) and those that lay in the open, the degree of preservation is very different. In the exterior deposits, whether in the plough zone or in the cultural deposits, most of the remains were either moderately (51%) or heavily (39%) weathered. Very few were lightly weathered. In the interior deposits, most of the remains were moderately (51%) or lightly (41%) weathered. There are some differences between specific deposits, but the effect of depositional context on preservation remains (Meadow 1978). Only one sample had information from Pannonia — Foeni-Salaş (Greenfield n.d. b; Greenfield et al. n.d.). It had a very different pattern of preservation. Most of the bones (81.6% — Table 2) were only moderately preserved, regardless of whether bones were recovered in pits or in the exterior deposits. This is a reflection of the sandy loam deposits of the site. It is difficult to gauge the pattern of bone preservation at tell sites (which had deeply stratified deposits that were advantageous for bone preservation — Obre I, Anzabegovo) since there is no recorded information on these sites. In general, preservation is worst in the Iron Gates (where they were heavily eroded), those from Pannonia were better, and those from central Serbia were the best. Bukovačka Česma had the best preserved assemblage of all. Little is known about preservation at the other sites in the region.

Discussion Despite all of the above issues, can anything be said about subsistence practices? Before this can be accomplished, two other issues must be examined.

Figure 3. Normal scale scatter-plot showing relationship between number of taxa identified and identified taxa assemblage size in each assemblage.

Sample size and taxonomic diversity The samples under consideration range quite widely in size. While large samples are best for reconstruction of subsistence practices, even small assemblages should be considered. They can contribute information otherwise unobtainable from other sources. Even though the number of samples and the diversity of environmental settings of Starčevo sites with analyzed fauna is extremely small, each 109

The Iron Gates in Prehistory

Figure 4. Normal scale scatter-plot showing relationship between percentage of wild animals in the assemblage and identified taxa assemblage size.

adds a new dimension to our understanding of Starčevo adaptations. Species counts are available by separate Starčevo phases only for Anza. At Obre, it is difficult to correlate the species frequencies with major occupational phases or strata (cf. Bökönyi 1974b; Greenfield 1991). The faunal samples from most sites were not published by separate levels, horizons, or phases within the Starčevo layer(s) (except at Starčevo, itself). As a result, it would be impossible to re-analyze the material, if and when the sub-phasing of the Starčevo culture, in general, or excavation units at particular sites, becomes more refined. The total fragment count of unidentified specimens (Table 1) is unknown from half of the sites (Anza, Divostin, Lepenski Vir, Obre). Since total fragment count cannot be used to examine the impact of total number of fragments upon taxonomic diversity, a proxy measure must be used instead — i.e. NISP. Only Obre and Blagotin have substantial samples of identified taxa (8000+ NISP). The rest have medium (2000–4000 bones) or small samples (c. 1000 or less). When the number of identified taxa are plotted against the number of bones identified to a genus or species in the sample on a semi-logarithmic scale, all the sites fall into a single large cluster (Fig. 2). A semi-logarithmic scale was chosen because the sample sizes are so disparate. It would appear that a curvilinear relationship exists between the two variables (y = 4005.6 Ln (x) – 7917.9). Statistically, when a curvilinear relationship appears and there is a low correlation (R2 = 0.1951), it is possible that the nature of the scale was incorrect. When the two variables are plotted on a normal scale, a different pattern emerges. Two separate distributions appear (Fig. 3). In the lower half of the graph, there is a strong positive correlation between identified sample size and species

diversity (R2 = 0.73). In other words, the greater the size of the identified sample, the greater is the number of identified taxa. The small number of identified taxa in small assemblages is, therefore, a reflection of sample size. Rare species may appear in small data sets, but less common forms appear regularly only when the sample reaches certain minimal size levels (usually several thousand identified fragments: Casteel 1973; Grayson 1984). Although the absence of rare species in small samples does not dramatically affect the overall analysis of subsistence systems, they are necessary for other types of analysis (e.g. ecological reconstruction: Hesse & Wapnish 1985). Several factors, such as recovery and preservation, may intervene to affect the normally linear relationship between the sample size of identified mammal bones and the number of mammal species. In the upper half of the graph, a different pattern appears, with the data from Blagotin and Obre. Major increases in taxonomic diversity can occur even when there are only modest increases in sample sizes. In these two cases, increases in sample size did not significantly alter taxonomic diversity. It would appear that once sample sizes reach a certain level (above 5000 identified specimens), other variables affect taxonomic diversity. In these cases, location would have limited the range of taxa available for exploitation. Obre is found overlooking a major river, while Blagotin is found in the highlands and away from any major water source. Hence, location and available surrounding resources can be a compounding issue. Domestic:wild proportions Is there a variable governing the relative frequency of domestic and wild animals in assemblages? Initial comparison of the domestic:wild species percentages from Starčevo sites reveals that half of the assemblages have relatively high and the other half relatively low percentages of wild animals. There is no overlap between the two groups. All three of the Iron Gates sites have very high wild species frequencies (>61%). But three of the sites outside of the Iron Gates (Ludoš-Budžak and Golukot in the Pannonian plain, and Bukovačka Česma in central Serbia) also have very high wild species counts. The remaining eight sites outside of the Iron Gates have a preponderance of domestic animals (>74%). How can these distributions be interpreted? In the past, sites outside of the Iron Gates were assumed to be those of intrusive food producers because most of the fauna was domestic. In contrast, sites in the Iron Gates were assumed to be from indigenous hunter-gatherers because: 1. Even when food production appeared, it was never a mainstay of the local economy (cf. Bökönyi 1971, 1978), and 2. The human skeletal record has been consistently interpreted to represent population continuity in the Iron Gates from the Terminal Palaeolithic through the Mesolithic and into the Early Neolithic (cf. Živanović 1975, 1976, 1979).3 In fact, the reality is more complex; sites with high and low frequencies of wild fauna are found outside the Iron Gates. If we continue to use the logic of most researchers on the Early Neolithic of the region, does this mean that even sites outside of the Iron Gates belong to hunter-gatherers? This is a point

110

Haskel Greenfield: Early Neolithic faunal assemblages from the central Balkans

to be returned to later. When the number of identified fauna from each of the sites is plotted against the percentage of wild animals identified in the assemblage, three clusters appear (Fig. 4). Using the xaxis, the samples can be divided on the basis of percentage of wild animals — one group with a high percentage of wild animals and two groups with low percentages. Each of the major groups can be subdivided on the basis of sample size: 1. The high wild percentage group contains sites with both low and medium sample sizes. This group comes from a variety of environmental settings. It includes the sites from the Iron Gates. The other sites in this group are found in the region beyond the narrow seclusion of the Iron Gates, in a variety of environmental contexts. The implications of these distributions appear fairly obvious. The high percentage of wild fauna found in the Iron Gates sites reflect the isolation, topographic ruggedness and abundance of wild fauna, and the lower energy returns of pastoralism versus hunting-fishing. The topography and environment of the gorge with its steep slopes and thick forests make pastoralism difficult and less profitable as a subsistence strategy when compared with the hunting of wild ungulates. The high percentage of wild fauna in the sites outside the Iron Gates (Bukovačka Česma, Golukot, and Nosa) is clearly influenced by the location of the sites near or on major riverine environments. These sites are more similar to those from the Iron Gates. This is a very different pattern than in the other sites found outside of the Iron Gates sites rather than the generally accepted pattern for sites outside of the Iron Gates. 2. The low wild percentage group can be divided into two sub-groups: (i) one sub-group is composed of sites with small and medium sample sizes; (ii) the other sub-group is composed of sites with large sample sizes. On the surface, it would appear that the difference in wild:domestic ratios may be affected by sample size. The smaller samples tend to have higher percentages of wild animals. The situation is more complex. One variable is recovery method (sieving or hand recovery). The partially sieved samples include both those with large (Blagotin) and medium sample sizes (Anza) among the low wild group. Sieving does not appear to have much of an effect on wild: domestic ratios. The preponderance of wild fauna cannot be attributed entirely to recovery method since wild animals are generally not larger than domestic animals. If anything, many wild taxa are smaller (e.g. beaver and fish) and would be more difficult to recover by hand. Hand collection would under-represent the wild component of the assemblage. Fish remains, in particular, are poorly collected by hand and therefore could be predicted to have been present in larger quantities before excavation. For example, they already represent 10% of the Bukovačka Česma sample, a hand collected site located above the floodplain of the Morava river. They were probably present in even greater frequencies before excavation. Another possibility is the surrounding environment and its relationship to the subsistence economy. Most of the sites with high frequencies of wild fauna are found in a limited set of environments — in the temperate environmental zone, and close to rivers, streams and floodplains that even today abound with wildlife. The sites with smaller sample sizes also 111

happen to be in such environments. Therefore, their high wild frequencies may not be a reflection of sample size, but of their exploitation of the rich wild resources in the surrounding wetlands. The faunal remains of Bukovačka Česma, Golukot and Nosa are more similar to those from the Iron Gates than from other sites spread through the hill country of Serbia, Bosnia or Macedonia because they also overlook resource-rich aquatic environments. Much of the source of variation between samples can therefore be attributed to a limited set of variables: recovery methodology, sample size and environmental location. Therefore, the percentage of wild or domestic fauna may still be used as a general reflection of overall environment. Recovery can alter the details significantly, nonetheless. Much of the source of variation between assemblages can therefore be attributed to a limited set of variables: recovery methodology, sample size and environmental location. The quantity of wild remains in sites outside of the Iron Gates can be relatively high indicating that traditional assumptions concerning the nature and rate of spread of the transition to a food producing economy rate in Southeast Europe need to be re-evaluated. The spread of food production may be a more complex process than previously envisioned. Some thoughts on the transition to food production in Southeast Europe If we accept the assumption that sites with wild resources in the Iron Gates are the remains of indigenous foragers, then it may even be that some of the sites outside of the Iron Gates with a predominance of wild resources may also be the remains of acculturating foragers. In fact, we know too little about the relationship between Mesolithic and Early Neolithic populations in this region to be able to make such blanket assumptions. While Early Neolithic sites are widespread across the region, so much is assumed in Mesolithic studies on the basis of a few highly concentrated localities (Iron Gates, Montengro). For much of the Early Neolithic of Central and Northern Europe, indigenous Mesolithic foragers probably coexisted with initial and intrusive food producers. All we need to do is to look at the evidence from other better-studied regions for comparison. For the first phase of the Early Neolithic of Central and Northern Europe, indigenous Mesolithic foragers coexisted with initial and intrusive food producers (e.g. Bogucki 1988). There is also evidence for acculturation by indigenous foragers to an agricultural lifestyle (Price 1987). In contrast, most studies of Southeast European prehistory implicitly assume that either the indigenous population immediately adopted food production upon its arrival or the area was not occupied by foragers at the advent of the Neolithic. The Mesolithic forager sites are not visible in the areas of intensive Early Neolithic settlement (in the great river valleys, beyond the Iron Gates and Montenegrin highlands). The widespread presence of Mesolithic sites (in areas that have been systematically surveyed, i.e. Iron Gates — Whittle 1985; or in regions with little vegetative cover, e.g. Montenegro — Djuričić 1991) and the demonstrable continuity between Mesolithic and Early Neolithic human populations in the more mountainous and less accessible areas of the peninsula (i.e. the Iron Gates — Živanović 1975, 1976,

The Iron Gates in Prehistory

1979; y’Edynak 1978) indicate that it is likely that foragers were also present in the lowlands at the advent of the Neolithic. If indigenous foragers were present at the advent of the Neolithic, it might be more realistic to assume that they coexisted with early food producers for a period of time, only slowly abandoning foraging subsistence economies? During this transition, they may have retained much of their commitment to foraging while at the same time incorporating some domestic food producing strategies and technologies into their subsistence round (as occurred in Southwest, Northwest and Northeast Europe — Price 1987; Rimantiene 1992), such as domestic animals, pottery production, or other hallmarks of Starčevo sites. If this is the case, we might expect foraging communities to retain their mobility, while adopting and incorporating domestic animals into their mobile subsistence system before adopting other Early Neolithic accoutrements (such as pottery) or vice versa as in northern and Western Europe (cf. Zvelebil & Dolukhanov 1991). If this was the case, many of the sites defined as Mesolithic on the basis of lithic typology and the absence of ceramics may, in fact, date from the Early Neolithic. The paucity of radiocarbon dates from ‘Mesolithic’-type sites makes this a difficult hypothesis to test immediately. However, there are no cases where domestic animals are found in clear association with a Mesolithic-type stone tool technology (and where Early Neolithic pottery is absent) outside of the Iron Gates. The few early Holocene sites without pottery, but with a Mesolithic stone assemblage, do not contain any domestic animals. But there are several sites outside of the Iron Gates with Early Neolithic material culture, but with assemblages dominated by wild animals (Greenfield, this volume). The question then arises — to whom do these settlements actually belong? 1. Are they settlements of indigenous foragers slowly acculturating to a food producing economy by selectively incorporating components of typical Starčevo material culture? 2. Are they a specialized wild food collection site for a larger community committed to domestic economies? 3. Do they represent new settlements established by recently arrived food producers who have yet to build up the agricultural component of their local economy? 4. Or, could they be settlements of food producers who have moved into an agriculturally marginal area? Each alternative has an emphasis upon wild resources as a logical explanation for the patterning in the data. How to choose between alternatives? The second and third choices seem less likely considering the size and degree of permanency of the settlements, as reflected by the material culture and domestic architecture, and the absence of any evidence for specialized production and exchange between contemporary sites involving foodstuffs at this time or even later in the Neolithic. Estimations of the seasonality of death of various species based upon the ageat-death of the various species at several sites also indicate year-round occupation (Clason 1980; Arnold & Greenfield 2006). It is more difficult to distinguish between the first and fourth choices. The choice varies depending upon the evidence chosen to consider. In the Iron Gates, there is evidence

for human population, material culture, and subsistence continuity between Mesolithic and Early Neolithic levels, even with the appearance of new subsistence strategies and material culture. But comparable studies have not been undertaken upon assemblages from Mesolithic sites outside of the Iron Gates. Also, virtually all sites with a Mesolithic-type artefact assemblage from the region outside of the Iron Gates are not dated by radiocarbon techniques. They are assumed to date earlier than Starčevo on the basis of the similarity of their material cultural assemblages to other Mesolithic sites in the Iron Gates and on the absence of Starčevo ceramic material culture.

Conclusion Previous studies of the earliest food producing communities in the more temperate zones of Southeast Europe, such as the central Balkans, have generally attempted to show that subsistence economies were largely orientated towards domestic animal economies (e.g. Murray 1970; Barker 1985). Faunal data from Early Neolithic sites in the central Balkans are reconsidered here to demonstrate that animal exploitation strategies during the Early Neolithic of the region was a relatively complex matter and that taphonomic issues must be considered if prehistoric economies are to be understood. Even though such faunal assemblages can increase our understanding of Early Neolithic subsistence strategies, it is not an easy leap from bones to human behaviour. A number of studies have demonstrated the obvious and not so obvious pitfalls of extrapolating from patterning in bone assemblages to prehistoric human behaviour (e.g. Gifford 1981; Greenfield 1986, 1988a, 1988b, 1991; Koster & Chang 1986). All of the analyses of Early Neolithic faunal assemblages and the ensuing comparative use of their results have ignored the possible distorting effects of bone assemblage attrition (e.g. Champion et al. 1984; Lazić 1988). There is a great deal of variability from site to site which affects the representativeness of the sample and utility of the final analysis, such as variability of the surrounding environmental conditions (temperature, precipitation), local sedimentary conditions (including soil type and pH), recovery procedures, and methods of quantification. As a result of the re-evaluation of the quality of the Early Neolithic faunal data derived from such an approach, it is hoped that this will force a reconsideration of the significance and value of conclusions derived from previously analyzed faunal assemblages. Notes 1. Manson (1990, this volume) established an age range of 6100–5100 cal BC, based on the archaeomagnetic intensity analysis of Starčevo and Körös ceramic sherds. 2. But note that Bökönyi’s (1988) faunal report does not deal with this issue. 3. Skeletal series from Starčevo sites outside of the Iron Gates are few and far between. However, they are now being re-analyzed and show continuity with Mesolithic populations (see this volume).

Acknowledgements I should like to express my sincere gratitude to my numerous friends

112

Haskel Greenfield: Early Neolithic faunal assemblages from the central Balkans and colleagues throughout the central Balkans who made available their assemblages for analysis or who helped in the excavation and analysis. Unfortunatel, there are too many to list separately here. Funding for this research came from a variety of sources, including the National Science Foundation (#BNS8105358), the WennerGren Foundation for Anthropological Research (#4210), the Fulbright-Hays Foreign Language and Area Studies Program (Program no. 84.022; Project no. 022AH10048), the International Research and Exchanges Board (1981–82, 1993–94), the University of Manitoba, Winnipeg (1990–99), and the Social Science and Humanities Research Council of Canada (1990–1996). I am grateful to all of them for supporting the fieldwork, analysis and writing up of this paper. In particular, I thank my wife, Tina Jongsma, who shared much of the field and laboratory analysis that led to this paper. Any errors, however, are my responsibility.

References Ammerman, A. & Cavalli-Sforza, L.L. 1971: Measuring the rate of spread of early farming in Europe, Man n.s. 6: 674–688. Aranđelović-Garašanin, D. 1954: Starčevačka Kultura. Ljubljana: Universitet v Ljubljana. Arnold, E.R. & Greenfield, H.J. 2006: The Origins of Transhumant Pastoralism in Temperate Southeastern Europe: A Zooarchaeological Perspective from the Central Balkans. British Archaeological Reports, International Series 1538. Oxford: BAR Publishing. Barker, G. 1975: Early Neolithic land use in Yugoslavia. Proceedings of the Prehistoric Society 41: 85–104. — 1985: Prehistoric Farming in Europe. Cambridge: Cambridge University Press. Bogucki, P. 1988: Forest Farmers and Stockherders: Early Agriculture and its Consequences in North-Central Europe. Cambridge: Cambridge University Press. Bökönyi, S. 1970: A new method for the determination of the number of individuals in animal bone material. American Journal of Archaeology 74: 291–292. — 1971: Animal remains from Lepenski Vir. Science 167: 1702–1704. — 1974a: History of Domestic Mammals in Central and Eastern Europe. Budapest: Akadémiai Kiadó. — 1974b: The vertebrate fauna. In Gimbutas, M. (ed.) ‘Obre I and II. Neolithic Sites in Bosnia’. Wissenschaftliche Mitteilungen des Bosnisch-Herzegowinischen Landesmuseum, Arheologie 3, part B: 55–154 (Sarajevo). — 1976: The vertebrate fauna from Anza. In Gimbutas, M. (ed.) Neolithic Macedonia as Reflected by Excavation at Anza, Southeast Yugoslavia. Los Angeles: The University of California, Los Angeles (Monumenta Archaeologica, vol. 1), 313–363. — 1978: The vertebrate fauna of Vlasac. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Science and Arts, 35–65. — 1988: The Neolithic fauna of Divostin. In McPherron, A. & Srejović, D. (eds) Divostin and the Neolithic of Central Serbia. Pittsburgh: University of Pittsburgh, Department of Anthropology (Ethnology Monographs no. 10), 419–446. Bonfield, W. & Li, C.H. 1966: Deformation and fracture of bone. Journal of Applied Physics 37: 869–875. Casteel, R.W. 1973: Some biases in the recovery of archaeological faunal remains. Proceedings of the Prehistoric Society 39: 382–388. Champion, T., Gamble, C., Shennan, S. & Whittle, A. 1984: Prehistoric Europe. London: Academic Press. Childe, V.G. 1929: The Danube in Prehistory. Oxford: Clarendon

Press. — 1958: The Prehistory of European Society. Harmondsworth: Penguin. Clason, A.T. 1980: Padina and Starčevo: game, fish and cattle. Palaeohistoria 22: 142–173. Clason, A.T. & Prummel, W. 1977: Collecting, sieving and archaeozoological research. Journal of Archaeological Science 4: 171–175. Dennell, R. 1984: European Economic Prehistory: A New Approach. London: Academic Press. Djuričić, L. 1991: Surveying Paleolithic and Mesolithic Sites (in Serbian). Glasnik Srpsko Arheolosko Društvo (Journal of the Serbian Archaeological Society — Belgrade) 7: 43–51. Garašanin, M. 1973: Praistorija na Tlu SR Srbija, 2nd edition. Belgrade: Srpska Kniževna Zadruga. — 1979: Centralbalkanska zona, In Basler, D., Benac, A., Gabrovec, S., Garašanin, M., Tasić, N., Cović, B. & VinskiGasparini, K. (eds) Praistorija Jugoslavenskih Zemalja II: Neolitsko Doba. Sarajevo: Akademija Nauka i Umjetnosti Bosne i Hercegovine, 79–212. — 1983: The Stone Age in the central Balkans. In Cambridge Ancient History, volume 3, part 1. Cambridge: Cambridge University Press, 75–135. Gifford, D.P. 1981: Taphonomy and paleoecology: a critical review of archaeology’s sister disciplines. In Schiffer, M.B. (ed.) Advances in Archaeological Method and Theory, volume 4. New York: Academic Press, 365–438. Gilbert, A.S., Singer, B.H. & Perkins, D. 1982. Quantification experiments on computer-simulated faunal collections. Ossa 8: 79–94. Gimbutas, M. (ed.) 1976: Neolithic Macedonia as Reflected by Excavation at Anza, Southeast Yugoslavia. Los Angeles: The University of California, Los Angeles (Monumenta Archaeologica, vol. 1). Grayson, D.K. 1984: Quantitative Zooarchaeology. New York: Academic Press. Greenfield, H.J. 1986: The Paleoeconomy of the Central Balkans (Serbia): A Zooarchaeological Perspective on the Late Neolithic and Bronze Age (4500–1000 BC). BAR International Series 304. Oxford: British Archaeological Reports. — 1988a: Bone consumption by pigs in a contemporary Serbian village: implications for the interpretation of prehistoric faunal assemblages. Journal of Field Archaeology 15: 473–479. — 1988b: The origins of milk and wool production in the Old World: a zooarchaeological perspective from the central Balkans. Current Anthropology 22: 573–593. — 1991: Fauna from the Late Neolithic of the central Balkans: issues in subsistence and land use. Journal of Field Archaeology 18: 161–186. — 1993: Zooarchaeology, taphonomy, and the origins of food production in the central Balkans. In Jamieson, R.W., Abonyi, S. & Mirau, N.A. (eds) Culture and Environment: A Fragile Coexistence. Proceedings of the 24th Chacmool Conference. Calgary: University of Calgary Archaeological Association, 111–117. — 1994: Faunal remains from the Early Neolithic Starčevo settlement at Bukovačka Česma, Starinar n.s. 43/44: 103–114. — 2000: Integrating surface and subsurface reconnaissance data in the study of stratigraphically complex sites: Blagotin, Serbia. Geoarchaeology 15: 167–201. — n.d. a: ‘Vertebrate Fauna from Blagotin’. Unpublished manuscript on file, University of Manitoba, Department of Anthropology. — n.d. b: ‘Vertebrate Fauna from Foeni-Salaş’. Unpublished manuscript on file, University of Manitoba, Department of Anthropology.

1

113

The Iron Gates in Prehistory Greenfield, H.J. & Jongsma, T.L. n.d.: Early Neolithic Blagotin: a summary of recent research. In Bailey, D.W., Whittle, A. & Cummins, V. (eds) Sedentism in the Central and East European Neolithic. Oxford: Oxbow Books. Greenfield, H.J., Jongsma,T.L. & Jezik, S. n.d.: Sedentary pastoral gatherers in the Early Neolithic — architectural, botanical, and zoological evidence for mobile economies from Foeni-Salaş, SW Romania. In Bailey, D.W., Whittle, A. & Cummins, V. (eds) Sedentism in the Central and East European Neolithic. Oxford: Oxbow Books. Halstead, P. 1989: Like rising damp? An ecological approach to the spread of farming in south east and central Europe. In Milles, A., Williams, D. & Gardner, N. (eds) The Beginnings of Agriculture. BAR International Series 406. Oxford: B.A.R., 23–53. Hesse, B. & Wapnish, P. 1985: Animal Bone Archaeology: From Objectives to Analysis. Washington D.C.: Taraxacum Press. Jongsma, T. & Greenfield, H.J. 2001: Architectural technology and the spread of early agricultural societies in temperate southeastern Europe. In Tupakka, S., Gillespie, J. & de Mille, C. (eds) Untrampled Ground — Untrammelled Views: Human Exploitation of, and Settlement Patterns on, New Landscapes. Proceedings of the 31st Annual Chacmool Conference, 1998. Calgary: University of Calgary, Department of Archaeology, 181–200. Klaić, B. 1982: Rjecnik Stranih Rijeci. Zagreb: Naklodni Zavod Matice Hrvatske. Koster, H. & Chang, C. 1986: Beyond bones: toward an archaeology of pastoralism. In Schiffer, M.B. (ed.) Advances in Archaeological Method and Theory, volume 9. Orlando: Academic Press, 97–148. Lazić, M. 1988: Fauna of mammals from the Neolithic settlements in Serbia. In Srejović, D. (ed.) The Neolithic of Serbia. Belgrade: University of Belgrade, 24–38. Leković, V. 1985: The Starčevo mortuary practices — new perspectives. Godišnjak (Sarajevo: Centar za Balkanaloška Istraživanja) 15: 157–172. Lyman, R.L. 1994: Vertebrate Taphonomy. Cambridge: Cambridge University Press. Lyneis, M.M. 1988: Antler and bone artifacts from Divostin. In McPherron, A. & Srejović, D. (eds) Divostin and the Neolithic of Central Serbia. Pittsburgh: University of Pittsburgh, Department of Anthropology (Ethnology Monographs no. 10), 301–319. Manson, J.L. 1990: A Reanalysis of Starčevo Culture Ceramics: Implications for Neolithic Development in the Balkans. Unpublished Ph.D. dissertation. Southern Illinois University, Carbondale, IL. Meadow, R. 1978: Effects of context on the interpretation of faunal remains. A case study. In Meadow, R. & Zeder, M.A. (eds) Approaches to Faunal Analysis in the Middle East. Cambridge, MA: Harvard University Peabody Museum Bulletin no. 2, 15–21.

Milojčić, V. 1949: Chronologie der Jungeren Steinzeit Mittel- und Südosteuropas. Berlin: Gebruder Mann. Murray, J. 1970: The First European Agriculture: A Study of the Osteological and Botanical Evidence until 2000 B.C. Edinburgh: Edinburgh University Press. Naval Intelligence Division. 1944: Jugoslavia. Geographical Handbook Series, vol. 1. London: Naval Intelligence Division. Payne, S. 1972: Partial recovery and sample bias: the results of some sieving experiments, In Higgs, E.S. (ed.) Papers in Economic Prehistory. Cambridge: Cambridge University Press, 49–64. Pounds, N. 1969: Eastern Europe. Chicago: Chicago University Press. Price, T.D. 1987: The Mesolithic of western Europe. Journal of World Prehistory 1: 225–306. Rimantiene, R. 1992: The Neolithic of the eastern Baltic. Journal of World Prehistory 6: 97–143. Schwartz, C. 1976: The vertebrate fauna from Rug Bair. In Gimbutas, M. (ed.) Neolithic Macedonia as Reflected by Excavation at Anza, Southeast Yugoslavia. Los Angeles: The University of California, Los Angeles (Monumenta Archaeologica, vol. 1), 364–374. Srejović, D. 1979: Protoneolit kultura Lepenskog vira. In Basler, D., Benac, A., Gabrovec, S., Garašanin, M., Tasić, N., Cović, B. & Vinski-Gasparini, K. (eds) Praistorija Jugoslavenskih Zemalja II: Neolitsko Doba. Sarajevo: Akademija Nauka i Umjetnosti Bosne i Hercegovine, 33–78. Srejović, D. & Letica, Z. (eds) 1978: Vlasac: A Mesolithic Settlement in the Iron Gates, 2 volumes. Belgrade: Serbian Academy of Arts and Sciences. Tringham, R. 1971: Hunters, Fishers and Farmers of Eastern Europe, 6000–3000 B.C. London: Hutchinson University Library. Turnock, D. 1989: The Human Geography of Eastern Europe. London: Routledge. Whittle, A. 1985: Neolithic Europe: A Survey. Cambridge: Cambridge University Press. y’Edynak, G. 1978: Culture, diet and dental reduction in Mesolithic forager-fishers of Yugoslavia. Current Anthropology 19: 616–618. Živanović, S. 1975: A note on the anthropological characteristics of the Padina population. Zeitschrift für Morphologie und Anthropologie 66: 161–175. — 1976: Cromagnon in the Iron Gate Gorge of the Danube. Nature 260: 518. — 1979: Further evidence on Cro-Magnon in the Iron Gate Gorge of the Danube. Current Anthropology 20: 805–806. Zvelebil, M. & Dolukhanov, P.M. 1991: The transition to farming in eastern and northern Europe. Journal of World Prehistory 5: 233–278.

114

Section 2

SITE STUDIES

Lepenski Vir animal bones: what was left in the houses? Vesna Dimitrijević

Abstract: A collection of animal bones from the eponymous site of the Lepenski Vir culture in the Iron Gates is stored in the National Museum in Belgrade. Formerly, four units from this collection were described, two of them related to the houses, another two to the areas outside houses (Dimitrijević 2000). In this paper, faunal assemblages collected from the houses are analyzed. By analyzing contextual units with animal bones that come from the houses an attempt is being made to understand activities of the house dwellers and significant aspects of their relations to animals. A distinction is made between the remains related to the lifespan of a particular house, and those the inhabitants left, intentionally, or unintentionally, behind them at the time of house abandonment. The animal species composition and skeletal parts distribution are presented. The distribution of dog bones inside houses, the evidence of food preparation and consumption, skinning procedure, bone tool use and manufacturing, as well as the circumstances of the deposition and symbolic meaning of red deer antlers in the houses of Lepenski Vir are discussed. Key words: Lepenski Vir, animal bones, Mesolithic, Early Neolithic, house abandonment event, red deer antlers, dog

Introduction

What was left in the houses?

The eponymous site of the Lepenski Vir culture in the Iron Gates was excavated from 1965 to 1971 by Dragoslav Srejović. Over the excavated area of 2500 m2, a 3.5 m-thick cultural layer with a sequence of building horizons of Mesolithic and Early Neolithic settlements was identified (Srejović 1966, 1969; Radovanović 1996). Radiometric dates associated with the houses indicate a time span from c. 6200 to 5500 cal BC (Quitta 1975; Whittle et al. 2002). The excavations were undertaken as a part of the vast Ðerdap (Iron Gates) I rescue project in advance of the construction of the dam near Kladovo. The time constraints imposed by the construction project influenced the excavation strategy. Collecting animal remains was not among the priorities. Consequently, only a small, hand-collected, sample of animal bones was saved. Animal remains from the earlier excavation campaigns were analyzed and published in a preliminary report (Bökönyi 1969). These were a sample of the bones recovered and comprised 2999 identified specimens. A selection of this sample, as well as animal remains collected in subsequent campaigns, are stored in the National Museum in Belgrade. The animal bones curated in the National Museum in Belgrade are packed in bags, with field data labels designating their origin. There are two main types of contextual unit with osteological material — units related to structures, and those related to various site deposits but unconnected with the architectural structures. The latter are recorded by arbitrary excavation layers and their spatial position across the site. Those units related to the architectural structures mainly include the designations: ‘house’, ‘on the floor of the house’, ‘below the floor of the house’, ‘between the floors of the houses’, ‘between the houses’, ‘hearth’, and ‘grave’ written on the bag labels. This paper focuses on the contextual units that were related exclusively to the houses.

By analyzing contextual units with animal bones from the houses, we hope to approach the life of the house dwellers and important aspects of their relations to animals. We shall try to differentiate between those remains that are related to the lifespan of a particular house, and those which dwellers of the houses left, intentionally or unintentionally, behind them, when abandoning a house. It is necessary to remain aware of numerous obstacles posed by excavation technique, stratigraphical problems, and the nature of the archaeological record itself. The answer to the question in the title of the paper is thus complex and ‘multi-layered’. In the course of the excavations at Lepenski Vir not all the bones were collected. This fact can best be affirmed by the quantity of collected bones in relation to the large excavated area. Bones were not collected from every house area. If nothing else, some animal bone remains can be seen on the photographs taken during the excavations that were neither collected nor curated. Within single houses, not all the bones found were gathered, also evident in unsaved material seen on published photographs. A large part of the osteological material analyzed by Bökönyi was discarded after his analyses, and only a small sample was curated in the collection of the National Museum. Furthermore, we have to remain aware that not a single contextual unit of collected animal bones from Lepenski Vir represents a complete recovery of the remains. Also, field labels designating the contextual origin of animal bones in relation to structures are not very precise. Many labels indicate the number of a house only. It may be supposed that these are bones that were found on the floor of a house. However, it is difficult to be sure that these remains do not relate to a layer above the floor that formed later, which is certainly very relevant for their interpretation. In this re-

117

The Iron Gates in Prehistory

Figure 1. Plan of the Lepenski Vir I houses (from Radovanović 1996).

spect, the units related to superimposed houses are particularly problematic, with designations ‘between houses’ or ‘between house floors’ on the labels. It remains uncertain whether the bones collected from these units relate to the floor of the earlier house, or the foundation of the later house. Even in those instances when field data confirm that bones were found lying directly on the floor, in what capacity can we relate that specific assemblage of bones to the life of the house? In this respect, osteological evidence may often help us to differentiate between what may be a heterogeneous assemblage, coming from a mixture of animal bones of the layer deposited over an assemblage that represents a meaningful or structured deposition within a house. During the life of a house, animals, their carcasses or parts of the carcasses, or even bones themselves, would be present on particular occasions. Also, sometimes these remains would have particular meaning and would merit a specific explanation. Reasons for the presence of animal remains inside the houses may relate to a wide range of possibilities, from the creatures which shared the space of a house with humans, carcass parts as food that was consumed or left inside the house, accumulated food debris, and bone artefacts or raw material discarded in the process of tool manufacturing. However, bones left by similar activities could be found, and commonly are found, outside houses; while in the process of burying a house the bones may be deposited on its

floors or in the layers above the floor. It may be supposed that an abandoned house was either suddenly covered with a layer of sediment when the base for the overlying house was being constructed, or slowly buried over a period of time. In either case, sediments containing animal remains would overlie the house. Being peculiar and extraordinary in many respects, the houses of Lepenski Vir seem to contain in several instances animal remains — offerings, that were intentionally left by their occupants. The circumstances of these findings will be emphasized in the following discussion. Other assemblages of animal bones that can be reliably related to the houses will also be described. The faunal assemblages are presented according to the groups of superimposed houses and their locations as defined by Radovanović (1996) (Fig. 1), followed by a description of assemblages from non-superimposed houses. Any additional data found on field labels apart from house numbers is mentioned. Only mammalian remains are analyzed, while bird, fish and invertebrate remains are noted. Remains of the following species were found in the houses of Lepenski Vir: Erinaceus concolor Martin (hedgehog), Castor fiber Linnæus (beaver), Canis familiaris Linnæus (dog), Ursus arctos Linnæus (brown bear), Martes martes (Linnæus) (pine marten), Meles meles (Linnæus) (badger), Sus scrofa Linnæus (wild boar), Cervus elaphus Linnæus (red deer), Capreolus capreolus (Linnæus) (roe deer), Bos primigenius Bojanus (aurochs), Rupicapra rupicapra

118

Vesna Dimitrijević: Lepenski Vir animal bones

5 6 7

11 13 14

6(2)

+

30

1(1)

+

22

3(3)*

47

1(1)

36

1(1)

35

1(1)

7(2)

1(1)

1(1)

+

1(1) +

3(1)

5(2)

1(1)

+

+

+

+

+

+

+

1(1)

37

1(1) 1(1)

34

2(2)

27

1(1)

1(1)

18 23

molluscs

fish

birds

Capreolus capreolus

8(1)

4(1) 1(1)

1(1)

12(2)

1(1)

2(2)

2(1)

1(1)

46

1(1)*

20

1(1)

32

1(1)

2(1)

+

26'

2(1)

1(1)

+

+

26

5(1)

+

+

1(1)

+

+

1(1)

+

25

1(1)

3(2)

1(1) 2(1)

1(1)

10(2)

44

2(1)

48

2(2)

1(1)

2(1)

1(1)

49

1(1) 1(1)

61 62

+

+

+

+

+

+

+

+

1(1)* 2(1)

65

1(1)

69 70

+

1(1)

40

54

3(2)

1(1)

28 non-superimposed houses

Cervus elaphus 17(1)

42 10

Rupicapra rupicapra

7(1)

38 8

Bos primigenius

Sus scrofa

Meles meles

Martes martes

13

Ursus arctos

4

Canis familiaris

9

Castor fiber

House

2

Erinaceus concolor

Location

Table 1. Animal remains found within the houses of Lepenski Vir: number of specimens, (minimum number of individuals), + presence confirmed, * supplemented from the literature.

2(1)

1(1) 1(1)

3(1)

71

3(1)

+ +

119

+

The Iron Gates in Prehistory

(Linnæus) (chamois), birds, fish and molluscs (Table 1).

Superimposed houses

Figure 2. Dog lumbar vertebrae from house 9: a. first or second lumbar vertebra, arrow pointing to butchering marks, b. the last or next to last lumbar vertebra, arrow pointing to exostosis.

