Early Urbanism in Europe: The Trypillia Megasites of the Ukrainian Forest-Steppe 9783110664959, 9783110664935

Table of contents :
Contents
Preface
List of Contributors
List of Participants
Acknowledgements
1 Introduction
2 Theory and Practice for Trypillia Megasites
3 Landscape Studies
4 Site Studies
5 The Finds
6 Discussion
7 Conclusions
References
List of Figures
List of Tables
Index

Citation preview

Bisserka Gaydarska (Ed.) Early Urbanism in Europe The Trypillia Megasites of the Ukrainian Forest-Steppe

Bisserka Gaydarska (Ed.)

Early Urbanism in Europe The Trypillia Megasites of the Ukrainian Forest-Steppe

Managing Editor: Katarzyna Michalak Associate Editor: Łukasz Połczyński

ISBN 978-3-11-066493-5 e-ISBN 978-3-11-066495-9

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License. For details go to http://creativecommons.org/licenses/by-nc-nd/3.0/. © 2020 Bisserka Gaydarska and chapters’ contributors Published by De Gruyter Poland Ltd, Warsaw/Berlin Part of Walter de Gruyter GmbH, Berlin/Boston The book is published with open access at www.degruyter.com. Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Managing Editor: Katarzyna Michalak Associate Editor: Łukasz Połczyński www.degruyter.com Cover illustration: D. Hale and J. Watson

Contents Preface — XVI List of Contributors — XIX List of Participants — XXI Acknowledgements — XXII Bisserka Gaydarska & John Chapman 1 Introduction — 1 Bisserka Gaydarska & John Chapman 1.1 Introduction to the Trypillia Group — 2 Bisserka Gaydarska & John Chapman 1.2 History of Megasite Investigations — 8 John Chapman & Bisserka Gaydarska 1.3 Project Biography — 12 Bisserka Gaydarska & John Chapman 2 Theory and Practice for Trypillia Megasites — 19 Bisserka Gaydarska 2.1 The Theoretical Debate on Urbanism — 20 2.1.1 The Global Debate on Urbanism — 21 2.1.2 Trypillia Megasites – the Theoretical Debate — 23 2.1.3 The Social Formations of the Trypillia Megasites — 28 2.1.4 The Maximalists, the Minimalists and the Middle Way — 32 2.1.5 The Possibility of Trypillia Megasites — 36 John Chapman & Bisserka Gaydarska 2.2 Changing Perspectives – Stable Research Questions — 47 2.2.1 Summary and Assessment — 59 Marco Nebbia & Joe Roe 3 Landscape Studies — 60 Marco Nebbia  3.1 Introduction — 61

Marco Nebbia 3.2 Remote Sensing — 61 3.2.1 Introduction — 61 3.2.2 High-Resolution Imagery — 62 3.2.3 CORONA Imagery — 62 3.2.4 WorldView-2 Imagery — 64 3.2.5 Mapping Archaeology in Ukraine — 65 Marco Nebbia 3.3 Fieldwalking — 75 3.3.1 A New Methodological Agenda for the Ukrainian Forest-Steppe — 75 Joe Roe 3.3.2

Intra-Megasite Collection — 76

Marco Nebbia 3.3.3 Trypillia Off-Megasite Survey: A Combined Adaptive Sampling Strategy — 80 3.3.4 Peri-Fluvial Survey Investigations — 83 3.3.5 Site Sampling Strategy — 86 3.3.6 Assessing Site Sizes — 88 3.3.7 From Space to Field: Ground-Truthing Remote Sensing Interpretations — 90 Marco Nebbia 3.4 GIS Settlement Patterns – Trypillia — 93 3.4.1 Introduction — 93 3.4.2 The Data from the Encyclopaedia of Trypillia Civilization — 94 3.4.3 Megasite Locational Strategies: Why Were They Where They Were? — 99 3.4.4 Site Size Hierarchies — 99 3.4.5 Spatial Distribution of Trypillia Settlements: Site Clustering and Megasite ‘Centrality’ — 103 3.4.6 Megasite Micro-Hinterlands — 106 Marco Nebbia 3.5 Concluding Remarks — 109

Bruce Albert, Jim Innes, Konstantin Kremenetski, Andrew Millard, Marco Nebbia, Bisserka Gaydarska, John Chapman, Dan Miller, Duncan Hale, Brian Buchanan, Stuart Johnston, Mykhailo Videiko, Manuel Arroyo-Kalin, Tuukka Kaikkonen, Svetlana Ivanova, Stoilka Terziiska-Ignatova, Patricia Voke, Natalia Burdo & Natalia Shevchenko 4 Site Studies — 111 Bruce Albert, Jim Innes, Konstantin Kremenetski, Andrew Millard, Marco Nebbia, Bisserka Gaydarska & John Chapman 4.1 Palaeo-Environmental Studies — 112 4.1.1 The Nebelivka P1 Core — 112 4.1.1.1 Introduction — 112 4.1.1.2 Megasite Human Impacts — 112 4.1.1.3 The Age – Depth Model — 114 4.1.1.4 A Sedimentological Hiatus? — 115 4.1.1.5 Assessment of Ecological Impact — 115 4.1.1.6 Conclusions — 118 Dan Miller 4.1.2 The Molluscan Evidence — 119 John Chapman 4.1.3 Summary — 121 Duncan Hale 4.2 Geophysical Investigations and the Nebelivka Site Plan — 122 4.2.1 Introduction — 122 4.2.2 The Site Plan — 127 4.2.2.1 Perimeter Ditch — 127 4.2.2.2 House Circuits — 128 4.2.2.3 Assembly Houses — 133 4.2.2.4 The Quarters — 139 4.2.2.5 Inside the Inner Circuit — 139 4.2.2.6 Features Outside the Outer Circuit — 146 4.2.2.7 Summary — 148 Duncan Hale, John Chapman, Bisserka Gaydarska, Marco Nebbia & Brian Buchanan 4.3 Architectural Analyses — 148 Duncan Hale, John Chapman, Bisserka Gaydarska & Marco Nebbia 4.3.1 House Size Analysis — 148 4.3.1.1 The Total Sample — 149 4.3.1.2 The Zonal Analysis — 151

4.3.1.3 4.3.1.4 4.3.1.5 4.3.1.6



The Sector Analysis — 151 The Analysis of the Quarters — 151 The Analysis of House Sizes in Neighbourhoods — 153 Summary — 155

Brian Buchanan 4.3.2 Visibility Graph Analysis — 158 4.3.2.1 Introduction — 158 4.3.2.2 The Built Environment — 158 4.3.2.3 Computational Approaches to Space and Place — 160 4.3.2.4 Nebelivka and VGA — 162 4.3.2.5 Methodology — 164 4.3.2.6 VGA Analysis of the entire Quarters — 165 4.3.2.7 The Distributed Governance Model (Model A) and the Assembly Model (Model B) — 171 4.3.2.8 Discussion and Conclusion — 175 Stuart Johnston 4.4 The Experimental Programme — 181 4.4.1 Introduction — 181 4.4.1.1 Issue 1: The Creation of a Ploshchadka — 182 4.4.1.2 Issue 2: Detailed Interpretations of House Features — 184 4.4.1.3 Issue 3: Construction Materials and Fuel for House-Burning — 186 4.4.1.4 Issue 4: Deliberate House Burning – the Alternatives — 193 Bisserka Gaydarska, Marco Nebbia, Mykhailo Videiko, John Chapman, Manuel Arroyo-Kalin, Tuukka Kaikkonen & Svetlana Ivanova 4.5 Joint Excavations — 194 Bisserka Gaydarska, Marco Nebbia, Mykhailo Videiko & John Chapman 4.5.1 The Mega-Structure — 194 4.5.1.1 Introduction — 194 4.5.2 Interpretation — 211 John Chapman, Manuel Arroyo-Kalin, Tuukka Kaikkonen & Svetlana Ivanova 4.5.3 The Barrow — 212 Bisserka Gaydarska, Marco Nebbia, Stoilka Terziiska-Ignatova, Patricia Voke & John Chapman 4.6 Excavations, Durham Side — 214

Bisserka Gaydarska, Marco Nebbia & John Chapman 4.6.1 The Test Pits — 214 Bisserka Gaydarska, Stoilka Terziiska-Ignatova, Patricia Voke, Marco Nebbia & John Chapman 4.6.2 The Pit in Sondazh 1 — 228 Mykhailo Videiko, Natalia Burdo & John Chapman 4.7 Excavations, Ukrainian Side — 233 Mykhailo Videiko & John Chapman 4.7.1 Ditches — 233 4.7.1.1 Introduction — 233 4.7.1.2 Ditch Coring — 234 4.7.1.3 Sondazh 2 — 234 4.7.1.4 Sondazh 4 — 235 4.7.1.5 Sondazh 10 — 236 Mykhailo Videiko & John Chapman 4.7.2 House A9 — 236 Mykhailo Videiko & Natalia Burdo 4.7.3 Houses B17 and B18 and Their Pits — 239 Mykhailo Videiko & Natalia Burdo 4.7.4 The ‘Industrial Structure’ and Its Pit — 241 Andrew Millard 4.8 The AMS Dates — 246 4.8.1 Aims — 246 4.8.2 Initial Dating — 246 4.8.3 Simulations — 247 4.8.4 Sample Collection — 248 4.8.5 Radiocarbon Results — 250 4.8.6 Results and Discussion of Modelling — 250 4.8.6.1 Circuits and Streets — 252 4.8.6.2 Ordering Within and Between Quarters — 252 4.8.6.3 Ordering Within Rows — 253 4.8.6.4 Radial Structure — 253 4.8.6.5 Overall Occupation — 253 4.8.6.6 Comparison with Other Sites — 255

4.8.6.7 4.8.7

Coeval Houses — 255 Conclusions — 256

Natalia Shevchenko & Bisserka Gaydarska 4.9 Analyses of Building Materials — 256 4.9.1 Introduction — 256 4.9.2 Methods and Materials — 257 4.9.3 Clay Materials Used in Building — 258 4.9.4 Daub and Daub Impressions — 260 4.9.5 Building the Mega-Structure Walls — 260 4.9.6 Constructing the Podium and the Platforms — 260 4.9.7 The Production of a Storage-Jar — 261 4.9.8 The Destruction of the Mega-Structure: The Evidence from Daub Firing Temperatures — 261 Bisserka Gaydarska & John Chapman  4.10 Summary — 262 Bisserka Gaydarska, Marco Nebbia, John Chapman, Edward Caswell, Sophia Arbeiter, Eduard Ovchinnikov, Dmytro Gaskevych, Cătălin Lazăr, Theodor Ignat, Adrian Boyce, Amanda Dolan, Jason Newton, Dmytro Kiosak, Mykola Belenko, Oliver E. Craig, Harry K. Robson, Matthew von Tersch, Alexandre Lucquin, Zsuzsanna Tóth, Alice Choyke, David Orton, James Nottingham, Giselle Rainsford-Betts, Kim Hosking, Andrew Millard & Galyna Pashkevych 5 The Finds — 265 Edward Caswell, Sophia Arbeiter, Eduard Ovchinnikov, Bisserka Gaydarska, Marco Nebbia & John Chapman 5.1 Pottery — 266 5.1.1 Introduction — 266 5.1.1.1 Sampling and Comparative Method — 266 5.1.2 Taphonomy — 270 5.1.2.1 Burnt Houses in Test Pits — 270 5.1.2.2 Assembly Houses — 271 5.1.2.3 Pit, Sondazh 1 — 273 5.1.2.4 Summary — 273 5.1.3 Pottery Production, Consumption, Refitting and Post-Depositional Evidence — 274 5.1.3.1 Pottery Production — 274 5.1.3.2 Sherd Refits — 277 5.1.4 The Analysis of the Nebelivka Pottery Assemblages — 279 5.1.4.1 Sherd Numbers, Weights and Mean Sherd Size (Analysis 1) — 279

5.1.4.2 5.1.4.3 5.1.4.4 5.1.4.5 5.1.4.6 5.1.4.7 5.1.4.8 5.1.4.9 5.1.4.10 5.1.4.11 5.1.4.12 5.1.4.13 5.1.4.14 5.1.5

Pot parts (Analysis 2) — 285 Fabrics (Analysis 3) — 285 Fabrics vs. Form (Analysis 4) — 288 Vessel Form (Analysis 5) — 288 Comparisons Between Open and Closed Forms (Analysis 6) — 292 Estimation of the Minimum Number of Vessels (MINV) (Analysis 7) — 293 Vessel Size (Analysis 8) — 295 Distribution of Vessel Types in the Mega-Structure (Analysis 9) — 297 Decorative Style by Vessel Type (Analysis 10) — 299 Distribution and Placement of Decorative Motifs by Excavation Unit (Analysis 11) — 299 Decorative Motif Combinations (Analysis 12) — 304 Motif Linkage (Analysis 13) — 304 Discussion of the Pottery Analyses — 310 Comparisons with Other Trypillia Pottery Assemblages — 314

Bisserka Gaydarska, John Chapman, Marco Nebbia, Dmytro Gaskevych, Cătălin Lazăr, Theodor Ignat, Adrian Boyce, Amanda Dolan, Jason Newton, Oliver E. Craig, Harry K. Robson, Matthew von Tersch & Alexandre Lucquin 5.2 Special Finds — 326 Bisserka Gaydarska, John Chapman & Marco Nebbia 5.2.1 Figurines — 327 5.2.1.1 Making — 327 5.2.1.2 Types — 328 5.2.1.3 Gender — 329 5.2.1.4 Fragmentation — 331 5.2.1.5 Context of Deposition — 331 5.2.1.6 Condition — 333 5.2.1.7 Comparisons with Other Assemblages — 333 5.2.1.8 Summary — 335 Bisserka Gaydarska, John Chapman & Marco Nebbia 5.2.2 Tokens (Counters) — 336 John Chapman, Bisserka Gaydarska, Dmytro Gaskevych, Cătălin Lazăr, Theodor Ignat, Adrian Boyce, Amanda Dolan, Jason Newton, Oliver E. Craig, Harry K. Robson, Matthew von Tersch & Alexandre Lucquin 5.2.3 The Group of Miniature Vessels from the Mega-Structure — 338 Dmytro Gaskevych 5.2.3.1 Graphite in the Production of Pottery in the Ukrainian Para-Neolithic  — 338

Cătălin Lazăr & Theodor Ignat 5.2.3.2 Graphite Painted Ware Analogies in the East Balkans — 341 Adrian Boyce, Amanda Dolan & Jason Newton 5.2.3.3 Sourcing the Nebelivka Graphite — 345 Oliver E. Craig, Harry K. Robson, Matthew von Tersch, Alexandre Lucquin & John Chapman 5.2.3.4 The Interior of the Miniature Vessels — 346 5.2.3.5 Summary — 349 John Chapman, Marco Nebbia & Bisserka Gaydarska 5.2.4 Other Miniature Vessels — 350 Dmytro Kiosak, Mykola Belenko & John Chapman 5.2.5 Chipped Stone — 352 5.2.5.1 Introduction — 352 5.2.5.2 Raw Materials — 352 5.2.5.3 Technological Modes — 353 5.2.5.4 The 2009 Assemblage — 353 5.2.5.5 The 2012 Assemblage — 354 5.2.5.6 The 2013 Assemblage — 360 5.2.5.7 The 2014 Assemblage — 363 5.2.5.8 Discussion — 366 5.2.5.9 Conclusions — 368 John Chapman 5.2.6 Ground Stone — 370 Zsuzsanna Tóth & Alice Choyke 5.2.7 Worked Bone — 370 5.2.7.1 Introduction — 370 5.2.7.2 Description — 370 5.2.7.3 The Manufacturing Continuum — 374 5.2.7.4 Evaluation — 375 5.2.7.5 Conclusions — 376 John Chapman, Marco Nebbia & Bisserka Gaydarska 5.2.8 Other Special Finds — 377 John Chapman 5.2.9 Summary — 381

David Orton, James Nottingham, Giselle Rainsford-Betts, Kim Hosking & Andrew Millard 5.3 Animal Bones — 388 5.3.1 Introduction — 388 5.3.2 Areas and Assemblages — 388 5.3.3 Methodology — 391 5.3.3.1 Diagnostic and Non-Diagnostic Specimens — 391 5.3.3.2 Quantification — 393 5.3.3.3 Burning and Taphonomic Modification — 393 5.3.3.4 Measurements — 393 5.3.3.5 Age Data — 393 5.3.4 Taxonomic Frequencies — 394 5.3.4.1 Identification Issues — 394 5.3.4.2 Regional Comparisons — 398 5.3.4.3 Intra-Site Comparisons — 401 5.3.5 The Mega-Structure — 404 Andrew Millard 5.3.6 Isotopic Dietary Information — 407 David Orton 5.3.7 Summary — 408 John Chapman, Galyna Pashkevych & Dan Miller 5.4 Plant Remains — 409 John Chapman 5.4.1 The 2009 Season — 409 Galyna Pashkevych 5.4.2 The Mega-Structure (2012) — 409 Dan Miller 5.4.3 The 2013 and 2014 Seasons — 411 5.4.4 Summary — 412 John Chapman 5.5 Summary — 412 Bisserka Gaydarska & John Chapman 6 Discussion — 415 6.1 The Nebelivka Megasite — 416 6.1.1 Introduction — 416

6.1.2 6.1.2.1 6.1.2.2 6.1.2.3 6.1.2.4 6.1.2.5 6.1.3 6.1.3.1 6.1.4 6.1.5 6.2 6.2.1 6.2.2 6.2.2.1 6.2.2.2 6.2.3 6.2.4 6.2.5 6.2.5.1 6.2.5.2 6.2.5.3 6.2.6 6.3 6.3.1 6.3.2 6.3.3 6.3.4 6.4 6.4.1 6.4.2 6.5

Four Nested Spatial Levels — 417 The Household — 417 The Neighbourhood — 421 Quarters — 423 The Whole Site — 426 Production, Distribution and Consumption — 430 Modelling the Growth of Nebelivka — 433 Assessment of the Models — 439 The Origins of the Megasites — 445 The Demise of the Megasites — 455 Megasites – a Comparative Approach — 458 Introduction — 458 Megasite Plans — 459 Taljanki — 459 Majdanetske — 461 Daily Resources — 469 Building and Burning Resources — 472 Testing the Alternative Models at Taljanki and Majdanetske — 476 The Assembly Model at Majdanetske — 477 The Pilgrimage Model at Taljanki and Majdanetske — 478 The Distributed Governance Model at Taljanki and Majdanetske — 479 Summary — 480 Low-Density Urbanism – a Global Approach — 482 An Example of a Small Trypillia Settlement — 482 An Analytical Construct for ‘Urban’ Sites — 484 Summary — 497 Comparisons with Other Trypillia Megasites — 498 Low-Density Urbanism — 499 The Low-Density Urbanism Model — 499 Summary — 505 Summary — 505

John Chapman & Bisserka Gaydarska 7 Conclusions — 508 7.1 Principal Results — 509 7.2 What People Thought About the Nebelivka Megasite — 516 7.2.1 The Viewpoint of a Nebelivka Guardian — 516 7.2.2 The Viewpoint of an Early House-Builder — 518 7.2.3 The Viewpoint of a Clan Leader — 519 7.2.4 The Viewpoint of a Visitor to the Assembly — 520 7.2.5 The Viewpoint of an Organizer of the Assembly — 521 7.2.6 The Viewpoint of a Pilgrim — 522

7.2.7 7.2.8 7.2.9 7.2.10 7.3 7.3.1 7.3.2 7.4

The Viewpoint of an Adolescent Visiting Nebelivka for the First Time — 522 The Viewpoint of a Nebelivka Ritual Leader — 523 The Viewpoint of a Trader Visiting Nebelivka at Assembly Time — 524 The Viewpoint of One of the Last Generations of Residents at Nebelivka — 525 A Future Research Agenda — 526 Issues for Trypillia Studies — 526 Issues for Nebelivka — 528 Endwords — 528

References — 529 List of Figures — 558 List of Tables — 566 Index — 568

Preface This open access monograph is the fruit of a collaborative Project between Ukrainian specialists from the NAS Institute of Archaeology, Kyiv, specialists in European prehistory from Durham University Department of Archaeology (UK) and many other friends and colleagues who have been working with us on Trypillia archaeology, urbanism and other cognate fields. The central topic of investigation was whether the Trypillia megasites of the 4th millennium BC could be considered as the first urban settlements in Europe, if not the world. Long before the Project began, Roland Fletcher (1995) had highlighted these megasites as the only known exception to his global model of the limits to agrarian settlement growth. But relatively little had been published on the megasites in English, French and German, with the important exception of Linda Ellis’ discussion of these sites in her monograph on Cucuteni-Trypillia pottery (Ellis 1984). In view of their principal publications in Russian and Ukrainian, it is hardly surprising, then, that Trypillia megasites have been excluded from discussions on early urbanism until the late 2000s and that, even in 2011, sentences such as “The first cities in the Near East – Mediterranean Basin appeared in Southern Mesopotamia, or Sumer, the creation of a people we call the Sumerians” (Gates 2011, p. 30) could be published in supposedly serious works on early urban developments. It is also pertinent that the question of megasite urban status has also divided Ukrainian archaeologists, with a distinct minority contemplating a notion that has received regular attacks from their colleagues. However, the high aesthetic levels reached by Trypillia potters in making their fine wares and figurines have led, in the 2000s, to a series of exhibitions in major museums (Royal Ontario Museum, the Institute for the Study of the Ancient World, New York, the Ashmolean Museum, Oxford, the National Museum of Archaeology, Kyiv, etc.), which not only prompted wider research questions about Trypillia settlements, pottery production and even megasites but also succeeded in introducing these questions to the general public and a wider archaeological audience. There was another glaring gap in urban studies that made the Trypillia case interesting: the lack of interpretative archaeological engagement with the urban question and the general lack of theoretical engagement by Ukrainian prehistorians dealing with Trypillia studies. As regards the former, post-processual, interpretative and ontological approaches have often eschewed the big questions of prehistory, feeling more comfortable with the évènement rather than the conjoncture1 – the quotidian rather than the medium-term, let alone the longue durée. With only a few exceptions, such approaches have steered well clear of urban origins, let alone at a Eurasian scale. As regards the latter, again with important exceptions, the dominant approach to Trypillia studies has been founded on the culture history of the Russian

1  The three temporal scales used by the Annales historian Ferdnand Braudel (1975).

XVIII 

 Preface

school, with one of the most widespread assumptions about the archaeological record being that it constitutes a direct reflection of past lifeways. All of these strands of positive and negative thinking came together in a Project designed to take our understanding of the Trypillia megasites beyond the traditional debates on terminology (‘was this site a city?’) and material culture (‘does this ornament hoard represent an elite deposit?’). The AHRC-funded Project which informs the title of this monograph ran from 2012 to 2016. Although our views on Trypillia megasites and urban origins have diverged so strongly from our partners in this Project – Drs. Mykhailo Videiko and Natalia Burdo – that it has regrettably not been possible to publish our results together, we still wish to record our debt of gratitude to our partners for their help in setting up the Project, for their hard work on the field seasons and for their constant supply of provocative materials forcing us to clarify our (alternative) views on Trypillia megasites. We have reached a point in our investigations of Trypillia megasites which we hope has advanced the debate from its position in the late-2000s. New issues have arisen which have largely replaced the older positions and we are confident that the research of this Project and the Ukrainian - German Project working at Taljanki and Majdanetske has indeed created a second methodological revolution, setting up a new fieldwork agenda that will endure for two decades or more. However, without new theoretical insights into urban megasites, we were never going to be able to convert the methodological revolution into a more profound re-evaluation of the Trypillia megasites. We offer this monograph in the hope that this re-formulation of critical issues will have moved the field further in terms of theory as well as method. The Project’s mode of publication is also novel insofar as we have created a twin-track, open access publication, consisting of the interpretative materials of the Project in this monograph and the basic excavation and fieldwork data in a Project Archive hosted by the University of York’s Archaeology Data Service (https:// doi.org/10.5284/1047599). A similar approach was taken by the Tundzha Regional Archaeology Project (TRAP) in Bulgaria, who have provided basic Project data in an open context Archive entitled the TRAP Digital Archive (DOI: https://doi.org/10.6078/ M7TD9VD3) to support a hard-copy publication (Ross et al. 2018). BISSERKA GAYDARSKA

List of Contributors Dr. Bisserka Gaydarska, Honorary Research Fellow in Archaeology, Durham University ([email protected]) Professor John Chapman, Emeritus Professor of Archaeology, Durham University ([email protected]) The late Professor Tony Wilkinson, erstwhile Professor of Archaeology, Durham University. Dr. Andrew Millard, Associate Professor of Archaeology, Durham University ([email protected]) Professor Thomas Higham, Deputy Director, NRCF, Oxford ([email protected]) Professor Geoff Bailey, University of York ([email protected]) Dr. Nikolaos Galiatsatos, Durham University and Bath University Departments of Geography ([email protected]) Dr. Marco Nebbia, PDRA, UCL - Institute of Archaeology (m. [email protected]) Dr. Bruce Albert, independent palaeo-environmental consultant ([email protected]) Dr. Jim Innes, Department of Geography, Durham University ([email protected]) Dr. Konstantin Kremenetski, Chaffey College, California ([email protected]) Dr. Manuel Arroyo-Kalin, Lecturer, UCL Institute of Archaeology ([email protected]) Mr. Tuukka Kaikkonen, PhD candidate, Australian National University, Sydney ([email protected]) Mr. Duncan Hale, Senior Archaeologist, Archaeological Services, Durham University ([email protected]) Dr. David Orton, BioArch, Department of Archaeology, University of York ([email protected]) Mr. James Nottingham, BioArch, Department of Archaeology, University of York ([email protected]) Ms. Giselle Rainsford-Betts, Department of Archaeology, University of York ([email protected]) Ms. Kim Hosking, Department of Archaeology, University of West Yorkshire Archaeology Services ([email protected]) Professor Oliver Craig, BioArCh, University of York ([email protected]) Dr. Harry Robson, BioArCh, University of York ([email protected]) Mr. Matthew von Tersch, BioArCh, University of York ([email protected]) Dr. Alexandre Lucquin, BioArCh, University of York ([email protected]) Professor Adrian Boyce, SUERC, Glasgow University ([email protected]) Ms. Amanda Dolan, SUERC, Glasgow University Mr. Jason Newton, SUERC, Glasgow University Mr. Stuart Johnston, Archaeologist, ASDU ([email protected]) Mr. Dan Miller, independent palaeo-environmental consultant ([email protected]) Dr. Brian Buchanan, Eastern Washington University ([email protected]) Dr. Zsuzsanna Tóth, Institute of Archaeology, Budapest ([email protected]) Dr. Alice Choyke, Central European University, Budapest ([email protected]) Mr. Joe Roe, University of Copenhagen ([email protected]) Ms. Kirrily White, PhD candidate, University of Sydney, Australia ([email protected]) Mr. Vlad Litkevych, independent consultant ([email protected]) Ms. Patricia Voke, Archaeologist, Wessex Archaeology ([email protected]) Mr. Côme Ponroy, former PhD candidate, Durham University. Mr. Ed Caswell, PhD candidate, Durham University ([email protected]) Ms. Sophia Arbeiter, PhD candidate, University of Pittsburgh ([email protected]) Mr. Ed Treasure, PhD candidate, Durham University ([email protected]) Dr. Cătălin Lazăr, ArchaeoScience#RO Platform, ICUB, University of Bucharest ([email protected])

XX 

 List of Contributors

Dr. Theodor Ignat, Department of Archaeology, Bucharest Municipality Museum (theodor_ignat@ yahoo.com) The late Dr. Charles Schwartz, Los Angeles. Ms. Christina Unwin, freelance illustrator ([email protected]) Lauren Woodard, British Museum ([email protected]) Janine Watson, ASDU, Durham University ([email protected]) Dr. Valentine Pankowski, NASU Institute of Archaeology, Kyiv ([email protected]) Vince Cherubini, freelance excavator ([email protected]) Nikita Ivanov, Masters candidate, Taras Shevchenko University, Kyiv ([email protected]) Maria Gurova, BAS Institute of Archaeology, Sofia ([email protected]) Yvonne Beadnell, freelance illustrator ([email protected]) Kirsty Harding, SHARE, Cardiff University ([email protected]) Professor Mykhailo Videiko, Professor of Archaeology, Taras Shevchenko National University of Kyiv ([email protected]) Dr. Natalia Burdo, formerly NAS Institute of Archaeology, Kyiv ([email protected]) Dr. Galyna Pashkevych, NAS Institute of Archaeology, Kyiv ([email protected]) Dr. Edvard Ovchinnikov, NAS Institute of Archaeology, Kyiv ([email protected]) Dr. Natalia Shevchenko, Scientific Department of Physico-Chemical Investigations of the National Scientific-Research Restoration Centre of Ukraine ([email protected]) Dr. Dmytro Kiosak, I.I. Mechnikov Odessa National University ([email protected]) Dr. Dmytro Gaskevych, NASU Institute of Archaeology, Kyiv ([email protected]) Dr. Svetlana Ivanova, NAS Institute of Archaeology, Kyiv ([email protected]) Dr. Olena Sekerskaya, K.D. Ushynskyi South Ukrainian National Pedagogical University Odessa ([email protected]) Mr. Vitalii Rud, PhD candidate, NAS Institute of Archaeology, Kyiv ([email protected]) Mr. Mykola M. Belenko, Junior Research Fellow, NASU Institute of Archaeology

List of Participants Mayor Bobko (Nebelivka) Ana Shyanova Julia Ushkova Olya Zaitseva Stoilka Terziiska-Ignatova Stanislav Ţerna Victoria Singleton Andy Platell Vince Cherubini Ben Cullen Graeme Laidlaw Slawomir Szyszka Ruth Hatfield Marco Pietrovito Richard Villis Natalie Swann Nathan Thomas Andrew Blair Ashley Bryant Igor Polishchiuk Larisa Polishchiuk Vladimir Gutnev Olesya Galkina Rostislav Bandurist Andriy Bilogoloviy Valentina Bidenko Maxim Petrov Sergei Slisarenko Sergei Polishchiuk Katya Mokryak Katya Kucherenko Yuliya Romashenko Tatiana Telih Igor Bobko Sergei our most reliable driver O. Lemeshtak V. Stanislavchuk M. Tkchenko I. G. Angileski N. Lisokonj R. Lisokonj S. Chernega V. Chernega A. Melnik Alina Ptashnik Ya. Koshelniy O. Koshelniy

Ronan O’Donnell Cathryn Reusch Lisa Snape-Kennedy Sam Wilford Chris Charmley Jessica Greenhalgh Ashild Vagene Sean Hamer Ruaridh Ellison Emily Dutton Monika Dimitrova Caroline Smith Jie Taylor Yee Min Gan Ed Treasure Tatiana Purton Emily Flach Esther Lusty Alvin Chua Tom Patterson Mathias Jensen Mark Houshold Iunn Jenn Ong Max Ratcliffe Bryony Rogers Kate Swinson Joe Ryder Tom Bergquist Alex Aflalo Tom Wright Alexandra Ames the late Chris Bond Lisa-Ellen Meyerling Sabrina Ki Zu Ning Tay Ben Spillane Nathan Sleaford Carrie Armstrong Louisa Gidney Charlotte O’Brien Richard Allen Tina Jakob Ann Marie Clinnick Václav Vondrovský

Acknowledgements It is our pleasure to acknowledge all of the contributors to this volume (listed overleaf, pp. XIX–XX) and all of the participants in Project research (listed overleaf, p. XXI). Without your hard and devoted work, we should never have been able to complete the fieldwork and excavation and bring the Project to publication. We are happy to acknowledge Project funding from many sources: the British Academy (Small Research Grant for the 2009 summer season); the Arts & Humanities Research Council (Grant No. AH/I025867: 2012–2016: principal Project funding); National Geographic Society (Grant No. 2012/211: excavation of the Nebelivka Megastructure); Society of Antiquaries of London, the Prehistoric Society and the Rosemary Cramp Fund of Durham University (the excavation of the experimental burnt house 2017); William Rust Family Foundation grant (illustrations for publication, 2017); the UK National Radiocarbon Facility (the majority of the AMS dating programme); and the University of Sydney and the Amerind Foundation (travel costs to the Amerind Foundation workshop, May 2014). The Project is most grateful for the institutional support of our home institutions: Durham University and the successive Chairs of Archaeology, Professors Chris Scarre and Chris Gerrard, as well as Rob Witcher, all of the much-appreciated Departmental secretarial and technical support and the expert financial advice of Sandra Robinson; the Institute of Archaeology, NAS, Kyiv, and especially their Director, Professor P. P. Tolochko and the members of their Prehistory Section; Volodymyr Vynnychenko Kirovograd State Pedagogical University and especially Dr. Iryna Kozyr, who co-ordinated her student participation; Kirovograd County Council, the successive Governors of Kirovograd Domain, the members of their Cultural Heritage section, Mr. Valentin Sobchuk and Mrs. Nadia Lisnyak, and the staff of the Kirovograd Historical Museum and their former conservator, Mr. Stanislav Fedorov; Novoarkhangelsk District Council and their leaders; the Local Mayor (Mr. Bobko), successive school directors and villagers of Nebelivka, who made us so welcome in their village and all of those Nebelivka farmers who kindly gave us permission to walk across/dig up their land. We also wish to acknowledge the help given to us to mount two international conferences – one in Kirovograd in 2015 and one in Durham in 2016 – as well as our thanks to all of those who presented papers or posters at the two meetings. In addition, we should like to thank the Directors and staff of all those Museums and Universities who hosted the Project Travelling Exhibition in 2015 and 2016 (Kirovograd Historical Museum; Archaeological Museum Varna; National Museum of History of Moldova, Chişinau; Eötvös Lórand University School of Archaeological Sciences, Budapest; Institut für Ur- und Frühgeschichte, Christian Albrechts University, Kiel; Durham University Museums; and University of Southampton Department of Archaeology). This Project has benefited from research conversations, short or sustained over many years, with a host of friends and colleagues to whom we offer our grateful thanks: the many friends and colleagues who continue to discuss ‘Eurasian urbanism’

Acknowledgements 

 XXIII

with us – in particular, Roland Fletcher and the members of the Hawaii/Tucson ‘Big Sites’ group (Nam Kim, Tim Pauketat, Eduardo Neves, Mike Parker Pearson, Innocent Pikiraye, Akin Ogundiran, Tom Moore, Kirrily White and Patricia McAnany), the participants at the Durham 2016 Conference, David Wengrow, Peter Bogucki, Richard Bradley and Geoff Bailey as the external member of our Project Panel who offered us much valuable advice; the Durham colleagues for their discussion of urban developments in their areas of specialization; the late Dan Monah and the late Gheorghe Dumitroaia for many hours of discussion about Cucuteni-Trypillia matters; E. Ovchinnikov, N. Ivanov, S. Ţerna, and particularly to A. Diachenko for providing valuable information and advice on the Trypillia group; Dmitriy Chernovol for discussions of house-burning; and Pál Raczky for kind discussion of Hungarian Mediaeval kilns and fine wines. We are also grateful to Dr. Knut Rassmann (RömischGermanisch Kommission, DAI, Frankfurt-am-Main) for permission to re-publish his geophysical plans. Finally, the production of the monograph has been greatly aided by the patience and skill of Katarzyna Michalak and Łukasz Połczyński at De Gruyter, as well as their anonymous referees – to all of you, we are most grateful. BISSERKA GAYDARSKA

Bisserka Gaydarska & John Chapman

1 Introduction

In the first chapter, we introduce readers to the world of Central and East European prehistory and, in particular, the Cucuteni-Trypillia group, which covered up to 250,000km2 and lasted over two millennia. This group was one of the largest networks ever to develop in European prehistory prior to the Bronze Age and one of the challenges is to identify mechanisms which enabled the survival of such a network. The key aspect of this massive network concerns settlement size, which reached a range of 1–5ha in the Cucuteni zone but expanded to the largest settlements yet known in 4th millennium Eurasia in one part of the Trypillia zone. The largest of these massive sites – known as ‘megasites’ – ranged from 100ha to 320ha and were comparable in size to the first Near Eastern cities. The lack of a prior inter-disciplinary project to investigate Trypillia megasites led to the establishment of an agreement between the Institute of Archaeology NAS (Dr. Mykhailo Videiko) and the University of Durham Department of Archaeology (Professor John Chapman) for a four-year project funded by the AHRC. In this Chapter, we introduce the Project by way of a Project Biography, with an explanation of the origins and growth of the Project, the starting positions of all concerned and the ways in which our theoretical differences led to alternative interpretations. This monograph is the Project’s final report.

2 

 Introduction

Bisserka Gaydarska & John Chapman 1.1 Introduction to the Trypillia Group The time-place distribution of the Trypillia2 – Cucuteni groups  – over two millennia (5000–2800 cal BC) and between 225,000 and 250,000km2  – makes them one of the largest and most long-lasting groups in Neolithic and Copper Age Europe (Figs. 1.1–1.2). The two names stem from the AD 19th century practice of naming a group of similar pottery after the first important site where such pottery was found. Thus, the distinctive painted ‘Cucuteni’ pottery was named after the promontory site ‘Cucuteni-Cetăţuia’, some 60km West-North-West of Iaşi, Moldavia, North-East Romania, first excavated by N. Beldiceanu in 1885 (Monah D. & F. 1997, p. 21). Within 15 years, Vikentiy Khvoika had found broadly similar painted pottery in his excavations near the village of Trypillia, some 40km South of Kyiv, Ukraine (Khvoika 1901, 1904). Similar material was also found in a third country – Moldova – at the site of Petreni, where von Stern published a remarkable site plan and excavated a number of what he thought to be pottery kilns (von Stern 1907). It was not until 1932 that the analogous ceramic assemblages from the three areas of Moldavia, Moldova and Ukraine were recognised to derive from the same ‘culture’ – henceforth named the ‘Cucuteni-Trypillia culture’ (Schmidt 1932) (Figs. 1.3–1.4).

Figure 1.1: Map of Cucuteni-Trypillia distribution (by M. Nebbia).

2  The term ‘Trypillia’ is the Ukrainian name for the group, as compared to the Russian ‘Tripolye’.



Introduction to the Trypillia Group 

Figure 1.2: Timeline of Cucuteni-Trypillia group.

Figure 1.3: Cucuteni A pottery, Drăguşeni (by B. Gaydarska, based upon Crîşmaru 1977, Fig. 20).

 3

4 

 Introduction

Figure 1.4: Trypillia BII – CI pottery, Bug-Dnieper Interfluve (by L. Woodard, based upon Ryzhov 2012, Figs. 6.4–6.5).



Introduction to the Trypillia Group 

 5

Those few syntheses of large regions that concern the prehistory of Central and Eastern Europe have often been written by outside specialists such as Gordon Childe (1929), Graham Clark (1969), Sarunas Milisauskas (1978), Marija Gimbutas (1982) and Jan Lichardus/Marion Lichardus-Itten (1985). In each case, the CucuteniTrypillia group has been considered as a major component of Late Neolithic/ Chalcolithic ‘Climax Society’ (in Nandris’ helpful term of 1978). While the Lichardus/ Lichardus-Itten approach is a straightforward culture history, Milisauskas develops a thematic, processual approach to Trypillia. In her heavily gendered model of the Kurgan transformation of East European Climax society,3 Gimbutas (1982) recognises Cucuteni-Trypillia as the pinnacle of a matriarchal society, with female priests and temples where goddesses and gods were depicted by a wide range of fired clay figurines (Fig. 1.5). It has rarely been underlined that of all the figurine-rich, settlementbased Climax societies, Cucuteni-Trypillia continued for over a millennium after the transformations of all other such Balkan regional groups (e.g., the KodzhadermenGumelniţa-Karanovo VI group, which hardly lasted long into the 4th millennium BC) (Lazar, Chapter 5.2.3). Accepting that there were minor regional variations in the longlived Cucuteni-Trypillia phenomenon does not deflect our attention from its immense stability and its conservative resistance to those major changes that had transformed all other Balkan Chalcolithic groups. Three key points stand out from the long history of Trypillia-Cucuteni studies – the utter predominance of the domestic domain over the mortuary sector in both groups, the closely related near-absence of the materialization of hierarchies in either group and the differential development of massive sites (the so-called ‘megasites’) in certain zones of the Trypillia group but not in others and not at all in the Cucuteni sites. Indeed, the Trypillia megasites stand out from the rest of Eastern, South-East and Central European Neolithic and Chalcolithic settlement, which was normally limited in size to 10ha, whatever the settlement form – tells, open sites or enclosed sites (Fig. 1.6). In European prehistory, there is a marked contrast between groups where the domestic domain was strong and groups with an often monumental mortuary zone (Chapman 1992). Regional sequences often show a change from one mode to the other, suggesting that an opposition to the previous dominant ideology is partly responsible for the change. There can be no doubt that the Trypillia-Cucuteni group is one of the most strongly household-oriented groups in European prehistory. Not only were extramural cemeteries absent, except in the latest stage of Trypillia in the North Pontic steppe zone and the Dniester valley, but there were hardly any intramural burials or fragmentary bone deposits in the hundreds of excavated settlements4. The

3  The term 'kurgan' is the Russian word for 'barrow'. 4  A striking exception concerns the Scânteia settlement, where House 9 was clearly a mortuary house containing the fragmentary remains of at least 33 individuals (111 bones or teeth: Bem 2007).

6 

 Introduction

Figure 1.5: Cucuteni-Trypillia figurines (by Y. Beadnell, based upon Monah D. 1997).



Introduction to the Trypillia Group 

Figure 1.6: Settlement model for Central and Eastern Europe: key – darker shades show higher densities of a site type, lighter shades lower densities (by C. Unwin).

 7

8 

 Introduction

absence of human bone remains has been linked to both the ubiquity of figurines (Bailey 2010) and house-burning (Kruts 2003; Chapman 2015). In other parts of Central and Eastern Europe, the appearance of formal intramural burial and, especially, cemeteries has often been related to the onset of richer, more diverse material culture, sometimes conceptualised as ‘elite’ or ‘prestige’ goods (Bailey 2000). A good example is the contrast between the Early Lengyel group, with little copper and no community group burials, and the Late Lengyel group, with many community group burials, often containing copper grave goods. The excavation of the huge Lengyel settlement-and-mortuary complex at Alsónyék supports this contrast, for copper ornaments appear early in the Lengyel burial sequence (Bánffy et al. 2016) and rarely if ever in such early settlements. Thus, a tenable position for the scarcity of prestige goods in the Trypillia-Cucuteni group is the rarity of human burials, whether intra-mural or extra-mural: Trypillia burials appeared late in the sequence. While prestige goods are more common in Cucuteni than in Trypillia, their context takes the form of the deposition of hoards, of which the most significant was the Karbuna hoard, dated to Trypillia Phase A and containing a rich assemblage of copper, Spondylus and red deer tooth pendant ornaments (Dergachev 1998). While the deposition of other ornament hoards and, indeed, also hoards of exotic long flint blades, occurred in Trypillia, they are notable for their rarity. As yet, not a single settlement hoard has been found on a Trypillia megasite, although a copper axe of Mareş’ Type A.B.10.1.4 (Mareş 2002, Pl. 51/8–10 & Harta 12) was found in House Zh-2 at Majdanetske (Shmaglij & Videiko 2001–2, Fig. 54/15 & 55/1). Moreover, the first gold find in Trypillia settlements was found at Nebelivka (see below, Section 5.2.8). Until the recent investigations of megasites, there was also a general rarity of architectural and ritual differentiation in Trypillia settlements. The current view identifies a paradox in Trypillia exchange networks – massive megasites with potential demand for huge resources, with possible hierarchical developments, in contrast to the paucity of exotic prestige goods. The third characteristic of the Trypillia group – its megasites – is of central importance to this Project and deserves its own history of investigations.

Bisserka Gaydarska & John Chapman 1.2 History of Megasite Investigations There are now several accessible accounts of the discovery and investigation of Trypillia megasites (chapters in Menotti & Korvin-Piotrovskiy 2012; Chapman et al. 2014b, 2015; Chapman & Gaydarska 2016; Kruts 2012; chapters in Müller et al. 2016b) to complement accounts in Russian or Ukrainian (e.g., Videiko 2012, 2013). In terms paralleling those first proposed by Thomas Kuhn (1970), the history of investigations comprises three phases of innovative fieldwork practices (viz., ‘scientific revolutions’) separated by two long periods of ‘normal excavation’ (viz., ‘normal science’), in which progress was dictated by the finances available for summer fieldwork.



History of Megasite Investigations 

 9

The discovery of the eponymous site of Trypillia was made by Khvoika in the 1890s. The distinctive architectural remains of burnt houses defined by a mass of fired clay daub were first interpreted as ‘mortuary houses’ but were later given the correct designation of dwellings (Kruts 1990). It is important to acknowledge that Ukrainian prehistorians were the first in Europe to recognise the deliberate burning of wattleand-daub houses as part of ritual practice. The elaborate and beautiful bichrome and polychrome painted pottery brought the Trypillia group into the European Neolithic family of painted pottery ‘cultures’ (Dimini, Butmir, etc.: von Stern 1907; Childe 1945). A succession of Trypillia sites was excavated using Khvoika’s techniques, which have been passed on to the fourth generation of those currently excavating megasites. The largest scale of excavation of a Trypillia settlement was reached by Passek in her postWW2 excavations of Volodymyrivka and Kolomiishchina (Passek 1949, 1949a). The second phase of innovation – termed ‘the first methodological revolution’ (Chapman et al. 2014b) – followed a 60-year period of ‘normal excavation’ of smaller and medium-sized Trypillia settlements. Given the normally tight restrictions on aerial archaeology in Eastern Europe (Braasch 1995), it was remarkable that Dudkin was able to use military photographs for the purposes of archaeological investigation (Dudkin 1978). These images revealed not only massive settlements but also hinted at regular concentric planning of the kind first noted by von Stern at Petreni (Dudkin 1978; first re-published in the West by Ellis 1984) (e.g., Yatranivka: Fig. 1.7a). Three stages of fieldwork were needed to confirm the association of the Trypillia group with the aerial images: (a) ground-truthing showed that Trypillia pottery was found in the areas covered by the aerial images; (b) a pioneering use of geophysical investigation pinpointed house-sized magnetic anomalies on the same sites (Fig. 1.7b); and (c) the excavation of a sample of these anomalies showed the typical Trypillia mass of burnt daub (the so-called ‘ploshchadka’). These innovations characterised the ‘first methodological revolution’ of the late 1960s – early 1970s. They form the basis for everything that later scholars achieved. This cluster of innovations set a new agenda for the next 35 years – a period of ‘normal excavation’ which produced a mass of new fieldwork and excavation data from sites such as Majdanetske (Shmaglij & Videiko 2001–2) and Taljanki (Kruts 1990). This was a period of incremental growth in the understanding of many aspects of megasites, not least the planning principles underlying megasite spatial development, the way in which domestic houses were built and deliberately burnt down and the subsistence basis of the huge populations (Kruts 1990). Excavation of 43 structures at Majdanetske and almost 50 structures at Taljanki (Shmaglij & Videiko 2001–2; Kruts et al. 2005; cf. Burdo et al. 2013) has provided detailed architectural plans and offered reconstructions of 1- and 2-storeyed houses full of ceramics, figurines and animal bones. However, the excavation of houses on its own was not likely to provide an accurate or reliable internal site chronology. The cumulative results of fieldwalking projects attached to major excavations and the systematisation of Trypillia settlement data led to an early attempt by Linda Ellis (1984) to produce

10 

 Introduction

Figure 1.7: Early remote sensing of Yatranivka: (a) plot of air photograph; (b) geophysical plot (by L. Woodard, based upon Videiko 2013).

the first regional settlement plans, interpreted as a 3-level, size-based site hierarchy in the Uman area. Twenty years later, continuing accumulation of field data led to an important synthesis – the ‘Encyclopaedia of Trypillia civilization’ (Videiko 2004; for critical analysis, see Chapter 3.3). A division among Trypillia specialists emerged on the fundamental nature of megasites – proto-urban or urban and comparable to the first cities in the Near East (Videiko 1996) or ‘large villages’ that fell far short of urban status (Kruts 2003; KorvinPiotrovskiy 2003). This debate has been recently critically reviewed (Gaydarska 2015). However, this debate remained at a classificatory level – what to call the megasites – rather than a model of how they developed or functioned. There was general acceptance of the diffusionist process of site populations moving from one megasite to another in succession, itself rooted in Ryzhov’s generally accepted, complex typology of Trypillia painted pottery (Ryzhov 1990, 1999, 2012, 2012a). The ‘second methodological revolution’ of the late 2000s and the 2010s (Chapman et al. 2014b) came on the heels of a decadal reduction in excavation of megasites, following the funding cuts suffered by archaeology after the fall of the Soviet Union. The key element in the second revolution was geophysical investigation, which was now capable of producing more accurate plans at a much greater rate (Chapman et al. 2014a, 2014b; Hale et al. 2017; Rassmann et al. 2014). Cesium magnetometry on vehicle-drawn carts was used effectively at Taljanki, Majdanetske and Apolianka (Rassmann et al. 2016), while Archaeological Services (Durham University) produced the only complete plan of a megasite so far, using



History of Megasite Investigations 

 11

pedestrian fluxgate gradiometry (Hale et al. 2017; Chapman et al. 2014b; see Chapter 4.2). Mikhail Videiko’s response to the 2009 magnetometric plan of part of Nebelivka (Fig. 1.8) summed up the change – “it looks like an excavation plan!” (John Chapman, witness statement). The ‘new geophysics’ enabled the recognition of new types of individual features – including larger-than-usual structures termed ‘Assembly Houses’, unburnt houses, pits, perimeter ditches, kilns and perhaps paths – as well as the study of new relationships between individual features, whether as groups of houses (‘Neighbourhoods’), groups of Neighbourhoods ('Quarters') or clusters of pits. But, even more significantly, the complete plan of Nebelivka permitted the detailed analysis of the constituent parts of the overall plan in terms of the divergences from, as much as the accordances with, the overall plan. It is this advance which has enabled a clearer picture of the growth of a megasite which was simply not possible with the older geophysical plans.

Figure 1.8: Geophysical plan of the 2009 season overlain on satellite image of Nebelivka (by M. Nebbia, based on Hale et al. 2010).

12 

 Introduction

In comparison with these methodological advances, there has been little concurrent development of the Trypillia megasite theoretical research agenda (but see Diachenko 2016, 2016a), despite its ultimate aim of the identification of explanations for the origins, maintenance and decline of the largest settlements known in 4th millennium BC Europe. It has long been maintained, if recently published (Chapman & Gaydarska 2018), that the gains of the 2nd methodological revolution would be forfeited unless comparable theoretical developments were made in the next decade. Such developments soon became central to the AHRC project.

John Chapman & Bisserka Gaydarska 1.3 Project Biography Perhaps the most successful, and certainly the most entertaining, account of the development of an archaeological project was Andrew Fleming’s (1988, 2008) publication on his ‘Dartmoor Reaves’ project, South-West England. Fleming gave a sensitive treatment of the moor which he investigated while integrating moorland places into the story of daily project routines and the ‘academic’ results. While we cannot emulate Andrew’s success, which won him the Book of the Year award in 1990, instead, in this short section, we shall evaluate the intellectual journey5 which we took while leading the Project, from its inception in 2003 to the present. Our good friend Dan Monah – a great Moldavian prehistorian who died in 2013 – managed to wangle Bisserka and John an invitation to a 2003 conference about the Trypillia megasites to be held in the home village of the largest – Taljanki. The best way to recover from an epic 18-hour trip from Iaşi to the conference, shared by three Romanian and one Moldovan colleague6, was to join the conference host Alexei Korvin-Piotrovskiy in a vast feast for all the delegates. One Ukrainian colleague at the conference stood out for his vigour, knowledge and English skill – Dr. Mykhailo Videiko. During our conference discussions, we decided to try to put together a project focussing on the Ukrainian megasites. As often happens, one thing did not lead to another and so it was not until 2007 that we were stimulated by the visit of Professor Roland Fletcher as a Visiting Fellow of the Durham University Institute of Advanced Study to renew our plans to work on the megasites. Reminding us that the Trypillia megasites were the only exceptions in the world to his global 100ha limit to agrarian settlements (Fletcher 1995), Roland urged us to develop the Trypillia contacts and to create a project to investigate these extraordinary sites. These sites, he believed, would be the earliest examples of a new

5  To which Andrew Fleming may well reply: “Pretentious - toi??!!” 6  The details of this trip would fill a chapter which perhaps deserves a separate publication.



Project Biography 

 13

class of urban site – the low-density urban site – which he was researching during his Durham Fellowship (Fletcher 2009). Renewed contact with Mikhailo Videiko led to a visit to Ukraine in summer 2008 to discuss a project, examine museum collections and visit a number of Trypillia sites to identify a potential megasite or perhaps two (!!!) for intensive, multi-disciplinary investigation. After discussions in Videiko’s home Institute in Kyiv, we decided that the site of Nebelivka, in Kirovograd Oblast, would make an excellent choice. British Academy Small Research Grant funding was obtained for a trial season in summer 2009, in which we tested our abilities to work together (viz., make necessary compromises) and tried out approaches to field archaeology in the Ukrainian foreststeppe. This included several firsts – intensive, systematic fieldwalking of Sovietscale fields, gridded collection within a megasite, dry-sieving and bucket flotation of a Trypillia house excavation, modern geophysical prospection – as well as intensive post-excavation finds processing and sediment coring at newly-discovered wetland sites in the region. The results of the 2009 season were promising enough to submit an application to the AHRC for a four-year project, which failed in 2010 but was funded in 2011, with a start-up date of March 2012. At this stage, we hired a Post-Doctoral Research Assistant (Dr. Bisserka Gaydarska – the only person on the Durham team with an excellent grasp of Russian) and a Project Ph.D. student (Mr. Marco Nebbia – a remote sensing and GIS specialist). The first major Project field season was planned for July – August 2012, during which we gave ourselves the immense challenge of excavating the whole of the largest known structure in the Trypillia world – the so-called Megastructure (56 x 20m, including a built-up area of 36 x 20m) – in one 8-week season7. This decision implied a change of direction, since nowhere in the Project application had we even mentioned ‘finds’. The first planned strategy for recovering samples for AMS dating had involved daub coring, with very limited finds recovery, rather than test excavation, which produced masses! Further major fieldwork seasons in the summer (2013 and 2014) alternated with winter and spring post-excavation recording seasons of the huge ceramic assemblages that we recovered from House A9 (2009), the Mega-structure (2012) and the over 80 test pits which contained often large ceramic assemblages (2013–4). The happenstance of taking a mature Durham student with building experience – Stuart Johnston – on the 2013 season led to the idea that the Project could begin an experimental programme under Stuart’s leadership, in which we would build (2014), burn (2015) and excavate (2017) the burnt remains of, two smaller-scale ‘Trypillia’ houses – both of floor plan 4 × 3m, one 1-storey and the other 2-storey (Fig. 1.9). This led to an international conference visit to Nebelivka village in

7  The Ukrainian side was convinced that we had to complete the excavation in one season, for fear of extensive looting of the Mega-structure in the months after September 2012. Supplementary funding for the Mega-structure excavation was kindly provided by National Geographic Society.

14 

 Introduction

2015 for the burning of the 2-storey house8, as well as a one-week season in 2017 to excavate part of the experimental burnt house remains.

Figure 1.9: Two experimental ‘Trypillia’ houses in the process of construction (by S. Johnston).

As time went on, our views on the interpretation of our common field results9 increasingly diverged from those of our Ukrainian colleagues, making it harder to write joint articles and book chapters and, ultimately, explaining why the current monograph contains regrettably few contributions from the Ukrainian specialists who worked on the Project. Every team member brings some previous intellectual baggage – positive and negative – to a new Project. The Ukrainian side brought an unrivalled knowledge of Trypillia, accumulated over many years of research experience but which inevitably introduced an element of prior certainty about what the Project would achieve. They also brought a traditional way of excavating and recording Trypillia burnt houses,

8  For reasons of village politics, it was not possible to burn both houses for a comparative experiment. 9  Perhaps the most startling example was published in the Journal of Neolithic Archaeology, when the two sides’ conflicting interpretations of the Mega-structure were published side by side (Chapman et al. 2014.).