Location 2, houses 17, 8, 7 and 9 Bones originate from house 9. Some bones bear traces of fire with varying degrees of carbonization/calcification. Remains of dog and red deer are present, and numerous smaller fragments of large mammal bones. The red deer remains consist of a fragmented tine, many smaller antler fragments, and the fragment of a second phalanx. All pieces are carbonized. The dog remains comprise a carbonized fragment of lower jaw, with the first premolar and preserved alveoli for the second and the third premolar, an isolated third lower incisor, two rib fragments, and a calcified metapodial fragment. In addition, there are two dog lumbar vertebrae (Fig. 2), with no traces of fire. The damage on their processes is recent. One of them, probably the first or the second in the lumbar row (Fig. 2a), bears butchery marks. Another, probably the last or next to the last in the lumbar row (Fig. 2b), shows a pathological feature — exostosis on the ventral side of the fusion between the caput and the body of the vertebra, indicating an unhealthy condition of the animal, possibly related to old age. The assemblage contains red deer remains that could relate to tool manufacturing and dog remains possibly originating from one or two skeletons. Dogs were probably eaten. Location 4, houses13, 14 and 15 Animal bones come from the earliest house at this location, house 13. Red deer and wild boar remains were present, while several identifiable fragments are of large mammals, most likely these two identified species. The red deer remains comprise two antler fragments, a fragment of the occipital part of the skull, two fragments of thoracic vertebrae, a fragment of a lumbar vertebra, a fragment of a metatarsal shaft, and a carpal bone. Wild boar is represented by an upper third molar, fragments of a proximal humerus and a humerus shaft, a proximal ulna, a distal scapula, and the first phalanx. The red deer antler fragments bear marks made by artefacts. The wild boar remains could belong to at least two animals, on the basis of the heavily worn third upper molar of an aged animal, in contrast to the scapula which comes from a young animal. Two humeral fragments bear butchery marks. The assemblage is heterogeneous: the bones may originate from carcass parts that were butchered for food consumption, and skeleton parts used as raw materials.

Figure 3. Right red deer antler from house 22.

Location 5, houses 30, 29, 22 and 21 At this location animal bones come from the floor of the earliest house in the group, house 30, and also from house 22, placed above house 29, and below house 21. On the floor of house 30, a fragment of a red deer ilium corpus fragment was found, with butchery marks on its dorsal side. Large fish bones and teeth, and mollusc shells were present too. An interesting situation is attested on the floor of house 22. Two pairs of red deer antlers were arranged here. First, antlers of a young red deer attached to the skull were found beside an extended human skeleton — burial 7 (Srejović & Babović 1983: 18, upper right). This burial was cut through 120

Vesna Dimitrijević: Lepenski Vir animal bones

Figure 4. Carbonized scapulae from house 47: a. wild boar right scapula; b. roe deer right scapula; c. red deer right scapula.

the floor of house 21, which partly covered house 22 in its rear part. Both the skeleton and red deer antlers were practically lying on the floor of house 22. Although the skeleton and antlers are seemingly in connection, as shown on the published photo, one of the antlers is just beneath and beside the stone construction of house 21, while the other one is stretching further in the direction below the floor of house 21. It seems more probable that it was placed on the floor of house 22, before house 21 was built and less probable that the pit cut for burial 7 was of a width to permit the antlers to be enclosed in the grave. Another pair of red deer antlers, together with the skull can be clearly seen on a photograph (Srejović & Babović 1983: 136, lower) beside the hearth of house 22, lying directly on the floor. The right (Fig. 3) and left antlers are preserved in the collection of the National Museum and correspond in size and shape to those seen on the photograph, but the skull was not saved. Most of their original length is preserved; the surviving length of the left antler is 560 mm, and of the right antler 630 mm. Brow tines, bez tines and trez tines are in analogous positions and the antlers are of similar size. There is a crown on the right antler with two tines that lack their

tips. The circumference of the right antler burr is 225 mm, and its greater diameter 70 mm. The size of the antlers, number of tines, and branching of the crown on the right antler, point to an animal at least 5 years old; this is indicated especially by the presence of the bez tine, which first appears at that age, but could also appear when an individual is six years old (Dragišić 1957). Location 6, houses 58, 53, 47' and 47 The preserved animal bones come from the earliest house at this location, house 47, more specifically, in its west (i.e. narrow) side, as recorded on the field label. There are three right scapulae of three different species: red deer, wild boar and roe deer (Fig. 4). All three were carbonized and fragmented in a similar way: the glenoid part was preserved together with a smaller portion of the blade above the scapula neck. Most likely these scapulae indicate remains of meat consumption. Location 7, houses 36 and 35 In the hearth of house 36 the third phalanx of a wild boar and fish remains were found.

121

The Iron Gates in Prehistory

Figure 5. Dog mandibles from house 34: a. left mandible of an adult animal; b. right mandible of a puppy.

In the later of the two houses, house 35, a large sample of mammal bones was collected, consisting of 96 fragments of various skeletal parts that can be attributed to several animal species. Bird, fish and mollusc remains were collected, too. Red deer remains consist of a tip fragment of an antler tine, an upper fourth premolar, fragments of the left and right pelvis, and a fragment of an unfused radius distal epiphysis. The aforementioned fragment of left pelvis and the fragment of a radius distal epiphysis are gnawed. The antler fragment bears many scars made by a blunt artifact. These bones belong to at least one adult animal, and one young individual whose body was still growing. The wild boar is represented by a fragment of cranium, a fragment of scapula blade that bears gnawing marks, and the third phalanx, all of them attributable to a juvenile animal. The dog is represented by two rib fragments, a proximal ulna with gnawing pits at the olecranon, a proximal metacarpal fragment, and the first phalanx. Two bones of a neonate, a humerus and a radius, belong to a puppy. Also, there are a carpal bone (intermedial) of a roe deer, a marten’s lower jaw, and a beaver’s second phalanx. Two fragmented artefacts were also found, one made on a fragment of a large mammal long bone, the other on a medium size mammal’s long bone. The assemblage in house 35 is heterogeneous, i.e. it consists of species whose body parts were utilized in different ways by humans. Yet, some remains, such as the neonate dog humerus and radius, and the bones of possible juvenile pig,

may originate from the same animal and would imply that their disposal was contemporaneous. It should be noted that a burial was discovered at this location (Radovanović 1996: 185; Stefanović & Borić, this volume: note 6). This is burial 70, described as the secondary disposal of human bones with a dog mandible (Radovanović 1999). Location 8, houses 41, 38, 37 and 42 In house 38 a fragment of a carbonized aurochs scapula was found. In house 37, which cuts one corner of house 38, a dog femur shaft fragment came to light. In house 42 a cattle lower molar was found. The shaft fragment of the dog femur from house 37 bears many short, irregular cuts, probably from the procedure of disarticulating the skeleton, while both ends are gnawed. The cattle lower molar from house 42, the first or the second in the molar row, is of a size not distinctive for the aurochs/domestic cattle separation. Location 10, houses 34, 27, 51 and 43 Analyzed animal bones originate from the earliest house in this group of houses, house 34, and the latest, house 27. Two lower jaws of dogs were found in house 34, a red deer antler fragment, and a metapodial fragment of a small ruminant. The lower jaws of dogs are of an adult dog and a puppy (Fig. 5). The lower jaw of the adult animal (Fig. 5a) has all teeth present in the alveoli except for the first incisor and the last molar. Incisors and canine crowns are broken,

122

Vesna Dimitrijević: Lepenski Vir animal bones

while the tips of the first three premolars are without enamel. Longitudinal cracks are visible on the bone surface of the corpus of the jaw. The jaw was not in direct contact with fire; however, it suffered from substantial heating, which is evident from the reddish colouring of the basal part of the corpus of the jaw, with cracking of the bone and ‘peeling’ of enamel as a consequence. The lower jaw of the puppy (Fig. 5b) has all three milk molars present in the alveoli. There are gnawing marks on the vertical ramus processes, which may have been made by a small animal. In the incisive part, the first milk incisor alveolus is present; also fragments of the broken second and third milk incisors as well as milk canines are in the alveoli. The molars are completely unworn, while behind the last milk molar, the crypt is opened and the permanent molar germ may be seen in the process of developing. The milk dentition in dogs is completed by the age of 5–6 weeks, while already at the age of 3–4 months, the crowns of milk molar teeth show considerable wear of prominent cusps and ridges; permanent teeth start appearing at this time, too (Habermehl 1975). Accordingly, the mandible from house 34 is of an animal less than three months old. Another two items found in house 34 are probably fragments discarded in the process of tool manufacturing. A fragment of a red deer antler bears pit-like depressions and short grooves scattered over its surface, and traces made by some artefact at the place where probably a tine was cut off from the main branch. Another fragment originates from the metapodial of a small ruminant, possibly roe deer. The surface of a proximal joint is present on this fragment with the shaft piece of 75 mm in length. There are traces made by an artefact on the lower part of the shaft fragment: two halfcircular incisions, and two deep longitudinal cuts above the upper incision. Obviously the bone was split longitudinally, with possibly a tool manufactured out of the bone, while this surviving fragment may have been a by-product. A beaver lower incisor was found in house 27. Its aboral part is broken. There is a natural lingual wear-mark, but in the continuation behind it there is a polished surface, probably intentionally modified. Location 11, houses 19, 31, 18 and 23 Bones come from two later, superimposed houses in this group, houses 18 and 23. House 23 partly covers house 18 in its rear end. The bag label indicates a more precise contextual position for the bones found in house 18 — it says ‘below corner A’. Four bear vertebrae were found here: a thoracic vertebra, probably the last one in the thoracic row (Th 14), and three lumbar vertebrae, probably the first, the second and the fourth. These vertebrae are from the same individual and were probably buried together. This may be attested by the same coloration, fossilization pattern, and similar age shown by the unfused cranial and caudal vertebral body epiphyses. There is a single filleting mark — longitudinal and sharp, cut in the base of the neural spine of the first lumbar vertebra. In house 23, the first lower molar of a dog was found. It is a tooth crown, with roots incompletely developed, indicating that the molar was in the process of erupting; the animal was 3.5–5 months old (Habermehl 1975). The sample collected ‘between houses 18 and 23’, as noted

on the bag label, consists of mammal bones, large fish bones, fish teeth, and molluscs. Among the mammals dog remains are the most numerous. Dogs are represented by a lower jaw, fragments of an upper jaw, an isolated canine, an isolated molar, an epistropheus, two left ulnae fragments, a fragment of radius, a fragment of pelvis and two metacarpals (the second and the fourth). The olecrani of the ulnae are gnawed, as well as one of the metacarpals. The metacarpals are of the same individual, while two left ulnae confirm the presence of at least two dogs. Remains of other species include the lower jaw of a hedgehog, a lower incisor fragment of a beaver, a proximal tibia of a marten, gnawed phalanxes of a wild boar (an adult first phalanx and a juvenile second), a third phalanx and a tooth fragment of an aurochs, a partly burnt thoracic vertebra body of a red deer (with multiple longitudinal and diagonal cuts made by a sharp artefact), and one fragmented artefact made on a small ruminant metapodial. The bear vertebrae ‘below corner A’ in house 18 are perhaps related to the event of house construction. The assemblage ‘between houses 18 and 23’ consists of a large number of species and different skeletal parts, obviously discarded as a consequence of different activities (meat consumption, tool manufacturing, possibly skinning). However, some bones were probably deposited in a single event (most likely the dog bones and perhaps the wild boar phalanxes), and these, if not the whole assemblage, might have been related to house 18. Location 13, houses 55, 56 and 46 Deer antlers attached to the fragmented skull were placed over the floor of house 46. This is shown on the published photograph of the hearth area of this house (Radovanović 1996: fig. 3.33). Location 14, houses 20, 33 and 32 Bones were collected between houses 20 and 33, as noted on the bag label, and from the floor of house 32. ‘Between houses 20 and 33’ a red deer lower molar was found, two left astragali of red deer, two right astragali and a metacarpal shaft fragment of roe deer, a fragmented proximal ulna of a dog, the artificially modified thoracic vertebra of a wild boar (Fig. 6), as well as bird bones, many fish bones and mollusc shells. On the floor of house 32, remains of a dog (an upper canine), red deer (lower second incisor and a proximal metatarsal), and fish were found.

Non-superimposed houses House 25 Long bones of small carnivores, which can be characterized as fur-bearing animals, were collected in this house. The following bones are present: a femur of a badger, and two long bones of a marten (a femur and a humerus). The femurs of both species bear cuts made by a sharp artefact. The beaver’s femur shows two subparallel arched cuts on the cranial side of the bone, below the proximal articulation. The marten’s femur bears very narrow, longitudinal incisions on the shaft.

123

The Iron Gates in Prehistory

Figure 6. Wild boar modified thoracic vertebra related to house 20: a. cranial view, arrow points to a groove made in the vertebra body; b. lateral view — the position of an imaginable cord is shown, the arrow goes through the arch foramen.

House 26' A tine of a red deer antler that was cut off at the base and remains of a dog (the first upper molar and distal femur) were found in house 26'. Fish remains and shells were also collected, as well as many small fragments of unidentifable mammal bones. House 26 A red deer antler on the pedicle and a fragment of a calvarium were found in ‘house 26, beside the sanctuary’, as noted on the bag label. Brow, bez and trez tines are cut off at their bases, while the beam was broken above the trez tine, probably during the excavation. This cutting off of the tines was patiently performed, by gradual thinning, as indicated by flaking in the direction of the tine, and circumferential furrows above them. Also, two fragmented tines are present, with analogous traces of cutting off from the beam. On the skull fragment, marks made by blows designed to separate the antler from the rest of the skull are noticeable on the anterior basis of the pedicle. There are also multiple, deep, elongated cuts in the area where the pedicles of the two antlers were closest to one another, indicating skinning. Dog remains were also found: an epistropheus, an upper canine, an upper first molar, a distal femur and a distal tibia, all possibly from the same individual. House 28 Antlers attached to the skull of a young deer were found on the floor of house 28 (Fig. 2). The skull was damaged when it was lifted from the position where it was laying to such an

extent that it was not possible to reconstruct it. The frontal, temporal and occipital parts of the skull were present, together with the left and right upper jaws with tooth rows. The frontal, temporal and occipital parts of the skull were present, together with the left and right upper jaws with tooth rows, while the bones of the face, the nasals and praemaxillars, were crushed (Dimitrijević 2000). Fish and molluscan remains were deposited in the same place, evidenced by fish bones and a gastropod shell fragment attached to the crushed skull bones. The antlers are preserved for a length of over 50 centimetres. They are asymmetrical: the left one has a brow tine and simple crown with two tines, while the right antler consists of a single branch, with just a slight protuberance at the site of the brow tine. Although variability in red deer antlers is well known (Dragišić 1957), antlers from house 28 should be characterized as unusual in their length and asymmetry. Due to their age they are extremely thin (right burr circumference =126 mm, left burr circumference = 123 mm, right beam circumference 100 mm from the burr = 80 mm; the same measurement in the left beam = 76 mm). The animal’s age is clearly indicated by its dentition, which is in the last stage of milk/permanent tooth replacement and with the last permanent teeth erupting. Both third milk molars still rise above the crowns of permanent fourth premolars, although half of the crown of the one on the right side of the jaw is worn; it would have been a matter of days before they fell out. The second and third premolars are erupting, as well as the third molar, preserved only in the left jaw. This stage of dental development in the upper jaw should be analogous to

124

Vesna Dimitrijević: Lepenski Vir animal bones

that in the lower jaw, and related to an age of 27 months (Brown & Chapman 1991). Since deer give birth in May/June (Bützler 1986), this means that the animal was hunted in the early autumn, most probably in late September or early October (Dimitrijević 2000).

Figure 7 Young red deer antlers from house 28, frontal view.

Figure 8 Wild boar left astragalus, calcaneus and distal fibula in articulation from house 40: a. dorsal view; b. plantar view, arrows point to the breakage of the fibula and cut marks at the distal end of the calcaneus.

House 40 Animal bones were found on the house floor, along its longer right wall (corners B–C). There were wild boar remains (fragment of an atlas, final lumbar vertebra, left scapula, fragments of right and left pelvis, right proximal and left distal radius, left ulna, astragalus and calcaneus (Fig. 8), distal fibula and tibia, left distal femur with unclosed epiphyseal lines (indicating a younger animal), together with a fragment of a sacrum and gnawed distal radius of red deer, and an aurochs’ scapula. The wild boar bones are not just fragments, but complete or almost complete bones of the cranial and sacral parts of the backbone, parts of shoulder and pelvic girdles, and mostly the upper parts of the fore and hind legs. None of the skeletal elements is represented by more than a single specimen, however, the assumption was made that the remains possibly come from two animals. This is largely based on the distal femur with the epiphyseal line of distal articulation, clearly observable along its whole length, which would belong to an animal that may be aged as almost adult, but not fully grown. In contrast, all other wild boar bones are fully grown, and of the size and proportions corresponding to a large adult male individual. The tarsal bones, astragalus and calcaneus, together with the distal fibula, remained in articulation (Fig. 8a–b). In addition, the distal tibia and fibula’s shaft were found to fit perfectly the astragalus and the distal fragment of the fibula, indicating that those bones were a continuation of the joint described. There are no traces of synarthrosis, which would indicate pathological fusing, so it is certain that they were buried in situ as articulated, and not removed from the site afterwards, when the organic tissue would have decomposed. The quality of the deposit in which they were embedded, probably enriched by carbonates, made their connection firm even after their excavation. The final lumbar vertebra, as well as the left and right pelvic girdle fragments, were attached to the sacrum, while two long bones of the left front leg, the ulna and radius, were also in articulation. Another important observation regarding this part of the skeleton is that the ulna is complete, except for some damage to its central portion, while the upper part of the radius is broken. The breakage of both bones was caused by a strong blow, and probably happened while the tendons, and probably also muscular tissue, were still holding the bones together. The single fragment of long bones in the right front leg, the proximal radius, points to a similar breakage, a consequence of the patterned manipulation of the carcass parts. Butchery marks are clearly observable on the tarsal bones, in the form of several deep transversal grooves on the calcaneus (Fig. 8b), and a few short grooves on the astragalus made in the same direction, and by the same action indicating disarticulation of the lower hind leg. There are no cut marks on other bones, but there are indications that other butchery techniques apart from cutting might have been applied, such 125

The Iron Gates in Prehistory

as a blow to the middle of the radius and ulna shafts. The left and right pelvis fragments attached to the sacrum were probably broken in the course of the disarticulation of the left and right flanks. While the primary butchery was performed, it seems that the butchering process was not continued to the point of completely stripping the meat from the bones. There are no traces of filleting, which would be very difficult if not impossible to perform, by avoiding the contact of sharp artefacts with the bone, which is necessary for this operation. This is especially true of the shoulder blade, on which filleting leaves characteristic long, longitudinal grooves (Binford 1981: fig. 4.06). It is suggested that large carcass parts of a wild boar were left in the house (Dimitrijević 2000). House 44 A diaphysis shaft of a dog humerus and a fragment of a gnawed proximal ulna of a dog, a third molar of a wild boar, and a fragmented lower jaw of a roe deer were collected in house 44. House 48 Two dog lower jaws were found in house 48, one left, and one right, but certainly from two different individuals, since wear traces are more advanced on one of them. House 49 An upper jaw of a roe deer and a red deer vertebra fragment originate from house 48. Fish remains were also present. House 54 A dog’s lower carnassial tooth and many fish remains (mostly teeth) and molluscs were collected in house 54. House 61 A deer’s skull with antlers is seen on the published photograph of this house (Srejović & Babović 1983: 151, lower). The skull is damaged, and antlers are partly behind a sculpture and stones, so their morphology is not very clear, although it is possible to say that they belonged to an adult animal. House 62 A radius and tibia shaft fragments of a dog, both with traces of gnawing, come from house 62. House 65 An antler fragment and a splinter of a metacarpal shaft of a red deer, as well as a partly carbonized horn-core of a chamois were found in house 65, together with fish teeth and a sea snail. House 69 According to the bag labels, a second metatarsal of a wild boar was found behind the hearth, while fish remains and mollusc shells come from the hearth of house 69. House 70 In the hearth of house 70, mammal and fish bones were

found. The following mammal remains were found: an atlas fragment of a dog, and several remains of red deer and chamois. A fragmented epistropheus, a distal unfused epiphysis of a radius, and a second phalanx are of red deer, and chamois is represented by a tip fragment of a horn-core with traces of manufacturing or use, and a complete second phalanx. A fragment of a distal metapodial of a ruminant could also belong to chamois. House 71 Fish remains and a single bird bone were recovered from house 71.

Interpretation of animal bone remains in the houses of Lepenski Vir Lepenski Vir houses by their size, solid floor construction, and elaborately built hearth constructions might have been multifunctional dwellings, where many activities were performed regularly, being highly organized. Food preparation and consumption, manufacturing and use of bone tools are among those regular activities we would hope to find as evidence among the studied animal remains collected in the houses. Stone sculptures, altars, A-signs, partial burials and other reflections of the highly developed symbolism of the Lepenski Vir culture indicate that apart from practical activities, those related to complex social, symbolic and cult practice might have taken place within the houses. Animal remains may show special ‘status’ of a particular species, and provide a material proof of an offering deposited in the house. The question remains, whether the archaeological record will reflect these activities, and how faithfully? For example, if a highly organized life was carried out in the house, piles of discarded animal bones left after the meal on the floor would be a less plausible outcome then regular cleaning of food debris. Of course, when a particular house was finally left, the remains of the last meal could have been cleaned off, or not, depending on the circumstances of departure or common habits of the dwellers. It is important to stress that, when we are dealing with finds that are certainly related to floors, we are looking upon the end of the house’s life, i.e. at the time of its abandonment. Intentional or forced, sudden or prepared departure from the house may be reflected in the animal bones left on a house floor. The composition of the animal species found in the houses, as well as the distribution of skeletal parts, may indicate some of these activities, too, if different from those unrelated to houses. In the following sections I shall discuss the pattern of the appearance of dog bones and other evidence about the presence of dogs in the houses, evidence of food preparation and consumption, skinning procedure, use and manufacturing of bone tools, and meaningful deposition of red deer antlers within the houses of Lepenski Vir. ?Dogs in houses Describing the faunal remains from Lepenski Vir, Bökönyi (1969) points to a high frequency of dog remains (4.95% in Lepenski Vir I, 11.6% in Lepenski Vir II, and 5.91% in

126

Vesna Dimitrijević: Lepenski Vir animal bones

Lepenski Vir III) and their undoubtedly domesticated character. Furthermore, he notes that dog bones are mostly unbroken and from adult animals. His observations that in the houses of Lepenski Vir I and II ‘undisturbed skeletons of body parts in anatomical arrangement are very often found’, and that ‘dog bones are often found together with red deer skull and scapula as well as wild boar bones’ are of special importance here. He does not deny the possibility that dogs were eaten, but supposes that their main role was as hunting companions and watch-dogs, and that dog was also used as a sacrificial animal (Bökönyi 1969). A special human attitude toward dogs is also revealed by human burials associated with dog remains. A grave with a secondary disposal of human bones and a dog mandible is described by Radovanović (1999). In her study of dog burials and human burials associated with dog remains in the Iron Gates Mesolithic, Radovanović supposes that a change occurred at the second half of the VII millennium BC, in a period coinciding with the first contact with Early Neolithic groups, when a symbolic meaning was associated with dogs, reflected in human burials with dog remains. At the same time, throughout the Iron Gates Mesolithic ‘dog remains scattered around the settlement show not only that they were eaten, but also that they were not paid any special respect (since their bones were left to be scavenged)’ (Radovanović 1999). It is questionable how deeply we can understand the complexity of human/dog relationship on the basis of the archaeological record in general, and especially in the light of ‘what was left in the houses’ of Lepenski Vir. However, osteological remains indicate that dogs were accorded special treatment. The abundance of dog remains among the bones collected in the houses of Lepenski Vir is striking. Only red deer remains are slightly more numerous in houses (Table 1). Often, remains of more than a single dog were found (houses 35, 34, 23, 48, possibly house 9), and often different parts of a skeleton that are certainly (houses 35 and 23) or most probably from the same individual (houses 9 and 26). This is in agreement with Bökönyi’s note of ‘undisturbed skeletons of body parts’. In four houses, lower jaws of dogs were found, which could be of a particular importance in light of their deposition as grave goods in two burials at Vlasac and burial 70 at Lepenski Vir (Radovanović 1999). There is a possibility that this skeletal part was chosen to represent the presence or the essence of the animal itself. Two mandibles were found in house 34, and in house 48, originating from two different animals respectively. One of those found in house 34 is of a puppy, while the bones of a newborn were found in house 35. That dogs were eaten is also highly probable. Bones with butchering marks were collected in houses 9 (Fig. 2) and 37. Cut marks on a dog vertebra from house 9 derive from the removal of the tenderloin, which is regularly found on the butchered vertebrae of meat-bearing animals (Binford 1981). Many dog bones found in the houses are gnawed. Although dog is not the sole animal that gnaws bones (some rodents do, as well as pigs and humans), this habit is by far most characteristic for canids. The size of the gnawing marks on the bones from houses of Lepenski Vir — pitting, grooves

and compressions made by canines, as well as the pattern of the epiphyses being gnawed in long bones strongly suggest that they are made by dogs. Gnawed red deer and wild boar bones are found in houses 35, 23, and 40. Dog bones gnawed by dogs are found in houses 35, 37, 34, 23, 44, and 62. A dog femur shaft fragment from house 37 with both ends gnawed, bears many short, irregular cuts indicating disarticulating, which means that only after butchering the animal did a dog get his share. There is something intriguing about the gnawed bones in the houses of Lepenski Vir. When we find animal bones within a house area, this does not mean that this animal was ever in the space of the house alive. However, the gnawed bones may suggest the opposite. Since there is really no reason for humans to gather gnawed bone refuse and bring it into a house, it would be reasonable to suppose that gnawed bones were brought into the houses by dogs! Parts of animal carcasses and remains of food consumption Generally, bones that bear good portions of meat, and originate from species that are of primary importance for meat consumption, may be presumed to have been largely disposed of as food remains. This presumption is further strengthened in those instances when bones are grouped, found in a meaningful context, and supported by the presence of butchery and filleting marks on bones. It is much more difficult to find proof of the disposal of parts of animal carcasses. Yet, it seems that this is the case with remains of wild boar in house 40, where evidence for primary butchery is present on the bones of lower extremities, related to dismemberment, but further operations necessary for separating individual bones and removing meat from the bones were not carried out. On this evidence, it may be concluded that carcass parts were left in the house, either as raw meat, or dried, or prepared for drying (Dimitrijević 2000). Another example of bones that bear large portions of meat comes from house 18. However, this time, signs of stripping meat from the bones are present. The carcass part comprised a thoracic-lumbar part of the backbone of a brown bear. A filleting cut mark was made on a single vertebra. Particularly intriguing is the exact provenance of this skeletal part ‘below corner A’ of the house, which signifies different reasons for its deposition from the offerings on the floors and within houses. This last instance possibly relates to the building of the house. An indication of the ‘offerings’ left in the houses can be found in Bökönyi’s note (1969) about seven scapulae of red deer found in the houses and ‘often together with wild boar bones and dog remains in anatomical position’. In the analyzed assemblages unbroken red deer scapulae were not found, while a single red deer scapula was found in house 47, together with the same skeletal part of a wild boar and a roe deer (Fig. 4). The scapula is among those parts of the skeleton that bear the most desirable meat portions. When the bone is preserved well enough, filleting marks, in the form of longitudinal cuts made by a sharp artefact, may confirm the stripping off of meat. Also, the fragmentation of a scapula discarded after meat consumption is rather characteristic: the distal, solid bone part may remain undamaged, while the rather thin blade

127

The Iron Gates in Prehistory

is regularly broken. In this respect, it is safe to conclude that the three scapulae from house 47 are remains of meat consumption, which took place in the house and were probably connected with a single event. The same degree of carbonization probably indicates that they were burnt at the same time. On the other hand, the carbonization itself need not be related to meat preparation; the bones might have been thrown in the fire after the meat was consumed, or as a consequence of house burning. The context where these were found, in the west (i.e. narrow) side of the house, does not relate them to the hearth. A carbonized distal scapula from an aurochs was found in house 38, and another was used as raw material in house 40. Apart from scapulae found in house 47, bones that originate from carcass parts, either those bones that carry good portions of meat or confirmed as butchered for consumption, were found in house 13 (red deer and wild boar), house 30 (red deer) and house 35 (red deer and wild boar). Skinning evidence Skinning is recognizable only by cut marks on the bones. The cut marks are not morphologically different from those made for achieving other goals, such as butchering. Yet, their position on certain bones is patterned (Binford 1981; NoeNygaard 1995). The presence of particular species, traditionally valued for their fur, is also indicative. Clear evidence of skinning is found on the red deer skull fragment with antlers in house 26. On the anterior basis of the pedicle, impact marks of blows are found, directed to separate the antler from the rest of the skull, while multiple deep elongated cuts stretch in the area where the pedicles of the two antlers were closest to each other. A similar skinning technique for deer has been recognized in the archaeological and ethnoarchaeological records (Binford 1981; Becker 1986: Abb. 41–43; Noe-Nygaard 1995: 183, 188). The presence of fur-bearing animals characterizes the assemblage of house 25. Although these species were probably eaten, too, their prime economic importance ought to be in connection with fur acquisition in particular. There is a single bone from a badger, a femur. Two sub-parallel arched cuts have been made on its cranial side below the proximal articulation. Of two bones of a marten, a femur and a humerus, only the femur bears cut marks. They are in the form of very narrow, longitudinal incisions on the shaft. Remains of badger and marten are considered as being hunted mainly for their fur at Vlasac and Padina, too, on the basis of the distribution of skeletal parts and skinning marks (Bökönyi 1978; Clason 1980). Bones of marten were also found in houses 23 and 35, in both cases associated with beaver remains, which is regarded as another important fur-bearing animal. Artefacts, by-products and bone and antler raw material We may expect to find artefacts in houses, since their use could be related to the life of a house, and houses are appropriate places where they could have been kept, or left and forgotten. In the assemblages described it was mostly fragmented artefacts and parts of bones rejected as useless for tool manufacturing that were found, since recognized artefacts were treated and packed separately during the

excavations. Evidence of tool manufacturing comes from house 26 — a red deer antler on the skull fragment, had all its tines cut off at the base. The tines were obviously successfully separated from the beam to be used probably as massive points. The work had been done laboriously, possibly in order to be continued at the antler base and beam, which still contains parts that can be used for tool manufacturing. Another example is an aurochs scapula from house 40. It is cut off transversely approximately on the distal third of its length. On the central part of the blade, where the bone is the flattest, a deep halfcircular cut is made in order to get material, probably for a very specific purpose, perhaps for something like an amulet. Yet, the most interesting item relates to house 20. It is a thoracic vertebra of a wild boar, with an artificial groove on the body of the vertebra (Fig. 6a). The groove is inclined at an angle so as to enable some sort of rod to be inserted and to lean, with its upper side, on the neural spine (Fig. 6b), where a similar, but shallower groove was made. There is a conspicuous polish on the neural spine at the place where an imagined rod would have been leaned on. The polish is also observable on the entire lateral surfaces of the spine, as well as around arch foramina (openings on transverse processes). These openings are probably the reason that the thoracic vertebra of wild boar was chosen, since they are found only in suids. They are positioned in the bases of transverse processes and connect dorsal and ventral sides of the processes with the vertebral channel. Via those openings a rod or a cord might have been pulled vertically through the transverse processes (Fig. 6b), while a cord (but not a rod, because of a slightly different height in their position) might have been pulled transversely through the vertebra. The bone around the openings is also slightly polished, and the openings themselves somewhat widened. It is possible to imagine several possible ways in which rods and cords were pulled through those openings. It seems that this modified vertebra was a part of a complex device with a special function. A wild boar thoracic vertebra modified in almost exactly the same manner has been recognized by the author in the assemblage from Padina. Modified vertebrae were also found at Vlasac, where they were interpreted as cult instruments (Srejović & Letica 1978, I: 106, pl. LXXI, CXXI no. 6, CXXIII no. 6, CXXIV no. 1). The specimens from Vlasac are different from those of Lepenski Vir and Padina in that their transversal processes have been removed, and the neural spine ornamented. Fragmented artefacts or fragments of bones with traces of artificial modifications were also found in houses 23, 26', 27, 34, 35, and 70. In several houses (9, 13, 20, 65 and 70) bones of skeleton parts traditionally used in tool manufacturing were found, which could be regarded either as remains of possible raw material, or as material rejected from such a skeletal part when a desirable part was taken. These are antlers and ruminant metapodials in the first case, and most often carpals and tarsals, as well as phalanges in the second. Antlers in houses Not all the antlers and antler fragments found in the houses at Lepenski Vir are taken into account here, but only those of unquestionably symbolic significance. Red deer antlers placed on the floors of houses are one of

128

Vesna Dimitrijević: Lepenski Vir animal bones

the specific features of Lepenski Vir, whose significance must be taken into consideration when trying to understand the symbolism of the Lepenski Vir culture, together with its unique architecture, stone sculptures, burial rites and other material evidence. In the early phase of research on the Lepenski Vir culture, Srejović (1969: 137) emphasized that red deer antlers possess unique features of growing and being rejected in a yearly cycle, as well as their amazing ability to grow bigger, stronger and more beautiful every year and in this way symbolising an existing connection between death and renewed birth. Whether or not we relate their deposition in the houses with the expression of the belief that the house will live again, their presence in the houses of Lepenski Vir does represent unique evidence of the house abandonment event. It shows that this event was planned, prepared and symbolically justified. In addition, animal bones and antlers offer a possibility to find out about the season of the abandonment event. Red deer skulls with antlers have been found in at least 13 houses at Lepenski Vir (Bökönyi 1969). Not all of them were saved. In the collection at my disposal, pairs of antlers from houses 28 and 22 are examined; on the published photographs two additional pairs of antlers on the skull can be detected in house 22 (Srejović 1969: fig.69; Radovanović 1996: fig. 4.3; Srejović & Babović 1983: 136, lower), and house 61 (Srejović & Babović 1983: 151, lower). Antlers related to burials or below the floors are not taken into consideration here (for example, those from house 61, Srejović 1969: fig. 71). The antlers attached to the skull found on the floor of house 28 are of a young animal. They are very thin, with small brow tine and crown on the right, and without tines on the left beam (Fig. 7). The age is estimated from the dentition as 27 months, and accordingly, the time of the kill to the early autumn (Dimitrijević 2000). We believe this is determining the season of the house abandonment, too. One of the pairs of antlers from house 22 (Srejović 1969: fig. 69; Radovanović 1996: fig. 4.3) has a similar appearance, and possibly a similar age and season of kill. Another antler belonged to a prime adult (Srejović & Babović 1983: 136, lower) (Fig. 3). Shedding of red deer antlers is related to a restricted period at the end of winter (February/March), while fully grown antlers on the skull cover the season from the end of August to the time of shedding in February/March. Still, the easiest time to hunt is thought to be the rutting season (September/October), when the deer population is gathered within limited areas and the males are in excellent feeding condition (Noe-Nygaard 1995). This is also the time of kill already established for the deer skull with antlers from house 28, and possible time of kill for the skull with antlers in the rear of the house 22.

Conclusion It is necessary to begin these concluding remarks by repeating a sentence from the introductory chapter: “not a single contextual unit of collected animal bones from Lepenski Vir has complete recovery of the remains”. This paper thus con-

siders fragments of already fragmented archaeozoological remains, while the overall degree of the fragmentation remains unknown. Yet, the collection of animal bones related to the houses of Lepenski Vir, offers some indications about human/animal relationships and related activities. Moreover, it offers some valuable reflections on particular events in the life of these houses. Probably the most important are those reflections to be made in relation to the events of house abandonment. It is attested that the abandonment of a particular house was a prepared action pregnant with symbolism. Undoubtedly, red deer skulls with antlers were placed in houses at the moment of their inhabitants’ departure, most probably as symbols of a belief in their return and renewal of the life of a particular house. The pattern of this symbolic act may be similar in the majority of the 13 recorded cases, but might have been different in the case of house 22, where more than one item was deposited. The abandonment of some houses was probably marked symbolically in another way. Part of a wild boar carcass in house 40, as well as dog mandibles, if not their whole bodies, were possibly left in the houses as acts of abandonment, too. It is certain that the repeated deposition of fish remains in graves and houses of Lepenski Vir needs also to be taken into account. However, this subject will receive more attention when analyses of the fish remains have been completed. The same is true for bird bones and invertebrate remains, which could in addition offer supplementary data on the seasonality of occupation. Determination of seasonality on the basis of animal remains is the only reliable source of detecting possible seasonal abandonment of particular houses and the settlement of Lepenski Vir as a whole. The early autumn date of the abandonment is confirmed in house 28, indicated in house 22, and possible in houses 46 and 61. Specific methods of ageing animal remains and identification of seasonality will show whether this pattern is of general significance for the whole site. Research in this direction is in progress (Borić 2002). Some of the assemblages described offer evidence of various activities that took place within a particular house, such as food preparation and consumption, use and manufacturing of bone tools, and skinning of animals. Does this evidence imply that some of the houses were not abandoned after elaborate preparations and symbolic acts, but suddenly, unwillingly, or without according special respect to the house left behind? This remains unclear, mainly as a consequence of the incompleteness of the assemblages, and the stratigraphic uncertainties already noted. Finally, analysis of animal bones from the houses of Lepenski Vir confirms that some animal species were accorded a special status. In the case of dog, its close relationship to humans and connection with the house area is reflected in the large number of dog remains, and probable deposition of its skeletal parts or even whole bodies in the houses. It may also be supposed that they were present alive in the houses, too. In fact, the extreme importance of the animal world for the occupants of Lepenski Vir is beyond doubt, and most of the animal species mentioned here certainly received some sort of special treatment, special status and place in the symbolic system of the Lepenski Vir culture. It is for us to try to

129

The Iron Gates in Prehistory

find out the details of these relations in order to enhance our understanding of the character of the Lepenski Vir culture. Acknowledgements I am very grateful to Ljubinka Babović, curator in the National Museum in Belgrade, for allowing me to analyze osteological material from Lepenski Vir. I also thank Ivan Stefanović for his skilful restoration of antlers and conservation of bones, Predrag Filipović for technical improvement of photographs, and Dušan Borić for valuable comments on the paper.