Project Biography 

 15

which had originated with the first excavator, Khvoika, and has been passed down to each successive generation of prehistorians (now the fourth!). By comparison, the Durham team was less well-grounded in knowledge about Trypillia, which gave them a more open field in which to develop ideas and alternatives to the dominant narrative. This meant the risk that some of the alternative interpretations were off the rails – but perhaps some were not ... The Durham team did, however, bring a wider knowledge about the Balkan Neolithic and Chalcolithic, as well as a deeper theoretical perspective grounded in the last three decades of British theoretical debate – a debate which has largely bypassed Eastern Europe. Lastly, the Durham team introduced specialist skills and methods which were often new to Ukrainian prehistory and could generate unexpected results. In addition, an important element in the Project’s ongoing research was the development from 2013 onwards of an Ukrainian-German project at the nearby megasites of Taljanki (20km from Nebelivka) and Majdanetske (23km from Nebelivka). This Project in turn brought its own contributions to megasite research – in particular, the vehicle-based geophysical data capture which covered so much ground at Majdanetske, Taljanki, Dobrovodi and Apolianka. Many of the preliminary results of this parallel project have been published in an EAA Monograph and subsequent book (Müller et al. 2016b, 2017). It goes without saying that the research goals of the Ukrainian-German project and their insights into Trypillia megasites have proved to be a major stimulus to the evolving research interests of this Project. The Project began with a title which set the overall agenda in terms of the question ‘Were Trypillia megasites urban in nature?’ Although many facets of our understanding have changed in the last decade, that question still resonates, albeit in a very different way from in our earliest formulation. We also naturally accepted the starting-point of our Ukrainian colleagues – that megasites were permanently occupied by a very large population, which meant that we had to explain the management of resources such as food, salt and lithics10 and the avoidance of environmental degradation. We also accepted Linda Ellis (1984) claims for a three-level settlement hierarchy in the Uman region – a region which included Nebelivka (!) – and looked favourably on the claims for secondary products usage at Trypillia sites. To the extent that we were initially paid-up members of the Ukrainian model for megasites, it was relatively easy to frame new data so as to conform with the current thinking. As late as the Arizona workshop of 201411, we were invoking Ellis’ settlement model, ploughing and large-scale, long-distance salt exchange to account for megasite growth and survival. But four

10  Chapman remembers vividly the conversation with Videiko in which the latter claimed that 2 tons of exotic flint were being exchanged into the Majdanetske megasite per annum! 11  The “Cities or Big Villages? New Approaches to 'Anomalous Great Sites' workshop was organised by Roland Fletcher and Kirrily White and involved four days of discussions of large, mostly lowdensity, sites in the idyllic surroundings of the Amerind Foundation.

16 

 Introduction

anomalies led us to operationalise the archaeological version of what Hemingway termed ‘a built-in shock-proof shit-detector’ – the most important element in any Project’s armoury. The first was the discussion of what became known as 'Anomalous Great Sites' – massive nucleated sites, usually low-density in character, which had a patchy global distribution (e.g., Angkor, Co Loa, Kelheim and Cahokia). We started to engage in these discussions at the SAA 2013 meetings in Hawaii and continued in an intensive one-week workshop at the Amerind Foundation’s Dragoon Centre in Arizona. Here, Roland Fletcher had assembled a global team of archaeologists who were working to understand massive low-density sites (urban or not) in all their heterogeneity. During this workshop, we learnt not only that low-density urbanism (LDU) was a diverse global phenomenon of uncertain origins but also that Trypillia megasites were its earliest manifestation in the world. This helped us to counter the ‘large village’ interpretation of megasites and led to Gaydarska’s challenges to conventional urban thinking (2016, 2017, 2019a). The second starting-point was the new geophysical investigations at megasites such as Nebelivka, Taljanki, Majdanetske and Dobrovodi, as much as the completion of the Nebelivka plan in October 2013. The high-resolution detail of the Durham University Archaeological Services plan enabled us to go beyond the overall planning principles of megasites defined in the 1970s and 1980s and confront the variability of the layouts and local differences at each scalar level (house size, character of Neighbourhoods, nature of Quarters) which pointed to a bottom-up element in what had previously been conceived of as a top-down settlement plan12. But this insight could still be consonant with Kruts’ idea of 40 separate settlements coalescing into a single megasite (see below, p. 39). The third element in our re-thinking came from the results of the Nebelivka Core P1 pollen analysis produced by Bruce Albert in 2015. What we expected from the multiproxy analyses was signs of a massive human impact from a very large population of tens of thousands of Nebelivkans – after all, the coring site was only 250m downwind of the megasite. However, to everyone’s surprise, there was no such massive human impact – in fact, most of the minor impact peaks were dated to before the settlement of the megasite. There is no doubt that the dating of pollen diagram has its problems – strong criticism on this ground was the main reason why the paper submitted to The Holocene in 2016 was rejected13. Nonetheless, it is a fact that there was no major human impact peak in the entire diagram, so wherever the megasite occupation is placed in relation to the pollen core, a modest human impact was all that was caused. We clearly needed to find an explanation for this absence.

12  There is surely a parallel here between post-WWII Soviet planned economies and post-Soviet bottom-up economic growth, with its booms and busts. 13  A revised version has been accepted for Vegetation History and Archaeobotany (Albert et al. 2020).



Project Biography 

 17

The fourth insight came from the Project’s only major addition to the initial nine objectives – the development of an experimental programme of house-building, house-burning and the excavation of the experimental burnt house remains. Stuart Johnston organised the building programme of two 4 × 3m ‘Trypillia houses – one 1-storey and the other 2-storey – in the 2014 season, while the 2-storey house was burnt down as part of the Kirovograd-Nebelivka international conference in May 2015. Despite initial opposition from a financially prudent co-director, Johnston persuaded the Project to purchase 30m3 of timber to fill the house before firing. Together with the good weather, this fuel was one of the principal factors in achieving what is believed to be the first creation of a Trypillia look-alike ploshchadka and the production of genuine vitrified daub (A. Diachenko, pers. comm.; Burdo 2011). But the implication for house-burning was far-reaching – a Trypillia household would have needed much more firewood to burn their house than to build it – perhaps as much as 10 times more. Given that over a thousand houses had been burnt at Nebelivka, this made the expectation of a major human impact even greater. This finding was linked to all of the other data which had hitherto been broadly supportive of the current ‘maximalist’ hypothesis to define a ‘tipping-point’ in our thinking about megasites. The implications of the tipping-point were developed in 2015 and 2016, too late for the Project-organised EAA Session on “Re-assessing urbanism in pre-Roman Europe” in Istanbul in September 2014. This session had two published outcomes, which neatly reflected the diverging interests of the Ukrainian, German and Durham teams. While Johannes Müller co-ordinated a data-rich EAA monograph on Trypillia megasites (Müller et al. 2016b), Bisserka Gaydarska (2017) guest-edited a special issue on Urbanism in the Journal of World Prehistory. Here, a critique of current approaches to urbanism framed the implications of the tippingpoint for Trypillia megasites. But if it was now logical to reject the maximalist position, what would take its place? The Arizona workshop was the crucible in which alternative ‘minimalist’ or ‘middle-way’ explanations were forged14 in the spirit of Gaydarska’s ‘relational’ approach to urbanism. Another important strand in our thinking developed from our meetings with David Wengrow, whose novel insights into the earliest stages of Near Eastern urbanism (Wengrow 2015) helped us to make appropriate comparative comments for megasites. These interactions led to theorising three different models for a smaller megasite, two using seasonal dwelling and the third a smaller but permanent settlement approach. The operationalisation of these models began in 2016, with presentations at the Southampton TAG (December 2016) and the Vancouver SAA meetings (April 2017). Their full evaluation continues in recent publications and is summarised in this monograph.

14  The ‘pilgrimage’ model was actively discussed at Arizona.

18 

 Introduction

Meanwhile, what has become of urbanism and, especially, low-density urbanism? To put it baldly, (how) can a ‘pilgrimage centre’ or an assembly site be called a ‘city’? Can even a slimmed-down, permanent ‘middle-way’ model of Nebelivka be called a ‘city’? Perhaps this last is the only model where the term low-density urban is appropriate? Escaping from the domain of pinheads and angels, we can only emphasise that it depends what you mean by a ‘city’. The remainder of this monograph is a complex attempt to answer these questions.

Bisserka Gaydarska & John Chapman

2 Theory and Practice for Trypillia Megasites In this chapter, we outline recent trends in the global debate on urbanism and seek to locate Trypillia megasites within that debate. We pinpoint a tipping point in our understanding of megasites, leading to a definitive break from the Maximalist position of very large, permanent, all-year-round occupations to alternative, shorterterm or seasonal positions based upon three models of Nebelivka settlement – the Distributed Governance Model, the Assembly Model and the Pilgrimage Model. Using Ben Anderson’s concept of ‘imagined communities’, we try to imagine the possibility of creating a megasite for the first time, leading to the development of a theoretical framework for such a creation. We develop the notion of the Trypillia Big Other, relating it to Bourdieu’s habitus. We also introduce the methodologies specific to each of the eight Project research questions, which sought to deliver a complete geophysical plan of Nebelivka, a welldated internal sequence for the megasite, the local and regional settlement contexts for the megasite, an assessment of the human impact of a megasite on its landscape, the experimental construction, burning and excavation of a ‘Neolithic’ house, an interpretative model of the growth and decline of Nebelivka, its architecture and artifacts, and the placing of Trypillia megasites in the context of global urbanism.

20 

 Theory and Practice for Trypillia Megasites

Bisserka Gaydarska 2.1 The Theoretical Debate on Urbanism The AHRC-funded Project “Early urbanism in prehistoric Europe?: the case of the Trypillia megasites” concerns the growth of early urbanism in prehistoric Europe. The fundamental approach to stages of human development was enhanced in the 1950s by Childe, who argued in a diffusionist manner for the priority of Near Eastern complexity over that of Europe (Childe 1950). The current view remains that the earliest states in Europe – the Minoans – were ‘first-generation secondary states’ (Parkinson & Galaty 2007, 118), dating to after 2000 BC. This view has consistently ignored the appearance of Trypillia ‘megasites’ in 4th-millennium Eastern Europe, the largest of which were bigger and earlier than Late Chalcolithic Phase 5 at Uruk (Müller & Pollock 2016, 282). In ‘The limits of settlement growth’, Fletcher (1995) identified the Trypillia megasites as the sole exception to his global model of constraints on agricultural settlement expansion. It is clear that the megasites have been neglected in the narratives of urban change and that a targetted investigation of one megasite and its hinterland would greatly aid our understanding of their settlement complexity. The primary aim of the project was thus a re-evaluation of Trypillia social and settlement developments through the inter-disciplinary study of a single megasite in its local, regional, and Eurasian settlement contexts. A second, theoretical aim of the project was the development of interpretative archaeologies dealing with urban developments, since, with a few exceptions (e.g., Christophersen 2015; Smith, M.L. 2003), interpretative archaeologies since 1990 have largely ignored one of the ‘Big Questions’ of social evolution – urban origins. This ‘Big Question’ of social evolution has been dominated by ‘top-down’ hierarchical approaches rather than a ‘bottom-up’ approach building on local household nodes, Neighbourhoods and networks. This project seeks to redress that imbalance by combining recent approaches to landscape and community identities, scientific methodologies and social modelling. In this chapter, which considers the theory and the practice of Trypillia megasite archaeology, we begin by cutting a pathway through the jungle of current urban theories, before discussing the Ukrainian theorization of the megasites and examining the theoretical approaches which we use to understand this phenomenon. A good place to start is with an explanation of the term ‘megasite’. The term ‘megasite’ has been taken up in modern business parlance as a land development by private developers, universities, or governments to promote business clusters. These organizations develop the land so that it is “shovel ready” for big business, by improving the infrastructure (roads, utilities, and landscape). The first university megasite was developed by Stanford University as an industrial park in the 1950s and evolved into Silicon Valley (https://stanfordresearchpark.com/about). Particularly large sites appeared and then disappeared at different times and places throughout the general trajectory of increasing settlement size noted from the



The Theoretical Debate on Urbanism 

 21

Palaeolithic to Megalopolis, usually without leaving any obvious settlement legacy. But these so-called ‘megasites’ constituted exceptions to the normal settlement size of any given period/region, often being larger than 10 times the usual settlement size. In this sense, the megasites were relationally significant in their social contexts. An early use of the term in prehistoric research was Gary Rollefson’s (1989) use of ‘megasite’ to describe ‘Ain Ghazal and other PPNB15 sites in the Southern Levant that were larger than 10ha, in comparison with ‘usual’ settlements of 1–2ha. Similar unusually large sites are now known in other areas of the Near East, such as Çatalhöyük (13ha) and Abu Hureyra (16ha). This shows that a relational appreciation of size coincided with the early use of this term. The relational approach was essential in the still earlier case of Levantine megasites in the Epipalaeolithic, with megasites such as Khareneh IV and Wadi Jilat 6 ca. 2ha in size in comparison with ‘usual’ sites of 0.2ha (Martin et al. 2010). Unusually large sites in the Neolithic and Copper Age of Western Europe have been called ‘megasites’ in recent years, such as the 3rd millennium BC Iberian Copper Age enclosures of Valencina de la Concepción (García Sanjuán et al. 2017), Perdigões and Marroquíes Bajos (Milesi García 2018). All of these sites present similar issues of population size and density, seasonality and site function to those featuring in our discussions of the Trypillia megasites. The same is true of urban megasites.

2.1.1 The Global Debate on Urbanism “The concept of ‘city’ is notoriously hard to define”. This is the opening statement of Childe’s (1950, p. 3) seminal article ‘The urban revolution’. Almost 70 years later, this task has not become any easier. The demand for global narratives and generalizations of human history (e.g., Service 1962; Trigger 2003) that is constantly fuelling crosscultural comparisons has cemented an essentialist and evolutionary perspectives of ‘the city’. The current sea of available archaeological literature on cities utilizes three major definitions of ‘city’: ‘archaeological’ – a re-statement or a re-worked understanding of Childe’s ten criteria; ‘sociological’ – variations of Wirth’s (1938) core definition of a city as a ‘relatively large, dense and permanent settlement of socially heterogeneous individuals’; and ‘functional’ – urban settlements are ‘centres whose activities and institutions, whether economic, administrative or religious, affect a wider hinterland’ (Smith, M. E. 2007, p. 4). Despite excellent scholarship to the contrary, there are several stubborn associations with urban development that are proving very hard to break. In addition to the evolutionary framework, these are the pairing of states and cities, urban

15  ‘PPNB’ stands for the Pre-Pottery Neolithic B period, in which domesticated plants and animals are used but no pottery was made (Simmons 2007).

22 

 Theory and Practice for Trypillia Megasites

and civilizations, urban and inequality, the urban-rural dichotomy and urban and social complexity. To these we can add the ‘size’ of both the settled area and/or number of occupants, that is particularly exacerbated in the uncritical equation of ‘sherds=people=population number=site size’ in Big Data sets (for an exception, see Whitelaw 2013). Recent decades saw a surge of diverse themes and regions in urban debates, within which three dominant themes and three dominant regions have, willingly or not, affected the tone of the discourse. The themes are origins, functions (political, economic and ritual) and, most recently, scaling. The dominant regions are the Near East/Mesopotamia/Egypt, North America and Europe, the last-named being almost exclusively represented by the Greco-Roman evidence. More often than not, other regions and periods are trying to ‘fit’ the dominant themes or the generalized narratives perpetuated in the dominant regions to their own particular cases. However, the greater the diversity of urban cases across the globe, the more difficult it is to fit them all into any definition of urbanism, as a number of dissenting voices have been claiming (e.g., Kusimba et al. 2006; McIntosh, S. 1995; Wilkinson et al. 2014). This meant that our understanding of urbanism had to change. The identification of a new and important category of ‘city’ was prompted by the Medieval Khmer capital at Angkor, that showed the importance of ‘low-density dwelling’ in global perspectives of urbanism (Coe 2003; Fletcher & Evans 2012). The low density of structures beyond a small high-density core covered a huge area of 1,000km2, with the extensive open space in the Angkor landscape making it almost impossible to define its boundary (Pottier 1999; Fletcher & Pottier 2002; Evans et al. 2007). The low-density landscape of Angkor became the stimulus for a current research agenda for identifying, mapping and ultimately explaining the rise and fall of such globally distributed sites. Low-density urbanism was initially defined in contradistinction to the classic highly nucleated urban capitals discussed by Childe, such as Uruk, Baghdad, Rome and Paris (Fletcher 2009). Low-density cities can now be found across all continents of the world except Antarctica, taking the form of sites whose obvious differences from high-density cities have led to names such as ‘Big Weird Sites’ or ‘Anomalous Great Sites’ (see above, pp. 15–16). The proposal here is that the term ‘megasite’ should not be restricted to certain classes of large sites, whether Epipalaeolithic sites in the Levant, Iberian enclosed sites or Trypillia settlements, but should instead be generalised to become the standard term for these anomalous settlements. The reasons for this proposal are fourfold, based upon widespread agreement that: (a) after a decade’s fruitless search for a common denominator, a new term is urgently needed for these places; (b) the term ‘urban’ no longer captures the essence of these big sites (5, 2013); (c) these settlements were unusually large for their local and regional cultural context; and (d) a relatively new term such as ‘megasite’, with far less intellectual baggage than ‘urban’, helps to unite these disparate places in a meaningful manner which allows comparative study. The practice of low-density urban dwelling is of particular



The Theoretical Debate on Urbanism 

 23

relevance to the Trypillia megasites, since they appear to be the earliest such cases in the world. But what was the social context of such megasites? Another approach to the impasse in urban studies has lain dormant, widely overlooked since 1938. Although most of Wirth’s sociological definition (see above, p. 21) is widely accepted, few have taken up his injunction that any definition of a city should be contextually informed (Wirth 1938, p. 6). This insight has suffered from its incompatibility with the favoured cross-cultural comparative approach; here, it is vindicated in the shape of a relational approach looking for meaningful markers within a region rather than for fixed universalities. In a relational framework, categories emerge, develop and integrate only in relation to each other rather than absolutely. The intensity of social practices at urban sites, some of which were predicated by their size, makes these sites very different from (in relation to) other smaller sites, and more importantly were experienced and perceived differently by their inhabitants, visitors and neighbours. At an early stage in the project, it became clear that there were several ways in which the Trypillia evidence did not fit the dominant urban accounts. Therefore, we decided to build on the critiques of, and alternatives to, the dominant model as discussed by authors such as Wirth (1938), A. Smith (2003), Cowgill (2004), Campbell (2009), Fletcher (2009), Wengrow (2015), Ur (2014), Jennings and Earle (2016), Hahn (2016) and Gaydarska (2016; 2017). In particular, we noted the relational difference between Tryillia megasites and smaller Trypillia sites. This difference was underpinned by the high intensity of social practices and network centrality, perhaps qualifying the megasites as different kinds of sites in a Trypillia context. This hypothesis invites consideration of Trypillia megasites as ‘cities’ in a global context.

2.1.2 Trypillia Megasites – the Theoretical Debate Archaeological theory behind the Iron Curtain was very different from its counterpart in the West. While this is not the place to praise, justify or criticise it, it is very important to underline a few of its characteristics that continue to have a devastating legacy. First and foremost is the now tacit, but nevertheless no less powerful, division between theory and practice. Cohorts of field archaeologists were compiling data, with their efforts rewarded as masterpieces of empirical studies. Explanations were provided by a few powerful theorists – explanations that in turn were firmly embedded back into the cohorts of archaeologists (Klejn 1982). There were no paradigm shifts, just shifts of powerful people. Secondly, the political regimes not only favoured but imposed social evolutionary thinking and revolutions as the prime drivers of social change. Even if archaeologists wanted to take a different route, the pool of theoretical resources was very limited. It is in this context that current thinking on Trypillia megasites should be set.

24 

 Theory and Practice for Trypillia Megasites

At the start of the project, the Ukrainian debate about the nature of the Trypillia megasites fitted well with the essentialist, evolutionary, ‘archaeological’ definition of urban. A formal definition of what did or did not constitute a megasite as a ‘(proto) city’ was never put forward. The small group of advocates of the proto-city hypothesis considered megasites as economical, political, military, administrative and religious centres of local groups. One member of this group, Mykhailo Videiko (initially with Shmagli; for full list see Videiko 2013), presented evidence for monumentality and welldeveloped crafts (e.g., pottery consumption and hence production at Majdanetske) as comparable to that of Uruk: Videiko 2013)16. Videiko argued that (proto)cities did not appear overnight and involved a developmental process, which started in the Southern Bug-Dnieper interfluve but was not restricted to that area. Indeed, sites larger than 100ha first appeared in the S. Bug-Dnieper interfluve but were soon followed by others in other areas (Videiko 2013, Fig. 86), with their distribution again contracting to the initial area in Phases CI–CII. Nonetheless, the largest settlements developed in this interfluve, so some special status was assigned to the people living there. For Nebbia (Chapter 3), there are no evident environmental reasons for why Trypillian groups should have settled and developed so many megasites in that area – however, the ‘mega-cluster’ apparently sat at a conjunction of two major hydrographic basins in Ukraine. Another visible pattern is that megasites appeared at the Southern edge of the currently known Trypillia area of influence so as to mark a sort of “border”. This would complement the notion that Trypillians gathering at megasites may have been interacting with bordering steppe communities. Critics of the ‘proto-city’ hypothesis emphasised that, in contradistinction to the Trypillia sites, true early cities were administrative, economic, cultural and religious centres of a rural hinterland17. Since there were no small sites around the megasites, all the population was believed to have been living in the megasites and thus, with no rural hinterland, there were by implication no cities. Childean criteria of urbanism, such as populations of over 5,000, writing and monumental architecture in the form of palaces and temples, were claimed to be absent at the large Trypillian settlements (Korvin-Piotrovskiy 2003, p. 5). Arguments for monumentality in the shape of twostorey buildings forming the so-called ‘living-walls’ and large geophysical anomalies were considered unsubstantiated. Despite the high number of inhabitants that points to urban population levels, the density of population was very low in comparison to smaller sites and there was no evidence for differentiation such as cult features or craft quarters. Another line of criticism took a diachronic and cross-cultural approach (Monah 2003). Quoting mainly Romanian examples, Monah argued that

16  It is clear that Videiko changed his view that the megasites were not centres of crafts and trade (Videiko 2007, p. 272). 17  It is important to point out that this is a conceptual rather than literal translation of the common understanding of proto-cities in the Ukrainian and Russian literature.



The Theoretical Debate on Urbanism 

 25

fortification and population size cannot be considered as criteria for urban life. Equally, some elements pertinent to towns and cities, like streets, markets, spaces and buildings for common use, were found also in rural Chalcolithic communities. Thus, for Monah, the key issue was the ratio between artisans and other units of population. Although he admits that Cucuteni (and by extension Trypillia) painted vessels were real masterpieces, this did not constitute evidence, even for Monah, for the kind of specialized production characteristic of urban economies. Indeed, Wengrow identified the replacement of elaborately decorated pottery by much more standardised ceramics at the transition to the first cities of the Late Uruk period in the Near East (Wengrow 2001). On a general level, the prevailing views of Cucuteni-Trypillia specialists about the nature of megasites were built on shaky theoretical foundations. The proponents of the ‘proto’-urban hypothesis have conflated Childe’s two revolutions – the Neolithic and the Urban – in an evolutionary, stage-based framework (Service 1962). Although the agrarian nature of Trypillian cities is certainly a research avenue to be pursued, it needs to be properly theorized, instead of relying on a implicit, narrow understanding of both revolutions. The absence of a settlement legacy of the megasites jeopardizes the evolutionary account towards urbanism, since there were remarkably few postTrypillia settlements known in this region until the Late Bronze Age. The opponents of the ‘proto’-urban hypothesis adopted a similar evolutionary stance in seeking to reconcile an essentially ‘Neolithic package’ with agglomerations of thousands of people – hence the term ‘settlement-giants’ (Korvin-Piotrovskiy 2003). Estimates for scaled-up operations were provided (Kruts 1989) but theorization was preoccupied with the origins of the megasite through internal or external conflict, migration and population pressure (Diachenko & Menotti 2017). The obvious contradiction of how small-scale Neolithic subsistence practices could possibly sustain such massive aggregations remained unaddressed. If subsistence was dependent on thousands of smallholdings (the basic Neolithic unit: Childe 1958; Chapman 2009), what was the underlying social structure and was it a Neolithic structure? A secondary issue concerned the environmental depletions caused by such massive agglomerations, which forced people to move on to another megasite. It should be emphasised that the reasons for the development of megasites were identical for those in favour of and contra Trypillia megasites as ‘proto-cities’ – internal or external conflict, migration and population pressure (Videiko 2007, p. 274; Diachenko & Menotti 2017). Echoing Chernysh (1977) and Gimbutas (1977), Kruts (1989) argues that the principal threat to Trypillia communities came from the Sredni Stog groups in the steppe zone to the South and East. However, even 10–20ha Trypillia sites would have been large enough to deter small groups of armed Sredni Stog raiders – removing the military need for much larger agglomerations. Dergachev (2002) uses his findings of a higher ratio of fortified to non-fortified sites, and higher numbers of arrowheads per site to suggests that Phase BI was a ‘society ... literally under siege’ (2002, p. 103), in a ‘state of war owing to outside

26 

 Theory and Practice for Trypillia Megasites

threat’ from the steppe (2002, p. 106). While this view can be used to support the appearance of early megasites, it offers no support for the military explanation of megasites in the more peaceful Phases BII and CI. By contrast, Videiko (2007, pp. 274–5) proposed an internal social conflict for the origins of megasites, describing Trypillia chiefdoms as ‘in a state of perpetual internecine war’ because of the expansive nature of Trypillia agriculture, with soil exhaustion causing megasites to move on to capture more arable land every 70 years. However, Videiko ignores the large unsettled areas in the Southern BugDnieper Interfluve, while the evidence for perpetual internecine war is limited to a single archery attack on the small site of Drutsi I, in Moldova (Ryndina & Engovatova 1990). More compelling evidence derives from the Verteba Cave, where 11 out of 25 buried crania have clear indications of trauma (Madden et al. 2018). Stratigraphic observations suggest that there were two episodes of interpersonal violence and subsequent skull deposition. However, none of the 11 examples has been directly AMS-dated and the long duration of neighbouring dated contexts (e.g., Stratum B – 3805–3707 cal BC and 3946–3774 cal BC, and above Stratum A, range between 3950 and 2578 cal BC) makes any suggestions for the timing of the two violent episodes highly speculative. In addition, the Verteba Cave is far from any megasite, thus jeopardizing any potential link between the two phenomena. It is clear that migrations can show how people moved across the landscape but do not provide a reason for any particular settlement form – say, megasites rather than village clusters. This leaves internally-driven or externally-imposed warfare as the principal traditional explanation for the rise of megasites – not the outcome predicted by Gimbutas’ (1977) peaceful matriarchal Cucuteni-Trypillia society! It is also worth noting that many of the problems with these traditional explanations are tied to basic maximalist assumptions about the megasites themselves. Paradoxically, the question about the collapse of the megasites is very often conflated with their origin. The two traditional responses to this question have focused on invasion and defence (Passek 1949a; Rassamakin 2004; Manzura 2005; Anthony 2007, pp. 279–282) and migrations caused by the unsustainability of such massive populations over a long period. Since the AMS-based gap of several centuries between the demise of the megasites and the earliest barrows makes an external threat untenable, we are left with Trypillia internal conflicts (Chapman et al. 2019) – an explanation subject to the same criticisms as for the origins of megasites. The second postulated cause of megasite collapse – the lack of a sustainable resource base – has been discussed more frequently in the last decade (Kruts 2012; Diachenko 2016a; Ohlrau et al. 2016). While Kruts et al. (2001) maintain that the combination of the high demand for construction timber and the daily demands for firewood for heating and cooking at Taljanki caused a resource crisis after 50 years, Videiko (2007, p. 276) avers that the disappearance of the megasites reflected the crisis of an extensive agricultural economy. In comparison with Shukurov et al.’s (2015) thorough and multi-faceted modelling of the Trypillia agro-pastoral system,



The Theoretical Debate on Urbanism 

 27

other studies claiming to be investigations of Trypillia carrying capacity (Diachenko 2016a; Ohlrau et al. 2016) are based upon poorly quantified palaeo-environmental data (cf. Section 4.1) and a weak grasp of the complex interactions between a farming population and the carrying capacity of the soil on which those people depended (see Chapman 1988 and critique in Gaydarska & Chapman 2016, pp. 183–4). In any case, environmental over-exploitation has been based upon maximalist population estimates, which we shall show to be unrealistic (see below, pp. 32–34). This is not to say that environmental over-exploitation did not occur – only that it has not yet been clearly documented. We should not throw the environmental baby out with the demographic bathwater. Currently, the only two well-documented local palaeo-environmental records derived from Nebelivka and Majdanetske. The post-megasite sediments in Nebelivka Core P1 fall in Zone 8 (Albert et al. 2019). Here, forest composition significantly changed to a gradual dominance of Quercus, perhaps related to cooler climatic conditions (Harper 2016). Cultivation indicators in the post-megasite period were not so obvious as in the megasite phase but there is spore evidence for the continued presence of grazing animals in the catchment. The Zone 8 data does little to support the interpretation of an environmental collapse, nor even a gradual decline. The second proxy record, from Majdanetske, concerns the creation of a ‘cultural steppe’ landscape through the megasite’s impact on the mosaic vegetation patterns of the forest steppe (Kirleis & Dreibrodt 2016, p. 177). The main evidence in support of this notion is the change in soils from a fertile forest cambisol before the start of dwelling at Majdanetske to a chernozem developing under steppic conditions (cf. Kruk 1980 for the TRB period in Little Poland). Kirleis & Dreibrodt imply the loss of soil fertility from cambisol to chernozem at Majdanetske, but this idea is contradicted by the high fertility of Ukrainian chernozems (Kubiena 1953). With their dense rootlet system, chernozems may have been harder to cultivate than cambisols (Chapman 1990) but once the practice of hoeing or ploughing in the interval soon after rain was mastered, the fertility of the chernozems was unmatched. The absence of any long-term improvements in Trypillia agriculture (Kruts 1989; Pashkevitch 2005) was perhaps related to the high fertility of the chernozem soils. In summary, there is little or no palaeo-environmental evidence for such degradation as to force megasite abandonment, although this may be found in future. We cannot claim that the same minimalist arguments applying to Nebelivka should automatically apply to all other megasites but one obvious reason for the lack of environmental degradation would have been a much smaller population than envisaged by the maximalists for Taljanki and Majdanetske (see Section 6.2). Two internal social scenarios for the demise of megasites stem from Müller and Rassmann’s (2016) observation that the greater social complexity that comes with higher populations caused greater vulnerabilities that had to be addressed by some form of adaptation. Although these authors never amplify what these vulnerabilities actually meant at Majdanetske, they consistently refer to the problem of scalar stress

28 

 Theory and Practice for Trypillia Megasites

that would have adversely affected massive megasite populations (Müller et al. 2016a, p. 265; Chapman 2017; Hofmann et al. 2019). Müller and Rassmann (2016, p. 5) suggest that the lack of writing in Trypillia society limited the range of adaptations they could develop in the face of internal conflict or tension. The first internal scenario is thus an expression of the recurrent political and/or administrative failure to deal with huge population size and complex lifeways. However, this explanation is incompatible with the cyclic re-appearance of megasites over a period of 800 years. The second scenario concerns Diachenko’s (2016a) proposal that the end of the megasites involved a transformation in the Late Trypillia political economy from centralized chiefdoms to dispersed chiefdoms based upon a shift from centralized agricultural settlements to semi-autonomous households, decentralized staple production and network-based exchange systems. Thus the need for megasites disappeared because it was impossible to control large territories with such dispersed networks. But Diachenko allows that social hierarchies were maintained through practices causing material inequality in the latest Trypillia (CII) phase. Despite several other problematic assumptions (for critique, see Gaydarska & Chapman 2016), this model has the potential of explaining the end of the megasites with social arguments as well as questionable traditional accounts involving migrations and climatic fluctuations. It is not, however, clear how the model explains the decline of earlier cycles of megasites. The view of Trypillia megasites as overgrown villages or settlement-giants remains the preferred explanation till this very day. Its inherent contradictions are exacerbated by the results of modern high-precision geophysics, which, in some cases, have increased the number of houses at megasites by up to 48% (e.g., for Majdanetske, compare 1,575 (Kruts 1989) with 3,000 (Ohlrau et al. 2016)). The resultant eyebrow-raising population number of 46,000 quoted for that site (Rassmann et al. 2014), later reduced to a maximum of 31,700 (Ohlrau et al. 2016) and now to 10,000 (Müller et al. 2018), is still believed to have lived in rural settlements that at best were qualified as ‘social experiments’ aided by a developed transport system with sledges allowing larger agglomerations (Müller & Pollock 2016, p. 286). A very useful point of comparison here is de Vries’ (1984, Table 3.4) conclusion that, in AD 15th century Europe, there were only 18 cities with a population exceeding 40,000. What was the social foundation of such megasites?

2.1.3 The Social Formations of the Trypillia Megasites The traditional view of the social structure of the inhabitants of megasites is a reflection of their size, permanence and significance as central places in the forest steppe landscape (Masson 1990; Kruts 2012; Videiko 2013). These views on society are ultimately founded on the Sahlins – Service evolutionist paradigm (Sahlins 1958; Service 1962) and focus on the level of social complexity represented at the megasites



The Theoretical Debate on Urbanism 

 29

– whether simple chiefdoms, complex chiefdoms or proto-urban, stratified elites. Masson (1990) considers the megasites as ‘early complex societies’ that have taken a non-urban route of development with its political equivalent being a chiefdom. One of the clearest statements on Trypillia social structure came from Kolesnikov (1993), who divides the group into tribes, communes, descent groups and multifamily households. In his view, the multi-family household consisted of two – four families (10–12 people), who developed neighbourly relations with other households and jointly owned land, animals and tools. The descent group consisted of many households and performed tasks such as house-building, preparation of the land for cultivation, rituals and ceremonies and also conflict-resolution within and between descent groups. The commune consisted of many descent groups, amounting to 350–500 people, and was run by the representatives of each descent group. The main functions of the commune were regulation of land and relationships, organization of rituals and defence. The highest level is the tribe of ca. 4,500 people, who regulated conflicts between communes over land ownership, since the supreme land-owner was the tribe. The tribe was also responsible for regulation of inter-tribal relationships, for example flint supplies from the Volhynian tribe. This summarises a classic socialist interpretation of prehistoric social relations (for variations on a similar theme for the Bulgarian Copper Age, see Raduncheva 2003). Overlapping in part with Kolesnikov but developing the spatial scale more widely, Videiko (2013) proposed a three-tier settlement system with a chiefdom controlling an area of 10–20km in radius and consisting of a ‘capital’ (viz., a megasite) covering an area from 50 to 200ha, dependent towns ranging from 10 to 40ha and villages of 2–7ha. However, the scale of these chiefdoms seems far too small to accommodate the widespread interactions of the Trypillia Big Other (for definition, see below, pp. 37–39), which covered 250,000km2. In a recent article, the German expedition working at Majdanetske published their most comprehensive view of the social organisation of what they continue to regard as massive rural settlements (Müller et al. 2018). Although, for the maximalists, megasites could not have emerged without favourable environmental conditions and technological innovations (p. 254), the authors propose that the creation of the megasites was “primarily a political decision, made by several groups of people, previously not so closely connected, to live together in the same settlement” (p. 253). Five forms of socio-spatial groupings are recognised at Majdanetske (2018, Fig. 11) – the household as the basic unit; the Neighbourhood as a group of houses, the Quarter as a group of 50–150 houses centred on a Mega-structure18 which may have been the focus of political decision-making; a supra-household economic grouping based upon the distribution of kilns (p. 257); the lineage, each of which matched a single concentric circle (p. 258); and the entire settlement for collective decision-

18  The Kiel group uses the term ‘mega-structure’ to mean what we have termed ‘Assembly House’.

30 

 Theory and Practice for Trypillia Megasites

making. Given the absence of archaeological evidence for stable central institutions, there were no hierarchical relations at the megasites but, rather, each individual was bound into the five different social groups whose residential structure created the social organisation of the megasite (p. 258). With the exception of the lineages and the supra-household economic groupings, this social model rests on a number of entities conceptualised for the Nebelivka megasite (household, Neighbourhood and Quarter: Chapman et al. 2014) and repeats the household model proposed for Majdanetske House 51 (Müller et al. 2016, Fig. 4). There are two major problems with the later version of this model – a macro-issue and a micro-issue. The macro-issue concerns the lack of any explanations of how the overlapping socio-spatial entities relate to each other. This is most serious for the Quarters and the supra-household economic groupings, which are almost coterminous in some cases while quite distinct in others; it is also problematic for the relations between lineages and Quarters, whose architectural manifestations crosscut each other. The micro-issue concerns the unfortunate Majdanetske individual whose identities were split into many different parts. There are alternatives to such individual conditions as appear at Majdanetske but they have not been discussed in the 2018 model (see below, pp. 31–32). In short, this over-complicated social model actually operates at the structural level rather than considering real residents, who suffer from the polarities of the structures rather than making day-to-day decisions about them. In this volume, we reiterate our support for the basic social units of households, Neighbourhoods and Quarters (Chapman et al. 2014; for architectural details, see below, Sections 4.2 and 4.3.1) and discuss two forms of descent groups linking different settlements – the clan and the lineage. There has been a long debate on the variable relationships between ‘houses’ and ‘households’ (for a summary, see Souvatzi 2008). In the Trypillia case, we make the distinction between small structures used for storage or production but not for residence (living, sleeping), houses above a certain size (a general threshold of 20m2) which were large enough to contain the full range of dwelling practices, and larger structures, termed ‘Assembly Houses’, which were built and burnt as public buildings. An important question concerns the relationship between ‘families’ and ‘individual dwelling houses’: it is possible that families were spread over several houses/households, perhaps even small Neighbourhoods. To the extent that households were relatively independent from the megasite as a whole (Ur 2014), households formed vital building blocks of the entire social order at Nebelivka and the practices which contributed to the materialisation of their identities must have constituted a significant part of the heterogeneity of the megasite as a whole (see VGA, Section 4.3). The Neighbourhood is a social term for a group of houses defined as separate from other house groups and having a minimum of three houses (the Nebelivka maximum for a Neighbourhood is 27 houses). The term makes the assumption that those living in the houses in the same Neighbourhood were more closely related to each other



The Theoretical Debate on Urbanism 

 31

than to those living in different Neighbourhoods. There are architectural, spatial and artifactual ways of assessing this assumption (see below, Chapters 4.3.1–4.3.2 & 5.1). Another assumption is that the number of houses in a Neighbourhood is a proxy for its overall duration – a proposition currently untestable with AMS dates. Michael Smith (2010, 137) has emphasised the importance of Neighbourhoods in urban development, due to ‘their status as communities with social ties among members (‘‘neighbors”), and the diverse functional roles they play within a city’. The third term – the Quarter – has an amount of historical baggage (e.g., Wilkinson et al. 2014) but it was found important to define a supra-Neighbourhood grouping smaller than the entire site level. A complex, multi-dimensional mechanism for defining the boundaries of the Nebelivka Quarters has been proposed (see below, Chapter 4.1), which gives a result of 14 Quarters of considerable heterogeneity. The construction of one or more Assembly Houses in the vast majority of the Quarters suggests their centrality to public practices (see VGA, Section 4.3). These three social groupings formed parts of a nested social arrangement, such that each house could be part of only one Neighbourhood and each Neighbourhood was part of only one Quarter. Such an arrangement was compatible with a broader, regional, descent-based social formation of the types familiar from social anthropological studies of kinship. Cutting across the nested social groupings was the term ‘Limited Interest Group’ (or ‘LIG’), introduced by T. Taylor in the 1990s (for discussion, see Chapman & Dolukhanov 1993). The term refers to any group of persons with shared skills or interests – nowadays, left-handed piano players, ice-skating enthusiasts or owners of Saab 90 cars: in the past, copper metallurgists, bone tool-makers or deer-hunters. The importance of a LIG would have increased with the diversification of a craft tradition, since LIGs provide useful channels for the dissemination of skills or innovations. LIG members would also be members of a household and a descent group. The classic form of lineage is an unilineal descent group that can demonstrate their common descent from a known apical ancestor. Unilineal lineages can be matrilineal or patrilineal, depending on whether they are traced through mothers or fathers, respectively (Friedman & Rowlands 1977). Clans are less clearly defined as to origin than lineages, sometimes lacking in a founding ancestor but also sometimes claiming an ancestor who symbolises the clan’s unity. Conical clan symbols often constitute an important visual differentiation of clan identity, whether as a cultural or biological referent (Friedman & Rowlands 1977). In their model, Friedman & Rowlands (1977) link lineages as constituent units to conical clans and the same is proposed for the Trypillia group. Residents at, and visitors to, the Nebelivka megasite would have been members of one of several lineages, each lineage itself forming part of a conical clan. It seems most likely that the majority of people in a Quarter were members of the same lineage or at most two lineages. In the wider settlement network, the tendency of both descent-group forms to practice exogamous marriage contrasted inclusivity at a regional level with exclusivity at the site level. The higher-

32 

 Theory and Practice for Trypillia Megasites

level conical clan structure would have underpinned the Trypillia Big Other through spatially extensive co-ordination of clan rituals and other practices involving the mobilisation of local lineage leaders. In this way, the nested Trypillia social formation would have generated centrally-placed megasites where major social practices integrated a large number of smaller sites. But how big were the populations of the Trypillia megasites?

2.1.4 The Maximalists, the Minimalists and the Middle Way A recent article comparing the structurally similar sites of Çatalhöyük and the South-West American pueblo of Awat’ovi reaches conclusions of direct relevance for the Trypillia megasites (Bernardini & Schachner 2018). Despite the similar architectural footprints, reconstructions of population for Çatalhöyük and Awat’ovi differ dramatically by a factor of 4x–6x in favour of the former. On these calculations, Awat’ovi remained below the threshold of 2,500 people (+/−500 people) above which formal, hierarchical leadership was almost always present but Çatalhöyük exceeded the threshold. The authors discuss several factors causing what they regard as an over-estimation of the Çatalhöyük population, of which the most important for us is the under-estimation of communal open space and the unwarranted assumptions of coeval occupation of all or most of the Neighbourhood clusters. All of the objections to the over-estimates for the Çatalhöyük population are matched in the Trypillia megasite case. We have termed the characterisation of megasites as long-term, permanent occupations with tens of thousands of people as the ‘maximalist’ view (Fig. 2.1). Such ‘maximalist’ views build uncritically on the existing models and interpretations of settlements-giants, claiming for one megasite (Majdanetske) contemporaneity of all buildings on the basis of 0.3% of the total number of houses dated with AMS determinations (Müller et al. 2017). Because of the adherence to the ‘archaeological’ definition of cities (see above) with its inherent check-list, the failure to tick the boxes of the urban-rural dichotomy and elaborate urban management (e.g., seals and property declarations) has relegated the megasites to the status of massive rural settlements (Müller & Pollock 2016). The views presented in the current volume challenge the premise of the city exclusively as a massive, high-density, permanent occupation, positing instead that emergent cities are very different from developed cities and thus much closer to preurban social formations. The alternative to the ‘maximalist’ approach (Müller et al. 2016) is the ‘minimalist’ or the ‘middle’ way developed here in terms of three different models of distributed governance, assembly and pilgrimage. Each model involves bottom-up agency and non-hierarchical social power and all seek to reconcile the footprint of 1,445 houses built at Nebelivka with the environmental evidence of low



The Theoretical Debate on Urbanism 

 33

human impact and lack of massive fires to account for so many ploshchadki (see below). They were also designed to fit the models of coeval house numbers developed from the AMS dates.

Figure 2.1: The ‘Maximalist’ model (by C. Unwin).

34 

 Theory and Practice for Trypillia Megasites

The choice of modelling megasite household and population trajectories stemmed from the difficulties of creating an internal site chronology caused by the wiggle in the calibration curve in the early 4th millennium BC (see pp. 49–50 and Section 4.8). The absence of a precise AMS-based model of changing house numbers left us with a choice of models which retained the key factor of the temporality of dwelling at Nebelivka. Any general statement about coeval house numbers (e.g., the claims for Majdanetske in Müller et al. 2016, 2017) would be immediately susceptible to challenge from the lack of diachronic changes in the life of the settlement. It was a requirement of two models – the Assembly and the Pilgrimage models – to build in temporal changes using the device of five 30-year generations; the permanent occupation of the Distributed Governance model makes generational time less significant (for definitions of the three models, see below, pp. 32–35). After that stage, the success or otherwise of a model could be explicitly measured against four criteria: (a) the number of houses built in the model could not exceed, or fall far short of, the footprint of 1,445 houses built at Nebelivka; (b) the number of houses burnt in the model could not exceed, or fall far short of, the footprint of 1,077 houses burnt at Nebelivka; (c) the effects of building and burning houses at any stage (generation) in the model could not produce a major human impact on the surrounding foreststeppe; (d) the number of houses in the model could not exceed the number of houses modelled using the AMS dates for each decade. It was clear from the outset that any model failing to meet all four criteria would be rejected. The three models used to account for the Nebelivka megasite are all variants on the theme of ‘assemblies’ but it is clear that each model characterizes ‘assembly’ (or ‘aggregation’ or ‘congregation’) in a different way. In her recent review of archaeological claims for ‘assembly sites, Milesi García (2018, Section 13.1) concludes that the term ‘assembly’ is not very useful as an all-encompassing analytical term, requiring closer definition in terms of the three key variables of time, space and social process. Milesi García’s complaint that the term ‘assembly’ has rarely received explicit definition is confirmed in a recent World Archaeology issue on ‘Gatherings’ (Semple 2018), in which contributors (ourselves included!) offer a guide to what happened at an assembly place without providing an explicit definition of the key term. The two processes Milesi García identifies in her Iberian Copper Age examples are settlement population concentrations and short-term gatherings in monumental places – two processes which have often been used to define opposed interpretations of these enclosures. However, Milesi García shows how both long human occupations and seasonal stays were found at Iberian megasites but at different times. In view of her strictures, we offer a variant definition of ‘assembly’ for each of the three models. What all three models share is a large population aggregation at a very large site that, if it is not monumental at the start of dwelling, becomes monumentalised within a decade or so through the large-scale construction of one- and two-storey houses and Assembly Houses. Where the models differ is in their temporality and



The Theoretical Debate on Urbanism 

 35

underlying social processes. Two of the three models portray the megasite assembly as a combination of a small ‘local’ population who lived there permanently and seasonal ‘visitors’ who came to Nebelivka from their own home communities for a month or several months per annum (the Assembly Model and the Pilgrimage Model). The Distributed Governance model factors in a larger permanent local population, one segment of which provided resources for all the residents for a year in concert with their home community. Thus each version of assembly hinges on the relations between ‘locals’ and ‘visitors’. The Distributed Governance model (Gaydarska, submitted) works on the premise of a permanent but substantially scaled-down contemporary occupation of 400 houses, with each of 10 clans living in 40 houses and each clan providing resources and leadership for one year in a 10-year cycle. The Assembly model is based upon a month-long seasonal congregation of increasing number of visitors at a centre maintained through the year by a small number of permanent guardians (Nebbia et al. 2018). The third model considers a concept hitherto rarely developed in prehistory – the megasite as a pilgrimage centre (Chapman & Gaydarska 2019). An intensive initial building phase to create the centre is followed by a long phase of gradual expansion of pilgrims’ dwelling houses, which cumulatively created the plan of the megasite as we know it. The choice of testing three alternative models of megasite assembly was made for two reasons. A larger number of models would have introduced undesirable overlaps which are, by and large, avoided in the three versions selected. However, the choice of only one model, or even two, would have given the impression of a certain confidence about the outcome – which we did not share then and do not share now. Therefore, the compromise solution of three models provides both appropriate variety and ambiguity. While each of these models is based upon severe reductions in megasite population estimates, the same size and scale of the settlement layout remains central to each model. The initial assumption of each of these models remains that a megasite is possible, that a combination of pre-existing layouts, households, Neighbourhoods and Quarters would result in a site form where successive generations of people would live in relative harmony until a perhaps inevitable decline. It is time, now, to step back from this comfortable assumption which, although strongly supported by hindsight, at any time from the mid-5th millennium BC onwards may not have worked out. In other words, with Benedict Anderson (1991) our guide, we turn to the possibility of Trypillia megasites.

36 

 Theory and Practice for Trypillia Megasites

2.1.5 The Possibility of Trypillia Megasites In Imagined Communities – the influential study of the anomaly of modern nationalism19, Anderson (1991, p. 4) reminds us that all communities larger than a single village are ‘imagined communities’, because separate communities have, by definition, never lived together with a second group. We suggest that integration of people beyond their normal, face-to-face groups required a vision of how those diverse communities could live together to derive benefits from the new settlement form that were considered greater than the difficulties this linkage may have brought. After all, there is a long tradition, beginning with Childe (1958), of praising the advantages of autarky – living in independent, face-to-face communities – which has, by and large, limited the scale of settlement nucleation in prehistoric Europe. Nonetheless, the existence of the Trypillia megasites clearly transcended small-scale communities, while their scale and size engendered an equally sizeable problem of how such communities were imagined in the first place. For let us be under no illusions: on the Eurasian continent of the 5th–4th millennia BC, the Trypillia megasites were unique in size and scale. There was nothing anywhere else on the planet, at 4200 BC, to compare with the Phase BI megasite of Vesely Kut, covering an area of 150ha – no analogies from which to derive this extraordinary place. We should never forget the unprecedented nature of Trypillia megasites, which have created immense problems of explanation and understanding but, first of all, problems of imagination. This discussion of how the Trypillia megasite communities were imagined will examine three crucial issues: what was the cultural background from which they emerged?; what were the changes in the Trypillia world to which megasites were a possible response?; and what (dis)advantages did Trypillia megasites bring to their world? Answers to these questions will frame what has become the Project’s approach to the Nebelivka megasite and its place in megasite developments in general. There are four inter-related elements of the cultural background of the Early Trypillia group (5000–4000 BC) which helped to create the possibility of the megasites: large-scale site planning, inter-regional networks, a pre-eminent symbolic order which we shall term the ‘Trypillia Big Other’ and everyday social practices. The general planning elements of oval house circuits, a wide inter-circuit space, the construction of large, public buildings, inner radial streets and a large central empty area were all known from Phase A–BI–I/II sites, although it is becoming increasingly likely that not one single large site earlier than Phase BII materialised all of the planning elements. The handful of Phase BI/II megasites20 invented new planning elements on sites of

19  We are not, of course, suggesting that Trypillia megasites were in any way reflected the development of Ukrainian nationalism. 20  Videiko (2007) mentions six Phase B I/II sites with sizes of 100ha or over, five of which are located



The Theoretical Debate on Urbanism 

 37

unprecedented size; it was only later that all the elements came together, through a process such as bricolage, in the creation of the massive BII megasites. Inter-regional exchange networks were particularly strong at the time of Trypillia A/Precucuteni III. One of the widest networks – the so-called Varna network – stretched from the Volga basin (the Hvalynsk cemetery) to Brittany, via Varna and North Pontic mortuary sites such as Giurgiuleşti (Chapman 2013). Prestige goods made of gold, copper, Spondylus and exotic lithics were exchanged over this network, one of whose nodes was the Karbuna site with its extraordinarily rich hoard dated to Phase A (Dergachev 1998). However, by the end of the Varna cemetery, ca. 4450 BC, this network had dramatically contracted to be dominated by two-way ceramic exchange between Skelya and Cucuteni-Trypillia communities, with rare Spondylus exports from the Black Sea to Lysaya Gora and an even rarer serpentine bracelet from the pre-Caucasus deposited in Novi Ruşeşti (Chapman 2002, Table 5.2). The exchange of high-quality Prut-Dniester flint Eastwards to the Bug-Dnieper Interfluve brought many communities into occasional but regular contact, contributing to Early Trypillia cultural interaction. The third widespread pre-existing linkage constitutes the domain of symbolic order, which we call the ‘Big Other’. The original notion of the Big Other is derived from Lacan’s (1988) perception of the world consisting of three registers – those of the Real, the Imaginary and the Symbolic. Although initially inspired by his training as a psychiatrist, Lacan’s ideas were disseminated mainly through his annual seminars, having a lasting influence on philosophers, anthropologists and other social scientists alike. According to Lacan, the ‘Big Other’ is a qua-symbolic order consisting of fictional ideas of anonymous authoritative power and/or knowledge such as Law, Nature, Science, God, the State or Ideology (Johnston, A. 2013). An underpinning theme in Lacan’s work is that the ‘unconscious is structured like a language’. The important implication of this claim is that the unconscious, i.e. the symbolic, is not chaotic and unruly but rather it is ordered and consistent. S. Žižek (2012, pp. 86–90) clarifies further the concept of the Big Other by discussing the efficiency of such a symbolic fiction which is constitutive of reality while being neither objective reality nor subjective inner experience. It is a virtual order that exists only through its subjects believing in it. Crucially, there are many cases in which individuals may stand for this symbolic order (Žižek 2012, p. 92). The implication for archaeology is that material culture may also stand for such a symbolic order. In the Trypillia context, the material world standing for the symbolic order comprises pottery, figurines and houses (Chapman & Gaydarska 2018) – the three most studied components in Trypillian archaeology and each contributing in important ways to everyday Trypillia life on all sites (Fig. 2.2). The sense of the Big Other recurs in Curta’s (2014, p. 2508) reminder that, for ethnic identities, it is not so much the group that endures as the idea of the group.

in the Southern Bug-Dnieper Interfluve.

38 

 Theory and Practice for Trypillia Megasites

Figure 2.2: The Trypillia ‘Big Other’ (by C. Unwin).