References Becker, C. 1986: Kastanas: Ausgrabungen in einem Siedlungshügel der Bronze- und Eisenzeit Makedoniens 1975–1979, Die Tierknochen-funde (Prähistorische Archäologie in Südosteuropa, Band 5). Berlin: Volker Spiess. Binford, L. 1981: Bones: Ancient Men and Modern Myths. New York: Academic Press. Bökönyi, S. 1969: Kičmenjaci (prethodni izveštaj). In Srejović, D., Lepenski Vir, Nova praistorijska kultura u Podunavlju. Belgrade: Srpska književna zadruga, 224–228. — 1978: The vertebrate fauna of Vlasac. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2. Belgrade: Serbian Academy of Sciences and Arts, 35–65. Borić, D. 2002: Apotropaism and the temporality of colours: colourful Mesolithic–Neolithic seasons in the Danube Gorges. In Jones, A. & MacGregor, G. (eds) Colouring the Past: The Significance of Colour in Archaeological Research. Oxford: Berg, 23–43. Brown, W. & Chapman, N. 1991: Age assessment of red deer (Cervus elaphus): from a scoring scheme based on radiographs of developing permanent molariform teeth. Journal of the Zoological Society of London 225: 85–97. Bützler, W. 1986: Cervus elaphus Linnæus, 1758 — Rothirsch. In

Niethammer, J. & Krapp, F. (eds), Handbuch der Säugetiere Europas, Band 2/II. Wiesbaden: Akademische Verlagsgesellschaft, 107–139. Clason, A.T. 1980: Padina and Starčevo: game, fish and cattle. Palaeohistoria 22: 141–173. Dimitrijević, V. 2000: The Lepenski Vir fauna: bones in houses and between houses. Documenta Praehistorica 27: 101–117. Dragišić, P. 1957: Jelen. Zagreb: Lovačka knjiga. Habermehl, K-H. 1975: Die Altersbestimmung bei Haus- und Labortieren: 2, vollständig neubearbeitete Auflage. Berlin: Paul Parey. Noe-Nygaard, N. 1995: Ecological, Sedimentary, and Geochemical Evolution of the Late-glacial to Postglacial Åmose Lacustrine Basin, Denmark (Fossils and Strata, no. 37). London: Taylor & Francis. Quitta, H. 1975: Die Radiocarbondaten und ihre historische Interpretation. In D. Srejović, Lepenski Vir, Eine vorgeschichtliche Geburtsstätte europäischer Kultur. Bergisch Gladbach: Gustav Lübbe, 272–285. Radovanović, I. 1996: The Iron Gates Mesolithic. Ann Arbor: International Monographs in Prehistory. — 1999: ‘Neither person nor beast’ — dogs in the burial practice of the Iron Gates Mesolithic. Documenta Praehistorica 26: 71–87. Srejović, D. 1966: Lepenski Vir — a new prehistoric culture in the Danubian region. Archaeologia Iugoslavica 7: 13–17. — 1969: Lepenski Vir, Nova praistorijska kultura u Podunavlju. Belgrade: Srpska književna zadruga. Srejović, D. & Babović, Lj. 1983: Umetnost Lepenskog Vira. Belgrade: Jugoslavija. Srejović, D. & Letica, Z. 1978: Vlasac: A Mesolithic Settlement in the Iron Gates. Vol. 1, Archaeology. Belgrade: Serbian Academy of Sciences and Arts. Whittle, A., Bartosiewicz, L., Borić, D., Pettitt, P. & Richards, M. 2002: In the beginning: new radiocarbon dates for the Early Neolithic in northern Serbia and south-east Hungary. Antæus 25: 63–117.

130

New-born infant burials underneath house floors at Lepenski Vir: in pursuit of contextual meanings Sofija Stefanović & Dušan Borić

Abstract: There are 41 infant (infans — an individual under one year of age) burials underneath houses at Lepenski Vir. Although there have been several studies of these individuals in the course of physical anthropological analyses, no special attention has been paid to the numerous burials of children at this site. Moreover, little attention has been focused on interpreting these interments in a wider social context. Our intention is to contextualize fully this phenomenon, primarily in connection with the stratigraphic and architectural features and paying particular attention to the osteometric properties of the buried bodies. The physical properties are: estimation of age at death, examination of possible pre-mortem injuries on the skeletons, a detailed evaluation of represented body parts, and a discussion about taphonomic and methodological problems in preservation and recovery of certain body parts in this context. Also, a general discussion of the mortality profiles of infants is contrasted with the picture of infant burials from Lepenski Vir concluding that these 41 individuals (with a predominant age at death of 38–40 gestational weeks) cannot represent a natural mortality profile of a human population. We analyze these burials as socially embodied and engendered individuals. In interpreting this phenomenon different spatio-temporal scales are considered: a building with a burial, the site of Lepenski Vir with its respective phases, the region of the Danube Gorges with its Mesolithic and Neolithic sequences, and the wider region of the central Balkans in the course of the Neolithic. Particular attention is paid to defining a chronological framework for the phenomenon. Key words: burials, neonates, houses, Lepenski Vir, Mesolithic, Neolithic

Introduction What dimensions of life at Lepenski Vir do the new-born infant burials open up? Buried mainly underneath the rear of trapezoidal buildings at Lepenski Vir, their large number and striking spatial patterning is unique to this site among the other regional sites with chronologically (c. 9500–5500 cal BC) overlapping sequences in the Danube Gorges (northcentral Balkans). Infant burials have remained underrecognized for a long time unlike those other unique and well-known features of Lepenski Vir — abundant artwork of sculpted boulders and limestone building floors. Some questions spring immediately to the mind. Do the uncovered infant burials reflect appalling religious/ritual practices of sacrifice made by family groups objectified through particular ‘houses’/‘shrines’, or is this determined inclusion into domestic space the expression of a specific care and of what kind? For what reasons did only certain buildings accommodate the deceased new-borns and how might this have related to the subsequent life history of these architectural spaces? How do infant burials relate to the other burials at this and the neighbouring sites? Where can we situate this phenomenon historically and what diachronic changes occurred? What sort of meanings might have surrounded the practice and how fixed might these have been? We shall aim at answering here a series of questions by looking into the contextual richness of the site and the area. However, this journey will take us away from Lepenski Vir, also aspiring to its relevance for other places and times, exploring ways of making sense of the web of relations we have started constructing below. It is only recently that a more dedicated approach has crystallized in the study of past infant lives, their social roles,

significance and embodiment (e.g. Meskell 1994; Moore & Scott 1997; Scott 1999). Burials are the most visible and most effectively discussed aspect of the archaeological record relating to children, although other material traces may have some potency for revealing dimensions of infants’/children’s actions (e.g. Finlay 1997; Politis 1999b) or actions specifically related to children (e.g. Politis 1999b). Within our discussion of Lepenski Vir we shall highlight a possible example of this kind, reaching to infants not only through the perspective of mortuary data. Infant burials, despite physical problems related to their survival, do not represent a rare phenomenon among numerous archaeological examples. Still it may be possible to identify reasons that might have deterred archaeologists from engaging fully in discussing and interpreting burials of this age group (which should rather be viewed as consisting of several age stages, cf. Sofaer Derevenski 1997). Possible reasons would account for a frequent low elaboration of lack of detail for infant burials when compared to adult burials, difficulties in attributing sex to children and subadults (up to the age of 15–20 years) on morphological grounds, which blurs picturing infants’ social personae clearly, and a general perception of infants’ lesser role in the (adult) life of a past society that is studied by adult archaeologists. In the case reported below, we have tried to overcome some of these problems, such as that of the determination of sex by using the newly acquired results of DNA analysis (PCR method) on the studied material (Čuljković et al., this volume). Other potential obstacles in the study of infants’ roles and meanings of their burials are solved by contextualizing their place and significance, i.e. by looking at relational features of the studied site and the stratigraphic contexts, and connecting existing evidence from neighbouring and con-

131

The Iron Gates in Prehistory

Figure 1. Lepenski Vir – houses and burials (supplemented after: Srejović 1981: 20–21, and field documentation).

temporaneous sites in the wider region. A more complex understanding is reached by examining how different temporal and spatial contexts of human societies, described through numerous ethnographic records, respond to, but also protect from, an infant’s death. This perspective, on the level of general analogies, explores emotional (e.g. a period of bereavement and a need for the physical proximity of the corpse), social (e.g. practices of negotiation of meanings in delaying the act of burial) and practical dimensions (e.g. the issue of reproduction in a descent group) that constitute culturally inscribed attitudes specifically related to infants, which are context sensitive in response to existential anxiety of this kind. A number of general similarities transparently visible across various ethnographic and archaeological contexts emphasize the need to bring and discuss these examples together and to compare them. While archaeological examples, such as the infant burials of Lepenski Vir discussed here, with the addition of those from the neighbouring and earlier site of Vlasac, can open up diachronic depths in viewing localized transformations of particular practices, ethnographic records with thick observations of social life can in turn indicate new possibilities for understanding the main themes of material culture rhetorics of the archaeological examples studied. The following discussion is in three main parts. Firstly, we provide a discussion from the perspective of physical an-

thropology with related issues in the study of osteological remains of infant burials. This includes taphonomic concerns, details of identification, metric procedures aimed at precise ageing of the burials and a discussion of subsequent construction and interpretation of mortality profiles, as well as an inquiry into the possible cause of death. Secondly, we examine the archaeological context of these burials and their relation to the architecture of hearths and trapezoidal building floors and cycles of their use and abandonment, other burials of different age groups at the site and the possible connection with the representational artworks of sculptured boulders. By drawing attention to the stratigraphic and chronological contexts of the infant burials at Lepenski Vir, we touch upon the problem of a historically contested significance of the phenomenon. Lastly, possible meanings of the phenomenon are sought through both a diachronic regional perspective and across a wider anthropological background, drawing upon local Balkan ethnographies.

Physical anthropology of infant burials The meaning of the Latin word infantia is ‘unable to speak’, and until recently infants remained quite silent in the field of physical anthropology. It remains an open question at what age infant becomes child, and definitions of these age stages

132

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir Table 1. Infant burials/houses at Lepenski Vir, their positions and orientations, and available long bone measurements used for providing estimated age in gestational weeks. House no.

Burial no. Position

1

3

112

‡

2

4′

109

Back/flexed legs

109a

‡

106

Back/legs splayed outwards

3 4

4

Orientation

Femur length (mm)

Humerus length (mm)

Age (gestational weeks) & Sex (PCR DNA)

‡

–

65

♂ 38–40

SE–NW

72

63

♂ 38–40

‡

82

–

♀ 41–43

c. SW–NE

81

74

♂ 44–46

5

106(1)

–

–

?

? 41–43

6

107

Back (legs missing)

N–S

–

67

♂ 38–40

7

108

Back/flexed legs

N–S

–

67

? 38–40

8

110

‡

‡

–

70

♀ 41–43

One bone only

9

13

116

‡

10

19

98

Flexed/left side

11

98a

‡

12

103

Back/contracted legs

94

Flexed/left side

14

95

‡

15

101

Flexed/right side

16

102

Flexed/left side

13

24

17

26

63

Extended/slightly bent legs

18

27

128

‡

19

129

20

130

21

131 29

‡ – these three burials were placed one over the other

119

‡

23

120

‡ – one placed on the top of the other

24

121

22

25

36

26 27

37

28 29

38

62 (fibula) –

61

♀ 35–37

‡

–

62

♂ 35–37

S–N

75

65

♂ 38–40

NE–SW

79

66

♂ 38–40

‡

79

–

♀38–40

NW–SE

79

71

? 38–40

SE–NW

79

67

♀ 38–40

SE–NW

83

70

♀41–43

‡

63 (tibia)

‡

78

68

♂ 38–40

‡

74

66

♂ 38–40

‡

75

65

♂ 38–40

‡

79

–

♂ 38–40

SE–NW

76

65

♂ 38–40

–

62

♂ 35–37

♀ 35–37

114

Back/legs splayed outwards

S–N

72

64

♂ 38–40

115

Back/legs splayed outwards

S–N

76

65

? 38–40

132

‡

‡

–

65

♂ 38–40

133

‡

‡

75

67

♂ 38–40

111

Back/legs splayed outwards

S–N

76

–

♀ 38–40

–

≈70

–

? ≈35–37

c. SE–NW

78

69

♂ 38–40

111(1)

30

♀ 38–40

‡ c. NW–SE

One bone only

31

43

96

‡

32

47

123

Contracted/left side

SE–NW

77

–

♀ 38–40

124

Extended on its back

S–N

74

64

♂ 38–40

125

‡

‡

74

65

♂ 38–40

127

‡

‡

–

69

♀ 38–40

NW–SE

77

66

♀ 38–40

80

–

♀ 38–40

33 34

54

35 36

62

118

‡

37

63′

113

Back/legs splayed outwards

♀ 35–37

38

63

117

‡

‡

39

48

134

‡

‡

40

See note 4

83b

‡

‡

–

82

? 48–50

41

See note 4

83b(1)

‡

‡

–

≈ 65

?≈38–40

‡

‡ — position or orientation unclear

133

Not preserved

The Iron Gates in Prehistory

vary across cultural contexts (Scott 1999: 2). Here, we use one of the most common definitions, that an infant is an individual under the age of one year (Saunders 1992). Most of the infant skeletons from Lepenski Vir discussed here belong to the category neonata, new-born infants, i.e. individuals who died during birth or soon after. We may expect some sort of special treatment in those instances when the number of neonates is high in the archaeological record. Although infant skeletons from Lepenski Vir were examined first by Zsuzsanna Zoffmann (1983) and subsequently by Mirjana Roskandić (1999), few details are given in those accounts. Zoffmann (1983) focused on the demographic facts, making the generalization that the “great number of newborn skeletons reflects the well-known high infant mortality in earlier times” (Zoffmann 1983: 131). A more detailed account was intended for the final publication on Lepenski Vir that has never appeared (Zs. Zoffmann, pers. comm.). Research on the Lepenski Vir human osteological material by Mirjana Roksandic (1999) led to a revision of the number of infant skeletons present in the assemblage whilst concentrating on other aspects of the collection. In this paper we offer a more detailed osteobiography of infants at Lepenski Vir through taphonomic analysis, more accurate estimation of age at death, new DNA findings on sex attribution (cf. Čuljković et al., this volume), plotting, comparing and interpreting their mortality rate, and indicating possible causes of death. Taphonomy To understand taphonomic processes and the extent of their severance in relation to osteological remains may help in determining sample bias (Eldredge 1991). For instance, the low mineralization of bones of infant skeletons may lead to poor preservation in the archaeological record. It is the relationship between mineral and mechanical properties of bones that explains the brittle nature of the bones of younger children (Guy et al. 1997) — their tensile strength is fairly low while their compressive strength and hardness are extremely low (Currey & Butler 1975). Under some pressure these skeletons would rarely escape being crushed in the ground. In addition, even at marginal soil pH ranges it may be expected that all or most of the infants would be eliminated systematically from the burial sample (Gordon & Buikstra 1981). These generally pessimistic remarks in relation to the preservation of infant skeletons do not apply to the infants discovered at Lepenski Vir. Bones of almost all infants excavated at Lepenski Vir are well preserved with a fairly high percentage of bones present in individual skeletons and with generally good preservation of single bones (even in the case of fragmentary skeletons single bones are in a very good state). Exceptions to these general rules are the bones of Burials 98a and 119, which are poorly preserved (Figs 13 & 14). Cranial bones are frequently crushed, related to the fact that these are not fused in the skull and are highly fragile. Also, hand and foot bones are very rare most likely due to hand-collecting and lack of sieving during excavation. As Zoffmann’s (1983) and Roksandic’s (1999) estimates of the number of infants present at Lepenski Vir differ, we need to resolve their exact number on the basis of the preserved material. According to Zoffmann (1983), there are 80

individuals buried at Lepenski Vir and “36 individuals belong to Inf.–Inf. I. age groups” (Zoffmann 1983: 131). Without indicating a precise number of new-born infant skeletons, Roksandic (1999: 78) notes the presence of “190 individuals from 134 graves plus 42 individuals from unidentified contexts. Among them, there are 51 infant skeletons, 27% of all buried individuals”. We re-examined only the skeletons of neonates and came up with 41 individuals at Lepenski Vir, accounting for over 20% of the entire burial assemblage (the exact percentage cannot be given before all the skeletons from the site are re-examined). As Zoffmann (1983) does not list individual burials, the discrepancy between our results and the two previous accounts can be discussed only in relation to Roksandic’s (1999: 78–80) findings. The difference in the number of neonates present relates to those buried under Houses 4, 4′, 24, 35 and 47 at Lepenski Vir. The differences come from the estimations made in the course of the osteological analyses when in a few cases individual bones of several individuals (often represented by only one ‘extra’ bone) are identified within the material of a single archaeological burial and subsequently marked as separate individuals (the identifications were made by M. Roksandic). However, as burials of neonates are found underneath building floors often close to one another, excavated at the same time and re-packed and analyzed several times we must account for all these potential biases. Thus, we concentrated on discerning how these ‘extra’ bones/individuals might relate not only to the bones of the burial within which they were found, but also to the bones of other burials underneath the same house. These instances are as follows: Houses 4 and 4′ (Burials 106, 107, 109, 109a, 108, 110) (see Fig. 7). Among the archaeologically identified burials of this building, the following individuals were added: 106(1) — one right humerus, 108(1) — one right femur, 110(1) — one right scapula. The additionally marked Burial 108(1), found among the bones of Burial 108, may belong to either Burial 107 or 110 — both lacking the right femur. Similarly, the right scapula marked as Burial 110(1) may belong to either Burial 109 or 109a, both lacking the right scapula. However, it is not clear how to explain one extra right humerus in Burial 106, marked as Burial 106(1), as this burial was found on the other side of House 4 (corner C). This may be a sorting mistake; however, we have to account for this individual (see Table 1). House 24 (Burials 94, 95, 101, 102) (see Fig. 16). Two more burials were separated: 101(2) — one right tibia and 4 ribs, and 102(1) — one pelvis. However, as there is no right tibia among the bones of Burial 102, the ‘extra’ right tibia in Burial 101 may well belong to Burial 102, buried close to one another (although initial sorting and subsequent re-packing and analyses must be accounted for again). The burial marked as 101(1) is represented only by a rib of an adult individual. The pelvis fragment (ischium) marked as 102(1) was rightly separated from Burial 102, but could belong to Burial 101. House 36 (Burials 114 and 115) (see Fig 9). Burial 114(1) — one right ulna clearly belongs to Burial 115. These two burials are buried one on top of the other. House 38 (Burial 111) (see Fig. 5). The burial marked as

134

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

111(1) is a left femur and since Burial 111 has both femora and is the only burial in House 38, this is clearly an ‘extra’ bone whose inclusion can be explained by some of the factors already mentioned. However, we shall treat it as a separate individual (see table 1). House 47 (Burials 123 and 124) (see Fig. 4). Two more individuals were separated: Burials 123(1) — a fragment of a right femur and 124(1) — a fragment of a humerus and a fragment of a left ulna. A fragment of a right femur marked as Burial 123(1) most likely belongs to Burial 124 where there is no right femur. Also, a fragment of humerus marked as 124(1) refits with the humerus of Burial 124. A fragment of a left ulna most likely belongs to Burial 123, which is missing both ulnae. Although Burials 123 and 124 are not buried in the immediate vicinity of one another, some of the factors discussed may have caused the evident mixing. Having in mind these corrections, the number of new-born infants present in the assemblage is 40, to which may be added infant burial 134 (House 48) which was noted and described in the field documentation (see below) but was not physically preserved after the excavation (noted in Field Journal, 5/10/1970). Although there remains a possibility that some of the ‘extra’ bones described do represent separate individuals, disturbed through activities upon the floors of the Lepenski Vir’s buildings, the data currently available indicate 41 burials of neonates at this site. We see no reason to suggest a different scenario to explain these instances (e.g. the possibility that the extra bones were incorporated into the burials intentionally). Age estimation Bone size is a good indicator of age in perinatal skeletons. Bone growth in the foetus may be influenced by extrinsic factors, such as poor maternal nutrition, but these factors may influence bone growth less after birth (Mays 1998). Maternal malnutrition has to be quite severe to retard growth in the developing foetus as in this situation the foetus is protected at the expense of the mother (Bagchi & Bose 1962). Scheuer et al. (1980) indicated the relationship between length of long bones and age, using foetal material from medical collections in England. Measuring the length of limb bones, they came up with a relationship between age at death and long bone length in foetal remains, thus giving an estimated age at death within about two weeks (Mays 1998). Some additional information on the growth of long bones in infants and children has been gathered from the study of the Indian Knoll skeletal sample by Johnston (1962) and a consideration of the problem of estimating age at death only from a diaphysis by Stewart (1968). It has been accepted that the gestational ages of the foetal skeletons ranged from 27–46 weeks, where an average gestational age at birth is 38–41 weeks (Tanner 1989). This contrasts with Scott’s (1999: 67) scepticism that it remains difficult “to determine the precise gestational age of infant skeletal material from antiquity”. Estimates of the gestational age of infants from Lepenski Vir were made according to the maximum length of both/either femora and/or humeri (Bass 1987; Mays 1998). Only in the cases of infant skeletons 116 (House 13) and 128 (House 27; Fig. 12), where femora and humeri are not present, was the age estimated according to the maximum

length of either fibula (Burial 116) or tibia (Burial 128), which were then compared to the length of these bones respectively in other skeletons where estimations were based upon femora and humeri. On the basis of these metrical comparisons, the greatest number of those present indicate an age of 38–40 gestational weeks (± two weeks), i.e. an average gestational age of an infant at birth, indicating neonates. Generally the mandibles of neonates are not well preserved in the material studied and it was not possible to infer age on the basis of dental development. There still remains the possibility of employing X-ray analyses on some of the wellpreserved mandibles of this age group. In addition, we examined two other individuals (children) buried within buildings — Burials 92 (House 28; Fig. 18) and 97 (House 31; Figs 19 & 20). As Burial 92 is well preserved, it was possible to infer its age on the basis of dental development (Ubelaker 1978: fig. 62) and the age is around 2 years and 8 months. Burial 97 indicates age around 2–6 years on the basis of the postcranial skeleton, as the mandible is not well preserved. Sex determination The reliability of morphometric analyses in sex identification of infants is low, and although pelvic sex differences are present in infants at birth, they become obscured soon after birth (Mays 1998). As a reliable substitute, DNA-based sex identifications of 30 sampled infant burials from Lepenski Vir were conducted. Of the 30 samples tested, 29 provided results indicating 16 males and 13 females (see Table 1; Čuljković et al., this volume). For further discussion on the sex identifications and their relations to the spatial patterning at the site see the discussion below. Mortality It is commonplace in discussions of infant mortality to emphasize high infant mortality in the past, and especially in earlier prehistory. Similar generalizations indicate that “high infant mortality in earlier and modern developing societies contrasts with lower infant mortality in modern Westernized societies” (Malhorta 1990: 315). Obviously, we can hardly estimate the real number of infant deaths within the first year of their lives in a population on the basis of the archaeological sample. Also, a burial sample may be biased as in a number of societies infants may be excluded from cemetery burials (Mays 1998), with their bones usually being underrepresented in the studied assemblages (Roberts & Manchester 1995). High chronological resolution of excavated burials would be desirable before any attempt is made at forming a synchronic picture of population demography. Therefore, any attempt to calculate infant mortality rate for an archaeological cemetery/burial site is a risky procedure, as this hardly reflects a true picture of a past population, bringing under scrutiny the relevance of sophisticated palaeodemographic statistics (contra Roksandic 1999). However, we may try to offer a qualitative, comparative and fairly general assessment of the high percentage of infant burials at Lepenski Vir. For instance, in archaeological cemeteries the proportion of infants often fluctuates around 5–6% of the whole assemblage (Guy et al. 1997: 221). We have already indicated that the presence of infants at

135

The Iron Gates in Prehistory

Lepenski Vir is around 20% out of the total number of burials uncovered (bearing in mind the considerable time span probably represented by these burials, see below). It seems that more new-born infants were chosen to be buried at Lepenski Vir in comparison to all other age groups. But it can hardly prove a higher mortality rate of neonates at Lepenski Vir over a particular period of time. And as it is “hard to imagine that the scarcity of infant remains in our cemeteries can be a true reflection of a demographic fact” (Guy et al. 1997), we find it hard to imagine that a large percentage of neonates at Lepenski Vir is a true reflection of a demographic dictum. Although generalizations about high infant mortality in prehistory due to diseases and accidents (Scott 1999) may hold some truth, with considerable fluctuations over historic periods, we have to remind ourselves that the reasons for an evident bias in the presence of any age group in the archaeological burial record are primarily functions of particular and context-sensitive cultural practices. Cause of death Perinatal deaths indicate foetal deaths plus deaths of infants who die within the first 28 days of life, while neonatal deaths occur after 28 days of life but within the first year. Today, some major factors associated with infant deaths include: congenital malformations, perinatal conditions and infections, and SIDS (sudden infant death syndrome), with peaks between the second and fourth month of life remaining unexplained. It may be assumed that in the past the percentage of infant deaths attributable solely to congenital disease was probably low and that many infants probably died from widespread infections of gastroenteritis and pneumonia, deaths which affected both normal and congenitally malformed infants (Roberts & Manchester 1995). Since infants remain vulnerable to acute gastrointestinal and respiratory infections today, deaths in this age group may have been much higher in antiquity without the availability of antibiotics (Roberts & Manchester 1995). None of this would leave any visible trace on the osteological material. Another natural cause of infant deaths may be nutritional deficiency diseases (scurvy, rickets and iron deficiency anaemia). Possible traces of classically described nutritional deficiency are present. Burials 106 and 107 (House 4) had irregular growth of claviculae. Burial 120 (House 29) displays cribra orbitalia. Some other signs (irN

regular growth of long bones and porosity) were noticed on Burials 108 (House 4), 109, 109a (House 4′), 114 (House 36) and 121 (House 29), and also on infant Burials 83b and 83b(1). However, the archaeologically visible palaeopathology in the case of anaemia is still questionable (StuartMacadam 1989). Another cause of infant deaths may be infanticide. It has been suggested that some societies practise infanticide as a method of population control owing to the traditional requirement to control the number of females, or for general population control (Roberts & Manchester 1995). Scott (1999: 12), indicating appalling feelings held in relation to the question of infanticide, notes: “It has struck me on a number of occasions how discussion of infants in the past, and especially infant burials, invariably turns to discussion of infanticide”. As some sort of infanticide is for the most part legal today through abortion, we can hardly reject the fact that infanticide occurred in the past, in various forms and with different reasons and connotations. There remains a problem with its detection on the osteological material since this might have included drowning, smothering, strangling, exposure, or neglect (cf. Scott 1999). One of the indicators of infanticide may be a strong peak of individuals aged about 38–40 weeks, following the logic that infanticide usually occurs immediately after birth, and “the age distribution of victims of this practice would be expected to mimic the gestational age of all live births”, i.e. between 38–40 weeks (Mays 1998: 66). Although the gestational age of infants from Lepenski Vir has a strong peak between 38–40 weeks (see Fig. 2), we cannot comment with any certainty about infanticide here. Cultural and contextual reasons for this opinion are provided below. Although it is to be expected that most births in the past as well as in the present took place commonly between 38–40 gestational weeks, there may be a significant discrepancy in the survival rate of new-born babies in the past and today. It is almost certain that in the past the rate of infant mortality in the course of delivery was much higher. In addition, as some authors have suggested, infanticide is evident through the skewing pattern in relation to one sex. Data from Lepenski Vir with 16 males and 13 females of successfully DNA-analyzed infants (see Table 1; Čuljković et al., this volume) do not fit this proposition. Presently, the only ambiguous trace is a cut mark on the left humerus of Burial 112 (in House 3 — male, 38–40 gestational weeks). The cut mark is located above the distal end of the humerus, at its lateral side (length 10 mm, breadth 3 mm). It was made by a sharp tool, possibly flint. However, rather than seeing this as proof of a violent death of this neonate, it may be related to the practice of defleshing the body. Equally, the skeleton may have been affected by postburial disturbance.

Stratigraphic and contextual relations Gestational Week Intervals Figure 2. Absolute number of infant individuals (N=38) studied at Lepenski Vir across stages of estimated age in gestational weeks (Burials 83b and 83b[1] excluded).

The following observations on the stratigraphic and contextual position of the burials of new-borns at Lepenski Vir are made on the basis of published information from the site (Srejović 1969a, 1969b, 1970, 1972, 1981; Srejović & 136

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Babović 1983; Radovanović 1996) and unpublished archive field documentation from the excavation campaigns at Lepenski Vir.1 Most of the infant skeletons from Lepenski Vir were uncovered in the course of the protection-excavation work throughout August, September and the start of October of 1970. During this period, a number of building floors and hearths uncovered at Lepenski Vir were lifted from their original in situ location and relocated to a higher terrace where the Museum of Lepenski Vir is situated today (Čanak-Medić 1970, 1971).2 During that campaign c. 40 new-born infant burials were found at Lepenski Vir below the level of building floors in the rear parts of 19 buildings at the site. Although these burials were mentioned briefly on several occasions by the excavator (e.g. Srejović 1969b: 15, 1981: 40, 42 — attributing this group of burials to phases LV Ib-e) and listed in relation to buildings by I. Radovanović (1996: 175–185), based on her insights into unpublished burial descriptions, the whole phenomenon has escaped further attention. The infant burials were not mentioned in the monographs published about Lepenski Vir. One reason was the fact that their discovery came after the first version of the Lepenski Vir monograph — the Serbian edition published in 1969 — was completed, with subsequent translations (into English, 1972; and German, 1975) largely following the original text with only minimal additional editing of paragraphs and with appendices to the previous text. However, the fact that infant burials regardless of their large number and striking patterning at Lepenski Vir, have not been studied and described with the same care and detail as some other features (such as artworks), possibly reaffirms the already mentioned general pattern of neglecting infants as a subject of lesser archaeological importance and interest. ‘Houses’ vs ‘shrines’: the significance of domestic space In the following discussion, referring to places where infant burials were discovered we shall use the term ‘house’ to describe uncovered architectural features at Lepenski Vir with trapezoidal bases and centrally-located rectangular hearth slabs and front ‘thresholds’. This kind of terminology is not without problems. We still lack in a published form detailed knowledge of the contextual associations at these floor surfaces that would say more about both the character of occupational activities that took place in connection with these spaces and about scheduling of acts of abandonment of the structures, witnessed through the intentional deposition of animal bones and objects over the limestone floors or thrown into the infill of these features (cf. Bökönyi 1969, 1972; Borić 2002a; 2002b; Borić & Dimitrijević 2005; Radovanović 1996). One view is to see and interpret these features exclusively as ‘shrines’ (also reflected in the terminology used in some source publications, cf. Srejović & Babović 1983; see also Babović 1997) where the supposedly sacred character of activities and material culture associations completely exclude ‘mundane’ aspects of life (Babović 1997: 103). On the other hand, the excavator of the site, Dragoslav Srejović (1969a: 67–69; 1972: 66–67), in a rather structuralist fashion, interprets the space of a recurrent architectural form of the structures found across the site, seeing the sacred part con-

fined to the area around the hearth (sometimes surrounded by stone ‘tables’) of the trapezoidal buildings, with restriction of movement in this area related to social boundaries, and the rest of the building space serving profane functions — see below, for similar suggestions regarding the sacred/profane aspects of building space and social implications of indicated oppositions (e.g. Eliade 1957; Bourdieu 1990[1970]; Hodder 1990). Interpretative attempts of this kind remain largely synchronous and idealized, only a ‘frozen’ view of one moment — in the case of Lepenski Vir most likely the moment of abandonment. Fixing oppositions in this way puts little emphasis on the historical contingency of these features and their individualized characters, the varied practices involved in their furnishing, and the ways in which generations of humans who visited or inhabited these structures understood, made sense of and altered their embodied and objectified social and cultural norms. Other suggestions have been made along the lines of a more complex transformational character of these spaces, seeing their life cycle trajectories as going through a full circle — from domestic areas with a clearly profane character to tombs and perhaps spaces for special ritual purposes, all depending on their individual biographies, i.e. events that took place in the course of individual histories of these structures and their visitors/inhabitants (Borić 2003, 2005). We will seek to understand the complex significance and character of the features where infant burials occurred by examining contextual associations reported within these structures and by unfolding the web of possible meanings and narratives created in the sequence of events that took place during their histories — their life cycles. This is highly relevant in interpreting a large number of infant burials at this site and for tackling the issue of their restriction to only 19 out of at least 73 architectural structures uncovered at the site.3 The term ‘house’ will be retained as it perhaps can best capture the possible significance that these features may have had for the complexity of kinship relations at Lepenski Vir. Also, C. Lévi-Strauss’s (1983, 1987) original formulation of ‘house societies’ carries significant relevance, despite the degree of difference expressed on the matter of its accuracy, pointing to buildings as complex objects that frequently mirror social relations and cultural identity (cf. Carsten & Hugh-Jones 1995; Joyce & Gillespie 2000). We shall explore this perspective here in relation to the significance of the existing association of infant burials and buildings found at Lepenski Vir, especially regarding the fact that the buildings at this site are the first examples of elaborated and localized building forms that appeared in a specific historic period. Burials found in the course of protection works in 1970, when the floors were lifted, were assigned numbers between 94 and 134 (only neonate Burial 63 [Fig. 3] was found earlier — in 1968). Apart from 19 buildings with neonate burials underneath their floors, two more buildings are considered closely (Houses 28 and 31), with burials of children of other age groups. Spatial patterning The striking pattern in relation to the infant burials at Lepenski Vir is that all of them are found exclusively in con-

137

The Iron Gates in Prehistory

Figure 3. House 26 and neonate Burial 63.

nection with buildings. A number of burials of different age groups (134 in all) found at the site were either placed in relation to the trapezoidal buildings or were found in the space around these structures (see Fig. 1). However, burials of new-born infants were found below the rear ends of 18 buildings at Lepenski Vir only.4 There was no specific spatial clustering of neonate burials to any particular area of the site (Fig. 1). Also, there was no clear-cut patterning in relation to building size — they were equally present in fairly small and less elaborate buildings, such as House 43 (Fig. 11), as well as in much larger structures with elaborate architectural features, such as Houses 54 (Fig. 8) and 37 (Fig. 6), which are among the largest buildings at the site. One clear spatial pattern that is not only confined to this age group but relates to all other burials at the site, is their absence from the area with a cluster of individual and overlapping buildings of the north-east part of the site (superimposed Houses 10, 11 and 12, Houses 5 and 6, Houses 7, 8, 9 and 17, House 16 and House 64 — see Fig. 1; cf.