The last pre-existing element is linked to the Big Other but operates at a different, more intimate scale. Although the Big Other is a material part of the everyday life, it works at a larger scale well beyond the level of one settlement or a single person (Kohring 2012, Fig. 2). Bourdieu’s (1977) notion of the habitus, on the other hand, is practice-based and therefore very much part of the personal, communal and intercommunal engagement of each person with the world on a quotidian basis. The two concepts are nested in each other and complement each other. The concept of the habitus has been widely used in archaeology over the last 30 years, giving rise to a widespread practice theory (Hodder 1990; Barrett 1994; Dobres & Robb 2000; Thomas, J. 1999). Bourdieu’s (1985, p. 14) definition is that habitus is ‘a relationship of ontological complicity with the world’. In other words, social practices are implicit – they go without saying because they come without saying. Critiques of the habitus based on its supposed inability to explain change (e.g., King 2000) ignore Bourdieu’s later identification of ‘symbolic struggles and specific knowledge’ whose negotiation changes the social structure (1988, p. 21). Thus, the habitus enshrines the power of ‘world making’, and crucially to change the world means to change the ways of ‘world making’ (1988, p. 22). Each of the forms of objects that remained central to the Trypillia Big Other for two millennia were contributing to daily lifeways on every Trypillia site. We have seen how the house framed daily practices and materialised



The Theoretical Debate on Urbanism 

 39

hitherto invisible social relations. We have also noted the repetitive practices in which standardised figurines participated, as well as the influence of the form of pottery on appropriate and inappropriate behaviour. The participation of houses, figurines and pottery contributed reflexively to a widespread habitus, which formed the basis for shared meaning of cultural practices in different communities. One important element of the habitus was the spatial order produced as communities settled on the land and created their own settlements. This move from space to place has been much discussed in the social sciences, with a consensus around the dialectical relationship between the physical embodiment and the meaning of the built environment (Tuan 1977; Chapman 1988a; Tilley 1994; Ingold 2009; Hillier 2014). It will be claimed here that megasites became a monumental part of the habitus of Trypillia people (Buchanan, Section 4.3). This new kind of habitus was based upon non-competitive interactions between different small sites in one single large site. These four elements of shared cultural practice – settlement layout, interregional exchange, the symbolic order of the Big Other and the daily practice-based engagements with the habitus in significant places – were closely intertwined and hence could create a strong framework for people from different communities developing relations with each other. Yet this convergence of shared frameworks of meaning did not necessarily create the possibility of imagining such vast communities as the megasites: it was a necessary but insufficient cause of megasite emergence. For this, we turn to the second question – the changes in the pre-megasite world which may have generated the imagined communities. As we have noted (see above, pp. 25–26), the standard response to the changes leading to the emergence of megasites targets increasing levels of internal and/ or external conflict, migration and population pressure (Diachenko & Menotti 2017). However, there is little supporting evidence for any of these variables making a difference, with the increased levels of these factors being particularly unsubstantiated. An alternative approach is to consider the social and settlement changes in the early Trypillia Phase. There was a long-term increase in the clustering of small Trypillia sites from Phase BI onwards (Nebbia, Section 3.4), leading to pressure on agro-pastoral resources. One response was more intensive exchange within a site cluster to provide buffering; the other was the formalisation of nucleation on single, larger sites. This led to the consolidation of assembly places in Phase BI, which, by extension from Kruts’ (1989) model for Taljanki, constituted kin-based aggregations of up to 40 smaller social units, each corresponding to the population of typical sites of up to 10ha. Social gatherings of local, regional and inter-regional scale for the purposes of trade, exchange, exogamous marriages, ceremonies, feasts, celebrations, burials and conflict resolution became an important part of the Trypillia calendar – a notion supported by the overall material integrity of the Cucuteni-Trypillia network over a vast area. Those exchanges performed at a regional level as temporary gatherings at special places gradually offered the

40 

 Theory and Practice for Trypillia Megasites

possibility of conceiving of regular formal meetings at a place with accumulated ancestral place-value but on an even greater scale than before. Thus, there were two parts of a vision of the new site type: assembly and settlement. Formal assemblies could take place in the open central space of the megasite plan, while large numbers of people could settle in the dwelling zone outside the inner open area. The inner open area was a powerful residual reminder of the function of such places that fed into the symbolic order of Trypillian communities. Moving beyond the possibility of imagining this community to its actual creation leads us to the site itself and the (dis)advantages of its actual dwelling within the local and regional network (Fig. 2.2). An early advantage in the early months and years of a megasite was the huge prestige attached to a person’s association with the megasite because of its centrality. The centrality of megasites in local and regional networks was formed not in the traditional core-periphery sense but as places fostering different opportunities for ceremonial, political and exchange interactions. Such a network differed from the frequently reproduced link between urban centres and their hinterland in that sites did not need to be in the immediate vicinity of a megasite – in fact, we postulate that such a distance may span a 100km radius (see Chapter 3.4). What is important here is not the spatial but the social proximity of smaller settlements. Another advantage of the megasite was the heterogeneity that was built into the site. In a discussion informed by social anthropology, Hahn (2016, p. 176) maintains that urbanization created “the capability to practice the characteristics of heterogeneity as everyday life”. This may have occurred in two ways: first, the heterogeneity of practices, whereby urban occupants and visitors would experience some practices similar to those on small sites (for megasites, house-burning and pit-digging), but would be unfamiliar with other practices which were an urban prerogative (e.g,, regional assemblies and ceremonies or long-distance trade and exchange activities); and, secondly, the heterogeneity of identities – the crosscutting and complementary identities of members of various corporate groups, different genders and forms of personhood (Gaydarska 2019), households and Neighbourhoods, which were often in tension with the overarching urban identity of the imagined community. This rather different vision for the habitus in urban places meant an advantageous fluidity of practice and identity for all those coming to the megasite, offering a greater potential for meetings, exchanges and ceremonies than was available in the home community. But these advantages were not to the detriment of losing contacts with the home community: instead, the mobility offered by megasites allowed continuing contact with one’s roots as well as expansion into new networks. The long-assumed permanent occupation of megasites is in need of much more supporting evidence; if that is not forthcoming, seasonality and mobility can finally take centre stage (cf. Tkachuk 2010–11).



The Theoretical Debate on Urbanism 

 41

The early years of a megasite may not necessarily have introduced major social change at once, although there was an obvious difference between the identities of small, autonomous, local communities and the few places which hosted formal assemblies for groups of such small settlements. These two identities were in tension all year long at the meeting-place but became particularly potent at the time of the assembly. In the early years, the megasite identity may have been weak, lacking as it was the formal mechanisms of ritual, ceremonial and feasting practice that consolidated such an identity. But the advantages of developing an overarching identity to the success of the megasite meetings may have produced a response in the creation of new ceremonies, depositional opportunities or feasting events that reflexively strengthened the megasite identity. This emergent process would have lasted over years, if not decades, and if successful would have created an attractive identity for the ‘imagined community’ that lay at the heart of the megasite. The greatest disadvantage to living in such massive communities as megasites was undoubtedly the potential for scalar stress (Johnson 1982), which could drive constituent groups apart and cause the failure of the entire settlement. However, there were several mitigating factors which reduced this risk. We suggest that Kruts’ (1989) kin-based aggregation could manage scalar stress by allowing the operation of everyday activities, including disputes (Johnson 1982), at smaller operational units such as the Quarters (see Chapter 4.2) or even smaller units such as Neighbourhoods. These emergent social practices were congruent with the notion of egalitarian, selfgoverning and heterarchical forms of society, rather than hierarchical forms. It was in the interests of small cooperative groups to control status competition at incipient cites (Jennings & Earle 2016) – an idea supported by the current lack of evidence for materialized hierarchy at megasites. We concur with Ur’s (2014) suggestion that the household was a key metaphor in the social organization of incipient urban societies and his conclusion that scalar transformation at such sites was not a mere sum of its constituents. A further mitigating factor of scalar stress was the low residential density of below 10 houses per ha found at all Trypillia megasites. In addition to the well-argued global advantages of low-density occupation (Fletcher 2012), the immediate benefits of such habitation should be emphasised, e.g., the creation of social space comparable to that found in megasite visitors’ home communities, with room for gardens, smallholdings and animal keeping. Lastly, the short times of such assemblies allowed deferral of tensions until the end of the season. If scalar stress constituted the principal threat to megasite survival, the opposite problem was how to continue to integrate such large numbers of residents and visitors. We should not forget the unprecedented size of Trypillia megasites – never seen before in Eurasian prehistory. What were the integrative principles and practices that produced the megasites and what held these disparate Neighbourhoods and Quarters together? The straightforward answer is the same four inter-related elements which helped to create the possibility of the megasites: large-scale site planning,

42 

 Theory and Practice for Trypillia Megasites

inter-regional networks, the ‘Trypillia Big Other’ and everyday social practices. The construction of a megasite can be viewed as an expression of cohesion, co-operation, obligation and inter-relationship within a larger community than an immediate kinship group. Perhaps the most critical element of the four was the Trypillia Big Other, which integrated communities through an already closely linked suite of material practices centred on the house (Fig. 1.8), the figurine (Fig. 1.4) and pottery (Fig. 1.3). The predominant and preferred study unit in Trypillian archaeology is the burnt house, referred to as ‘ploshchadka’ in the Russian and Ukrainian literature. This term is very revealing, as it is explicitly linked to the physical remains with no mention of the sociability implied by words like ‘house’ or ‘dwelling’. People are conspicuously missing from the discussion of house-building and -burning, while house assemblages are taken literally with little appreciation of taphonomy and site-formation processes (Müller et al. 2017). It is our aim here to bring the members of the house back to the fore. The production of an individual house can be viewed as a symbolic fusion of the different elements that made up the Trypillia landscape. Creation and fusion was achieved through combining clay from the earth with straw from the steppe or as a by-product of agriculture, wood from the forest and reeds and water from the rivers and lakes. For Bailey (2000 p. 268), the building of houses established otherwise invisible sets of social relations as explicit and materialised, as well as increasingly durable and in many cases permanent. The personality of each house would have emerged out of the combination of persons living there – the persons who were at the same time developing their own forms of personhood according to household and wider, corporate principles. The ways in which these principles of personhood were worked out in daily practice were strongly grounded in Trypillia household practices and relationships, only some of which required materialisation. It was largely out of the household setting that gendered (in)dividuals emerged through reiterated practices of cultural transmission. Bradley (2005, p. 120) maintains that “In prehistory, ritual gave domestic life its force and domestic life in turn provided a frame of reference for public events”. This was equally true of figurines as it was of houses. Figurines are traditionally viewed as sacred objects or deities. The concept of the Big Other advocated here is much more flexible and encompassing, allowing for symbolic renderings to be linked to all aspect of social life, not exclusively to cult and religion. The importance of figurines to the Trypillia world is shown by their ubiquity on sites, with ca. 10,000 known examples of anthropomorphic figurines (Ţerna 2017). Their frequent fragmentation (often over 90% of figurines on a site were deposited as broken) stimulated the enchainment of persons as well as households to each other (Chapman & Gaydarska 2007, 2015). Their flexibility of design made figurines an ideal constituent of the Big Other, with designs ranging from highly stylised to realistic ‘portrait heads’ (Fig. 2.3). The latter showed how



The Theoretical Debate on Urbanism 

Figure 2.3: Realistic and stylised figurines (by B. Gaydarska).

 43

44 

 Theory and Practice for Trypillia Megasites

Trypillia personhood was sometimes materialised through individual features, while the former is reminiscent of Orphanidis’ theory of repetition (2010, p. 109), in which people used figurines to mediate community agreement on critical concepts of being or one’s position in a social group. In this sense, figurines were material codes with which people re-shaped reality (Kokkinidou & Nikolaidou 2010, p. 76). Last but not least is the pottery, which fulfils the two basic features associated with the Big Other – a long-running significance in daily practices and the potential for variability which allows a great diversity of people to continue using ceramics. A key role in these new social environments were the things that came to dominate daily life – expressive things made of a huge variety of materials, in a vast range of shapes and sizes, and with an almost unlimited durability (Bailey 2000, p. 270). These things were dominated by pottery, which became a key element in the ‘container revolution’ (Gamble 2007), leading to new opportunities for food storage and allocation and creating a range of appropriate practices which formed part of the habitus. Techniques for making new things such as pottery added to the expanding rules of social tradition (Childe 1949). The widespread distribution of the means of aesthetic production of pottery (Fig. 2.4) resulting from its predominantly household production created a strongly enchained system of social relations between consumers and producers, because they were often members of the same household (Wengrow 2001). Dramatic changes in the gendered structure of household work would have followed changes in pottery production into separate workshops within the settlement (e.g., Trypillia sites such as Varvareuvka XII: Ellis 1984) and/or by the introduction of pottery kilns on Trypillia megasites (KorvinPiotrovskiy et al. 2016). We have seen that all the objects which materialised the Big Other were central to daily Trypillia social practice, exhibiting the most significant visual effects on everyday performance. The long-term structuring effects of houses extended beyond the control of movement within the megasite to the creation of the households that produced and deposited the pottery and the figurines. The imagining of megasites in the mid-5th millennium BC would have been impossible without the existence of the shared meanings of the house, the figurine and the pottery that linked so many communities. What megasite households did with their houses, figurines and pottery made a further, enormous contribution to the perpetuation of megasites as successful examples of the biggest sites in 5th–4th millennia BC Eurasia.



The Theoretical Debate on Urbanism 

Figure 2.4: The Scânteia vessel with multiple symmetries (source: Monah D & M 1997, Fig. 46).

 45

46 

 Theory and Practice for Trypillia Megasites

Table 2.1: Summary of Project activities. Number of units

Activity Unit

6

Years of fieldwork and excavation

1

Geographic zone (Northern tributaries of Southern Bug)

14

Field activities

7

Field seasons

137

Participants

286

Ha of geophysical investigation

1,445

No. of houses identified

850*

No. of Pits identified

24

No. of Assembly Houses identified

160

No. of geophysical anomalies cored

10 (2,619m )

Area excavations (area excavated)

88 (225m2)

Test pit excavations (area excavated)

86 / 25

AMS dates / radiocarbon dates

6

Units tested with soil micro-morphology

25

On-site soil investigations

39

Transect soil cores

233

Ha surveyed with intensive systematic fieldwalking

178 (574)

Sq km of linear valley walked by extensive survey (ha fieldwalked)

4

No. of sites with intra-site gridded collection

34

Surface concentrations documented

800

Burial mounds registered

5

Sediment cores collected

1

Full multi-proxy analysis of sediment core

15

Units investigated in megasite hinterland

2

No. of experimental houses built

1

No. of experimental houses burnt

1

No. of excavations of burnt house remains

2

No. of building experiments, Durham

37,000*

No. of sherds excavated

21,300 (504 kg.)

No. (Weight) of sherds fully studied

6,316

No. of animal bones excavated

150

No. of chipped stone studied

85*

No. of ground stone tools

291 (143)

No. of Special Finds (figurines)

5,248

No. of Digital Images in ADS Project Archive

394

No. of Documents in ADS Project Archive

2

Key: * – estimated value



Changing Perspectives – Stable Research Questions 

 47

John Chapman & Bisserka Gaydarska 2.2 Changing Perspectives – Stable Research Questions This project has been the first multi-disciplinary investigation of megasites ever conducted21. We sought to deliver field data and interpretation on a scale never attempted on megasites, whose very size makes them extraordinarily difficult to investigate. At the outset, we defined eight major research questions for project investigation. Despite the changes in research perspectives outlined in Chapter 1 (‘A Project biography’), the research questions have, for the most part, remained stable, largely because they constitute the most basic research questions to be asked of a Trypillia megasite. The Project’s wide-ranging research activities have been summarised above (Table 2.1) to allow the reader to relate them to the research questions which we now discuss. It should be recalled that the number of visual images and documents presented on the ADS Project Archive represents but a small fraction of the total available images and documents. Question 1: What was the settlement plan at the Nebelivka megasite? At the outset of the Project, no complete plan of a megasite had yet been produced using modern geophysical methods. The derivation of a complete settlement plan of Nebelivka through large-scale geophysical prospection of the 238ha site would allow the identification of a wider range of anomalies than had previously been recorded, a detailed spatial analysis of the settlement plan as well as insights into small-scale variations in plan at various spatial scales (see Chapter 4.2). Methods for Question 1 - Data collection Measurements of vertical geomagnetic field gradient were determined using Bartington Grad 601-2 dual sensor fluxgate gradiometers. A zig-zag traverse scheme was employed and data were logged in 30m grid units. The sample interval was 0.25m and the traverse interval was 1m, thus providing 3,600 sample measurements per 30m grid. Approximately 3,600 grids were surveyed at Nebelivka, providing almost 13 million geomagnetic data points. The complete data archive is held at Durham University. - Data processing Geoplot v.3 software was used to process the geophysical data and to produce continuous tone greyscale images. The basic processing functions applied to the geomagnetic data typically included clip, zero mean traverse, de-stagger and interpolate (typically to 0.25m × 0.25m intervals). The principal data processing issue

21  The multi-disciplinary projects at Majdanetske, Taljanki, Dobrovodi and Apolianka began in 2011 (Rassmann et al. 2014).

48 

 Theory and Practice for Trypillia Megasites

was the suppression of the strong magnetic signal from the underlying Ukrainian ‘granite shield’ by the use of a high pass filter with Gaussian weighting. - Notes on structural features There are two types of rectilinear geophysical anomaly that have been interpreted as ‘houses’: intense anomalies (e.g., -30 to +80nT), considered to be burnt houses, and weak anomalies (e.g., +1 to +6nT), considered to be unburnt houses.22 A third category comprises anomalies which are considered likely to reflect houses, but which are amorphous to varying degrees, and either weak or strong or a combination of both. Using criteria of size, orientation of anomaly, strength of anomaly and location, none of the geomagnetic anomalies at Nebelivka is entirely consistent with what would be expected of a kiln anomaly. - Testing the geophysical results Several field tests were performed to check the accuracy of the geophysical results. The first took place in 2009, when the first intra-site gridded collection on a Trypillia megasite was carried out over a 15-hectare area in the South-West corner of Nebelivka. A total of 138 30 × 30m grid squares (12.5ha) was surveyed, using the same grid as the geophysical survey of the same area. The collection was a timed pick-up of surface material, overwhelmingly burnt daub and potsherds, with thirty person-minutes allowed per square. A total of 32 grid squares was not surveyed due to time constraints, and priority was given to those squares which included magnetic anomalies on the geophysical survey. Additionally, four squares have missing data for at least one class of find. Consideration was given to cultural and natural transforms affecting the field data, as well as surface visibility. The results were compared with the distribution of house-shaped magnetic anomalies, with a good correlation between surface daub concentrations and house-shaped anomalies (see Chapter 3.2.2). The second test of the geophysical results involved soil coring, performed on a limited scale in 2009 and much more extensively in 2012. In 2009, daub was retrieved from the two cores placed over house-shaped anomalies, while dark, organic-rich soil was found in the cores placed over pit-shaped anomalies. In 2012, all of the 91 cores placed over house-sized anomalies yielded daub which was later scrutinised for charred botanical remains. The last coring was conducted at two places over a linear anomaly interpreted as the perimeter ditch; in both cases, deep ditch-fill was recovered in the cores. Full-scale excavations of burnt houses, adjacent pits and ditches in 2012–2014 showed good correlations between the size of the anomalies and the eventual size of the excavated features. Finally, all of the 88 house-shaped anomalies which were investigated by test pitting showed burnt or unburnt house remains. The results of the testing of the geophysical data showed a remarkably high accuracy in the identification of magnetic anomalies at Nebelivka (see Chapter 4.2).

22 Pichartz et al. (2019) use inverted soil magnetization mapping to demonstrate that "house masses ... are basically daub masses".



Changing Perspectives – Stable Research Questions 

 49

- Analysis of the plan The limitation on analyses of the Nebelivka plan is our initial assumption that not all houses were built and settled coevally. This limits the numbers of useful analyses; for example, in the publication of Ohlrau’s (2015) Masters dissertation, all eleven of the excellent analyses of Majdanetske were based upon the assumption of coeval occupation of all houses. Comparative spatial analyses of the location and composition of Neighbourhoods and Quarters have been complemented by the comparative analysis of Assembly Houses. An analysis of house sizes by Neighbourhood, by Quarter and for the whole site has been conducted, together with a GINI Coefficient analysis of house sizes by Quarter (see Chapter 4.3.1). An extensive Visibility Graphic Analysis (VGA) has been conducted for ten Quarters using 10 different kinds of analysis (see Chapter 4.3.2). In addition, VGA of the three stages of each of two Models (the Distributed Governance Model and the Assembly Model) has been completed for six Quarters. The selection of 10 Quarters from the original 14 provides a sound sample of the spatial variability found across the megasite. The parameters of the Pilgrimage Model were developed too late to allow extensive VGA analysis; the application of the diachronic analysis of the two Models to only six Quarters was predicated upon the high probability of redundant results from extending the analysis to four more Quarters. Question 2: What was the internal chronology of the Nebelivka layout? What was the chronological place of Nebelivka relative to other megasites? The production of an internal chronological sequence for Nebelivka was vital, since is is difficult to gain a full understanding of megasites without an estimate of the number of coevally occupied houses. The central methodological issue is how to place what turned out to be 1,445 structures on the megasite into a chronological sequence. Four key questions were posed about the chronology of Nebelivka: (1) How long was the occupation of an individual segment of the inner or outer circuit?; (2) Were adjacent houses and segments constructed, occupied, and destroyed sequentially or coevally?; (3) How many segments/groups were constructed, occupied, and destroyed coevally across the whole site?; and (4) How do the radial streets inside the circuits relate chronologically to the circuits? Following on from the internal chronology of Nebelivka was the question of placing Nebelivka in a secure chronological relationship to other neighbouring megasites, such as Taljanki and Majdanetske, as well as to the sediments of the multi-proxy Nebelivka P1 core, with its own age-depth model. Methods for Question 2 Initial dates were obtained from the structures excavated in 2009. The majority of dates were obtained following a sampling strategy based on the geophysical plan of the site. The aim was to sample different sectors of the site, including the inner and outer circuits of houses and the inner radial streets. The sampling strategy followed the highly organized spatial arrangement of houses. We aimed to sample houses from an inner circuit segment, from an outer circuit segment and from each of two groups inside the circuit to address the question of contemporaneity of houses within these

50 

 Theory and Practice for Trypillia Megasites

groups, the duration of use of a segment/group, and whether the construction dates were spatially structured. The first method for obtaining datable samples was the auger coring of ploschadki to obtain daub samples from which charred plant remains could be extracted for dating. However, the daub firing temperature was so high that the charred cereals were destroyed; 130 cores from 91 houses yielded just one charred grain. Instead, a programme of test pit excavation was undertaken to recover stratigraphically reliable bone samples from houses. Bone samples were screened for collagen preservation using overnight acid demineralisation of a chip of bone, and only those yielding visible collagen were considered for submission to the Oxford Radiocarbon Accelerator Unit. A grant-in-kind from the National Radiocarbon Facility funded 80 dates. A sub-set of samples was divided and parts sent to the Poznań Laboratory as well as to Oxford for the measurement of inter-laboratory consistency. The final set of dates was measured at higher precision in an attempt to overcome the problem of the wiggle in the radiocarbon calibration curve at ca. 4000–3700 BC. Attempts were also made to model the duration of individual houses. Question 3: What was the distribution of Trypillia, as well as earlier and later, sites in the Nebelivka hinterland and micro-region? We needed to understand the settlement pattern in the Nebelivka hinterland (5km radius) and its micro-region (25km radius), since many urban studies had posited an inter-dependent relationship between an urban centre and small sites close to the the centre (Adams 1965). It was also important to investigate off-site discard in the hinterland to identify the incidence of manuring scatters as a mark of agricultural intensification. Wider investigations of the settlement pattern would shed light on Linda Ellis’ (1984) contention of a three-level settlement hierarchy in the Uman region. A diachronic perspective for both areas would set Trypillia settlement in a longer time-frame, helping us to understand continuities or otherwise in settlement location and thus providing ideas for establishing the reasons for the location of Nebelivka. Methods for Question 3 Two methods never previously used in Trypillia research were utilised to recover settlement information: remote sensing and intensive, systematic fieldwalking. The free availability of CORONA satellite images over a wide area enabled coverage of the Nebelivka micro-region and beyond (see below, Question 4). The high commercial costs of WorldView2 satellite images limited coverage to a 25km block around the megasite. The results of each set of images were compared to assess their respective merits in a heavily cultivated, residual forest-steppe environment. Intensive, systematic fieldwalking was conducted using two sampling strategies: (a) coverage of all parts of the landscape in a 50% sample of all fields within a 5km radius of Nebelivka; and (b) judgmental sampling along streams within a 25km radius. The first technique was applied to fields in 2009, with GPS recording of every surface object found during a 20m-wide spacing transect. This labour-intensive data recording was abandoned in 2012 and 2013 for a recording of single finds (‘off-



Changing Perspectives – Stable Research Questions 

 51

site’) by 20m × 50m transect, with greater attention given to surface scatters (‘sites’). The restriction of all scatters to a zone close to existing streams led to a judgmental coverage of such zones in 2014 (see Chapter 3.4). A third more intensive method was used to investigate the alleged sizes of Trypillia settlements, which were suspected of being over-estimates of the actual site sizes. This method involved collecting surface sherds and daub along a transect from the centre of a site to well beyond the limits as judged by aerial photography or geophysical survey. In this way, a better size estimate of four Trypillia sites was achieved (see Fig. 3.16). Question 4: How did megasites relate to their wider settlement context? The huge area of the Cucuteni-Trypillia group (up to 250,000km2) meant that no single region had been subject to intensive, systematic fieldwalking. The data on site location and size was regionally variable, yet this data set was important for placing Nebelivka and its micro-region in a wider regional settlement network of interaction and mobility covering the whole of the Trypillia area. The well-known concentration of megasites in the Southern Bug-Dnieper Interfluve raised the question of environmental differences between this region and other Trypillia areas, while a regional investigation of site sizes would be a second test of Ellis’ hierarchical settlement model. Finally, if Trypillia megasites exerted influence over smaller sites, how far did this influence extend and were there regularities between megasites? Methods for Question 4 Remote sensing was also applied on this wider spatial scale but with limited success. The main method was a source-critical analysis of the site database included in the Encyclopaedia of Trypillia Civilization (Videiko 2004), with its synthesis of information for all known sites, using a gazetteer with site phasing and general co-ordinates (“the site lies 6km NE of the village of XXX”). This critical filtering of the Encyclopaedia’s database led to a much smaller set of sites with reliable locational and size data. Marco Nebbia’s engagement with exploratory spatial data analysis led to a comparison of location and environmental parameters of megasites with smaller sites to assess environmental influence on megasite location and the use of Incremental Global Moran’s I index and Anselin’s Local Moran’s I Index to investigate site hierarchy and the scale of megasite influence (see Chapter 3.3). Question 5: How can we best characterise the Nebelivka landscape? What was the human impact of the megasite on the surrounding landscape? The vegetation history of the semi-arid forest-steppe landscape of Southern Ukraine has not received widespread attention, mainly because of the scarcity of long-term peat deposits outside of the principal Ukrainian valleys (e.g., the Dnieper, the Southern Bug and the Dniester: Kremenetski 1995, 2003). Our challenge was to locate coring sites which could be chronologically related to the Nebelivka megasite, so as to place the site in the context of its long-term vegetation history, as well as to make a detailed assessment of the human impacts caused by megasite dwelling. A

52 

 Theory and Practice for Trypillia Megasites

second approach concerned the information on the local, site environment provided by the molluscan remains found in all excavation units. Methods for Question 5 The discovery of a Ukrainian Government database of peat-extraction sites revealed the existence of two peat-filled depressions within 50km of Nebelivka. A preliminary investigation of these sites in 2009 and 2012 yielded one potentially valuable core for regional vegetational history at Onopriivka but this site was too far from Nebelivka to provide any idea of human impacts. A survey of the stream valleys around the megasite revealed three coring alluvial sites which were tested in 2012, using the particular expertise of Dr. Bruce Albert, who had been working on small alluvial basins for a decade. The only core to sample a palaeo-ecological archive coeval with the megasite occupation was the 6m Core P1, 250m from the NE edge of the megasite. The archive sampled by Core P1 was subjected to multi-proxy analysis, including Loss on Ignition analysis, particle size analysis, pollen and non-pollen palynomorph analysis, microcharcoal counting and sedimentological analysis. Bayesian modelling of 11 AMS dates provided an age-depth model. The results proved an important catalyst for the Project’s re-conceptualisation of the nature of a megasite (see Chapter 4.1.1). The Project’s flotation programme of a sample of every context in the four excavation units yielded a varied suite of remains, including lithic chips and pottery fragments. The sizeable molluscan assemblage studied by Dan Miller (see Chapter 4.1.2) showed that the majority of molluscan species was characteristic of steppic rather than wooded environments, whether deriving from test pit samples below or above the living floors, the soils buried under the sole barrow constructed on the megasite area and the soil pits designed to provide a long-term soil sequence. The only limitations to the molluscan analysis concerned the lack of AMS dates attributed to most of the samples; this was particularly important in the attempt to reconstruct the environment of the Nebelivka promontory before the occupation of the megasite (see Chapter 4.1.2). Question 6: How can we reach a better understanding of house architecture and finds assemblages at Nebelivka? One of the key strategies for the investigation of megasites was the Ukrainian predilection for the excavation of complete burnt houses. It was anticipated that this strategy would be pursued at Nebelivka, as it was in two cases – Houses A9 & B17. The Project Biography (see above, p. 12ff) alluded to the two key decisions of excavating the Mega-structure and conducting extensive test pitting for AMS samples, which led to the recovery of far greater quantities of Trypillia pottery and Special Finds than we had anticipated at the start of the Project. These decisions led to changing Research Question 6 from a better understanding of ‘house architecture’ to ‘house architecture and finds’. This positive development necessitated the integration of our thinking on houses and other features (pits, Assembly Houses) with our approaches to their



Changing Perspectives – Stable Research Questions 

 53

contents, especially through questions of taphonomy. Three issues of house practice were of particular relevance to the interpretation of household pottery assemblages: whether a house was one-storey or two-storey; whether the finds in a burnt house were a reflection of the residents’ living assemblage or a staged, placed deposit constituting a ‘household death assemblage’, comparable to offerings in a grave; and how the houses were actually burnt down. In taphonomic discussions, the three basic terms introduced by Schiffer (1976) were extended by Kuna’s (2015, Fig. 22.1) elaboration of three more categories of refuse: – primary refuse – the remains of artifacts left at the locus of activity where artifacts were produced or used (Schiffer 1976, p. 30) – secondary refuse – damaged or destroyed artifacts intentionally moved to refuse areas (Schiffer 1976, p. 129) – de facto refuse – discarded refuse that is still usable (Schiffer 1976, p. 33) – tertiary refuse – artifacts that had found their way into their place of deposition not in the form of individual, damaged artifacts but together with the material of the layer in which they were originally deposited as refuse (Kuna 2015, p. 281) – internal residue – material that has accumulated in the cultural level during the existence of a given activity area (Kuna 2015, p. 281) – external residue – the remants of preceding, unrelated components or phases of the site (Kuna 2015, p. 281). These additional terms do much to offer precision in the way that artifact deposition can be described. The artifacts found at Nebelivka were so numerous and so diverse that they offered many opportunities for investigation – in terms of intra-site distribution, formal variability and differences in association and context. A final class of biological finds (or ecofacts) comprised the animal bones and macro-botanical remains discarded as food waste. The faunal remains were studied to answer three questions: (a) what was the relative contribution of hunted vs. herded animals at Nebelivka, and how does this fit into wider trends noted for the Trypillia period?; (b) are there any detectable differences in animal use (or at least bone deposition) between areas of the site and/ or between different context types (e.g., houses and their associated pits)?; and (c) what was the nature of bone deposition in the Mega-structure? Methods for Question 6 (houses) Four methods were adopted to improve our understanding of Trypillia houses – excavation, soil micromorphological analysis, the analysis of building materials and an experimental programme. In line with the expansion of types of magnetic anomalies found in the geophysical investigations, it was decided to excavate at least one complete example of each of the units – a house, an Assembly House and a pit – in addition to the large number of test pits in which we targetted a small sample (3–5%) of a burnt or unburnt house or an Assembly House (1–2%). A major goal of this strategy was to make comparative

54 

 Theory and Practice for Trypillia Megasites

analyses of the pottery, animal bones and Special Finds discovered in each of the four excavation units in order to gain a deeper understanding of the contexts of deposition of the portable finds. While total excavation of Trypillia burnt houses has been the standard field method for Ukrainian specialists, it was clear that the utility of this method for a fouryear Project on a site with 1,445 houses was not cost-effective23. Instead, we decided upon the use of test pits – a strategy that was far less common in Ukraine. An initial test pit size of 5m × 1m was abandoned because of high labour costs. Instead, in the course of 2013, test pits of 2m × 1m were used to recover animal bone samples for AMS dating; if none was discovered, the test pit was extended to 3m × 1m and, in a few cases, 4m × 1m. While it is conceded that a 3–5% sample of a house could hardly be considered to be a representative sample of that house’s material assemblage, the excavation of over 80 test pits across all main elements of the Nebelivka plan (outer and inner circuits, inner radial streets and Squares) gave a good sample of the total range of material culture used across the site, as well as an unparalleled set of snapshots into household architecture, both building and burning, across the whole site. What the test pit approach lost in depth, it gained in breadth. It has now been recognised as a indispensible method by the Ukrainian-German project at Majdanetske (Müller et al. 2017) (see Chapter 4.6.1). The soil micromorphological study was based upon a total of 22 undisturbed block samples and associated bulk samples collected from nine separate contexts at Nebelivka. In order to retrieve archaeological information and evidence of past formation processes from the soil archive, block samples were made into thin sections for soil micromorphological analysis (impregnated and cut at the UCL Institute of Archaeology; crafted into polished thin sections by Spectrum Petrographics Inc.). Bulk samples were employed for the measurement of different physical and chemical parameters (pH, organic matter, carbonates, and magnetic susceptibility). Among the questions we hoped to answer were: i) What microscopic indicators of human activity could be detected using soil micromorphological analysis? ii) How were the different occupation deposits formed? iii) What factors affected archaeological preservation – including of charcoal and bone – within these deposits? Overall, the study of these research materials illuminated important differences between the different contexts sampled (see individual excavation unit reports in Chapter 4: barrows – 4.5.2; test pits – 4.6.1; pit, Sondazh 1 – 4.6.2; House B17 – 4.7.3). Dr. Natalia Shevchenko’s programme of analysis of building remains focussed on the daub remains from the Mega-structure, for which she carried out detailed laboratory tests designed to shed light on three areas of research interest: (1) the establishment of the technical – typological characteristics of the samples, including

23  And also time-consuming: according to our estimates, the excavation of two complete houses per annum would take until 2743 to complete!



Changing Perspectives – Stable Research Questions 

 55

their physical-chemical properties, their composition, their structure, their context and their recipes; (2) the determination of the function of the different building materials; and (3) the classification of building materials based on point (1). A combination of macro- and micro-level methods was used, including morphological analyses, qualitative analyses (composition), quantitative analyses (ratios between different elements in clay mixture), metric analysis (fractions of content, thickness of layers of building material) and textural-structural analyses. The most characteristic samples were selected for further laboratory analysis, including stereo-microscopic investigation and the making of thin-sections for polarising microscopy to recover more specific information on mineral contents, structural traits, ratios of elements in the clay mixture and mineral temper. Micro-probe analysis was used for the clay minerals in some of the samples, while polished sections (Russian: Anschliff) were studied in reflected side light to gain further stratigraphic information (see Chapter 4.9). The least precise method of visual inspection of daub fragments during excavation was also an important aspect of the recording protocols, which involved characterisation of the daub on plans and sections in one of three ways: (1) destruction daub – daub produced by the burning of a structure at low to medium temperatures (this was by far the commonest type of daub); (2) vitrified daub – daub produced by high-temperature firing of a structure; and (3) construction daub – daub forming an interior feature of a building, usually a platform, a podium, a hearth or a bin (for an example, see Fig. 4.40). Experimental house-building and -burning has been part and parcel of CucuteniTrypillia archaeological methodology, but there have been ambiguous results, especially over two important questions: the construction of 1- or 2-storey houses; and the way in which houses were actually burnt down. An experimental programme of building, burning and excavation of the burnt remains was designed in order to seek answers to these questions (see Chapter 4.4) and to provide the Project with insights into house-burning taphonomy. Two small-scale experiments in the Durham University Botanic Gardens woods were carried out to investigate (a) Korvin-Piotrovskiy’s hypothesis of ‘construction burning’ (Korvin-Piotrovskiy & Shatilo 2008; Korvin-Piotrovskiy et al. 2012) and (b) ways of producing the cracked daub surface effect found during excavation of fired clay platforms. The main experimental effort was devoted to the construction, burning and excavation of two timber-framed, wattle-and-daub-walled ‘Trypillia’ houses with a footprint of 4 × 3m in the centre of Nebelivka village. The construction team took initial advice from Dr. Videiko on Trypillia construction methods, although the use of modern tools and the delivery of building materials to the building site failed to match Neolithic practice. Mr. Johnston’s detailed recording of the construction process and estimates of building resource required formed the basis of his Durham University B.Sc. undergraduate dissertation (Johnston, S., n.d.).

56 

 Theory and Practice for Trypillia Megasites

The burning of only one of the two experimental houses (the two-storey house) was carried out at the time of the Kirovograd-Nebelivka International Conference (May 2015), thus vitiating the possibility of a comparative burning experiment. The entire burning procedure was carefully recorded through photography and a timed diary of burning stages, starting with the filling of the house with 30m3 of firewood the day before the firing and concluding with the collapse of all but one of the house walls six hours after ignition. There was one aim for the Project team’s final fieldwork season (2017): the geophysical investigation and excavation of the experimental burnt house remains of 2015. A Kyiv colleague, Dr. Kseniya Bondar, tested two methods: a cesium magnetometer test of the house mound, using a PKM-1M (Geologorazvedka, Russia) instrument, and a ground-penetrating radar test of the same, with a VIY-2-300 (Transient technologies LLC, Ukraine) instrument. The Nebelivka villagers had solicitously covered the remains of the experimental burnt house with earth to form a low mound. The mound was divided into quadants and two quadrants were excavated – the West Quadrant and the North Quadrant. Hand excavation produced a mass of structural detail, much of which was recorded by photography and some of which was plotted by photogrammetry24. Methods for Question 6 (finds) The analysis of artifacts was conducted in a variety of methods, depending upon the material and the excavation unit(s). The largest finds assemblage – the pottery – was studied in terms of its production, fragmentation and discard taphonomy before a series of 13 analyses investigated a range of questions, from basic statistics on sherd number, weight and mean weight to the fabrics chosen for specific vessel shapes. Most of these analyses compared the samples from each of the four main excavation units. There was also a comparative analysis of the Nebelivka pottery with other megasites (Taljanki and Majdanetske) as well as with the Kaniv group in the Dnieper valley (see Chapter 5.1). Detailed reports were made on many of the Special Finds (especially the lithics and worked bone), including their placing in a broader comparative perspective (see Chapter 5.2). The group of miniature vessels from the Mega-structure received special attention because of its unique character, with lipid analysis of their contents and isotopic analysis of the graphitic decoration and washes on some vessels (see Chapter 5.2.3). Another form of analysis made use of the large numbers of Test Pits, which produced spatial data on Special Finds, as well as pottery decorative motifs, from all over the megasite. The Project’s flotation programme began in 2009 with the use of bucket flotation, continued with the use of a Legge Mark IV flotation tank in 2012 and resumed with

24  There is a plan for a short fieldwork season in the future to collect soil micromorphological samples from the burnt remains of the experimental house.



Changing Perspectives – Stable Research Questions 

 57

bucket flotation in the 2013 and 2014 seasons (see above, p. 52). The programme produced a small macro-botanical sample studied by Dr. Galyna Pashkevych with the help of a x25 microscope (see Chapter 5.4). Flotation also produced many small faunal remains from the majority of excavation units, with this difference in retrieval methods integrated into the overall faunal report. A feature of the faunal report concerned David Orton’s elegant solutions to the problem posed by the contribution of several faunal analysts. The lowest common denominator approach focussed on broad questions of subsistence, set in the wider context of Trypillia economy, with particular attention to differences in faunal composition between different features and feature types, and especially to understanding the nature of bone deposition in the Mega-structure (see Chapter 5.3). Question 7: What were the main characteristics of the origins, development and decline of the Nebelivka megasite? Much of the earlier part of this Chapter concerns a theoretical critique of earlier approaches to the origins and development of the megasites (see Section 2.1). The alternatives to the maximalist view have been the development of three interpretative models of the foundation, growth and decline of Nebelivka – the Assembly Model, the Distributed Governance Model and the Pilgrimage Model. Methods for Question 7 The coincidence of the occupation of Nebelivka with a wiggle on the radiocarbon calibration curve meant that we have been unable to provide a tight internal chronology for the megasite. This has led to the alternative approach of modelling of the megasite development. Each of the three models grew out of Project fieldwork results, while at the same time being rooted in recent general discussions of aspects of human interaction and agglomeration. The Assembly Model has been presented in a thematic World Archaeology issue debating ‘Temporary places, gatherings and assemblies’ (Nebbia et al. 2018). The Distributed Governance Model comes out of discussions of heterarchical power relations, as widely discussed in the volume ‘Power from below’ (Gaydarska, submitted), and inspired by the distant links between sites in the Latin American model of altepetl (Hirth 2008). The Pilgrimage Model (Chapman & Gaydarska 2019) draws on recent discussions of assemblies, as well as considerations of pilgrimage in mostly historical contexts (e.g., the Durham Lumbini Project) but also picks up on Loveday’s (2015) formulation of the characteristics of prehistoric pilgrimages. Each model has the initial goal of matching the number of houses built and houses burnt to the known footprint of the Nebelivka site, as well as its estimated duration of more than 150 years or five 30-year generations, and the absence of major human impacts as recorded in the Nebelivka P1 core. All models use a generational framework to calculate the number of houses standing, the number of new houses built, the number of houses burnt and thus the number of houses standing at the end of the generation. Only in the Pilgrimage Model is particular attention paid to the initial two-year period of intensive construction. Falsification of the model comes

58 

 Theory and Practice for Trypillia Megasites

when the projected house statistics fail to meet the Nebelivka footprint (as happened with one rejected version of the Assembly Model) (see Chapter 6.1). Question 8: How do the Trypillia megasites fit into an account of comparative urbanism at the global scale? In Section 2.1, we introduced the comparative study of urban origins elsewhere in the world and critiqued the over-reliance on check-lists in the usual comparative studies. The first approach concerns the relational approach to cities, which has been proposed by Gaydarska (2016, 2017). Here, the main point is the difference between a claimed urban site and other, smaller and preceding or coeval sites in the vicinity. While in the classic high-density urban sites with well-developed hinterlands, the relationship between the core site and the peripheral settlements can be characterised in hierarchical terms, this may not be the case with big, anomalous sites. The relational approach is at the cutting edge of urban research. Our early comparative studies of the megasites showed that megasites were very different from the classic, high-density urban agglomerations which dominate the urban narrative (e.g., Uruk, Babylon, Rome, Byzantium). This was hardly a revelation but what was interesting was that, around this time, other anomalous sites akin to the megasites were becoming the focus of attention (Fletcher 2012; Fletcher & Kim, submitted). This disparate collection of sites, initially termed 'Anomalous Great Sites', had little in common but their extreme size and low-density nature, exactly two defining features of the Trypillia sites. The only useful comparative exercise involving the megasites has therefore been a comparison with other low-density sites. Methods for Question 8 The first step in finding the place of Nebelivka, and Trypillia megasites more generally, within the global development of settlement forms was to establish whether or not Trypillia megasites could be classified as ‘urban’ in a relational sense. This meant the development of a relational framework for the study of megasites so as to provide an archaeological equivalent of ‘measurement’ as suggested by Cartwright and Runhardt (2014) for the social sciences. Their approach to measurement in social science has three interlinked components - characterization, representation and procedures. ‘Characterization’ involves defining the quantity or category that is to be measured using four principles: it needs to be useful for the purposes of the enquiry, it is socially constructed, it does not have rigid boundaries but it is also not too general. ‘Representation’ is the way the category/quantity is represented, with procedures to describe what needs to be done to measure the category/quantity successfully. They both can be expressed either in absolute numbers or as measurements according to scale. The application of this approach in archaeology is of course, not straightforward and needs adjustment. Thus, the characterization of the category ‘urban’ in the Trypillian context is not just a word or a phrase but a range of locally specific and interlinked factors. The second step is the investigation of the similarities and differences between the Trypillia megasites and the other ‘Anomalous Great Sites’. A structural comparison



Changing Perspectives – Stable Research Questions 

 59

between these sites identified seven overlapping characteristics, suggesting that such sites are structurally less random as may look at first sight. A proposal is made that all members of this category of sites should be called ‘megasites’.

2.2.1 Summary and Assessment By the end of the Project, the approaches and methods in use had moved a long way beyond those envisaged in the Project application. Many initial methods worked well without any change (e.g., the geophysical investigations). The application of one fieldwalking sampling strategy was changed as an outcome of early results: in the intensive, systematic fieldwalking programme, overall landscape coverage shifted to a focus on the stream valleys where a high proportion of sites was located. Other methods worked well after minor changes made to adapt the method to Trypillia circumstances (e.g., the Mont Beuvray pottery processing system, with its use of the sherd as the basic unit of analysis, was adapted to the vagaries of Trypillia fine and coarse wares). Yet other methods of which we had high hopes worked less well and were abandoned (e.g., the use of CORONA satellite imagery proved unsuited to heavily cultivated post-Soviet landscapes, and was replaced by another satellite package). The method which had seemingly offered great potential for obtaining samples for AMS dating from hundreds of burnt houses (viz., daub coring and the retrieval of well-preserved charred cereal grains) failed because the grains were burnt out in the high-temperature house fires or destroyed by pedogenic processes; this method was replaced by the test pitting approach, which reduced the number of houses under investigation and therefore the number of samples recovered for AMS dating. Despite the widespread use of flotation, poor recovery of charred botanical remains continued to limit our understanding of Trypillia agriculture, leading to the question of whether the finding of so few cereal remains was indicative of the nature of Trypillia crop processing and discard practices rather than the failure of the flotation method25. Our understanding of the excavated remains of burnt houses was so limited that it could only be mitigated by the innovation of an experimental programme  –  the biggest methodological change introduced in the middle of the Project. We are sure that these successes, adaptations and failures are typical of any largescale Project. But the challenges of the failure of a method with which we had started with great confidence proved to instigate some of the most valuable methods by the end of the Project. In the next three chapters, we present the principal results of the application of these methods.

25  One approach not tested at Nebelivka has been utilised at Majdanetske: phytolith analysis has provided valuable details of crop-processing and the use of cereal by-products, although no information was gained on the scale of arable intensity (Kirleis & Dal Corso 2016, pp. 201–204).

Marco Nebbia & Joe Roe

3 Landscape Studies In this chapter, Marco Nebbia presents the multi-scalar results of landscape investigations ranging from the critical analysis of Ukrainian heritage data for Trypillia settlement, through the interrogation of satellite images for a 25km zone around Nebelivka, targetted fieldwalking in the 25km radius and systematic, intensive fieldwalking in the 10-km territory to site-based gridded collection at four Trypillia sites. Neither of the principal two features visible on the satellite images could be closely dated – a palaeo-hydrological network of small, anastomosing streams and a spread of burial monuments (barrows, dating from the EBA to LIA) which proved to be the only site form dispersed across interfluvial areas. The fieldwalking programmes yielded a spatially precise sequence of settlement distributions from the Trypillia period to the Modern era which showed remarkably consistent choices of site location close to the confluence of river courses. Dr. Nebbia’s critical ‘cleaning’ of the ‘Trypillia Encyclopaedia’ settlement data produced a secure data set for innovative spatialstatistical analyses of Trypillia settlements, focussing on the Southern Bug-Dnieper interfluve with its concentration of megasites. The paucity of small Trypillia sites within a 25km radius of Nebelivka showed the value of settlement analysis at the 100-km scale.

Introduction 

 61

Marco Nebbia 3.1 Introduction This Chapter will look into the way the project investigated the wider landscape that surrounds the megasite of Nebelivka. For the first time in Trypillia studies, a systematic analysis of the contextual environment that constitutes the settings of megasites has been planned and carried out. The scope was of course to use a number of tools and field methods that sheds new light on the environs of megasites and how these influenced the emergence, development and demise of these large settlements. The investigation followed a trans-scalar approach where micro- and macro-regions have been defined and then studied, ‘from the air’ by means of remote sensing analysis, and from the ground through intensive and extensive field survey. The data collected and processed have been integrated with the bulk of existing grey literature of known Trypillia sites. Primary field and remote sensing data were combined and used for the first thorough accuracy assessment of the legacy archaeological data, while a new spatial database of known Trypillia evidence has been produced for the Ukraine. This dataset was then studied by means of advanced statistical analytical tools with the aim to identify large-scale settlement pattern changes throughout the five phases of Trypillia occupation. The results of these analysis shed new light on the nature of Trypillia megasites and their possible relationship with the coeval settlement pattern. New insights on the scale of social movements that Nebelivka and other large settlements involved in their development and occupation were drawn upon for the first time in Trypillia studies. Clearly, the benefit of the landscape approach integrated with a small-scale site investigation brought a more nuanced understanding of the function and organisation of megasites. Moreover, new research questions arose from the wider context of small sites, the sites which we shall term “mega”- off sites” and their relevance in the emergence of megasites. Overall, the project benefitted from the multi-disciplinarity of the underpinning methodology.

Marco Nebbia 3.2 Remote Sensing 3.2.1 Introduction In this section, the main contributions of remote sensing to Trypillia studies will be elucidated. A number of available satellite imagery, ranging from 1960s CORONA images to the more recent WorldView-2 datasets, will be evaluated for their potential value in researching megasites, as well as smaller Trypillia (and other periods) features. Results of the photo-interpretation will also be discussed.

62 

 Landscape Studies

Since the 1960s, Trypillia settlements have been mapped from aerial photographs, starting with the military topographer Shishkin’s flights in 1964. Shishkin later surveyed several sites, including the major megasites of Majdanetske, Taljanki and Dobrovodi (Shishkin 1973, 1985). The visibility of archaeological sites and features from satellite images varies according to the shape and size of the site/feature, the nature of the site/feature, the location of the site/feature, the contrast between the land cover of the surroundings and of the site/feature itself, the state of preservation of the site/feature, the spatial and spectral resolution of the image, the time of acquisition of the image and the different satellite sensors (Sabins 2007). All these variables have even more impact on the recognition of archaeological remains that are completely buried under the ground surface, as is the case for Trypillia sites. Hence, two further factors have to be taken into consideration in order to understand what we can see from the imagery, namely the depth of the archaeological horizon and the land cover of the area at the time of the data acquisition.

3.2.2 High-Resolution Imagery In order to detect and map buried archaeological remains such as Trypillia sites, a set of high-resolution imagery was needed, for which an assessment of the potential of various data sets was undertaken. Only high spatial resolution allows for the detection of archaeological features and the so-called “anomalies” indicating the presence of something underground can be of various types; the ways archaeological features (or other phenomena) manifest themselves on aerial images can be cropmarks (when there is a differential growth in the crops), soilmarks (when there is a differential moisture content in the soil), shadowmarks (when the object topographical expression produces a shade) and several others (Wiseman and El-Baz 2007; Lillesand et al. 2014). In the Ukrainian study, two sets of images have been tested and only one of them turned out to be useful for the identification of archaeological features in the study area around Nebelivka.

3.2.3 CORONA Imagery The CORONA was the first American satellite programme, operational from 1959 until 1972, devoted to a photographic surveillance of the Soviet Union, People’s Republic of China and other key areas. The great potential of such imagery in arid and semi-arid areas has been established over 15 years of research; CORONA images have been used with particular success in exploring the archaeology of the Near East, where the remains of tells, flat sites and route ways are clearly visible on the images (Wilkinson 2003; Ur 2003: 2005; Hritz 2014). The three main advantages of CORONA are, 1) the relatively low cost for a vast coverage of each frame (200km ×



Remote Sensing 

 63

15km); 2) the fact that they give a time-depth of the archaeology before recent urban or agricultural encroachment; and 3) the availability of time-series of images which provides the possibility of seeing features that might be masked in different years. This gives CORONA the potential to be the favourite satellite dataset for archaeologists. Nevertheless, they are not always the ideal dataset for archaeological reconnaissance and its potential has to be evaluated for each specific research area. A set of panchromatic CORONA full-frame images acquired in 1967 with a spatial resolution of 2m was purchased for the entire study area in order to acquire suitable quality for archaeological applications (Fig. 3.1).

Figure 3.1: Coverage of CORONA imagery for the study area (by M. Nebbia; copyright: The Project).

The assessment of the imagery potential for archaeological reconnaissance consisted of two tests of the mapping of archaeological features in the Ukrainian steppe; 1) a predictive photo interpretation in terra incognita of the area of interest; and 2) a post-dictive comparison of known sites clearly discernible on other high-resolution datasets, such as WorldView-2 (for the micro-region) and Google Earth sets of images (for the regional domain). The knowledge of the archaeology in the territory assisted in the second postdictive assessment; after checking the visibility of known Trypillia megasites on the CORONA images, it turned out that only one settlement (Nebelivka) out of eight manifested itself on the image as a slight anomaly that can be interpreted as a potential archaeological site.

64 

 Landscape Studies

In conclusion, the contribution to the study made by CORONA imagery is, unfortunately, very limited, and therefore the project necessarily purchased higher resolution, more expensive datasets, such as WorldView-2, for targeted areas.

3.2.4 WorldView-2 Imagery The ratio of cost to coverage always dictates the research strategy. Therefore, a first sample set of images was acquired for the 5km hinterland around Nebelivka. WorldView-2 was judged the best dataset on the market for archaeological application, both in terms of its spatial and spectral resolution, and its ability to overcome the above-mentioned visibility limitations (Fig. 3.2).

Figure 3.2: Coverage of the acquired WorldView-2 satellite images (8-band multispectral 1.85 m and panchromatic 0.46 m) for the Nebelivka micro-region (5 km radius) and panchromatic (0.46 m) satellite image for the Nebelivka macro-region (25 km radius) (by M. Nebbia; copyright: The Project).

WorldView-2 is a commercial Earth observation satellite owned by the company DigitalGlobe26 which provides panchromatic imagery of 0.46m resolution and 8-band multispectral imagery of 1.85m resolution. It was launched in October 2009

26  https://www.digitalglobe.com/about-us/content-collection (accessed 4th June 2018).



Remote Sensing 

 65

and it takes a photograph of any spot on the Earth almost every day. Having a multispectral dataset allows the performance of a number of image processing and feature enhancements, thus stressing the vegetation growth and soil moisture content as key factors in detecting buried archaeological remains. At first, the focus was on the nearhinterland (5km radius) of Nebelivka for which the project purchased a mosaic of multi-spectral and panchromatic images which included a small element (0.034%) of cloud cover. After a pansharpening27 process, a higher resolution of 0.46m has been achieved for the multi-spectral imagery as well, thus producing the best satellite imagery available commercially, in terms of both spatial and spectral resolution. The choice of selecting an area of 25km radius around Nebelivka as the study area has been dictated mainly by cost, as it would have been too expensive to purchase WorldView-2 images for a wider territory. Therefore, a full coverage of panchromatic (0.46m resolution  –  0.001% cloud cover) imagery was purchased for the Nebelivka macro-region (Fig. 3.2)28. The dataset covers a territory where a Trypillia presence is known in the Western part, whereas the South-Eastern area was archaeologically poorly explored. This presented the opportunity to test the potential of the remote sensing data both on an archaeologically known area and in terra incognita.

3.2.5 Mapping Archaeology in Ukraine The feature mapping process has been developing since the beginning of photointerpretation analysis, and new ways of storing mapped data have been theorized and applied. The best toolkit to manage spatial data are GIS platforms and their embedded geodatabases, which develop from a bespoke data model for the topological nature of the information. A geodatabase customized for remote sensing analysis has been structured so that every useful piece of information regarding the buried features could be stored and represented geometrically as a point, a polyline or a polygon. Each object is represented as a polyline, because what is being mapped is conceived as the interface (therefore a line) between the anomaly produced by the object and the image background.

27  Pansharpening is the process of combining a high-resolution panchromatic image with a lowresolution multi-spectral image in order to obtain a high-resolution multi-spectral image. In most remote sensing software packages such as ERDAS Imagine (used in this study), this procedure is easy, automated, time-effective and accurate. 28  The data purchased were collected at different dates: 30th March 2014, 27th April 2012, 14th April 2008, and 6th September 2011.

66 

 Landscape Studies

The data recorded describe each mapped object with various attributes. All the values of each attribute are predetermined – on the basis of a general knowledge of what kind of information can be extracted from an image – and organized in different lists of close vocabularies, termed Domains, thus preventing redundancies in the database. The intensive analysis of the multi-spectral WorldView-2 image produced over 300 features, mapped using a natural colour display (WorldView-2 multispectral acquired on 17 September 2011) (Fig. 3.3).

Figure 3.3: Distribution of anomalies mapped on the WorldView-2 satellite image, Nebelivka microregion (by M. Nebbia; copyright: The Project).



Remote Sensing 

 67

Figure 3.4: Distribution of anomalies interpreted as traces of a palaeo-channel network, the Nebelivka micro-region (by M. Nebbia; copyright: The Project).

The results showed that 51% of the features were manifested as cropmarks – viz., the presence of buried features restrained the crop growth – and 43% as soilmarks, i.e., the presence of features affected soil moisture. Only 21% of the features have been attributed to an anthropogenic origin, whereas the majority of the rest show an intricate but undated palaeo-hydrological network (Fig. 3.4). Considering only those features relating to potential archaeological sites (which include the Nebelivka megasite, burial mounds and other potential smaller sites), 76% showed up as soilmarks (Fig. 3.5). The rotating agriculture regime allows us to appreciate the different visibility of

68 

 Landscape Studies

Figure 3.5: Anomalies mapped on the WorldView-2 pansharpened multispectral 8-band (0.46 m) satellite image, which have been interpreted as having anthropic origins (by M. Nebbia; copyright: The Project).

the same feature under two different land uses;29 therefore, it was possible to state that, in cultivated fields, the archaeological anomalies are virtually invisible. As for the case of Nebelivka, where the site extends across more than one field, it was possible to verify the decrease in the visibility of the anomalies indicating the Trypillia settlement on the cultivated areas. Moreover, the analysis of the archaeological deposit depths across the site shows how the features in the NorthEastern part are more visible because the anthropic horizon is closer to the current ground surface (see below, p. 217 & Fig. 4.47). A combination of tillage erosion and water run-off arguably explains the high visibility of features on the satellite image in

29  This is possible whenever we have two images, taken at different times and covering the same area.



Remote Sensing 

 69

the North-Eastern part of the site, in correspondence with the ground surface sloping downhill towards the river valley bottom (Fig. 3.6). Another WorldView-2 image, taken on 14 April 2008, shows the good visibility of the South-Western part of the site in crop-free conditions, since the features appear as soilmarks. Nevertheless, the higher visibility of the North-Eastern part of the settlement is confirmed, as the run-off effect is the major contributor to the shallowness of the topsoil. For another Trypillia megasite (Perehonivka, Cherkassy region – 7km South of Nebelivka), we have a double coverage of panchromatic WorldView-2 images, one taken in April 2008 and the other in September 2011. Since the whole of the 25ha site lies within a single field, the land use evenly affects the visibility of all the features constituting the settlement (Fig. 3.7). The comparison of the two images clearly demonstrates that, while the April 2008 crop-free image shows the site in its entirety, the September 2011 image does not even suggest the presence of buried features as the field is totally covered in crops. The results of the assessment show the ways in which archaeological remains of Trypillia settlements can be detected and mapped on satellite imagery under very specific geomorphological and vegetation conditions. Moreover, ploughing and natural drainage erosion contributed significantly to archaeological visibility in remote sensing data. This may well explain why historical images like CORONA, which were taken 50 years ago, are not well suited for the identification of Trypillia sites in this area. Paradoxically, erosion through intensive and deep ploughing over the last century did not just destroy the sub-surface archaeological remains, but helped to make them more visible on the satellite imagery. Since the visibility of archaeological features as cropmarks is very poor in this area, the most effective image-processing techniques involved the enhancement of visual capacity for distinguishing features of interest from the background. A number of procedures have been tested and the results showed how a PCA applied to the full spectrum of eight bands slightly enhanced the visibility of most of the features, but did not show more traces than a standard natural colour display. A decorrelation stretch applied to the RGB bands significantly enhanced the visibility of the anomalies, but no extra features could be detected compared to natural colour visualization. Finally, a decorrelation stretch using the near IR – 1 (band 7), red (band 5) and yellow (band 4) wavelengths gave the best results in terms of the visibility of features showing as anomalies. The number of features detected and mapped, however, has not increased, meaning that all the anomalies are potentially visible on the natural colour stretch. This outcome has to do with the particular sensitivity of the near-infrared wavelength to soil moisture level, which is affected by the presence of buried archaeology. The systematic photo-interpretation of the WorldView-2 imagery resulted in different archaeological features being mapped both within the Nebelivka hinterland (5km radius) and within the macro-region (25km). There are relatively similar patterns in terms of object types detected in the two areas, with a significant difference regarding the presence of Trypillia evidence. The majority of the features can be interpreted as traces of an old hydrological system constituted by dry gullies

70 

 Landscape Studies

Figure 3.6: Shallow depth of the top of the anthropogenic deposit on the North-Eastern part of Nebelivka, allowing a better visibility of the anomalies compared to the rest of the site (by M. Nebbia).



Remote Sensing 

 71

Figure 3.7: Two views of the Trypillia site of Perehonivka (BII): clearly visible in crop-free field conditions (upper) and totally invisible when the field is cultivated (lower): WorldView-2 panchromatic images (0.46m resolution), acquired on April 2008 (upper) and September 2011 (lower).

72 

 Landscape Studies

and relict palaeo-channels connected with still active, major rivers (Fig. 3.4). The scenario represented by these natural features suggests a denser network of rivers and streams, which was active in the past. Unfortunately, there is no chronological evidence to date the older features 30, although we can argue that the fact that one of the palaeo-channels runs across the two outer circuits of the site in Nebelivka and that the layout of the dwellings respects its limits, could suggest that it was active during the occupation of the Trypillia megasite (Fig. 4.14 lower). Most of the other archaeological features mapped within the Nebelivka hinterland refer to burial mounds, which can date from the Early Bronze to the Late Iron Age. Burial mounds were preserved differently as they are situated in currently cultivated fields and therefore ploughing activity has levelled some of them out. Their height varies from 0.30m to 4–5m but even the subtler ones can be detected and mapped on satellite imagery (Fig. 3.8). The eroded mound tops reveal the subsurface soil composition which has a spectral signature distinguishable from the background field.

Figure 3.8: Comparison between two extreme examples of barrows as they appear on the WorldView-2 satellite image and on the ground (by M. Nebbia; copyright: The Project).

30  No OSL dating has been scheduled within the project for these features.



Remote Sensing 

 73

Barrows are one of the major categories of archaeological evidence that marks the territory under study at all scales. They are primarily situated on natural ridges or high topographical zones and mostly occupy inter-fluvial areas (Fig. 3.9). The mapping of the macro-region of Nebelivka yielded around 800 anomalies, which can with 95% certainty be referred to as burial mounds of different time periods. Although their shape is quite discernible from the background across different land uses, there is a number of false positives – i.e. anomalies which originate from modern human activities such as spoil heaps of agricultural waste, that resemble the shape and the topography of burial mounds. Trypillia megasites are quite visible on the panchromatic WorldView-2 image, although, as already mentioned, only in crop-free conditions. More Trypillia sites have been mapped in the macro-region and the patterns of their visibility on the imagery are similar to that of Nebelivka, where the parts of the sites lying on the slope towards the river valleys are considerably clearer. Sites like Sushkivka, Yatranivka and Volodymyrivka present a similar layout as Nebelivka, with two outer concentric circuits of dwellings and radial rows leading to the inner open area at the centre; they are all situated in comparable geomorphological locations at the junctions of two river valleys or along a sharp river bend. The run-off areas show the archaeological remains more clearly than the top of the field where the sites are nested. Sites like Majdanetske and Taljanki (the biggest Trypillia megasite) are not so easy to detect on the imagery, probably due to two main factors. The first is that, arguably, the archaeological deposit is deeper compared to the other sites and therefore the presence of buried remains does not affect soil moisture content enough to produce a clear anomaly. Secondly, the simple fact is that one single image cannot guarantee the best condition for the identification and mapping of a feature. The results of the remote sensing analysis show how only 8 out of 24 sites are visible – whether completely or partially – on the images within the macro-region; and only 15 out of 500 sites recorded in the whole Ukraine.31 However, smaller sites like Apolianka can also be mapped if the land use conditions are favourable. This site is an example of a scatter of potsherds and building materials indicating very low-density settlement occupation, where the mean average distance between dwellings is nearly 40 m. From the photo-interpretation, the structures can be detected and mapped individually; therefore, we can argue that the anomalies on the image represent in situ archaeological features. This can be argued also for the megasites where the more regular planning shows clusters of dwelling nested one next to the other, thus revealed on the image as a continuous linear feature. In the case of megasites, the high proximity of archaeological remains produce an uninterrupted anomaly but single structures can still be detected from the surface scatter of potsherds and building materials confined to each structure (Roe, n.d.; see below, Chapter. 3.2.2) (Fig. 3.11).

31  The assessment was performed with Google Earth which provides a wide range of images, taken in different times of the year for different years.

74 

 Landscape Studies

Figure 3.9: Distribution of all the anomalies that can be interpreted as burial mounds mapped within the Nebelivka macro-region, with Trypillia sites by Phase (copyright: The Project).

Fieldwalking 

 75

Overall, the remote sensing analysis has been shown not to be the best method for discovering sites in the Nebelivka micro-region. This is due to several factors, but mainly to the limited availability of satellite datasets. In fact, a wider range of land uses, different times of data acquisition and different sensors could help in overcoming some of these limitations. The tendency is now to use multiple data sources combined in order to produce images containing a high level of information for archaeological applications. Only the availability of more data and a more accurate assessment of the best temporal window for data acquisition can improve the applicability of remote sensing analysis in this area (Aqdus et al. 2012; Agapiou et al. 2013). The great contribution of Very High Resolution (VHR) remote sensing datasets resides in the direct correspondence between the anomalies and the archaeological features, which allows for a better estimation of the site size and occupation density. Individual structures are visible as in situ features, rather than homogeneous halos representing surface and sub-surface scatters of material. Therefore, site layouts, the number of dwellings and their orientation, and site limits can be recorded and estimated more accurately.

Marco Nebbia 3.3 Fieldwalking 3.3.1 A New Methodological Agenda for the Ukrainian Forest-Steppe Since the outset, the research agenda of Trypillia studies has not included systematic field survey investigations. The Ukrainian literature yielded very few reports on unsystematic surveys, carried out mostly as “supplements” of major excavations on Trypillia sites and megasites. Archaeologists mostly relied on local farmers’ knowledge of Trypillia potsherd scatter locations in the fields, and most of the sites were found thanks to sporadic and unsystematic field surveys (in Russian, razvedki). Data recording systems have been developing, and the introduction of GPS has considerably improved the site location process in the last few years. Nevertheless, data collection methods do not follow procedures that are now standard in Western European archaeology and they are still inadequate for the level required by the scientific community. Ukrainian methods of investigations in the field have always been quite traditional and not inclined towards technological and methodological innovations. As mentioned above, Ukrainian archaeologists have been using remote sensing since the 1960s without developing a tailored strategy for the specific case of Trypillia sites. The same conservative approach has been pursued with field survey, so that a new methodological agenda for field investigation was needed. The Danish-Dutch-Ukrainian Dzarylgaz Survey Project (2007–2008) first introduced modern field survey methods into Ukrainian archaeology, including extensive, intensive and systematic investigations in the region around Lake Dzarylgaza

76 

 Landscape Studies

and the hinterland of the Greek site of Panskoe I on the Tarchankut Peninsula, NorthWestern Crimea (Guldager Bilde et al. 2012). Although the project’s main focus was population dynamics and interactions during the Greek period between colonisers and indigenous people, the research adopted a long-term diachronic approach to the study of landscape (Guldager Bilde et al. 2012, p. 13). As for Trypillia studies, no systematic investigation of the landscape has been carried out so far. Therefore, one of the most fundamental impacts of the current Trypillia Megasites Project was the introduction of field survey methodologies into the Ukrainian research agenda, thus refining the understanding of settlement patterns dynamics in the forest-steppe belt in continental Ukraine before, during and after the Trypillia period. Field survey has been conducted for four main reasons: (1) to assess the potential of field survey recovery for detection of Trypillia settlements and establish the intrasite structure of Nebelivka; (2) to establish patterns of archaeological evidence in the off-site domain of a Trypillia megasite, both during the occupation of the settlement and in other time periods; this information will help in understanding the complex processes of site emergence, development and abandonment; (3) at a smaller scale, field survey data have been used to assess archival information regarding site locations and sizes; since the majority of site sizes reported by the literature are based on surface collection, it was vital to cross-check the potential of the surface scatters as proxies for site extent estimation; and (4) to ground-truth the results of remote sensing mapping; it is essential to check a sample of features mapped from satellite imagery in order to establish the reliability of photo-interpretation and to assess the potential of the different types of datasets in the specific territory. In the following sections, these four strategies will be discussed as they have been used in the field, and their contribution to the research elucidated.

Joe Roe

3.3.2 Intra-Megasite Collection In addition to the program of fieldwalking in the landscape surrounding the site, an intensive survey of surface material on the megasite was conducted as part of the project’s initial field season in 2009. The aim of this survey was to collect independent data on intra-site spatial patterning to compare to the pilot geomagnetic survey. Full details of the methodology and results of the results of this surface collection may be found in the Project Archive (https://doi.org/10.5284/1047599 Section 4.7.1) and were also the basis of an undergraduate dissertation (Roe n.d.). The collection was undertaken on the same area covered by the 2009 geomagnetic survey, consisting of approximately 15 hectares on the Eastern edge of the megasite (Figs. 3.10 upper & 3.11). The area was divided into 30 × 30m grid squares but, due to time constraints, only about 85% of the area could be surveyed; squares that

Fieldwalking 

 77

contained geomagnetic anomalies were prioritised. In those squares that were surveyed, a random, timed collection was carried out allowing thirty person-minutes per square. All visible material on the surface was collected, labelled by square, and taken back to the field lab to be classified, counted and weighed. Diagnostic potsherds that were not of a Trypillia type were identified and recorded as a separate class of find. Bone was recorded, but can be dismissed as being unlikely to be of any antiquity. Otherwise, it was assumed that all the material was associated with the Trypillia site. Particular attention was given to accounting for post-depositional factors that may have distorted the surface distribution, namely the depth of soil and consequent biases in dispersion due to ploughing, geomorphological slumping, and surface visibility during the survey itself. We were therefore able to rule out any of these factors having had a significant bearing on the surface collection data. Vast quantities of pottery and daub were recovered from the surface of the megasite. Despite the significant ‘noise’ expected due to ploughing and other taphonomic effects, the distribution of these finds closely matches the sub-surface remains as seen on the geophysical plot (Fig. 3.10 upper). The two concentric rows of houses are clearly visible, as are the loosely-spaced interior houses and, to a lesser extent, the street dividing the rows into ‘Neighbourhoods’. The space between the rows was all but barren, indicating that whatever activities took place there during the use of the site, it left little material trace. The resolution of our surface collection was not sufficient to distinguish individual houses, although it is possible that one conducted using a smaller sampling grid would do so. We were not able to use the surface collection to distinguish burnt and unburnt houses, either by the density of surface material or the proportion of it that was vitrified. Pits were also only evident on the geophysical plot. Still, the macro-spatial layout of the megasite is readily apparent. Interestingly, the two Assembly Houses in this area of the megasite are not at all visible in the surface collection data. Two conclusions may be tentatively drawn from this fact. One is that, despite their size, the Assembly Houses used relatively little daub compared to the average Trypillia house, adding weight to the interpretation that these were relatively light, open buildings. Similarly, the lack of pottery in comparison to houses suggests that they were used in a different way, i.e., not as domestic structures. Although these conclusions, drawn from unstratified and poorly preserved surface material, carry little interpretive weight in comparison with data from the subsequent excavations, they do immediately caution against the assumption that a structure with a “mega” footprint is necessarily substantial in an architectural sense. Other than daub and Trypillia pottery, there were few other finds of interest. Only four ceramic sherds were identified as being non-Trypillia in origin, suggesting that, for the most part, this part of the site had only one phase of occupation, though plough action may also have affected our ability to diagnose sherds on the surface (Taylor, J. 2000). Bone and groundstone was sparse, probably due to poor surface preservation and resistance to ploughing respectively. The relative absence of chipped stone may

78 

 Landscape Studies

Figure 3.10: Upper: Interpretation of the geomagnetic survey, South-East corner of Nebelivka. Visible are parts of the two concentric circuits of houses (by J. Roe); Lower: sherd densities across the surveyed land units (by M. Nebbia).

Fieldwalking 

 79

Figure 3.11: Interpolated contour plot of daub (upper) and pottery (lower) densities by number of fragments (by J. Roe).

80 

 Landscape Studies

be more meaningful. Certainly if there was knapping in this part of the megasite, we would expect to see more concentrated scatters on the surface. Some Trypillia sites can produce large lithic assemblages, but more frequently in the Western part of their range and in earlier periods (Zbenovich 1996, p. 224). The provision of lithics to Cucuteni-Trypillia sites in general is not well understood (Chapman 2002; see below, Section 5.2.5); quite possibly at Nebelivka and other Eastern sites, flint and chert were relatively scarce imports that were not readily discarded, similar to copper. Other miscellaneous finds included a possible figurine and a conical fired clay counter. One interesting isolated find was a fragment of a Greek amphora handle, some 220m from the barrow (see below, Chapter 4.5.2). Scythian and Sarmatian groups on the Black Sea coast and Ukrainian forest-steppe had contact with the Aegean from the first millennium BC and Greek artifacts were frequently deposited in their barrows (Videiko 2008, pp. 207–11). The coincidence of these two observations (the amphora fragment and the possible ploughed-out mound) might therefore allow the tentative dating of an Iron Age barrow in that part of the site. A piece of slag was also recovered from the centre of the area, perhaps suggesting metal production activity of unknown date in the locality. The overall conclusion is that the intra-site gridded collection demonstrated that it is informative of the broader pattern of spatial organisation of a megasite, especially considering the low investment of time and resources required to carry it out. Moreover, the ‘mismatches’ between the surface collection and geophysics, as for example the absence of finds over the Assembly Houses, highlights the complementary nature of the two datasets, and opens avenues for further investigation in the field.

Marco Nebbia

3.3.3 Trypillia Off-Megasite Survey: A Combined Adaptive Sampling Strategy Since neither a systematic nor a designed field survey has ever been conducted in continental Ukraine (especially within Trypillia studies), the project deemed it necessary to plan a sampling strategy targeted at assessing the potential for recovering archaeological sites in general and to investigate the hinterland of a Trypillia megasite (Nebelivka). The sampling strategy is a fundamental step in the research design and key to understanding the data collected, particularly how representative and reliable they are “within the bound of their (researchers’) restricted time and monetary resources” (Binford 1964, p. 427; Redman 1987). The strategy was a combination of informal and formal sampling (Orton, C. 2000, p. 2). The first, informal choice was to investigate a radius of 5 km around the megasite to verify the presence of Trypillia sites and to assess the general site locations for all time periods. The choice was to conduct a block survey covering the entire extent of each field, thus recovering a representative sample of the

Fieldwalking 

 81

landscape. Considering the limited amount of time available, a random coverage of fields with higher ground visibility has been sampled in order to have the most reliable results from a single survey. The rotating agricultural regime guarantees that, every year, every field changes crops; hence, the influence of modern land use on post-depositional processes and preservation of surface scatters is randomized. Therefore, we can consider the choice of surveying one field against another one independent from their influence to the potential of recovering surface finds aside from the current ground visibility at the time of the survey. We can then say that the choice of surveying only the fields with high visibility represents a formal random sampling with the highest potential to recover archaeological presence in a single survey season. Furthermore, this strategy allowed for the assessment of sites recovery on both inter-fluvial and peri-fluvial areas of different slope and aspect. Embracing an adaptive ‘non-site’ sampling strategy (Thomas, D. 1975), where the smallest unit of investigation is the artefact and not the site as whole, the 2009 survey season focussed on collecting all the materials scattered on the surface of the walked field and plotting them using a hand-held GPS device32 (Fig. 3.12). This gave an idea of how the finds were distributed on the surface and, therefore, helped the definition of site from off-site scatters. Students participating in the field season carried out the survey by walking transects across each field. After a first test of different spacing between transects, it turned out that 20 metres is the most cost-effective distance, which allows recognition of the different range of site scatters found in the surveyed area. The definition of site scatter has been broadly discussed since the beginning of modern archaeological surveying (Gallant 1986; Schofield 1991; Bintliff 2000; Waagen 2014), and numerous methods have been established in order to achieve the best results. In the case of Ukrainian ploughed fields, the differentiation of site scatter from off-site distributions of material resulted quite straightforwardly from the outcome of the first season of fieldwalking around Nebelivka. From 2012 onwards, the adopted regular grid allowed for clear recognition of the concentration of potsherds defining a site against the average sherd density per ha. Field 25, for instance, has an expected sherd density of 2.6 per ha, whereas, in the South-Western corner, we identified a cluster of surface material with a density of 60 sherds (88% of the total number of sherds found within the field limits) per ha. Another example is Field 39, where the expected sherd density is 1.6 per ha, and the site identified on the Northern corner presents a density of 32 sherds per ha. If we compare surface finds densities across the 30 different surveyed fields we can clearly see how the four fields containing archaeological sites stand out, thus providing a threshold of 1 sherd per ha for land units containing scatters definable as sites (Fig. 3.10 lower).

32  See https://doi.org/10.5284/1047599 Section 2.2.

82 

 Landscape Studies

Figure 3.12: Fields surveyed with a non-site sampling strategy in 2009, 2012 and 2013, Nebelivka hinterland, to assess the definition of ‘sites’ from surface scatters (by M. Nebbia).

Fieldwalking 

 83

The data collection followed an adaptive cluster sampling (Orton, C. 2000, pp. 34–38). The starting sample unit is the regular transect until a high level of material concentration is found. Thereafter, the walking strategy adapts to the local variability of potsherd density by investigating the adjacent Neighbourhood until an expected density is found again. The localized anomaly in material density corresponds to what we defined as a site. The limited amount of time available did not allow any evaluation of the nature of these sites, although most of them returned mainly pottery and rarely any building material. Sampling strategies in successive years built on this first assessment and aimed at confirming site locations, thus achieving a more accurate estimation of intra-site organization and establishing a methodology to document site size.

3.3.4 Peri-Fluvial Survey Investigations The results of the first season of field survey suggest that a high proportion of archaeological sites sit on riverbanks at the junction of two or more river branches, very close to watercourses, with the inter-fluvial areas mostly free of settlements. The outcome of the first assessment suggested the planning strategy for further investigations of the Nebelivka hinterland. No Trypillia settlements were found within 5 km of Nebelivka, and no Trypillia potsherds were classified as off-site scatters. This result has multiple archaeological implications that will be discussed below (see pp. 106–108). After surveying 2,744ha of ploughed fields within a 40km2 total area in 2012 and 2013, the four scatters defined as archaeological sites all reflect the same geomorphological settings, as mentioned above. Therefore, further investigations were planned along major and minor watercourses, some still active, others dried out and currently used as pathways. Table 3.1 summarizes the main periods and cultural labels that relate to the material collected during the field survey.

Table 3.1: Main periods and archaeological cultures found during the field survey (by M. Nebbia). Period or cultural label

Dates

Bronze Age

3000/2900–1050/1000 BC

Iron Age

1050/1000 BC–2nd century AD

Cherniahov

2nd–5th century AD

Slavonic

5th–10th century AD

84 

 Landscape Studies

Figure 3.13: Areas covered by the field survey in the 2012–2014 seasons (by M. Nebbia).

Fieldwalking 

Figure 3.14: Distribution map of all sites recovered during the 2012–2014 field survey seasons, including the locations of known Trypillia sites in the Nebelivka macro-region (by M. Nebbia).

 85

86 

 Landscape Studies

In 2014 and 2015, we surveyed a total of 158.5km along river courses covering an area of 574ha during a short three-week season, finding 30 sites33 of all time periods (two of which were Trypillia) (Figs. 3.13–3.14). These results showed a shift from a site density of 0.001 per ha in the first season to 0.05 sites per ha in the fourth season. This result confirmed that settlement locational strategies were essentially the same from the Copper Age to the Post-Medieval period. The investigated area included a transect leading from Nebelivka towards the megasite of Volodymyrivka (Trypillia BII), situated to the South-East along the Synukha River. This choice was dictated by the question of looking for Trypillia settlements between two contemporary megasites in an archaeologically understudied area. Furthermore, the territory covering 25 km radius from Nebelivka has been chosen as wider hinterland to be investigated in terms of Trypillia settlement patterns, thus defining the megasite macro-region.34 The macro-region comprises the counties (Ukrainian = oblast) of Cherkassy and Kirovograd and the border between the two crosscuts the study area more or less diagonally from North-East to South-West. This left the South-East quadrant of the macro-region totally within the Kirovograd county, which has never been properly investigated archaeologically. Therefore, the field survey focussed on the right bank of the river Synukha and all the right tributaries, both active and dried out. In this way, the South-East quadrant of the macro-region represents a sample for both completing the knowledge on Trypillia presence and grasping insightful data on long-term settlement patterns trajectories (Fig. 3.14).

3.3.5 Site Sampling Strategy Another goal of the field survey conducted along river courses has been to establish a method of sampling single sites, in order to understand their extension and gain some insights on the internal structure. Given that site scatters are clearly discernible from the background and they are mostly single phase (at least from the surface scatters), a sampling strategy intended to establish the shape, extent and internal organization of the site scatters has been adopted. Very few off-site materials have been found. The survey technique was a mix between extensive and intensive – a combination of extensive survey of long river branches and intensive sampling of each site scatter. The tight schedule and the limited manpower dictated the operative procedures. A first assessment of the scatter extent defined sampling intervals ranging from 20 metres, for small sites, to 80 metres, in the case of 1.8 km long sites located along the lower riverbank. In this way, it is still possible to compare the sherd density between

33  See https://doi.org/10.5284/1047599 Section 2.2. 34  This decision was driven mostly by cost and time factors.

Fieldwalking 

 87

Figure 3.15: Surface material counts by site sampling transects at the newly discovered Trypillia settlement of Kutsa (20 km North-East of Nebelivka) (by M. Nebbia).

sites, by simply reducing the 20m sampling to 40m or 60m or 80m. We walked multiple transects across each scatter and picked up surface material within 3-metre radius samples until two consecutive samples were empty (Fig. 3.15). Samples were located using a hand-held GPS device and the finds database – with information regarding sample number, quantity and chronological horizon, type of material, part of the vessel (for potsherds), dimensions (for building material like daub) and comments for special finds – was merged with the points layer in GIS35. The plotting of samples with material quantities and material types enabled the definition of the edges of the scatter and therefore its shape, permitting the differentiation of core areas with a higher density of material from built-up areas from open spaces, with the additional comparison of the distribution and density of daub against sherds. For some Early Medieval sites, the distribution of metal slag enabled the location of production areas, usually at the downstream end of the settlement. Despite the low percentage of multiphase settlements, in some cases it was possible to detect expansions, contractions and shifts of the settlements through time.

35  See https://doi.org/10.5284/1047599 Section 2.2.

88 

 Landscape Studies

3.3.6 Assessing Site Sizes The second way field survey contributed to Project research was through the assessment of information regarding site size derived from the literature. The major publication – the Encyclopaedia of Trypillia Civilization – contains all the Trypillia sites known in Ukraine compiled by gathering all available information; from unpublished amateur archaeological reports, to sporadic findings by local farmers, to scientific reports of excavated features (Videiko 2004). The sheer diversity of sources raised some issues about the reliability of the information reported. Therefore, it was crucial to make a thorough assessment of the reliability of the information regarding site size, with field survey playing a fundamental role in ground-truthing. Very few Trypillia settlements have ever been excavated in their entirety, mostly because of their massive extent (the principal exceptions are Kolomiishchina I (Passek 1949a) and Ozheve-Ostriv (Chernovol & Radomskiy 2015). Therefore, the estimate of site size has always been derived from the dimensions of the surface scatter of material. Different techniques have been used to measure and to calculate the area of the halo. Generally, people have measured the two diagonals of the surface scatter and then calculated the extent by using the rectangle area formula. This method has been criticised when scholars realized that the shape of Trypillia settlements was oval rather than rectangular. Thereafter, archaeologists started calculating the sizes of surface scatters using the ellipse area formula, thus reducing previous estimations of site extent (Diachenko & Menotti 2012). Nevertheless, all these measurements, whether they were taken using a tape measure or GPS, do not take into account the density of the intra-site features and the effect that ploughing has had on the dispersal of surface material (Haselgrove et al. 1985). The first issue has been addressed in the first season of the project in 2009, when intra-site gridded collection on the megasite of Nebelivka found a good correspondence between surface material and sub-surface features (cross-validated with geomagnetic anomalies). The same results were obtained while surveying other, smaller Trypillia sites where the density of dwellings is lower and therefore each sub-surface structure shows up on the surface scatter quite neatly as a dense cluster of material. Even though we could not double-check the correspondence between surface and sub-surface features, we can argue that, overall for Trypillia sites, the surface material is an accurate proxy of the internal layout of the built-up area, based on the results of the gridded survey conducted at Nebelivka (Chapter 3.2.2) and visual assessment on the ground36. This is probably due to the fact that the deep ploughing reaches the shallow eroded top of the archaeological horizon, which is quite rich in material and therefore the sheer amount of sherds and daub that come to the surface is in strong contrast with the background soil. Therefore, since the micro-halo of surface material of a single buried

36  It was almost possible to see the rectangular shape of the structures from the surface scatter.

Fieldwalking 

 89

Figure 3.16: Sampling transects on the Trypillia site of Volodymyrivka. The colours indicate the quantity of both pottery and burnt daub collected at each sample. The red line shows the actual limit of the built-up area (by M. Nebbia).

Figure 3.17: Material counts of the 6 transects walked on the BII megasite of Volodymyrivka (by M. Nebbia).

90 

 Landscape Studies

structure is generally very minimal, it is possible to work out a way of establishing the edge of the built-up area. This method was first tested on the site of Volodymyrivka, since the South-West limit of the site is quite clear from the satellite imagery and it has already been determined that the anomalies visible on the imagery are in-situ features. By walking six parallel transects from the middle of the site towards the outside, using a 40m spacing, we counted the number of sherds and daub in each sample (3m radius) (Fig. 3.16). Plotting the results in a histogram showed different patterns in the middle of the site and the edge of the site. In the former, six transects crossed alternatively, but not regularly, both structures and open areas, producing an irregular trend in sherd and daub counts. By contrast, the off-site part of each transect saw a drastic drop in material counts and an overall descending trend towards the off-site (Fig. 3.17). The interface between these two trends can be considered as the limit of the built-up area, because, outside the site edge, the amount of ploughed/dragged surface finds gradually decreases further away from the in-situ archaeology. This also means that the overall halo of surface material goes well beyond the site limits, thus leading to a general over-estimation of the site size. After testing this method on two other Trypillia megasites (Nebelivka and Perehonivka) under different geomorphological conditions, it is clear that the over-estimation is not consistent and not dependent only on the size of the built-up areas. Arguably, multiple factors can affect the spread of material on the surface and therefore it is hard to define a fixed percentage of overestimation. In conclusion, as much as the surface artefacts are a good proxy for the definition of a single dwelling and therefore the internal layout of the settlement, they are not reliable for the estimation of the built-up area without the use of systematic, intensive sampling.

3.3.7 From Space to Field: Ground-Truthing Remote Sensing Interpretations The last contribution of the field survey activity to the project concerns the crossvalidation on the ground of the features detected and mapped from the satellite imagery. When looking at a satellite image, a great variety of features are seen but no one is completely sure of what one is looking at – hence the term “interpretation” of the aerial image. The only secure way of defining the nature of a feature is to go and visit it on the ground. Then, future interpretations of a similar feature on a satellite image may be more accurate. If this rule is valid even for a well-known region, because the same feature can change the ways of manifesting itself on the image, depending on a number of factors discussed in Section 3.2.1, it must be endorsed even more for a completely unknown territory. The latter was the case of the Ukrainian forest-steppe zone of Nebelivka.

Fieldwalking 

 91

Figure 3.18: Upper: Boxplot of site areas reported in the Encyclopaedia by phase, showing the megasites sizes as outliers; Lower left: plot of GINI coefficients of Trypillia site sizes by phase; Lower right: plot of Trypillia megasites vs. smaller sites by phase (by M. Nebbia).

92 

 Landscape Studies

After a first mapping and interpretation exercise, based on the knowledge of fundamental concepts of archaeological photo-interpretation and with the experience of having worked with a range of different datasets, the team needed to check on the ground the number of features detected on the satellite imagery. The ground-truthing exercise served for the assessment of the potential of the different available remote sensing datasets and to understand what has been mapped. Since the choice of the best dataset has already been broadly discussed in Section 3.2, I shall concentrate here on the understanding of what is visible from space. In the 2012 season, ground-truthing was attempted for seven sites identified by Mr. Stefan Smith from the Project’s set of CORONA images. Not one site turned out to be of archaeological significance, highlighting the problematic nature of CORONA images for forest-steppe research. Trypillia sites are only visible in a very limited range of land use conditions, satellite sensors and soil moisture. The fieldwalking results showed that there is a number of smaller and larger Trypillia sites, which are not visible on the World-View images, but very much detectable from surface collection. Nevertheless, when visible, Trypillia sites are quite unmistakable features on satellite imagery. Moreover, there are a number of other features that can be seen on the imagery which depend on understanding the geomorphological characteristics and the general archaeological characteristics of a territory (see Section 3.2.5 for detection of palaeo-channels). Furthermore, some bright patches that look like archaeological sites on the images, and are situated along river courses – like the large majority of sites, turned out to be outcrops of the brighter clayish chernozem C horizon that is exposed by the effect of the constant water and soil erosion. One positive result of this investigation is the great potential of remote sensing for the recognition of burial mounds as characteristic archaeological features of the study area. After the first season of field survey, it was clear how mounds of different diameters and heights (ranging from 0.3m to 5m) are highly visible on satellite imagery, both multispectral and panchromatic. This result yielded a database of nearly 800 barrows mapped within the macro-region (Fig. 3.9). Of course, the limitation of remote sensing does not allow for an estimation of the real height of these features, but it helped enormously, during the second season, in targeting specific field visits aimed at the recording of even very subtle mounds, almost invisible from a ground perspective. In summary, the field survey represented one of the key aspects of the field methodology developed within the Project and its results contributed enormously to a number of research aspects. The project benefitted from the field survey in four principal ways: 1) in establishing the layout of Trypillia megasites, 2) in establishing improved ways of determining site size, 3) in assessing site densities within the Nebelivka micro-region and a sample of the macro-region; and 4) in establishing the immediate hinterland of a megasite such as Nebelivka and the implications that this information had for the understanding of the settlement’s formation and development.



GIS Settlement Patterns – Trypillia 

 93

Marco Nebbia 3.4 GIS Settlement Patterns – Trypillia 3.4.1 Introduction The theoretical definition of megasites has neither been changed nor updated since the 1970s, when the first complete geomagnetic plans were produced (Shmaglij 1980). Since then, scholars have termed as ‘megasites’ those settlements whose areas extend beyond 100ha (Fig. 3.18 Upper) and whose layout is defined by concentric circuits of dwellings, radial rows of structures and a more or less extensive empty space in the middle (Chapman et al. 2014a; Rassmann et al. 2014; see above, Chapter 2.1.2). Further insights into the internal structure of megasites developed out of the “second phase of the methodological revolution” in Trypillia studies (Chapman et al. 2014b), which contributed to the discussion on how social structure was materialized inside these large settlements, thus providing new understandings within the wider discussion regarding their possible ‘urban’ nature. For many years, the focus of the research remained at the megasite level, trying to gather more information and data for the whole settlement. Only recently have archaeologists started to look at these large settlements in the wider context of the Southern Bug-Dnieper interfluve ‘system’ or “Western Trypillia Culture” (WTC) (Diachenko 2010). Diachenko modelled movements and migrations within the Southern Bug-Dnieper interfluve – though including a limited number of sites in his network analysis – in an attempt to revise the chronological sequence of megasites occupation (Diachenko & Menotti 2012, 2015). Manzura studied Cucuteni-Trypillia settlement dynamics within the framework of the “colonisation” of the North Pontic steppe territory, but referred to megasites (or “super-centres”) as evidence of a shift from an egalitarian tribal system towards a more complex societal organization in control of restricted resources from intensified production and exchange (Manzura 2005). Technological and methodological advances improved our understanding of the layout and internal structure of megasites, but what Manzura and Diachenko demonstrated is the importance of considering the megasite phenomenon in the broader context of coeval sites, rather than considering it as a separate social development. Unfortunately, we have not yet reached the same level of detail for the other Trypillia sites as we have for the biggest of the megasites. Nevertheless, it is worthwhile including in the research agenda the final database of sites derived from ‘cleaning’ and selecting the information contained in the Encyclopaedia of Trypillia Civilization (Videiko 2004). This hints at a new definition of megasites based on their spatial relationship with other settlements. Since the quality of the database is temporally consistent over the five Trypillia phases, and recalling that Cherkassy (where the majority of the

94 

 Landscape Studies

megasites is located) is not the only region with a good dataset, we can assume that diachronic analysis of settlement patterns for the two millennia of the Trypillia period are to be taken as real patterns and not excessively biased by data collection and data quality. The chronological resolution based on pottery typology is 300–400 years (Kruts & Ryzhov 1985; Ryzhov 1990, 1993, 1999, 2000, 2007), the site location accuracy is within 1–2 km, and the site size accuracy is difficult to assess but is random across the sample. A useful heuristic device to be adopted in order to overcome a number of problems in the survey data is a trans-scale strategy where patterns in the data are analysed at both different temporal and spatial scales and where the continuity across scales is considered and respected, even when using scalar categories such as macro-meso-micro to facilitate the analysis (Knappett 2011, p. 10). The scale of analysis helped to control for inaccuracies in the data, just as the final interpretative model has been designed to be dynamic so as to allow the possibility of the inclusion of new data coming from future research. Using an inter-scalar approach, four lines of investigation have been pursued in order to define the spatial relationship and the formation processes of the megasites in the territory of Ukraine: 1. Megasite locational strategies: is there a correlation between the locations of megasites and the locations of smaller Trypillia settlements? 2. Size hierarchies: does the first appearance of megasites in phase BI introduce a level of hierarchy in site sizes and what changes occurred during phases BII and CI as the megasite phenomenon itself developed? 3. Size clustering: are the settlements nucleating with the appearance of the megasites and at what scale is the clustering statistically occurring? 4. Megasite micro-hinterland patterns: what is the settlement pattern within a megasite’s micro-hinterland (5 km)? Overall, a regional and contextual perspective (Kantner 2008) is the only approach that allows for a full understanding of settlement patterns and, most importantly, the underpinning settlement systems of a specific social entity (Flannery 1976, p. 162).

3.4.2 The Data from the Encyclopaedia of Trypillia Civilization The core set of data derives from the publication of all the known Trypillia sites in modern Ukraine up to 2004. The collection of data has been built up since the 19th century when the first register and map of archaeological remains in Ukraine have been compiled. In the early 20th century, an official register of archaeological monuments was collected in 1925 (V.U.A.K) and then updated until 1950, but never published. The first publication of Trypillia sites was in the middle of the 20th century, when Passek listed 94 entries (Passek 1949a; Childe 1951). A decade later, Passek published



GIS Settlement Patterns – Trypillia 

 95

an updated version which included some sites along the Dniester in Moldova, making a total of 125 Trypillia sites (Passek 1961). A few years later, the first broad collection of archaeological sites in the country, including 960 Trypillia sites out of a total 7,000 recorded, was published as ‘Archaeological Monuments of Ukraine’ (Zbenovich et al. 1966). From the 1960s to the early 1990s, a national programme of recording archaeological sites developed a standard protocol of data collection by preparing a document for each site (the site ‘Passport’) including basic information regarding the type of site (settlement, burial mound, surface scatter), period and dimensions. During this period of investigation, many previously unrecorded Trypillia sites were found, although some counties, such as Vinnitsa, Cherkassy and Kirovograd, remained poorly investigated. In 1971, the first map of 171 Trypillia sites was published within the ‘Arheologia Ukrainskoi RSR’ volume (Berezanskaya 1971). Since the fall of the Berlin Wall and the end of the Soviet era, the national programme of archaeological investigations has been decentralized and each county has developed its own plan for site recording and mapping. In 1995, a series of regional maps was published for the counties of Chernivtsi, Vinnitsa, Ternopil, Khmelnits and Odessa. This included the plotting of site locations on a map as well as recording other basic information. From this moment, the development of registers of archaeological monuments has been under the control of each county, with some central monitoring by the Institute of Archaeology in Kyiv. The non-systematic and decentralized way of managing the archaeological heritage has led to a level of uncertainty regarding the number of known Trypillia sites, so that some archaeologists (e.g., Videiko) proposed a total of 1,500, whereas others (e.g., Ryzhov) are more optimistic, with around 5,000 Trypillia sites (Videiko 2004, p. 564). Finally, in 2004, a comprehensive collection of all the information regarding known Trypillia evidence has been published as an edited volume called the Encyclopaedia of Trypillia Civilization (Videiko 2004). The maximum information has been taken from all possible sources and updated with new discoveries from 19 counties, to constitute a total of 2,042 Trypillia entries; however, some Ukrainian archaeologists claim that they are up to 4,400 sites including both Cucuteni and Trypillia sites (Videiko 2004, p. 565). Unfortunately, the information collected since the beginning of the 20th century has not been assessed nor evaluated with field visits or excavations but rather taken as granted and reported while the main focus has been devoted to finding more (and possibly bigger!) sites. Meanwhile, as field methodologies and theories advanced and developed, the information collected since the 19th century has not been double-checked or updated with improved recording procedures. Therefore, the Encyclopaedia represents a massive amount of information compiled using a varied range of methods and field procedures that has produced an uneven and inconsistent dataset.

96 

 Landscape Studies

The information reported consists of: 1) a description of the site location; 2) a brief history of site investigations; 3) an estimation of the size of the site; and 4) the chronological horizon of the material found on site. Unfortunately, these data are not provided for every site – indeed, the information is often piecemeal and we have full and detailed information for only a few entries of the Encyclopaedia. Moreover, the inaccuracies, derived from the adoption of old and, by now generally obsolete, field methods and theories, have been transmitted into the final version of the publication. The compilation of metadata regarding how data have been collected in the field is lacking for almost all the entries. It was, therefore, necessary to evaluate the reliability of the information by trying to understand how people recorded sites in the field and elaborated reports. A long and severe data cleaning process assessed the accuracy as well as the reliability of the information reported and resulted in a strict selection of “usable” data. The entries that have been deemed sufficiently reliable for the research have been plotted on the map. A total of 499 sites (from the complete list of 2,042 contained in the Encyclopaedia) has been considered and mapped with a location accuracy that ranges from few metres to 2 km, depending on the details provided in the descriptions. The overall distribution of known Trypillia sites in modern Ukraine is shown below (Fig. 3.19) (for list, DOI https://doi.org/10.5284/1047599 Section 2.1). The figure itself shows how site densities changes across the whole territory of occupation, and the data cleaning process confirmed that these differences are actually reflecting diverse research intensity rates. In fact, the two counties which are archaeologically best investigated are Vinnitsa and Cherkassy. For this reason, most of the spatial analysis and interpretations have been conducted diachronically and on a large scale. The database37 represents the data that have been used for the research on Trypillia settlement patterns and the table (Table 3.2) here displayed reports the type of information that has been recovered from the Encyclopaedia and how it has been organised. An initial data mining process ‘extracts’ patterns in the data collected and managed, which constitutes the basis for the interpretative model that provides a nuanced explanation of the nature and function of Trypillia megasites.

37  The database can be accessed at: http://archaeologydataservice.ac.uk/archives/view/trypillia_ ahrc_2018/downloads.cfm?group=1244



GIS Settlement Patterns – Trypillia 

 97

Table 3.2: List of all the fields of the attribute table of the Trypillia database (by M. Nebbia).

FIELD

DESCRIPTION

ID

Unique ID identification number.

Name

Name of the nearest village or main watercourse (as it appears in the Encyclopaedia).

Oblast

Name of the county where the site is located.

Region

Name of the municipality where the site is located.

Phase

Trypillia phase attributed to the site.

Area (ha)

Site area as reported in the Encyclopaedia for the majority of the sites and corrected where possible.

A

Boolean value of presence/absence.

BI

Boolean value of presence/absence.

BII

Boolean value of presence/absence.

CI

Boolean value of presence/absence.

CII

Boolean value of presence/absence.

Remote_sensing

Level of certainty for site visibility on satellite imagery (from 0 to 1).

Stage_code

A numerical value for the assigned Trypillia phase.

Location_certainty

Level of certainty for the location assigned to the point (from 0 to 1).

Elevation

Elevation of the point derived from the SRTM data (30m) in metres.

Notes

Notes on the pottery group assigned to the site (if present) and other general notes.

Annotations

Annotation on where the area value has been derived from that specific site.

98 

 Landscape Studies

Figure 3.19: Distribution of the 499 Trypillia sites used in the study, derived from the Encyclopaedia (Videiko 2004) (by M. Nebbia).



GIS Settlement Patterns – Trypillia 

 99

3.4.3 Megasite Locational Strategies: Why Were They Where They Were? A simple plot of all the Trypillia settlement data on a map shows that there is a concentration of megasites in the Southern Bug-Dnieper (henceforth SBD) interfluve, within the region of Cherkassy (Fig. 3.20). A total of 20 out of 23 megasites are located in a megasite cluster (henceforth mega-cluster), whereas 3 are situated in a more marginal38 area spanning the whole Trypillia territory. At first glance, it would seem that the SBD interfluve constitutes the “core” of the Trypillia people, whereas the rest of the sites represent the “periphery”39, and probably that is one of the reasons why archaeologists focussed their attention mostly on the cluster rather than considering the whole volume of available data. Regardless of the nature of the relationship between the two macro-patterns, it is clear that there is a predominant locational strategy for the development of megasites, which prompts the question: why were they where they were? Is there an environmental reason why they developed in that specific territory? Or maybe other explanations are possible? A two-step logistic regression test has been performed on the datasets, based on four comprehensive and general environmental variables; elevation, slope, distance from rivers, and soil type. At first, all Trypillia sites have been tested against the four independent environmental variables and the results showed how their locations are statistically dependent on the variables. Secondly, the locations of SBD sites have been compared against the rest of the Trypillia settlements and the results of the regression show how the settlement strategies inside and outside of the SBD interfluve are the same when tested against environmental variables. Overall, the results of the combined logistic regression test suggested that the settlement strategies that megasites and small sites followed in the SBD territory are to be sought in the social rather than the environmental sphere.

3.4.4 Site Size Hierarchies Most archaeologists have talked about site size hierarchies in the context of synchronous variability in site sizes within the same ‘cultural’ and ‘political’ context, and, in the absence of monumental architecture, took this as evidence for political or social stratification or hierarchy (Creamer & Haas 1985; Earle 1987;

38  Here the term marginal is conceived as “not part of the cluster” and not in terms of overall importance. 39  See Wallerstein (1974) for core-periphery theory and Friedman and Rowlands (1977) for the first applications to archaeological research.

100 

 Landscape Studies

Figure 3.20: Distribution map of known Trypillia settlements, highlighting the location of megasites in the Southern Bug-Dnieper interfluve (by M. Nebbia).



GIS Settlement Patterns – Trypillia 

 101

Gilman et al. 1981; Johnson 1977; Liu 1996; Peregrine 2004; Ellis 1984). More recently, scholars have moved away from this static and consolidated paradigm, and argued for multiple explanations for the formation of site size hierarchies, thus moving beyond a constraining correlation between size and centrality (Flannery 1976, 1976a; Keswani 1996; McIntosh, R. 2005; Parkinson 2002; Peterson & Drennan 2012). The development of alternatives for explaining site size hierarchies is also fundamental for a more thorough understanding of Eastern European settlement patterns (Galaty 2005; Kowalewski 2008), and new theoretical frameworks alongside new methods are needed in order to fully comprehend settlement patterns as more primary data is produced. As Duffy clearly shows in his synthesis of settlement data, there can be a number of different reasons and processes that can lead to the development of a site size hierarchy within the same cultural and political framework (Duffy 2015). In this research, the analysis of site size hierarchies has been used for data exploratory purposes and the definition of different settlement patterns in the SBD interfluve and the rest of the Ukrainian forest-steppe territory, particularly in relation to megasites development. The key questions are: have megasites affected site hierarchies? And if so, how? A series of histograms representing site sizes for the five Trypillia phases indicates how each phase, at different scales40, shows a degree of size hierarchy (Fig. 3.21). The data appear to follow a “Primate Pattern” distribution – a typical indicator of social hierarchy and of regional centralization, where there is a peak for small areas and a long tail of fewer, larger sites (Milisauskas & Kruk 1986, p. 25; Drennan & Peterson 2008, p. 360). It seems that, for each phase, there are one or more very large sites (i.e., the megasites) which stand out from the “expected” unimodal Poisson distribution. Simple visual inspection of the histograms prompts the conclusion that, from a system theory viewpoint, those are the centres of the system (Bentley & Maschner 2003). Drennan and Peterson criticise the fact that this interpretation is generally adopted without any statistical testing of whether the largest values “depart” from the normal Poisson distribution enough to be considered ‘dominant centres’ (Drennan & Peterson 2008, p. 361). The approach followed in this research is to consider broader patterns in the data at different scales across the Trypillia phases. The starting point is to consider the megasites as statistical outliers in the overall site hierarchies. Furthermore, histograms show the gap between the high frequency of smaller sites and the very large ones with the absence of middle ‘tiers’.

40  Xlims and binwidths are optimised and therefore different for each plot.

102 

 Landscape Studies

Figure 3.21: Histograms displaying Trypillia site sizes by phase. Phase A (N=33 sites); Phase BI (N=46); Phase BII (N=176); Phase CI (N=234); Phase CII (N=85) (by M. Nebbia).



GIS Settlement Patterns – Trypillia 

 103

The next step is to assess the effect that the development of megasites had in terms of site size hierarchies. In order to do that, we used the GINI coefficient as a measure of data ‘inequality’ (Gini 1912). The coefficient has been applied in archaeology in recent years mostly to estimate social inequality within a single site or between sites, based on grave goods as a proxy of wealth (Bowles et al. 2010; Windler et al. 2013). In the present case, it has been used to assess the influence of the megasites over size variability/hierarchy in each phase (Fig. 3.18 Lower left). The graph shows how the initial development of megasites during Trypillia BI prompted an increase in size ‘inequality’, thus establishing a hierarchical pattern in site sizes. Throughout Trypillia Phases BI, BII and CI, size ‘inequality’ remains quite stable if not showing a slight decrease, even though the number of megasites increases exponentially and proportionally to the number of smaller settlements (Fig. 3.18 Lower right). This suggests that, during the middle Trypillia period (~ 4100–3400 BC), the appearance and development of such large sites did not coincide with a parallel development of a strong hierarchy in settlement sizes. Regardless of whether site size hierarchies are considered evidence of political organization in the case of the Trypillia, we are not facing a materialization at the regional level of any supra-local political or administrative structure. This pattern is visible at a global scale, but, within the SBD interfluve, there is a cluster of megasites which are not ‘at the centre’ of a traditional hierarchical scenario of smaller sites where sparse ‘central places’ create a landscape of multi-tiered settlements. Instead, the SBD interfluve represents a unique panorama of ‘central places’ whose ‘hinterland’ has yet to be identified.

3.4.5 Spatial Distribution of Trypillia Settlements: Site Clustering and Megasite ‘Centrality’ Following the exploratory spatial data analysis, we can now discuss the distribution patterns of Trypillia sites across the landscape, how they change at different scales and how they changed in time. In order to do that, the followed approach analysed the data by using two basic principles of point pattern analysis (PPA); namely the analysis of the first and second orders characteristics of a given point pattern (Bailey & Gatrell 1995, pp. 32–35). In brief, the first order characteristics describe the average point density (or intensity) and distribution patterns across the whole region under investigation; whereas the second order characteristics describe the density of points relatively to their internal spatial organization, reflecting internal interactions of attraction and inhibition (Bevan et al. 2013 p. 31). Understanding the behaviour of data at different scales and their internal interactions provides insights for the archaeologist who has the major challenge of interpreting these patterns, by providing an explanatory historical narrative.

104 

 Landscape Studies

Figure 3.22: Spatial distribution of Trypillia settlements by Phase: upper left – Phase A; upper right – Phase BI; middle left – Phase BII; middle right – Phase CI; lower left – Phase CII (by M. Nebbia).

The simple visual assessment of the five distribution maps of Trypillia sites shows an increasing level of settlement clustering until phase CI – especially after the appearance of the first megasites in phase BI – and a trend to a more dispersed settlement pattern in phase CII (Fig. 3.22). Moreover, by phase BII, the full Dniester-Dnieper interfluve is occupied by Trypillia settlements and remarkable further expansions are not visible until phase CII. During phase CI (the period of the maximum size of megasites), the level of site clustering reaches its peak, but the areas occupied remain roughly the same. It is only with phase CII – the time of the end of megasites – that a remarkable overall dispersal can be observed and the Trypillia area of influence reaches its maximum extent.



GIS Settlement Patterns – Trypillia 

 105

Bearing in mind the development of the overall site patterns throughout the five Trypillia phases, attention should now be focused on the internal characteristics of the data and the variation of the variable ‘size’ across the landscape. This analysis will provide insights into the scale at which certain phenomena are occurring and will suggest the perhaps surprisingly large distances at which we need to look for understanding the range of Trypillia human interactions. The study of data spatial variability or data spatial autocorrelation was incorporated into archaeology in the late 1980s to early 1990s, when it was adopted to investigate the spatial distributions of tombs (Attwell & Fletcher 1987), settlement locational strategies (Kvamme 1990), the origins and spread of languages (Piazza et al. 1995), and the “collapse” of the Classic Maya ‘state’ (Neiman 1997; Bevan et al. 2013a). The fundamental tenet underpinning spatial autocorrelation analysis is Tobler’s first Law of Geography, which states that “everything is related to everything else, but near things are more related than distant things” (Tobler 1970). On this basis, the analysis of the spatial dependency of archaeological data – in this case Trypillia settlement size – could help in understanding the spatial relationship of megasites with other smaller settlements. One of the most commonly used methods of analysing spatial autocorrelation of a given value is Moran’s I index, developed by Patrick Moran (1950). The index describes the combined behaviour of point location (viz., site location) and point value (viz., site size) and whether the pattern is random, clustered or dispersed. Moran’s index calculates the global spatial autocorrelation of the data, thus measuring the overall clustering of a given point dataset. A further development of the test has been suggested by Luc Anselin (1995), who proposes “local indicators of spatial association” (LISA) in order for a statistical evaluation of the clustering occurring in a local spatial unit, starting from the principle that, if there is no statistical evidence for global clustering, this does not exclude the possibility of local clustering. Moreover, Anselin’s local Moran’s I statistics allows for the differentiation of clusters of low and high values and of spatial outliers. In this way, it is possible, for instance, to identify a high value (in this case a big site or megasite) within a Neighbourhood of low values (viz., smaller Trypillia settlements) and measure the statistically significant scale of the Neighbourhood41. In other words, the data will set the scale at which a site becomes statistically ‘mega’ compared to its neighbouring sites. An incremental Global Moran’s I test has been performed on the dataset by phase, and at ‘incremental’ scales. The results indicate that the optimal distance at which the data show a clustered behaviour is around 100 km for each phase but different levels of clustering are suggested by z-scores. Considering the difference in number of sites and megasites (see Fig. 3.18 Lower) in each phase, the almost consistent optimal

41  In this Chapter, the term ‘Neighbourhood’ is used in its spatial-statistical meaning of the space surrounding a given point or feature.

106 

 Landscape Studies

distance acquires an even higher significance as a common pattern between the three phases that cover almost 1,000 years. We can, therefore, argue that ~ 100 km has a statistical significance in terms of settlement pattern nucleation and will be used in the next set of analyses aimed at showing where these clusters are and how they behave in the presence of megasites. The Anselin’s local Moran’s I test performed on each phase shows how the megasites in Trypillia BI, BII and CI represent spatial ‘outliers’ that are significantly bigger compared to their coeval smaller site outside the SBD territory at the scale of ~ 100 km, which suggests the human scale at which these settlements could be seen as ‘central/special’ places. But what was happening closer to the megasites, within what we can call their micro-hinterland?