Radovanović 1996; Borić 2002a: fig. 7). On the other hand, there was again no clear-cut spatial patterning in the distribution of infants in relation to their sex (enabled by recent DNA analyses; see Table 1, cf. Čuljković et al., this volume); neither in relation to a possible clustering across the site nor within those buildings that contained more than one infant interment. Only in two buildings, with more than one infant burial, were individuals of exclusively one sex found — House 29 (3 males) and House 37 (2 males). Seven buildings contained infant burials of both sexes — House 4′ (one male and one female), House 4 (two males and one female), House 19 (two males and one female), House 24 (two females and one male), House 27 (three males and one female) House 47 (one male and one female) and House 54 (one male and one female); in addition, in House 36, with two burials, there is one male and the other of unknown sex, and another nine buildings had only one infant burial respectively. The number of male infants slightly outnumbers the female ones (16 males and 13 females). There are no indica-

138

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

tions of a burial preference in relation to one sex among the studied individuals. This fact may find its relevance within an overall interpretation of the phenomenon. Another striking spatial pattern is that all infant burials, except one (Burial 63), were found below the ‘rear’ of the buildings (the area behind the rectangular hearth). The neonates and children were found a) underneath the limestone floorings at the rear of buildings, b) in the area next to the limestone floorings at the rear of buildings, sometimes below the construction stones, which at certain buildings surrounded the limestone floors, and c) in one case (Burial 63) below the south-west corner (corner A), i.e. in the front part of the building (House 26; Fig. 3). a) Experience of building space at Lepenski Vir In order to add a contextual dimension to these patterns, it would be necessary to describe briefly the original topography of the site and some of the constructional elements of the buildings at Lepenski Vir. The site is placed on a sloping alluvial terrace covered by deposits of loessic sand (for the further discussion on the geological formation of this deposit see MarkovićMarjanović 1969, 1978; Brünnacker 1971). As the excavator described (Srejović 1969a: 62, 1972: 62) and as emphasized recently (Borić 1999, 2002a, 2002b), the structures with the trapezoidal footprints were actually objects cut into the slope. In this way, their builders made a semi-subterranean levelled space that had a trapezoidal shape, making an elaborate reference to the trapezoidal Treskavac Mountain that is situated across the Danube in front of Lepenski Vir, being an impressive landmark (for discussions on the significance of this landmark, see Srejović 1969a, 1972; Chapman 1993, 2000; Borić 1999, 2003). The base of the space dug out in this way was further furnished with a limestone flooring of red to white colour c. 5–25cm thick (Srejović 1969a: 57, 1972: 54; for a discussion on the possible technology used for this flooring, see Ney 1971). In a number of instances the limestone flooring did not cover the entire area levelled for the floor. Instead, rows of irregular stone blocks surrounded the floors in a number of buildings assigned by Srejović to the Lepenski Vir I phase. On the other hand, the Lepenski Vir II phase, as defined by the excavator, represents a later phase (Srejović 1969a: 92 calls it ‘maniristic’), overlying the settlement of phase I. This phase was defined as consisting of buildings with a trapezoidal shape that is defined by rows of irregular stone blocks only, but without specific limestone flooring and mostly lacking rectangular hearths. However, it has been suggested that this stratigraphic interpretation is perhaps misleading (Borić 2002a). This can be shown clearly by superimposing outlines of stone walls of supposedly Lepenski Vir II phase buildings (marked in Roman numbers) over the outlines of limestone building floors of Lepenski Vir I phase. As the outlines of phase II buildings strikingly match the outlines of limestone building floors of phase I, the features identified as belonging to the Lepenski Vir II phase can rather be seen as walls made in a dry stone technique to protect the dug loessic sides of the cuts of Lepenski Vir I buildings, forming part of the lower wall portion of a building. Also, probably, these served to support the upper part of the construction and the roof, made of more perishable materials

such as wood and possibly reeds and/or hides (Borić 2002a: figs 7–9). This description of the three-dimensional outline of the main architectural features of the buildings at Lepenski Vir, especially in relation to the offered re-interpretation of some of their constructional elements, is intended to better contextualize the possible significance of neonate and child burials in the rears of the buildings. Although the rear of the building might have been inscribed with meanings in relation to the structural normative division of space at Lepenski Vir (prescribed for certain embodied actions, although with possibly fluid and transient boundaries in space division), it might have gained some significance relating to its ‘cave atmosphere’ (Srejović 1981: 22; Srejović & Babović 1983: 43), placed deep into the slope, i.e. cut into the virgin, sterile deposits of the original sloping terrace. The fact that infant (and other) burials were not found within the buildings of the north-east part of the site may relate to the relatively flat terrain and shallow deposits (quickly reaching bedrock) in this area, where buildings would have lacked the darkness and protection of steeply cut sides. The importance of this spatial dimension becomes even more evident through the example of the only deviation from the rule of placing infants in the rear of the building. This is the case of Burial 63 interred at the front corner (corner A) of House 26 (Fig. 3) — the only building at Lepenski Vir (except miniature House 49, which will be discussed below) that was orientated parallel to the Danube, in contrast to the other buildings at the site the wider front parts of which mainly faced the Danube. This unusual orientation may have been a consequence of the topography of the north-central area of the site where House 26 was situated, which slopes slightly toward the south-east. Also, several very large rocks originally surrounded the building on its north-east side (see Fig. 1), i.e. the side its front part would have been had it followed the way other buildings were oriented. Thus, we have some reasons to speculate that the way this building was oriented is related to the unusual and exceptional interment of Burial 63 in the south-west, front corner of the building. A part of the floor in this corner was cut in order to place the burial into the virgin sandy loessic soil of the slope (Borić 1999: fig. 20). Having in mind a topographic spatial dimension, placing Burial 63 in this area of the building is not different from other infant interments, i.e. away from the Danube and cut into the slope of the terrace. This instance indicates the importance of the ‘topographic rule’ behind the way of reasoning of where to inter an infant’s body at Lepenski Vir. b) Engendered space and its interpretations The last point does not undermine the fact that all other infants were restricted exclusively to the rear end of the buildings. This may indicate the presence of embodied social, age and gender boundaries in the use of building space (both in life and death), restricting activities and social rights in relation to differences that were mapped onto the building space. Ethnographic records abound with examples indicating spatial relations of this kind (e.g. Bourdieu 1990[1970]; Lane 1994; Carsten & Hugh-Jones 1995; Joyce & Gillespie 2000). Also, for instance, on the basis of detailed micromorphological analyses of floor use at the Early Neolithic site of

139

The Iron Gates in Prehistory

Figure 4. House 47 and neonate Burials 123 and 124.

Çatalhöyük (south-central Anatolia), similarly prescribed engendered restrictions in the use of a building are suggested (Hodder & Cessford 2003). Different building parts may have had very different meanings and complex connotations, as will be discussed in more detail below in the context of the scale of building associations. Despite the suggested general patterning, where the rear of the building at Lepenski Vir frequently served for infant interments, in the case of each building the decision of where to place a burial was made according to the topography of the surrounding terrain and constructional elements of each house. Moreover, specific meanings assigned to particular areas of ‘houses’, their constantly shifting and individualized re-interpretation and the events that took place in the course of the house history, all influenced decisions about where to inter a deceased infant. Although people at Lepenski Vir must have obeyed culturally determined spatial norms, burial practices following an infant’s death probably reflected a

particular constellation of events within the house and the community. More general or collective reasons for the special and patterned burial rite in the case of this age group at Lepenski Vir, shared among the houses and over generations, are constituted by the noise of individual actors. The provision of burials may release to some extent the noise spectrum of individual/‘house’ (re-)interpretations of these collective and shared norms, including dimensions of diachronic changes. Body treatment and elaboration of burials The neonate and child burials at Lepenski Vir are relatively poorly furnished. This is not surprising as generally this age group is rarely accompanied by rich grave goods (e.g. Scott 1999: 90ff.). None of the infant burials at Lepenski Vir was accompanied by any sort of non-perishable grave offering. However, we can focus on other elements of the burial, such as the orientation and position of a skeleton and elements of

140

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Figure 5. House 38 and neonate Burial 111.

burial construction. There were problems in determining the exact orientation and especially the position of the easily disturbed infant burials (see Table 1). This is due to the taphonomic processes that may have affected these burials regardless of their good preservation, as indicated in the taphonomic description above. Also, in several cases it seems that the placing of a new interment in the same area disturbed some of the earlier skeletons, affecting their preservation, as in the cases of Burial 98a disturbed by Burial 98 (House 19; Fig. 13), Burial 110 disturbed by Burial 107 (House 4; Fig. 7) and Burial 109a disturbed by Burial 109 (House 4′; Fig. 7); this is also evident from the nomenclature used in marking these burials with an additional letter ‘a’ to the number of burial that disturbed them. Moreover, even in those houses where there was only one infant burial, we may assume that their occasional poor preservation is a consequence of later occupational and

building activities that disturbed them. These later disturbances might have caused the reportedly poor preservation of Burials 95 (House 24; Fig. 16), 112 (House 3; Fig. 10), 117 (House 63), 118 (House 62), 119 (House 29; Fig. 15), 132 and 133 (House 37; Fig. 6), 125 and 127 (House 54; Fig. 8) and 128 (House 27; Fig. 12). In the case of Burials 129, 130 and 131 (House 27; Fig. 12) and Burials 120 and 121 (House 29; Fig. 15), it remains an interesting possibility that their interments took place at the same time, although the very complicated and to some extent messy position in which they were found may still be a consequence of sequential interments, with a time lapse, at the same spot within the building. a) Re-enacting the body position In those cases where it was possible to detect and reconstruct the anatomical position of infant skeletons, the excavators noted the following positions: a) extended, b) contracted/

141

The Iron Gates in Prehistory

Figure 6. House 37 and neonate Burials 132 and 133.

flexed and c) a position with legs splayed outwards (see Table 1). Only two burials were found in the extended position — 63 (House 26; Fig. 3) and 124 (House 47; Fig. 4; cf. Radovanović 1996: fig. 4.4). In addition, around 2 years and 8 months old child Burial 92 (House 28; Fig. 18) was found in the extended position on its back. A number of skeletons were found in flexed/contracted positions — Burials 94 (flexed/left), 101 (flexed/right) and 102 (flexed/left) in House 24 (Fig. 16), Burials 98 (flexed/left) and 103 (on the back, contracted legs) in House 19 (Fig. 13), Burial 108 (on the backs, contracted legs) in House 4 (Fig. 7), Burial 109 (on its back, flexed legs) in House 4′ (Fig. 7) and Burial 123 (contracted/left) in House 47 (Fig. 4). From the description of these burials it is difficult to discern any pattern that would relate to the difference between those flexed burials lying on the side and those lying on their back with only the legs con-

tracted. Also, there is no clear pattern in the choice of the side for placement of the flexed/contracted burials. The third group of burials are infant skeletons found on their backs with legs splayed outwards and heals below the pelvis (in the 1970 Field Journal noted as sitting) — new-born Burials 114 and 115 (House 36; Fig. 9), 113 (House 63′; Fig. 17), 111 (House 38; Fig. 5) and 106 (House 4; Fig. 7). Child Burial 97 (2–6 years old) in House 31 was found in this last position too (Figs 19 & 20). To interpret this variety of positions we find the procedure of re-enacting and re-fleshing the body a very useful one (cf. Hawkes & Molleson 2000). The manipulation of the body prior to its burial (e.g. excarnation, dismemberment, binding/wrapping or burials in bags and baskets), processes of decay as well as later disturbances may all be relevant for the position in which we find the body (Hawkes & Molleson

142

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Figure 7. House 4 and neonate Burials 106, 107, 108 and 110 and House 4′ and neonate Burials 109, 109a.

2000). In addition, it is reasonable to assume that in some instances the position of the deceased can be explained by a chronological difference between interments, where customs about the way the deceased infants, as well as adults, are

placed change through time. The position of infants with their legs splayed outwards at Lepenski Vir indicates that possibly we need to account for some sort of treatment of the body prior to the burial. We suggest that at least some infants

143

The Iron Gates in Prehistory

might have been placed in a bag (which would hardly leave any visible trace) prior to their interment in a small burial pit. In the process of decay when a movement of the body parts reduces a corpse to bones that slump in a certain way, depending on the way the bag with the neonate’s body was placed in the ground, we may expect to find the deceased infants in varied positions. This would solve the unusual position of the legs splayed outwards, as in these burials an infant’s body was put contracted in a bag and might have been placed on its back; in the process of the decay of the bag and the body the bones of the legs would slump either outwards to both sides or both legs to one side. The latter instance, also, explains burials with the torso on the back while only the legs were found in contracted positions, as well as the reported unclear body positions. Furthermore, the bag might have reduced the immediate disturbance of the body, thereby accounting for the relatively good preservation of infant skeletons at Lepenski Vir. However, there remains a possibility that not all neonates and older infants were placed in bags prior to burial and treated in the same way, with or without chronological significance. Extended burials may have been placed directly in the ground. Also, is it a coincidence that three infant burials in House 24, all three found in the flexed/contracted position, are in the same building where two adult burials were placed on the floor (or possibly one in the infill) in flexed/contracted positions (Burials 8 and 9, cf. Srejović 1969a: fig. 72)? In chronological terms, the contracted/flexed position largely connects to the burial practices that characterize the Early Neolithic of the wider region (cf. Stanković 1992; Borić 1999; Minichreiter 1999). On the other hand, in House 47, Burial 123 was placed in the contracted position (perhaps placed in a bag) and Burial 124 in the extended position (cf. Radovanović 1996: fig. 4.4). Can this mean a change of burial custom/norm during the life cycles and rebuilding of House 47? Also, burials found with their legs laying splayed outwards were found mainly in possibly older buildings, completely or partly overlapped by later buildings (Burials 114 and 115 in House 36, Burial 113 in House 63′, Burial 111 in House 38 and Burial 97 in House 31 — see Fig. 1, cf. Radovanović 1996). In this case does the treatment of the infant’s body by putting it in a bag prior to burial represent a relatively older burial practice? All these questions related to the chronological significance of any patterning can best be answered through the application of AMS dating to infants’ bones. However, later we shall discuss the significance of the practice possibly identified here of putting the dead infant’s body into a bag in relation to an instance from recent Balkan ethnography. b) Orientation In those cases where it was possible to establish their orientation, it seems that a number of burials were oriented with the head toward the south east or the south, i.e. approximately parallel with the Danube with the head pointing downstream (see Table 1). This is important in connecting infant burials to other burials at this and other sites in the Danube Gorges. This particular orientation may have a chronological significance and may relate to an older phase of burial practices (Radovanović 1996), or even may indicate a hybrid re-

invention that kept its potency through narratives about past practices, preserving certain beliefs and conveying particular meanings. However, in several cases infant burials were placed with their heads pointing to the north west, north or north east. These variations were also noticed among the skeletons within the same building. These differences possibly reflect the ‘noise’ of individualized actions, realizations, and re-interpretations of memories about past practices, their effectiveness and potency (cf. Borić 2003). c) Burial pits Infant burials at Lepenski Vir were placed in pits that were dug a) mostly immediately behind the rear part of the limestone flooring (a number of instances), b) through the damaged parts of flooring in the rear end of a building, such as Burials 128–131 (House 27; Fig. 12) and 96 (House 43; Fig. 11), c) through the limestone floors, for example in the cases of Burials 111 (House 38; Fig. 5), 113 (House 63′; Fig. 17), 103 (House 19; Fig. 13), or d) within the stones of the rear of the building, as Burials 94, 95, 101 and 102 (House 24; Fig. 16). A number of burial pits were clearly visible in the virgin soil and had different shapes — the pits for Burials 113 (House 63′) and 94 (House 24) were rectangular, Burial 103 (House 19) was placed in an oval burial pit, while pits for Burials 106 and 107 (House 4), 128–131 (House 27) and 132 (House 37) were reported by the excavators as having irregular shapes. Burial pits were most clearly visible in those cases where a skeleton was placed in the virgin soil. It was much more difficult to spot the outline of a burial pit in the case of Burials 123 and 124 (House 47) that were dug through the floor of House 47 and placed in deposits that were covering an older building (House 47′) (see photo in Radovanović 1996: fig. 4.4). The same goes for those burials interred into deposits of occupational activities, possibly considerably older from those of the building floors (cf. Borić 1999: 53–54; Borić & Dimitrijević 2005; Whittle et al. 2002), such as neonate Burials 116 (House 13; Fig. 10), 119 (House 29), 133 (House 37) and child Burial 97 (House 31; Fig. 19). The burials were placed at differing depths, measuring from the level of the floor — from as little as 10 cm in the case of Burial 63 (House 26), 20–30 cm in the case of Burials 128–131 (House 27), 118 (House 62), 132 (House 37) and 134 (House 48), up to 60 cm in the case of Burial 116 (House 13). d) Use of stone in burials In several instances stones were used in different ways in the elaboration of a burial. In the case of Burial 63 (House 26) stones were placed vertically, two on the sides of the head and one on the side of the legs. In two more instances special concern for the infant’s head was expressed — the head of Burial 106 (House 4) was fixed between two stones and two small stones were placed next to the head of Burial 125 (House 54). There was a stone construction in the case of the rear of House 24 (Fig. 16) where Burials 94, 95, 101 and 102 were placed. A similar stone construction was found within the rear of House 43 where Burial 96 was interred. In the case of Burial 113 (House 63′), where the rectangular burial pit was cut through the floor of the building, the burial was covered by a stone slab after placing the body (Fig. 17). Also,

144

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Figure 8. House 54 and neonate Burials 125 and 127 and mandible Burial 126.

child Burial 92 (House 28; Fig. 18) was partly lying on a small stone slab of irregular shape while above it, on the level of the building floor, there was a large stone block. The excavators note that a piece of floor, with a depression on its upper side, was used to cover Burial 103 (House 19, Fig. 13; 1970 field diary).

Discussion An important conclusion to be drawn from the described instances of elaboration of infant burials at Lepenski Vir suggests that probably all of these burials were dug from already existing buildings with furnished limestone floors. On the basis of data presented, it is difficult to sustain the opposing

145

The Iron Gates in Prehistory

Figure 9. Houses 35 and 36 and neonate Burials114 and 115.

view that the burials were placed before the buildings were furnished, i.e. before the floors were laid (contra Radovanović 2000: 340, note 7), as some kind of foundation deposits (similar to the examples from Neolithic settlements in the Carpathian Basin, cf. Makkay 1983), or even as some sort of ritual sacrifice, indicating a practice of infanticide — for instance, Benac (1973) offered an interpretation along these lines for infant burials at the Early Neolithic site of Obre in central Bosnia. Although burial pits for infant interments were not recognized immediately from the floor level during the excavation of Lepenski Vir’s buildings, this can be attributed to the fact that the burial pits were small, perhaps

wide and deep enough just to place an infant’s body (possibly in a bag). Also, as indicated above, in many instances infants were buried just off the floor area in the rear of a building, frequently under the stone wall, which would tend to obscure an unexpected and small burial pit. To these reasons, one should add the occasionally crude excavation method. Here, we have to disagree with Srejović’s opinion that infant Burial 63 (House 26) (see Borić 1999: fig. 20) and an older child Burial 61 (House 40) were buried prior to the construction of these buildings (Srejović 1969a: 136; 1972: 119). Both burials were clearly recognized and excavated in the course of the excavation of these buildings. This would suggest that the

146

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Figure 10. House 13 and neonate Burial 116.

burials were noticed from the level of the building floors, and indicates that they were dug from this level. Furthermore, a representational boulder was placed above Burial 61 (see below), which is not a coincidence (contra Srejović 1969a: 136), and the limestone floor in the rear of House 40, where the burial took place, was significantly damaged (see Fig. 1; Srejović 1969a: fig. 65). It seems that in some instances infants were placed in already existing damaged areas of the building floor (e.g. House 27 — Fig. 12, House 43 — Fig. 11). These damaged areas may indicate that buildings were used over a long period of time prior to the burial of an infant. On the other hand, it is difficult to ascertain if perhaps these damaged parts of the floor expanded exactly as a consequence of digging a burial pit. Also, at present we have no way of telling if the buildings continued to be used after the burial of an infant took place. Combined AMS dating of the abandoned animal bones on building floors and infants’ skeletons may be a solution to the problem.

A final point in this part concerns two small fragments of Early Neolithic pottery found in the fill of Burial 113 (House 63′). As already indicated, this pit was covered with a stone slab and by the floor of later House 63 (Fig. 17). In the original field notes it was indicated that the pottery fragments within the burial infill were found in the part of the burial pit that was not covered with the stone slab and the possibility was advanced that these were deposited by the Danube. In our opinion, there is a slim chance for this scenario, and the fragments may represent either an accidental content of the pit infill or some sort of (fragmented) grave goods (cf. Chapman 2000). This instance, also, opens up the very contentious issue of the chronological position of the infant burials and buildings of Lepenski Vir I phase (cf. Borić 1999, 2002a, 2002b; Borić & Dimitrijević 2005; Whittle et al. 2002; Radovanović 2000). We shall return to this question. The contextualization of infant burials at Lepenski Vir takes us further to consider other features found within the

147

The Iron Gates in Prehistory

Figure 11

House 43 and neonate Burial 96.

buildings. Here we focus on two types of finds in particular which may relate to the infant burials on the level of building associations. These are articulated and disarticulated burials of other, non-infant, age groups found placed on or below limestone floors and the artworks of ornamented boulders and ‘altars’ (most likely serving as pestles) found in connection with building floors. a) Adult burials/disarticulated bones in buildings Disarticulated bones of adult individuals were found frequently in relation to the rectangular hearths in buildings. For instance, a human mandible (Burial 22) was found next to Proto-Lepenski Vir Hearth a (see Fig. 1; cf. Srejović 1969: 132). A striking case is Burial 21, a human mandible placed next to the hearth slab of House 40, the teeth facing down with a flat stone plaque blocking the open area between the caudal rami of the mandible (Srejović 1969a: fig. 70, 1972: 119, fig. 64; Radovanović 1996: fig. 3.48). This made a recognizable ‘A-support’ which, in the case of all other buildings, is made out of stone plaques only. These supports were placed in varying numbers around a number of the building hearths at Lepenski Vir.5 A human mandible (Burial 126) was found lying on a large rock below the rear of the hearth of House 54, with neonate Burials 125 and 127 buried under its rear (Fig. 8). Below the floor of House 31, where child Burial 97 was found, another isolated human mandible (Burial 105) was found approximately below the area of the hearth (Fig. 19). Also, a human skull with no mandible (Burial 122) was found in the rear part

of the hearth of House 47′ (Fig. 4), underneath the floor of House 47 that had neonate Burials 123 and 124 (cf. Radovanović 1996: fig. 4.4). It is important to emphasize the connection of human mandibles/skulls with building hearths. In other buildings with infant burials, disarticulated or semi-articulated and partly preserved extremital bones of human adults were found, again closely related to building hearths, and only rarely found in the other parts of buildings. In House 43, where neonate Burial 96 was buried, partly preserved and probably articulated bones of a human torso (Burial 104 consisting of clavicle, sternum and ribs) were found below the floor on the rear part of the hearth (Fig. 11). Also, in House 19 with neonate Burials 98, 98a and 103, on the sides of a shallow pit dug for the sculpted boulder and placed behind the hearth, disarticulated human bones (Burial 99 consisting of scapula and clavicle on one side and broken humerus on the other side of the pit) were found (Fig. 13). In House 24, with neonate Burials 94, 95, 101 and 102, besides three adult burials placed on the building floor or alternatively dug into the already existing infill of the building (these are articulated contracted Burials 8 and 9 and human skull Burial 10 — Srejović 1969a: fig. 72, 1972: fig. 63), partly preserved and semi-articulated human adult bones (Burial 100) were found below the rear of the building (Fig. 16). Another disarticulated human humerus (no burial number) was found in the occupation layer of dark soil (with animal bones) below the floor of House 29 with neonate Burials 119–121. Bones of a human foot were found below the floor of House 62′ in the north corner (corner B). This instance

148

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Figure 12. House 27 and neonate Burials 128, 129, 130 and 131.

may relate to neonate Burial 118 found below the rear of House 62 that overlapped older House 62′. Also, Srejović (1969a: 134–135, 1972: 118) mentions that in both Houses 3 (neonate Burial 112) and 54 (neonate Burials 125 and 127) one human femur was embedded in the floor behind the hearth construction. In some instances, disarticulated human bones were placed over building hearths. A disarticulated humerus (Burial 23) was placed over the hearth of House 48 with neonate Burial 134. On the floor of House 36 (neonate Burials 114 and 115), at the rear side of the hearth, disarticulated human adult bones were found (Burial 70) together with a dog mandible (cf. Radovanović 1999: 74).6 Certainly, some of these examples of disarticulated and

semi-articulated burials found below the building floors may come from burials considerably older than the occupation inside these spaces (Borić 1999: 59), disturbed by subsequent building activities. For instance, Burial 100, below the rear of House 24, or the bones of a human foot found underneath House 62′ may perhaps represent older, disturbed burials. However, the intentional placement of disarticulated human bones is indicated by their frequent placement in the area at the rear, narrow end of the hearth (below or over it). These bones possibly both circulated during the life of a building and/or are a consequence of an intentional deposition indicating the event of building abandonment (Borić 2003; 2005). Considering the choice of manipulated body parts, as pre-

149

The Iron Gates in Prehistory

Figure 13. House 19 and neonate Burials 98, 98a and 103.

viously noted (Srejović 1969a: 143–144), one can underline a special concern in the manipulation of heads and mandibles of the deceased. Mandibles and skulls of the deceased in many cases were found separated, as mentioned in some of the described examples of disarticulated burials. However,

some kind of dismemberment of skulls from mandibles was noticed in articulated burials too. For instance, a skull of child Burial 92 (House 28) was detached from the mandible and adult Burial 26 (House 34) was found with the skull detached from the body and turned to face the east, i.e. the

150

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Figure 14. Neonate Burials 98, 98a and 103 after lifting the floor of House 19.

Danube,7 while the mandible slumped as if left in the anatomical position (Srejović 1969a: 135; cf. Radovanović 1996: fig. 4.5). Also, articulated Burial 28 (‘House’ XXXIII) was found with the mandible, but no skull (Srejović 1969a: fig. 68). At Lepenski Vir, several burials were found in their anatomical position but with complete skulls detached from their bodies, such as extended Burial 54e (House 65/XXXV, Srejović 1969a: 137, fig. 67), extended Burials 15 and 16 (marked as in ‘House’ XXVII of Lepenski Vir II phase, Srejović 1969a: 137, fig. 18) and contracted Burial 19 (House 54/XLIV) where the skull of the deceased was placed on the stone slab that covered the burial (Srejović 1969a: 165, 1981: 8). All these instances, together with the practices mentioned in relation to infant burials, point to the possible metonymic importance of the head and the mandible, standing for the deceased individuals. The importance of human (but also animal) skulls and mandibles that stand for the whole body and the person, and their manipulations though cultural practices were widespread in the whole eastern Mediterranean during the Mesolithic and Neolithic periods (e.g., Mellaart 1967; Hodder 1990; Cauvin 1994). We may find some relevance in Srejović’s (1969a: 140, 1972: 117) inclination to see some disarticulated burials connected to building hearths as ‘hearth guardians’. Indeed, these might have been individuals whose death occurred long before the time of their deposition in buildings of Lepenski Vir. Their actual antiquity and possible place(s) of origin may be deduced in the future by applying AMS dating and trace

element analyses. Although skulls and mandibles had a special significance among the body parts, other skeletal parts were also manipulated and similar practices have been observed in other contexts of the European Neolithic (cf. Whittle 1996). At Lepenski Vir, we may claim with some certainty that these might have been observed as ancestors, whose identity and recognition varied extensively — from those very recognizable to members of a building through direct memory and/or generational narratives to almost anonymous individuals from a distant past. Furthermore, we may speculate that placement of parts of ancestors’ bodies next to the hearths might have meant that the area of the hearth was particularly inhabited with ancestral potency. These bones might have acted as second-class agencies (cf. Gell 1998) with powerful apotropaic potencies (Borić 2003). These, possibly protective, ancestors might have had special significance and importance for buildings and their members/visitors, perhaps especially for those building lineages where infants and children were dying frequently or were faced with other misfortunes. We shall elaborate further on this argument, but first we need to pay attention to the sculpted boulders that might have anchored ancestral potencies in a similar way, focusing social and collective attention. b) Artworks and their potencies Sculpted and aniconic boulders and pestles (‘altars’) were found mostly in the apparently fixed contexts of buildings at Lepenski Vir and only occasionally in a loose context (cf. Srejović & Babović 1983). They were found concentrated

151

The Iron Gates in Prehistory

Figure 15. Houses 21, 22 and 29 and neonate Burials 119, 120 and 121.

152

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Figure 16. House 24 and neonate Burials 94, 95, 101 and 102.

mainly at the rear of the hearths (similar to the disarticulated burials). In several cases, boulders were found also at the rear of the building and only rarely in its front part. Also, some buildings (including a few of those with infant burials) lacked both/either boulders and/or pestles. Their placement in buildings at Lepenski Vir, as well as their spatial patterning, must have depended on specific events related to the prac-

tices that took place immediately before a building was abandoned. The representational, purely ornamented and aniconic boulders and so-called ‘altars’ (i.e. pestles, sometimes ornamented too) might have been conceived as important heirlooms of individual buildings, closely fixed to the building hearths. We may underline the monumentality of some of the

153

The Iron Gates in Prehistory

Figure 17. House 63′ and neonate Burial 113.

boulders, being architectural parts of buildings, as in the case of the hearth of House 43 (Srejović & Babović 1983: 133) where the boulder acted as a side of the stone hearth. Srejović (1969a, 1972) drew attention to the perfectly spherical and ‘closed’ boulder as pregnant with a mysterious sacredness. We may also view them as bodies with animate characteristics, objectifying narratives and myths about past events, related to the whole community as well as individual buildings (Borić 2005). In particular, they might have been connected to mythical dimensions of anadromous fish (cf. Radovanović 1997), but also to heroes and petrified ancestors from the distant past. Being artworks with a special investment of intention and desire (cf. Gell 1998), these objects might have embodied special powers important for building members, both deceased and living. Here, we shall focus on those instances where sculpted boulders were placed on the level of the building floor immediately above the head of the deceased buried within buildings. In the case of Burial 7/I, a decorated boulder (Srejović & Babović 1983: 136, cat. no. 30; Srejović 1967) was placed on the forehead of the deceased male. The already indicated connection of sculpted boulders and disarticulated

burials as pregnant with ancestral (protective) potencies may be further affirmed, as a detached human (ancestral?) skull (Burial 7/II), with strong eyebrow ridges and heavily worn (possibly from handling), was placed next to the head of the deceased, facing him. Also, an aurochs skull was placed over the other shoulder of the deceased (Srejović 1969a: 136–137, fig. 69, 1972: fig. 61, 1981: 43a; Srejović & Babović 1983: 136; Radovanović 1996: fig. 4.3). Both the boulder and the human and animal skulls may have accompanied the deceased for apotropaic reasons. In addition, he was placed next to the hearth, behind its rear. Also, the burial pit for 7/I was clearly cut through the floor of House 21, while the skeleton was placed in the extended position with the head pointing downstream. In another instance already discussed, child Burial 61, around 7 years old (see above and note 7), was placed in the extended position with the head pointing downstream in a similar way, within a burial pit cut through the floor in the rear of House 40. Above the head of the deceased, a small sculpted boulder was placed, representing a human (child?) face with schematic mouth, nose and engraved eye circles (Srejović 1969a: fig. 65; 1972: fig. 59; Srejović & Babović 1983: 108). In both examples (with the

154

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Figure 18. House 28 and child Burial 92.

addition of Burial 26 cut through the floor of House 34 and placed in the extended position with its head pointing downstream, see above and note 7) the building floor was cut to place the deceased. This may be relevant for the way infants were buried by cutting through the floor or though the virgin soil in the area next to the floor, in a similar fashion and in some cases with the same orientation of the body, i.e. with the heads pointing downstream. We may underline that Burials 7/I, 61 and 26 were treated in a particular way and were specifically furnished/manipulated. Only child Burial 92 (House 28; Fig. 18) clearly shares this orientation, position and association with a sculpted boulder. Approximately above Burial 92 two sculpted boulders were placed on both sides of a large stone block in

the rear of House 28 (Srejović 1969a: fig. 24, 1972: fig. 14; Srejović & Babović 1983: 107). This skeleton was found without its skull (only the mandible was present), placed in the extended position with the head pointing downstream. The representational boulder ‘Adam’ (Srejović & Babović 1983: 107, cat. no. 1) placed in relation to this burial is of particular interest. It shows a human (child’s?) face with a schematic representation of its mouth and nose similar to those represented on the sculpture above the head of Burial 61. What appears different is the way the eyes are represented. Instead of carved circles, only two horizontal lines were carved, possibly indicating shut eyes. In this context, one must emphasize the fact that all other representational boulders found at Lepenski Vir showing a human face bear

155

The Iron Gates in Prehistory

Figure 19. House 31 with child Burial 97 and isolated mandible Burial 105.

eyes as carved circles (cf. Srejović & Babović 1983). The differences mentioned may have some importance in indicating cultural attitudes toward infants’ lives and deaths at Lepenski Vir through the medium of their representation, breaking down age stages of childhood at this site. There may be an important connection between the way the eyes are represented in the case of boulders above Burials 61 and 92 and the respective age stages of these individuals. Seen from an anthropological perspective in the treatment of artworks, eyes are ‘orifices’ that first open up the aniconic form of an idol and do not have representational significance only (cf. Gell 1998: 132, 147). First and foremost they serve to penetrate the mind, into the invisible ‘inwardness’ of an idol (Gell 1998: 132). From all the representational features of the sculpted face, eyes most strongly express the personhood represented. Hence it is no coincidence that an older child Burial 61 had eyes carved as circles, possibly indicating that on death it had reached a socially recognizable personhood, as a consequence of the rites of passage he experienced on the way to adulthood (cf. van Gennep 1960; Turner 1974). Although other representational boulders found at Lepenski Vir were not associated with any burial in particular, it may be possible to assume that they represent metamorphosed adults/elders/ancestors and this state of being might have been emphasized with two semi-circled carved lines that appear immediately below the engraved eyeholes on these sculptures (Srejović & Babović 1983: 113, 116, 118). Younger child Burial 92 in House 28 can be meaningfully connected to the sculpted boulder with ‘shut’ eyes, represented as two carved horizontal lines. At the time of the child’s

death, it might have been lacking socially achievable personhood, although recognizable in the social world of the community. Metaphorically and socially at the time of death, its worldly persona was not fully formed and hence the eyes could not have been represented opened up, facing the world. We may possibly indicate that it might have been still partly conceived as ‘boneless’ (cf. Astuti 1998; see below), prone and vulnerable to the malevolent forces abounding in the world. c) Identification of ‘ancestral’ potencies The examples of disarticulated burials and boulders described above indicate a possible connection of these features in the buildings of Lepenski Vir with the function of anchoring ancestral powers and elaborating the arena of domestic space with strong apotropaic associations. The placement of infant burials in a patterned way may have accounted for these potencies in the expression of lineage and communal care toward the deceased. Moreover, this care for the descent group is obvious in examples of the burials of other age groups, possibly depending on the constellation of events that surrounded their deaths. Considerable age differences of these deceased individuals and the infants were played out through spatial patterning of their respective burials within the building space. Some of the child and adult burials were commemorated with sculpted boulders, the iconographic representational features of which were context sensitive. Infants of neonate age were never accompanied with a boulder and the memory of their short existence was almost neutralized by placing them back into the earth (possibly in a

156

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

Figure 20. Child Burial 97 (House 31).

bag). However, significantly this happened within the space of the building. d) House 49 — visible infants? One example that possibly opens a perspective on the visibility of infants in the society of Lepenski Vir and, at the same time, makes reference to the elaborate building space is House 49. This is a miniature structure (1.4 m2) that is a proportionally reduced image of a normal-size building (Srejović 1969a: fig. 37, 1972: fig. 27). House 49 was placed below a large rock on the north-west side of House 47 and was conspicuously and exceptionally oriented with its front part facing the north-west, i.e. almost completely opposite to the way all other buildings at the site were oriented (see Fig. 1). The function of this building, along with another example of a miniature building, though less well preserved and slightly bigger, House 55 (1.9 m2), remained an enigma for the excavator of Lepenski Vir (Srejović 1969a: 71, 1972: 69). Srejović notes that nothing was found on the building floors although the small plaques, that formed the hearth, were burnt by fire. He suggests that these examples are architectural prototypes for other buildings at the site, representing blueprints (Srejović 1969a: 71, 1972: 69). However, a dif-

ferent interpretative account is offered here. Their construction may relate to children at Lepenski Vir with various reasons in mind. Some indications for the scope of reasons can be gathered from an ethnographic case describing the Nukak hunter-gatherers from the Amazonian rainforest (cf. Politis 1999a, 1999b). An instance was recorded among the Nukak where they constructed a miniature hut at a residential camp which in every specific constructional detail followed the outline of a normal size hut. Also, a photograph was taken of a child sitting within this miniature hut. The anthropologist notes a special reason for the construction of this hut and placement of the child under the hut as it may have been sick or faced with some other sort of danger, indicated by the child’s red painted face (Politis, pers. comm.; see the picture on the back cover of Cárdenas & Politis 2000). In the case of House 49 we may speculate about a possible window into practices at Lepenski Vir that might have related to children during their lives, involving issues of protection with a strong reference to the building space. Before continuing to draw upon wider meanings of similar cultural practices through other ethnographies, we turn to the diachronic perspective of the phenomenon of infant burials at Lepenski Vir, their regional and historical context.