3.4.6 Megasite Micro-Hinterlands The reason for investigation of the immediate hinterland of such massive settlements is that of site sustainability. Thus, if we argue for a large coeval population living in megasites (e.g., Rassmann et al. 2014; Müller et al. 2016), they would have exploited the surrounding territory for a number of economic activities. Several scholars have studied the Trypillian economy (Bibikov 1965; Gaydarska 2003; Harper 2011; Nikolova & Pashkevich 2003; Pashkevych 2012; Shukurov et al. 2015; Ohlrau et al. 2016) and all the models propose an agro-pastoral mixed economy based on archaeological evidence. Carrying capacity models, however, propose that in order for such massive sites to be sustained by land products, even in a mixed economy, they needed the help of manuring and ard tillage (Shukurov et al. 2015, p. 280). Ethnographic data suggests that the maximal limit of arable land around a site remains within 1.5–2 hours of walking distance (Jarman et al. 1982, pp. 30–31), which translates to approximately 5 km in terrain such as the Ukrainian forest-steppe. This distance is also the estimate proposed for the size of infields and outfields around Bronze Age tell sites in Mesopotamia (Widell et al. 2013). The most recent carrying capacity model developed by Ohlrau et al. (2016) also proposes a minimal radius of approximately 5 km for arable land around megasites like Majdanetske and Taljanki (with respective sizes of 200 and 320ha). All the models recently developed are based on the contemporaneous occupation of the majority of the dwellings in the megasites (Ohlrau et al. 2016, Table 5; Shukorov et al. 2015, pp. 239–240), thus relying on a maximalist population estimation for each site. Shukurov et al. (2015) argue that even a fertile soil like chernozem needs manuring and ard tillage to support the agricultural regime needed to supply even small sites (10–20ha).



GIS Settlement Patterns – Trypillia 

 107

Therefore, the investigation of the immediate (5 km) hinterland of Nebelivka has been conducted in order to check any archaeological evidence of farming activities and more specifically, manuring practices. Evidence of manuring has been identified in scatters of worn potsherds spread on the fields as part of debris coming from the settlement (Bowen 1962, p. 6; Wilkinson 1982, 1989; Gaffney et al. 1985; Bintliff & Snodgrass 1988; Alcock & Cherry 2004; Chapman et al. 2010). The ability to identify manuring scatters was built into survey methodology at Nebelivka. The first systematic field survey ever conducted in Ukraine has been carried out in the near hinterland of the site of Nebelivka (Trypillia BII), covering around the 42% of the available fields (Fig. 3.12). During the 2009 season, we adopted a non-site sampling strategy (Thomas, D. 1975) in order to assess the definition of ‘site’ derived from surface material. This involved picking up and positioning every single small find with an accuracy of 3 metres in the 55ha coverage. The results show four major surface scatters, which can be identified as archaeological sites (Fig. 3.12) among a huge quantity of off-site material. Of approximately 1,000 potsherds collected in the 30 km2 covered in the megasite hinterland, only a single sherd can be reported as dating to the Trypillia period. The results of the field survey show that there is very little, if any, archaeological material presence in the immediate surroundings of a 238ha settlement with almost 1,500 dwellings. Therefore, it can be argued that this is negative evidence for such intensive land exploitation as proposed by the carrying capacity models developed by Ohlrau et al. (2016) and Shukurov et al. (2015). In particular, there was no evidence for manuring scatters in the Nebelivka micro-hinterland. Finally, looking at the immediate hinterlands of the other megasites in the SBD interfluve, it is noticeable how the absence of other coeval sites is a common characteristic for phases BI, BII and CI (Fig. 3.23). This could suggest that those territories were devoted to farming, although the systematic, intensive survey of the Nebelivka micro-region seems to contradict this hypothesis. More systematic field survey around megasites would help understanding whether this is an isolated pattern or a common trait for these big settlements. As for Nebelivka, further evidence demonstrating a very small impact of the settlement on the local environment derives from the results of the pollen sequence obtained by a core located 250m North-East of the edge of the megasite at the bottom of the river valley (https://doi. org/10.5284/1047599 Section 3.4). The results show how, during the occupation of the settlement, the quantities of cereals are even lower than during either the pre- or postoccupation period of Nebelivka (Albert et al. 2020).

108 

 Landscape Studies

Figure 3.23: 5-km hinterlands of megasites in the Southern Bug-Dnieper interfluve (by M. Nebbia).

Only more data derived from further systematic fieldwalking around other megasites combined with more pollen analysis conducted on core locations near megasites as well as smaller sites could confirm or reject this pattern. The two models developed by Ohlrau et al. (2016) and Shukurov et al. (2015) assume that the single megasite or smaller settlement is entirely occupied at the same time, but this is a ‘megaassumption’ since no-one yet has established a fine-grained chronological sequence of megasite development. Most of the arguments in favour of the coeval occupation of all the dwellings derive from the assumption that the formal layout of these settlements (including the whole range of archaeological features, such as kilns, mega-structures and house rows) has to be developed by a single top-down decision-making process (Müller et al. 2016a). Nonetheless, other possibilities for explaining the formation processes of such structured layouts can be derived from a broader context of Trypillia settlements in general (Nebbia et al. 2018). The extraordinary dimension and density of megasites cannot be really appreciated if we do not include their relative capacity for accommodating people coming from coeval smaller settlements.



Concluding Remarks 

 109

Marco Nebbia 3.5 Concluding Remarks The aim of this Chapter was to explore a range of possible formal analyses of the Trypillia datasets, newly included in the discussion on megasites. The aim was to provide new perspectives derived from the investigation of settlement data from different points of view, which can be summarized as follows: 1. There are no clear environmental-related reasons why the megasites are located in the mega-cluster within the Southern Bug-Dnieper interfluve; hence motivations must be sought in the social sphere. 2. With the development of megasites, there are two clear macro-settlement patterns that are established by the beginning of phase CI, clearly identifiable in the SBD mega-cluster of megasites and the rest of the Trypillia settlement distribution. With the increasing number of sites, including megasites, and the development of site nucleation throughout Phases BI, BII and CI, there is no evidence for the development of a single, structured settlement hierarchy. On the contrary, the megasites stand out as exceptionally bigger than the rest of the Trypillia settlements, without a solid middle tier of sites. 3. There is a global and developing nucleation process of Trypillia settlements during the first four Phases A, BI, BII and CI, whereas the last CII phase is characterized by a return to a more dispersed settlement pattern. On a local scale, the ‘centrality’ of megasites is measured by a statistically significant difference in site size at a scale of approximately 100 km. This figure seems to be consistent for around 1,000 years, without being affected by the increasing number of sites and their development and nucleation. 4. The immediate hinterlands of megasites seem archaeologically ‘empty’ and the human impact on the local environmental very marginal. In conclusion, the great potential of the landscape approach has been shown for the first time in Trypillia studies. The main contributions of this Chapter are both methodological and interpretative: - It represents the first systematic assessment of the potential that remote sensing can have in the recovery of Trypillia sites (mega- and non). - It proposes a new agenda for field investigation in Ukraine where systematic field survey is applied for the first time. - It represents the first attempt to use the remarkable amount of legacy data in the study of Trypillia megasites. - New insights on the immediate hinterland of Nebelivka demonstrated the peculiar nature of such a big settlement that seems to exist within a thinly occupied landscape. The implications become more striking when the exploration of the

110 

-

-

 Landscape Studies

broader hinterland (macro-region) shows how other similarly large settlements are located (and possibly co-existed, see p. 255) within a few kilometres. A remarkable large-scale effect of site clustering alongside the emergence of megasites cluster is highlighted for the BII–CI phases, thanks to the inclusion of a larger settlement dataset. A probable scale (100km) of people’s movements across the landscape is proposed and the implications that this has on the wider interpretation of the nature of megasites will be discussed in depth in Chapter 6.

Overall, the combination of large-scale landscape analysis with detailed site-based investigation can only bring to a wider and more accurate picture the archaeologist is constantly seeking when trying to answer their research questions. In this case, the contribution of the landscape side of the project will be integrated with a fine-grained analysis of the site of Nebelivka to produce a better understanding of the Trypillia megasite phenomenon.

Bruce Albert, Jim Innes, Konstantin Kremenetski, Andrew Millard, Marco Nebbia, Bisserka Gaydarska, John Chapman, Dan Miller, Duncan Hale, Brian Buchanan, Stuart Johnston, Mykhailo Videiko, Manuel Arroyo-Kalin, Tuukka Kaikkonen, Svetlana Ivanova, Stoilka Terziiska-Ignatova, Patricia Voke, Natalia Burdo, Natalia Shevchenko

4 Site Studies

In this chapter, we draw together and integrate the studies relating to the Nebelivka megasite at the site level. Three sources of palaeo-environmental evidence – pollen analysis, soil micro-morphology and molluscan analysis – are used to build up a picture of the landscape on and around Nebelivka before, during and after the occupation of the megasite. The pollen, charcoal and non-pollen palynomorphs from the Nebelivka P1 core provide crucial insights into the unexpectedly low level of human impact on the landscape, which has had such profound effects on our approach to the understanding of the megasite. In a fundamental part of this chapter, Duncan Hale presents the only complete geophysical plan of a Trypillia megasite to date, enabling Brian Buchanan’s analysis of movement in and through the site by Visual Graph Analysis and a series of nested analyses of the social space comprising the megasite – the Quarters, Neighbourhoods and houses. Stuart Johnston summarises the results of the experimental programme of house construction, house-burning and excavation of the burnt house. A lengthy section by Bisserka Gaydarska summarises the results of the Ukrainian-British excavations at Nebelivka. Andrew Millard presents the Bayesian analysis of the over 80 AMS radiocarbon dates for the megasite, while Natalia Shevchenko reports on her analyses of the building materials from the Megastructure.

112 

 Site Studies

Bruce Albert, Jim Innes, Konstantin Kremenetski, Andrew Millard, Marco Nebbia, Bisserka Gaydarska & John Chapman 4.1 Palaeo-Environmental Studies 4.1.1 The Nebelivka P1 Core 4.1.1.1 Introduction The geology of the Nebelivka area consists of Pre-Cambrian granite and gabbro. This bedrock is deeply incised by valleys, including all tributaries to the local stream network. The study is focused in the valley of an unnamed tributary with steep transportational slopes (ca. 30°) on the eastern edge of the megasite. Interfluves are draped with Pleistocene loess of variable thickness – typically 1.5m but deeper in gulley fills. Carbonate-rich soils, including chernozems, developed out of the loess at various points in the Holocene. Modern climate is temperate and moderately continental, with a mean July temperature of 20°C and a mean January temperature of -6°C, mean annual precipitation being ca. 550 mm. Vegetation in the study region is classified as forest-steppe. The interfluves today are largely cultivated with a variety of non-cereal crops, with forested boundaries of large fields also consisting of many introduced species. Primary mixed-oak woodland with Tilia cordata elements is found at the South-Eastern edges of the Nebelivka Cooperative Farm in the direction of Borschova village. The study focuses on Holocene sediments in a six-metre sediment core where pollen is preserved under reducing and neutral conditions, designated P1 and located at 48°38’59.4” N 30°33’38.1” E (Fig. 4.1). The core was taken at the South-East edge of a small basin ca. 200m in length and ca. 80m in width. The topography indicates that the coring site was an alluvial site rather than a cut-off side-channel with fen or bog characteristics. On the West side of the basin, nearest the megasite, slopes approached 30° – the steepest slopes of any area on the edge of the megasite – with more gradual slopes of 15–20° on the East side of the valley. The methods involved in the coring are fully described in Albert et al. (2020).

4.1.1.2 Megasite Human Impacts Investigations of the palaeo-environment at Nebelivka provide an excellent opportunity for testing the human impact of a megasite, the more so since the coring site is only 250m from one edge of the megasite. Previous investigations of Trypillia lifeways (Videiko 2013; Pashkevitch 1997, 2005; Pashkevych 2012; Pashkevitch & Videiko 2006) and their ecological implications indicate an expectation of five kinds of cumulative impact: (1) forest clearance to provide land for intensive or extensive farming, with timber for building the hundreds of houses as well as cooking and heating (Kruts 1990); (2) intensive micro-charcoal concentrations marking the regular



Palaeo-Environmental Studies 

Figure 4.1: Pollen diagram, Nebelivka P1 core (by B. Albert).

 113

114 

 Site Studies

burning of houses at the end of their life-cycle (Chapman 2015); (3) agricultural and/or pastoral indicator species; (4) high soil erosion caused by megasite settlement which would have resulted in high sedimentation rates; and (5) the stress that a supposedly large site population would place on the water supply provided by a network of small streams. The aim of this section is to assess the scale of the human impact identified in the Nebelivka P1 core with a view to adjudicating between the ‘minimalist’ and ‘maximalist’ models of Trypillia settlement and lifeways. The results of previous studies (Albert & Pokorný, 2012) translate into the following generalised predictions for Nebelivka: cereal pollen levels up to 4% (Total Land Pollen, or ‘TLP’) or less would suggest minor extensive cultivation, pollen levels above 5% would be typical of short-fallow agriculture, while total cereal pollen of 10–15% TLP would suggest major, extensive cultivation of a majority of interfluve soils. This third scenario might be an expected result of very high population levels at the megasite, for example, with permanent populations in excess of 10,000 individuals. The three cultivation scenarios would also yield predictors of sedimentological change, in particular erosion.

4.1.1.3 The Age – Depth Model Dating was conducted on eleven samples from the core. An iterative approach was used to focus efforts on the sections most likely to be coeval with the megasite occupation. Where samples contained both fine organic material and plant macrofossils, the fine organic material was dated as the macrofossils appeared to be roots intruding from plants growing at a higher level. Initial samples were sieved to remove macrofossils and the fine organic sediment fraction was dated. As some dates produced by this approach were out of chronostratigraphic order, later dates were measured on ‘pollen residue’: the fine residue was prepared by the Department of Geography, Durham University. All results are calibrated using IntCal13 (Reimer et al. 2013) and reported according to international conventions (Millard 2014). Chronology construction was performed in the OxCal software (Bronk Ramsey 2009a) using the P_Sequence age depth model (Bronk Ramsey 2008; Bronk Ramsey & Lee 2013). Outlier analysis with a prior outlier probability of 10% (Bronk Ramsey 2009a) was used in a preliminary model with low-resolution interpolation to confirm which dates were out of sequence, and those definitively identified as outliers, with posterior probability 100%, were omitted from the final modelling. The chronology was interpolated at 5 mm intervals to match the pollen sampling, and durations and sedimentation rates for each pollen zone were calculated. Although some of the AMS determinations date sediments secondarily re-deposited from the valley slopes, there are sufficient dates in stratigraphic order to broadly align the sediments with the well-dated Nebelivka megasite. It is important



Palaeo-Environmental Studies 

 115

to differentiate between precision and resolution in the dating – between the high resolution of the sampling interval and the low precision of the dating.

4.1.1.4 A Sedimentological Hiatus? The ambiguities in the age-depth model and the possible absence of the predicted environmental impact of human populations led us to consider two hypotheses concerning sedimentation in the small valley where the pollen core was taken: the Hiatus Hypothesis, in which sediments contemporary with the megasite occupation are absent due to an interruption in sedimentation or subsequent erosion, and the self-explanatory Continuous Sedimentation Hypothesis (Albert et al. 2020). The overall conclusion from Loss on Ignition, Particle Size and Carbonate analyses show that there is no positive evidence in favour of a hiatus in core P1. Although we cannot rule out short hiatuses in the core, it is unlikely that there was an absence of sediment representing a continuous period of several hundred years.

4.1.1.5 Assessment of Ecological Impact The general assessment of each of the five kinds of human impacts we expected to find in the Nebelivka P1 diagram – cereals, charcoal, agro-pastoral indicators, erosion and hydrology – is based upon a tabulation of the key depths at which changes can be noted (Fig. 4.2) (for detailed results, see Albert et al. 2020). Forest cover was highest before the megasite occupation, reaching peaks of 55% total land pollen in Zone 1 and 45% in Zones 2 and 3 (Fig. 4.1). There was a gradual, cumulative decline in forest cover with cycles of forest clearance and re-afforestation before, during and after the megasite occupation. Minor episodes of elm decline are dated to the megasite period but both were reversed within a period of decades. Both episodes were of a magnitude found before and after the megasite occupation. In the entire diagram there is no single forest clearance event indicating a massive phase of building and/or burning. The micro-charcoal sequence shows a series of peaks in either two or all three size ranges but nothing on the scale of the major 5210 mm fire event before the megasite was settled. Although cereal indicators pre-date this event, it is likely that it represented a significant opening-up of the Nebelivka landscape for agro-pastoral activities through widespread burning of the primary forest. Age-depth modelling puts this fire event at least 100 years before the foundation of the megasite. Thereafter, the periodicity of minor micro-charcoal peaks before, during and after the megasite period, rarely matched those of Cerealia pollen, suggesting the cause lay in house-burning rather than burning of primary forest.

116 

 Site Studies

Figure 4.2: Human impact proxies, Nebelivka P1 core: pollen zone, pollen zone depth and human impact depth columns show darker shades with increased depth (by C. Unwin).



Palaeo-Environmental Studies 

 117

The age-depth model suggests a long duration of cultivation over at least 500 years and probably 900 years from initial cultivation before the megasite to a maximum of 5.1% cereals during the late megasite occupation. This time-span is much longer than the modelled occupation period of the megasite. After an early Zone 3 spike in Cerealia pollen at 5305 mm, a continuous curve of Triticum, Hordeum and Cerealia pollen lasted well into the megasite period. The sequence of cereal pollen findings indicates variable levels of cultivation rather than increased agricultural intensification, with the poverty of definitive annual weed flora suggestive of a four – five-year fallowing system. There is a decline in cereal indicators post-megasite in Zone 8. Pastoral indicators began to be important before the start of dwelling on the megasite and increased during the occupation. Further increases in Zone 8 indicate the likelihood of greater reliance on pastoralism in the post-megasite period. Identification of sedimentation from A-horizon sources relied on the high levels of the spore Glomus, suggesting that moderate land erosion was a long-term effect of megasite dwelling. Our indicators for palaeo-hydrology reveal a striking pattern of increasing water depth and flow through the pre-megasite occupation, with a fall in water quality and a drop in the water table during the megasite occupation. The megasite demand for drinking water, water for the construction of houses and animal use of rivers and streams must have placed a strain on the relatively small water resources of the Nebelivka basin. We have identified four ‘impact events’ where three of the five classes of impact information changed at the same depth (Fig. 4.2). All but the fourth ‘impact event’ occurred before the occupation of the megasite. The fourth ‘impact event’, in the late megasite occupation, is defined by a combination of a fall in arboreal pollen – mostly Quercus – with the re-start of a continuous curve for Cerealia and Glomus (an erosion and therefore cultivation indicator). One significant conclusion is that the impact events are by no means limited to, or indeed correlated with, the megasite dwelling period, reinforcing the suggestion that there were pre-megasite settlements in the Nebelivka area which have not yet been discovered during fieldwalking. A surprising finding is the lack of synchronisation between the micro-charcoal peaks, of which ten were found in the core, and any of the impact events. Only three microcharcoal peaks were found in layers coeval with the megasite occupation and, of these three, only one peak was found in all three microcharcoal sizes. It is important to underline what is perhaps the most important conclusion of our investigations – namely that there is no sign of major human impact at ANY point in the P1 sequence. The well-preserved pollen, charcoal and NPPs42 permit an assessment of ecological impacts and there is no concentrated impact of all five forms of evidence anywhere in the sequence. Based on the age-depth model, what we find are increases in erosion rates and a decline in water quality during the megasite occupation. Given

42 NPPs refer to 'non-pollen palynomorphs'.

118 

 Site Studies

the variations in population numbers even between the various ‘minimalist’ models (currently modelled at 1,500–4,000), not to mention the ‘maximalist’ estimates of 10,000–46,000, the increase in land erosion rates and the decline in water quality in what is, after all, a small hydrological network, come as no surprise. What is more significant is the lack of sustained high cereal values and high pastoral indicators, as much as the absence of spikes in the micro-charcoal curve, especially at the end of the megasite occupation. It is these absences which suggest that the minimalist models are far more likely to be correct and that it is time to consider them seriously as the ecological impact required by maximalist models is not observed. The second most important conclusion stemming from the acceptance of the age-depth model concerns the high probability that there were indeed mixed farmers living in the Nebelivka area before the settling of the megasite, even though their artifactual footprint has not yet been found. There is ceramic evidence for an earlier (Phase A) settlement only 20 km from Nebelivka. Intriguingly, the persistence and strength of pre-megasite agro-pastoral indicators were comparable to those found in the megasite period. However, the strongest evidence for a major fire event indeed dated to the pre-megasite period, as the probable opening-up of the area for agropastoral activities. Nothing on the same scale as the 5210 mm event has been detected in the micro-charcoal record for the megasite occupation. It seems highly likely that this pioneer clearance event increased the attractions of the Nebelivka promontory for subsequent dwelling (see below, pp. 119–122, for the molluscan evidence for a cleared, grazed-grass promontory).

4.1.1.6 Conclusions Palynology and allied sediment studies at the Nebelivka megasite have enabled the reconstruction of environmental changes in close proximity to this large site. The reconstruction of a moderate impact according to palynology and local sedimentology is in some ways surprising, and indicates that the trajectory towards agglomeration of settlement populations here is not related to agricultural intensification, although natural hydrological conditions were favourable to such an agglomeration in Phase BII. Pre-megasite environmental conditions attest to the creation of a small-scale ‘cultural steppe’ in the process of extensive farming in the Nebelivka area. The Nebelivka core is the first sediment core ever to have been recovered so close to a Trypillia megasite with detailed pollen, loss on ignition, particle size analysis and microcharcoal analyses. It is also the first core from outside major valleys, such as the Southern Bug or the Dnieper, to provide a vegetation history of the loesslands of South Central Ukraine. The eight AMS dates in stratigraphic order provide the basis for a reasonably robust age-depth model, linking the core to the megasite occupation. The moderate human impact recorded at the time of the megasite occupation, combined with the stronger evidence for fires both before and after this occupation, come as a considerable surprise to Trypillia specialists and those interested in early



Palaeo-Environmental Studies 

 119

urbanism in Europe. It should be emphasized that the absence of a major ecological impact at any point in the sequence means that, despite dating issues, the results on human impacts are reliable. Three possible scenarios should be considered in the light of such findings. The first and most obvious implication of these results is that the ‘maximalist’ scenario of truly massive populations in Trypillia megasites – ranging from 12,000 to 46,000 people living coevally at Majdanetske – must be rejected. In the second scenario, Pashkevitch’s (2005) traditional interpretation of Trypillia megasite agricultural practices as extensive, inefficient, low-yield arable farming with a modest level of pastoralism, implied a large area of cultivated land reaching 10 km or more from the megasite and the necessity for wheeled vehicles and animal traction. However, this scenario is potentially fatally threatened by the lack of off-site discard anywhere within the Nebelivka micro-region. The third scenario requires us to reckon with the possibility of smaller-scale permanent settlement or seasonal agglomerations at Trypillia megasites rather than long-term, massive, permanent populations. This challenge requires a new approach to Trypillia taphonomy, house-burning and object deposition (see Chapters 4 and 5). The moderate human impact on the Nebelivka environment and the dispersed nature of burning episodes would support the third scenario of smaller or less permanent agglomerations (for comparison of the three models, see Chapter 6) more than the maximalist scenario or the second scenario of extensive agriculture and pastoralism.

Dan Miller

4.1.2 The Molluscan Evidence Holocene molluscan assemblages relating to the poorly dated period before the establishment of the megasite, are strongly indicative of heavily grazed grassland. Even allowing for the potential complex taphonomy of chernozem soils with their krotovina and other pedogenic processes, and the potential post-depositional mixing of Holocene/Pleistocene strata, the absence of woodland species and the predominance of grassland/steppe taxa is stark in the time interval after the end of the Pleistocene and the start of the megasite occupation. This conclusion may have ramifications for the landscape where the megasite was settled. Although mollusc shell was often quite sparse, a total of ca. 4,000 shells has been recovered and classified (for details, see https://doi.org/10.5284/1047599 Section 3.5). Samples were collected from excavated pits, ditches and house deposits, as well as test pits into the Holocene soils and substrate, and samples from the strata sealed by an Early Bronze Age barrow. Given the low abundance of shells, many samples could be immediately scanned and assessed at the sieving stage. The majority of molluscs has been scanned and examined off-site.

120 

 Site Studies

The dominance of Heliocipis striata and Vallonia sp., in combination with the regular findings of Chrondula tridens and Tuncatellina cf. cyclindrica, indicates a diagnostically steppe-like assemblage in all contexts and periods. In fact, the highest densities of these steppe-grassland species occurred in the buried Holocene soil and underlying layers, suggesting the persistence of steppe biota over the entire examined sequence. Notably absent is virtually the whole suite of ‘classic’ Holocene climaxforest species. Clausiliidae and Zonitidae, which are rare here, are only suggestive of limited shrub-like habitats or isolated stands, and not necessarily woodland per se. Caryichium tridentatum is also present, suggesting some shrub-like/minimallywooded shady areas in the open steppe, but not really a forest-steppe as such. The aquatic species are obviously imported to the site. It should be noted that the freshwater mussel, Unio sp., is not a suitable foodstuff but some shells have been used as temper in ‘shell-tempered’ coarse ware (see below, Chapter 5.1.3). Unio sp. should be considered along with Theodoxus fluviatalis and Pisidium sp., in terms of viable transport mechanics. The possibilities include transport in nets with large fish brought to the site. Importation of reeds and maybe other long grasses such as the Succinea sp. elements found in ‘floor’ and pit deposits could also be considered in relation to the use of hay/straw-like materials in fodder and bedding, and possibly also the presence of herbivore dung, mud, and ‘hoof-trample’. Peaks of Succinea in ‘house floor’ deposits hint that bundles of reeds may have been used in house construction. This opens a new approach to Trypillia house-building, especially in the light of the arched rooves of some Trypillia clay house-models. The highest density of Unio sp. fragments occurs in the daub-rich destruction layers but only a single (terrestrial) shell has been observed as an inclusion in all the available daub sampled or seen on site. Finally, samples from the ditches provide some interesting suggestions for archaeological interpretations. Species variance from the local Pleistocene substrate is suggestive that the Northern Ditch segment was indeed perhaps up to ca. 1.4m deep (Fig. 4.55 upper), and appears to have been an open feature, creating a favourable sheltered habitat for Succinea sp. This model is not inconsistent with additional structural features such as a palisade reinforcing the deep, open ditch. This situation differs from the Southern Ditch segment, where each ditch appears much shallower (maximum ca. 0.4 to 0.7m) (Fig. 4.55 lower) and with an associated fauna that suggests the ‘ubiquitous steppe-grassland’ was uninterrupted, with no evidence for the development of a distinct microhabitat for molluscs. In summary, there were periods in the Holocene when the landscape of the Nebelivka megasite appeared significantly open but it is not yet possible to date these periods to a specific part of the Holocene. The existence of local open conditions would have favoured the selection of the promontory for the location of the megasite.



Palaeo-Environmental Studies 

 121

John Chapman 4.1.3 Summary

In the general, comparative analysis of the landscapes of Trypillia settlements based upon the four variables of elevation, slope, distance from rivers and soil type, the conclusion was reached that there was no significant difference between the environs of megasites and smaller sites (see above, p. 99). In terms of human practices, this may be taken to mean that Trypillia groups of all sizes had a clear concept of the kind of site which they preferred to inhabit – a finding supported by the Project’s fieldwalking programme (see above, Chapter 3.2). It is noteworthy that Nebelivka fits precisely into that ‘template’. But what were the characteristics of the actual landscape before the megasite was settled? Two different views emerge from two kinds of evidence. The picture of the wider pre-megasite landscape of a pollen catchment of perhaps 2–3km radius is of a wellwooded landscape, with arboreal pollen of over 50%, which was dramatically altered by a major ‘fire event’ a century or more before the megasite occupation. This fire event is thought to have created a small-scale ‘cultural steppe’. Thereafter, a series of clearance episodes and re-afforestation phases indicated a cyclical pattern of landscape modification throughout and indeed after the megasite occupation. The more localised picture from the molluscan evidence collected from the megasite only showed that, at some time in the Holocene, the Nebelivka promontory was a heavily grazed steppe grassland, with a complete absence of woodland species and the occasional indications of shrub-like habitats and isolated stands of trees. There is no close dating of the significantly open landscape, which may well date to long before the pre-megasite ‘fire event’. It is not surprising that molluscs characteristic of open habitats continued all through the megasite occupation. But we cannot currently document a continuously open landscape on the Nebelivka promontory from the Early Holocene through to the megasite occupation. There can be little doubt that the Nebelivka area was an open steppe landscape at the end of the Pleistocene and that increasing Holocene temperature and precipitation would have stimulated the re-afforestation of this landscape into a forest-steppe, as supported by the pollen data. The extent to which the poorly dated molluscan evidence can give us a picture of an open landscape shortly before the occupation of the megasite is hard to assess. One way to resolve the variance in the two views of the Nebelivka landscape is to invoke the notion of scale. It is feasible to have a more open Nebelikva promontory on a local scale (molluscs) at the same time as a largely wooded catchment around the promontory at the landscape scale (pollen). This account is reminiscent of the ecological term ‘wooded grassland’ (Agnoletti & Emanueli 2016, pp. 78–79) rather than the open natural steppe of Gradmann’s (1933) Steppenheide hypothesis – a human creation of a parkland which required frequent small-scale burning to prevent its reversion into woodland. Such a parkland may well have included the largely open Nebelivka promontory. A wooded grassland would

122 

 Site Studies

have provided the Nebelivka inhabitants with an open space for their megasite and plentiful cleared croplands as well as the forest resources vital for house-construction, house-burning and everyday firewood.

Duncan Hale 4.2 Geophysical Investigations and the Nebelivka Site Plan 4.2.1 Introduction Between 2009 and 2013, a team of archaeological geophysicists from Archaeological Services Durham University completed a detailed magnetometer survey over the Nebelivka megasite, creating the first virtually complete geophysical plan of a Trypillia megasite (Chapman et al. 2014a). The only parts that could not be surveyed were small wooded areas along the extreme Eastern edge of the site. The survey covered approximately 286ha. The results are shown both as a plan of magnetic gradient data (Fig. 4.3) and an interpretative plan (Fig. 4.4). The vast majority of magnetic anomalies reflect the remains of burnt, partly burnt or unburnt buildings; other features include soil-filled pits, ditches, palaeo-channels and possible kilns. Almost 1,500 buildings have been identified in the geophysical survey, all but 23 of which are believed to be dwelling houses. Three-quarters of the structures appear to have been burnt at the end of their ‘use-lives’, with one-quarter unburnt. The Nebelivka megasite conforms to the five key planning principles established by Ukrainian archaeologists in the first methodological revolution in the study of Trypillia megasites (Dudkin & Videiko 2004; Videiko 2012, 2013): 1. at least two, and possibly as many as four, principal concentric circuits of structures 2. an open space in the centre of the site, inside the inner circuit 3. an open space between the two circuits, constituting a buffer zone of varying widths 4. the construction of some structures inside the inner circuit 5. the construction of some structures outside the outer circuit However, the second methodological revolution (Chapman et al. 2014b), exemplified by recent international projects such as the Kyiv-Durham and Kyiv-FrankfurtKiel collaborations, has enabled the identification of several additional planning principles and classes of feature (Fig. 4.4). These recent projects have used new generations of magnetometers to cover huge areas with higher sampling densities and greater spatial precision. Several new elements have been identified at Nebelivka: perimeter ditches, internal ditches, palaeo-channels, pit clusters, household clusters (‘Neighbourhoods’), bounded unbuilt spaces and buildings which are much larger



Geophysical Investigations and the Nebelivka Site Plan 

 123

than typical dwelling houses (‘Assembly Houses’43) (Figs. 4.8–4.9). Many elements of the Nebelivka site plan have also been identified at other megasites, such as Majdanetske and Taljanki (Müller & Videiko 2016). These new elements reveal a far greater degree of internal spatial ordering than was ever detectable on the older plans. The result is that we have begun to understand much more clearly the spatial components of megasites and their combinations and recombinations in ‘Neighbourhoods’ and ‘Quarters’ (Chapman et al. 2014a, 2014b). A Neighbourhood has been defined as a minimum of three structures in a group with a gap at either end separating the unit from adjacent Neighbourhoods (Chapman & Gaydarska 2016). Application of this definition has identified 150 Neighbourhoods at Nebelivka. The number of structures in a Neighbourhood ranges from three to 27, with a strong trend towards fewer instances of the larger Neighbourhoods: over half of all Neighbourhoods comprise only three to seven structures. The locations of the Assembly Houses, as focal points for small communities (described below), are considered to be of critical importance for the spatial division of the house circuits and inner radial streets into a level of spatial order termed Quarters (Fig. 4.5). The plan of the settlement can thus be examined on four levels or scales: the entire site, Quarters, Neighbourhoods and individual features, such as houses and pits. Eight criteria have been used to partition the Nebelivka megasite into Quarters (Chapman & Gaydarska 2016): (1) natural features, such as palaeo-channels (BUT there are only two palaeo-channels); (2) the border half-way between Assembly Houses (BUT this ignores local topographical variation); (3) the boundary between (pairs of) Assembly Houses (BUT sometimes there are three Assembly Houses or only one); (4) any large gaps between Neighbourhoods (BUT there is often a continuous spread of houses (e.g., E–F, F–G); (5) kinks in circuits (BUT these are absent in many parts of the circuits); (6) major variations in the width of the middle (intercircuit) space; (7) gaps in the ditch (BUT some one-third of the outer circuit has no surviving ditch); and (8) ‘obvious’ entrances and passageways (BUT these gaps are not always obvious). Note that no Assembly House has been detected in Quarter N, though it seems likely that at least one may have stood in an unsurveyable part of this area. The judicious combination of as many of the multiple criteria as possible has led to a partition of the Nebelivka megasite into 14 Quarters, labelled A to N (Fig. 4.5).

43  We acknowledge with thanks Tim Pauketat’s suggestion of the term ‘Assembly House’, instead of ‘Clan House’, during the Amerind Foundation workshop of 2014.

124 

 Site Studies

Figure 4.3: Geophysical plan of Nebelivka showing magnetic gradient data (by J. Watson).



Geophysical Investigations and the Nebelivka Site Plan 

Figure 4.4: Interpretative geophysical plan of Nebelivka (by J. Watson).

 125

126 

 Site Studies

Figure 4.5: Interpretative geophysical plan of Nebelivka, showing boundaries of Quarters (by Y. Beadnell on the basis of D. Hale's geophysical plan).



Geophysical Investigations and the Nebelivka Site Plan 

 127

These themes will be explored further below (see pp. 139–140). Suffice it to say here that Quarters appear to have developed in markedly different ways, perhaps indicating a localised, bottom-up decision-making process within overall planning constraints.

4.2.2 The Site Plan At the site level, the outermost feature is a boundary ditch, within which are two concentric circuits of houses (Table 4.1). Further houses inside the inner circuit are typically arranged along radially oriented streets or around open squares. The large central area of the site appears completely devoid of any structures. Occasional larger buildings are present between the house circuits and outside the outer circuit. Probable entrances have been identified in the Eastern and South-West parts of the megasite. The principal features which provided a framework for the whole site were the perimeter ditch and the two house circuits (Fig. 4.4). Between these features were three largely open spaces: 1. the outer space, between the perimeter ditch and the outside of the outer circuit, with occasional Assembly Houses, domestic houses, smaller structures, pits and perhaps gardens 2. the middle space, between the inside of the outer house circuit and the outside of the inner circuit, perhaps an area for garden plots or a processional zone with Assembly Houses 3. the inner space, inside the inner radial streets or the inside of the inner circuit in the absence of radial streets; a major space for congregations, perhaps also used seasonally for animal keeping, whether domestic livestock or wild horses

4.2.2.1 Perimeter Ditch This ditch was detected as a very weak curvilinear positive magnetic anomaly around the edge of the megasite. This type of anomaly reflects slightly higher magnetic susceptibility materials relative to the natural subsoil; these are typically sediments (within features such as ditches and pits) whose magnetic susceptibility has been enhanced by decomposed organic matter or by burning. The anomaly was relatively narrow and discontinuous, with apparent breaks up to 55m across, reflecting a segmentary or ‘interrupted’ ditch. The ditch was typically between 40–70m from the houses in the outer circuit, a considerable distance from the more intense occupation areas, which probably accounts for its relatively weak magnetic signal. It was not possible to detect the ditch around the entire perimeter of the site due to patches of woodland and erosion down a steep slope in the East. Assuming the ditch originally continued around the Eastern side of the site, it would measure

128 

 Site Studies

approximately 5.9km in length and enclose an area of approximately 238ha. Although interrupted in places, some particularly long stretches of uninterrupted ditch were also detected. For example, a length of approximately 720m was recorded along the Northern part of the boundary (Quarters D-F) and another continuous length of approximately 640m was recorded along the South-Western side of the site. The ditch appears to have defined the megasite’s extent rather than served a defensive function, since: 1) there appear to be several, often very wide, causeways across it; 2) on excavation, it was found to be relatively shallow, measuring less than 1.5m in depth and up to 4m in width; and 3) at 5.9km, the perimeter is too long for an effective defence against attack. The ditch is significant, however, in that it demonstrates that this Trypillia community was concerned to define the limits of its settlement, to distinguish ‘inside’ from ‘outside’, an idea previously discussed by Harding et al. (2006).

4.2.2.2 House Circuits Approximately half of all the buildings identified at Nebelivka are arranged in two oval circuits (Table 4.2). Two broad types of rectilinear geophysical anomaly have been interpreted as buildings: intense anomalies (typically in the range -30 to +80 nT), considered to be burnt houses, and weak anomalies (typically +1 to +6 nT), considered to be unburnt houses. The anomalies of the burnt houses reflect rectilinear or rectangular deposits of burnt daub and other fired clay structures such as platforms, benches, ovens or hearths, bins and thresholds. In very rare instances, one end of a building appears burnt while the other end does not. Another category of anomaly comprises those which are considered likely to reflect houses, but which are amorphous to varying degrees, and either weak or strong, or a combination of both (Fig. 4.3). What anomalies in this category have in common is that they reflect discrete magnetic variation at specific locations where houses might be expected, often on ‘streets’ where the other houses are better defined. All three types of house anomaly can be identified on some of the radial streets (for example in Quarter H, Fig. 4.12 lower), whereas almost all of the houses in the circuits appear to be burnt. Table 4.1 provides details of the numbers of all structures (partial/complete houses and Assembly Houses) within each Quarter and in the circuits and ‘spaces’ at Nebelivka (Fig. 4.4). A greater proportion of houses within the circuits are burnt than in the radial streets. For example, 94% of houses in the inner circuit and 83% of houses in the outer circuit are burnt, compared with 63% of houses inside the inner circuit (Table 4.2).



 129

Geophysical Investigations and the Nebelivka Site Plan 

Table 4.1: Numbers of all structures (partial/complete houses and Assembly Houses) within each Quarter and in the circuits and ‘spaces’ at Nebelivka (OC – outer circuit, IC – inner circuit) (by D. Hale). Quarter

A

B

C

D

E

F

G

H

I

J

K

L

M N

Total

Outside Burnt OC Unburnt

1

1

0

1

0

7

4

2

1

2

1

5

0

0

25

0

1

0

1

1

3

1

4

1

1

1

0

0

0

14

Probable

0

5

0

1

0

2

1

0

1

11 14

4

0

0

39

Burnt

8

20

9

27

13 38

37

27

35

21 27

23

17 16

318

Unburnt

0

1

0

3

0

0

1

0

0

3

0

0

0

0

8

Probable

4

19

11

2

0

0

0

0

1

1

2

7

7

2

56

4

3

0

1

1

1

2

1

1

1

1

1

0

4

21

0

0

0

0

0

0

0

0

1

0

0

0

0

0

1

Probable

0

0

1

1

0

0

0

0

0

0

0

0

1

1

4

Burnt

14 38

15

25

7

27

23

33

34

12 38

22

21 27

336

Unburnt

0

0

0

0

0

0

0

1

0

1

1

1

0

0

4

Probable

0

0

6

2

3

0

0

2

0

1

4

0

0

1

19

Burnt

18 47

50

17

14 46

60

24

18

0

16

39

4

24

377

Unburnt

1

1

0

6

5

4

7

2

2

1

4

0

1

36

Probable

11 17

8

22

5

22

14

6

8

1

10

18

10 35

Outer Circuit

Between Burnt OC & IC Unburnt

Inner Circuit

Inside IC

Total

2

187

61 154 101 103 50 151 147 107 103 57 116 124 60 111 1445

Table 4.2: Summary of all structures (partial/complete houses and Assembly Houses) in the circuits and spaces at Nebelivka (OC – outer circuit, IC – inner circuit) (by D. Hale). Outside OC

Outer Circuit

Between OC & IC

Inner Circuit

Inside IC

Total structures

Burnt structures

25

318

21

336

377

1077

74.5%

Unburnt structures

14

8

1

4

36

63

4.4%

Probable structures

39

56

4

19

187

305

21.1%

Total structures

78

382

26

359

600

1445

5.4%

26.4%

1.8%

24.9%

41.5%

100%

130 

 Site Studies

The area of every complete burnt house-plan in the megasite has been measured from the interpretative plan (see below, Section 4.3.1). Partial houses (i.e. of unknown length/width) and Assembly Houses have not been included in these measurements. At 63m2 for the measured houses at Nebelivka, the average house size is very similar to that estimated for the megasite at Dobrovodi (64m2) and slightly smaller than the estimates for Majdanetske (67m2) and Taljanki (71m2) (Rassmann et al. 2016).44 In terms of house density, the area within the perimeter ditch (238ha) minus the central open space (65ha) gives an ‘occupied’ area of 173ha; this provides a density of 8.3 houses per hectare. This is also very similar to the house densities reported at other megasites: Majdanetske – 8 houses/hectare; Taljanki – 7 houses/hectare (Chapman et al. 2014b; Diachenko 2016). Inclusion of the central open space means that the number of burnt houses per hectare falls to 4.5, in comparison with Diachenko’s figures for Taljanki (5.1) and Majdanetske (7) (Diachenko 2016, p. 188). The layout of the two house circuits shows both segmentation and irregularity, with three kinds of layouts: (1) groups of closely-set structures (some of which may even have shared walls); (2) small groups of parallel structures (both (1) and (2) have been termed ‘Neighbourhoods’); and (3) widely spaced structures with perhaps more individual than group identity (Fig. 4.6). Several abrupt kinks or small breaks are evident in both house circuits, where the course of a circuit on one side of the break is displaced either in towards the interior or out towards the perimeter (Fig. 4.7 upper– middle). These shifts in alignment can measure between 25–40m either clockwise or anti-clockwise. There are two instances in the North-West of the site and one instance in the South where slight kinks are present in each circuit at matching locations (Fig. 4.7 upper–middle). These kinks probably relate to different groups of house-builders, perhaps operating at different times. In any event, it appears that the kinks are associated with deviations from a pre-determined site plan, or planning principle, with subsequent adjustments in alignment made as necessary. There are many slight changes in the course of each circuit, resulting in considerable differences in the width of the space between the two circuits, but these variations are typically smoothed out rather than an abrupt kink being created. Indeed, a principal characteristic of the space between the two house circuits is its great variability in width, ranging from 60m in some parts to 160m in the East. This characteristic has also been recorded at other megasites. For example, at Dobrovodi the distance between the two outer house circuits is recorded as about 50–150m (Rassmann et al. 2016), while at Taljanki the distance between the two principal circuits is recorded as 100–150m (Rassmann et al. 2014).

44  If the unburnt houses at Nebelivka are included, the density per ha rises to 6.1.



Geophysical Investigations and the Nebelivka Site Plan 

Figure 4.6: Tightly-spaced and loosely-spaced houses, Nebelivka (by J. Watson).

 131

132 

 Site Studies

Figure 4.7: Upper & middle: gaps and kinks in house circuits; lower: converging inner radial streets, Nebelivka (by J. Watson).



Geophysical Investigations and the Nebelivka Site Plan 

 133

Two wide breaks were detected in both house circuits on the East side of the site (Quarters A–B) and at the South-West (Quarters I–J). These are likely to be the principal entrances to the megasite (Fig. 4.11 upper). The breaks in the house circuits on the East side (i.e. the distances between houses) measure approximately 64m in the outer circuit and 72m in the inner circuit. A particularly large structure (AH1, below) occupies part of the gap within the inner circuit. The gaps between identified houses in the South-West measure approximately 108m (outer circuit) and 84m (inner circuit). There is no evidence for either ditches or palisades towards the sides of these gaps, which would better define the extents of these probable entrances. Extremely weak positive magnetic anomalies detected beyond the outer house circuit at both locations appear to reflect short lengths of perimeter ditch. In the South-West, at least, the ditch does not extend across the whole width of the potential entrance, but has an apparent causeway at each end which could still allow access. Survey coverage of the perimeter ditch at the Eastern entrance was fragmentary and it is not known if the ditch also had causeways here or not. Since the Neighbourhoods within the house circuits are relatively small, with gaps at either end, many of the causeways through the perimeter ditch are aligned with, or are very close to, the gaps between Neighbourhoods. It is likely that some of these will also have served as access routes into the site.

4.2.2.3 Assembly Houses A new category of building has been identified at Nebelivka, distinguished by both its large size and its location relative to standard dwelling houses and to other large buildings. Three such buildings were detected in the initial survey at Nebelivka in 2009 (Hale et al. 2010) and now 23 of these buildings have been identified (Figs. 4.8– 4.9). These large structures are located at varying intervals around the house circuits, often standing between the house circuits or outside the outer circuit. Some of these buildings have been identified in apparent pairs, with one structure located between the circuits and the other located outside the outer circuit. Indeed, five such pairs are located at regular intervals around the South-Western side of the megasite. There are also several instances of single large structures. The buildings are most commonly, but not always, aligned parallel to the house circuits and perpendicular to the dwelling houses. These large structures are presumed to have served as public places, with integrative functions perhaps including meetings and rituals, and have been termed ‘Assembly Houses’ (Chapman & Gaydarska 2016)45. A common characteristic of the Assembly Houses, but relatively rare among the dwelling houses, is the presence of a strong, typically rectilinear, magnetic anomaly (which reflects the bases of the walls) but a noticeable absence of almost any other

45 NB the recent summary of evidence for Assembly Houses (Hofmann et al. 2019).

134 

 Site Studies

Figure 4.8: Assembly Houses 1–11, Nebelivka (by J. Watson).



Geophysical Investigations and the Nebelivka Site Plan 

Figure 4.9: Assembly Houses 12–23, Nebelivka (by J. Watson).

 135

136 

 Site Studies

strong magnetic anomalies (e.g., AH2 in Quarter B (Fig. 4.8) and AH19 & 20 in Quarter K (Fig. 4.9). There are often no concentrations of strong anomalies to indicate the presence of burnt wall, floor or roof remains, in contrast to the association of such anomalies and remains with dwelling houses. If these Assembly Houses had been burnt, and then the burnt remains had been cleared away, one might still expect to see anomalies reflecting the underlying soil which had been burnt during the fire; if the temperature of the underlying soil exceeded its Curie point (perhaps 600⁰–700⁰C), the soil would have acquired a thermo-remanent magnetism which could be detected during survey. Given the absence of such anomalies, it would appear that these Assembly Houses were not burned down in the manner of most dwelling houses. The almost ‘empty’ rectilinear anomalies which were often detected in this survey may instead reflect fired clay slots or troughs used as foundations to support upright timbers or planks directly, or perhaps to hold sleeper beams for upright timbers. Such fired clay slots were found during excavation of the largest Nebelivka ‘Mega-structure’ (see below, pp. 136, 199) (Fig. 4.35 lower); although the Western part of this particular structure was burnt, it provides an excavated, proven, example of the fired clay slots which appear to be present at many of the other Assembly Houses. The discovery of these slots has implications for both the construction and use of the Assembly Houses. In the absence of burnt soil around the slots, it is likely that they were pre-fabricated and fired nearby and then brought to the construction site and installed. Perhaps the walls of these large structures were primarily made of upright timbers, with little or no burnt daub present, consistent with some of the magnetic anomalies. Rather than being burned at the end of their use-life, like the dwelling houses, this type of construction would facilitate the erection and dismantling of the super-structure, and could perhaps indicate a more occasional or seasonal use of these large structures. The relative lack of anomalies could supports the possibility that Assembly Houses may mostly have been unroofed. Additional anomalies were detected within most of these large buildings, almost certainly reflecting internal features. Some linear examples appear to reflect internal divisions or podia (e.g., AH1, and possibly also 3, 8, 14, 19 and 22), while single discrete, small, strong anomalies, invariably located near one end of the building, almost certainly reflect platforms or ovens/hearths (e.g., AH2, 5, 7, 9, 10, 12, 13, 15, 17 and 19: cf. Figs. 4.8–4.9 with Fig. 4.46/3). The details of Assembly House sizes and features are presented below (Table 4.3). The largest Assembly House detected at Nebelivka (AH1) is exceptional in two respects: 1) it is located within a broad gap in the Eastern side of the inner house circuit (Quarter A–B) rather than between or outside the circuits; and 2) it measures approximately 56m x 20m46 (2.5 times larger than the next largest structure). In fact,

46 The original estimated dimensions, based upon the geophysical plot of 60m x 18m, were corrected following the excavation of the Mega-structure.



Geophysical Investigations and the Nebelivka Site Plan 

 137

this ‘Mega-structure’ is currently the largest structure yet to be found on any site within the Trypillia-Cucuteni culture. There is considerable variation in the size of Assembly Houses, with an average size of 276m2 (including the Mega-structure) or 238m2 (excluding the Mega-structure). Assembly House sizes between the house circuits range from 123m2 up to 435m2, with an average size of 282m2, while those outside the outer circuit range from 132m2 to 247m2, with an average size of 174m2. The average size of the Assembly Houses between the house circuits is considerably larger than that of the Assembly Houses outside the outer circuit. The variation in the strength of the magnetic anomalies at the Mega-structure largely reflects the varying amounts of burnt daub and fired ceramics recorded during subsequent excavation in 2012 (Chapman et al. 2016). The Eastern part of the large rectilinear anomaly therefore reflects an open, featureless, enclosed yard or entrance area; parts of the built area indicate a series of rooms, possibly roofed, in the Eastern part, with another open area in the centre and probable further rooms to the West. The Mega-structure is aligned broadly East-West along what appears to be a band of relatively near-surface granite rockhead; one other Assembly House (AH2) sits on top of this same geological feature. This may have been more evident as a ridge before the loess was deposited, giving greater prominence to these two large structures as they are approached through the East entrance into the site. The only other Assembly House that sits on top of a similar presumed near-surface band of granite is AH15 in the West of the site. Table 4.3: Assembly Houses at Nebelivka (OC – outer circuit, IC – inner circuit) (by D. Hale). Assembly Quarter House

Dimensions / Area (located between IC–OC)

Dimensions / Area (located outside OC)

1

A

56 x 20m / 1120m2

at North edge of Quarter A, in E entrance; burnt in W, unburnt in E; internal features in central and W

2

B

21.9 × 13.5 m / 296 m2

burnt; small internal feature near N end

3

B

27.9 × 13.0 m / 363 m2

burnt; internal features

4

C

22.0 × 13.1 m / 288 m2

poorly preserved

5

D

17.0 × 8.6 m / 146 m

burnt; small internal feature near W end

6

D

2

14.2 × 9.3 m / 132 m2

Notes

unburnt

138 

 Site Studies

Table 4.3: Assembly Houses at Nebelivka (OC – outer circuit, IC – inner circuit) (by D. Hale).

Continued

Assembly Quarter House

Dimensions / Area (located between IC–OC)

7

E

14.3 × 8.6 m / 123 m2

8

F

9

F

24.7 × 13.9 m / 343 m2

burnt; small internal feature near NE end

10

G

16.9 × 11.2 m / 189 m2

burnt; small internal feature near N end

11

G

12

H

H

15

I

20.1 × 7.1 m / 143 m2

17.6 × 9.1 m / 160 m2 24.6 × 17.7 m / 435 m

19

burnt; apsidal S end; ?small internal feature at S end

21.1 × 9.1 m / 192 m2

unburnt in N, burnt in S; internal division

24.0 × 14.6 m / 350 m2

unburnt; small internal feature near NW end

19.3 × 13.7 m / 264 m2

18 22.6 × 15.3 m / 346 m2

20

burnt; size of two dwelling houses end-to-end burnt; small internal feature near NW end

14.8 × 9.3 m / 138 m2 K

unburnt

20.0 × 10.5 m / 210 m2

32.9 × 7.5 m / 247 m2

J

burnt; ?internal division

burnt; oval; small internal feature near N end

2

16

17

Notes burnt; small internal feature near W end

13

14

Dimensions / Area (located outside OC)

unburnt burnt; small internal feature near W end

16.0 × 10.4 m / 166 m2

burnt

17.5 × 10.0 m / 175 m2

burnt

21

L

22

L

20.8 × 14.9 m / 310 m

burnt; internal division; small internal features

23

M

21.3 × 10.2 m / 217 m2

poorly preserved

2



Geophysical Investigations and the Nebelivka Site Plan 

 139

4.2.2.4 The Quarters The 14 Quarters at Nebelivka represent an analytical construct defined on the basis of eight multi-dimensional criteria; as such, another analysis may produce a different division of the site into a different number of Quarters. However, constituting a local optimal solution rather than a global optimal solution, the current division is stable and can be operationalized for the purposes of the comparison of the 14 examples. Quarters also form the basis for one of the three megasite explanatory models – the Assembly Model (see Chapter 6.2). Even a cursory glance at the Nebelivka plan will indicate the opposing structural tendencies of the Quarter – an overall similarity of ‘pie-slice’ form to include part of each concentric zone of the megasite, from perimeter ditch to inner open space, which sits in tension with considerable diversity in the form and size of individual Quarters (Fig. 4.5). The principal factor affecting similar overall shape is the presence of the Southern palaeo-channel, which truncates the inner open component of Quarter J by displacing it into Quarter K. In terms of size, Quarters vary from 5.3ha (Quarter E) to 21.8ha (Quarter B), with a mean of 13.1ha. The distribution of Assembly Houses also varies, from none to three per Quarter, although it is possible that one Assembly House was located in what is now wooded, and hence unsurveyed, terrain in Quarter N. The number of Neighbourhoods also varies by Quarter from a low of five (Quarter E) to a high of 18 (Quarter N). This variation is carried over into the number of houses per Quarter (Table 4.1), with a low of 50 houses in Quarter E and a high of 151 houses in Quarter F – a mean of just over 100 houses per Quarter. These basic statistics provide a picture of a key structural element of the megasite concealing massive variation across the plan. The most obvious explanation for such variation is the building-up of the Quarters from the bottom-up rather than as a series of top-down, hierarchical decisions. There would be no bottom-up logic which would have led to an emphasis on regularities in the number of Neighbourhoods, houses or Assembly Houses in each Quarter.

4.2.2.5 Inside the Inner Circuit Many more dwelling houses, amounting to some 600 structures, were also identified inside the inner circuit, comprising 41.5% of the total number of structures at the megasite. Many of these houses (459 or 31.7% of the total) were arranged along radially oriented streets (‘Inner Radial Streets’) with the long axis of each house parallel to the house circuits. As with the houses in the main circuits, it is the gable end of the house that fronts onto the open area or street. There are 45 Inner Radial Streets of four or more houses, rising to 52 if rows of two or three houses are included. The greatest number of houses on one street is 26, in Quarter L, where the street extends approximately 280m towards the interior. Whilst radial streets are present around much of the interior, the highest density of radial streets is in the North-

140 

 Site Studies

west (Quarters E–H) and the lowest in the South-West of the site (Quarters I–K). No radial streets were identified in Quarter J, though this Quarter is constrained by a palaeo-channel. Similarly, only one radial street, of only three houses, has been identified in the adjacent Quarter K. The composition of the streets shows as much variability as that of the circuits, with some houses clustered into Neighbourhoods and some more widely spaced. If the inner open area was a key social assembly place, the radial streets would have defined processional routes towards the sacred centre of the site. Whilst many of the radial streets are parallel to their neighbours, there are also instances where the streets converge, such as in Quarter L, and two instances where two adjacent streets actually merge into one street, in Quarters F and G (Fig. 4.7 lower).

Figure 4.10: Squares and short inner radial streets, Nebelivka (by J. Watson).



Geophysical Investigations and the Nebelivka Site Plan 

Figure 4.11: Upper: megasite entrances and the main palaeo-channel; lower: short inner radial streets and blocking structures, Nebelivka (by J. Watson).

 141

142 

 Site Studies

Although the majority of houses within the inner circuit were arranged along these radial streets, 141 houses (9.8% of the total structures at the site) have also been identified which are not located on radial streets. These houses are oriented radially, that is, with their long axis oriented towards the centre of the site, perpendicular to the houses on the radial streets. Again, there is a great deal of variation in the spatial organisation of these houses. Many of these houses do not have close neighbours. For example, there are individual houses in Quarters B, G and I which are over 40m away from their nearest neighbour. However, many are also grouped into small clusters of parallel houses or Neighbourhoods, as evident in Quarters A, B, G and N, for example. Indeed, four such Neighbourhoods in Quarters N–A are located just inside and parallel to the inner house circuit, resembling a small arc of another internal circuit. Other small groups of radially oriented houses appear to span gaps between radial streets (Fig. 4.10). These short transverse or tangential streets can also have the effect of creating small open ‘Squares’. In Quarter N, many of the internal Neighbourhoods appear to form parts or sides of Squares, enclosed areas which do not contain further houses but may contain pits (Fig. 4.10). In Quarters B & C, there appear to be similar open spaces, but with houses on only three sides. Other short transverse streets appear to block the end of a radial street. For example, in Quarters B, C, D, F and G, groups of between 2–5 houses are located across the inner end of a street, perhaps being deliberately located to mark the end of a street, to define the point at which no further houses are built and prevent further incursion into the central open area (Fig. 4.11 lower). The large central part of the site (approximately 65ha) appears to be devoid of any structures or cut features; this may perhaps have been used as a central congregation area but also as a seasonal area for animal husbandry or the corralling of wild horses. Another feature of the Inner Radial Streets is the occasional occurrence of very short or ‘failed’ streets, streets which were not continued for some reason (Figs. 4.12 upper and 4.13 upper).

4.2.2.5.1 Pits Small, typically weak, positive magnetic anomalies were detected throughout all parts of the site where houses were detected, as well as in a few small areas where no houses were detected. The anomalies generally measure between 3–4m across and reflect small areas of slightly enhanced magnetic susceptibility. They were initially interpreted as soil-filled pits – an idea subsequently confirmed by excavation. Over 850 of these features have been identified at Nebelivka; investigation of a sample of pits has indicated that they contained highly organic deposits, some with little cultural material and others with large quantities of animal bone, pottery and figurines (Chapman & Gaydarska 2016). The majority of these pits are associated with dwelling houses and are located outside one end of a house (Fig. 4.12 lower).



Geophysical Investigations and the Nebelivka Site Plan 

Figure 4.12: Upper: blocking streets; lower: poorly burnt dwelling houses and pits, Nebelivka (by J. Watson).

 143

144 

 Site Studies

Figure 4.13: Upper: linear pits and short inner radial streets; lower: strong anomalies possibly representing ‘kilns’, Nebelivka (by J. Watson).



Geophysical Investigations and the Nebelivka Site Plan 

 145

The placement of pits either outside or inside the house circuits varies by Quarter and even by Neighbourhood in some places. For example, in Quarter B, the pits associated with the outer circuit of houses are located on the outside of the circuit, whereas, in the adjacent Quarters C and D, the pits are located by the gable ends inside the circuit. Where two parallel rows of houses extend radially towards the centre of the site, the associated pits are typically located on the outer side of the house rows, thus leaving an uninterrupted space between the two house rows, perhaps for procession (for example, Quarter H; Fig. 4.12 lower). Conversely, in the apparent house Squares in Quarter N, the pits are generally located within the squares rather than outside. In places where the houses are very close to one another, in both the circuits and the radial streets, the pits appear to overlap or merge, appearing to form ‘linear’ pits (Fig. 4.13 upper). Similar features have also been detected at Taljanki, Majdanetske and Dobrovodi. In some instances, the linear pits extended around the sides of houses. Investigation of one such feature at Nebelivka in 2013 revealed part of the feature to be a linear pit and part of it to be a layer of cultural material (Chapman et al. 2014b). It is believed that the pits were originally excavated to provide clay for use in house construction. In some parts of the site, however, there are groups or lines of probable pits which do not appear to be associated with houses. It seems plausible that clay extracted from these pits was for use in pottery production; such groups of pits could therefore indicate the areas where pottery was being made. In the East of the site, there appear to be pit groups in Quarters L, N and A; other groups of pits have been detected in Quarters F through to I in the West. Future investigations may identify kilns in those areas, though they do not currently correspond to the sample of possible kiln sites below.

4.2.2.5.2 Kilns Given the quantities of pottery found on megasites, it was anticipated that pottery kilns would be present on site. There are many small strong geomagnetic anomalies which could potentially reflect the remains of such features. However, using a combination of criteria including size, orientation of anomaly, strength of anomaly and location, none of the geomagnetic anomalies at Nebelivka appeared entirely consistent with what might be expected of a well-preserved kiln. Many of the anomalies could nevertheless reflect small fired structures or the remains of such features. Several possible kilns were therefore identified and selected as targets for intrusive investigation in 2014 (Fig. 4.13 lower). Although they did not appear to be surrounded by concentrations of fired debris, as might be expected of a kiln, the anomalies could still reflect small ovens or hearths and were considered worthy of further investigation. Though distributed around the megasite, an increased frequency of possible kilns was noted in Quarter H.

146 

 Site Studies

Three anomalies were targeted by the Ukrainian team in 2014 (Fig. 4.13 lower). The first two (in Quarter D) were reported as ‘modern features with numerous iron items’ (Burdo & Videiko 2016), which had given rise to the strong magnetic anomalies. However, the third target (in Quarter C) proved to be a fired structure. This was interpreted as the remains of a pottery kiln by the Ukrainian team (ibid.) and as a communal cooking feature by the Durham team (John Chapman, pers. comm.). The structure comprised a 2 × 2m clay platform with four low walls creating three channels or flues (see below, Chapter 4.7.4 & Fig. 4.58). The structure was located at the end of a very short radial street. Other three-channel fired structures of somewhat different design, interpreted as pottery kilns, have been found at Majdanetske and Taljanki (Korvin-Piotrovskiy et al. 2016).

4.2.2.6 Features Outside the Outer Circuit Of the 78 structures identified outside the outer circuit of houses, eight have been interpreted as Assembly Houses; these are generally located in the West of the megasite between Quarter F in the North-west and Quarter L in the South. A few of the other structures appear to be larger than the average house, many are similar in size to the average house and many more are particularly small features but detected as quite strong magnetic anomalies (Fig. 4.14 upper). The latter anomalies are most apparent in the South of the site, in Quarters J to L; several similar features were also detected inside the inner circuit in Quarter K (Fig. 4.14 lower). Whilst these features are considerably smaller than typical dwelling houses, they appear to comprise similar burnt daub and could perhaps be small workshops or sheds, perhaps storage structures for construction materials such as withies, reeds or chaff (see Section 6.2), or indeed they may have had many other varied functions. There are slight indications in the survey of many more possibly similar small features, represented by very weak magnetic anomalies. Indeed there are a great many extremely weak positive magnetic anomalies between the outer house circuit and the perimeter ditch (Fig. 4.14 upper). These anomalies have been detected around much of the outer space and reflect areas of slightly enhanced magnetic susceptibility, almost certainly variation in the soil as opposed to burnt or fired materials, and could be associated with garden features such as beds and paths.

4.2.2.6.1 Palaeo-Channels Three probable palaeo-channels have been identified in the magnetometer survey; these are natural former stream beds, detected as weak curvilinear positive magnetic anomalies. One of the palaeo-channels in the South of the megasite corresponds to a topographic feature observed on the ground, which extends Southwards into a narrow wooded ravine (Fig. 4.14 lower). This channel forms the boundary between



Geophysical Investigations and the Nebelivka Site Plan 

 147

Figure 4.14: Upper: anomalies outside the Northern end of the megasite; lower: main palaeochannel with kinks in house circuits, Nebelivka: numbers refer to Assembly Houses (by J. Watson).

Quarters J–K. In the geophysical survey, the linear positive magnetic anomaly measures approximately 5m in width, at its widest, and has been detected for just over 400m within the site, extending through both house circuits and into the interior space. The locations of the buildings close to the channel indicate that it was a significant landscape feature at the time of the megasite; however, it is not currently known if the stream was an active water source at that time. On the Eastern side of the palaeo-channel, both the inner and outer house circuits appear to be diverted slightly Northward, to avoid the hollow and accommodate more buildings, rather than simply continuing on their alignment and stopping at the edge of the hollow.

148 

 Site Studies

Further probable palaeo-channels have been detected in the South-East of the site (taken as the boundary between Quarters L–M) and to the West, outside the perimeter ditch in Quarter I. A similar soil-filled feature detected to the North of the megasite beyond Quarter F could reflect another palaeo-channel or ditch. These features are all similar in nature and measure up to approximately 4m in width.

4.2.2.7 Summary The Nebelivka plan is a classic example of cumulative, local decision-making leading to an emergent, regular layout which conforms to recognisable planning principles. Substantial variations can be found in Neighbourhoods and Quarters, the Inner and Outer Circuits, the inner radial streets and the width between the Outer and Inner Circuits, leading to the conclusion that there was a prevalence of bottom-up decisionmaking based on local groups of people. A key, presumably early, construction was the digging of the perimeter causewayed ditch, defining ‘inside’ from ‘outside’ rather than acting as a defensive barrier, providing two main entrances and with the positioning of the causeways often linked to gaps between Neighbourhoods in the Outer Circuit. It is surely significant that the proportion of burnt houses in the Outer and Inner Circuits was much higher than that of the inner radial streets. It is also interesting that one of the newly discovered plan elements – the Assembly House – was not only constructed in a different way from dwelling houses but also was itself differentiated into the larger buildings located between the Circuits and smaller structures built outside the Outer Circuit. The open inner space has been identified as a positive space for congregation – retained as central to Trypillia culture – rather than only as a negative space lacking structures and therefore abandoned to nature (animals – even wild horses!).

Duncan Hale, John Chapman, Bisserka Gaydarska, Marco Nebbia & Brian Buchanan 4.3 Architectural Analyses Duncan Hale, John Chapman, Bisserka Gaydarska & Marco Nebbia 4.3.1 House Size Analysis

The analysis of house sizes based on the interpreted geophysics plan was initiated by Duncan Hale (see above, Chapter 4.2). Because of the uncertainty of the dimensions of the unburnt and possibly burnt houses (coded ‘green’ and ‘purple’) (Fig. 4.4), Hale restricted his CAD-based measurements to burnt houses (coded ‘red’), considering a sample of 36 houses in each of two Quarters (Quarters G and L). In the second analysis, we measured all the burnt houses, which limited the sample size to 1,077 out of a possible 1,445 houses (for the full table of house sizes, see https://



Architectural Analyses 

 149

doi.org/10.5284/1047599, Section 4.6). We do not consider the analysis of the total group of burnt houses useful in our understanding of the development of Nebelivka in the sense that there is a very low probability  –  approaching zero  –  of all of the houses in coeval use. Nonetheless, since the measurements of the burnt houses at Nebelivka constitute one of the largest samples of house dimensions, the analysis has some comparative value. It is our view, however, that house size analyses at the Neighbourhood or Quarter level is more meaningful. The analyses were conducted in three ways: (a) a length by width bivariate plot (Fig. 4.15/1); (b) histograms of house area using ranges of 10m2 for the total sample (e.g., 15–25m2; 26–35m2; 36–45m2, etc.) and ranges of 20m2 for Quarters and Neighbourhoods (e.g., 15–35m2; 36–55m2, etc.) (Fig. 4.15/2–3 & 4.16); and (c) a GINI co-efficient analysis of house sizes by Quarter (Table 4.4). A trial analysis of the second method using slightly different ranges (viz. 10–20m2, 21–30m2, etc.) produced almost identical results, confirming the robustness of the results.

4.3.1.1 The Total Sample The bivariate plot of the sizes of all the burnt houses produced a clear central cluster with outliers in all directions (Fig. 4.15/1). The trendline of this distribution showed a typical length of 7.5m for a width of 4m (a L:W ratio of 1:2), with 15m lengths for 5m-wide houses (a ratio of 3:1) and lengths of 23m for 6m-wide houses (a ratio of 4:1). This shows that Nebelivka houses tended to add ‘modules’ of 7.5m of length for each metre of increased width. However, it is important to note that these are only general trends. The total variation in length for 4m-wide houses is 7m–18m, with 7m–19m for 5m-wide houses and 10m–24m for 6m-wide houses. Any increase in house width would result in a directly proportional increase in the height of the house; a design choice for a more monumental, higher house would have been predicated on selecting a wider house. Needless to say, the reasons underlying individual house sizes are complex, requiring much further discussion (see below, pp. 151–155, 417–418, 422). The plot of the sizes of all burnt houses approximates to a Gaussian distribution with a unimodal peak of 56–65m2 house area (Fig. 4.15/2). This continuous distribution shows the difficulty of dividing houses into different categories of house sizes (viz., ‘large’, ‘medium’ and ‘small’), with the exception of structures with an area smaller than 20m2, which we have interpreted as ‘huts’, ‘sheds’ or ‘workshops’ – probably not for residential use in the general sense of ‘houses’47.

47  It should be noted that structures of 20m2 were frequent in the West Balkans in the 4th millennium BC, where they have been considered as ‘houses’ (Chapman, in prep.).

150 

 Site Studies

Figure 4.15: (1) bivariate plot of house sizes; (2) histogram of all house sizes; (3) house sizes by Outer Zone, Nebelivka (by J. Chapman).



Architectural Analyses 

 151

4.3.1.2 The Zonal Analysis For a more detailed analysis, the Nebelivka plan was divided into three Zones  – an outer Zone (including the Outer Circuit and houses outside the Outer Circuit); the middle Zone (the Inner Circuit and houses between the Inner and Outer Circuits); and the inner Zone (all houses inside the Inner Circuit). There were more differences between the Outer Zone and the other Zones than between the Middle and Inner Zones. The main difference was that the Outer part’s size peak of 66–75m2 (Fig. 4.15/3) was higher by one 10m2 range than in the other Zones. The Inner part differed from the other Zones through the higher frequency of houses of area 36–45m2. There was a steeper drop in large houses in the Outer Zone, after 75m2, than in the other Zones. These variations indicate differences in the cumulative development of the three Zones which require further discussion at a more detailed level.

4.3.1.3 The Sector Analysis The further subdivision of the Nebelivka plan into more defined areas brings further clarity to the overall pattern. Outside the Outer Circuit (‘OOC’), the 17 burnt houses were typically small structures, with a size peak of 26–35m2 (Fig. 4.16/1) – the joint smallest, with the houses between the Inner and Outer Circuits (n=6), of all sectors. Both these sectors had a narrow size range, with only one house in the OOC larger than 85m2 – a massive 143m2 structure that may well have been an Assembly House. The cross-streets (‘XS’) which potentially blocked radial streets showed a double size peak, at 36–45m2 and 56–65m2 (Fig. 4.16/2), while the houses in the Squares showed an intermediate peak of 46–55m2 (Fig. 4.16/3). With the exception of the Outer Circuit houses (‘OC’), with their size peak of 66–75m2, all of the other main sectors, as well as houses inside the Inner Circuit (‘IIC’), conformed to the total sample size peak of 56–65m2. These more detailed analyses confirmed the mean size differences between the Outer, Medium and Inner parts but more detail shows the tendency of most houses outside the main three plan elements – the Outer and Inner Circuits (‘IC’) and the Radial Streets (‘RS’) – to be built as smaller structures. This was particularly the case in the OOC and IC–OC sectors, suggesting different functions and meanings for houses in these sectors.

4.3.1.4 The Analysis of the Quarters The calculation of GINI co-efficients for a data series such as house sizes provides information on intra-group differences often explained through inequality (Kohler et al. 2017). These co-efficients have been calculated for house sizes by Quarter (Fig. 4.18/3 and Table 4.4). The results show that, while the majority of Quarters built houses of a size that was very similar to that of the overall site average, Quarters in the Southern half of the megasite showed greater house size diversity, with a resulting

152 

 Site Studies

Figure 4.16: (1) house sizes outside the Outer Circuit; (2) house sizes in cross streets; (3) house sizes in Squares, Nebelivka (by J. Chapman).



Architectural Analyses 

 153

increase in 'inequality'. This in turn suggests that groups within Nebelivka were trying to create and re-create similar visual architecture all over the site – as the growth of an architectural habitus or even part of the ‘Big Other’. The overall result is a picture which, overall, supports egalitarian tendencies at this megasite, while confirming some Quarters built houses larger than others. Table 4.4: GINI Co-efficients for House Size by Quarters (by M. Nebbia). Quarter

GINI

No of Dwellings

Area

Density

A

0.2700198

58

122016.9

2103.74

B

0.2444237

146

218093.7

1493.792

C

0.1906812

101

115849.8

1147.028

D

0.196473

104

102904.2

989.4635

E

0.179496

50

53401.83

1068.037

F

0.2172476

151

138439.9

916.8205

G

0.1703885

147

139005.3

945.6143

H

0.2211206

106

146844.7

1385.327

I

0.2180508

104

139008.2

1336.617

J

0.2438355

54

79303.9

1468.591

K

0.3010071

118

155454.8

1317.414

L

0.2531448

121

163780.4

1353.557

M

0.2708698

60

63444.85

1057.414

N

0.2389944

113

166724.8

1475.441

Nebelivka

0.2340455

1433

238000

166.0851

4.3.1.5 The Analysis of House Sizes in Neighbourhoods Since the measurement of house sizes included only burnt houses, the number of Neighbourhoods with appropriate sample sizes fell to 105; the smallest number of Neighbourhoods in a Quarter was six (Quarter M), while there were 18 Neighbourhoods in Quarter N. The most obvious conclusion is the massive variability of individual Neighbourhoods – there are no two Neighbourhoods that are similar in terms of house size. This result merely emphasises the already formulated conclusion (see above, p. 148) that Neighbourhoods were developed from the bottom up through myriad decisions about what and where to build new houses. Although these dynamics were

154 

 Site Studies

mainly rooted in the local rather than the global, we can recognise common patterns in the range of house sizes found in Neighbourhoods in different Quarters, suggesting shared responses to those local dynamics. The analyses focussed on two aspects of the complex data: (a) the maximum size class in a Neighbourhood48; and (b) the mean breadth of size classes in a Neighbourhood49. The maximum class sizes fall into two groups: a larger group where Class 3 is dominant (11 Quarters) and a smaller group where Class 3 is equal to, or less frequent than, other size classes (Quarters A, I and N) (Fig. 4.17/1). The larger group is divided into three sub-groups: those with Class 4 but no Class 5 (Quarters C, F, H, J, K and M) (Fig. 4.17/2); those with Classes 4 and 5 (Quarters B, G and L) (Fig. 4.17/3); and those with neither Class 4 nor 5 (Quarters D and E). The dynamics of choice of house size are complex but it may be suggested that Neighbourhoods with a dominance of Class 4 or 5 houses betoken a greater sense of size-based emulation or competition than in other groups, reinforcing the number of residents living in such large houses. Scores for the mean breadth of size classes varied between 2.5 to 4.5, with low scores indicating a narrow range of house sizes and high scores a wide range (Fig. 4.18/1). These mean breadth scores can be arranged in ascending order (lower group with 2.5–3.4: Quarter G–C /A–I–F–E–H / N; upper group with 3.5–4.5: Quarter L–K–B–D / J–M). It is worth noting that two of the three Quarters with the highest mean breadth scores (J and M) also have the smallest number of Neighbourhoods. There is a general correlation between Quarters with higher maximum house sizes and those with higher mean breadth scores. Several patterns are apparent which link the size breadth trends of different Quarters. The least common pattern is the spread of size classes 1 to 2 or 1 to 3 – found in only four out of 14 Quarters. In the opposite direction, the spread of size classes 1 to 4 or 1 to 5 occurs in all but one Quarter, as is also the case with size classes 2 to 4 or 2 to 5 (Fig. 4.18/2). The presence of all size classes in a Neighbourhood shows that the maximum potential for size differentiation has been acted upon and selected, with all the implications for acquisition of building resources and, eventually, firewood for a successful burning. The decision not to enlarge the breadth of existing size classes was taken so regularly that we can assume that size differentiation was considered a positive, if not standard, feature of Nebelivka Neighbourhoods.

48  Five size classes were distinguished: Class 1: 15–35m2; Class 2: 36–55m2; Class 3: 56–75m2; Class 4: 76–95m2; and Class 5: >96m2. 49  Mean breadth of size classes was measured as follows: a score was assigned to each Neighbourhood (1 point for 1 size class; 2 points for a size range of 2 classes, etc.) and the total divided by the number of Neighbourhoods.



Architectural Analyses 

 155

4.3.1.6 Summary The analysis of Nebelivka burnt house sizes shows a clear overall trendline with a core module of 7.5m in length and 4m in width and successive 1m increases in width producing 7.5m increases in length. However, it is important to note that these are only general trends and the range of house lengths for a 4m width is 7m–18m, for a 5m width 7m–19m and for a 6m width 10m–24m. Different house widths have important implications for the design of roofing and the materials used. The mean house area for all burnt houses is 63m2, similar to that for Dobrovodi (64m2) and slightly smaller than for Majdanetske (67m2) and Taljanki (71m2). The house size plots show a unimodal peak (56–65m2), with a continuous distribution that makes it hard to divide houses into ‘small’, ‘medium’ and ‘large’ classes. The analysis of house size by Zones shows an important trend towards larger houses built in one of the three main plan Zones (Outer and Inner Circuit, Inner Radial Streets) and smaller houses built outside these Zones, especially outside the Outer Circuit and between the Outer and Inner Circuit. The GINI Co-efficient analysis of house size by Quarters shows how the GINI value of most Quarters lies close to the mean value for the whole megasite, suggesting an attempt to reproduce visually similar living zones wherever people settled at Nebelivka (Fig. 4.18/3)50. Looking at the variability of house sizes by Neighbourhoods within Quarters re-emphasised the importance of the bottom-up principle of building Neighbourhoods all over Nebelivka. Nonetheless, some of the patterns of the mean breadth of house sizes by Neighbourhood were found in most of the Quarters, suggesting tendencies wider than the local group were in action. Moreover, there was a general appreciation of house size differentiation within Neighbourhoods, as measured by the breadth of size classes. The conclusion from all of these house size analyses is that the cumulative effect of many local decisions was a complex, varied settlement plan formed through the tension between general principles of house size and local initiatives.

50 Minor differences in the totals of houses by Quarter found between Tables 4.1 and 4.4 are explained by the fact that only complete houses could be measured for the GINI calculations.

156 

 Site Studies

Figure 4.17: Maximum class sizes by Quarter; (1) Class 3 equal to, or fewer than, other size Classes; (2) Class 3 dominant, with Class 4 but no Class 5; (3) Class 3 dominant, with Classes 4 and 5, Nebelivka (by J. Chapman).



Architectural Analyses 

 157

Figure 4.18: (1) mean breadth of all house size classes; (2) spread of house size classes; (3) No. of houses vs. GINI House Size Co-efficient plot by Quarters, Nebelivka ((1) & (2) by J. Chapman; (3) by M. Nebbia).

158 

 Site Studies

Brian Buchanan

4.3.2 Visibility Graph Analysis 4.3.2.1 Introduction The high-precision geophysical plan of Nebelivka provides a distinctive dataset to analyse the origin and formation of a Trypillia megasite in a wider context. This complete plan details the organisation of the structures of the site over its lifetime of use as a gathering place. The plan presents a unique opportunity to analyse the entirety of a Trypilia megasite’s structural organisation to improve our understanding of the development and use of this type of site and to investigate spatial developments within smaller Quarters or in different models of site development (Burdo and Videiko 2016; Chapman et al. 2014a; Chapman and Gaydarska 2016). An innovative use of Visibility Graph Analysis (VGA) allowed the investigation of the geophysical plans and the visual characteristics of the spaces and places of Nebelivka to gain insights into the structuring organisational principles at different time periods. VGA investigates the visual characteristics of a built environment by combining aspects of visibility fields (Benedikt’s isovist), space syntax theory, and small-world network approaches to “(…) derive a visibility graph of an environment – the graph of mutually visible locations in a spatial layout” (Turner et al. 2001, p. 104). This technique has increasingly been used in archaeological contexts and provides a unique method to interrogate the complexities of Nebelivka’s plan. The groups that inhabited or interacted with Nebelivka over its lifetime may have perceived ‘empty’ space (the areas of the site between built forms) differently owing to a variety of social, environmental, and cultural factors. The basic premise of why VGA was used is that the more visually integrated parts of a built environment (the ‘empty space’) have greater chances of attracting human movement and activities, and it is an application that can quantify these areas and compare if there were significant differences in the overall plan of Nebelivka or in the models of development (Turner 2003). The VGA adaptation employed here differs from the original use of the technique, which is focused on understanding human movement based on visibility fields within interior architectural space (Turner et al. 2001; Turner 2001). Instead, VGA expands the method beyond the interior of houses to the intra-site level to improve our understanding of the visual characteristics of Nebelivka’s overall built environment, constituting dwelling houses, smaller structures (huts), Assembly Houses and broad, but related, open spaces. This chapter discusses the scholarly background of intra-site analyses of the built environment and how VGA was used at the site before turning to the results.

4.3.2.2 The Built Environment Space and place are important concepts for considering archaeological communities, as the spatial organisation of these terms reflect how groups of individuals interact



Architectural Analyses 

 159

with and move through their local and regional environments. These somewhat abstract terms have become critical areas of study across the humanities and social sciences and refer, at a basic level, to where something is located (a place) and where something is not (a space). These concepts are interrelated and they tend to define one another, with the more abstract spaces often demarcated and defined by their relationship to places (Tilley 1994; on the transformation of space into place, see Chapman 1988a). Examples include the empty areas between walls or buildings at a small scale or the vacant regions between settlements, cities, or states at larger scales. Together, space and place comprise the built environment: the cultural alterations to the natural environment where human practice and events take place. Although the duality of space and place as noted above has been disputed (for example Ingold, 2009, p. 38), there is broad agreement on the importance of the built environment by social scientists due to its influences on social groups’ behaviours and practices while at the same time its reflection of the socio-cultural norms of societies (Agnew 2011; Fisher 2009; Gieryn 2000; Goodchild & Janelle 2010; Hillier 2014; Hillier & Hanson 1984; Ingold 2009, 1993; Lawrence & Low 1990; Tuan 1977). Archaeologists have long been attracted to many of these approaches and their focus on the physical effects of walls, ditches, structures and monuments on individuals’ perceptions, movements, and activities at the household, settlement, and landscape levels of analysis (Allison 1999; Aslan 2006; Delle 1998; Ferguson, L. 2012, Ferguson, T. 1996; Flannery 1976; Hastorf 1991; Steadman 1996; Tringham and Krstić 1990; Wilk and Rathje 1982). The built environment structures day-to-day interactions within and between cultural groups in socially-specific ways. Edward Hall’s notion of proxemics argues that individuals’ interactions with spaces and places are specialised and dependent on social group membership (Hall 1966) – an idea that has influenced much of the scholarship examining the relationship between spaces and places within the built environment. This idea allows an understanding not only of the development of the built environment but also its alterations due to continual use by ever-changing social interactions. Amos Rapoport built on these ideas, arguing that individuals’ interactions and use of space within the built environment are linked to how people understand the social meaning, cues, and conditions of the constructed world (Rapoport 1982). His cognitive congruence model demonstrates that the organisation and usage of the built environment reflects the social norms and practices of cultural groups (Rapoport 1982, pp. 287–289). Thus, people arrange and alter their environment according to their specific shared social memories, norms, and practices as well as making changes and alterations to these places based on changing normative frameworks. Based on these ideas, the organisation and development of the built environments at Trypillia sites can be thought of as a form of non-verbal communication that was inherently understood by members of the social group(s) that utilised these settlements. Therefore, a study that quantifiably addresses the built environment of Nebelivka can produce better understandings of social interactions and cultural norms.

160 

 Site Studies

One of the more popular as well as practical methodological frameworks for accessing the built environment was Bill Hillier and Julienne Hanson’s space syntax theory. Space syntax argues that there are underlying rules to the built environment related to patterns of movement or practice, and these rules can be mapped and understood (Hillier & Hanson 1984). It quantitatively investigates the configurational properties of the built environment using a variety of techniques that identify how humans proceed through and process the complexities of the built environment (Hillier 2005, p. 5). Originally designed to improve understanding of urban planning and development, space syntax has evolved and been utilised by a variety of disciplines. It has appealed to archaeologists working on past built environments, and it has been used to better understand differential zones within structures and settlements as reflecting distinctive activities, norms, and statuses within past social groups (Chapman 1990a; Bowser & Patton 2004, p. 170; Deetz 1996; Fairclough 1992; Ferguson, T. 1996; Steadman 1996; Van Nes 2009). The analytical techniques developed by Hillier & Hanson have arguably been accepted more by archaeologists than the theoretical underpinnings of space syntax theory. T.J. Ferguson (1996), for example, used the access and axial maps proposed by space syntax theory to examine historic-period Zuni architecture in the American South-West. Although he argued that the techniques were a valuable analytical tool, he noted that the theoretical models of space syntax needed to be adjusted for archaeological research needs (Ferguson, T. 1996, p. 152). In addition, although space syntax theory has been an extremely useful methodological framework for investigating the archaeological built environment it is a time-intensive technique to both prepare the hand-drawn access and axial maps correctly and to interpret the results of these processes.

4.3.2.3 Computational Approaches to Space and Place Arguably the most significant innovation for investigating the built environment over the last twenty years has been the integration of spatial analysis in Geographic Information Systems (GIS) with archaeological datasets. GIS has the unique abilities of combining highly accurate cartographic map layers, detailed databases and powerful analytical tools for the investigation of the spatial world (Conolly & Lake 2006). As Wheatley has noted, the integration of GIS into archaeological investigations has provided a ‘meaningful archaeology of place’ due to its powerful computing abilities and high accuracy (Wheatley 2004). One of the GIS techniques that has received a relatively large amount of attention is visibility, or viewshed analysis (see Brughmans et al. 2015; Conolly & Lake 2006; Gaydarska 2007; Gillings 2017, 2015; Llobera 2003; Llobera et al. 2010; Wheatley & Gillings 2002). Viewshed analysis interpolates digital elevation models into ground surfaces and, based on elevation surface changes affecting line of sight, models what is visible from any given point. This technique has revolutionised landscape studies and reinforced the importance of intervisibility between sites and/or monuments for understanding past cultural practice (Gillings 2017, 2015; Wheatley & Gillings, 2002).



Architectural Analyses 

 161

In general, viewshed analysis has investigated intervisibility between sites or monuments across broad areas of the landscape instead of assessing visibility within archaeological sites. Although a powerful tool for investigating viewsheds in the landscape, this technique is not suitable for investigating the visual characteristics within the built environment of structures or settlements. It is preferable, therefore, to use VGA, which combines aspects of viewshed analysis, space syntax analytical frameworks, and social theory for analyses of visibility fields and their effects on movement within the built environment of households and settlements. VGA models the intervisible connections between grid points of a graphical plan of the built environment. As structural forms alter visual perceptions of a place, VGA can model visual fields that alter movement and practice (Buchanan 2017; Turner et al. 2001; Turner 2004). VGA charts the most and least visible areas of a graphical representation of a household or settlement, using a combination of concepts derived from Hillier and Hanson’s space syntax theory with the graphical representations of visibility fields in architectural space which are known as isovists (Benedikt 1979; Turner et al. 2001). Originally introduced by Braaksma and Cook (Braaksma & Cook 1980), VGA was refined and redesigned by Alasdair Turner and the Space Syntax Laboratory of University College London in the software package Depthmap, which calculates VGA and produces measurements of integration, connectivity, and depth by quantifying the visual connections of grid points within a graph (Turner et al. 2001; Turner & Penn 1999). In effect, it quantifies how space within an environment is culturally demarcated by examining the intervisible connections in a graph. VGA determines the areas of a graphical plan of a built environment that are the most and least visible to other portions of an area and can see or not see the largest number of other areas. It investigates these areas based on how the structural elements of a built environment impede visibility fields, thereby producing areas of a plan that are more public or more private. These more public or more private areas would have impacted movement and activities practiced within these areas. Whilst natural features such as topography, vegetation and waterways also affect practice, VGA is here considered a useful proxy for understanding how the physical arrangement of structural forms affect individuals’ interactions. In addition, this investigation has focused on the archaeological remains identified during the geophysical survey. Undoubtedly there were more ephemeral features of the site such as temporary dwellings or fence-lines that are both more invisible to the geophysics and would have altered perceptions and use of the site. It was felt that a focus on the known and identified features provided a valuable understanding of the social structuring of the site regardless of the more temporary (and unknown) features of the site. Future research could run simulations on these features if and when they are better understood. Depthmap processes VGA by attempting to connect the vertices, or nodes, within a graph to all of the other visible locations within the graph. It examines the graph node by node and then interprets the results using a variety of measurements (Turner 2001,

162 

 Site Studies

p. 3). Every node within a graph has its own unique number of connections to the other nodes, known as a vertex’s Neighbourhood (Turner 2001, p. 3). The Neighbourhood connections can be analysed using global measurements or local measurements, with global measurements calculating information from all the vertices in the graph and the local measurements using information from the immediate Neighbourhood of a vertex (Turner 2001, p. 4). The analysis at Nebelivka used global measurements, as these were deemed more appropriate for the open spaces of the site (Turner 2004, p. 14). Depthmap calculates the above VGA comparisons, and the results are colour-shaded images of the most and least-connected areas of a built environment based on a variety of measurements related to the visual connections of the grid nodes. In addition to the colour-shaded imagery, Depthmap provides measurement scores based on ideas from space syntax theory that can be used in comparison between different plans of the built environment (Buchanan 2015, 2017). Turner (2004, pp. 14–15) notes that the integration, entropy, and mean depth scores are the most reliable for analysis, and the connectivity measurements are the best for visualisation of the graphical plans. VGA has received limited usage within archaeological investigations, with much of the scholarship focusing on the interior architectural space of standing ruins. For example, David Chatford Clark used VGA to investigate the visual integration and spatial relationships of sanctuaries with the broader assembly places of Byzantine-era churches in present-day Jordan, demonstrating that the churches reinforced visual and spatial separation between the clergy and the masses related to different roles and interactions in the services (Chatford Clark 2007). Archaeologists have more recently adapted this methodology to investigate the spatial organisation of archaeological settlements and grave plans (Brookes et al. 2017; Buchanan 2017, 2015). This adapted methodology has increasingly been used in archaeological research and focused on how the built environment was spatially and visually organised, and as such provides an innovative method to understand the complexities and potential patterns in the organisation of the site. It does so through a quantifiable investigation of the built environment based on the visual and spatial arrangement of structural elements and the interpretation of these findings within a framework of established ideas regarding how individuals perceive and use space and place within their households, their communities, and in the landscape (Buchanan 2015, p. 23).

4.3.2.4 Nebelivka and VGA The structural plan of Nebelivka has been divided into fourteen Quarters on the basis of eight characteristics of the site (Chapman & Gaydarska 2016; here, Chapter 4.2.1 and Fig. 4.5). The Quarters generally centred on at least one Assembly House that was located within empty areas between two parallel, concentric rings of structures. Within the ring, there were clusters of linearly arranged structures that extended, like spokes of a wheel, into the relatively empty centre of the site. These arrangements created radiating avenues – termed ‘Inner Radial Streets’ – that extended from the Assembly



Architectural Analyses 

 163

Houses inwards. The selected boundaries of the Quarters were divided based on the spatial patterning and orientation of the structures as well as the topographical features of the landscape (Chapman & Gaydarska 2016). Ten of the fourteen Quarters were examined using VGA to examine if there were any similarities in different parts of the site in the visual arrangement of the Quarters. Ideally, all of the Quarters would have been surveyed, but due to time and computational constraints a sub-sample was used in the analysis. VGA tested the extent of differences or similarities between the plans of the Quarters. Quarters B, C, D, F, G, H, I, L, M, and N were selected to provide a valid sample size of the spatial patterning of the site. It has been previously argued that the buildings at the site were not all standing at the same time (Chapman & Gaydarska 2016; Gaydarska 2019, 2019a). The lack of overlapping buildings as shown in the geophysical survey plans, therefore, suggests either a physical awareness or social memory of where buildings had previously stood as well as a respect shown to these zones which prevented buildings from disturbing these spaces (Nebbia et al. 2018; Gaydarska 2019). This was somewhat fortuitous for the VGA analysis, for the acknowledgement that all the structures were not standing at one time could be investigated as a proxy for understanding visibility and movement, given the avoidance of previously built-on space. Six of the Quarters (B, C, F, I, L, and N) were further analysed in VGA using two models of the evolution of the Quarters’ plan at three temporal stages of development (Nebbia et al. 2018; Gaydarska 2019). Models A and B were developed to examine the origins of the site and as a way to rationalise the limited material culture found at the site alongside the high number of burnt houses (n=1,445) (Gaydarska 2019). The two models were divided into three stages – Early, Middle and Late – representing time-slices of the life-cycle of the Quarters and denote two competing ideas on how the plan of Nebelivka developed. Model B demonstrates that a small core population permanently lived at the site, and the population grew seasonally due to large assembly gatherings. In contrast, Model A contends there was a larger permanent occupation of around 400 houses at any one time, and seven to ten houses were burnt annually, and the same number were built to create the large plan of burnt houses seen in the geophysics plan (Gaydarska 2019). The identified Quarters ranged in size and composition but had basic similarities in the spatial arrangement of the structural evidence in size, scale, and organisation. The Quarters were divided based on observable differences in the topographic characteristics of each area along with the position and arrangement of the structures (Gaydarska 2019). VGA was performed to address a few key questions about the visual arrangement of Nebelivka structural evidence, and, more specifically, how this connects with the division of the site into the above-discussed Quarters. First, were the Assembly Houses key to the integration and connectiveness of the Quarters? The large Assembly Houses appear to command a special place within both the concentric ring of houses and within each Quarter, separated from the other structures by large areas of empty space. By testing the Quarter plans using VGA, quantitative results can be reached

164 

 Site Studies

that can suggest the degree of visual connection of the Assembly Houses to other parts of the Quarter, i.e., by public vs. private space. Secondly, VGA allows us to test if there were significant differences in the organisation of the Quarters using the various models and temporal stages. Similar or significantly different measurements could imply either differences or similarities in how Nebelivka developed over time. Finally, VGA could also be used to examine quantitatively if there were similarities or differences in various parts of the site by providing visual and statistical evidence to test many of the observations that have previously been made on the structural evidence (Chapman et al. 2014a; Chapman & Gaydarska 2016).

4.3.2.5 Methodology The steps required to accurately run Depthmap to assess Nebelivka’s built environment amounted to a lengthy and involved process. The geo-rectified geophysical survey plans of Nebelivka’s Quarters were converted from digitised, GIS shapefiles into AutoDesk drawing exchange format (.dxf) files and then imported into Depthmap. These files preserved the scale, shape, and orientation of Nebelivka’s structural forms. Each Quarter was examined with all of the known buildings present and with reduced numbers within the models and stages. In addition, the proposed Quarter boundaries were imported into Depthmap. Once a Quarter was fully imported into Depthmap, a rectilinear grid was overlaid on the entirety of the Quarter plan. The grid was overlain on the imported files at a defined interval of 2m. This interval allowed a fine resolution of VGA outputs, accurate results, and a satisfactory processing speed. A flood-fill algorithmic command filled all of the space within the Quarter that were not within one of the structures from the plan. The now filled grid plan was used to calculate VGA connections from each part of the Quarter based on visibility (Turner 2004, p. 1). Every grid node within the graph has a unique number of connections to the other nodes, which are then displayed using a colour-range from indigo for low values through magenta for high integration values (Turner 2001, p. 3). Global measurement averages of these connections were produced using calculations derived from space syntax. All of the global measurement averages were recorded (Fig. 4.21/1 & Table 4.6). However, the analysis focussed on three categories as they have been shown to be the most useful for investigating the visual arrangement of the built environment (Turner 2004, pp. 14–15). – Mean depth determines how many turns are required to connect grid points to the other visible points within the graph. The results are totalled and then divided by the total sum total of grid points in the graph to derive a mean depth score for each grid node (Turner 2004, p. 14). Graphic colours: darker blue – more easy movement; yellow–green – more circumscribed movement (e.g., Fig. 4.22). – Integration investigates how visually connected the grid points are to the other visually accessible grid points and approximates the relative “depth” or permeability of a grid point to all of the other points in a graph. This important



Architectural Analyses 

 165

measurement is a normalised form of the mean depth measurement and has been found to correlate with the intensity of pedestrian movement (Turner 2004, p. 14). Graphic colours: magenta–red–orange – high permeability of space; indigo–blue – low permeability of space (e.g., Fig. 4.23). – Entropy refers to the overall complexity of an analysed area by calculating the depth distribution within a graph. Entropy was developed to analyse the distribution of locations near every node to determine a relative measure of complexity to avoid prioritising open spaces within a graph (Turner 2004, p. 15, Turner 2001, p. 7). VGA produces two measurements of entropy: visual entropy and visual relativised entropy which is a normalised measurement of the entropy results. Graphic colours: yellow–green – complex movement; dark blue – simple movement.

4.3.2.6 VGA Analysis of the entire Quarters The division of the geophysical plan of Nebelivka into Quarters was based on visual observation of similarities in the spatial organisation of the build environment, in the topographical features of the landscape, similarities in the clustering of Neighbourhoods and buildings to at least one assembly house, and the orientation of the structural evidence. VGA was conducted on 10 Quarters using the entirety of structural evidence. As there is little to no evidence of the buildings overlapping one another, it is suggestive that the structural arrangement of each Quarter respected the location of previously occupied buildings, implying a long after-life of buildings in social memory and the planning of the Neighbourhoods and Quarters, and as such VGA could be used as a proxy to identify avenues of movement and areas of visual integration. Table 4.5: Explanation of 10 forms of VGA analysis (see Fig.4.19) used at Nebelivka (by B. Buchanan). Type of map

Summary

Connectivity

The number of connections each grid node can ‘see’ in a graph, this measurement is a useful visualisation of the connectedness of Nebelivka based on the structural plan. However, as Turner notes, connectivity does not produce as meaningful a measurement for comparative studies as the integration, entropy, and mean depth measurements (Turner, 2004, p. 10). This is because connectivity is highly influenced by the total size of an area, with larger areas naturally more connected using VGA

Point First Moment

Point first and second moments are mathematical measurements of the graph. Point first moment is the sum of the visible distances from every node to other nodes in the graph (Al Sayad et al., 2013, p. 35)

Point Second Moment

This mathematical measurement calculates the sum of the dispersal of a graph’s isovists, the visibility fields of the graph interpreted as geometrical polygons (Krukar & Conroy Dalton, 2013, p. 16)

166 

 Site Studies

Table 4.5: Explanation of 10 forms of VGA analysis (see Fig.4.19) used at Nebelivka (by B. Buchanan).

Continued

Type of map

Summary

Visual Entropy

Entropy defines the overall visual complexity of a graph based on the distribution of depths within an area. It was developed specifically as VGA in Depthmap appeared, when the program was first developed, to prioritize open spaces (Turner, 2004, p. 15). Entropy is ‘[…] a measure of the distribution of locations in terms of their visual depth from a node rather than the node itself (Turner, 2004, p. 15). Therefore, if the depth is evenly dispersed, the entropy measurement is high and if more areas are close to a node, the entropy scores are low

Visual Integration [HH]

Integration is a normalised calculation of mean depth, and as Hillier et al. note, correlates well with pedestrian movement within the built environment (Hillier et al, 1993). It determines how visually connected every grid point is to the other nodes in a graph, and approximates the permeability of a point to all the other points in a graph. This normalisation is based on efforts Hillier and Hanson proposed when normalising axial maps in order to make areas of different sizes comparable to one another (Hillier & Hanson, 1984). This normalisation measurement proposed by Hillier and Hanson (1984) produces the Visual Integration [HH] measurement in VGA

Visual Integration (p-value)

This integration score was a normalization suggested by De Arruda Campos & Fong (2003, 35.9) as a more appropriate measurement for analysing VGA than Visual Integration (HH). The P-value measurement is similar to the D-value but divides an area into pyramid-shaped patterns of analysis and, as such, may be better suited for examining broader spaces (Hillier & Hanson 1984, p. 114)

Visual Integration [TEK]

The Visual Integration [TEK] measurement is similar to the Visual Integration [HH] measurement in that it is a normalised calculation of mean depth and determines the permeability of a graph. It differs in that it utilises an integration normalisation variant proposed by Teklenberg et al. (1993), which Turner notes is a simpler normalisation of integration than HH (Turner, 2004, p. 15)

Visual Mean Depth

This measurement calculates the fewest number of turns needed to connect all of the grid points within a graph to all of the other grid points. The shortest route to traverse through the graph based on the least number of turns is calculated, and these calculations are divided by the total number of vertices to provide a mean depth score for every node in the graph (Turner, 2004, p. 14)

Visual Node Count

This measurement is the total number of nodes within a graph as defined by the area of the graph and the grid spacing set by the user. It is the basis of all the other global measurements (Turner, 2004, pp. 14-15)

Visual Relativised Entropy

Like the visual entropy measurement but considers the expected distributions around a node in order to normalise the entropy measurement



Architectural Analyses 

 167

Figure 4.19:Visibility Graph Analysis of Quarter L: (1) Connectivity; (2) Point First Moment; (3) Point Second Moment; (4) Visual Entropy; (5) Visual Integration: HH; (6) Visual Integration: P-value; (7) Visual Integration: TEK; (8) Visual Mean Depth; (9) Visual Node Count; (10) Visual Relativised Entropy (by B. Buchanan). For explanation of these models, see Table 4.5.

168 

 Site Studies

Figure 4.20: Connectivity analysis of all 10 Quarters: (1) B; (2) C; (3) D; (4) G; (5) H; (6) I; (7) L; (8) F; (9) N; (10) M (by B. Buchanan).



Architectural Analyses 

 169

Figure 4.21: (1) Average measurements of VGA analysis of entirety of structural evidence; (2) Average VGA measurements of Model A; (3) Average VGA measurements of Model B (by B. Buchanan).

170 

 Site Studies

The comparison of the VGA results using the entirety of the structures within the Quarters suggests similarities in the spatial and visual arrangement of the buildings across the site, with no significant differences noted in the measurements as discussed above (Fig. 4.21/1). The colour-shaded images of the VGA results demonstrate that the portions of the site outside of the concentric ring of structures were significantly less permeable/visually integrated than the rest of the Quarters, suggesting the more connected areas of both visibility and activities were within the site. The outer ring of structures effectively blocked visibility to regions outside the site from areas within the Quarters. The interior of the parallel concentric rings, on the other hand, tended to be the most integrated areas of each Quarter, with these zones around the assembly houses being highly visible to the other parts of the Quarters. The perceived similarities on the colour images were further tested using an Analysis of Variance statistical test (ANOVA) of the average global measurements, which demonstrated there were no significant differences in the VGA measurements between the 10 Quarters that were analysed. This implies similar long-term visual access across the Quarters throughout the site. Although there were no significant differences, there were outliers in the measurements. Quarters M and N, in particular, were more visually integrated than the other Quarters due to having fewer structures overall and clustering these buildings within distinct and smaller portions of each Quarter (Table 4.6; Fig. 4.21/1). This increased the number of large, open areas within Quarters M and N, which led to outlying VGA scores for these areas. This is particularly true for Quarter N, the only Quarter analysed without a large Assembly House within the empty space between the concentric rings of structures. This made this already relatively empty space more open in Quarter N and, within the VGA analysis, made it very well connected and integrated to the other regions of the Quarter. Quarter M, on the other hand, did not have the same number of inner radial streets, which in turn made a large and well-integrated public area in the Northern half of the Quarter. Fig. 4.20 shows how much more connected Quarters M and N are than the other Quarters using the Connectivity measurement51. Although these were outliers, the results still demonstrate a high degree of similarity in the spatial and visual organisation of the Quarters across the site when examining all of the structures. The analysis of the ten Quarters (Fig. 4.20) suggests that the inhabitants of the site were utilising similar patterns of spatial organisation based on visibility and movement and that the successive occupations of the Quarters were replicating traditional spatial structures. As previously discussed, it was unlikely all of the structures existed at once, but they appeared to respect one another by not being built on top of one another, and thus represent cultural memory and social cues to avoid or interact the previously burnt structures in specific ways. The image results

51  The set of maps for every VGA analysis has ten plots; however, owing to limits of space, not all are presented in this chapter. The complete set of maps can be found at the Archaeological Data Service https://doi.org/10.5284/1047599



Architectural Analyses 

 171

also indicate that the most visually connected areas of the investigated zones were near the Assembly Houses, with highly connected linear corridors leading to these visually connected, and presumably quite public, structures. The least visible areas were on the outer perimeter of the site, with the outer house circuit almost operating as a screen of visibility and/or movement, producing in effect a demarcation between the place of Nebelivka and the space near the edge of the settlement, marked by the perimeter ditch. These results are excellent barometers to compare with the analysis of the two proposed models of development.

4.3.2.7 The Distributed Governance Model (Model A) and the Assembly Model (Model B) Six of the Quarters were more closely interrogated in VGA, comparing the results of the above investigations of the maximal structural plan with two models of the development of the site – Model A (the Distributed Governance Model) and Model B (the Assembly Model). These models were further sub-divided into three separate temporal stages. The models and stages had differing numbers of houses and placements of these structures, with the total size and number of buildings in each Quarter/stage shown in Table 4.7. As noted by Diachenko (2016) and Gaydarska (submitted), the number of structures and corresponding number of potential inhabitants (about six per house) fit the range of populations of smaller and mediumsized Trypillia settlements in the region (0.3–2ha to 35ha) and reinforce the division of the site into Quarters and developmental models helping us to a better understanding of the overall plan. The Quarters, therefore, can be seen as a structural practice, forming smaller units of habitation to help deal with the scalar stress of living at a megasite (Gaydarska 2019). Table 4.6: Average VGA measurements of the entirety of structural evidence (by B. Buchanan). Area

Visual Integration Visual (TEK) Entropy

Visual Mean Depth

Visual Relativised Entropy

B

0.958857

0.795993

1.77479

2.52364

C

0.965055

0.8468

1.71953

2.42759

D

0.956214

0.834379

1.77355

2.48599

F

0.947765

0.80666

1.85775

2.5821

G

0.94765

0.830496

1.86102

2.56293

H

0.95971

0.842899

1.75848

2.46494

I

0.965468

0.858762

1.71666

2.41388

L

0.950528

0.813058

1.82634

2.5501

M

0.974847

0.938742

1.63974

2.26744

N

0.974684

0.89327

1.66087

2.32861

172 

 Site Studies

Table 4.7: Number of houses in each Quarter by Model and Stage (by B. Buchanan). Quarter/Size B (21.8ha)

C (11.6ha)

F ( 13.8ha)

I (13.9ha)

L (16.3ha)

N (16.6ha)

Total No. of Houses 154

101

151

103

124

111

Stage 1

Model A (Distributed Governance): No of Houses 47

Model B (Assembly): No. of Houses 30

2

48

52

3

46

28

1

33

26

2

34

41

3

26

27

1

44

50

2

44

60

3

55

11

1

34

20

2

32

36

3

35

20

1

36

45

2

32

54

3

52

26

1

29

55

2

29

65

3

47

44

First, and not surprisingly, the VGA results of three temporal stages of Models A and B were significantly different from the VGA measurements of the entirety of the structural evidence (Gaydarska 2019) (Fig. 4.21/2–3). At the same time, the VGA analysis of the two models did not display significant differences between each other or between the three stages of development. The significant differences between the models and the entirety of the structural evidence is most likely due to the fact that each stage of each model had markedly different numbers of structures and therefore spatial arrangements (for more details, see Figs. 4.22–4.25 and respective Tables). The global measurements of the two models at the varying stages displayed remarkably similar scores, although there were outliers in the measurements of Quarters F and L (Table 4.8). These outliers had scores demonstrating Quarters F and L were more visually complex and less integrated. Both Quarters were within the Southern half of Nebelivka, which suggests that there was regional differentiation of spatial complexity within the site.



Architectural Analyses 

 173

Table 4.8: Average VGA measurements of Models A and B (by B. Buchanan). Area

Group Number

Visual Integration (TEK)

Visual Entropy

Visual Mean Depth

Visual Relativised Entropy

B

Model A Stage 1

1.03538

0.921318

1.38045

2.02497

B

Model A Stage 2

1.00242

0.973094

1.49918

2.1008

B

Model A Stage 3

1.04784

0.883064

1.33075

2.00705

B

Model B Stage 1

1.02738

0.939518

1.40561

2.03501

B

Model B Stage 2

0.99202

0.980965

1.54357

2.13864

B

Model B Stage 3

1.04765

0.873754

1.35681

2.04099

C

Model A Stage 1

1.04392

0.896309

1.3517

2.01703

C

Model A Stage 2

1.00296

0.978189

1.49429

2.09165

C

Model A Stage 3

1.03284

0.923707

1.38569

2.02823

C

Model B Stage 1

1.01759

0.956007

1.43841

2.05413

C

Model B Stage 3

1.04672

0.88247

1.39493

2.02622

C

Model B Stage 2

1.00457

0.98002

1.48639

2.08216

F

Model A Stage 1

0.960927

0.844794

1.75093

2.45508

F

Model A Stage 2

0.963307

0.854241

1.73131

2.42954

F

Model A Stage 3

0.955328

0.809199

1.79242

2.52499

F

Model B Stage 1

0.965995

0.868157

1.71373

2.40042

F

Model B Stage 2

0.963713

0.859313

1.72641

2.42102

F

Model B Stage 3

0.965817

0.865543

1.71038

2.40131

I

Model A Stage 1

1.04854

0.88119

1.33261

2.0105

I

Model A Stage 2

1.01324

0.968356

1.44852

2.054

I

Model A Stage 3

1.03835

0.917564

1.35694

2.00452

I

Model B Stage 1

1.01461

0.956051

1.44923

2.06492

I

Model B Stage 2

1.03604

0.915204

1.36988

2.0195

I

Model B Stage 3

1.00686

0.977027

1.47553

2.0738

L

Model A Stage 1

0.977925

0.924914

1.6329

2.27367

L

Model A Stage 2

0.979952

0.919601

1.62669

2.27124

L

Model A Stage 3

0.957762

0.844009

1.76998

2.47248

L

Model B Stage 1

0.969386

0.894585

1.68709

2.35169

L

Model B Stage 2

0.965054

0.865666

1.7129

2.40324

L

Model B Stage 3

0.973354

0.908682

1.65816

2.31316

N

Model A Stage 1

1.01471

0.954462

1.44904

2.06599

174 

 Site Studies

Table 4.8: Average VGA measurements of Models A and B (by B. Buchanan).

Continued

Area

Group Number

Visual Integration (TEK)

Visual Entropy

Visual Mean Depth

Visual Relativised Entropy

N

Model A Stage 2

1.0221

0.944956

1.41902

2.04387

N

Model A Stage 3

0.98969

0.963286

1.56376

2.17341

N

Model B Stage 1

0.988429

0.937367

1.57879

2.21016

N

Model B Stage 2

0.993534

0.947114

1.55086

2.17407

N

Model B Stage 3

1.02371

0.946736

1.41001

2.03342

Overall, the Model A plans were more visually complex than Model B for each temporal stage, with corridors of more easily integrated space. The Model B plans tended to produce less visually complex and more open spaces, which created broader permeable (public) areas instead of smaller, grouped integrated space (Gaydarska 2019). More visually complex areas and less integrated (private) spaces in both models were located close to structures. The Assembly Houses were located in more public and well-integrated areas. These measurements reflect the differences in the morphological layout of the Quarters and their structural arrangements. Although there were observable differences when they were mapped out, there were not significant differences within the stages of the model – only between the models. The VGA results are summarised above and the average global measurements are detailed in Table 4.8.