157

The Iron Gates in Prehistory

Chronological and regional perspectives We arrive at the question of the chronological position of infant burials at Lepenski Vir. Clearly, this is inseparable from the question of the chronological position of buildings with trapezoidal limestone floors of Srejović’s Lepenski Vir I phase. It has already been noted that the previously suggested schemes for stratigraphic phasing of buildings and other contexts at this site by both Srejović (1969a, 1972) and Radovanović (1996) are problematic in a number of points (see above, notes 3 and 6; Borić 1999, 2002a, 2002b; Borić & Dimitrijević 2005). Also, their schemes partly ignore the existing radiometric 14C dates on charcoal (Quitta 1975; Borić 1999: fig. 7, 2002a: fig. 5) which indicate a time span of c. 6300–5500 cal BC, i.e. the period that coincides with the appearance and the duration of the Early Neolithic in the wider region of the central Balkans. This chronological position of occupation on the building floors was further reaffirmed by new AMS dates (Whittle et al. 2002). Among the new dates, one date (OxA-9055: 8445±60 BP), also, gives the first clear indication of much older deposits concealed by limestone floors and represented by a number of contexts across the site. There have been a few published accounts that discussed the presence of Early Neolithic pottery at Lepenski Vir in relation to the trapezoidal buildings, particularly by analogy with the neighbouring and contemporaneous site of Padina (Jovanović 1969; Borić 1999; 2002a), but also by revealing photos of in situ pottery on building floors of Lepenski Vir (Garašanin & Radovanović 2001). This contextual association that has a chronological significance may further be supported by the instance of the Early Neolithic pottery fragments found in neonate Burial 113 in House 63′ which, as already mentioned, is difficult to understand as intrusions from a separate and overlying Early Neolithic layer. By analogy with the buildings at Lepenski Vir and considering the fact that the deceased infants were buried by cutting through their floors, we can largely relate these interments to the same period that is attributed to the dated building, i.e. 6300–5500 cal BC. A recent extensive AMS dating programme even suggests that the period can be confined to 6300–5900 cal BC (Borić & Dimitirjević 2007). Localized diachronic changes However, changes in cultural practices of infant burial rites considering much deeper time depths may be observed in the micro-regional context of Lepenski Vir, within the Upper Gorges of the Danube. This is related to the site of Vlasac, situated c. 3 km downstream from Lepenski Vir. The deposits of this site are of considerable antiquity, especially in view of some of the AMS dated burials (Bonsall et al. 1997; Borić 2006) and possibly cover the period from at least 9500 to 5900 cal BC, indicating a Mesolithic as well as transitional/Early Neolithic development in the region. Around 87 burials with more then 119 individuals of different age groups were found at Vlasac (Srejović & Letica 1978: 53–82), buried over the indicated time span. There are 22 burials of infants (re-examined by S.S.). However, their mortality profile is different from and less uniform than the infant interments from Lepenski Vir, but also the treatment of their bodies prior to the burial and their burial contexts are

considerably different (Borić & Stefanović 2004). At Vlasac, the mortality profile is not dominated by any particular category of infants expressed in gestational weeks on the basis of the morphometric analyses (see above) and we shall discuss elsewhere the results of these findings. It is sufficient to say that the age in gestational weeks of the infants from Vlasac varies between 26–28 and 35–37 weeks for those that would correspond to the foetal category, and 38–40 to more than 47 weeks for neonates and older infant categories. Infants of foetal age most likely indicate miscarriages and in some instances this is reflected through the burial rite. The most dramatic example is Burial 67, a female who probably died during pregnancy or in giving birth, found with the bones of a foetus in the pelvic area. The skeleton was covered with ochre, also buried in the red earth (Srejović & Letica 1978: 57). In other cases, infant bones were frequently found within the burials of adults (Burials 4b, 6a — on the shoulder of Burial 6 — ochre, 50/1 — ochre, 50a(1) — ochre and Cyprinidae teeth, 49(2) — completely covered with ochre, 55(2), 58b —adjacent to the legs of Burial 58, 60(1) — pelvic area covered with Cyprinidae [fish] teeth). These infant burials were sprinkled with red ochre over the pelvic area or over other body parts and/or placed in red soil, with addition of Cyprinidae teeth in several cases. On the basis of previous morphological analyses some of these burials were sexed as males and it remains necessary to re-examine their sex determination in this light. For instance, in Burial 60, determined as male and with the pelvic area covered with Cyprinidae teeth, a foetus — Burial 60(1) — was found. Furthermore, this individual was lying on the right side, facing three small pits in the immediate proximity with infant burials completely covered with red ochre — Burials 59, 61 and 62, containing more individuals (Srejović & Letica 1978: 60–61). Burial 62–62a with an adult individual and an infant was covered with red ochre and Cyprinidae teeth, with the addition of a flint nucleus (Srejović & Letica 1978: 61). In other instances there was no obvious connection between the burials of infants and those of older individuals (infant Burials 21, 42, 5, 10, 12b, 35a–b, 19 and 66a). Some of these were furnished in a particular way. For instance, particles of graphite were found within Burial 5 (Srejović & Letica 1978: 68). Infant Burial 42 was found without its head and accompanied with ochre and Cyprinidae teeth over its pelvis (Srejović & Letica 1978: 58, fig. XXXV/2). Also, infant Burial 21 was accompanied by some Cyprinidae teeth around its abdomen and with a necklace of 50 perforated snails (Cyclope neriteia) on its chest (Srejović & Letica 1978: 58). Infant burials from Vlasac indicate cultural practices that are different from those at Lepenski Vir, and among other things this is the function of diachronic changes. We may account for a considerable importance of expressing a mother–child tie in a number of burials at Vlasac, an emphasis on the danger of pregnancy and possibly a communal focus on a misfortune in respect to miscarriages. Some of the accents of practices, thus, differ from the underlying conceptual schemes at work in Lepenski Vir. Yet we may recognize a much wider underlying theme in these practices that might have been shared over a considerable time span with powerful elements of signification. Red ochre in particular, as well as some other elements present, such as Cyprinidae teeth and

158

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

possibly graphite, might have been used as powerful means of acting upon the body of the deceased infant and the accompanied adult (in some cases possibly mothers) with the intention of guarding their liminal experience of death, guiding them toward the underworld. The use of red ochre at Vlasac almost exclusively relates to infants and adult burials accompanied by an infant. Signifying qualities of red colour imminently relate to the body (cf. Turner 1967). The concepts of care, protection, but also danger in connection with birth, infants and pregnant women, are clearly present at Lepenski Vir too. We can advance the hypothesis that there remained a residual significance of red, used at Vlasac through the use of red ochre and reproduced at Lepenski Vir through the quality of red limestone floors, acting as the elaborated and powerful domestic arena. Still, this connection might have rather come from an unconscious production of signification, than as a conscious and meaning-laden realization (cf. Deleuze & Guattari 1984). Wider regional context Presently, no other infant burials are known from contemporaneous sites in the Danube Gorges. Infant burials do appear among a number of other sites in the central Balkans and across Southeast Europe, dated to the Early Neolithic (e.g. Benac 1973; Stanković 1992; Minichreiter 1999; Borić 1999). These sites are largely contemporaneous with the elaboration and occupation of trapezoidal floors at Lepenski Vir, also sharing similar pottery styles (cf. Whittle et al. 2002). For instance, a new-born infant burial was placed in a pithos that was found at the Early Neolithic site (c. 6100–5500 cal BC) of Anzabegovo in FYR Macedonia (Gimbutas 1976: plate 47; Garašanin 1982: 89; Nemeskéri & Lengyel 1976: 396, fig. 242). The pithos itself, with intentionally broken bottom and handles, was placed beneath the two adult burials arranged symmetrically in contracted positions. Also, infant bones were found in a central hole of one of two smaller quadrangular buildings of tamped clay excavated at this site. Another striking example is the Early Neolithic site of Obre in central Bosnia (Benac 1973). At this site 8 burials were found, and all belonged to infants and children (Nemeskéri 1974). Burials 1–4 were placed in contracted or flexed positions, Burial 5 was represented by a skull, while Burials 6–8 were largely scattered or it was difficult to determine their exact positions and orientations (Benac 1973: 347–363). It was suggested that Burial 7 was placed in the seated position (Benac 1973: 351). Grave offerings were found possibly accompanying Burial 1 (Starčevo type clay ‘altar’) and Burials 5 and 6 (most likely residual remains — potsherds, flint tools and animal bones), and clearly in Burial 7 (‘sun discs’ — one from fired clay and one from stone and potsherds) and Burial 8 (broken amber ornament, semicircular Spondylus ornament, two polished stone axes and two pots — one of the Adriatic type and another ornamented with ‘barbotine’ decoration with stone pebbles inside). Also, Burial 3 was placed over the area of burned clay and, as well as Burial 4, covered with a thin layer of clay. In the case of Burial 2, stones were placed around the skeleton, and Burial 5 was also surrounded with a circle of burned stone slabs.

Burial 7, even more elaborately, was placed on a stone base and surrounded by stone slabs (representing a stone construction of some kind). Scattered bones of Burial 8 were found on the clay floor, surrounded with stone slabs and the whole area covered by small stone pebbles, almost forming a small tumulus. Burial 8 with finds of two stone axes, largely influenced the excavator to interpret all the burials as sacrifices, possibly related to a sun cult (Benac 1973: 357–363). This interpretation may be regarded as rather naïve and without a serious grounding in the described instances. In the context of the present discussion we may assume some significance for the elaboration of clay floorings and, perhaps, the practice of covering some of the buried infants with clay. Single infant burials were found at several other Early Neolithic sites in the Balkans. At Obrež in the region of Vojvodina, a contracted burial of an infant with no grave offerings was found, being the only burial found at the site (Brukner 1960). Also, an infant burial was discovered at the site of Divostin in central Serbia (Zoffmann 1988). In the same region, at the site of Blagotin, an infant burial was found in the infill of a large pit-dwelling, covered with a thick layer of ash. This burial is AMS dated (OxA-8609: 7270±50 BP; Whittle et al. 2002), being contemporaneous with the earliest dates for occupation remains on building floors of Lepenski Vir (see above). In the layers of this feature some other finds were exceptional — two large figurines of fired clay, a grain model with incisions, a clay ‘altar’ above the layer of ash and amulets (Stanković & Leković 1993; Whittle et al. 2002). In the described examples across the central and northern Balkans, roughly in the same historical context, an occasional focus on infants and children is evident. At some of the sites infant burials, as at Lepenski Vir, are related to building floorings or less elaborated areas, but frequently to ‘special’ features of some sort. The focus of these practices may be of wider regional as well as cross-cultural importance. At other sites of the eastern Mediterranean infant burials are similarly found in buildings and below their floors in the pre-Neolithic–Early Neolithic phases. At the Early Neolithic site of Nea Nikomedeia in northern Greece, a woman and two children were found underneath the floor of one building (Rodden 1965; Hourmouziadis 1973; Gallis 1996). At Franchthi Cave in southern Greece, burials were found in both Mesolithic and Early Neolithic levels. Among nine articulated Mesolithic burials two were infants (Cullen 1995). However, in the Early Neolithic levels of the cave, eight burials belonged to infants or children. Also, one of these burials, a several weeks old infant, was accompanied by a fine marble vessel and a broken-in-half clay pot (Jacobsen 1969: 373–374, 380–381, 1976: 140, 142). In south-central Anatolia at the site of Çatalhöyük (c. 7200–6300 cal BC) infant burials were found during excavations in 1960s (Mellaart 1967). Describing burials with red ochre painting over the skull or the body (mostly females), Mellaart (1967: 207–208) singles out a burial of a prematurely born infant (VI.A.14) and a female burial with a child on top of her (VIII.1) which, among other goods in the burial, had “fresh water mussels filled with red ochre”. Also, red ochre covered the whole body of a young girl who had suffered a broken femur, with addition of cinnabar paint over

159

The Iron Gates in Prehistory

her skull (IX.1). Another child’s skull and the upper part of the body were again painted with red ochre (E.IV.8). In the course of new excavations at this site, more burials were found underneath building floors. For instance, in Building 1 in the North Area of the site a very large number of infants, young children and juveniles (more than half of all buried in this building) were found (especially confined to the early phases of the building’s use) among around 60 buried individuals (Hamilton 1996, 1997; Molleson and Andrews 1997: fig. 32–34; Hodder 1999; Archive Reports 1996–2000, http: //www.catalhoyuk.com/). In Building 1 under the north-west platform, an adult burial (1924) with an infant on top of it was uncovered (Hamilton 1996) while, similarly, in the foundation layers of Building 1, a neonate (2532) was placed on the head of an adult (2527), with a fragment of red painted plaster found next to the head of this individual (Hamilton 1997; Hawkes & Molleson 2000: fig. 13.6). Infant burial 2105, found under the north-west platform, was possibly wrapped with several hundred white and pink limestone beads (Hamilton 1997). It was suggested that the posture of neonate burial 2197, found buried on a grinding stone in the foundation layer of Building 1 (Hamilton 1997), with legs and arms splayed outwards, indicates that it was placed in a bag (Hawkes & Molleson 2000: 160). Another practice related to infants in this building is the placement of three neonates (2199, 2197 and 2515) in a row at the threshold of the crawl hole (connecting two parts of the room). A specific physical intimacy is also seen in placing an infant (1450) facing an old female skeleton (1450) (Molleson & Andrews 1997). Also, continued work within the South Area of Çatalhöyük (where Mellaart’s excavation took place) confirmed the same pattern of a relatively high proportion of neonates/infants found in buildings (Hamilton 1999). In Building 6 of the South Area, six infants, two adults and one adolescent were excavated. Two infants had red pigment (ochre) applied over them, while three of them were placed in baskets. One of these (4406), covered with ochre too, had strings of beads around each wrist and an ankle. Another infant placed in a basket (4927) was covered by a yellow substance (ochre?). In Building 17, two other babies were covered one with a yellow substance and the other with red pigment; the latter was placed in a basket. Also, it was indicated that in several instances infants were placed behind ovens and fire installations (such as the only neonate burial in Building 18, i.e. Mellaart’s shrine X.8). In the South Area, in Space 112 (Mellaart’s Shrine VII.9) a partly preserved infant 2017 and, also, infant 2779 were found associated with a fire installation. Within the same space, again, the posture of infant burial 2728 indicated that it was possibly placed in a bag prior to its burial. Within Space 109, a neonate (2772) was placed between the platform and the fire installation. Its legs were splayed and it was overlaid with a horncore (Molleson et al. 1998). It is interesting to note that some of the burials with traces of chronic bone pathologies had red ochre applied to their skulls (Molleson et al. 1999). A very similar pattern of a relatively large number of buried infants, children and juveniles has been confirmed also in the course of the ongoing excavations of Building 3 (BACH Area) at the site (Hamilton 1999; Molleson et al. 2000). At sites of the Natufian culture, such as Mallaha (Eynan)

(Boyd 1995), infant burials were found associated with architectural features too. There is an important connection in the accentuated appearance of infant burials within elaborated structures/buildings, spaces with both domestic and sacred elements, across the eastern Mediterranean in the period prior to and in the course of the development of the Neolithic phenomenon. All the differences and specific developments involved in these shared practices and meanings would merit a much more detailed discussion elsewhere. Now, we need to go beyond the temporal and spatial context of our case study and beyond archaeological vestiges that limit accessing immaterial aspects of social life in their contextual richness and diversity.

Possible meanings: Balkan ethnographies and beyond It does not require moving far from the hinterlands of the Danube Gorges and the central Balkans to start an ethnographic pursuit for the meanings of infant burials at Lepenski Vir. The ethnography of the Saracatsans and the Vlachs, pastoralist groups of the various mountainous regions of the Balkans (eastern Serbia, Thrace, Macedonia, Epirus, Thessaly), who culturally (and possibly genetically) represent one of the oldest surviving populations of the Balkans, is extremely rich (Høeg 1925; Kavadias 1965; Campbell 1964; Antonijević 1982; for a general history of the Vlachs see Winnifrith 1987). In our opinion, this ethnography carries specific significance for understanding numerous practices and patterns seen throughout the prehistory of the Balkans. Although it would be difficult or even impossible to pin down the ways of transmission of distant cultural practices and memories through narrative and material forms or to follow their genealogies in fine detail, it seems that in many instances the former practices stay rooted in the local landscapes long after the generations of past inhabitants are gone and their dwelling places abandoned (cf. Connerton 1991; Fentress & Wickham 1992). We are left with two legitimate possibilities in this context, that is either to create imaginary genealogical narratives (cf. Tringham 2000: 126)8 or to study rich details of local ethnographies against archaeological examples in a more conventional way. We explore the latter possibility here. The religious beliefs of the Saracatsans and Vlachs are overwhelmingly occupied with issues of death and protection from evil forces. Their mythology, although under strong influence from Christianity, took a specific eclectic form with numerous hybrid features. Firstly, there are intrinsic elements that connect domestic hearth and fire to mother and new-born child in the ethnography of these groups. Thus, among the customs of the Saracatsans a woman would give birth on the ground next to the hearth and fire, which is believed to protect the mother and the child. Another noted custom is to draw a hearth on the body of a new-born child with a stone, previously put in fire. These customs, together with many others, may fall under the rites of protection and it has been noted for some time now that among the Bulgarians and the Slavs in general, but also other European peoples, the emphasis was

160

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

frequently placed on practices that serve to shelter pregnant women, foeti and children from the dangers of malevolent forces (van Gennep 1960: 45). For these reasons, a number of rites of separation and incorporation that frequently relate to buildings, i.e. to the time of their occupation, show incredible richness across ethnographic contexts. The main aim of these practices is, first, to introduce a woman and the society into the ‘abnormal’ state caused by a woman’s pregnancy and, after the birth, to slowly incorporate the mother back into the normal flow of life, as a “social return from childbirth” (van Gennep 1960: 46). However, in the event of a new-born’s death, more dramatic practices may take place. According to one recorded custom that was widespread among the Saracatsans up to 1940 (Hatzimichali 1957: 68, cited by Antonijević 1982: 134), in the event of the death of a new-born or a miscarriage the child would be put in a rawhide bag full of salt and hung above the parents’ marital bed for approximately forty days, until the child’s corpse dried out. A rite of separation of the mother (she could not leave the hut during this period) from the community is involved in the custom. Also, every morning, the mother would draw a cross (a sign believed to possess special protective powers) on the earthen floor of the hut. After this period the corpse of the deceased child in the bag is buried inside the hut, in a burial pit dug in one corner. This custom strikingly resembles the pattern of the previously described infant burials from Lepenski Vir. But apart from a possible formal similarity, we may discern elements that can be of wider importance. Firstly, the burial takes place inside the hut, where life continues, and may relate to a belief that by this practice the fertility of the conjugal couple would be increased and facilitate a future delivery of a healthy child. The practice of keeping the deceased child in the domestic hut and spatially close to the parents for forty days (but also its subsequent burial inside this space) indicates a strong tie of the deceased child with the parents, and especially the mother. It is possible to interpret this as an act of specific care and solidarity, among the Saracatsans and the Vlachs very pronounced in relation to all the dead (Antonijević 1982: 137) and specifically in relation to children. Secondly, although the salt used in this custom may have the practical function of embalming the child’s body to some extent (Antonijević 1982: 135), in relation to local beliefs the salt also has a strong demonic power and through its use positive contact would be established with the demons (on the similar significance of salt in other parts of the eastern Mediterranean, see Welters 1999c: 64). Hence, the issue of protection is involved again. The third element in this particular ethnographic example that may be of some relevance for our archaeological case relates to the meaning of the earthen floor of the hut under which the corpse of the deceased infant is buried. As the earth may be conceived of as the place of the dead, by burying the deceased through/under the earthen floor of the hut the deceased infant is returned to the place of its origin. Van Gennep (1960: 52) points out that similar practices of protection in relation to mother and infant may relate to the fact that the infant who dies before its introduction to and incorporation into both social and celestial worlds lacks spiritual powers, i.e. a ‘layered’ social persona gained through

social experience. It relates primarily to the age of an infant, and across ethnographic examples world wide (including the Saracatsans and the Vlachs) special care is expressed in relation to the liminal (transitional) period after the birth. Thus, by customs such as preserving the umbilical cord or placenta after the birth and burying them in a distant place (where nobody can touch it) or under the threshold, hearth, etc., it was intended to preserve the child’s personality, its soul (van Gennep 1960: 52, see below). In various ethnographic examples, the crucial emphasis is on the time that elapses before an infant acquires full individuality, which can take days or weeks after the birth, and sometimes also can be relevant much later throughout the period of the child’s growth toward adulthood. During this whole period, the child remains highly vulnerable and different rites of protection are needed. This may equally apply to the soul of the deceased infant that requires special care in the liminal/transitional phase on the journey to the underworld. Another significance of the custom whereby the deceased new-born stays above the marital bed for forty days may relate to the belief held across the Balkans that forty days represent the first phase on the way to the underworld, a voyage that ends only after one year (e.g. Bandić 1997). For instance, a crucial event after birth that determines the treatment of an infant among the Slavic inhabitants of the Balkans and Eastern Europe relates to whether the infant is baptized or not at the moment of its death (Čajkanović 1924: 60-61, 128). Beliefs among the Saracatsans and the Vlachs focus on the supernatural and on practices that involve black magic, with a specific belief about the ‘evil eye’ as a source of danger (Antonijević 1982). The belief in a harmful ‘evil eye’ (especially threatening those in liminal stages of their lives, such as new-borns, newly-married, pregnant women, etc.) is widespread across the Mediterranean (cf. Welters 1999c: 64ff.; Mladenović 1999: 101ff.). As already emphasized above in the discussion on the significance of the representation of eyes on the artworks of Lepenski Vir, these orifices of the body can both expose the hidden interior of the being/mind (Gell 1998: 147) but also penetrate the body and possibly harm it. The world of the Balkan pastoralists and their whole mythology can be described as inhabited by demons, ghosts, monsters and fairies, with both malevolent and benevolent intentions, although the Saracatsans believe that malevolent forces are dominant. Such forces inhabit the landscape and are present in the air, mountains, rocks, streams, wells, trees, leaves — almost everywhere – and are called different names and possess powers that can be in varying degrees harmful to human health, good fortune or life (Antonijević 1982: 162–170). It has been recorded that in the region of Homolj (eastern Serbia) among the Christian Serbs, during the periods when frequent deaths of new-borns occur in a buildinghold, immediately after a birth, a midwife takes the new-born and passes it though the circle frame of a wheel, sways it three times toward the east and asks the mother while standing on the building’s threshold (where the ancestors of the building reside): ‘Do you like the sun or the moon?’. And the mother each time answers from the building: ‘the sun’ (Čajkanović 1924: 130), i.e. this world. In the mythology of the Balkan

161

The Iron Gates in Prehistory

pastoralists, the moon, although sometimes signifying fertility, equally relates to the underworld since it goes through periodical changes, indicating the termination of life cycles and rebirth (Antonijević 1982: 60–61). The practice of passing the child through, in the previous example, indicates the performance of an early, ‘prematurely’ evoked rite of passage to achieve the ‘transference of evil’ (cf. van Gennep 1960: 52, 59). It ensures the child’s smooth incorporation into the society as the first step in preparation of later rites of passage on the way to gaining a personhood. However, a different connotation of the rites of protection among Balkan pastoralist groups, but also the Slavic inhabitants of these areas, in relation to new-borns and pregnant mothers is the notion that these individuals can be dangerous for other members of the society (Bandić 1980, 1997). This comes from the ambiguous associations of pregnant women and infants who, although vulnerable and in need of protection, can be of considerable danger for the community, causing serious illness and even death. In order to protect the collectives, a number of customs relate to the isolation of a mother and child during the period of pregnancy and for some period after the birth (Bandić 1980, 1997). The custom of presenting children to ancestors of the building among some groups (van Gennep 1960: 61) is grounded in the notion of apotropaic powers held by the ancestors, that can protect the child. In Balkan ethnographies the threshold and the hearth are the main places in a building where the spirits of the ancestors reside (e.g. Čajkanović 1924: 127ff.; for a cross-cultural example of the significance of the threshold, see Eliade 1957: 25–27, 181). The cultural attitudes toward pregnant mothers, new-born babies and deceased infants in the aforementioned examples from Balkan ethnographies indicate a very strong emphasis on rites of protection. Similarly, a recent comparative study of folk dress in numerous examples across the Balkans and Anatolia, also explicitly emphasizes the obsession with the intertwined issues of protection and fertility, especially related to mothers and children (Welters 1999a, 1999b). Going beyond the Balkan ethnographies, there are two relevant issues that can enrich the present discussion with a wider cross-cultural and comparative perspective. The first concerns the powerful place of architecture, domestic spaces and buildings in transcending issues about protection and apotropaic potency anchored at these places. As already noted, if we accept that buildings objectify social units, serve the role of outer shells to descent groups that belong to these places, both dead and alive, reminding of complex biographies of places and past inhabitants, their multifaceted nature reaffirms the significance of the notion of ‘building societies’ formulated by C. Lévi-Strauss (cf. Carsten & Hugh-Jones 1995; Gillespie 2000a). Houses encapsulate tensions of complex social webs of blood and affinal kin relations and transcend long memories, in a way fixing the ‘essence’ of a descent group. Still their gendered and androgynous images remain unfixed as they move through transformations that alter their ‘essence’, inseparable from the flux of the lives of people that inhabit or visit these places. The building of a new building is the birth and growth of an objectified descent group and this notion is transparent in the rituals involving the placenta of new-born babies. For in-

stance, in Tana Toraja, South Sulawesi, Indonesia, through successive generations, fathers would bury placentas (seen as a ‘twin’ to the baby) of new-born babies always at the same spot at the east side of a building, which is associated with life and the rising sun (Waterson 1990: 198, 2000: 180, 182). In other Indonesian societies, such as the Tetum, the umbilicus and placenta would be placed in a bag and hung on the central (ancestral) pillar of the main room in a building. Similarly, in Tanimbar the placenta would be buried in the building floor (Waterson 2000: 180). In another example from the same region, in the village of Ara, among the Makassarese of South Sulawesi, Indonesia, the birth is surrounded by extensive magic rituals intended to facilitate the delivery and protect the baby (but also the mother) from evil forces and spirits that can harm the baby while still “in the extremely vulnerable condition of having an unhealed navel and soft fontanelle” (Gibson 1995: 137). The following birth customs make a reference to the building: “The bloodied banana leaf on which the child was born is folded up and placed in a bamboo pole with a basket at one end called a tompong. ... The tompong is planted in the ground beneath the building with its ‘mouth’ open to the sky. It is thus able to catch the dalle, ‘good fortune’, that descends from heaven. ... The tompong is left under the building for seven days, by which time the child’s navel is healed and the next ritual can be performed. During this time, the child must remain in the building, and the mother is not supposed to leave it either, for fear her ‘open’ state may attract evil spirits to enter her which she would then bring back with her into the building.” (Gibson 1995: 136–137)

In Ara, a number of apotropaic rituals that are performed in the construction, i.e. ‘closure’ of the building are closely analogous to those rituals that are performed with the newborns in attempts to ‘close’ their navel and fontanelle (Gibson 1995: 144–145). Away from Indonesia, among the Mam Maya of Chimaltenango in highland Guatemala, to bury the afterbirth in the father’s sweat building (the focal ritual structure of a family) emphasized the patrilateral extended family in which future ritual activities would take place (Gillespie 2000b: 219, note 8). Differently in terms of kinship associations, among the Zafimaniry of Madagascar a woman, who is pregnant for the first time, leaves the husband’s building and goes to her natal building. Here, another protective layer is made by creating a ‘building’ on the bed after the woman gives birth. The child is connected to the mother’s building as the ritual of putting the soot from the hearth of the building on the child’s forehead and burying of the placenta (here, as in Indonesia, believed to represent a twin) are done in the mother’s parent’s building. Only the third child would be born in the parents’ own building and through the same ritual materially associated with that building, which at the same time becomes complete (Bloch 1995: 77). In the same society, Bloch (1995: 83) records an instance where the ill new-born was cured through a ritual medicine soup cooked on the place of the hearth of the holy building (where previously the head of the village was buried) by feeding the baby with the soup and pouring it in places where the posts once stood. Frequently, reasons to abandon a building are related to events that are interpreted as the result of misfortune. Thus, among the Ye’cuana of Venezuela, the building is abandoned if either illness or the deaths of infants or a village leader is

162

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

associated with the building. Also, the deceased can be buried under the floor of the building (Rivière 1995: 197). The second relevant cross-cultural issue relates to the social construction of age and gender in human societies. As the examples from Lepenski Vir indicate, there was a significant difference in the way that neonates of the same age are buried in comparison to the burial rite that accompanied other age groups. As emphasized, the construction of these age (and possibly gender) differences (Borić 2005) might have been inscribed in the boulder artworks found at this site. It is of relevance here to describe the notion of a ‘boneless’ baby that comes from the Vezo of Madagascar (Astuti 1998). The Vezo think that people are not born as humans, but they become humans, with a rather non-essentialist view on identity. They see new-born babies as strongly tied with mothers, vulnerable and not fully human. Moreover, new-born babies are malleable and plastic, especially in regard to their facial features, and their bodies are soft and weak, i.e. ‘boneless’, and only have the potential to become ‘fully-boned’ adults (Astuti 1998: 36). Both mother and baby are vulnerable after the event of birth as their bodies can be penetrated by the deadly ‘air’ and the babies must be protected by wrapping up in layers of clothes (Astuti 1998: 35). The babies must not be left alone as they are unprotected from the harmful influence of passing spirits known as angatse, and the angatse are the reason that eyelids of babies tremble and roll sideways every time the spirits are around. Another kind of danger comes from vengeful ancestors who can be bad-tempered and can easily harm a ‘boneless’ baby. If the baby dies before it is one year old, it cannot be buried in a family ancestral tomb as it has not become a human, i.e. its social person has not been ‘created’ yet (Astuti 1998: 36–37). It is interesting to note that in the case of the Zafimaniry of Madagascar, buildings, like people, also acquire ‘bones’ through their slow build up with more permanent material (from woven bamboo to massive wooden planks) and elaboration of specific decorative wood carvings, where the process of ‘beautifying’ the timbers forms a part in the process of the growth and successful marriage of the founding couple (Bloch 1995: 78ff.). The views held by the Vezo in regard to infants are instructive and this example triggers a more complex discussion on the way the social person is constructed in different cultural contexts, relations of sex and gender, individuality and dividuality of personhood, etc., which we do not pursue directly here. However, this example shows some aspects of a non-essentialist view of identity that the Vezo themselves maintain, related to both collective and individual selfrealization, with the body as the arena of this expression. It points to a constant flux of becoming in the construction of a social person, in this respect similar to the way the Melanesian personhood is constructed (cf. Strathern 1988; LiPuma 1998).

Conclusions The study of burials of neonates and children from Lepenski Vir, together with apparent diachronic changes at the microregional level in burial rite and its elaboration in relation to these age groups at the neighbouring and earlier site of Vlasac, brings an invaluable corpus of data that triggers

questions of much wider significance. We have tried to trace an ‘ethnography’ of neonates’ and children’s burials in the Upper Gorge of the Danube that probably involves at least part of the timerange from 9500 to 5500 cal BC. We can point to several important findings that the data and their contextual analyses imply: 1. On the basis of osteometric analyses of 38 infant burials found underneath the floors of buildings at Lepenski Vir, it is possible to confirm a high peak of individuals aged at 38–40 gestational weeks (± 2 weeks) (see Fig. 2), indicating that most if not all of the infants buried underneath the rear of building floors at this site were neonates, i.e. newborn babies who died soon after birth. 2. The neonates were confined to 19 (out of c. 73) buildings at Lepenski Vir, with the addition of two younger children — Burials 97 (House 31) and 92 (House 28) — buried in two other buildings in a similar way. As with other burials, neonates and infants are absent from groups of buildings in the north east of the site. Houses accommodated from one to six neonate burials and there was no clear patterning related to the deceased’s sex, as mostly individuals of both sexes were found in buildings with more than one interment. 3. Almost all neonates were buried in the rear of buildings, either through their floors or immediately next to the limestone floors into the virgin deposits of the loessic slope into which the trapezoidal buildings were dug. The only exception, Burial 63 (House 28), relates to the unusual orientation of this building, indicating that the topographic perspective was of considerable significance concerning the burial rite of this age group. On the basis of present findings, we reject the possibility that the infant burials were interred before the building floors were furnished (contra Srejović 1969a: 136; Srejović 1972: 119; contra Radovanović 2000: 340, note 7). 4. There were no grave offerings in neonates’ burials. On the basis of the varied positions in which they were found, and with the help of the perspective of re-enacting the buried body, we suggest that some, if not all neonates, and possibly younger children as well, might have been buried in bags. These were buried in small burial pits, sometimes covered with stones and stone slabs, or in one case with part of the floor. Both their positions and orientations varied even within the same building. 5. In a number of buildings with neonate burials, both articulated burials on the floor level and, especially, disarticulated bones of adult individuals were found in the area around the hearth of the building. In the latter case, these were detached skulls and mandibles, but also other disarticulated bones of the postcranial skeleton. We suggest that these bones might have encapsulated powerful, apotropaic forces and agencies that might have served varied purposes in the lives of the inhabitants of, and/or visitors to, these spaces, and possibly specifically relating to the buried remains of the neonates. However, significantly, no disarticulated adult (ancestral?) bones were found in the neonate burials. 6. Apart from the neonates and children, only three other burials at Lepenski Vir (7/I, 26 and 63) were dug through the building floor. Strikingly, all three were placed in ex-

163

The Iron Gates in Prehistory

tended positions on their backs oriented with the heads pointing downstream, and behind/next to the rear of the hearths of the buildings, respectively. We suggest that this patterning in their placement significantly relates to their age (probably socially fully embodied persons), hence the difference compared to the neonates’ and younger children’s burials. However, rites involved in their burial might have carried the same prophylactic concerns toward the deceased’s corporeal selves, similar to the neonates’ burials. Incorporation of ‘ancestral’ bones, such as the skull (Burial 7/II) in Burial 7/I points in this direction. 7. There were no representational/decorated boulders (mainly concentrated around hearths) commemorating any of the buried neonates. In contrast, above two- to three-year-old child Burial 92 two sculpted boulders were found, one with a unique iconic expression (visible but ‘closed’ eyes — carved horizontal lines), possibly related to his unfinished social embodiment. Significantly, Burial 63 (House 40) of an approximately seven-year-old child was commemorated with a small boulder, representing a (child?) face with eyes ‘opened up’ as engraved circles, as on all other representational boulders (cf. Srejović & Babović 1983), possibly indicating stages of socially inscribed personal embodiment. 8. At the neighbouring site of Vlasac, there are proportionately fewer infants and children among the discovered burials (Borić & Stefanović 2004). On the basis of osteometric parameters there is no dominant infant category expressed in gestational weeks. A number of foetuses and neonates were found, in several cases spatially closely tied to adult individuals, possibly mothers (also within the mother’s pelvic area), indicating miscarriages that may have caused the death of the mother too. A number of whole infant burials and frequently groin areas of females were sprinkled with red ochre and/or in several cases fish (Cyprinidae) teeth. The red ochre might have been of apotropaic significance. We may assume that there is intrinsic, if only not fully consciously-made, connection between the ochre of red colour used in Mesolithic burials at Vlasac and the red colour of later limestone floors at Lepenski Vir. The Vlasac examples indicate significant diachronic changes within the same cultural milieu. 9. The neonates and children discovered at Lepenski Vir were buried through already existing building floors. Considering a number of radiometric dates for the occupation and wooden posts of upper structures of these buildings, we may assume an Early Neolithic date for the discussed burials, i.e. 6300–5500 cal BC. This is also confirmed by the reported presence of fragments of Early Neolithic Starčevo-type pottery in the hardly disturbed context of neonate Burial 113 (House 63′). This conclusion can be of significance in the attempt to situate historically (and possibly culturally) the particular phenomenon of infant/child burials and buildings, i.e. their placement within the elaborated domestic space. This is related to a number of examples of infant/child burials found underneath building floors across Pre-Neolithic and Early Neolithic societies of the eastern Mediterranean. 10. Cross-culturally, C. Lévi-Strauss’s notion of ‘building societies’ brings fruitful perspectives in considering the

connection of the paradoxical and illusory struggle to achieve ‘fixed’ descent groups, objectified through buildings (more precisely through the acts of their creation/becoming), and the ways in which social persons are constructed in a society, involving negotiation of personal and group identities and their engendered embodiments. These issues are central to the practices that surrounded the burials of infants and other age groups at Lepenski Vir. In addition to the ‘exclusivity’ of the conclusions listed above, we offer several less sharply defined concluding comments on the basis of the forgoing discussion. In his discussion of Lepenski Vir, Ian Hodder (1990: 25) indicates that death dominates the buildings of this site. In his opinion, at Lepenski Vir, like Çatalhöyük, ‘the domestication of death’ takes place by bringing the dead body into the building (Hodder 1990: 29). The power discourse is involved in this interpretation and the dead body is controlled and acted upon. Hodder (1990: 40) sees a significant relationship between the dying and decaying individual body and more permanent social units. As Meskell (1996: 7) argued in her critique of the social constructionists’ view of the body, especially in relation to the works of Michel Foucault, this is “the body as the scene of display”. Meskell (1996: 9) suggests that archaeological inquiries into lived and experienced bodies should escape the discourse of domination and control, something that was so frequently applied in interpreting evidence from the Danube Gorges and Lepenski Vir (e.g. Handsman 1991; Chapman 1993, 2000). The emotional force that might have surrounded the death of a new-born baby in a building of Lepenski Vir, causing feelings such as grief, bereavement and rage (cf. Rosaldo 1989[1993]: 1–20), although not directly penetrable for an archaeological study, must be emphasized, indicating a possible porosity of all previously established ‘fixed’ cultural patterns. We offer an alternative to the discourse of power and control over both the dead and the living body. We emphasize that the proliferation of neonate burials in buildings of Lepenski Vir expresses particular care in the face of a disrupting existential anxiety of facing the death of a new-born baby. On the other hand, it is believed that the deceased infant, already partly embodied, needs the protection of the building, ancestral powers and apotropaic potencies anchored in these structures, their hard limestone floors, hearths and sacred heirlooms, such as sculpted boulders. Personal embodiment probably played a significant part in determining the rites performed at Lepenski Vir. For instance, several older age burials placed within cuts made through the building floor were close to the hearths and were literally accompanied by ‘ancestral’ bones (Burial 7/I), while neonate and child burials were always placed away from the hearths, avoiding their direct contact with ‘ancestral’ bones. This may indicate a realization of a harmful influence of a direct physical contact between the deceased neonates and younger children with possibly easily angered ancestors (cf. Astuti 1998). We may assume that neonates and very young infants at Lepenski Vir were not born with an already formed social persona. It is possible that individuals who died immediately after birth were seen by society as not fully human and, had

164

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir

they lived longer, only with the passage of time would they have layered their respective social personae, creating the ‘essence’ of their selves and socially formed personhood. Still, there must have been some ‘essence’ of their embodied selves that was socially ‘invested in’ by burying them within buildings. Future work on the DNA-based formulation of relatedness among the burials of Lepenski Vir will provide further clues to the direction of what influenced the structuring of infant and other burials at the site in particular buildings. It may also be that although perceived as humans, neonates, infants and younger children were not considered capable of being protected on a dangerous journey to the underworld. This brings us to the question, what elements might have constituted a person at Lepenski Vir? Some of the ethnographic examples (e.g. Astuti 1998; LiPuma 1998) as well as those archaeological examples aided by textual information (e.g. Meskell 1996) indicate that persons might have been “multiply constituted throughout life and death and that the spiritual and corporeal selves were inextricably bound, rather than the Classical notion of the soul as entity imprisoned within flesh” (Meskell 1996: 13). This perspective gives significant weight to the materiality of the dead body as simultaneously it belongs both to the world of the living and to the ‘other’ world. As the body is invested through varied forms of social actions (e.g. initiation rites, tattoos) during its life, and rites that transform and guide the individual through the stages of social becoming (cf. Gell 1998: 86ff.), it remains similarly important to help the deceased body, inextricably connected with immaterial components that constitute it, to pass the thresholds of its metamorphosis in death. Although Meskell (1996: 11) critiques the preoccupation of current archaeological discourse with the exteriority, surfaces, treatment, elaboration and decoration of the body, which in her opinion neglects the embodied individuality, exactly these practices are central in the creation/becoming of the socially embodied person. All these practices of layering and wrapping of the human body (also by the building) bring to the forefront the processes of becoming and indicate the apotropaic salience of cultural practices. Particular examples of burials at Lepenski Vir may indeed indicate “drama and performance of some plot” (Hodder 1990: 29). This drama may be obvious in adult Burial 7/I as it was accompanied by animal and human skulls that might have been intended to aid his journey to the underworld. We may only assume that this individual might have died in a violent or unusual way, not ready for the perils of the journey to the ‘other’ world, desperately needing the guiding of this possibly ‘ancestral’ skull in particular. It seems that similarly the burials of infants and children at Lepenski Vir needed the sheltering proximity of the building and anchored potencies that were spatially played out in a specific way for this age group. As Balkan and other ethnographies inform us, a safe journey to the underworld was not of importance for the deceased only but also for the living, as the spirits of those deceased that are trapped in the liminal plane of existence, between the world of the dead and the living, can be of a considerable danger to the living. The notion of apotropaism underlying a number of cultur-

al practices is emphasized in the offered interpretation of neonate and infant burials at Lepenski Vir. Prophylactic concerns related to the human body are also significantly connected to the construction and elaboration of building and domestic area. This may be related to the pronounced emphasis on the descent-group and concerns on its ‘fixation’, historically constituted (Borić 2003). However, we must not stay detached, only looking at the transcended cross-cultural issues of political and social strategies at large. It is also necessary to approach empathy, admitting that practices such as events of infant burials at Lepenski Vir, were constituted by acts that are individual, emotional and experiential (cf. Meskell 1994). Thus, the ontology of care is highlighted in this interpretation, connecting concerns of larger social structures, buildings as objectified lineages on one side, with very individual and ‘noised’ actions of experiencing and emotionally driven individuals on the other. People at Lepenski Vir intentionally engaged in significant repetition of cultural practices, subscribing to and believing in their potencies. Notes 1. The archive field documentation used is stored at the Centre for Archaeological Research of the Department of Archaeology, Faculty of Philosophy, University of Belgrade. The late Prof. D. Srejović, principal excavator of Lepenski Vir, was affiliated to this Department. Permission to use the archive documentation was kindly granted by the Head of the Department of Archaeology, Prof. Živko Mikić. 2. The floors of almost all excavated buildings were relocated in the course of protection work in 1970 (Čanak-Medić 1971). It is important to note that the floors of Houses 13, 18, 27, 32, 37, 48 and 51 were moved as complete blocks (whereas some of the buildings owing to their large size were cut into several blocks), together with c. 0.75 m thickness of the deposits underneath the floor (Čanak-Medić 1971: 14). Although infant burials were also found underneath some of the buildings moved in this way (13, 27, 37 and 48), one could speculate that those buildings like Houses 18 and 51, where the whole building was moved in a single block, still conceal unexcavated infant burials together with other finds. 3. It is not possible to indicate the exact number of buildings for several reasons. First, some of the open-air hearths (e.g. Hearth a and d, see Fig. 1) have no floor around them and perhaps are much older than the later limestone floorings (Srejović’s Lepenski Vir I phase) and although these do not correspond to the notion of an elaborate ‘house’, they may indicate a domestic space. Second, in the cases when an older feature is overlaid with a later flooring (e.g. Houses 62/62′, 63/63′, 37/37′, 47/47′ and 26/26′; see Fig. 1) it is difficult to specify if the earlier features should be regarded as separate buildings. And third, as we largely reject Srejović’s Lepenski Vir II phase (Borić 2002a) the separate Roman numbers that he assigned to the phase II buildings do not indicate separate buildings but dry stone walls around the cuts of Lepenski Vir I buildings (see discussion in the text). 4. The only exception to this is Burial 83a–b. This burial was found in the occupation layer (quad. d/II, next to the corner of point c, excavation level XI, in the virgin soil; Field Diary 30/06/1970). Burial 83a is an older child (around 5 years old) placed in a contracted position, NW–SE orientation. Several potsherds, two pieces of flint, and snail shells were found in the burial. Together with this individual, bones of two infants (83b and 83b(1); see Table 1) were also found, partially preserved. After a taphonomic

A

165

The Iron Gates in Prehistory

5.

6.

7.

8.

examination of the bones from this burial by one of us (S.S.), it was noted that their fossilization (colour and bone consistency) is markedly different from the rest of the infants at this site. There is again no clear-cut pattern in the relationship between these supports and buildings with infant burials. For example, in some buildings, such as Houses 24, 47, 27, 43, 19 and 4, there were a number of these supports surrounding the hearths and, at the same time, there were from one to four or more new-borns buried in each building. On the other hand, infants were also buried in buildings with no supports found around the hearths (Houses 54, 37, 38, 29, 62, 26, 36 and 48). We may simply observe that relatively more infants were buried in buildings with more hearth supports (Houses 19, 4, 24 and 27). Srejović (1969: 140–141, 1972: 121–122) suggested that these supports stand for the deceased members of a building regardless of whether or not they were buried in the buildings. In his opinion, the community cared for the deceased who were represented by the supports (as schematic mandibles) and by placing the supports around the hearth the deceased would be able “to enjoy the warmth of the hearth and the food prepared on it” (Srejović 1969: 141, 1972: 122). Radovanović (1996) used the supports and the form of building thresholds as the main architectural features of the buildings for her re-phasing of Lepenski Vir’s stratigraphy, assuming uniform change through time reflected especially through the quantity of these supports around hearths. In her division, buildings with a large number of hearth supports would represent chronologically later features, while their smaller number or absence would suggest older buildings. This can only be partly relevant as it is obvious from the published photographs that a number of buildings that once had stone plaques were left with triangular holes as the stone plaques were removed (Borić 2000b). Although some profane function for these supports cannot be excluded (cf. Jovanović 1969), it may be that they were not placed once and forever and their presence or absence to some extent could indicate the changing status of a building in its life history, while indeed they can perhaps be connected to the burials and/or the deceased members of a building (Borić 2003, 2005). Ivana Radovanović (1999: 74) discusses Burial 70 in connection with dog ‘burials’, i.e deposition of dog bones at Lepenski Vir and Vlasac. Radovanović incorrectly attributed, discussed and phased this burial in relation to House 32 at Lepenski Vir, although it was actually found in House 36 (see Radovanović 1996: 185). A similar practice of turning the head to the east, i.e. to the Danube, is expressed differently in relation to Burial 61 in House 40. Although the skull of this burial is in the anatomical position, the sculpted boulder in the form of a human head that was placed directly above the skull of the deceased child was turned to face the Danube (cf. Srejović 1969a: fig. 65–66; Srejović & Babović 1983: 108) in the same way as the skull of Burial 26. Since the local Vlachs represented the excavation team in the course of excavation work at Lepenski Vir, this fact could have been used to view their interaction with and interest in the excavated residues of the past excavated in their local surroundings. This possibility is now largely lost.

Acknowledgements We are grateful to Prof. Dr Živko Mikić for his kind permission to work on the infant burials from Lepenski Vir. We also thank the organizers and participants of the conference, The Iron Gates in Prehistory (Edinburgh, 31 March–2 April, 2000) for their reactions and comments on the presentation of this paper. For valuable comments and information we would like to thank Katarina Novaković.

References Antonijević, D. 1982: Obredi i običaji balkanskih stočara. Belgrade: Balkanološki institut. Astuti, R. 1998: ‘It’s a boy’, ‘it’s a girl!’: reflections on sex and gender in Madagascar and beyond. In Lambek, M. & Strathern, A. (eds) Bodies and Persons. Comparative Perspectives from Africa and Melanesia. Cambridge: Cambridge University Press, 29–52. Babović, Lj. 1997: Položaj i funkcija svetilišta na Lepenskom Viru. In Antidoron Dragoslavo Srejović completis LXV annis ab amicis collegis discipulis oblatum. Belgrade: Centre for Archaeological Research, University of Belgrade, 95–108. Bagchi, K. & Bose, A.K. 1962: Effect of low nutrient intake during pregnancy on obstetrical performance and offspring. American Journal of Clinical Nutrition 11: 586–592. Bandić, D. 1980: Tabu u tradicionalnoj kulturi Srba. Belgrade: Beogradski izdavački zavod. — 1997: Carstvo zemaljsko i carstvo nebesko. Ogledi o narodnoj religiji. Belgrade: Biblioteka XX vek. Bass, W. 1987: Human Osteology: a Laboratory and Field Manual (3rd edition). Columbia, Mo: Missouri Archaeological Society. Benac, A. 1973: Obre I. A Neolithic settlement of the Starčevo–Impresso and Kakanj culture at Raskršće. Wissenschaftliche Mitteilungen des Bosnisch–Hercegowinisch Landesmuseums IIIA: 327–430. Bloch, M. 1995: The resurrection of the building amongst the Zafimaniry of Madagascar. In Carsten, J. & Hugh-Jones, S. (eds) About the House, Lévi-Strauss and Beyond. Cambridge: Cambridge University Press, 69–83. Bökönyi, S. 1969: Kičmenjaci (prethodni izveštaj). In Srejović, D. Lepenski Vir: Nova praistorijska kultura u Podunavlju. Belgrade: Srpska književna zadruga, 224–228. — 1972: The vertebrate fauna. In Srejović, D. Europe’s First Monumental Sculpture: New Discoveries at Lepenski Vir. London: Thames and Hudson, 186–189. Bonsall, C., Lennon, R., McSweeney K., Stewart C., Harkness, D., Boroneanţ, V., Bartosiewicz, L., Payton, R. & Chapman, J. 1997: Mesolithic and Early Neolithic in the Iron Gates: a palaeodietary perspective. Journal of European Archaeology 5: 50–92. Bourdieu, P. 1990 [1970]: The Kabyle House, or the world reversed. In Bordieu, P. (ed.) The Logic of Practice. Cambridge: Polity Press, appendix Borić, D. 1999: Places that created time in the Danube Gorges and beyond, c. 9000–5500 BC. Documenta Praehistorica 26: 41–70. — 2002a: The Lepenski Vir conundrum: reinterpretation of the Mesolithic and Neolithic sequences in the Danube Gorges. Antiquity 76: 1026–1039. — 2002b: Seasons, Life Cycles and Memory in the Danube Gorges, c. 10,000–5500 BC. Unpublished Ph.D. Dissertation, University of Cambridge. — 2003: ‘Deep time’ metaphor: mnemonic and apotropaic practices at Lepenski Vir. Journal of Social Archaeology 3: 41–75. — 2005: Body metamorphosis and animality: volatile bodies and boulder artworks from Lepenski Vir. Cambridge Archaeological Journal 15: 35–69. — 2006: New discoveries at the Mesolithic–Early Neolithic site of Vlasac: preliminary notes. Mesolithic Miscellany 18(1): 7–14. Borić, D. & Dimitrijević, V. 2005: Continuity of foraging strategies in Mesolithic–Neolithic transformations: dating faunal patterns at Lepenski Vir (Serbia). Atti della Società per la Preistoria e Protostoria della regione Friuli-Venezia Giulia 15 (2004–05): 33–80. Borić, D. & Dimitrijević, V. 2007. When did the ‘Neolithic package’ arrive at Lepenski Vir? Radiometric and faunal evidence. Documenta Praehistorica 35:.53–72.

A

166

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir Borić, D. & Stefanović, S. 2004: Birth and death: infant burials from Vlasac and Lepenski Vir. Antiquity 78: 526–546. Boyd, B. 1995: Houses and hearths, pits and burials: Natufian mortuary practices at Mallaha (Eynan), Upper Jordan Valley. In Campbell, S. & Green, A. (eds) The Archaeology of Death in the Ancient Near East. Oxford: Oxbow Books, 17–23. Brukner, B. 1960: Rezultati zaštitnog iskopavanja lokaliteta ‘Baštine’ kod sela Obreža. Rad vojvodjanskih muzeja 9: 81–111. Brünnacker, K. 1971: Geologisch–pedologische Untersuchungen in Lepenski Vir am Eiserner Tor. In Schwabedissen, H. (ed.) Die Anfänge des Neolithikums vom Orient bis Nordeuropa, Teil 2: Östliches Mitteleuropa. Köln: Böhlau, 20–32. Campbell, J.K. 1964: Honour, Family and Patronage. Oxford. Cárdenas, D. & Politis, G. 2000: Territorio, movilidad, etnobotánica y Manejo del Bosque de los Nukak orientales. Bogotá: Universidad de los Andes. Carsten, J. & Hugh-Jones, S. (eds) 1995: About the House, LéviStrauss and Beyond. Cambridge: Cambridge University Press. Cauvin, J. 1994: Naissance des divinités, Naissance de l’agriculture: La Révolution des symboles au Néolithique. Paris: Éditions CNRS. Chapman, J.C. 1993: Social power in the Iron Gates Mesolithic. In Chapman, J. & Dolukhanov, P. (eds) Cultural Transformations and Interactions in Eastern Europe. Aldershot: Avebury, 71–121. — 2000: Fragmentation in Archaeology. People, Places and Broken Objects in the Prehistory of South-eastern Europe. London: Routledge. Connerton, P. 1991: How Societies Remember. Cambridge: Cambridge University Press. Cullen, T. 1995: Mesolithic mortuary ritual at Franchthi Cave, Greece. Antiquity 69: 270–289. Currey, J.D. & Butler, G. 1975: The mechanical properties of bone tissue in children. Journal of Bone and Joint Surgery 57A: 810–814. Čajkanović, V. 1924: Život i običaji narodni. Studije iz religije i folklora. Belgrade: Srpska Kraljevska Akademija. Čanak-Medić, M. 1970: Projekat za spasavanje Lepenskog Vira. Saopštenja 7: 7–12. — 1971: Radovi na spasavanju Lepenskog Vira u 1970. godini. Dopunska istraživanja. Saopštenja 9: 5–17. Deleuze, G. & Guattari, F. 1984: Anti-Oedipus. London: Athlone. Eldredge, N. 1991: Fossils: the Evolution and Extinction of Species. New York: Harry N. Abrams. Eliade, M. 1957: The Sacred and the Profane. The Nature of Religion. New York: Harvest. Fentress, J. & Wickham, C. 1992: Social Memory. Oxford: Blackwell. Finlay, N. 1997: Kid knapping: the missing children in lithic analysis. In Moore, J. & Scott, E. (eds) Invisible People and Processes. Writing Gender and Childhood into European Archaeology. London: Leicester University Press, 201–212. Gallis, K. 1996: Burial customs. In Papathanassopoulos, G.A. (ed.) Neolithic Culture in Greece. Athens: Nicholas P. Goulandris Foundation, 171–174. Garašanin, M. 1982: The Stone Age in the central Balkan area. In Boardman, J., Edwards, I.E.S., Hammond, N.G.L. & Sollberger, E. (eds) The Prehistory of the Balkans; the Middle East and the Aegean world, Tenth to Eighth Centuries BC. (The Cambridge Ancient History, Volume III, Part I. 2nd edition). Cambridge: Cambridge University Press, 75–135. Garašanin, M. & Radovanović, I. 2001: A pot in building 54 at Lepenski Vir I. Antiquity 75: 118–125. Gell, A. 1998: Art and Agency. An Anthropological Theory. Oxford: Clarendon Press. Gibson, T. 1995: Having your building and eating it: buildings and siblings in Ara, south Sulawesi. In Carsten, J. & Hugh-Jones, S.

(eds) About the House, Lévi-Strauss and Beyond. Cambridge: Cambridge University Press, 129–148. Gillespie, S.D. 2000a: Beyond kinship: an introduction. In Joyce, R.A. & Gillespie, S.D. (eds) Beyond Kinship. Social and Material Reproduction in House Societies. Philadelphia: University of Pennsylvania Press, 1–21. — 2000b: Maya ‘nested buildings’: the ritual construction of place. In Joyce, R.A. & Gillespie, S.D. (eds) Beyond Kinship. Social and Material Reproduction in House Societies. Philadelphia: University of Pennsylvania Press, 133–160. Gimbutas, M. (ed.) 1976: Neolithic Macedonia as Reflected by Excavation at Anza, Southeast Yugoslavia. Los Angeles: Institute of Archaeology, University of California. Gordon, C. & Buikstra, J. 1981: Soil pH, bone preservation and sampling bias at mortuary sites. American Antiquity 46: 566–571. Guy, H., Masset, C. & Baud C-A. 1997: Infant taphonomy. International Journal of Osteoarchaeology 7: 221–229. Hamilton, N. 1996: The figurines and other small finds. Çatalhöyük 1996 Archive Report. http://www.catalhoyuk.com/archive_reports/ — 1997: Figurines. Çatalhöyük 1997 Archive Report. http://www.catalhoyuk.com/archive_reports/ — 1999: Figurines, burials/grave goods and beads. Çatalhöyük 1999 Archive Report. http://www.catalhoyuk.com/ archive_reports/ Handsman, R.G. 1991: Whose art was found at Lepenski Vir? Gender relations and power. In Gero, J. & Conkey, M.W. (eds) Engendering Archaeology: Women in Prehistory. Oxford: Blackwell, 329–365. Hatzimichali, A. 1957: Sarakatsanoi [Sarakatsans], 2 volumes. Athens. Hawkes, L. with Molleson, T. 2000: Refleshing the past. In Hodder, I. (ed.) Towards Reflexive Method in Archaeology: The Example at Çatalhöyük. Cambridge: McDonald Institute for Archaeological Research, 153–166. Hodder, I. 1990: The Domestication of Europe. Structure and Contingency in Neolithic Societies. Oxford: Blackwell. — 1999: Symbolism at Çatalhöyük. In Coles, J., Benley, R. & Mellars, P. (eds) World Prehistory. Studies in Memory of Grahame Clark (Proceedings of the British Academy 99). Oxford: Oxford University Press, 177–191. Hodder, I. & Cessford, C. 2004: Daily practice and social memory at Çatalhöyük. American Antiquity 69: 17-40. Høeg, C. 1925: Les Saracatsans. Une Tribu nomade Grècque. Paris–Copenhague. Hourmouziadis, G. 1973: Burial customs. In Theocharis, D.T. (ed.) Neolithic Greece. Athens: National Bank of Greece, 201–212. Jacobsen, T.W. 1969: Excavations at Porto Cheli and vicinity, preliminary report, II: the Franchti Cave. Hesperia 38: 343–381. — 1976: 17,000 years of Greek prehistory. In Lamberg-Karlovsky, C.C. (ed.) Hunters, Farmers and Civilizations: Old World Archaeology. Readings from Scientific American. San Francisco: W.H. Freeman, 133–151. Johnston, F.E. 1962: Growth of the long bones of infants and young children at Indian Knoll. Human Biology 23: 66–81. Jovanović, B. 1969: Chronological frames of the Iron Gate group of the Early Neolithic period. Archaeologica Iugoslavica 10: 23–38. Joyce, R.A. & Gillespie, S.D. (eds) 2000: Beyond Kinship. Social and Material Reproduction in House Societies. Philadelphia: University of Pennsylvania Press. Kavadias, G.B. 1965: Pasteurs-nomades Méditerranéens, Les Saracatsans de Grèce. Paris: Gauthier–Villars. Lane, P.J. 1994: The temporal structuring of settlement space among the Dogon of Mali: an ethnoarchaeological case study. In Parker Pearson, M. & Richards, C. (eds) Architecture and Order. Approaches to Social Space. London: Routledge, 196–216. Lévi-Strauss, C. 1983: The Way of the Masks. London: Jonathan Cape.

A

167

The Iron Gates in Prehistory — 1987: Anthropology and Myth: Lectures 1951–1982. Oxford: Blackwell. Lipuma, E. 1998: Modernity and forms of personhood in Melanesia. In Carsten, J. & Hugh-Jones, S. (eds) About the House, LéviStrauss and Beyond. Cambridge: Cambridge University Press, 53–79. Makkay, J. 1983: Foundation sacrifices in Neolithic buildings of the Carpathian Basin. Valcamonica Symposium 3 (1979): 157–167. Malhorta, K.C. 1990: Changing patterns of disease in India with special reference to childhood mortality. In Swedlund, A. & Armelagos, G. (eds) Disease in Populations in Transition. Anthropological and Epidemiological Persectives. London: Bergin & Garvey, 313–332. Marković-Marjanović, J. 1969: Geologija i stratigrafija. In Srejović, D. Lepenski Vir – Nova praistorijska kultura u Podunavlju. Belgrade: Srpska književna zadruga, 185–192. — 1978: Geologija i stratigrafija. In Srejović, D. & Letica, Z. (eds) Vlasac: A Mesolithic Settlement in the Iron Gates, vol. 2 Belgrade: Serbian Academy of Science and Arts, 11–27. Mays, S. 1997: ‘New directions in the scientific study of infant skeletons from archaeological sites’. Paper presented to the TAG Conference, University of Bornemouth 1997. — 1998: The Archaeology of Human Bones. London: Routledge. Mellaart, J. 1967: Çatal Hüyük: a Neolithic Town in Anatolia. London: Thames & Hudson. Meskell, L. 1994: Dying young: the experience of death at Deir el Medina. Archaeological Review from Cambridge 13: 35–45. — 1996: The somatization of archaeology: institutions, discourses, corporeality. Norwegian Archaeological Review 29: 1–16. Minichreiter, K. 1999: Ranoneolitički ukopi i pogrebni običaji u naseljima starčevačkog kulturnog kompleksa. Prilozi instituta za arheologiju u Zagrebu 15/16: 5–17. Mladenović, V. 1999: Threads of life: red fringes in Macedonian dress. In Welters, L. (ed.) Folk Dress in Europe and Anatolia: Beliefs about Protection and Fertility. Oxford: Berg, 97–110. Molleson, T. & Andrews, P. 1997: The human remains. Çatalhöyük 1997 Archive Report. http://www.catalhoyuk.com/archive_reports/ Molleson, T., Andrews, P., Boz, B., Sofaer Derevenski, J. & Pearson, J. 1998: Human remains up to 1998. Çatalhöyük 1998 Archive Report. http://www.catalhoyuk.com/archive_reports/ Molleson, T., Andrews, P. & Boz, B. 1999: Report on human remains recovered from the South Area, together with a summary of material from the BACH area and the KOPAL trench. Çatalhöyük 1999 Archive Report. http://www.catalhoyuk.com/archive_reports/ Molleson, T., Andrews, P., Boz, B. & Hager, L. 2000: Human remains. Çatalhöyük 2000 Archive Report. http://www.catalhoyuk.com/ archive_reports/ Moore, J. & Scott, E., eds. 1997: Invisible People and Processes. Writing Gender and Childhood into European archaeology. London: Leicester University Press. Nemeskéri, J. 1974: Outline on the anthropological finds of a Neolithic site. Wissenschaftliche Mitteilungen des Bosnisch–Herzegowinischen Landesmuseums, Archäologie IVA: 37–46. Nemeskéri, J. & Lengyel, L. 1976: Neolithic skeletal finds. In Gimbutas, M. (ed.) Neolithic Macedonia as Reflected by Excavation at Anza, Southeast Yugoslavia. Los Angeles: Institute of Archaeology, University of California, 375–410. Ney, P. 1971: Mineralogische Analyse einer Fussbodenprobe von Lepenski Vir I. In Schwabedissen, H. (ed.) Die Anfänge des Neolithikums vom Orient bis Nordeuropa, Teil 2: Östliches Mitteleuropa. Köln: Böhlau, 33–34. Politis, G.G. 1999a. Plant exploitation among the Nukak huntergatherers of Amazonia: between ecology and ideology. In Gosden, C. & Hather, J. (eds) The Prehistory of Food. Appetites

for Change. London: Routledge, 99–125. — 1999b. La actividad infantil en la producción del registro arqueológico de Cazadores-Recolectores. Revista do Museu de Arqueologia e Etnologia Suplemento 3: 263–283. Quitta, H. 1975: Die Radiocarbondaten und ihre historische Interpretation. In Srejović, D. Lepenski Vir. Eine vorgeschichtliche Geburtsstätte europäischer Kultur. Bergisch Gladbach: Gustav Lübbe, 272–285. Radovanović, I. 1996: The Iron Gates Mesolithic. Ann Arbor: International Monographs in Prehistory. — 1997: The Lepenski Vir Culture: a contribution to interpretation of its ideological aspects. In Antidoron Dragoslavo Srejović completis LXV annis ab amicis collegis discipulis oblatum. Belgrade: Centre for Archaeological Research, University of Belgrade, 85–93. — 1999: ‘Neither person nor beast’ — dogs in the burial practice of the Iron Gates Mesolithic. Documenta Praehistorica 26: 71–87. — 2000: Houses and burials at Lepenski Vir. European Journal of Archaeology 3: 330–349. Rivière, P. 1995: Houses, places and people: community and continuity in Guiana. In Carsten, J. & Hugh-Jones, S. (eds) About the House, Lévi-Strauss and Beyond. Cambridge: Cambridge University Press, 189–205. Roberts, C. & Manchester, K. 1995: The Archaeology of Disease. Ithaca (N.Y.): Cornell University Press. Rodden, R.J. 1965: An Early Neolithic village in Greece. In Avenues to Antiquity. Readings from Scientific American. San Francisco: W.H. Freeman, 151–159. Roksandic, M. 1999: Transition from Mesolithic to Neolithic in the Iron Gates Gorge: Physical Anthropology Perspective. Unpublished Ph.D. thesis, Department of Anthropology, Simon Fraser University, Burnaby, B.C. Rosaldo, R. 1989 [1993]: Culture and Truth. The Remaking of Social Analysis. Boston: Beacon Press. Saunders, S.R. 1992: Subadult skeletons and growth related studies. In Saunders, S.R. & Katzenberg, M.A. (eds) Skeletal Biology of Past Peoples: Research Methods. New York: Wiley-Liss, 1–20. Scheuer, J.L., Musgrave, J. & Evans, S.P. 1980: Estimation of late foetal and perinatal age from limb bone lengths by linear and logarithmic regression. Annals of Human Biology 7: 257–265. Scott, E. 1999: The Archaeology of Infancy and Infant Death (BAR International Series 819). Oxford: BAR Publishing. Sofaer Derevenski, J. 1997: Engendering children, engendering archaeology. In Moore, J. & Scott, E. (eds) Invisible People and Processes. Writing Gender and Childhood into European Archaeology. London: Leicester University Press, 192–202. Srejović, D. 1967: Lepenski Vir. Katalog izložbe. Belgrade: Narodni muzej. — 1969a: Lepenski Vir: Nova praistorijska kultura u Podunavlju. Belgrade: Srpska književna zadruga. — 1969b: The roots of the Lepenski Vir culture. Archaeologia Iugoslavica 10: 13–21. — 1970: Lepenski Vir, Boljetin — predneolitska i neolitska naselja i nekropole. Arheološki pregled 12: 72–73. — 1972: Europe’s First Monumental Sculpture: New Discoveries at Lepenski Vir. London: Thames & Hudson. — 1981: Lepenski Vir: Menschenbilder einer frühen europäischen Kultur. Mainz: von Zabern. Srejović, D. & Babović, Lj. 1983: Umetnost Lepenskog Vira. Belgrade: Jugoslavija. Srejović, D. & Letica Z. 1978: Vlasac: A Mesolithic Settlement in the Iron Gates. Vol. 1, Archaeology. Belgrade: Serbian Academy of Sciences and Arts. Stanković, S. 1992: Kultna mesta centralnobalkanskog područja. Unpublished Ph.D. dissertation. Department of Archaeology, University of Belgrade.

A

168

Sofija Stefanović & Dušan Borić: New-born infant burials at Lepenski Vir Stanković, S. & Leković, V. 1993: Neolithic settlement at Blagotin. Glasnik Srpskog arheološkog društva 9: 177–179. Stewart, T.D. 1968: Identification by the skeletal structures. In Camps, F.E. (ed.) Gradwohl’s Legal Medicine, 2nd Edition. Bristol: Wright, 123–154. Strathern, M. 1988: The Gender of the Gift. Problems with Women and Problems with Society in Melanesia. Berkeley: University of California Press. Stuart-MacAdam, P.L. 1989: Nutritional deficiency diseases: a survey of scurvy, rickets and iron-deficiency anaemia. In Işcan, M.Y. & Kennedy, K.A.R. (eds) Reconstruction of Life from the Skeleton. New York: Alan Liss, 201–221. Tanner, J.M. 1989: Foetus into Man (2nd edition). Ware: Castlemead. Tringham, R. 2000: The continuous building: a view from the deep past. In Joyce, R.A. & Gillespie, S.D. (eds) Beyond Kinship. Social and Material Reproduction in House Societies. Philadelphia: University of Pennsylvania Press, 115–134. Turner, V. 1967: The Forest of Symbols. Aspects of Ndembu Ritual. Ithaca: Cornell University Press. — 1974: Dramas, Fields and Metaphors. Symbolic Action in Human Societies. Ithaca: Cornell University Press. Ubelaker, D.H. 1978: Human Skeletal Remains: Excavation, Analysis, Interpretation. Chicago: Aldine. Van Gennep, A. 1960: The Rites of Passage. Chicago: University of Chicago Press. Waterson, R. 1990: The Living House: an Anthropology of

Architecture in South-East Asia. Kuala Lumpur: Oxford University Press. Waterson, R. 2000: House, place and memory in Tana Toraja (Indonesia). In Joyce, R.A. & Gillespie, S.D. (eds) Beyond Kinship. Social and Material Reproduction in House Societies. Philadelphia: University of Pennsylvania Press, 177–188. Welters, L. (ed.) 1999a: Folk Dress in Europe and Anatolia: Beliefs about Protection and Fertility. Oxford: Berg. — 1999b: Introduction: folk dress, supernatural beliefs, and the body. In Welters, L. (ed.) Folk Dress in Europe and Anatolia: Beliefs about Protection and Fertility. Oxford: Berg, 1–12. — 1999c: The Peloponnesian ‘Zonari’: a twentieth-century string skirt. In Welters, L. (ed.) Folk Dress in Europe and Anatolia: Beliefs about Protection and Fertility. Oxford: Berg, 53–70. Whittle, A. 1996: Europe in the Neolithic. The Creation of New Worlds. Cambridge: Cambridge University Press. Whittle, A., Bartosiewicz, L., Borić, D., Pettitt, P. & Richards, M. 2002: In the beginning: new radiocarbon dates for the Early Neolithic in northern Serbia and south-east Hungary. Antaeus 25: 63–117. Winnifrith, T.J. 1987: The Vlachs: The History of a Balkan People. London: Duckworth. Zoffmann, ZS. 1983: Prehistorical skeletal remains from Lepenski Vir. Homo 34(3/4): 129–148. Zoffmann, ZS. 1988: Human skeletal remains from Divostin. In McPherron, A. & Srejović, D. (eds) Divostin and the Neolithic of Central Serbia. Pittsburgh: University of Pittsburgh, 447–455.

169

DNA-based sex identification of the infant remains from Lepenski Vir Biljana Čuljković, Sofija Stefanović & Stanka Romac

Abstract: At Lepenski Vir 41 infant skeleton were excavated from under the houses. Infant mortality could be affected by a broad range of economic, social and health conditions, and could be regarded as an indicator of overall quality of life in past communities. Sex determination presents a difficult problem in anthropology in cases of fragmentary or child remains. We performed DNA-based sex identification of 30 infant remains. Of the 30 samples tested, 29 specimens provided results: 16 were found to be males and 13 females. Key words: DNA, sex determination, infant skeletons, Lepenski Vir

Introduction Human remains from Lepenski Vir consist of 190 individuals from 134 graves plus 42 individuals from unidentified contexts (Roksandic 1999: 78). Among them there are 51 infant skeletons, most of them found underneath the house floors when these were lifted in 1970. Infants are defined as individuals under 1 year of age and the study of their skeletons in palaeodemography is important because infant mortality could be affected by a broad range of economic, social and medical conditions, and could be regarded as an indicator of overall quality of life of past communities. Although this material represents the most demographically variable and sensitive portion of the human life cycle (Roth 1992: 177), there are many problems associated with the assessment of infant mortality from skeletal material. The reasons why these skeletons might not be represented in osteoarchaeological material are: the effect of soil acidity (as soil acidity increases so the preservation of bone decreases, especially infant bones), burial practices (infants may be excluded from cemetery burial), and excavation strategy. In this context we can state that the number of infants at Lepenski Vir is high (27% of all buried individuals). We estimated the gestational age of infants from Lepenski Vir, according to maximum length of femora and humeri. Most of them belong to 38–40 gestational weeks. The anthropological criteria for sex determination are based on qualitative assessment of the morphological features of the skull, long bones and pelvis. The problem of sex determination persists when dealing with fragmentary and/or infant burials. The reliability of morphometric analyses for gender identification in infants is low, although pelvic sex differences are present at birth, they become obscured soon after birth. New developments in molecular biology, and especially in analysing DNA recovered from ancient bones, have provided reliable methods for gender determination based on amplification of DNA sequences specific to the X and/or Y chromosomes (Mannucci et al. 1994; Lassen et al. 1996). The development of the polymerase chain reaction (PCR) (Saiki et al. 1985) has allowed extremely small amounts of highly degraded DNA to be analysed.

In this study we performed DNA-based sex identification of 30 infant remains in order to obtain information on the sex distribution of infants at Lepenski Vir.

Materials and methods The extractions were performed from 0.2–0.3 g powdered bone (ribs). The bone was ground either in a coffee grinder or in a mortar. In order to avoid contamination with contemporary DNA, all instruments and working areas were exposed to UV light. Powdered bone was mixed with lysing buffer in the ratio: 1 g of powdered bone and 2 ml of lysing buffer (Hughes & Galau 1988). The sample was then incubated at 550 °C overnight followed by phenol/chloroform extraction and isopropanol precipitation. The same experimenter conducted all experiments. All buffers and water were autoclaved. A large amount of PCR premixes containing all the necessary components, except DNA and Taq polymerase, were prepared and frozen in small sealed aliquots at –200 °C. A ‘quality control’ PCR without added DNA but only Taq polymerase was performed on random aliquots. The preparation of bone samples and the DNA extractions were performed in one room and stored in a freezer. On arrival in the laboratory, the specimens were cleaned and UV irradiated (254 nm) for 60 minutes. Specimens were subsequently handled under sterile conditions. Small samples of bone were extensively cleaned by cutting off approximately 20 mm of the entire bone surface with scalpel blades. Bones ware then placed in disposable sterile tubes further used for DNA extraction. Preparations of buffers and PCR set-up were performed in a dedicated sterile hood under constant UV illumination (254 nm). Dedicated pipettes with aerosol resistant plugged tips were used throughout. PCR experiments as well as the analysis of PCR products were performed in a separate room. For each set of PCR experiments reaction blanks and mock extraction controls were conducted. The PCR was performed in a total volume of 25μl reaction mixture containing 10mM Tris-HCl pH 8.3, 50mM KCl, 1.5mM MgCl 2 , 1% Triton X-100, 0.2mM dNTPs each, 0.2mM each primer and 1.25U Taq DNA Polymerase

171

The Iron Gates in Prehistory

A

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

150bp 100bp

B

1

2

3

C

1

2

3

4

5

6

7

8

9 10

11

12

13

14

15

16

17

9

10

11

12

13

14

15

16

150bp 100bp

4

5

6

7

8

17

150bp 100bp

Figure 1. Acrylamide gel electrophoresis of PCR products of amelogenin loci for sex determination. A: lanes 1 and 2: sample 63, lanes 3 and 4: sample 71, lanes 5 and 6: sample 95, lane 7: blank reaction control, lanes 8, 9 and 10: sample 110, lane 11: 50bp DNA marker (Pharmacia LKB), lanes 12, 13 and 14: sample 111, lane 15: mock extraction control, lane 16: female’s contemporary DNA, lane 17: male’s contemporary DNA; B: lane 1: 50bp DNA marker, lanes 2 and 3: sample 102, lane 4: sample 103, lane 5: sample 109A, lane 6: sample 107, lane 7: mock extraction control, lane 8: female’s contemporary DNA, lane 9: blank reaction control, lane 10: 50bp DNA marker, lane 11: sample 113, lanes 12, 13 and 14: sample 112, lanes 15 and 16: sample 110, lanes 17 and 18: sample 111, lane 19: mock extraction control. C: lane 1: 50bp DNA marker, lanes 2 and 3: sample 113, lane 4: mock extraction control, lanes 5 and 6: sample 118, lanes 7 and 8: sample 119, lane 9: 50bp DNA marker, lane 10: sample 114, lane 11: sample 123, lanes 12, 13 and 14: sample 120, lanes 15 and 16: sample 133, lane 17: blank reaction control.

(Boehringer). Samples were amplified through 40 cycles in a Techne Progene Thermal Cycler. The PCR temperature profiles and sequences of primers for the X–Y homologous amelogenin genes (AMELX and AMELY) and CD4 gene were previously described (Mannucci et al. 1994; Edwards et al. 1991).1 PCR products were visualized on 12% polyacrylamide gels by silver staining.

Results and discussion For each DNA extract we performed at least three PCR reactions for the amelogenin gene. Out of 30 samples tested, 29 specimens provided results. Sixteen specimens were found to be males and 13 females. The results for 18 specimens are shown in Figure 1. The success rates of PCRs are given in Table 1. In total, data were obtained for 68 of 118 PCRs of 30 specimens. There were no inconsistencies or conflicting data for any of the specimens. The same type of analysis was successfully performed on aDNA samples from adult skeletons of known sex. DNA extracted from these samples was also analysed by agarose gel electrophoresis (data not shown), and all of them

showing smear pattern, indicating DNA degradation. A large portion of the isolated aDNA samples came from bacteria, so quantification does not guarantee success for PCR amplification, which probably is the reason why we observed low success rates of PCR amplification for many samples. The authenticity of DNA samples was verified by amplification of CD4 loci (commonly used for human identification — Edwards et al. 1991) and comparing these DNA profiles to the profile of the analyst who conducted all the experiments. Results of CD4 profiling for eight specimens from Lepenski Vir and the experimentator’s DNA are shown in Table 2. These experiments prove that the amplified fragments when aDNA is used as a template are not the result of contamination with present day DNA. If the amplified fragments were the result of contamination with contemporary DNA, we would obtain similar or identical results for the CD4 gene. Although isolation of DNA was performed under stringent conditions, which included scraping the surface of bone samples (see materials and methods), and DNA profiling showed differences among analysed samples and experimentator DNA, there is still a chance of contamination with contemporary DNA coming from excavators and all other people who handled the samples prior to aDNA analysis.