4.3.2.7.1 Entropy Across the six Quarters, this measurement produced the most standardised values, with a majority of visually simplistic space in the open areas and more complex space clustered near the houses. Entropy measurements did not show much intermediary space within the models or stages, which contrasts with the VGA of the entire structural evidence which showed much more intermediate space across the Quarters. These observations are fairly consistent across the temporal stages and within each model. The two models differ in both how they were arranged over time, and how they would have been inhabited, with Model B more likely to have been populated for small periods of time in a year and Model A with more permanent dwelling but in smaller numbers. These differences would have affected the activities that would have been performed at the site, and the movement corridors in the entropy measurements thus reflect the unconscious and conscious perceptions of the spatial arrangement of the site. These are perhaps unsurprising measurements, as entropy measures the overall complexity and both model A and B are less visually complex than the complete plans.



Architectural Analyses 

 175

4.3.2.7.2 Mean Depth Like the entropy measurements, there are inverse relationships between the more public and private zones within all of the Quarters in both models and all three stages within the mean depth maps. The more public, open areas shift over time depending on the model and/or stage, with the more private areas clustered near the structures. The mean depth measurements appear to show a diachronic relationship between easy and intermediate movement zones due to the visual dynamics of the structural arrangements of space. The results suggest in the mean depth scores that it would have been easier to traverse the Quarters in Model B than Model A, with more visually accessible areas allowing multiple avenues of movement and access. A more detailed analysis of the two models for Quarters F and L shows similar results (Figs. 4.22 & 4.24 & Tables 4.9 & 4.11), with more integrated and/or connected areas allowing easier mobility.

4.3.2.7.3 Integration The integration maps show no clear-cut correlation between the unbuilt areas and highly permeable spaces as compared to the mean depth and entropy measurements (e.g., Quarters F and L: Figs. 4.23 & 4.25 and Tables 4.10 & 4.12). Whereas entropy measures the complexity of a graph and mean depth measures how many turns needed to go in any one direction, integration notes how many grid nodes can be seen from any one point. Therefore, the unbuilt areas are not necessarily the most integrated areas, as visibility lines to these unbuilt areas can be blocked by the structures that demarcate that space – in effect, making some of the open areas more private. It appears that there is a general increase in high-permeability areas over time in the use of the integration measurement, which aligns with the differences in the models and stages of development.

4.3.2.8 Discussion and Conclusion The results of the VGA of the Quarters of Nebelivka have both confirmed some of the original hypotheses and raised new questions on the development and spatial organisation of the site. First, the results suggest that the inhabitants of the site were following similar patterns of spatial organisation of the structures, and each Quarter would have had similar structuring of visibility and movement across the entirety of the site and over long periods of time. The models and corresponding stages of use provide similar visual accessibility measurements, which reinforces the viability of these models in explaining the site’s development. The long-term similarities in visual accessibility across all parts of the megasite show the extent to which megasite spatial order emerged as a monumental part of the Trypillia habitus. At the same time, all of the outlying measurements were located in the Southern half of

176 

 Site Studies

Table 4.9: VGA Mean Depth analyses of all Stages of both Models, Quarter F, Nebelivka (see Figure 4.22) (by B. Buchanan). STAGE DG MODEL – MEAN DEPTH, QUARTER F

ASSEMBLY MODEL – MEAN DEPTH, QUARTER F

1

Easy Movement focussed along the OC–IC corridor, with Intermediate zones dominating the rest of the Quarter, and Circumscribed Movement restricted to house groups.

Areas of Easy Movement focussed on the OC–IC corridor, with Intermediate Movement in the rest of the Quarter and little expansion of Circumscribed Movement beyond house groups.

2

Expansion in Easy Movement along the Similar to Stage 1 but with smaller area of OC–IC corridor, with the rest of the Quarter Easy Movement in the corridor. as in Stage 1.

3

Fragmentation of spaces affording Easy Movement, with continuing dominance of Intermediate Movement and expansion beyond a greater number of house groups.

Areas of Easy Movement dominate the OC–IC corridor, with Intermediate Movement in the rest of the Quarter and little expansion of Circumscribed space beyond house groups.

Figure 4.22: VGA Mean Depth analyses of all Stages of both Models, Quarter F, Nebelivka: (1) Distributed Governance Model (A), Stage 1; (2) Model A, Stage 2; (3) Model A, Stage 3; (4) Assembly Model (B), Stage 1; (5) Model B, Stage 2; (6) Model B, Stage 3 (see Table 4.9) (by B. Buchanan).



Architectural Analyses 

 177

Table 4.10: VGA Integration-TEK analyses of all Stages of both Models, Quarter F, Nebelivka (see Fig. 4.23) (by B. Buchanan). STAGE

DG MODEL – INTEGRATION, QUARTER F

ASSEMBLY MODEL – INTEGRATION, QUARTER F

1

The area of High Permeability is focussed As in Stage 1 of the DG Model on the OC–IC corridor, with Intermediate Permeability dominating the OOC and IIC and major expansions of Low Permeability beyond house groups, especially in the IIC.

2

High-Permeability areas have fragmented across the OC–IC corridor, with continuing dominance of the OOC and IIC by areas of Intermediate Permeability and somewhat less expansion of Low-Permeability space beyond house groups.

As in Stage 2 of the DG Model

3

Increasing fragmentation of areas of High Permeability, with greater dominance of the OOC and IIC than in Stage 2 and even more expansion of areas of Low Permeability space beyond house groups.

Expansion of areas of High Permeability in the OC–IC corridor, with dominance of the OOC and IIC by areas of Intermediate Permeability and limited expansion of Low Permeability beyond house groups.

Table 4.11: VGA Mean Depth analyses of all Stages of both Models, Quarter L, Nebelivka (see Figure 4.24) (by B. Buchanan). STAGE

DG MODEL – MEAN DEPTH, QUARTER L

ASSEMBLY MODEL – MEAN DEPTH, QUARTER L

1

Small areas of Easy Movement in the West part – mostly in the OC–IC corridor – with Intermediate areas dominating the OOC and IIC and expansion of Circumscribed space beyond house groups mostly in the IIC.

Area of Easy Movement confined to East side of the OC–IC corridor and part of the IIC; zonal distribution of Intermediate space on the West side, with expansion of the Circumscribed area beyond most house zones.

2

Expansion of Easy space from the OC–IC corridor into the IIC, with a zonal distribution of Intermediate space on the East side. Little expansion of Circumscribed zone beyond house groups.

Changes in location of areas of Easy and Intermediate movement, with the latter a zonal distribution on the East side and the former a zonal distribution on the West side. Expansion in the areas of Circumscribed movement outside house zones.

3

Contraction and fragmentation of areas of Easy Movement, replaced by Intermediate zones in the OC–IC corridor and the IIC; continued expansion of Circumscribed zone beyond house groups.

The Intermediate-Movement zonal distribution has expanded to take in the Centre as well as the East side, with a further reduction in areas of Easy movement and the pattern of Circumscribed movement the same as in Stage 2.

178 

 Site Studies

Figure 4.23: VGA Integration-TEK analyses of all Stages of both Models, Quarter F, Nebelivka: (1) Distributed Governance Model (A), Stage 1; (2) Model A, Stage 2; (3) Model A, Stage 3; (4) Assembly Model (B), Stage 1; (5) Model B, Stage 2; (6) Model B, Stage 3 (see Table 4.10) (by B. Buchanan)

the site, suggesting possible differential arrangements based on location within the site complex. These differences could be due to these areas developing at different times, or to different social groups inhabiting or using these areas and thus altering the spatial arrangement slightly from their neighbours to the North. The VGA results also reinforce the importance and status of the Assembly Houses for the development and use of the site overall and more specifically the defined Quarters. These places show striking visual accessibility within the Quarters, with highly connected visual fields in the immediate vicinity of the Assembly Houses as well as corridors of movement and visibility leading to the large structures. Taken together, they suggest that the spatial location of the Assembly Houses is not



Architectural Analyses 

 179

Figure 4.24: VGA Mean Depth analyses of all Stages of both Models, Quarter L, Nebelivka: (1) Distributed Governance Model (A), Stage 1; (2) Model A, Stage 2; (3) Model A, Stage 3; (4) Assembly Model (B), Stage 1; (5) Model B, Stage 2; (6) Model B, Stage 3 (see Table 4.11) (by B. Buchanan).

accidental; they are located in the most integrated and public zones of the Quarters and thus located in the areas of easiest mobility and highest public interaction. Their locations may also have been related to sound, as nodes in an interconnected soundscape. Their different construction (see above, pp. 133–136) may have enabled a higher permeability of sound than in ordinary buildings. Although it has been hypothesised that these structures were important for the development of the site as a whole due to their size and spatial regularity across the site, the VGA of the Quarters suggests that their location was key to the development of the Quarters across time, as they remain highly integrated across models and temporal periods.

180 

 Site Studies

Figure 4.25: VGA Integration-TEK analyses of all Stages of both Models, Quarter L, Nebelivka: (1) Model (A), Stage 1; (2) Model A, Stage 2; (3) Model A, Stage 3; (4) Assembly Model (B), Stage 1; (5) Model B, Stage 2; (6) Model B, Stage 3 (see Table 4.12) (by B. Buchanan).

The geophysical plan of Nebelivka has inherent importance for understanding the development and use of Trypillia sites. Its completeness and high resolution have shown a complex plan of a large number of burnt and unburnt structures, ditches, and Assembly Houses. This high resolution has allowed the interrogation of the site through visibility graph analysis, which examined the spatial organisation and visual arrangement of the built environment and noted that, while there were outliers, there was no significant difference in the visual arrangement of the structures across the site. In addition, two models of development were tested using VGA and, although they displayed different zones of integrated space, the analysis of the models also displayed fairly similar average measurements of integration, entropy, and depth.



The Experimental Programme 

 181

Table 4.12: VGA Integration-TEK analyses of all Stages of both Models, Quarter L, Nebelivka (see Fig. 4.25) (by B. Buchanan). STAGE

DG MODEL – INTEGRATION, QUARTER L ASSEMBLY MODEL – INTEGRATION, QUARTER L

1

High-permeability space limited to West part of OC–IC corridor and SouthWest part of IIC; Quarter dominated by Intermediate space, with low permeability around buildings and expanded from houses in IIC, OC and OOC.

Large corridor of high-permeability space in East side of Quarter, with corridor of intermediate space immediately to the West; major expansion of low-permeability space around all houses, as in DG Model

2

Increase in high-permeability space but more fragmented, with patches in IIC and across the OC–IC corridor; decrease in intermediate space, with low-permeability space as in Stage 1.

Similar corridor of high-permeability space but shifted to West side of Quarter, with similar areas of intermediate space to Stage 1 but expansion of low-permeability space, especially in IIC and OOC

3

Similar fragmentation of highpermeability space but in different parts of OC–IC corridor and IIC; intermediate and low-permeability spaces as in Stage 2

Major expansion of low-permeability space to dominate the Quarter, especially in IIC, IC, OC and OOC; small area of high-permeability space in West side (shrunk from Stage 2) and intermediate space limited to OC–IC corridor and West side of IIC.

Both models suggest possible patterns in how Nebelivka could have developed and demonstrate the importance of public vs. private zones in our understanding of the site’s diachronic development.

Stuart Johnston 4.4 The Experimental Programme 4.4.1 Introduction One of the research goals of the Project was an improved understanding of the taphonomy of the Trypillia burnt houses and, in particular, those excavated at Nebelivka, where by far the largest excavated sample comprised houses, whether complete (House A9 – 2009; House B17 – 2013), parts (House B18 – 2013) or test pits in over 80 (mostly burnt) houses. Much of Cucuteni-Trypillia archaeology can be defined as the excavation of burnt houses. But the Project has been unable to find a thorough, detailed account of the taphonomy of a burnt house in what has become a rather traditional field of recording of excavated features and finds and their interpretation as a reflection of a living house assemblage that has collapsed in a narrow range of ways (Monah & Monah 1997; KorvinPiotrovskiy et al. 2012; Müller & Videiko 2016). Excavators of burnt Cucuteni-Trypillia

182 

 Site Studies

houses have accumulated vast databases of specific forms of remains, especially the mass of burnt clay known in Russian as the ‘ploshchadka’ (Fig. 4.26 upper) and various burnt clay features thought to be walls, podia, altars (or ‘platforms’) etc. Equally, although eight house-building experiments of eleven houses have been previously conducted (for summary, see Burdo 2011; Cotiugă 2009), only one excavation of an experimental burnt house has been made (the excavation is summarised in Chabaniuk 2008) in order to make direct comparisons with the excavated remains of 6,000-yearold houses. It was for this reason that the Project decided to build two smaller-thanlife-size Trypillia experimental houses – one single-storey and one two-storey – burn them down and excavate the burnt remains in order to make a comparison of different house remains. A report has been published for the construction of the onestorey and two-storey houses and the burning of the two-storey house (Johnston et al. 2019; https://doi.org/10.5284/1047599 Section 6), while the detailed report on the excavation of the burning of the experimental house (Johnston et al. 2018; https://doi. org/10.5284/1047599 Section 6.6) is the first of its kind. There were four major issues to which the Nebelivka house-burning experiment could make a useful contribution – Issue 1: whether the burning of an experimental two-storey house left traces that would be recognisable in excavations of CucuteniTrypillia ploshchadki; Issue 2: the comparative interpretation of features, fittings and objects in the experimental house and Trypillia burnt houses; Issue 3: the nature and quantity of fuel needed for a successful house-burning; and Issue 4: whether houseburning was a deliberate social practice. Because of its timescale, the experiment could not contribute to debates over the effects of soil formation processes, the results of krotovina action52 or the preservation of different constructional materials such as ash, charred material or weakly burnt daub that would eventually have reverted to clay or soil.

4.4.1.1 Issue 1: The Creation of a Ploshchadka The presence of a ploshchadka in the excavation of a Trypillia house has been a certain indication of a high-temperature fire from the earliest Trypillia investigations (Khvoika 1901). The converse – the recognition of unburnt houses – came much later and was based on a weak geomagnetic anomaly and the absence of a ploshchadka (e.g., in Nebelivka Test Pit 1/4: Burdo & Videiko 2016, p. 107, Fig. 9; see also below, p. 217, 221). It is, therefore, perhaps surprising that none of the previous Trypillian house-burning experiments have claimed to produce a convincing ploshchadka (A. Diachenko, pers. comm.; Burdo 2011). This was presumably because of the absence of one or more key ingredients or circumstances: well-dried clay, plentiful fuel to

52  Krotovina are the traces of underground movements by rodents whose effects on stratigraphy can be severe.



The Experimental Programme 

 183

Figure 4.26: Upper: Burnt house remains (ploshchadka), Mega-structure Context 55; lower: bone and ceramic scatter, Pit near House B17, Nebelivka (upper by J. Chapman; lower by M. Videiko).

produce a high temperature, conditions for the creation of a good draught, and/or a bright, sunny and windy day. In the case of the Kruts & Chabanuk experiment, the design choice of a log cabin consumed 2.5 times as much timber as a Nebelivka house of the same size and 1.5 as much clay (Chabaniuk 2008, pp. 219–220); therefore, the lack of clay cannot be invoked to explain the lack of a ploshchadka. Most of the previous experiments did not quantify the fuel utilised or the temperature of the house-burning. However, the personal observation by A Diachenko of several houseburning experiments showed that far less fuel had been used than in the Nebelivka

184 

 Site Studies

house-burning (30m3 of timber). A high temperature of over 700oC was essential to convert the low-conductivity clay into a burnt mass; without sufficient fuel, the temperature needed for a ploshchadka could not be achieved. It is important to note that, after the successful Cucuteni Experimental Archaeological Park experiments (Cotiugă 2009), the Nebelivka experiment was only the second recorded houseburning53 which produced a ploshchadka identical to those on excavated Trypillia sites (see DOI https://doi.org/10.5284/1047599 Section 6.6.2.1). However, the production of vitrified daub in the Nebelivka experiment was the first such in the house-burning experiments, confirming that a temperature of over 1000oC was reached.

4.4.1.2 Issue 2: Detailed Interpretations of House Features The Nebelivka experiments can contribute to three long-running debates concerning the interpretation of excavated Trypillia house remains: the use of fire in construction burning; the design choices for walls; and the presence of one- or two-storey houses. The claim from Korvin-Piotrovskiy et al.’s (2012) experimental programme is that controlled firing of the lower walls of a house or its internal features could strengthen without destroying them. A small-scale house-firing experiment organised in the grounds of Durham University (https://doi.org/10.5284/1047599 Section 6.3) attempted to re-create the distinctive cracked surface of the house platforms (aka ‘altars’) often found in Trypillia houses (Fig. 4.27 upper). This experiment showed that construction burning did cause the development of a cracked surface but that this construction burning would have damaged the house walls unless the firing was done before the building of the walls. This indicates that construction burning worked for the building of a platform provided it was built on the ground floor, before the wall construction. However, there is no obvious difference between the effects of construction burning and post-dwelling destruction burning on platforms or other features. This experiment raises serious doubts about the likelihood of construction burning in wall-building. The question of wall-design of Trypillia houses arises from the well-documented absence of post-holes in house excavations, as well as the paucity of wall remains. Both excavations of experimental burnt houses – Chabaniuk’s (2008) excavation of the Legedzine burning of a log cabin and the Nebelivka burning of a wattle-and-daub house with vertical posts set in sleeper beams – demonstrated that wall remains were preserved in burnt house remains. At Nebelivka, wall daub could be differentiated from floor daub through the small size and regularity of the withy impressions, as compared with larger, semi-circular or squared-off floor timbers. Several wall panels whose collapse was documented at the time of the house-burning were discovered in

53  By a ‘recorded house-burning’, we mean a graphic record of the objects placed in the house-tobe-burned and a timed photographic record of the burning, with analyses of daub firing temperature.



The Experimental Programme 

 185

the excavation in the expected state and place. Identical withy impressions to those in the Nebelivka experiment (Fig. 4.27 lower) have been found in burnt Trypillia houses such as Nebelivka House A9 and Dobrovodi House 4. The experimental results confirm that wall remains can be well preserved in, and form part of, Trypillia ploshchadki.

Figure 4.27: Upper: platform, Durham Experiment; lower: withy impressions, Nebelivka Experimental Burnt House Excavation (by S. Johnston).

186 

 Site Studies

Since local permission was denied us to burn down both the one- and twostorey houses at Nebelivka, there were no direct comparative results from the experiment. Nonetheless, useful information was gathered on both types of house. There were several lines of evidence supporting the effects of burning on two-storey houses. In this discussion, it is important to note the terminology applied to floors: for a one-storey house, the ground floor and the upper floor (dividing ground-floor room from loft) and, for a two-storey house, the ground floor, the middle floor (separating the ground-floor room from the second-floor room) and the upper floor (separating the second-floor room from the loft). Schematic reconstructions of the destruction of one- and two-storey houses provide a range of potential scenarios (Figs. 4.28–4.29). The most obvious effect was the way that our excavation located a total of 11 wall-panels, which fell inwards into the house (five cases) (house-burning example: Fig. 4.30 upper) or outwards (experimental burnt house excavation: Fig. 4.30 lower). Well-fired panels would have been preserved whichever direction they fell. For poorlyfired wall panels, inward-falling panels would have been protected by other house debris, while the lack of protection of outward-falling panels meant a lower chance of surviving to the present-day. The issue of whether middle-floor features such as hearths, podia and platforms could have survived intact a fall onto the ground floor was more complex. While we cannot rule out the possibility of an entire floor sliding downwards to show good preservation of internal features, the experimental evidence of hearth fragments and fragmented floor daub showed the impact of the fall. Wherever middle-floor fragments and fragments of features (especially platform fragments, as found in the Nebelivka excavations) appear in excavation, it is suggested that this is good evidence of a two-storey house; conversely, intact or fragmented but in situ features and floor remains indicate the high probability of a one-storey building (see below for Test Pits: Chapter 4.6.1).

4.4.1.3 Issue 3: Construction Materials and Fuel for House-Burning The production of an individual house can be viewed as a symbolic fusion of the different elements that made up the Trypillia landscape. Creation and fusion was achieved through combining clay from the earth with straw from the steppe or as a by-product of agriculture, wood from the forest and reeds and water from rivers and lakes. The coordinated construction of parts of the megasite can similarly be viewed as an expression of co-operation within and outside a supra-household group.



The Experimental Programme 

Figure 4.28: House collapse scenario 1 (1-storey houses) (by L. Woodard).

 187

188 

 Site Studies

Figure 4.29: House collapse scenario 2 (2-storey houses) (by L. Woodard).



The Experimental Programme 

 189

Figure 4.30: Upper: house panels falling inwards, Nebelivka House Burning Experiment; lower: house panel fallen outwards, Nebelivka Experimental Burnt House Excavation Contexts 425 & 426 (upper by M. Nebbia; lower by J. Chapman).

190 

 Site Studies

The quantity of wood necessary to build a Trypillia house has been discussed on many occasions (e.g., Kruts 1989; Kruts et al. 2001; Korvin-Piotrovskiy et al. 2012). This figure is dependent on several factors, principally the number of storeys and the type of architecture, whether a wattle-and-daub upper storey built on a log-cabin ground floor (e.g., Legedzine: Korvin-Piotrovskiy et al. 2012, Fig. 9.4) or a timber-framed house with wattle-and-daub walls on sleeper-beams (e.g., the two Nebelivka houses: Fig. 1.9). The estimates for the construction materials for the one- and two-storey houses are shown below, with values for full-size Majdanetske houses for comparison (Table 4.13). Scaling-up of timber requirements to the standard Nebelivka ‘Module’ of 100 15m × 5m houses would have meant the felling of over 13,000 trees for one-storey buildings and almost 25,000 trees for two-storey buildings – ca. 20,000 trees for a mixture of one- and two-storey houses. Even the coeval construction of 10 houses would have required the transport of almost 400 tonnes (170 m3) of clay, over 150m3 of reeds and the collection of timber resources from a 1km2 area of forest. The production of houses on the scale required to construct large parts of a megasite would have required co-ordinated management of skills, labour and the landscape. This includes woodland management, with coppicing of hazel trees beginning several years in advance to produce withies of suitable size. In the process of constructing a Neighbourhood or Quarter, the coordination of these activities and the formation and maintenance of a reliable supply chain would have had organisational and administrative effects beyond the comparatively modest efforts necessary to construct a single dwelling. Kirleis and Dal Corso (2016, p. 201) claim that de-husking took place within the settlement and the waste material was either discarded in pits or used in daub-making. Such a statement has important implications for temporality and storage. The largescale building of megasite houses in one site phase would have required large-scale storage of husks for daub-production (see below, Section 6.2). This is not impossible but such a scenario needs further evaluation in terms of planning and logistics. Alternatively, the building programme was much slower, lasting decades rather than a few years. This would have allowed the storage of the husk waste from one household for subsequent use in the next planned house building. If, by contrast, all the houses had already been built, the husks produced by the Majdanetske households would not have been needed and then all husks would have been discarded in pits. We suggest that a steady building programme that would incorporate and recycle the daily waste into useful building material was more likely; it also matches a broader strategy of woodland management, including the coppicing of hazel, for a constant supply of building material.  The large quantity of fuel placed in the house before firing was probably instrumental in achieving a complete combustion of the two-storey structure. We cannot, however, be certain whether complete combustion could have been achieved without the filling of the upper room with much fuel (cf. Kruts 2003). We maintain that the main fuel element was timber, because of the logistical problems of collecting



The Experimental Programme 

 191

large quantities of animal dung or reeds and their insufficient thermal properties when used alone. Table 4.13: Estimates for house-building of (a) Nebelivka experimental houses and (b) full-sized Trypillia houses, Majdanetske (by S. Johnston). a: Estimate of quantities of materials used in the construction of the single- and two-storey house models with a footprint of 3m × 4m at Nebelivka, 2014 (data from S. Johnston) House models

Single-storey

Two-storey

Total volume of timber (m3)

2.52

3.44

No of trees of 0.15m diameter and 4m length required

35

48

Total volume of wattle (m3)

0.29

0.56

Total volume of daub (m3)

2.95

5.61

Comprising:

 

 

Clay at roughly 80% by volume (m3)

2.36

4.49

Temper at roughly 20% by volume (m3)

0.59

1.12

Total volume of roofing material (m3)

3.00

3.00

b: Estimate of quantities of materials used in the construction of full-sized single- and two-storey houses with a footprint of 4.8m × 12.7m at Maidenetske (Müller & Videiko, 2016) Full sized houses of 4.8m × 12.7m (Müller and Videiko, 2016)

Single storey

Two storey

Total volume of timber (m3)

9.65

17.73

No of trees of 0.15m diameter and 4m length required

135

248

Total volume of wattle (m3)

1.20

2.37

Total volume of daub (m3)

14.65

28.52

Comprising:

 

 

Clay at roughly 80% by volume (m3)

11.72

22.81

Temper at roughly 20% by volume (m3)

2.93

5.70

Total volume of roofing material (m3)

15.24

15.24

192 

 Site Studies

The estimates made for the fuel requirements for a successful, deliberate houseburning show that the quantity of fuel for house-burning far exceeded the quantity of timber for house-construction. Extending the fuel requirements to the burning of a mix of 100 houses requires ca. two million trees or deciduous forest cover of 10km2. This is the first of the two most important conclusions of the Nebelivka experimental programme – the severe logistical implications of burning one house, let alone ten or a hundred. Prehistorians have barely begun to explore the social implications of this logistical requirement.54 By the same token, unburnt or weakly burnt houses, forming one-third of the total houses at Nebelivka, may have been a result of insufficient timber or other variants on the poor or rushed planning of a house-burning. This still implies a deliberate decision to burn the house down, even if this may not have been fully successful. Let us return to Ohlrau et. al’s (2016) figures of 20m3 of construction timber per dwelling, which, despite a higher estimate than that of Johnston, we can broadly accept for this exercise. Ohlrau et al. (2016) have excluded from their calculations the estimated quantity of timber required to burn the houses and produce a ploshchadka. Ohlrau et al. (2016) propose 91ha of woodland for the construction of all of the Nebelivka houses to be burned; our experimental figures suggest that up to 10 times the quantity of fuel was needed to burn the houses than build them, indicating an area of up to 910ha was needed to produce the fuel for this burning. Even if we halve that figure, potentially jeopardizing the production of the ploshchadka, this is still a very large area especially given that these are not densely wooded regions but a forest-steppe zone. We suggest that a single massive burning at the end of the settlement was far less likely than house-burning performed at a constant rate (for example, 10 houses per year), thus not compromising the availability of a scarce resource used not only for building and burning but also for cooking, heating and pottery manufacture. In such a scenario, a forest steppe area of 30ha, with a regeneration cycle of 30 years (Ohlrau et. al 2016) would provide enough fuel for the firing of the burnt Nebelivka houses. If we include the daily needs of timber, too, and increase the area to 50ha, this is a far more realistic figure than 910ha for the availability of this resource in the forest steppe and for the requisite woodland management.

54 In a very recent article, Dal Corso et al. (2019, p. 4) maintain that no extra fuel at all was needed for deliberate Trypillia house-burning, citing Harper's (2017, p. 46) claim that mature house-timber had dried sufficiently for ready combustion. However, no experimental research is quoted to support this claim.



The Experimental Programme 

 193

4.4.1.4 Issue 4: Deliberate House Burning – the Alternatives The debate over deliberate house-burning still continues in European prehistory, more than 40 years after the seminal work of Zinkovskiy (1975), extended by Stevanović (1997). Contrary opinions such as the following are still rife in Balkan prehistory: We could therefore think of the deliberate burning of some such, but perhaps finite, unit within the tell at Belo Brdo, as an ongoing coring programme and limited test excavation some 60m from the centre of the tell do not show burning everywhere in late Vinča levels. This could leave open the possibility of accidental burnings that got out of hand, rather than deliberate, malicious or otherwise aggressive fire-settings. But the formal chronological models presented here may weigh in favour of a deteriorating social context, with putative deliberately aggressive burnings repeated within a generation or so (Tasić et al. 2015, pp. 1078–9).

To the contrary, the Nebelivka experiment has demonstrated the requirement for large quantities of fuel to produce a ploshchadka, as did the Cucuteni Archaeological Park experiments (Cotiugă 2009). There are at least four problems with accidental house combustion or even burning as a result of a military attack: 1. the floor and wall daub coverings are poor conductors of heat, which would have protected the unexposed structural timbers from fire-damage; 2. the poor heat conduction of the wall daub would have prevented a fire from spreading to nearby timber-framed houses55; 3. only the roof of a neighbouring house would have burned but not the rest without additional fuel to maintain the fire; and 4. even if the houses were tightly packed, as in the case of Balkan tells, the outward collapse of walls onto neighbouring houses would not have sustained a damaging fire without additional fuel; while the walls may have baked, they would not have collapsed. All of these reasons make it highly improbable that a complete combustion of a timberframed, wattle-and-daub house leading to the creation of both a ploshchadka and vitrified daub would have been possible through an accidental fire or even a military attack. This is the second of the two most important conclusions of the Nebelivka experimental programme. In summary, the Project’s experimental programme combined three elements – the building of two 4m × 3m ‘Neolithic’ houses – one one-storey and the other twostorey, the burning of the two-storey house one year later and the excavation of its burnt remains two years later. The programme shed light on four important issues. The burning of the experimental house showed that it was possible to create a Trypillia

55  In the Nebelivka house-burning experiment, the one-storey house stood only 2m from the burning two-storey house yet its outside wall daub was barely warm to the touch and the inside of the wall was not warm at all.

194 

 Site Studies

ploshchadka as well as produce true vitrified daub – showing a firing temperature of close to 1000oC. This was the first such house-burning experiment to produce either of these results – probably because of the high density of timber for burning. Issue 2 focussed on the reproducibility of certain features regularly found in, or claimed for, Trypillia excavations. There was little evidence in favour of the hypothesis of construction burning. By contrast, the question of whether house-walls survived a high-temperature burning could be demonstrated both by the wall panels themselves and by the multiple daub impressions produced by the wattle. In addition, a reliable method was developed for distinguishing between one- and two-storey houses from excavated remains of daub. The nature and quantity of fuel needed for successful house-burning comprised Issue 3. Here, the surprising result was that a successful burning required many times more timber than for the house-building  –  in the Nebelivka experiment, up to 10 times the amount of timber. This result stimulates many important implications for European prehistory in any region, such as the Central Balkans, with a high proportion of burnt houses on excavated sites. The final, fourth issue tackled the fraught question of whether burnt house remains were the product of accidents, attacks or deliberate household action. Although there are several strong arguments for the third explanation, the issue of the high level of fuel required is the final argument in favour of deliberate house-burning in most cases. It is now time to turn to the accounts of the various excavations carried out under the aegis of the Project. It is appropriate to begin with the largest, joint excavation that the Ukrainian and Durham sides managed to complete – the excavation of the Mega-structure.

Bisserka Gaydarska, Marco Nebbia, Mykhailo Videiko, John Chapman, Manuel Arroyo-Kalin, Tuukka Kaikkonen & Svetlana Ivanova 4.5 Joint Excavations Bisserka Gaydarska, Marco Nebbia, Mykhailo Videiko & John Chapman 4.5.1 The Mega-Structure

4.5.1.1 Introduction Since Assembly Houses represented rare buildings, it was decided to explore the largest example through excavation in summer 2012. The preliminary description of the ‘Mega-structure’, as the building has been termed, has been presented elsewhere (Chapman et al. 2014), with an account of the drastically differing interpretations of the Ukrainian and British sides. In this account, we shall summarise the British view before making an assessment of the meaning of such buildings for overall megasite settlement order.



Joint Excavations 

 195

The excavation of the largest of the three large structures identified in the geophysical investigations in summer 2009 at Nebelivka (Chapman & Videiko 2011) took place over eight weeks in the summer of 2012. Samples of each context were collected for dry-sieving and froth-flotation. All finds in Weeks 2–7 of the excavation were plotted with Total Station co-ordinates. The burnt daub and features were drawn in the field, with later digitisation transferred to a GIS platform. The Mega-structure was built on a granite rockhead that would have been close to the surface, if not actually appearing as a surface feature. The large bi-partite structure covered an area of 1120m2, with 720m2 represented by burnt remains (Fig. 4.31; for GIS data, https://doi.org/10.5284/1047599 Section 5_1_4). The Mega-structure was divided into two large areas – the Eastern, unburnt part and the Western part, partly burnt and partly unburnt. In the former, there were relatively few features, which could not be differentiated into earlier or later phases. By contrast, the latter was defined in the Eastern area primarily by a mass of burnt daub normally interpreted as the remains of the deliberate burning of the structure. The unpicking of the sequence of construction remains and destruction debris proved to be the principal challenge in the excavation. The stratigraphy of the burnt part of the Western part of the Megastructure can be divided into four Phases: Phase 1 – pre-Mega-structure; Phase 2 – use of Mega-structure (Fig. 4.35 upper); Phase 3 – two phases (3 Lower and 3 Upper) of deposits representing the destruction of the Mega-structure (Figs. 4.32–4.33); and Phase 4 – the soil fill above the destruction deposits (for long section, see Fig. 4.37 upper) (cf. daub plot for Phases 2 and 3 combined: Fig. 4.34). Four animal bone samples were recovered for AMS dating, one of which produced an unacceptable result in the centuries AD. Bayesian modelling of the remaining dates within the overall model for the site (see Chapter 4.8.6 and Fig. 4.62) indicates the highest probability of the Mega-structure’s use covered 0–160 years, from 3970–3840 to 3900–3760 BC (all at 95.4% probability) (Fig. 4.38/2–5). The following account focuses on the principal features of each Phase of the Western part.

4.5.1.1.1 Phase 1: Pre-Mega-Structure Features There are currently three contexts indicating prior deposition in the area subsequently covered by the Mega-structure: a filled pit beneath the level of the base of the podium, a foundation deposit under a Platform and a post-hole below the central open area. These contexts indicate a minimal presence in the area where the Mega-structure was subsequently constructed  –  probably of less significance than the natural granite rockhead in this part of the megasite.

196 

 Site Studies

Figure 4.31: Kite photo of Mega-structure, Nebelivka: North to Right side; burnt area 36m East-West (by M. Houshold).



Joint Excavations 

Figure 4.32: Digitised remains, Nebelivka Mega-structure; Phase 3 Lower (by M. Nebbia).

 197

198 

 Site Studies

Figure 4.33: Digitised remains, Nebelivka Mega-structure; Phase 3 Upper (by M. Nebbia).



Joint Excavations 

 199

4.5.1.1.2 Phase 2: The Construction of the Mega-Structure The digitisation of all the construction daub shows the overall features of Phase 2 of the Mega-structure (henceforth, 'East' and 'West' refer to parts of the built area) (Fig. 4.35 upper; cf. kite photograph Fig. 4.31). Linear daub concentrations indicated the place of external and internal walls. Parts of the East area and the West wall revealed fired clay ‘slots’ to support vertical planks (Fig. 4.35 lower). It remains doubtful that a roof or rooves covered the whole of the Mega-structure. The Eastern end has by far the most intensive concentration of daub, which is broadly similar to a burnt house ploshchadka. In two parts in the East area, evidence of two-phase construction consisted of a threshold which had been blocked in Context 236. The initial interpretation of the complex stratigraphy of the Eastern part of the Mega-structure was based on a kite photo, allowing a bird’s-eye view of this rather large building (Fig. 4.31). This suggested that it consisted of seven somewhat irregular ‘rooms’. However, the experimental programme of building, burning and excavating of the Trypillia-like dwelling (see above, Chapter 4.4) made us reconsider many of our assumptions, including the interpretation of the Mega-structure. The complete digitisation of all features and destruction daub further reinforced the need for a fresh look into the possible layout of the Eastern part. Thus, what was initially identified from the kite photo as some kind of dividing lines suggesting the separation of space was outlined now with greater precision, revealing the existence of two large rooms (Fig. 4.35 upper). The ‘voids’ that were observed and documented during the excavations may derive from long-decayed sleeper beams that once formed the base of walls. The oblique voids or lines visible on the kitephoto that appear somewhat straightened on the digitized plan as a result of the regularisation of the geo-referencing may represent long-decayed, fallen timber uprights. Obliquely fallen timber was observed during the burning and excavating of the experimental house (Fig. 4.36). The rest of the Mega-structure was closer in structure to the Assembly Houses as shown in the geophysical plots (Figs. 4.8–4.9), viz., a largely open structure surrounded by wattle-and-daub walls whose destruction daub has remained in place (most of the North Wall), fallen inwards (the Western part of the North Wall) or fallen outwards (most of the South and West Walls). The daub lines in the Western part suggest a corridor running along the West wall, probably roofed over. A reconstruction of the Mega-structure according to the British view shows these features intact (Fig. 4.38/1).

200 

 Site Studies

Figure 4.34: Digitised remains, Nebelivka Mega-structure; Phases 2 and 3 Lower & Upper (by M. Nebbia).



Joint Excavations 

 201

Figure 4.35: Upper: digitised remains, Phase 2. Key to Context Numbers: Fired Clay Bin – 80; Platforms – 6, 46, 58, 89, 176 & 272; Podium – 29; Threshold – 120; Platform 257 was excavated on the last day of Week 7 and therefore not digitised; it is found in Grid Square C21–22; lower: fired clay slot, Nebelivka Mega-structure (upper by M. Nebbia; lower by J. Chapman).

202 

 Site Studies

Figure 4.36: Burnt timber remains, Nebelivka Experimental Burnt House Excavation: upper: burnt timber fallen obliquely, Context 419; lower: timber void, Context 226 (by J. Chapman).



Joint Excavations 

 203

Figure 4.37: Upper: long North-South section (the width of both walls reflects daub tumble out from the walls); lower: three sections across podium, Nebelivka Mega-structure (by C. Unwin).

204 

 Site Studies

Figure 4.38: (1) Durham reconstruction; (2)–(5) Bayesian plots of AMS dates, Nebelivka Megastructure (1 by C. Unwin; 2–5 by A. Millard).



Joint Excavations 

 205

A total of 10 interior features was found in the Mega-structure (Fig. 4.35 upper). All of these features were sufficiently well preserved to indicate construction on the ground floor rather than a upper floor construction followed by a fall to the ground floor. The first feature comprised a 2m-long dark-fired clay threshold in the Eastern end of the building, near massive daub fragments probably representing a monumental superstructure of the East entrance. The slope of 0.25–0.30m from the open unburnt part of the Mega-structure to the burnt area enhanced the dramatic effect of this entrance. This means that anyone approaching the Mega-structure from the East side would have been confronted by a 2m ± high wall, with a possibly monumental entrance at the top of a slope. A 10m-long fired clay ‘podium’, an estimated 15cm high, was built along the inside edge of the South wall. The podium was built of clay (Contexts 319 and later 219) on a sandy loess-derived soil (Context 419) (Fig. 4.37 lower). The upper surface of the podium became cracked as a result of the fire, as happened with the Platforms. In most parts of the podium, the daub slabs were found beneath destruction daub (Phase 3), showing that the South wall fell on top of the podium. However, in one area (Context 161), the finding of podium fragments above wall daub shows a reversal of the usual sequence. Eight fired clay Platforms of various sizes had been built up with two to four layers of fired clay, some decorated with incised motifs or a red painted wash. The total of eight Platforms in the Mega-structure is unprecedented for Trypillia buildings. Two of the four Platforms placed inside the East end rooms were found in each room (North room – Platforms 58 and 272; South Room – 89 and 176), while the remaining four were built in the central open space (Platforms 6, 46, 257 and 281). Since there were no signs of burning on the Platforms, it is suggested that they were not used as hearths or fireplaces but, rather, for temporary displays of special items during Megastructure ceremonial. Very few objects remained on the upper surface of the Platforms – exceptional finds were the two animal bones placed on Platform 6. However, the largest Platform 46, at 3.6m × 3.5m, shows evidence of the incorporation of cultural material in its construction. The lowest layer was built on yellow loess (Fig. 4.40). Each successive Platform layer of sandy clay was slightly larger than its preceding one. The basal Platform layer contained one large and two tiny animal bones; old, worn, small sherds, daub and Platform daub were found in the body of the second layer (Fig. 4.46/3). Platform 89 was the most highly decorated; it was lifted by a team led by the conservator Stanislav Fedorov and moved for reconstruction and display in Kirovograd Historical Museum (Fedorov 2015). Platform 281 was the smallest and most fragmented, with a diameter of 0.5m, and lay so close to the largest platform (Context 46) that it may have been made in a different time-interval.

206 

 Site Studies

Figure 4.39: Upper: general view from East; lower: Stage 4 placement of vessel from North; Fired Clay Bin, Nebelivka Mega-structure (by J. Chapman).



Joint Excavations 

Figure 4.40: Platform 46 from East, Nebelivka Mega-structure (by J. Chapman).

 207

208 

 Site Studies

The multiplicity of the Nebelivka platforms showed a complex, changing sequence of creations designed to stage ritual performances in the Mega-structure. These performances varied according to the size and setting of the Platforms, whether located inside or in the open courtyard. Perhaps no more than six persons would have stood round Platform 281, while up to fifty participants could have encircled Platform 46 (Figs. 4.40 & 4.46/3). The four expansion phases of Platform 46 combined a commitment to place over an as yet undefined temporality presenced by old objects with the desire to create increasingly impressive visual effects. However, at the end of the ceremony, most offerings were removed, either by the Nebelivka ritual leaders or by those who had brought the offerings there in the first place, to presence the Megastructure by retention of objects sanctified by their participation in the performance. A large fired clay ‘bin’ was constructed near the largest Platform; according to the field observations of one of the authors (JCC), it enjoyed a long and complex biography in ten stages, which did not include the storage of grain (Fig. 4.39): – Stage 1: after clearing and flattening of the area to form a stamped floor, a low fired clay wall was made in what appears to be a continuous manner; – Stage 2: a clay surface was constructed, which was found cracked as in a Platform, possibly over the whole of the walled area but perhaps in the central area only; – Stage 3: the construction of the central fired clay feature – an altar or a pillar? – up to 10cm high. This feature has been so badly damaged that no original walls were left. The primary use of this feature then followed. – Stage 4: some time after the stamped floor was made, a large vessel was placed in the South-West corner (Fig. 4.39 lower) but the dark brown fill between the vessel and the floor suggests this was not at the start of Stage 2. – Stage 5: traces of burning on the daub fragments around the central feature, on the feature itself and on the Stage 2 cracked clay surface must have occurred before the general collapse of the Mega-structure. – Stage 6: the collapse of the central feature onto the cracked clay surface (Stage 2), with burnt daub falling onto the cracked clay surface and possibly even the stamped floor. This also must have occurred before the general destruction of the Mega-structure. – Stage 7: The infilling of the now-disused bin with dark brown fill mixed with sherds and daub fragments up to the level of the top of the central feature. – Stage 8: following this infilling, the bin was ‘closed’ by the placing of a complete, upside-down grindstone in the North-West corner (Fig. 4.39 upper). – Stage 9: a daub scatter related to the destruction of the Mega-structure was deposited over the surface of the infilled bin, perhaps at the same time as the grindstone surface was cracked by the temperature of the fire. – Stage 10: in post-Trypillia times, the bin suffered damage by animal burrows, while plough damage removed the top of the bin and the top of the large vessel in the South-West corner.



Joint Excavations 

 209

This biography of a fired clay bin shows a temporality to which we rarely have access in the megasite. Although the time dimension is not defined, the stratigraphic sequence shows three main phases: a construction phase and an abandonment phase, both occurring before the destruction of the Mega-structure.

4.5.1.1.3 Phase 3: The Destruction of the Mega-Structure (Figs. 4.32–4.33) There is a fundamental assumption that Trypillia houses have been burnt down deliberately at the end of their lives (Kruts, 2003; Burdo et al. 2013; Chapman 2015). However, the distribution of fired clay daub across the Mega-structure is by no means continuous nor always massive, revealing patches of dense, often vitrified daub, zones of medium density daub with little or no vitrification and areas of low-density daub with no traces of vitrification (Fig. 4.41). These findings emphasise the combination of ‘dwelling-house destruction’ (the East end) and ‘Assembly House destruction’ (the central area and West end), so far unique in the Trypillia world. There seems little doubt that there were major variations in the temperature at which different parts of the Mega-structure burned down (see Shevchenko, Chapter 4.9). This may have been a by-product of the conditions of the fire or perhaps the different burning strategies designed to burn different areas in different ways. According to one of the author’s field observations (JCC), stratigraphic evidence from more than 10 contexts showed the covering of the Phase 2 living surface of the Mega-structure with a thin layer of dark soil prior to the first daub destruction deposits. The features where the thin soil layer was found include the podium, two Platforms and the fired clay bin. It seems probable that this soil was derived from the local chernozem and blew into the Mega-structure over a period of time whose duration is currently difficult to assess. The suggestion is that a Mega-structure that was relatively open may have been abandoned for a period of time before it was burned down. The digitisation of all destruction daub led to the insight that there were two stratigraphic stages of destruction daub – here termed Phases 3 Lower (Fig. 4.32) and 3 Upper (Fig. 4.33; cf. combined Fig. 4.34). These data show that different parts of the Eastern block, the North wall, the Western block and the Southern area fell ‘earlier’ than other parts of the building. However, these two stages of mostly wall collapse may well have happened during the same destruction event, as indeed happened during the recent house-burning experiment at Nebelivka (see Section 4.4). In Chapman’s view, there were three stages in the biography of the Megastructure’s destruction: (1) the temporary or permanent cessation of social practices inside the building; (2) a period of as yet unknown duration allowing the build-up of thin levels of chernozem-derived soil layers within the Mega-structure; perhaps the Mega-structure was not used in this period; (3) the burning and collapse of the building to produce the ploshchadka.

210 

 Site Studies

Figure 4.41: Plot of daub firing temperatures and vitrified daub, Nebelivka Mega-structure (by M. Nebbia based upon information from N. Shevchenko); Colour Key for daub firing temperatures: blue: 200–4000C; yellow: 400–9000C; orange: >9000C.



Joint Excavations 

 211

4.5.1.1.4 Phase 4: After the Destruction of the Mega-Structure The main characteristic of the period after the burning of the Mega-structure was a period of chernozem formation indicating an absence of cultural activity above where the Mega-structure once stood. Some 0.80–1m in thickness, this soil covered the entire megasite and was possibly partly caused by secondary erosion of loess blown into the site in the historic period. One can suppose that this period of soil formation was, at the same time, a period of little local deposition of artifacts or ecofacts; any finds in the post-Trypillia chernozem horizon would constitute an internal residue sensu Kuna (2015) (see above, p. 53). The ploughing of the soil above the megasite in general, and the Mega-structure in particular, was so deep as to leave traces of furrows in the top of the ploshchadka. It is this modern ploughing that has removed a large quantity of Trypillia pottery from its original location and created a large and varied plough-zone ceramic assemblage of at least 1,500 sherds with Total Station recording. The best guess that we can make for the source of this plough-zone assemblage is near the top of the ploshchadka – a notion that would lend support to Burdo et al.’s (2013) view that there was much deposition on the burnt remains of Trypillia houses, viz., on the top of the ploshchadka. Such deposition may have included the large number of unburnt bones found within the destruction layer, which may have been deposited on the house after the destruction event, only to become included in the daub layer by subsequent disturbance. The demonstration of the deliberate deposition of objects would represent secondary refuse sensu Schiffer (1987).

4.5.2 Interpretation A consideration of the principal features of the Mega-structure suggests that the basic elements of the Trypillia house have been borrowed and adapted to fit the great size of what remains a public building but one without the depositional characteristics of a ritual or administrative centre. The layout of the rooms and internal features in the Mega-structure do not fit any of the ‘typical’ domestic house layouts as defined by Chernovol (2012, Fig. 8.8). Korvin-Piotrovskiy (2015) proposes two alternative interpretations of the Mega-structure: the first, which he favours, is a complex of related domestic structures with household functions; the second as a ritual complex of seven rooms, each with an altar for an extended family. Instead, we propose that the Mega-structure was a monumental public building (Fig. 4.38/1), visible from several km on the South part of the micro-region, which created the potential for major congregations of several hundred people within the open courtyard in the Eastern part and the inner Central area of the Western part. The multiplicity of platforms was a result of the varied social groups (clans or lineages rather than families) participating in Mega-structure ceremonies. The destruction of the Megastructure by fire would have been one of the great ceremonial manifestations of the

212 

 Site Studies

Trypillia world - impacting on the lives of both Nebelivka inhabitants and visitors to the megasite. Offerings to the dead building may well have continued for years after the fire as a contribution to cultural memory and the heightened place-value of the Mega-structure. The Mega-structure can be perceived as a microcosm of the entire megasite, with its built-up space contrasting with the open inner areas in the same way as the megasite combined an inner, open area with an outer, dwelling area.

John Chapman, Manuel Arroyo-Kalin, Tuukka Kaikkonen & Svetlana Ivanova 4.5.3 The Barrow56 Barrow burials were common in the Eurasian steppe zone from the late 4th millennium BC onwards (Ivanova 2014), with examples known from the forest-steppe zone of South Central Ukraine. Many barrows were identified and recorded in the Nebelivka fieldwalking programme (see Chapter 3.2). While few of the Ukrainian barrows have been excavated, the earliest date to ca. 3200 BC, indicating an overlap of several centuries with the latest Trypillia settlements (dating to 29/2800 BC). The single barrow found on the area of the Nebelivka megasite is therefore likely to post-date the megasite by 500 years to a millennium – and maybe even longer (see below, p. 250ff.) (Ivanova, S. 2015). It was located ca. 65m from the closest house – the innermost burnt house in an inner radial street in Quarter F – and marked the ancestral (Trypillia) transition from inhabited area to inner open congregational space. The barrow was 16m in diameter and rose 2.2m from the current ground surface (Fig. 4.42). Looting of the barrow in the 1980s left a large rectangular pit excavated to a depth of 3m. The North profile of this pit was cleaned and prepared for geoarchaeological recording and sampling to a depth of 3.5m. The South-facing section of the test pit (https://doi.org/10.5284/1047599 Section 5_7_2_2) showed the following stratigraphy (Table 4.14). Worthy of particular interest was the fact that the buried A horizon did not resemble the deep chernozem A horizon of the current soil mantle. This could point to environment changes in the locale during and since the megasite occupation took place, or be an artefact of the considerable surface modification of the buried A horizon prior to burial by the building of the mound.

56  See https://doi.org/10.5284/1047599 Section 5.7.



Joint Excavations 

 213

Figure 4.42: General view of Nebelivka barrow from South-West (by J. Chapman).

Table 4.14: The barrow stratigraphy (by S. Ivanova). Layer Description 5

barrow fill – sediments identical to present chernozem B horizon (1.4m thick)

4

no clear sign of spoil removed from burial pit; traces of black and yellow inclusions suggest this occurred at a depth of 0.6m from top

3

post-Trypillia buried occupation layer in the form of a black chernozem A horizon – uneven with varied thickness (up to 10cm thick) and subjected to in situ trampling; different morphology from that of current chernozem A horizon

2

Trypillia-age buried soil – light brown chernozem with small pieces of yellow daub (60–65cm thick)

1

sterile yellow loess (sampled to a thickness of 1.1m)

214 

 Site Studies

The analysis of molluscan assemblages from the barrow profile suggests that grassland-type habitats have dominated the immediate location in poorly dated but in any case Holocene sediments deposited since the late Pleistocene (viz., the loess stratum). Vallonia excentrica dominates nearly every sample, and there are notable absences of many snail species indicative of climax Holocene forest. Indeed snail diversity seems slightly higher earlier in the Holocene, with the assemblages potentially indicative of a mix of established grassland and colonising woodland. By the time of the megasite, and probably significantly earlier, snail faunas became less diverse, and indicate uniform grassland over a significant local area. This finding is potentially linked with the important discovery that chernozem formation had already taken place on the Nebelivka promontory by the time of the megasite occupation (see Section 4.1.2). The discovery of a fragment of a Greek imported amphora dated to the Early Iron Age in the intra-site gridded collection (see above, p. 80) can be interpreted in at least two ways: (a) this sherd dates a visit to the Nebelivka promontory by a group using imported pottery who then built the barrow and discarded the sherd; or (b) the amphora-using group visited the promontory and discarded the sherd in awe of an ancestral barrow to which they could visually relate as part of their ‘heritage’.

Bisserka Gaydarska, Marco Nebbia, Stoilka Terziiska-Ignatova, Patricia Voke & John Chapman 4.6 Excavations, Durham Side Bisserka Gaydarska, Marco Nebbia & John Chapman 4.6.1 The Test Pits57

Preliminary test pit excavation of two test pits was started in the 2009 season, with a view to testing the precision of the geophysical investigation of apparently burnt houses; four more test pits were dug in 2012. The success of the 2009 and 2012 test pit trials (https://doi.org/10.5284/1047599 Section 4.8) led to the excavation of a further 82 test pits – 41 in 2013 and 41 in 2014/15 (Fig. 4.43). Not only did the test pit excavations produce vital organic samples for AMS dating but they proved to be a major source of information about techniques of house construction and the taphonomy of houseburning. The excavations also contributed to a large sample of Trypillia pottery, many animal bones and a range of special finds. All three types of house were investigated through test-pitting – burnt houses, unburnt houses and Assembly Houses (Fig. 4.44). Those Assembly Houses which were test-pitted showed few finds and low concentrations of destruction daub (e.g., Test Pit 27/3: Fig. 4.50/3).

57  See https://doi.org/10.5284/1047599 Section 5.3.



Excavations, Durham Side 

Figure 4.43: Distribution of all excavated areas, Nebelivka (by M. Nebbia).

 215

216 

 Site Studies

Figure 4.44: Distribution of burnt houses, unburnt houses and Assembly Houses (by M. Nebbia).



Excavations, Durham Side 

 217

The recording of the test pit excavations followed a standard format, with predefined contexts for ease of comparability. Five contexts were defined according to the ‘standard’ stratigraphic sequence (from top to bottom) (Fig. 4.45). The depths of the test pit stratigraphies are shown below (Fig. 4.47). – Context 1: the ploughsoil, equivalent to the A horizon of the chernozem soil. – Context 2: the lower part of the ploughsoil, where it can be distinguished by colour and texture from Context 1. In the rare cases where no destruction daub was found, the Context 2 unit could be very deep (e.g., Test Pit 26/7, where Context 2 is 0.25–0.6m in depth); – Context 3: the destruction daub which resulted from the burning of the structure. This was sometimes a single layer but there were many examples of two layers of destruction daub (3 Upper, 3 Lower) and occasional examples of three layers (3 Upper, 3 Middle, 3 Lower). – Context 4: the living surface or floor level of the structure. In an earlier excavation (2013 season), we had misinterpreted the lower layer of destruction daub as a solid baked clay floor; in fact, there were very rare examples of a solid floor level, with most structures having a stamped earth living surface. Instead of ascribing separate Context numbers to fired clay features (e.g., boxes) or dug features (e.g., pits), these were described with the site-wide Context numbering system but with descriptors such as Context 4/Feature 1. – Context 5: the deposits which pre-dated the construction of the excavated structure. Standard practice was to excavate 20cm below the base of Context 4, although deeper finds were occasionally made. In the case of Contexts where daub and pottery was scattered, it was generally inferred that material from the living surface had been pressed down into a marginally lower level (e.g., Test Pit 26/2). However, those rare Contexts where features were encountered or larger quantities of daub and pottery were found (e.g., Test Pit 24/4) were taken to indicate pre-construction activity in that part of the megasite. Soil micro-morphological investigations of a burnt house (Test Pit 1/3) and an unburnt house (Test Pit 1/4 – Sondazh 5, 2014: Burdo & Videiko 2016; https://doi. org/10.5284/1047599 Section 5.2.3) showed some surprising results. The section of Test Pit 1/4 was predominantly composed of chernozemic soils and had little evidence of anthropogenic modifications. However, in contrast to the burnt house, this test pit had far less burnt daub but, interestingly, more abundant microscopic charcoal. Although daub is present, it was in smaller quantities than in the burnt house. Furthermore, the daub fragments appear less rubified, possibly indicative of lower firing temperatures. These observations have implications for how house architecture and burning are understood at Nebelivka. Furthermore, the larger abundances of charcoal in the ‘unburnt’ house as opposed to the burnt house are significant for understanding contextually variable taphonomic processes at the megasite. One resolution of this issue is that the weak anomalies interpreted as ‘unburnt’ houses

218 

 Site Studies

Figure 4.45: Typical Test Pit stratigraphy, Test Pit 26/6, Nebelivka (by C. Unwin).