172

Biljana Čuljković, Sofija Stefanović & Stanka Romac: DNA-based sex identification at Lepenski Vir Table 1. Data on DNA-based sex identification of the infants from Lepenski Vir. Specimen

Successful PCRs*

Results

63 71 94 95 96 98 98A 102 103 106 107 109 109A 110 111 112 113 114 116 117 118 119 120 123 124 125 127 128 132 133

2 3** 0 2 2** 1 2 2** 2 2** 4** 2** 2 3 3** 3** 4** 1 2 4** 2 2 3** 2 2 2** 2 2 1** 3**

Female Male – Female Male Female Male Female Male Male Male Male Female Female Female Male Female Male Male Female Female Male Male Female Male Male Female Female Male Male

*3 PCRs were performed

**5 PCRs were performed

Table 2. Results of DNA profiles for the CD4 gene. Sample

CD4 profile

71 95 109 114 118 120 124 133 Experimentator DNA

5/4 9/9 6/6 9/9 10/10 5/10 9/5 7/7 4/4

Reasons for infant mortality can be various, from congenital diseases to infanticide. Whatever the reason for the deaths of infants at Lepenski Vir, we did not find significant differences in the sex distribution of those tested. Note 1. Twenty-two out of our 23 pairs of chromosomes are called autosomes and do not differ at the gross level between the sexes. The 23rd pair — the sex chromosomes (XX in females and XY in males) — determine the sex of an individual. Overall, the X A

173

chromosome is much larger and has many more genes than the Y chromosome, but one region, the pseudoautosomal region is shared between these two chromosomes. One gene located there, amelogenin (AMEL), encodes a tooth enamel protein. An ancient 6 base deletion mutation in the X chromosome version of the gene, AMELX, compared to the Y chromosome version, AMELY, has been exploited by researchers as a means to identify the chromosome complement, and hence the gender of any DNA sample. The polymerase chain reaction (PCR) is a technique that amplifies large amounts of a selected sequence even if it is initially present at vanishingly small quantities (such as ancient

The Iron Gates in Prehistory DNA samples). Sex determination works by using a PCR assay over the AMELX/AMELY region to amplify the DNA sample and then examining the length of the PCR products (short — deletion carrying — products indicate female DNA (XX), whereas a mixture of long and short products indicate male DNA (XY)). The advantage of this system is that rather than using separate X-specific and Y-specific tests and combining the results to determine the sex, the choice of amelogenin gene PCR necessitates just a single amplification process — more efficient and technically robust. [Ed.]

References Edwards, M.C., Clemens, P.R., Tristan, M., Pizzuti, A. & Gibbs, R.A. 1991: Pentanucleotide repeat length polymorphism at the human CD4 locus. Nucleic Acids Research 19: 4791. Hughes, D. & Galau, G. 1988: Preparation of RNA from cotton leaves and pollen. Plant Molecular Biology Reporter 6: 253–257. Lassen, C., Hummel, S., Herrmann, B. 1996: PCR based sex identification of ancient human bones by amplification of X- and Y-

chromosomal sequences: a comparison. Ancient Biomolecules 1: 25–34. Mannucci, A., Sullivan, K., Ivanov, P.L. & Gill, P. 1994: Forensic application of a rapid and quantitative DNA sex test by amplification of the X/Y homologous gene amelogenin. International Journal of Legal Medicine 106: 190–193. Roth, E. 1992: Applications of demographic models to paleodemography, In Saunders, S.R. & Katzenberg, M.A. (eds) Skeletal Biology of Past Peoples: Research Methods. New York: WileyLiss, 175–188. Roksandic, M. 1999: Transition from Mesolithic to Neolithic in the Iron Gates Gorge: Physical Anthropology Perspective. Unpublished PhD thesis, Simon Fraser University, Burnaby, B.C. Saiki, R.K., Scharf, S., Faloona, F., Mullis, K.B., Horn, G.T., Erlich H.A. & Arnheim, N. 1985: Enzymatic amplification of β-globin genomic sequences and restriction site analyses for diagnosis of sickle cell anemia. Science 230: 1350–1354. Saunders, S.R. 1992: Subadult skeletons and growth related studies. In Saunders, S.R. & Katzenberg, M.A. (eds) Skeletal Biology of Past Peoples: Research Methods. New York: Wiley-Liss, 1–20.

174

Dating burial practices and architecture at Lepenski Vir Clive Bonsall, Ivana Radovanović, Mirjana Roksandic, Gordon Cook, Thomas Higham & Catriona Pickard

Abstract: Previous attempts to establish a chronology for Lepenski Vir using three different methods (stratigraphy, radiometric 14C dating of bulk charcoal samples, and AMS 14C dating of human bone collagen) produced inconsistent results. Discrepancies between the human bone and charcoal ages were found to result from a reservoir effect in the bones of people who ate significant quantities of Danube fish. When a reservoir ‘correction’ is applied, the human bone 14C dates are consistent with the charcoal dates, and this raises questions about the excavator’s relative and absolute chronology based on stratigraphy and inter-site comparisons. Single-entity dating of surviving archaeological materials offers the best hope of constructing a reliable chronological framework for Lepenski Vir. This paper presents the results of a further programme of AMS 14C dating of human remains. Direct dating of 24 burials confirms that different burial practices characterized the Final Mesolithic and Early Neolithic. Previous attempts to assign burials to Mesolithic or Neolithic phases, based on stratigraphic observations, are shown to be broadly correct but not always accurate in detail. The evidence from radiocarbon dating and stratigraphy is used to calculate ‘minimum’ and/or ‘maximum’ ages for certain of the trapezoidal buildings, which suggest that this architectural form was in use during the Final Mesolithic and Early Neolithic. The implications of the human bone 14C dates and associated stable isotope measurements for the timing of the Mesolithic–Neolithic transition in the Iron Gates are also discussed. Key words: burials, architecture, Mesolithic, Neolithic, AMS 14C dating, stable isotopes

Introduction Lepenski Vir stands out among the Stone Age sites of the Iron Gates (Fig. 1) by virtue of its architecture characterized by distinctive trapezoidal structures with lime plaster floors containing numerous large stone sculptures and other symbolic artefacts (among other finds), together with a record of complex burial practices associated with the structures and the areas between them. Its unusual features have led many archaeologists, including the excavator Dragoslav Srejović to view Lepenski Vir as not just another settlement on the edge of the Danube, but as a site that came into being early in the Mesolithic and later assumed an important role in the lives of the local population as an aggregation, sacred and/or ceremonial site. The conventional interpretation of the occupation sequence at Lepenski Vir was proposed by Srejović (1969, 1972a), who recognized a succession of Mesolithic settlements: Proto-Lepenski Vir, Lepenski Vir I, and Lepenski Vir II, followed by the Neolithic settlements of Lepenski Vir IIIa and IIIb. The main Mesolithic occupation phase of Lepenski Vir I was divided into 5 sub-phases (Ia–e) based on stratigraphic observations and material culture content, while the Early Neolithic Lepenski Vir IIIa phase was divided into two sub-phases (IIIa1–2). Srejović (1969, 1972a) argued for a significant time gap between Lepenski Vir II and III. This periodization of Lepenski Vir became controversial even before the excavations were completed in 1970, due primarily (though not exclusively) to the discrepancies between Srejović’s stratigraphic observations and the first published radiometric 14C dates on charcoal samples collected from the Lepenski Vir I structures (Quitta 1969). The 14C results suggested that some of the earliest Lepenski Vir I

structures ‘stratigraphically’ were among the latest chronologically, according to their radiocarbon ages. Radiometric 14C dating of charcoals from the Lepenski Vir I–II structures gave 14C ages between c. 7360 and 6560 BP (Quitta 1972, 1975). These ages were similar to those for Early Neolithic (Starčevo–Körös–Criş sites in the surrounding regions and were rejected by Srejović (1971, 1972a, 1972b) as being too young. He suggested an age range for Lepenski Vir I–II of c. 500 years earlier, based on radiometric 14 C ages of c. 7930–7440 BP for charcoal samples from Mesolithic contexts at the neighbouring site of Vlasac. Other researchers have accepted the radiocarbon measurements for Lepenski Vir as valid and have interpreted phases I–II as either Early Neolithic (e.g. Jovanović 1969, 1972) or a Late Mesolithic survival in an area unsuitable for farming but rich in natural resources capable of supporting a hunter-gatherer population (e.g. Voytek and Tringham 1989). Many of the burials at Lepenski Vir were also assigned to the five main phases (Zoffmann 1983), but the basis on which this was done was never adequately explained in the literature. It is assumed to be partly on the basis of stratigraphy, including the stratigraphic relationship with buildings, and partly on the basis of burial type. The human remains from Lepenski Vir were re-analyzed by Roksandic (1999, 2000; Roksandic et al. 2006, this volume), but a reconsideration of the phasing of the burials was beyond the scope of her study. Subsequent AMS dating of human remains assigned to phases IIIa and IIIb, produced ages between c. 7770 and 6910 BP (Bonsall et al. 1997), which are earlier than expected on the basis of Srejović’s stratigraphic interpretation and older than the charcoal ages for preceding phases I and II. Bonsall et al. (1997) suggested, and subsequent research (Cook et al. 2001) has demonstrated, that the radiocarbon ages on the

175

The Iron Gates in Prehistory

Figure 1. Lepenski Vir and the Iron Gates.

bones of humans whose diets included regular consumption of aquatic food sources (mainly freshwater fish) are subject to a reservoir effect of up to approximately 500 years. Cook et al. (2001, 2002) devised a method for correcting the 14C ages made on human bone samples for this reservoir effect using: (1) the δ15N values for human bone collagen samples to estimate the percentage freshwater aquatic diet, based on a knowledge of the approximate δ15N end members for 100% freshwater aquatic and terrestrial diets, assuming a linear relationship between δ15N and percentage freshwater diet and (2) the age offsets between human bone ages and closely associated ungulate bone ages for particular δ15N values, based on excavated material from another Iron Gates site, Schela Cladovei, on the Romanian bank of the Danube. When this reservoir correction is applied to the human bone ages from Lepenski Vir obtained by Bonsall et al. (1997) the resulting age range (7310 to 6720 BP) is very similar to that of the bulk charcoals (7360 to 6560 BP). This age range and the fact that some of the Lepenski Vir III burials had ‘Mesolithic’ dietary signals raises questions about the phasing of the burials. It also raises doubts about the stratigraphic integrity of Lepenski Vir III, because the ages are no different from phases I and II. In parallel with the AMS 14C dating of human remains, Radovanović (1996a) attempted a revision of the Lepenski Vir stratigraphy. Using information in Srejović’s published accounts together with previously unpublished field documentation, she proposed a revised sequence of occupation phases: Proto-Lepenski Vir, Lepenski Vir I (comprising sub-

phases 1–3, with 11 building levels), Lepenski Vir II, and Lepenski Vir III. Subsequently, several authors have questioned the stratigraphic integrity or even the existence of some of Srejović’s phases (e.g. Borić 1999, 2002b; Radovanović 2000; Garašanin & Radovanović 2001; Bonsall et al. 2002a; Perić & Nikolić 2004). The greater part of the Lepenski Vir site was excavated in 1965–70 .With the impounding of the Danube by the Iron Gates I dam the site was flooded and is no longer accessible, apart from a number of structures from Lepenski Vir I that were rescued and preserved for display in 1970. Realistically, therefore, the disagreements over the relative and absolute chronology of Lepenski Vir can now only be addressed by a detailed programme of scientific analysis, including singleentity dating, of the surviving archaeological materials. To the authors’ knowledge, no charcoal or other carbonized plant material is available for analysis. However, large collections of animal and human bones are still available for scientific study, as are low-fired, organic-tempered Starčevo pottery sherds that, potentially, are datable by AMS 14C, archaeomagnetic intensity, and luminescence techniques (Bonsall et al. 2002b).

The dating programme Aims and objectives The main aims of the AMS 14C dating programme reported here were to:

176

Clive Bonsall et al.: Dating burial practices and architecture at Lepenski Vir

1. Investigate changes in diet and subsistence during the occupation of Lepenski Vir; 2. Establish the chronological contexts of particular forms of burial represented at the site; 3. Test the phasing of the Lepenski Vir burials proposed by Srejović (1969) and Zoffmann (1983) based on stratigraphy and burial custom; 4. Test previous hypotheses of the age and phasing of the trapezoidal buildings by dating human burials that were either cut through or sealed by their plaster floors; and 5. From the above, provide new information bearing on the timing of the Mesolithic–Neolithic transition in the Iron Gates gorge. Why date the burials? Two approaches have been used in an effort to establish a more reliable chronology for the trapezoidal structures at Lepenski Vir using single-entity AMS 14C dating of bone remains. One approach has involved the dating of terrestrial animal bones (e.g. Borić & Miracle 2004; Borić & Dimitrijević 2005, 2007), the other the dating of human bones from articulated skeletons (Bonsall et al. 2004). Both approaches rely on the existence of a clear stratigraphic relationship between the dated bone and the building. There are uncertainties associated with both approaches, but we suggest that the uncertainties are far greater with the first approach. The bases of the Lepenski Vir buildings were made by cutting more or less horizontally into the sloping bank of the Danube, effectively creating pit features, and these pits became infilled by one means or another after the buildings were abandoned. A disarticulated bone found on the floor of a pit could have reached that position in any of a number of ways. It may be in a primary context, deposited when the building was in use, or shortly after its abandonment. On the other hand, the bone could be older than the pit (i.e. in a secondary context) deposited there after its abandonment as a result of slumping at the sides of the pit, movement of material from upslope due to gravity or hillwash (probably a common event on steep valley side slopes in the Iron Gates gorge), or deliberate infilling of the pit with soil material containing earlier archaeological objects. It is possible for younger bones to be introduced into ancient pit features as a result of bioturbation, including animal burrowing, earthworm activity and root penetration, or of post-depositional disturbance by humans. Moreover, unless animal bones bear manufacturing traces or butchery marks, there will always be an element of doubt about whether their presence in an archaeological site is the result of primary human activity or natural processes. These are just some of the initial steps in the taphonomic process that can affect bones prior to deposition; further loss of information (e.g. water transport or the destruction of datable collagen) may be expected in the deposit itself during fossil diagenesis. In addition, secondary human effects in the form of errors of excavation or curation (see below) may result in older or younger material being wrongly attributed to an archaeological feature. A possible example of this from the Iron Gates was discussed by Bonsall et al. (2002b). It has been argued that the taphonomic biases associated with the dating of animal bones are lessened by dating bones

that are still in articulation (skeletons or partial skeletons) on the assumption that they are in a primary context and were deposited soon after the death of the animal (e.g. Borić & Dimitrijević 2007: 55). However, these are rare occurrences on most archaeological sites, including Lepenski Vir. This argument applies equally to human skeletons, which were much more frequent than animal skeletons/partial skeletons at Lepenski Vir. While 14C age measurements on human bones from this site are less precise because of the reservoir age associated with them (see above), often the articulated bones (more-or-less complete skeletons, rather than typical food remains from animals) are from clearly-defined, archaeological contexts and are subject to better stratigraphic control. Moreover, in contrast to the limited information available for animal remains, the stratigraphic relations between burials and buildings at Lepenski Vir are reasonably well documented by photographs, plans and written records. Thus, given that we have already developed a reservoir correction for 14C measurements on human bones from the Iron Gates (Cook et al. 2001, 2002, in press) this becomes our preferred material for a detailed dating programme. Curatorial issues Recovery methods and post-excavation practices can have as much taphonomic effect on bones as the diagenetic processes that led to their fossilization, and must be taken into account when assessing the 14C dates and stable isotope results from Lepenski Vir. The following observations are based on the study undertaken by Mirjana Roksandic in 1996–8 for her PhD thesis (Roksandic 2000). In many respects the Lepenski Vir excavation was a remarkable achievement. Between 1965 and 1969 in five campaigns lasting a total of 12 months, approximately 2500m2 of the site, with deposits averaging 3.5 m deep, were excavated to reveal architecture, monumental sculpture and graves of the ‘Lepenski Vir culture’. But all this was done at great speed, for the most part without the use of fine sieving,1 and much detailed information on burial practices and relationships between burials and architecture undoubtedly was overlooked. Judging from photographic evidence and testimonies of archaeologists who worked at Lepenski Vir as students, when burials were identified their excavation and recording was usually quite meticulous. But not all burials were recognized, or were only recognized after they had been disturbed by the excavators. In some cases (e.g. burials 29, 30, 49 and 60) only skulls and long bones were collected, even though the burials (according to the available photographs) were intact and the bones well preserved. In other cases, contextual data are imprecise. Some groups of bones were recorded only as coming from a particular ‘cultural layer’ and were not registered as ‘burials’; yet these are sometimes represented by more skeletal elements than some of the identified burials. Furthermore, the material was washed on site by unskilled workers, and it is possible that some smaller bones and bone fragments were lost in the process. However, the frequency of small-sized human carpal bones and phalanges in the collection would seem to exclude this as a major factor affecting bone presence and preservation. Since the excavations, the human bones from Lepenski Vir

177

The Iron Gates in Prehistory

have been housed in the Faculty of Philosophy, University of Belgrade, where they have been studied by generations of researchers. Frequent handling took its toll on the collections, but most damage was a consequence of the economic and social problems that afflicted the former Yugoslavia in the 1980s and 1990s. During this period cardboard boxes and packaging materials were in short supply and storage space was at a premium. The collections could not be housed together in one area. There were not enough boxes to store the skeletons individually, and bones from two or more skeletons were occasionally placed together in the same box. Moreover, shelf space was limited with the result that boxes were piled on top of one another, which caused some to collapse. For a time a large part of the collection was housed in a damp basement, causing boxes and wrappings to disintegrate and mould to develop on some of the bones. Labels were also affected to some extent, but in the majority of cases were nevertheless still legible. In 1997 with the support of the Wenner-Gren Foundation and logistical help from Prof. Ž. Mikić and students at the University of Belgrade, Roksandic was able to bring the collection together in one place, with proper shelving, and to embark on the task of restoring the collection to its original curatorial condition — work that took seven months. By 1998 she had reconstructed as much of the collection as she could, removed the mould from the bones, and recorded bone

presence and preservation for each of the skeletons. A number of burials were found to be missing; some skeletons that appeared more or less intact in excavation photographs had few bones remaining in 1996–8; and discrepancies with the documentation and probable mixing of bones from different skeletons were consistently recorded. This knowledge guided selection of samples for 14C dating and stable isotope analysis in 2000. Even so, not all problems were resolved it seems, as illustrated by burials 4 and 7/I (see below). Methods and results Bone samples were prepared for AMS 14 C dating at the Oxford Radiocarbon Accelerator Unit (ORAU) using routine collagen extraction procedures (Law & Hedges 1989; Bronk Ramsey et al. 2000). An additional ultra-filtration pretreatment step was used to further purify the bone gelatin and retain only the >30-kD molecular weight fraction for 14C assay (Brown et al. 1988; Bronk Ramsey et al. 2000). The 30-kD fraction was lyophilized and analyzed using a Europa Scientific ANCA-MS system consisting of a 20–20 IR mass spectrometer interfaced to a Roboprep CHN sample converter unit operating in continuous flow mode. CO2 from the combustion was trapped cryogenically and graphite was prepared by reduction of CO2 over iron within an excess H2 atmosphere. Graphite targets were then measured by AMS (Bronk Ramsey & Hedges 1997). Small samples of CO2 (100yr) and should, perhaps, be disregarded. The weighted mean of the three dates with errors of 100 yr or less is 7079±57 BP (c. 5950 cal BC).

Acknowledgements We would like to thank the Director of the Institute of Archaeology in Belgrade for permission to reproduce photographs from the Lepenski Vir archive. The senior author would also like to thank Dušan Borić for permission to use his plan of Lepenski Vir for Figure 2.

References Alley, R.B., Mayewski, P.A., Sowers, T., Stuiver, M., Taylor, K.C. & Clark, P.U. 1997: Holocene climate instability: a prominent widespread event 8200 yr ago. Geology 25: 483–486. Biagi, P., Shennan, S. & Spataro, M., 2005: Rapid rivers and slow seas? New data for the radiocarbon chronology of the Balkan Peninsula. In Nikolova, L. & Higgins, J. (eds) Prehistoric Archaeology & Anthropological Theory and Education. Salt Lake City: International Institute of Anthropology, 43–51. Bojadžiev, J. 1995: Chronology of prehistoric cultures in Bulgaria. In Bailey, D.W. & Panajatov, I. (eds) Prehistoric Bulgaria. Madison: Prehistory Press, 149–192. Bökönyi, S. 1969: Kičmenjaci (prethodni izveštaj). In Srejović, D., Lepenski Vir, Nova praistorijska kultura u Podunavlju. Beograd: Srpska književna zadruga, 224–228. — 1972: The vertebrate fauna. In Srejović, D., Europe’s First Monumental Sculpture. New Discoveries at Lepenski Vir. London: Thames and Hudson, 186–189. Bonsall, C. 2003: The Iron Gates Mesolithic. In Bogucki, P. & Crabtree, P. (eds) Ancient Europe 8000 B.C. to A.D. 1000: Encyclopedia of the Barbarian World. New York: Scribner, 175–178. — 2007: When was the Neolithic transition in the Iron Gates? In Spataro, M. & Biagi, P. (eds) A Short Walk through the Balkans: the First Farmers of the Carpathian Basin and Adjacent Regions. Trieste: Società per la Preistoria e Protostoria della Regione Friuli-Venezia Giulia, 53–65. Bonsall, C., Lennon, R., McSweeney, K., Stewart, C., Harkness, D., Boroneanţ, V., Bartosiewicz, L., Payton, R. & Chapman, J. 1997: Mesolithic and Early Neolithic in the Iron Gates: a palaeodietary perspective. Journal of European Archaeology 5: 50–92. Bonsall C., Cook G., Lennon R., Harkness D., Scott M., Bartosiewicz, L. & McSweeney, K. 2000: Stable isotopes, radiocarbon and the Mesolithic–Neolithic transition in the Iron Gates. Documenta Praehistorica 27: 119–132. Bonsall, C., Macklin M.G., Payton, R.W. & Boroneanţ, A. 2002a: Climate, floods and river gods: environmental change and the Meso–Neolithic transition in southeast Europe. Before Farming: the Archaeology of Old World hunter-gatherers 3_4(2): 1–15. Bonsall, C., Cook, G., Manson, J. & Sanderson, D. 2002b: Direct dating of Neolithic pottery: progress and prospects. Documenta Praehistorica 29: 47–59. Bonsall, C., Cook, G.T., Hedges, R.E.M., Higham, T.F.G., Pickard, C. & Radovanović, I. 2004: Radiocarbon and stable isotope evidence of dietary change from the Mesolithic to the Middle Ages in the Iron gates: new results from Lepenski Vir. Radiocarbon 46: 293–300. Bonsall, C., Horvat, M., McSweeney, K., Masson, M., Higham, T.F.G., Pickard, C. & Cook, G.T. 2007: Chronological and dietary aspects of the human burials from Ajdovska Cave, Slovenia. Radiocarbon 49: 727–740. Borić, D. 1999: Places that created time in the Danube Gorges and beyond, c. 9000–5500 BC. Documenta Praehistorica 26: 41–70.

— 2002a: Seasons, Life Cycles and Memory in the Danube Gorges, c. 10000–5500 BC. Unpublished Ph.D. Dissertation, University of Cambridge, Cambridge. — 2002b: The Lepenski Vir conundrum: reinterpretation of the Mesolithic and Neolithic sequences in the Danube Gorges. Antiquity 76: 1026–1039. Borić, D. & Dimitrijević, V. 2005: Continuity of foraging strategies in Mesolithic–Neolithic transformations: dating faunal patterns at Lepenski Vir (Serbia). Atti della Società per la preistoria e protoistoria della regione Friuli-Venezia Giulia, Trieste 15 (2004–2005): 33–107. Borić, D. & Dimitrijević, V. 2007: When did the ‘Neolithic package’ reach Lepenski Vir? Radiometric and faunal evidence. Documenta Praehistorica 34: 53–72. Borić, D. & Dimitrijević, V. In press. Apsolutna hronologija i stratigrafija Lepenskog Vira (Absolute chronology and stratigraphy of Lepenski Vir). Starinar. Borić, D. & Miracle, P. 2004: Mesolithic and Neolithic (dis)continuities in the Danube Gorges: new AMS dates from Padina and Hajdučka Vodenica (Serbia). Oxford Journal of Archaeology 23: 341–371. Borić, D. & Stefanović, S. 2004: Birth and death: infant burials from Vlasac and Lepenski Vir. Antiquity 78: 526–546. Borić, D., Grupe, G., Peters, J. & Mikić, Ž. 2004: Is the Mesolithic–Neolithic subsistence dichotomy real? New stable isotope evidence from the Danube gorges. European Journal of Archaeology 7: 221–248. Boroneanţ, V., Bonsall, C., McSweeney, K., Payton, R. & Macklin, M. 1999: A Mesolithic burial area at Schela Cladovei, Romania. In Thévenin, A. (ed.) L’Europe des Derniers Chasseurs: Épipaléolithique et Mésolithique. Paris: Éditions du Comité des Travaux Historiques et Scientifiques, 385–390. Bronk Ramsey, C. & Hedges, R.E.M. 1997: Hybrid ion sources: radiocarbon measurements from microgram to milligram. Nuclear Instruments and Methods in Physics Research B 123: 539–545. Bronk Ramsey, C., Pettitt, P.B., Hedges, R.E.M., Hodgins, G.W.L. & Owen, D.C. 2000: Radiocarbon dates from the Oxford AMS system: Archaeometry Datelist 30. Archaeometry 42: 459–479. Brown, T.A., Nelson, D.E., Vogel, J.S. & Southon, J.R. 1988: Improved collagen extraction by modified Longin method. Radiocarbon 30: 171–177. Budja, M. 1999: The transition to farming in Mediterranean Europe — an indigenous response. Documenta Praehistorica 26: 119–141. — 2004: The neolithisation of South-eastern Europe. In Knutsson, H. (ed.) Coast to Coast — Arrival: Results and Reflections. Uppsala: Uppsala University, 377–417. Ciugudean, H. 1996: Epoca Timpurie a Bronzului în Centrul şi Sud-vestul Transilvaniei. Bucureşti: Institutul Român de Tracologie. Clark, P.U., Marshall, S.J., Clarke, G.K.C., Hostetler, S.W., Licciardi, J.M. & Teller, J.T. et al. 2001: Freshwater forcing of abrupt climate change during the Last Glaciation. Science 293: 283–287. Coard, R. & Dennell, R.W. 1995: Taphonomy of some articulated skeletal remains: transport potential in an artificial environment. Journal of Archaeological Science 22: 441–448. Cook, G.T., Bonsall, C., Hedges, R.E.M., McSweeney, K., Boroneanţ, V. & Pettitt, P.B. 2001: A freshwater diet-derived 14C reservoir effect at the Stone Age sites in the Iron Gates gorge. Radiocarbon 43: 453–460. Cook, G.T., Bonsall, C., Hedges, R.E.M., McSweeney, K., Boroneanţ, V., Bartosiewicz, L. & Pettitt, P.B. 2002: Problems of dating human bones from the Iron Gates. Antiquity 76: 77–85. Cook, G.T., Bonsall, C., Pickard, C., McSweeney, K., Bartosiewicz, L. & Boroneanţ, A. In press: The Mesolithic–Neolithic transition

1

196

Clive Bonsall et al.: Dating burial practices and architecture at Lepenski Vir in the Iron Gates, southeast Europe: calibration and dietary issues. In Crombé, P. (ed.) Chronology and Evolution in the Mesolithic of Northwest Europe. Coplen, T.B. 1994: Reporting of stable hydrogen, carbon and oxygen isotopic abundances. Pure and Applied Chemistry 66: 273–276. Garašanin, M. & Radovanović, I. 2001: A pot in house 54 at Lepenski Vir I. Antiquity 75: 118–125. Gimbutas, M. 1991: The Civilization of the Goddess. San Francisco: Harper. Görsdorf, J. & Bojadžiev, J. 1996: Zur absoluten Chronologie der bulgarischen Urgeschichte. Eurasia Antiqua 2: 105–173. Grupe, G., Mikić, Ž, Peters, J. & Manhart, H. 2003: Vertebrate food webs and subsistence strategies of Meso- and Neolithic populations of central Europe. In Grupe, G. & Peters, J. (eds) Decyphering Ancient Bones: The Research Potential of Bioarchaeological Collections. Rahden (Westf): Marie Leidorf, 193–214. Gurova, M. 2008: Towards an understanding of Early Neolithic populations: a flint perspective from Bulgaria. Documenta Praehistorica 35: 111–129. Hedges, R.E.M., Clement, J.G., Thomas, C.D.L. & O’Connell, T.C. 2007: Collagen turnover in the adult femoral mid-shaft: modeled from anthropogenic radiocarbon tracer measurements. American Journal of Physical Anthropology 133: 808–816. Jovanović, B. 1969: Chronological frames of the Iron Gate group of the Early Neolithic period. Archaeologica Iugoslavica 10: 23–38. — 1972: The autochthonous and migrational components of the Early Neolithic in the Iron Gates. Balcanica 3: 49–58. Krajcar Bronić, I., Minichreiter, K., Obelić, B. & Horvatinčić, N. 2004: The oldest Early Neolithic (Starčevo culture) settlements in Croatia: Zadubravlje-Dužne and Slavonski Brod-Galovo. In Higham, T., Bronk Ramsey, C. & Owen, C. (eds) Radiocarbon and Archaeology: Fourth International Symposium, St Catherine’s College, Oxford, 9–14 April 2002. Oxford: Oxford University School of Archaeology, 229–245. Law, I.A. & Hedges, R.E.M. 1989: A semi-automated bone pretreatment system and the pretreatment of older and contaminated samples. Radiocarbon 31: 247–253. Magny, M., Bégeota, C., Guiot, J. & Peyrona, O. 2003: Contrasting patterns of hydrological changes in Europe in response to Holocene climate cooling phases. Quaternary Science Reviews 22: 1589–1596. McKinley, J.I. & Roberts, C. 1995: Excavation and Post-excavation Treatment of Cremated and Inhumed Human Remains (IFA Technical Paper 13). Birmingham: Institute of Field Archaeologists. Murray, M. & Schoeninger, M. 1988: Diet, status and complex social structure in Iron Age Central Europe: some contributions of bone chemistry. In Gibson, D.B. & Geselowitz, M.N. (eds) Tribe and Polity in Late Prehistoric Europe: Demography, Production and Exchange in the Evolution of Complex Social Systems. New York: Plenum, 155–176. Nemeskéri, J. 1969: Stanovništvo Lepenskog Vira. In Srejović, D., Lepenski Vir: Nova Praistorijska Kultura u Podunavlju. Belgrade: Srpska Knjizevna Zadruga, 239–257. — 1972: The inhabitants of Lepenski Vir. In Srejović, D., Europe’s First Monumental Sculpture. New Discoveries at Lepenski Vir. London: Thames and Hudson, 190–204. Perić, S. & Nikolić, D. 2004: Stratigraphic, cultural and chronological characteristics of the pottery from Lepenski Vir 1965 excavations. In Vasić, M., Vucković, P. & Cvetković, B. (eds) The Central Pomoravlje in Neolithization of South East Europe. Belgrade: Archaeological Institute in Belgrade, Regional Museum in Jagodina and Regional Museum in Paraćin, 157–217. Quitta, H. 1969: Datiranje radiokarbonskih proba. In Srejović, D.,

Lepenski Vir: Nova Praistorijska Kultura u Podunavlju. Belgrade: Srpska Knjizevna Zadruga, 229–238. — 1972: The dating of radiocarbon samples. In Srejović, D., Europe’s First Monumental Sculpture. New Discoveries at Lepenski Vir. London: Thames and Hudson, 205–210. — 1975: Die Radiocarbondaten und ihre historische Interpretation. In Srejović, D., Lepenski Vir, Eine vorgeschichtliche Geburtsstätte europäischer Kultur. Bergisch Gladbach: Gustav Lübbe, 272–285. — 1978: Radiocarbon dates and three chronological systems. Interdisciplinary Research 1: 12–24. [in Bulgarian] Radovanović, I. 1996a: The Iron Gates Mesolithic. Ann Arbor: International Monographs in Prehistory. — 1996b: Mesolithic/Neolithic contacts: a case of the Iron Gates region. Documenta Praehistorica 23: 39–48. — 2000: Houses and burials at Lepenski Vir. European Journal of Archaeology 3: 331–350. — 2006: Further notes on Mesolithic–Neolithic contacts in the Iron Gates region and the central Balkans. Documenta Praehistorica 33: 107–124. Roksandic, M. 1999: Transition from Mesolithic to Neolithic in the Iron Gates Gorge: Physical Anthropology Perspective. Unpublished PhD thesis, Department of Anthropology, Simon Fraser University, Burnaby, B.C. — 2000: Transition from Mesolithic to Neolithic in the Iron Gates gorge: physical anthropology perspective. Documenta Praehistorica 27: 1–100. Roksandic, M., Djurić, M., Rakočević, Z. & Seguin, K. 2006: Interpersonal violence at Lepenski Vir Mesolithic/Neolithic complex of the Iron Gates Gorge (Serbia–Romania). American Journal of Physical Anthropology 129: 339–348. Roksandic, M., Wood, C. & Vlak, D. 2007: Death in the line of duty: a late Medieval burial at the site of Lepenski Vir (Serbia). International Journal of Osteoarchaeology 17: 635–642. Sladić, M. 1986: Kula près Mihajlovac — un site préhistorique. Đerdapske sveske 3: 432–442. Srejović, D. (ed.) 1969: Lepenski Vir, Nova praistorijska kultura u Podunavlju. Beograd: Srpska Knjizevna Zadruga. — 1971: The roots of the Lepenski Vir culture. Archaeologia Iugoslavica 10: 13–21. — 1972a: Europe's First Monumental Sculpture: Lepenski Vir. London: Thames and Hudson. — 1972b: Kulturen des frühen Postglazials in südlichen Donauraum. Balcanica 3: 11–44. — 1988: The Neolithic of Serbia: a review of research. In Srejović, D. (ed.) The Neolithic of Serbia. Archaeological Research 1948–1988. Belgrade: Centre for Archaeological Research, University of Belgrade, 5–20. Srejović, D. & Babović, Lj. 1983: Umetnost Lepenskog Vira. Beograd: Jugoslavija. Taylor, T. 2002: The Buried Soul: How Humans Invented Death. London: Fourth Estate. Todorova, H. 1989: Das Frühneolithikum Nordostbulgariens im Kontext des ostbalkanischen Neolithikums. In Heger, N. & Hiller, S. (eds) Tell Karanovo und das Balkan Neolithikum. Gesammelte Beiträge zum Internationalen Colloquium in Salzburg, 20.–22. Oktober 1988. Salzburg: Institut für Klassische Archäologie der Universität Salzburg, 9–25. Tringham, R. 1971. Hunters, Fishers and Farmers of Eastern Europe 6000–3000 BC. London: Hutchinson. Vajsov, I. 1998: The typology of the anthropomorphic figurines from northeastern Bulgaria. In Stefanovich, M., Todorova, H. & Hauptmann, H. (eds) In the Steps of James Harvey Gaul, vol. 1: James Harvey Gaul – In Memoriam. Sofia: The James Harvey Gaul Foundation, 107–141. Vasić, R. 1986: Compte-rendu des fouilles du site préhistorique à

1

197

The Iron Gates in Prehistory Velesnica 1981–1982. Đerdapske sveske 3: 264–285. Voytek, B. & Tringham, R. 1989: Rethinking the Mesolithic: the case of South-east Europe. In Bonsall, C. (ed.) The Mesolithic in Europe. Edinburgh: John Donald, 492–500. Whittle, A. 1996: Europe in the Neolithic: the Creation of New Worlds. Cambridge: Cambridge University Press.

Whittle, A., Bartosiewicz, L., Borić, D., Pettitt, P. & Richards, M. 2002: In the beginning: new radiocarbon dates for the Early Neolithic in northern Serbia and south-east Hungary. Antæus 25: 63–117. Zoffmann, Zs. 1983: Prehistorical skeletal remains from Lepenski Vir (Iron Gate, Yugoslavia). Homo 34: 129–148.

198

Grid square A/VIII–IX

Building 34

Grid square b/VI, VII

Grid square b/VI, VII

Building 65/XXXV

Building 65

Near building 40 and Proto-LV hearth. Grid square B-C/11

Building 40

Grid square b/VI, VII

Grid square a/13

Grid square A/II–III, 2nd excavation level.

Grid square A/V, 8th & 9th excavation levels, relative depth 1.57–2.06m

Building 24

Building 24

Grid sq. c–d/1–I

14

26

29

30

54c

54e

60

61

62

69

2

4

8

9

32a

Dorsal decubitus

Crouched

Location

Building 21

Burial

199 On flat rock behind (upslope of) buildings in the central part of the site

0.30 m above floor in front part of building

Above the floor, in rear part of building

In ‘virgin soil’, downstream of the zone with trapezoidal buildings

n/a (medieval)

Burial cuts through floor, to rear of hearth

Next to corner A of building 40, ‘in virgin soil’

On the floor

Above floor of building 65 (field interpretation: cut through burials 54a–b)

n/a (medieval)

n/a (medieval)

Cuts through floor of building 34, to rear of hearth

Upstream of the zone with trapezoidal buildings

Burial cuts through floor, to rear of hearth

Relationship to trapezoidal structure

Descriptions of the burials included in the AMS 14C dating programme.

7/I

Extended

Body position

Appendix 1.

BR: Found with burials 32b–c. Body in crouched position, in burial pit bordered with several stones. [orientation and ‘side’ not recorded]

BR: N–S orientation, crouched on the right side.

BR: S–N orientation, crouched on right side, knees at level of upper body, arms bent, hands in front of face.

BR: In rectangular stone construction, covered by stones. Orientation W–E, head to W, crouched on the right side, well preserved. Around the head – lithics, river shells and animal bones; bone awl in area of legs.

BR: S–N orientation, crouched on left side; damaged – only mandible, maxilla and femur recovered.

BR: Covered with stones, W–E orientation, back almost vertically placed, legs bent, left foot over the right foot, arms along the body over the upper legs, left arm slightly bent.

BR: Extended W–E, lower right arm on pelvis, left arm bent with hand on the right shoulder.

BR: Body extended S–N, right arm along the body, hand on pelvis, lower left arm missing.

BR: Extended SW–NE, right arm bent touches the chin, lower left arm across the pelvis.

BR: Body extended S–N, lower arms across the pelvis (field interpretation: burial 54e disturbed skeleton 54d).

BR: Body extended S–N, right arm bent with hand on left shoulder, left arm along the body, no skull.

BR: Body extended W–E, lower right arm across the torso, left arm bent, hand on the pelvis.

BR: Body extended W–E, left arm along the body, right arm bent across the torso, hand touching the chin.

BR: Body extended S–N, hands on pelvis.

BR: Burial 14 occurred in an oval pit with bones of other individuals (it was originally labeled burial no. II from the ‘ossuary’); it is described as extended S–N, upper arms along the body, lower arms on the body.

BR: Burial no. 7, extended, S–N orientation, arms along the body, hands on pelvis; to the west of the skull of the skeleton was another human skull; to the east of the skull of the skeleton was a bovid skull and antlers. FJ: Oct 24 1967: skull belonging to skeleton no. 7 is labeled ‘Skull I’ ; the separate skull next to it (to the west) is labeled ‘Skull II’ (described as having prominent supra-orbital ridges).

Summary descriptions taken from the Field Burial Record (BR) and Field Journal (FJ)

Clive Bonsall et al.: Dating burial practices and architecture at Lepenski Vir

200

1

Grid square d/III, section d/II

35

Behind (upslope) buildings in the central part of the site

At rear of central part of the site, upslope from trapezoidal structures

On the floor; comprising disarticulated bones adjacent to articulated skeleton 54e (burials 54a–c occurred at higher levels in infilling of ‘house’ pit )

To the rear of the central part of the settlement, upslope of the zone with trapezoidal buildings

Relationship to trapezoidal structure

BR: a fragment of a skull, shoulder bone, and ribs from the right side of the body collected from section d/II in sq. d/III on 18th July 1967. The rest of the body remained within the section – not excavated.

BR: Comprised fragments of a skull and a tibia.

BR: Burial 89 consists of bones from 2 individuals, an adult (89a) and an infant 3–4 years old (89b); adult comprises fragments of skeleton and skull. [field interpretation: extended burial, S–N orientation, arms along the body, damaged, legs missing, Bos primigenius horns around the head]

BR: In burial pit bordered with stones, with burials 79b–c. Burial 79a described as comprising fragments of a skull and the left side of the mandible; 79c also described as parts of a skull and the left side of a mandible.

BR: Field interpretation: extended inhumation with S–N orientation dislocated by the interment of individual 54e; lower extremities preserved.

BR: In destroyed hearth construction with burials 45a and 45c (45a is recorded as a skull, 45b is recorded as a skull found within the hearth but below the level of the floor, and 45c is recorded as consisting of a pelvis, femur, and vertebrae).

BR: In pit. Comprises ulna, radius, part of pelvis, a few vertebrae, femur head

BR: Above pit 2. Burial 31 consisted of left mandible with 5 teeth, 3 separate teeth, and head of a femur.

BR: Burial pit bordered with stones. Skeleton crouched with E–W orientation, left arm bent below the backbone, right arm extended along the body touching a fragmentary bowl. [‘side’ not recorded]

Summary descriptions taken from the Field Burial Record (BR) and Field Journal (FJ)

The bones sampled for dating from burials in the ‘disarticulated’ and ‘uncertain’ categories (cf. Table 1) do not always correspond with the descriptions of the burials in the Field Burial Record or Field Journal. It is not clear if this reflects post-excavation curatorial errors, or errors in the field documentation.

Grid square A/I

Grid square e/4

79a

18

Building 65

54d

Uncertain

Building 61

45b

Grid square f/I–II, next to profile f

Grid square d/I

44

89a

Grid square a/VI

31a

Disarticulated 1

Grid sqare e/I, in a pit bordered with stones

Location

88

Burial

Descriptions of the burials included in the AMS 14C dating programme.

Crouched

Body position

Appendix 1 (cont.).

The Iron Gates in Prehistory

201 5972–5554

≤5554

5972

22

>5931

>6212

>6212

Buildings 29 and 30

Buildings 22, 29, and 30

Building 21, and burial 7/I

Burial 7/I

5931

From burial inserted through plaster floor

21

Bone

Buildings 32 and 33

Building 66

≤5571

Building 23

20

Buildings 31 and 19?

Buildings 9, 7 and 5

Building 31? >5867

>5978

19

18

17 ≤5565

Buildings 14 and 13

15 6820±100

Building 15

Building 13

14

16

Buildings 14 and 15

From floor

5978

13

Charcoal

≤5565

9

Buildings 7, 8 and 17

≤5565

8 Buildings 9 and 7

Buildings 8 and 17 Building 9

≤5565

7 >5978

Buildings 5 and 5a

Building 5

Building 1

Older than

6 >5978

Building 2

Younger than

x

Pottery

x

x

x

x

Building 6

>5982

5982

TPQ (maximum age) cal BC

5a

≤5572

≤5572

TAQ (minimum age) cal BC

?

5978–5565

5982–5572

Calibrated age range BC (2σ)

Buildings 5a, 6 and 17

6845±100

6860±100

age BP

5

From structural beam and floor

From floor

14C

x

Charcoal

Charcoal

Sample type Context

?

x

x

?

x

Borić & Dimitrijević (2007) reported an AMS date of c. 5800 cal BC on a red deer skull interpreted as a grave offering with burial 7/I

The existence of building 5a was disputed by Perić & Nikolić (2007)

Ground Comments edge tools

Chronological and stratigraphic relationships of the trapezoidal buildings at Lepenski Vir, and presence/absence of pottery and ground-edge tools.

4

2

1

Building

Appendix 2.

Clive Bonsall et al.: Dating burial practices and architecture at Lepenski Vir

202

Buildings 34 and 52?

Building 27b, 34 and 52

From burial inserted through plaster floor

>6056

Bone

34

5972–5554

≤5554

6820±100

Charcoal

34

From a beam

≤5571

33

>6233

5972?

>5867

Buildings 27, 27b and 43?

Buildings 20 and 66

Buildings 33, 20 and 66

5867

32 ≤5571

Building 19?

31

>6212

>5931

>6212

>6445?

5983

>5983

TPQ (maximum age) cal BC

Building 30

Charcoal

Charcoal

>5729

5729

TAQ (minimum age) cal BC

29

28

27b

27

27

26′

26

24a

From crouched burials in the infill of ‘house’ pit

Calibrated age range BC (2σ)

Building 24a

Bone

age BP

24

14C

Buildings 18, 31 and 19?

Sample type Context

Burial 26

Building 32

Buildings 18 and 23

Building 30

Buildings 21, 22 and29, and burial 7/I

Buildings 21 and 22, and burial 7/I

Building 27

Building 26

Building 24

Burials 8 and 9

Older than

?

x

x

x

x

x

Pottery

x

x

x

x

x

x

TAQ/TPQ are based on stratigraphic relationship to burial 26.

Conflict between charcoal age and age relative to burial 26. Association between charcoal sample and building is suspect.

TPQ based on charcoal age of building 27 (but context of charcoal uncertain).

TAQ/TPQ based on charcoal age of stratigraphically older building 34.

TAQ/TPQ based on charcoal age (but context of charcoal uncertain).

TAQ/TPQ based on 2-sigma calibrated age range of burial 8

Ground Comments edge tools

Chronological and stratigraphic relationships of the trapezoidal buildings at Lepenski Vir, and presence/absence of pottery and ground-edge tools.

23

Building

Appendix 2 (cont.).

The Iron Gates in Prehistory

6970±60

5983–5735

203

55

From a beam and hearh

7132±64*

6205–5849

≤5849

Charcoal

≤5376

54

5718–5376

≤5735

6620±100

53

From a beam and floor ≤5376

Charcoal

≤5735

52

51

48

47′

6205

>5983

>5718

5718

>5983

Building 52

Buildings 58 and 53

Buildings 47′, 58 & 53, and burial 122

47

Building 34?

Buildings 37, 38 and 41

Building 41

Buildings 38 and 41

Building 36

Younger than

Building 55 5983

?

>5890

>5890

>5890

5890

6374

5867

Buildings 32, 33 and 20

≤5571

>6365

>5983

TPQ (maximum age) cal BC

66

≤5735

TAQ (minimum age) cal BC

Burials 54e and 54c

6592–6006

Calibrated age range BC (2σ)

>6017

7430±160

age BP

65

From a beam

14C

Building 63

Charcoal

Sample type Context

Pottery

Context of charcoal sample (7430±160 BP) is uncertain.

Borić & Dimitrijević (2007: fig. 3) reported a calibrated 14C age of c. 5950 cal BC for burial 19 (crouched, headless) “found at the floor level” of building XLIV/57. The number of buildings at this location, and the stratigraphic relationship(s) between the building(s) and the burial, are unclear.

Ground Comments edge tools

Chronological and stratigraphic relationships of the trapezoidal buildings at Lepenski Vir, and presence/absence of pottery and ground-edge tools.

63′

63

62

58

57

56

Building

Appendix 2 (cont.).

The Iron Gates in Prehistory

The vertebrate fauna from Hajdučka Vodenica in the Danubian Iron Gates: subsistence and taphonomy from the Early Neolithic and Mesolithic Haskel J. Greenfield

Abstract: This paper presents the results of the analysis of the vertebrate fauna from Hajdučka Vodenica, an Early Neolithic site in the Danubian Iron Gates along the border between Serbia and Romania. The site was initially misreported as an Early Iron Age settlement, but recent studies have redated its lower levels to both the terminal Mesolithic and Early Neolithic. Vertebrate faunal remains from the site are divided between these two periods and analyzed from the perspective of subsistence and taphonomy. None of the other assemblages from the gorge belonging to these periods were analyzed from the perspective of taphonomy. Wild fauna dominate the assemblage in both periods, but domestic species (pigs and cattle) are found in both as well. The data point to the borrowing of domestic fauna by later Mesolithic occupants of the region, rather than indigenous domestication. The types and frequencies of both domestic and wild fauna from the lower levels are compared to other Early Neolithic and Mesolithic settlements in the gorge. Key words: Iron Gates, Hajdučka Vodenica, vertebrate fauna, subsistence, taphonomy, Mesolithic, Neolithic

Introduction Most considerations of the transition from the Mesolithic to the Early Neolithic among the indigenous societies of temperate Southeast Europe focus upon a series of sites in the Iron Gates of the Danube, where the river cuts through the mountains on its way to the Black Sea. The Iron Gates is the only area in temperate Southeast Europe where there is abundant archaeological and physical anthropological evidence for the continuous presence of the same population of humans from the end of the Palaeolithic into the Early Neolithic (Živanović 1976a, 1976b; y’Edynak 1978; y’Edynak & Fleisch 1983; Prinz 1987). During the Mesolithic, there is very little evidence for hunter-gatherer occupation in temperate Southeast Europe beyond the confines of the Iron Gates (e.g. Tringham 1973; Srejović 1975; Voytek & Tringham 1989; Radovanović 1996). With the advent of the Neolithic, the meagre evidence for Mesolithic cultures disappears except in the Iron Gates. According to radiocarbon dates, the terminal Mesolithic sites of the Iron Gates (Fig. 1) are contemporary with the earliest phases of the Early Neolithic (Starčevo-Körös-Criş) food producing cultures in the surrounding region (Radovanović 1996; Bonsall et al. 2000). It would appear that the Mesolithic societies of the Iron Gates eventually adopt the material culture of food production into their subsistence systems, while retaining much of their hunting-gathering-fishing Mesolithicstyle subsistence system. As such, they provide the only significant Southeast European example of an indigenous population of Mesolithic hunter-gatherers assimilating to an Early Neolithic lifestyle (Voytek & Tringham 1989; Greenfield 1993). In the restricted confines of the Iron Gates gorge, several sites dated to the later Mesolithic and Early Neolithic were found during the 1960s when large-scale salvage surveys of

the gorge were carried out preceding the construction of a large hydroelectric dam at the eastern end of the Iron Gates. The later Mesolithic occupation of the Iron Gates appears to be contemporary with the Early Neolithic Starčevo-KörösCriş cultures in the surrounding regions (Manson 1990; Radovanović 1996). Faunal remains were collected and analyzed from only a few of them — four from the settlements located on the Serbian side of the gorge (Lepenski Vir I–III, Padina IA and IB, Vlasac, and Hajdučka Vodenica) and at least one on the Romanian side (Icoana II) (Fig. 1). The fauna from all of the sites, with the exception of Hajdučka Vodenica, have been reported on in greater or lesser detail (Bökönyi 1971, 1978; Bolomey 1973; Clason 1980). This paper presents the results of the analysis of the vertebrate fauna from Hajdučka Vodenica. Even though Hajdučka Vodenica was one of the first of the Early Neolithic sites in the gorge to be located and excavated, its fauna only recently became available for analysis and the results are presented here.1 The site was initially misreported as an Early Iron Age settlement, but recent studies have redated its lower levels to the Terminal Mesolithic and the Early Neolithic. In the original faunal analysis, unfortunately, the faunal remains were not temporally subdivided since the stratigraphic information was not available (Greenfield 1984). The faunal remains were thought to belong to the Early Neolithic horizons because they were found in artefact bags that also contained Starčevo-type ceramics (which was considered to be a hallmark of Early Neolithic cultures). Many of the same bags also contained ornamented pebbles like those described from some other sites in the Iron Gates gorge (cf. Jovanovic 1969). Recent reanalysis of the Iron Gates assemblages has demonstrated that the appearance of pottery takes place during the second half of the 7th millennium BC throughout the region in otherwise Mesolithic cultural contexts (Radovanović 1996). As a result, the faunal assemblage is reconsidered in

205

The Iron Gates in Prehistory

Figure 1. Map of sites mentioned in the text: A–Anza; B–Blagotin; BC–Bukovačka Česma; D–Divostin; G–Gomolava; H–Hajdučka Vodenica; I–Icoana; L–Lepenski Vir; O–Obre; P–Padina; S–Starčevo; V–Vlasac.

the light of this new information. In this analysis the data are divided between the Mesolithic and Early Neolithic levels, and analyzed separately. The data from Hajdučka Vodenica form an important comparative database that can increase our understanding of the other Mesolithic and Early Neolithic assemblages from the Iron Gates gorge. The problems in how the data were collected are common to all samples from the Iron Gates. While the problems compromise its utility for traditional reconstructions of subsistence economy, the sample can be used to elucidate issues common to all of the excavations in the gorge, such as assemblage recovery and taphonomy. It can also be used to corroborate the evidence from the other sites, such as distinguishing between Mesolithic and Early Neolithic patterns of faunal exploitation. The types and frequencies of both domestic and wild fauna from the lower levels are compared to those from other settlements in the gorge.

The site Site location and surrounding environment Hajdučka Vodenica (‘brigands water mill’) is located on the right bank of the River Danube as it passes through the Iron

Gates on its way to the Black Sea. The Iron Gates is a 100-km long gorge cut by the Danube through the southern arm of the Carpathian Mountains where it crosses from Romania to Serbia. The gorge is bordered by high and rugged mountains. Steep cliffs often reach over the water’s edge. The mountains bordering the gorge effectively isolate the area from the outside world. Most traffic is forced into a narrow corridor parallel to the river. The Danube narrows in width from 1–2 km at the mouth of the gorge, to only a couple of hundred metres at places inside the gorge. The river becomes extremely rapid and turbulent inside the gorge. Its flow is punctuated by whirlpools that have long been favourite fishing spots. The rugged landscape is interspersed by small and large coves. It is in these coves that several prehistoric settlements were found during the surveys of the 1960s. Hajdučka Vodenica is located in the eastern half of the gorge, on the right (Serbian) bank of the Danube, in a broad cove between the Mali Štrbac hill and the Danube. The prehistoric section of the site is located on the riverbank and the first terrace above the river. The terrace slopes up gradually at first, but it ascends rapidly after a short distance into the mountains bordering the gorge. It extends both east and west of the Roman/Medieval castellum (Jovanović 1966a: 102).

206

Haskel Greenfield: Reanalysis of the vertebrate fauna from Hajdučka Vodenica

The local environmental zones range within a kilometre from riverine to shoreline, to small meadows in the mountains, to high altitude deciduous and evergreen forests. Soils in the region include alluvium along the river’s edge and skeletal podzols at higher altitudes. Most of the prehistoric settlements along the river’s edge are cut back into loess deposits on the river’s terraces (Mišić et al. 1970; Gigov 1972; Cârciumaru 1978). At the present time the area is used for the grazing of domestic stock (primarily sheep and goat, with some cattle). Wild game still abounds throughout the region. Hunting, gathering and fishing continue to be important subsistence supplements in the region. Shellfish collecting was important until the local ecology was devastated by massive flooding of the gorge when the hydroelectric dam was completed. Site description and excavation history Excavations at Hajdučka Vodenica took place between 1966 and 1969 coincident with the excavation of many other sites in the gorge. It is a multi-period site with deposits dating to the Roman, Early Iron Age (EIA), Early Neolithic, and possibly Terminal Mesolithic (Živanović 1976b; Srejović & Letica 1978; Jovanović 1984; Vasić 1983; Prinz 1987; Radovanović 1996). The chronological position and cultural affinities of the prehistoric layers at Hajdučka Vodenica have long been a source of controversy. Originally, all the prehistoric levels were reported to be from an Early Iron Age settlement and cemetery (Jovanović 1966a, 1966b, 1969a, 1969b). It is now recognized that the prehistoric part of the site can be divided between Early Iron Age, Early Neolithic and Mesolithic occupations (Srejović & Letica 1978; Jovanović 1984; Radovanović 1996). The excavations extended between 2.0 and 2.5 metres in depth (Jovanović 1966a: 102). The upper level (nivo 1) included an EIA settlement and necropolis, and associated architectural features and artefacts. The EIA level is stratigraphically superimposed above the Early Neolithic levels, but EIA burials occasionally intrude into them. The EIA horizon is stratigraphically and culturally distinct and separated by a sterile sandy layer from the Early Neolithic and Mesolithic cultural layers. This ultimately allowed the material remains from the lower levels

(Early Neolithic and Mesolithic) to be separated from the upper levels at the site (Iron Age and later). The soil matrix from the cultural levels was also mixed with sand. Recently, the lower horizons at the site have been temporally reassigned. Nivo 2 has been reassigned to the Early Neolithic Starčevo-Criş culture (Hajdučka Vodenica II). The finds from levels (nivo) 3–4 and the level with stone constructions have been reassigned to the Terminal Mesolithic (Jovanović 1984; Radovanović 1996) (see Appendix 3). The Mesolithic occupation has been subdivided into an earlier occupation (Hajdučka Vodenica Ia — end of the 8th and first half of the 7th millennium BC), and a later (Hajdučka Vodenica Ib — first half of the 6th millennium BC) (Radovanović 1996). The latter phase is contemporary with the earliest phases of the Early Neolithic cultures (StarčevoCriş) beyond the Iron Gates (c. 7450–6450 BP — Garašanin 1983; 6100–5100 cal BC — Manson 1990, this volume). The architecture and sculptures found in levels 2–4 include free standing stone sculptures, trapezoidal houses, and stone-lined rectangular hearths often used for inhumations (Jovanović 1966a, 1966b, 1969a, 1969b, 1984). These are typical of Mesolithic culture in the region and are found at Lepenski Vir I–II, Padina IA, and Vlasac I–III (Radovanović 1996). Early Neolithic Starčevo-type ceramics and lithic and ground stone tools are present, however. These are similar to those from Early Neolithic Starčevo culture levels at sites in the region which have been associated with pit houses (at Lepenski Vir III and Vlasac IV) and trapezoidal houses and stone-lined rectangular hearths (at Padina IB — Srejović 1970, 1972; Jovanović 1974; Srejović & Letica 1978). Early Neolithic levels at these sites are usually distinguished by the presence of domestic animals (Bökönyi 1970; Clason 1980: 143). Because of continuity between Late Mesolithic and Early Neolithic sites in the chipped stone tool assemblages and architecture (Voytek & Tringham 1989), the Early Neolithic in the Iron Gates is considered one of the few examples of the indigenous adoption by the Mesolithic inhabitants of Southeast Europe of an Early Neolithic lifestyle (Greenfield 1993). The appearance of pottery has often been used to distinguish Mesolithic from Neolithic occupations. Yet, pottery similar to that of the Early Neolithic Starčevo culture in the surrounding region appears in the first half of

Table 1. Percentage of tools in the assemblage compared with NISP. Period unknown # of Tools

NISP

% of NISP

Early Neolithic # of Tools

NISP

% of NISP

Mesolithic # of Tools

NISP

Grand Total % of NISP

# of Tools

NISP

% of NISP

Taxon 3

Bos taurus Cervus elaphus

7

22

31.82

5

23

21.74

1

Ursus arctos 1

Large-sized mammal 4

1

2

5

40.00

2

8

25.00

43

130

33.08

55

175

31.43

1

1

100.00

1

2

50.00

1

1

100.00

1

30

3.33

100.00

4

25.00

22

Bos/Cervus

1

Unknown

0

9

0.00

2

2

100.00

Grand Total

8

35

22.86

8

34

23.53

207

46

1

2

12

16.67

159

62

228

27.19

28.93

The Iron Gates in Prehistory

the 7th millennium BC (calibrated radiocarbon dates) in the Mesolithic sites of the Iron Gates (Radovanović 1996). Soon afterwards, architecture similar to that used by the Starčevo culture (pit houses) makes its appearance in the gorge (Lepenski Vir III).

Analytical methodology The following analysis will provide a description of the bone finds from Hajdučka Vodenica, and the foundation for some cautiously drawn conclusions based on the sample. All of the bones described here derive from the pre-Early Iron Age levels. Excavation and recovery methodology Excavation was conducted in two ways. First, after clearing of the overlying humus and Roman layers, excavation was conducted by following the architectural features within trenches. Most of the artefact bags are labelled by trench and feature. The most pronounced features were trapezoidal houses, with stone-lined rectangular hearths, and horizontal clay floors cut back into the terraces in a manner similar to the houses at Lepenski Vir, Padina, and Vlasac. Second, where architecturally defined units were not visible, strata were excavated by a system of horizontal excavation units of arbitrary depth (usually 10 cm) that potentially cut across the sloping natural strata. This may have resulted in the mixture of materials from different periods. Where possible, artefacts from different coloured soil deposits within the same level were collected and analyzed separately. As a result, the Early Iron Age material was separated from the Early Neolithic (by the excavator — B. Jovanović, pers. comm. 1982). Hajdučka Vodenica was one of the first early prehistoric sites to be excavated in the gorge. As such, its importance was not recognized. Artefacts were haphazardly collected without the aid of any mechanical devices (e.g. sieves). Consequently, most bones were probably not recovered. Bone tools and other bones with evidence of modification were preferentially collected (B. Jovanović, pers. comm. 1982). The bias in recovery is obvious from the high percentage of tools in the assemblage (Table 1). No wellcollected collection from the region has such a high percentage of bone tools (Greenfield 1986, 1991). As with most other Mesolithic and Early Neolithic sites from the Iron Gates (Vlasac, Padina and Lepenski Vir — as cited in Bökönyi 1978; Clason 1980: 142), the sample from Hajdučka Vodenica is small. The exception is Vlasac, which yielded a much more substantial faunal assemblage. Vlasac was the last of the Mesolithic and Early Neolithic sites to be excavated in the gorge. At the time of its excavation, there was an increased awareness on the part of excavators from the region of the importance of faunal assemblages. Excavations at Vlasac were therefore conducted with a more systematic rigour that had been practised during earlier excavations in the region (S. Bökönyi, pers. comm. 1982). Because of its relative size and the improved recovery techniques employed, Vlasac provides a comparative sample for the scale and quality of the collections that were lost or discarded at earlier excavations in the Iron Gates area. Recovery

of faunal remains, nevertheless, was not fully systematic in any of the samples (Greenfield 1993). Bone preservation A variety of factors, other than recovery methodology, affect the quality of the bone assemblage. Almost all bones show signs of surface erosion, many with deep pitting revealing the trabecular structure. Soil pH levels were not measured since the excavation took place prior to an interest in taphonomy in the region. Location of the site on a loess terrace, the high sand component of the soil matrix, and the poor preservation of the surfaces of recovered bones indicate that extensive attrition of the bone assemblage occurred prior to excavation. Other factors affecting the nature of the assemblage include disturbance and reworking of deposits through down slope movement, river flooding and erosion, and post-occupational interments. The presence or absence of the various bone elements from each species does not reflect the differential butchery or disposal practices of the ancient inhabitants of the site. Instead, they are a direct reflection of the recovery procedures used during excavation. Given the problems recognized above, it is necessary to ask why any attention should be devoted to such a poor data set? Several reasons exist. First, the gorge has been completely flooded since the completion of the hydroelectric dam and comparable additional data sets are unlikely ever to appear. Second, it is the last of the collections to go unreported and can add to our discussion of early cultural adaptations in the region. Third, the vast majority of the faunal remains from Lepenski Vir and Vlasac were discarded after they were identified by Bökönyi (1971, 1978 — personal communications from Sandor Bökönyi, Svetozar Stanković, and Dragoslav Srejović, 1982). Only some selected samples were curated for posterity. Only a preliminary report was published on the Icoana sample before the premature death of the analyst (Bolomey 1973). Aside from the limited published data available, little new information can be squeezed out of the reports. Only the faunal remains from Padina and Hajdučka Vodenica, of the Serbian sites, have been entirely curated. Therefore, Hajdučka Vodenica is only one of two faunal data sets from the region currently available for reanalysis. Hence, the utility of this report lies in describing potentially useful material still available for analysis. Finally, the faunal collection can, as will be shown below, contribute to the discussion of the chronological position of Hajdučka Vodenica. Quantification procedures Quantification of bone assemblages has been the subject of numerous debates (Grayson 1984). Several techniques have been proposed, but only two are commonly used for Early Neolithic assemblages in the central Balkans: number of identified specimens (NISP) and minimum number of individuals (MNI) (Greenfield 1986, 1991). Both quantification procedures were used for the Hajdučka Vodenica fauna. During the NISP calculations, each bone or bone fragment was separately identified and analyzed using procedures outlined in Greenfield (1986). In brief, when articulations between bones or fragments were recognized, they were analyzed as belonging to the same individual and quantified

208

Haskel Greenfield: Reanalysis of the vertebrate fauna from Hajdučka Vodenica

as equal to one specimen regardless of the number of bones or fragments that were present. Thus, each separate specimen or group of articulated specimens was quantified only once. MNIs were calculated using Bökönyi’s (1970) method. Similar quantification procedures were used for all other sites in the region. No whole skeletons were found in the assemblage.

The sample — general considerations Faunal sample and species diversity With a total of 354 separate bone fragments recovered (Appendix 1), the faunal sample from Hajdučka Vodenica is the smallest collected from the Early Neolithic/Mesolithic sites in the Iron Gates (Table 2). In contrast to other samples from the area, the Hajdučka Vodenica sample lacks diversity, although remains of at least 12 different taxa were identified. This is undoubtedly a reflection of the recovery techniques employed at Hajdučka Vodenica (Greenfield 1993, this volume). Identified species include wild and domestic cattle, wild and domestic pigs, domestic dogs, red and roe deer, brown bear, chamois, rodent, fish, and human. Only some of the major domestic species are present in both the Mesolithic and Early Neolithic assemblages from Hajdučka Vodenica (i.e. cattle, pig and dog — all of which may have had a local source of domestication: Bökönyi 1974, 1978). Notably lacking are sheep and goat, which would have been domesticated in the Near East. The wild taxa at Hajdučka Vodenica lack diversity when compared to the larger samples from the gorge. For instance, there were no mollusc shells in the

Hajdučka Vodenica sample, while assemblages from contemporary sites in the gorge include many. Based on this comparison, we can assume that the Hajdučka Vodenica people probably exploited molluscs as well. The absence of shell in the sample is a result of the previously described sampling bias. The same is likely for fish and other typically under-sampled taxa in haphazardly collected samples. Size categories A small quantity of remains were identified only to more general taxonomic size categories (Table 2), such as medium-sized mammals (N=4) and large-sized mammals (N=1). Most of the fragments relegated to these categories were ribs, vertebrae, and other fragmentary remains that are normally difficult to identify to a higher taxonomic level. More commonly, remains were identifiable to a narrower, but still relatively non-specific taxonomic grouping of genera (e.g. Bos taurus/Cervus: N =40). Very few remains were totally unidentifiable (N =14). The paucity of minimally identifiable material is a reflection of the selective recovery procedures employed during excavation. Domestic:wild ratio The Mesolithic has a much lower percentage of domestic (N=38, 16.3% — Table 3) and higher percentage of wild mammals (N=209, 83.7%) than the Early Neolithic levels (domestic: N=14, 33.3%; wild: N=28, 66.7%). The Early Neolithic percentages of domesticates are higher than those from contemporary sites in the Iron Gates (Padina and Lepenski Vir), but lower than most Early Neolithic assemblages outside of the Iron Gates (Greenfield, this volume). In contrast to sites outside of the Iron Gates, all of

Table 2. Sum of number of fragments and NISP by taxon and period. Period unknown Fragments NISP Domestic

Wild

Unknown

N/A Grand Total

Taxon Bos taurus Sus scrofa dom. Canis familiaris Bos primigenius Sus scrofa fer. Canis lupus Cervus elaphus Capreolus capreolus Ursus arctos Martes martes Rupicapra rupicapra Pisces (sp.) Sus scrofa Medium-sized mammal Large-sized mammal Bos/Cervus Unknown Sub-Total

4 1

22

2 4 9 15

3 1

22

Fragments NISP

Mesolithic % of NISP

Fragments NISP

4 9 15

41

Grand Total % of NISP

Fragments NISP

3 8 3

3 8 3

7.14 19.05 7.14

5 23 6

5 22 6

2.70 11.89 3.24

8 35 10

8 33 10

1 2

1 2

2.38 4.76

23

23

54.76

1

1

2.38

1

1

2.38

16 12 1 139 1 1 1 1 13

2 12 1 130 1 1 1 1 3

1.08 6.49 0.54 70.27 0.54 0.54 0.54 0.54 1.62

17 14 1 184 1 2 1 1 4

3 14 1 175 1 2 1 1 4

4

2

2 2 1 30 12 47

1 8 3 12

1 4 2 7

28 1 33

22 1 25

2 4 1 40 13 60

3

3

13

12

16

15

57

52

255

222

354

315

2

Homo sapiens 42

Early Neolithic

209

The Iron Gates in Prehistory Table 3. Summary of domestic and wild ratio by period. Period unknown

Domestic Total

Early Neolithic

NISP

NISP

% of NISP

Mesolithic NISP

% of NISP

Grand Total NISP

4

14

33.33

33

17.84

51

Wild Total

22

28

66.67

152

82.16

202

Total

26

42

100.00

185

100.00

253

Table 4. Temporal comparison of the NISP of Cervus elaphus with antler, without antler, and all elements by period. Period unknown

Early Neolithic

Mesolithic

Grand Total

NISP

% of NISP

NISP

% of NISP

NISP

4

14

33.33

33

17.84

51

All Elements (including antler)

22

28

66.67

152

82.16

202

All Elements (excluding antler)

26

42

100.00

185

100.00

253

NISP Antler

Table 5. Number of burnt bone fragments by taxon and element.

Domestic Wild

Taxa Sus scrofa dom. Cervus elaphus Bos primigenius

Period:

Unknown

?

Mesolithic

Burning:

Carbonized (black)

Slightly burnt (red)

Partly carbonized

Element Radius Antler Astragalus Cranium – horn core, frontal and occipital

Grand Total

1 1 1 1

the Early Neolithic samples from the Iron Gates have much higher percentages of wild than domestic fauna. The source of the difference lies with the red deer antlers. The antlers represent over half of the red deer remains (Table 4). When the antler remains are removed from the bone counts, wild fauna dominate the assemblage, albeit with a much smaller percentage. The proportion of wild in the Mesolithic level is much lower than in other Mesolithic assemblages (probably because the antlers are counted in those assemblages). In contrast, the proportions are much more similar to the Early Neolithic assemblages from the region as a whole.

Sources of assemblage attrition Burning Most Mesolithic and Early Neolithic faunal reports do not mention burnt bones at all. At Hajdučka Vodenica, only a very small fraction of the bone sample was burnt (N=2 — Table 5). In the indeterminate period sample, a subadult domestic pig distal radius and a red deer subadult/adult antler fragment were burnt. In the Mesolithic sample, a red deer subadult/adult astragalus and an aurochs adult male horn core

1 1

2

were burnt. Three possible explanations why so few burnt bones are found easily come to mind. First, if the low fraction of burnt bones is a reflection of their original proportions in the assemblage, we have to assume that most bones were probably covered with meat when exposed to fire, were boiled in a stew-like concoction, or had their meat removed prior to cooking. In this case, there would have been limited contact between the bone and the flame. A second possibility is that the burnt bone may have simply disintegrated in the soil, since burnt bone breaks up into fragments more readily than uncooked bone (Bonfield & Li 1966; Shipman et al. 1984). A third reason that few pieces of burnt bone are found could be due to differential disposal patterns. For example, burnt bone is not necessarily discarded in a homogeneous fashion across sites. Unless it is moved to a secondary deposit, most burnt bone would be expected to be found near cooking areas (e.g. hearths). If they are redeposited, they should be found in nearby middens. The depositional context of the burnt bone is unclear. Based on the provenance information, none of the burnt fragments come from any of the excavated hearths or features. While several hearths were found, it is not clear if any burnt remains were recovered from them (B. Jovanović, pers. comm. 1982). Did they come

210

Haskel Greenfield: Reanalysis of the vertebrate fauna from Hajdučka Vodenica Table 6. Distribution of butchered elements by period. Period unknown Butchering type: Taxon Bos taurus Cervus elaphus Ursus arctos

Element Metacarpus Cranium Astragalus Astragalus

Part Distal Frontal and antler Whole Whole

Mesolithic

Early Neolithic

4

5

9

5

1

1

1

1

7

Grand Total

9

4

5

7

2

1 1

6

6 3

1 19 3 1

2

2

6

9

24

1

Total

1

1

1

1

1

Butchering type 4 chop mark/severed lengthwise to main axis 5 chop mark/severed cross-wise to main axis 7 slice cut mark crosswise to main axis 9 slice cut mark and break in bone (bone broken subsequently)

Table 7. Distribution of NISP by weathering stages. Period unknown

Early Neolithic

Mesolithic

Grand Total

NISP

NISP

% of NISP

NISP

% of NISP

NISP

% of NISP

Slight

24

27

48.21

123

54.42

174

53.87

Medium

17

29

51.79

101

44.69

147

45.51

0.00

2

0.88

2

0.62

41

56

100.00

226

100.00

323

100.00

Weathering stage:

Heavy Grand Total

from food preparation areas or from middens where the burnt bone would have broken up and disintegrated with relative speed? Similarly, low proportions of burnt bone are found in other Early and Late Neolithic assemblages (Greenfield 1991, 1993). Butchering A small sample of remains shows evidence of butchering marks: the result of skinning or limb dismemberment (Table 6). Most butchered remains derive from the Mesolithic (N=17 out of 185, 9.2%). These include a variety of taxa. A domestic cattle metatarsus had its distal condyles chopped off with a butchering instrument, probably in an attempt to sever the foot bones from the rest of the carcass. Three red deer astragali (subadult/adult) exhibit slicing marks perpendicular to the long axis of the bone, which are also probably by-products of severing the feet from the lower limbs. A number (N=13) of red deer cranial remains (usually frontal bones with antlers still attached) show signs of chopping, cutting and/or slicing especially around the pedicles to either remove the skin or the antlers (for tool preparation). While the number of butchered remains during the Early Neolithic declines (N=4 out of 42), the percentage is similar to the Mesolithic (9.5%). Three red deer crania with antler remains show evidence of chopping and slicing. A bear calcaneus has a slice perpendicular to the long axis of the bone, on the lateral side, as part of the severing of the muscles and ligaments to detach the foot from the lower limb. There are no data from

other Early Neolithic sites with which to compare these data. Weathering While burning and butchering are more or less ambiguous indicators of the degree of assemblage attrition, weathering is less so. Table 7 shows the distribution of NISP elements by weathering stage. The differences between either period and the total are insignificant; the pattern is the same regardless of the period. In general, the assemblage is more highly weathered than it should be in order to have any expectation that the more fragile specimens (based on age or morphology) would be found in the numbers according to which they were deposited at the site (Lyman 1994). Weathering was analyzed using three general stages following the system proposed by Meadow (1978). In general, slightly more than half (53.87%) of the assemblage has slight weathering (surface is discoloured, but unpitted). Slightly less than half (45.51%) exhibited medium weathering (compact surface pitted, lightly eroded and cracked). A small percentage (0.62%) also exhibited heavy weathering (major destruction of compact surface; cancellous bones extensively exposed). While it is unclear to what extent weathering has affected the other assemblages from the Iron Gates, the assemblage from Hajdučka Vodenica is only slightly more heavily weathered than contemporary assemblages (Greenfield 1993; Greenfield & Jongsma n.d.). For example, at Blagotin 35.8% was lightly weathered, 57% was moderately weathered and 7.1% was heavily weathered. The Early

211

The Iron Gates in Prehistory Table 8. Distribution of fragments by size class and taxon. Period unknown Whole 3/4

1/4

Early Neolithic ?

Whole 3/4

Mesolithic

1/2

1/4

?