Excavations, Durham Side 

Figure 4.46: Sections across (1) Fired Clay Bin; (2) Contexts 215 & 310; (3) Platform 46; and (4) Contexts 210 and 310 (by C. Unwin).

 219

220 

 Site Studies

Figure 4.47: Depth of burnt houses in excavation units and Test Pits (by C. Unwin).



Excavations, Durham Side 

 221

may reflect structures burnt at lower temperatures which were insufficient to destroy all of the charcoal but also too low to create daub that survived six millennia of krotovina or other soil processes. The test pits subsumed much architectural variation, partly caused by the location of the test pit in a particular place in a house. Houses were divided into nine ‘zones’ and the location of the test pit was recorded. However, few cases of a significant difference at the 0.1% level were found in any chi-square test assessing the importance of the position of the Test Pit in a specific House Zone. One exception is the discovery of vitrified daub – generally regarded as a sign of a ‘hot-spot’ in a house fire. Almost ⅔ of all vitrified daub came from test pits located in the centre of the house (Zone 9) – a finding related to the concentration of added timber fuel placed in the centre of the house. However, there is no correlation between test pit zone and the discovery of either decorated daub or construction daub from living floor features such as podia or bins. A self-defining feature of burnt houses in the test pits was the presence of often large quantities of destruction daub (e.g., Test Pit 15/2: Fig. 4.48 lower). Conversely, the so-called ‘unburnt’ houses contained no destruction daub (e.g., Test Pit 26/7) or very few small fragments (e.g., Test Pit 1/4: Fig. 4.50/1). While unburnt houses showed no level of destruction daub, up to three levels were found on some houses but two levels of destruction daub were more common than three (e.g., Test Pit 33/1, with two levels: Fig. 4.49 upper). An important feature of one in nine investigated burnt houses was the production of a low mound of burnt debris which survived in test pit sections (e.g., Test Pit 24/4: Fig. 4.50/4). These mounds would have remained visible on the surface for some time after the collapse of the house, marking the place of the former house and preventing any further building on the place. We can term these mounds ‘memory mounds’, since they preserved the memories of the long-dead house into the future, preventing the abandonment of the megasite from becoming a loss of memory. The slow accumulation of memory mounds across the megasite effected a gradual transformation of the site from a living site into a site where the living and the dead were more in everyday contact. An important question for the megasite was the ratio of one- to two-storey houses on the site as a whole and in different Quarters. The experimental programme showed that the discovery of scattered platform daub meant it may have fallen from an upper floor, viz. of a two-storey house, while in situ platform daub built on the living floor indicated the existence of a one-storey house (see above, Chapter 4.4). Close study of the distribution of in situ and scattered platform daub in 38 of the test pits (Fig. 4.51) showed that the majority of these test pits contained insufficient fragments of platform daub to determine clearly that they had fallen from an upper floor.

222 

 Site Studies

Figure 4.48: Upper: section of thick pile of destruction daub, Test Pit 13/3; (a) South-facing; (b) Eastfacing; lower: sections of platform below destruction daub, Test Pit 15/2; (a) East-facing; (b) South-facing (by L. Woodard). Numbers in Figs. 4.48–4.50 refer to general pit stratigraphic sequence in Fig. 4.45.



Excavations, Durham Side 

 223

Figure 4.49: Upper: sections of two well-defined layers of destruction daub, Test Pit 33/1; (a) SSEfacing; (b) WSW-facing; lower: sections of two-storey house with platform above destruction daub, Test Pit 26/5: (a) West-facing; (b) North-facing (by L. Woodard).

224 

 Site Studies

Figure 4.50: (1): West-facing section of unburnt house, Test Pit 1/4; (2): North-facing section of burnt house with platform daub with pit under floor, Test Pit 17/1; (3) SW-facing section of platform in Assembly House, Test Pit 27/3; (4) section of mound of burnt house debris, Test Pit 24/4; (a) ESEfacing; (b) NNE-facing (by L. Woodard).



Excavations, Durham Side 

Figure 4.51: Distribution of Test Pits with in-situ vs. dispersed platform daub (by M. Nebbia).

 225

226 

 Site Studies

Detailed consideration of the type, fragmentation and location of platform daub can give valuable indications of the kind of house in which the daub had fallen. Unusually, there were signs of in situ platform daub in two ‘unburnt’ house (Test Pits 29/3 and 33/2), with the former containing only a single piece of platform daub. This suggests one-storey houses with platforms constructed on the ground floor. A second case – this time of two test pits with scattered platform daub (Test Pit 1/1 and 19/1) – showed platform fragments that had fallen from an upper floor, with the implication of a two-storey house. A third case found in three or four further test pits showed that scattered daub had probably fallen from an upper floor, again with the implication of a two-storey house. The fourth and final case showed four test pits with both types of platform daub – in situ and fragmentary – but with opposite results. In Test Pits 15/2 (Fig. 4.48 lower) and 17/1 (Fig. 4.50/2), in situ daub showed the existence of a platform built on the living floor but with plough or other post-depositional processes causing the scattering of daub from this feature. By contrast, some of the platform daub that had fallen from the upper floors in Test Pits 20/2 and 26/5 (Fig. 4.49 lower) fell into concentrated areas of daub resembling an in situ platform. These examples show how careful structural and contextual analysis can unpick the collapse sequence of a burnt house. Very rarely do we find stratigraphic evidence for the most obvious indication of a two-storey house – two layers of destruction daub separated by platform daub – (e.g., Test Pit 26/5: Fig. 4.49 lower). The variation in numbers of destruction daub layers (Fig. 4.52) gives an idea of the complexity of the house destruction process, with multiple layers suggesting walls falling on top of each other. The distribution of test pits with these varying numbers shows some regional and some zonal differences. There are no examples of three layers in the South-West half, with all four examples in the North-East part. In terms of zones, there was a clear domination of houses with two layers in the inner radial streets, while houses with one and two layers were evenly distributed in both house circuits. The inner circuit was the only zone lacking examples of no layers at all. The various destruction processes implied by this finding may suggest the prevalence of some modes of burning over others in the inner radial streets compared to the house circuits, although this is not a categorical difference. These results indicate that we can still distinguish between one- and twostorey houses on the basis of the distribution of platform daub where there are good quantities of such material. This suggests that, where such a judgment was possible, many more two-storey than one-storey houses were found in the test pit houses, with an estimated ratio of 5:1.



Excavations, Durham Side 

Figure 4.52: Distribution of number of layers of Destruction Daub by Test Pit (by M. Nebbia).

 227

228 

 Site Studies

The manner in which Nebelivka houses were destroyed has been discussed in terms of a dichotomy – unburnt vs. burnt (see above, p. 128). Although Roe (above, pp. 76–80) has noted a statistically significant difference between the proportion of burnt houses in the main house circuits over the inner radial streets, it is possible to investigate the distribution of ‘unburnt’ (or, better, ‘poorly fired’) houses at a Neighbourhood level (Fig. 4.54 lower). Using a sample of six Quarters containing 80 Neighbourhoods, it can be shown that four Neighbourhoods contain only poorly fired houses, 31 Neighbourhoods contain only burnt houses and the remainder – a majority at 60%  – contain a mixture of burnt and poorly fired houses. A division of these Neighbourhoods by Quarter shows that, in comparison with the mean proportion of mixed destruction practices, people living in one group of Quarters (F, I and N) tended to select mixed practices less frequently (ca. 40%) while mixed practices were much more frequent (ca. 80%) for those living in the other Quarters (B, C and L). The higher the proportion of mixed house firing practices, the greater the potential differentiation between houses in terms of labour control and timber/fuel acquisition.

Bisserka Gaydarska, Stoilka Terziiska-Ignatova, Patricia Voke, Marco Nebbia & John Chapman 4.6.2 The Pit in Sondazh 158 Sondazh 1, a trench measuring 6.70–6.95m East – West by 3.97–4.07m North – South, was set over a large geomagnetic anomaly, which turned into a pit so large that it required two seasons of excavations (2013 & 2014). In the South wall of the trench, an identical profile to that found in Sondazh 2 was encountered: an A horizon to 0.45m depth, with a B horizon dominated by carbonate in-washing to a depth of 1m (Fig. 4.53). The first 40cm from the surface were excavated over the entire area in two 20cm spits. Although there were sherds in both layers, clear concentrations of archaeological materials and pit boundaries were not identifiable. This imposed a change of excavation strategy, with subdivision of the trench into a North and a South sector and identification of finds concentrations to define the blurred edges of the pit. A combination of factors – collapsed pit walls in the past, ‘sinking’ of dense material in the loess, later re-cuts and very intensive animal burrowing activity (krotovina) – contributed to a situation whereby the ‘original’ pit walls were difficult, if not impossible, to define and archaeological materials were found ‘outside’ the pit.

58  See https://doi.org/10.5284/1047599 Section 5.4.



Excavations, Durham Side 

 229

Figure 4.53: Sections of Pit, Sondazh 1: (1) East-facing; (2) West-facing; (3) East-facing; (4) Southfacing; (5) West-facing (by C. Unwin).

230 

 Site Studies

A consideration of the pit stratigraphy prompted a division into five ‘Stratigraphic Units’ (SUs), numbered 1 to 5 from the earliest to the latest (Table 4.15). Table 4.15: Stratigraphic division, Pit, Sondazh 1 (by J. Chapman). Stratigraphic Unit

Relative Fill depth from Trench Datum (m) (Level depth (cm))

Depth of Episodes from surface (m)

5

1.75–0.44 (148,25– 149,56)

Chernozem A horizon (topsoil); No Episodes could be recognised in this much of this deposit is a SU. post-Neolithic soil build-up and therefore containing material ploughed up from the uppermost pit layer, including one anthropomorphic figurine.

4

2.3–1.75 (147,75/70– 148,25)

Crumbly black-brown fill in all Zones, with the uppermost fill representing the top of the pit in Trypillia times. The largest number and diversity of Episodes (n = 16), with occasional examples of only bone or only pottery, but often the combination of pottery + bone + chipped stone+ daub; 10 anthropomorphic figurines.

Zone NE/ 1.83–2.01 (148,17–147,99) Zone NE/ 1.84–2.11 (148,16–147,89) Zone NE/ 2.2–2.33 (147,8–147,67) Zone 2&3/ 1.95–2.06 (148,05–147,94) Zone 2&3/ 1.95–2.25 (148,05–147,75) Zone 2&3/ 1.96–2.27 (148,04–147,73) Zone 3/ 1.84–1.96 (148,16–148,04) Zone 3/ 2.01–2.14 (147,99–147,86) Zone 2/ 1.99–2.09 (148,01–147,91) Zone 2/ 2.13 (147,87) Zone 1/ 1.75–1.85 (148,25–148,15) Zone 1/ 2.05–2.15 (147,95–147,85) Zone 1/ 2.11–2.2 (147,89–147,8) Zone 1/ 2.15–2.24 (147,85–147,76) Zone 1/ 2.15 (147,85–147,85) Zone 1/ 2.19–2.21 (147,81–147,79)

3

2.65–2.3 (147,35– 147,75/70)

Crumbly black-brown fill in all Zones; eight Episodes with more varied deposits, including a Bos horn-core, 13 anthropomorphic figurines, grindstones, chipped stone, other bone, charcoal and daub, as well as much pottery.

Zone NE/ 2.31 (147,69) Zone NE/ 2.34–2.52 (147.66–147,48) Zone 3/ 2.59–2.65 (147,41–147,35) Zone 2/ 2.3–2.4 (147,7–147,6) Zone 2/ 2.33–2.44 (147,67–147,56) Zone 1/ 2.32–2.44 (147,68–147,56) Zone 1/ 2.42–2.52 (147,58–147,48) Zone 1/ 2.55 (147,45)



Excavations, Durham Side 

 231

Table 4.15: Stratigraphic division, Pit, Sondazh 1 (by J. Chapman).

Continued

Stratigraphic Unit

Relative Fill depth from Trench Datum (m) (Level depth (cm))

Depth of Episodes from surface (m)

2

2.8–2.18 (147,2– 147,82)

Crumbly black-brown fill Zone 1/ 2.5–2.59 (147,50–147,41) mostly in Zone 2; two Zone 1/ 2.63–2.73 (147,37–147,27) Episodes, containing overwhelmingly pottery but no figurines.

1

3.06–2.49 (146.94– 147.51)

Basal part of two test boxes – one in Zone 1 and the other in Zone 2 – (2013), with lower part of fill in Zone 2 (2014); four Episodes, containing a very high proportion of pottery. one anthropomorphic and one zoomorphic figurine.

Zone 2/ 2.49–2.62 (147.51–147.38) Zone 2/ 2.5–2.4 (147,50–147,60) Zone 2/ 2.7–2.85 (147,30–147,15) Zone 2/ 2.93–3.06 (147,07–146,94)

In terms of the soil micromorphological study, the pit fill and the sediment into which it was cut resemble the natural Chernozem A and B horizons respectively. However, bio-cultural inclusions in the former distinguish it from all the other contexts included in the soil micromorphological study. In contrast to the anthropogenically-enriched fill, the sediment into which the pit was cut lacks any cultural inclusions and closely resembles the natural Chernozem B horizon. The pit fill resembles the Chernozem A horizon in terms of its structure, contents and common bioturbation. However, in contrast to the Chernozem or the house features, the fill has a far higher abundance of coarse organo-cultural inclusions, including burnt daub, decomposed sherds, unburnt bone, excrement and many large charcoal fragments. However, it is not clear whether these inclusions originate from the houses, which are largely devoid of such materials, probably because of good house-keeping (Miller et al. 2010). Among all the samples studied, the charcoal fragments are the most abundant and best preserved in the pit fill. The fragments appear to be concentrated in loose clusters, and often the larger fragments appear to be undergoing further in situ fragmentation and comminution. The virtual absence of charcoal from the natural B horizon also suggests that the charcoal has been confined to the fill and its origin may be traced to occupation activities in the surrounding area.

232 

 Site Studies

Figure 4.54: Upper: plan of base of pit; Pit, Sondazh 1 (by L. Woodard); lower: proportion of burnt and ‘unburnt’ houses by Neighbourhood and Quarter (by J. Chapman).



Excavations, Ukrainian Side 

 233

Structurally, the pit lacks observable micromorphological evidence for distinctive infilling episodes, whether because the samples did not capture transitional boundaries indicative of discreet infilling episodes, the pit was filled in a single event or bioturbation has homogenised structural variation beyond microscopic recognition. Nonetheless, concentrations of finds which we term ‘episodes’ were readily distinguishable from the background noise of low-level sherd discard. A total of 30 episodes was identified (https://doi.org/10.5284/1047599 Section 5_4_4_1_ EPISODES), sometimes marked by indices of burning, with a high proportion of ceramic clusters and rather fewer concentrated animal bone deposits. These depositional episodes were created by cutting into the fill, placing material in the negative features and then re-filling it – often with the same material. Parts of the pit had little evidence for this re-cutting/re-filling cycle, while such cycles were common in other parts of the pit (Fig. 4.53). By far the higher frequency of Episodes came in the middle layers of the pit, especially in SU 4. The initial interpretation is that the beginning and the end of a fill episode was marked in material ways. Two AMS dates from the pit (a third sample had low collagen yield and a fourth date was an outlier) show an overlap at 1 sigma, and within the model for the overall duration of the site (see Chapter 4.8.6), OxA-29598 calibrates to 3950–3780 BC (95.4%) and OxA29599 to 3940–3830 BC (74.4%) or 3820–3760 BC (21.0%) (Fig. 4.63/3). There are so few dates that a sensible estimated duration of pit deposition cannot be made. We can cautiously suggest that the pit was oval in shape at the base (Fig. 4.54 upper) and middepth, while its upper part was much larger and amorphous in shape.

Mykhailo Videiko, Natalia Burdo & John Chapman 4.7 Excavations, Ukrainian Side Mykhailo Videiko & John Chapman 4.7.1 Ditches59

4.7.1.1 Introduction At Nebelivka, the perimeter of the site covers a linear distance of ca. 5.9km, of which 76% (ca. 4.5km) was available for geophysical investigation. The geophysical plot shows a single ditch over much of the available perimeter, specifically the North, West and South sides of the settlement; erosion down the steeper slope of the East side probably removed traces of the ditch in that area (Figs. 4.3 and 4.4). A triple ditch appeared to show up in the South part of the geophysical plot, in Quarter L, and was confirmed by excavation (Sondazh 10).

59  See Chapman et al. (2016); https://doi.org/10.5284/1047599 Section 5.6.

234 

 Site Studies

There are 13 well-defined gaps in the well-preserved parts of the perimeter ditch, with the width of the smallest gap being 10m and the largest 180m. Since there were no geological or pedological reasons to cause the magnetometry to miss existing stretches of ditch in these Quarters, we can assume that these gaps were genuine and thus resemble the kind of porous perimeter boundary well known to British prehistorians in the class of monument known as the ‘causewayed enclosure’ (aka ‘interrupted ditch enclosure’: Mercer 2006; Whittle et al. 1999).

4.7.1.2 Ditch Coring The initial exploration of the oval linear anomaly took place by coring and trial excavation in 2013. The first core was placed in the North-East part of the linear anomaly and reached a depth of 5.50m, without hitting any obvious ditch fill. Instead, there were two principal deposits in the core: a lower reddish silty clay deposit 1.79m in width (4.29m–2.50m) and an upper off-white silty clay deposit 1.80m in width (2.50m–0.30m). Informal testing of these clays showed that both were suitable for pottery-making. It is currently hard to explain how such thick clay deposits came to be present in a feature that may have been a ditch. The second core through the linear anomaly was placed in the North-West part of the megasite. At the base of the 4.50m-deep core, a buried chernozem C horizon had developed over 1.10m (4.50–3.40m), with a 1.40m-thick deposit of alluvial clay above the first C horizon (3.40–2m). Above the alluvial clay, a typical chernozem sequence developed with an A, a B and a C horizon. Intriguingly, the contents of both cores into the so-called ‘ditch’ differed markedly from each other, as did the types of clay found in the two cores.

4.7.1.3 Sondazh 2 This sondazh was laid out over a linear geophysical anomaly just North of Sondazh 1. Despite two extensions, no signs of a ditch profile were encountered (https://doi. org/10.5284/1047599 Section 5.6.1 SONDAZH_2_S-facing_profile). This meant that a priority for excavation in summer 2014 was at least one section cut across the linear anomaly. The initial excavation of sections across the Northern part of the perimeter ditch (Sondazh 4) and its Southern part (Sondazh 10) was accomplished by the Ukrainian side using ambitiously large trenches (Sondazh 4: 22 × 5m; Sondazh 10: 15 × 2m). In both trenches, the geophysical plans proved accurate guides of the location of the ditches but in neither trench were the ditches as deep as had been expected.



Excavations, Ukrainian Side 

 235

Figure 4.55: Upper: North Ditch profile from South-East; lower: Triple Ditch profiles 1–3 from East (by L. Woodard).

4.7.1.4 Sondazh 4 Trypillia sherds were recorded from the middle and upper fill of the Northern ditch, as well as from the cultural layer above the ditch, but not in the lowest fill, where daub was encountered; no animal bones were recovered from within the ditch. However, daub was also found outside the ditch in the supposedly ‘natural’ sediments. The width of the Northern ditch segment was ca. 1.5m, while there was considerable debate about the depth of the Northern ditch exposure, with different views recorded on Vince Cherubini’s section drawing (Fig. 4.55 upper). While the shallower depth

236 

 Site Studies

was believed to be 1.30m, the deepest ditch line was considered to be closer to 1.50m. Analysis of molluscs retrieved from bulk samples from the ditch fill indicated a distinctive habitat which persisted for some time – for example; an open, gradually infilling ditch, mainly dry, but holding significant pockets of moisture, with thick/ long grasses and other herbaceous plants, perhaps sparse trees, but in a landscape dominated by short grassland. Thus the debate over whether this shallow ditch contained a palisade has not entirely been settled, although there were no post-holes visible to document this kind of feature. 4.7.1.5 Sondazh 10 This sondazh was laid out across an area in which three parallel ditch sections were indicated by the geophysical plot (https://doi.org/10.5284/1047599 Section 5.6.3 SONDAZH_10_Plan). Each ditch was recognizable but their depths were less than the shallowest interpretation of the Northern ditch segment, in no case exceeding 1m in depth (Fig. 4.55 lower). One Trypillia sherd was found in the middle fill of Ditch 1, with one sherd loosely associated with Ditch 3. One animal bone sample was recovered from near Ditch 3 for AMS dating but proved to have insufficient collagen. The interim conclusion is that the shallowness of the ditch segments in the Northern and Southern areas was not commensurate with a defensive ditch but, rather, a marker of an enclosed space.

Mykhailo Videiko & John Chapman 4.7.2 House A960

In the first season of geophysical investigations (2009), parts of the Outer and Inner House Circuits defining the Nebelivka plan were revealed, enabling the choice of a dwelling house (A9) for complete excavation during that trial season. The contours of the excavated house rubble coincided closely with the archaeo-magnetic plot, which showed a narrower Southern part and a wider Northern part of the dwelling. In some cases, a geophysical anomaly was registered even where parts of the house were totally destroyed by ploughing. The total excavated area was 236m², with a recording grid set to 2 × 2m (Fig. 4.56). The remains of the building consisted of burnt daub found at depths of 0.25–0.4m. The investigated area was on a slope and the difference in height between the ends of the burnt daub was up to 1.3m (Fig. 4.57). The burnt daub scatter had a rectangular shape and was nearly 18m in length and 4.5–5.6m in width. This area consisted of two daub scatters of different dimensions:

60  See Chapman et al. 2015; https://doi.org/10.5284/1047599 Section 5.2.1.



Figure 4.56: Plan of House A9, Nebelivka (by L. Woodard).

Excavations, Ukrainian Side 

 237

238 

 Site Studies

Figure 4.57: Upper: sections of House A9: B–B1 – South-facing; C–C1 – North-facing; A–A1 – Eastfacing (by L. Woodard).

a Northern, trapezoidal daub scatter measuring 10.5m (North-South) in length and 4.1m (Northern part) 6 or 7m (Southern part) in width; and a Southern, rectangular daub scatter, measuring 5–5.5m (North-South) in length and 4.3m in width. The difference in widths of the two rooms is considered to reflect the additional wall tumble falling outwards (i.e., to the East) from the Northern room (the outer limit of wall tumble is marked by a dashed line on the plan in Fig. 4.56). In the SouthWest part of the scatter, an area of daub was found that remained outside the area of compact rubble. The edges of the excavated area lay 1.5–3.5m from the edge of the compact daub, allowing the investigation of the culture layer surrounding the building. For example, two sherd and bone scatters located opposite the short side of the building, 1–1.5m from its end, marked the position of two pits (Pits 1 and 2) – 1.2– 1.5m in diameter and up to 0.2m deep. A detailed discussion of pottery distributions in House A9 is presented below (p. 319 & Figs. 5.22–5.25). Burnt daub was identified over the entire area of the house. The building remains consisted mainly of pieces of burnt daub mixed with the remains of threshed cereals/ grasses. The surface and the cross-section of the daub contain visible traces of stubble, ears and grains of cereals and/or grasses. The smooth part of the daub faced up; some pieces have traces of smoothing by hand. On most pieces, the lower part of daub had the imprints of a wooden post, including longitudinally-cut beams of timber 15–25cm in diameter. In one area, matching pieces of daub with wood impressions



Excavations, Ukrainian Side 

 239

were traced over 2m in length, suggesting fallen walls of the kind observed in the excavation of the experimental burnt house. Fragments with impressions of rods 2–2.5cm in diameter were found in some places outside the main rubble area. It is possible that these fragments are the remains of walls that had fallen outside the house (cf. the experimental results). In addition, impressions of three vertical posts 8 to 12cm in diameter, probably supporting some sort of a structure, were identified in the Northern room. In general, there were no traces of postholes, which suggests that usually the posts were inserted into sleeper beams. Each room had its own platform constructed on the earthen floor. The Northern room had a larger platform, covering 2m × 2m, while the Southern room had an eroded platform measuring ca. 1m × 1m. Both platforms consisted of three layers of clay mixed with sand (each 20–25cm thick), the lowest of which had been laid directly on the ground surface. The next layers were laid during later repairs. The remains of a large vessel and a flask, as well as some stones, including one grindstone, were found next to the platform of the Northern room, while vessels and two small grinding stones were found near the Southern room’s Platform. The construction of House A9 is typical for Trypillia settlements in this region. It was destroyed by fire as is evident from the burnt daub, some of which was vitrified. The complete excavation of House A9 as a single unit allowed a broader perspective on dwelling-house taphonomy, with several features replicating both the test pit observations and also the findings of the excavation of the experimental burnt house. The diffuse scatter of destruction daub probably represents the falling of wall daub out from a core house area of two rooms, one no larger than 10m × 4m and the other no larger than 5m × 4m, rather than a house with rooms of differing widths. The presence of in situ platform daub in both rooms suggests that this was a one-storey house. The fallen wall panel marked by parallel withy impressions is well matched in the experimental burnt house excavation (Fig. 4.27 lower); its location suggests a section of fronton or wall that fell inside the house. The discovery of figurines and binocular vessels mixed with the destruction daub suggests they fell with the walls and ceiling from shelving or wall-pegs during the burning of the house.

Mykhailo Videiko & Natalia Burdo 4.7.3 Houses B17 and B18 and Their Pits61

The 2009 geophysical plot revealed part of the inner and outer house circuits which lay close to the Mega-structure excavated in 2012 (Quarter A). The Ukrainian side chose to excavate two adjoining houses (B17 and B18) which lay closest to the Mega-structure, as well as investigating the supposedly associated house pits for each house. One complete burnt house was excavated (House B17), as well as part of a second house

61  See Burdo & Videiko (2016); https://doi.org/10.5284/1047599 Section 5.2.2.

240 

 Site Studies

(House B18). The shapes of the houses corresponded to the shape of the anomalies on the geophysical plot. House B17 was constructed on a natural Chernozem A horizon. Its dimensions were 24m in length by 8 to 12m in width. Most of the area of the house was covered in destruction daub to a depth of over 0.30m (viz., the classic ‘ploshchadka’: Fig. 4.26 upper), with few organic remains other than charcoal. Micromorphological analysis showed the absence of cultural inclusions other than daub reinforces field observations of houses that the structures are relatively impoverished in biocultural materials (Chapman et al. 2014). The lack of microscopic cultural inclusions (such as straw or charcoal) other than daub suggests that the building was kept clean during its use-life. Degrees of heat alteration are indicated by the range of yellow to red colours and optically inactive, undifferentiated b-fabrics, probably resulting from the firing of clay in temperatures exceeding 800–850°C (Macphail & Goldberg 2010; Quinn 2013). However, no dewdrop-shaped quartz grains could be observed at the available magnifications, suggesting that temperatures did not greatly exceed these readings (cf. Courty et al. 1989). Interior details from House B-17 show a possible threshold in the West side. One example was noted of two sections of wall overlying each other – perhaps a sign of a two-storey building but also possibly one wall falling on top of another. Wall plaster showed occasional signs of incised and red-painted decoration; in particular, near the edge of the house, plaster fragments were found to have painted decoration. Under the daub were three Platforms decorated with incised ornament, comparable to that on the ‘Platforms’ in the Mega-structure. Two hearths were found near the Platforms. Up to eight sherd scatters in House B-17 corresponded to groups of once-complete pots, comprising smaller and medium-sized vessels but no storage-jars. It was observed that chernozem had accumulated on top of the burnt remains, indicating continuous soil formation under relatively stable conditions after the settlement was abandoned. Whether and how Trypillian or later pre-modern land management contributed to this soil formation remains for future study. The pit to the North-West of House B17 proved to be much bigger than the geophysical anomaly, amounting to 8m in diameter and 3.5m at its deepest point, near the Northern edge. The Southern edge of the pit came within 1m of the Northern edge of House B17 but this shallow area increased in depth as one moved North. The upper fill was a chernozem with a large number of small sherds and animal bones (Fig. 4.26 lower). There were many placed deposits in the pit, which was extremely rich in material remains. At 1.2–1.3m depth, a sloping layer 10–25cm-thick contained many large sherds and animal bones, including a Bos horn core. This layer sloped into the centre of the pit at 1.6–1.8m depth and contained 14 finds concentrations, some with anthropomorphic female figurines (a total of over 20 was found in the pit as a whole). Under this layer was a burnt daub layer 5–10cm thick, which overlaid a 2–3cm-thick charcoal layer; 10–20cm deeper was a second charcoal layer separated by a yellow



Excavations, Ukrainian Side 

 241

loessic sediment from the upper charcoal layer. Many sherds and animal bones were recovered from all of these layers. A special style of pottery consisted of many sherds with incised decoration, starting at 1.2–1.3m depth; interestingly, this type of incised decoration was rare in House B-17. The pit to the North of House B18 held fewer finds but was even larger than the B17 pit, although only an 8m-long section was excavated to a depth of 2.5 m. The upper fill, to 1.2m depth, was a chernozem; under this layer, the cultural layer shared the same properties as in the B-17 Pit. It is clear that the initial use of both pits was to extract clay for house construction. Both pits were filled in at the time of dwelling. The pits were still visible as negative features at the time of the end of the settlement: the upper fill consisted of a lower layer of soil mixed with cultural material, and an upper layer formed by chernozem. The linear pits were dug on three sides of House B17 but were much more shallow than the B17 or B18 pits. The primary use of all of these pits could have been to produce soil for mixing with clay in house construction, as in the LBK system of digging pits close to houses (Bickle 2013, pp. 167–9). These pits were much more shallow than the pits at the short end of the houses – usually no more than 30cm depth. The further the linear pit was laid out from the house, the fewer the finds that were discovered. Unlike the larger pits, by the time of the house-burning the linear pits had completely filled up to the general surface level.

Mykhailo Videiko & Natalia Burdo

4.7.4 The ‘Industrial Structure’ and Its Pit62 One of the most interesting developments in Trypillian megasite archaeology in recent years has been the discovery of high-intensity, concentrated geomagnetic anomalies which, upon excavation, turned out to be pottery kilns (Korvin-Piotrovskiy et al. 2016). The discovery of such geomagnetic features at Taljanki in 2013 prompted the question as to whether there were similar kilns at Nebelivka. Duncan Hale is of the view that, using criteria of size, orientation of anomaly, strength of anomaly and location, none of the geomagnetic anomalies at Nebelivka is entirely consistent with what would be expected of a kiln anomaly (see above, Chapter 4.2). Nonetheless, the Ukrainian side tested three features with strong, concentrated magnetic anomalies in the North-East part of the megasite. In the first Sondazh (Sondazh 7), modern iron-working was found in the upper levels, mixed with Trypillia pottery. In Sondazh 8, a pit-like feature containing a high concentration of Trypillia pottery was also a modern feature, with metal finds at a depth of 0.50 m. A historical map of the Nebelivka area shown to the Project indicates how a village street

62  See Burdo & Videiko (2016); https://doi.org/10.5284/1047599 Section 5.5.

242 

 Site Studies

extended Northwards from the main village focus, crossing the Eastern edge of the megasite. It is believed that the modern pits were associated with this street. Although the prehistoric finds were ‘contaminated’ with recent material, this experience is noteworthy in indicating that we cannot automatically make the assumption that all of the features identified on the megasite geophysical plot are datable to the Trypillia occupation. The most successful Ukrainian trench was Sondazh 9, in which an enigmatic daub feature was found close to a pit. This Trypillia feature was a 2m × 2m fired clay platform with four walls forming three channels (0.20–0.25m in width, 0.25– 0.30m in height) (Burdo & Videiko 2016, Figs. 4–6) (here: Fig. 4.58). The walls were protected by fire-proof plaster and the whole complex was covered by destruction daub fired to a green colour mixed with soil, pottery and bones. The green colour indicates the presence of ferrous oxide (Fe2O) in the clay (see below, Section 4.9), which had been fired to a high temperature. Near the feature were found fragments of fire-hardened clay which have been interpreted as the mobile covers for the channels. Currently unexplained daub lines projected to the NW and SW from the main feature.

Figure 4.58: General view of Industrial feature, Nebelivka (by M. Videiko).



Excavations, Ukrainian Side 

 243

The daub feature was sub-rectangular, with some granite lumps built into the outside edges of the corners. Three channels some 25–30cm in depth and width were cut into this feature. Many of the daub fragments had a greenish tinge, marking the greatest concentration of greenish daub on the whole megasite (e.g., one or two greenish daub fragments have been encountered on test pits). In the postabandonment phase of this feature, many small sherds were found on the upper surface of the feature – perhaps the kind of placed deposit found sometimes on the ploshchadka of the Mega-structure. There has been a difference in the interpretation of this feature, with the Ukrainians proposing that this was indeed a pottery kiln (Burdo & Videiko 2016) and the Durham team suggesting that there was no signs of a separation of fuel from pots and that this feature was more likely to be the remains of a communal food preparation area. By analogy with excavated features at Taljanki and Majdanetske (cf. Korvin-Piotrovskiy et al. 2016, Figs. 17 and 32), the three channels have been interpreted as a three-chambered firebox or combustion chamber under a firing chamber with a dome. Burdo & Videiko (2016) interpret this feature as the poorlypreserved remains of an updraught kiln. Several metres to the North of the ‘kiln’ was a 5m × 4m pit whose total depth was not established. The pit was cut from a depth 0.40–0.60m below the current ground surface. Mixed into all of the four layers, which sometimes contained charcoal lenses, were sherds, figurines, animal bones, stone, a few flints and fragments of ‘channel cover’. The strongest argument for the interpretation of the Nebelivka feature as a kiln comes from the Majdanetske and Taljanki analogies. However, the Nebelivka channels were far shallower than the other examples, showing that there was little space for fuel for a kiln. Moreover, there was a striking absence of vitrified daub or secondarily burnt pottery at Nebelivka, as was found at the kilns of the other megasites. On this basis, there remains the possibility that the ‘industrial’ feature at Nebelivka was not so much a pottery kiln as a communal cooking place for the provisioning of the periodic feasts whose animal bone remains were often discarded in adjacent pits (e.g., the Pits near House B17 and in Sondazh 1). The channel covers would have functioned just as well in a communal oven as in a kiln and the former interpretation would explain the absence of vitrification and pottery ‘wasters’. The closest analogy found so far comes from Hungarian Mediaeval brick ovens with their triple flues (Jakab 2011) (Fig. 4.59). However, these ovens have two essential elements both missing at Nebelivka – a floor separating fuel from bricks and a domed superstructure. A second structural parallel for the Nebelivka feature is the integrated set of three corn-drying ovens excavated at the Keston Romano-British villa in Kent, UK (Philp et al. 1991, pp. 87–88, Figs. 21 & 22 & Plates XI–XIV) (here Fig. 4.60). The best-preserved – the South Oven – was dug into the chalk and had three parallel-sided channels, with a stoke-pit outside the oven providing hot air to the central channel from which it passed to the side channels by means of lateral vents. The surface area of the platform above the ovens of

244 

 Site Studies

Figure 4.59: Hungarian Mediaeval brick kilns as analogies for Nebelivka ‘industrial’ feature; (1) Békéscsaba–Mezőmegyer; (2) Debrecen–Józsa Pláza (by B. Gaydarska, based upon Jakab 2011).



Excavations, Ukrainian Side 

Figure 4.60: Corn-drying ovens, Keston Roman villa (internal width of South Oven – 4.15m) (by B. Gaydarska, adapted from Philp et al. 1991, Fig. 22).

 245

246 

 Site Studies

ca. 20m2 enabled the ‘processing of fairly large quantities of grain’ (Philp et al. 1991, p. 88). A structural feature essential to the process – in this case, a covered platform separating fuel from grain – was missing from the corn-drying ovens, as it was from the Nebelivka example. We are still some way from a proper understanding of one of the most enigmatic features found at Nebelivka.

Andrew Millard 4.8 The AMS Dates 4.8.1 Aims This part of the project at Nebelivka aimed: 1. to develop an internal chronological sequence for the Trypillia BII megasite using a programme of radiocarbon dating and Bayesian modelling; and 2. to use the chronology to estimate the number of coevally occupied houses at any one stage of the megasite’s occupation. Specifically, the four key questions were: 1. how long was the occupation of an individual segment of a circuit? 2. were adjacent houses and segments constructed, occupied, and destroyed sequentially or coevally? 3. how many houses, and segments/groups were constructed, occupied, and destroyed coevally across the whole site? and 4. how do the houses inside and outside the circuit relate chronologically to the circuits?

4.8.2 Initial Dating Twenty-five radiocarbon dates were obtained from the Kyiv (conventional) and Poznań (AMS) Laboratories on material collected from the excavation of Nebelivka House A9 in 2009 (17 dates), and other burnt houses (8 dates). The use of conventional radiocarbon techniques on daub and pottery containing small quantities of organic material, and on bone led to very poor results. Ten bone dates ranged from 4130±60 BP to 4710±80 BP, seven daub dates from 2740±60 BP to 5970±70 BP and three pottery dates from 3720±180 BP to 4430±180 BP. Only three AMS dates  – on bone (5010±40 BP), cereal (5030±40 BP) and daub (5180±60 BP) respectively – fell within the expected date range of 5300 to 4900 BP, while AMS dates on pottery were much younger (3310±35 BP and 4040±35 BP). Even if the expected range for Trypillia BII is



The AMS Dates 

 247

incorrect, the occupation of the site falls entirely within a single stage in the ceramic relative chronological scheme (Ryzhov 2012) and must have been relatively short. These results show the importance of using AMS dating of carefully selected material for accurate and consistent results.

4.8.3 Simulations Prior to sampling, simulations were undertaken to establish the number of dates likely to be required to resolve the chronological questions. The approach is to consider the widest possible range of chronologies, simulating possible combinations of radiocarbon dates and then calculating the parameters of the chronology, such as start and end dates, using a Bayesian model of the same type as was intended for the actual dataset. This allows an assessment of the efficacy of the proposed programme of dating before any expenditure on dates. The efficacy of the sampling process and the number of dates to be obtained can be optimised to resolve the chronological questions to hand (Bayliss 2013). The range of possibilities was based on prior radiocarbon dating, previously published geophysical plans of other sites (e.g., Glybochok: Videiko 2007), and expert judgements about the nature of the site. Four zones were expected: the inner house circuit, the outer house circuit, the area internal to the inner circuit and the area external to the outer circuit. Three possible sequences of occupation of zones were simulated: (1) consecutive occupation of the four zones, (2) parallel occupation of all four zones for the entire duration, (3) a mixture of parallel and consecutive occupation. For each sequence, three possible chronologies were simulated: minimal (A: 160 years, 3980–3820 BC), medium (B: 300 years, 4050–3750 BC) and maximal (C: 600 years, 4250–3650 BC). For each of these nine sequence and duration scenarios, a dating effort of (i) 52, (ii) 100 or (iii) 200 dates distributed equally between the zones was considered. These 27 scenarios, with multiple possible orderings within sequence scenarios (1) and (3), capture the range of chronological and sequence possibilities. For each scenario, simulated dates were analysed in an OxCal model with independent bounded phases for each zone, to yield posterior estimates of zone start and end dates, the order of the zone start dates and the overall duration of occupation at the site. The results showed that 52 dates would be inadequate; they would neither resolve the order of zones with a 160-year chronology nor distinguish between 160- and 300year chronologies. The length of the 95.4% highest posterior density (hpd) region for the duration was used as an index of the precision of the calculations. This averaged 212 years (range 138–329) for 52 dates, 155 (115–213) for 100 dates and 112 (81–168) for 200 dates. These results show that at least 100 dates were needed to be able to estimate the duration within ±100 years at 95% probability, though, as will be seen below, issues with sampling led to the modification of this strategy.

248 

 Site Studies

4.8.4 Sample Collection In order to obtain reliable dates, dating of short-lived single-entity samples (Ashmore 1999) is widely recognised as necessary. From previous excavations, it was known that burnt house floors on Trypillia sites lie at less than 1m depth and do not overlap (Videiko 2004). The daub used in construction contained significant quantities of seeds (Pashkevich 2005) which therefore have an unambiguous association with the construction. Within destruction layers, animal bones offer greater certainty of association than charcoal or charred grain, as they are less likely to be reworked by bioturbation. The initial strategy was, therefore, to target burnt seeds for dating the construction of houses, and bones for their destruction. Using the geophysical survey (see above, Chapter 4.2) to locate burnt structures, mechanical coring to about 1m would be used to recover burnt daub samples from which charred seeds would be extracted for dating, and test pits on a subset of houses would verify the coring stratigraphy and recover samples from destruction deposits. Although coring during the 2012 season was successful in recovering 130 samples of burnt daub from 91 houses, the almost total absence of charred material within the daub meant that an alternative strategy was required (see above, p. 50). The new strategy adopted was to excavate test pits in order to recover stratified samples of animal bone or charred plant macrofossils. To tackle the questions about association between chronology and spatial structure, sampling followed the identified components of the arrangement of houses based on the geophysical survey results (Figs. 4.3–4.4). This comprised: – Two transects from outside the outer circuit to the innermost part of a radial street including an Assembly House (Test Pit groups 26, 29–30, and in part 27–28). – Rows of three to six consecutive or near consecutive houses along the circuits distributed around the whole site (Groups 1, 23, 22, 24, 25 and 13; Group 15 with House A9) – Groups of four houses, two on either side of a radial street or the two circuits (Groups 21, 20, 19, 18 and 16) – A group of houses arranged in a square rather than a street (Groups 32–34). In total, test pits were excavated on 82 houses over the 2013–2014 seasons (Fig. 4.61), recovering over 500 bones and six charcoal samples from houses63. In addition, 45 animal bone samples were recovered from excavations in pits and structures. To aid selection of samples for dating, all the bone samples were subject to zooarchaeological identification (see below, Chapter 5.3) and then screened for collagen preservation.

63 We are grateful to Charlotte O'Brien for the species identification of the charcoal samples.



The AMS Dates 

Figure 4.61: Site plan showing Test Pits with(out) AMS dates, Nebelivka (by M. Nebbia).

 249

250 

 Site Studies

Full excavation of two houses (see above, Chapter 4.7.2 & 4.7.3) and the test pits (Chapter 4.6.1) demonstrated that there was a consistent stratigraphic sequence for each house, consisting of pre-construction, house floor, destruction deposits, postdestruction deposits and topsoil. Samples were retrieved from all these levels in the test pits, yielding respectively 20, 23, 30, none and one samples. In total, there were 80 dates on bone samples and six on charcoal samples, measured at Oxford (80) and Poznań (6). The routine methods of both laboratories follow the same pretreatment procedures (Brock et al. 2016). This was reduced from the initial quota of 100 dates as it became clear from early results that the timespan of the site was short and that it lay on a wiggle in the calibration curve. In an attempt to resolve this, in the later batches fewer samples were measured but with longer counting times to give higher precision.

4.8.5 Radiocarbon Results Of the 86 measurements, six were duplicated on the same sample at Oxford as a routine quality control measure and four were replicated between Oxford and Poznań. These duplicate dates were combined by the method of Ward and Wilson (1978). Three samples were rejected as unreliable: one Poznań bone yielded C:N ratios outside the acceptable range of 2.9–3.6 (DeNiro 1985) and failed a Ward and Wilson test for combination with its Oxford duplicate, one bone dated to AD 1962 or 1975, and one bone with a low collagen yield gave a date in the Bronze Age. All dates are reported in the site archive (see https://doi.org/10.5284/1047599 Section 4.9). Thus, there were 83 measurements on 74 samples for statistical analysis. The 83 measurements were all very similar and a subset of 71 were statistically indistinguishable using the Ward and Wilson test. All results reported here follow the conventions given by Millard (2014).

4.8.6 Results and Discussion of Modelling The homogeneity of the dates and the fact that they fell on a wiggle in the calibration curve (Fig. 4.62/1) posed problems for statistical modelling, as did a lack of strong prior information. No two excavated structures had stratigraphic relationships between them and samples were recovered with stratigraphic ordering from only eight test pits. Consequently, the modelling strategy was based on testing and comparing models based on different hypotheses derived from the spatial structure of the site and on equating stratigraphic units between test pits. All modelling was performed in OxCal (Bronk Ramsey 2009).



The AMS Dates 

 251

Figure 4.62: Bayesian plots of AMS dates, Nebelivka: (1) all AMS dates plotted on calibration curve; (2) start and end dates for occupation at Nebelivka; (3) duration of occupation at Nebelivka (by A. Millard).

252 

 Site Studies

4.8.6.1 Circuits and Streets The large-scale spatial structure of the site is into the inner and outer circuits and the radial streets. Dates were assigned to these three groups and all six possible orderings of these three groups were modelled. The range of OxCal agreement indexes, Amodel, was 6% to 44%; since acceptable Amodel values are over 60%, none of these models gives an acceptable fit to the data. A model where these three groups are allowed to overlap and the ordering of their start and end dates is investigated is more revealing. Figure 4.63/1 shows that the start dates cannot be distinguished, but the end date for the outer circuit is later than the other two groups. The duration of the inner circuit comes at 0–145 years at 95% and 0–50, 70–120 years at 68%, and none of the dates can be ordered. So although the houses in the inner circuit could have been built over a century, they could also all have been built simultaneously. As noted above, in each test pit a similar stratigraphic sequence was observed. However, only one reliable sample was recovered and dated from a pre-construction context and none from a topsoil context, so these contexts were not modelled. A model grouping dates as house floor, destruction deposits, and post-destruction deposits yielded Amodel of 20%. Re-running the model with outlier analysis did not identify any samples with posterior outlier probabilities exceeding 50%. A model treating the destruction deposits and post-destruction deposits as a single stratigraphic unit produced similar results, though with some possible outliers identified.

4.8.6.2 Ordering Within and Between Quarters Analysis of the layout of the site has identified a series of Quarters labelled A–N (see above, Chapter 4.2.2 and Fig. 4.5). Grouping dates by Quarter and treating each Quarter as a separate phase with a start and end date, we investigated whether the Quarters can be ordered in time, and whether there is any discernible chronological structure within Quarters. Ideally such a model would also have an overall start and end date for the site but, with this structure, OxCal failed to initialise the model. Quarters F and K were omitted as they had no dates. The posterior distributions for the start and end dates of the Quarters (Fig. 4.63/2) only yield probabilities of ordering greater than 90% for (a) Quarter E starting before Quarters A, B, C, G, H, I, L and M, (b) Quarter E ending after Quarter B, and (c) Quarter G ending after Quarters A and B. Quarter E therefore seems to predate most of the rest of the site and to continue after some parts of it. Within each Quarter, the ordering of dates was also tested. Probabilities greater than 90% were obtained only in two Quarters. In Quarter E, Test Pit 20/1 is earlier than Test Pit 20/3 and 35/1. In Quarter G, the sample from Test Pit 25/3 post-dates all other samples excepting that from 25/1 which itself predates most other samples, and has probabilities of 87–89% for the others (Test Pit 24/1, Test Pit 24/4, and Test Pit 25/4 OxA 31665). These sequences suggest variations on the order of several decades in the construction dates for houses in the same Quarter – perhaps an indication of the temporal duration of Quarters across multiple human generations.



The AMS Dates 

 253

Quarter E therefore stands out as the earliest Quarter, and Quarters E and G as likely to be the latest Quarters to be occupied. It is probably their longevity which allows the detection of ordering of events within them. This sequence suggests that not all of the Quarters were built at the same time but that there were multi-decadal differences in not only their starting date but also their duration. 4.8.6.3 Ordering Within Rows At the more local geographical scale within the site, there are clearly Neighbourhoods, some of which are separated from each other by kinks or gaps. Models were constructed considering both (a) possible orderings along a row and (b) differences in date between houses grouped either side of a kink or gap. Continuous rows were explored for Test Pit Groups 1, 24 with 22, and 12, and rows with breaks for Test Pit Group 13 with houses B17 and B18, and groups 22, 23, and 31. In all but one case, no ordering was distinctively better than another and the duration of any break had its highest probability at zero years. The exception concerned the three dates in Test Pit Group 12, which suggest a clockwise order around the circuit. Given the number of models tested, this is likely to be a chance finding. The absence of a sequence in any of these Neighbourhoods suggests that the houses were built as an integrated group within a short period of time. 4.8.6.4 Radial Structure Some of the test pits were deliberately located to investigate any radial structure to the dating of houses. Using similar models to the rows of houses, this possible structure was investigated for Test Pit Group 26, Groups 29, 30 and 31, and for 15, 13 and 16. Only the first group produced any possible evidence for structure, primarily because the date from test pits 26/2 is earlier than several others in the group, and therefore an inner to outer ordering is consistent with the dates. This counterfactual finding is currently hard to explain. 4.8.6.5 Overall Occupation The timing and overall duration of the entire occupation at Nebelivka is a key question. A series of models was used to investigate this and to check the sensitivity of the answer to possible variations in the approach used. Models were constructed with an overall start and end for the site, using the eight stratigraphic sequences of dates and omitting them, and with and without outlier analysis. Figures 4.62/2 and 4.62/3 show the results for the least precise model, without the stratigraphic sequences and with outlier analysis. This is the most conservative estimate, though the other models all produced similar results. The site is estimated, with 95% probabilities, to have started in the range 3990–3840 BC and to have ended 3850–3740 BC, having lasted 50–230 years; all models agree the most likely dates are ca. 3970 to ca. 3770 BC with a duration of ca. 200 years.

254 

 Site Studies

Figure 4.63: Bayesian plots of AMS dates, Nebelivka: (1) start and end dates for circuits and streets modelled independently; (2) start and end dates for Quarters; (3) AMS dates from the Pit, Sondazh 1; (4) start and end dates for three megasites (by A. Millard).



The AMS Dates 

 255

4.8.6.6 Comparison with Other Sites In addition to Nebelivka, two other similar sites have significant numbers of radiocarbon dates, Majdanetske (n = 86: Müller et al. 2017, p. 75; Dal Corso et al. 2019, Fig. 2) and Taljanki (n = 11; Rassamakin & Menotti 2011)64. Modelling the overall sequence of these dates and comparing the start and end allows investigation of whether these three large sites were occupied simultaneously or consecutively, which is an important question when considering regional population sizes. Based on ceramic typology, Nebelivka is assigned to Phase BII while the others are assigned to the succeeding Phase CI, with Taljanki preceding Majdanetske (Müller et al. 2016). Despite the pottery typology, Diachenko & Menotti (2012) proposed that the occupations of Nebelivka and Taljanki overlapped by 100 years. The results (Nebbia et al. 2018) show that occupation at Majdanetske started before the end of occupation at Nebelivka (Fig. 4.63/4).65 Occupation at Taljanki probably started before the end of occupation at Nebelivka, and certainly before the end of occupation at Majdanetske. The uncertainties in the radiocarbon dating do not allow us to determine which of the three sites was occupied first. This implies that two out of the three largest megasites were partially coeval, and all three may briefly have been coeval. Testing a more constrained model in which the typological ordering is used as a constraint showed that this is not consistent with the radiocarbon dates.

4.8.6.7 Coeval Houses The estimate of overall duration can be used to investigate the likely number of coeval structures. Figure 4.62/3 shows two peaks of probability for the duration of around 100 and 200 years, which provide convenient round values to explore the likely range of possibilities. A very simple model posits a constant rate of house-building, with each house having a defined average use-life, followed by cessation of building. This leads to a linear growth of the number of houses with time, a period with the maximum number of coeval houses followed by a gradual decline after building ceases, and a total of 1,500 houses built and destroyed. If the use-life of a house was 25 years, then a 100-year duration would require 20 houses to be built per annum with a maximum of 500 coeval houses, but a 200-year duration would reduce those figures to a rate of 8.6 per annum and a maximum of 210 coeval structures. If houses were used for an average 50 years, then these figures increase to 30 per annum and 1,500 coeval for a 100-year duration, and 10 per annum and 500 coeval for a 200-year duration. A short

64  It should be noted that the set of AMS dates from Taljanki is much smaller than those from Nebelivka and Majdanetske, with several of the dates based upon long-lived species and some dates based upon samples from insecure contexts. 65 Bayesian modelling of the overall dwelling period at Majdanetske, at 3935–3640BC (Dal Corso et al. 2019, p. 3 & Fig. 2), confirms the results of this modelling.

256 

 Site Studies

use for a house of only 10 years reduces the figures to 17 per annum and 170 coeval or 7.9 per annum and 79 coeval structures. Although the model is simple and the approach is broad-brush, it can nonetheless indicate the scale of occupation. As it is most likely that the duration of the site was 200 years and the lifetime of a house constructed of timber is likely to be a few decades66, we should probably envisage between one-sixth and one-third of the houses at Nebelivka being occupied simultaneously. Simultaneous occupation of the whole site could only have occurred if the houses were long-lived and the building rate was very high.

4.8.7 Conclusions The study of the chronology of Nebelivka is hampered by the short occupation of the site, especially as the relevant period falls on a wiggle in the calibration curve that limits the precision that can be obtained and the possibility of ordering events in time, whatever Bayesian models are used. The preservation of charred material was poor and thus investigation of the duration of occupation of individual houses was not possible. Nevertheless, some robust conclusions may be drawn. It is clear that Nebelivka was a relatively short-lived site occupied for between 50 and 230 years between 3990–3840 BC and 3850–3740 BC (all at 95.4% probability). Most likely this was ca. 3970 to ca. 3770 BC with a duration of ca. 200 years. Quarter E started earlier and ended later than other parts of the site. It is unlikely that all the houses were occupied simultaneously but, quite probably, one-sixth to one-third of them were in coeval use.

Natalia Shevchenko & Bisserka Gaydarska 4.9 Analyses of Building Materials 4.9.1 Introduction The building of houses with timber, clay and reeds has been at the centre of Trypillia archaeology since the earliest excavations (Khvoika 1904; for summary, see Videiko 2013). However, it is surprising to note that very few results of ploshchadka research after more than a century have so far focussed on physical-chemical analyses of the building materials (Kulska & Dubitska 1940; Shevchenko & Ovchinnikov 2003;

66  The TOTL Project has investigated the duration of many Neolithic houses and their overall finding is that the duration of houses varied between 15 and 80 years (e.g., the Uivar tell: Draşovean et al. 2017, Fig. 7).



Analyses of Building Materials 

 257

Shevchenko 2015, pp. 61–68; Diachenko & Harat-Strotsen 2016), This is presumably because of poor preservation thanks to the effects of burning and ploughing. Thus, no ‘original’ template has been preserved from which to derive past practices. It can be queried whether house-models (Shatilo 2005) combined artistic license with impressions of actual houses or faithfully replicated original house forms and building techniques. The burning of a Trypillia house preserves the burnt daub mass (the ‘ploshchadka’) in plan form but has destroyed much, often leaving only shapeless daub remains. However, close attention to this building material allows the recovery of much useful technological information. The reasons for house-burning are multiple and should be separately investigated, in part through emphasis on taphonomy (e.g., the surface deposits on building materials and pottery: see Chapter 5.1). It is important to recall that the nature of the clay affects the secondary changes to these materials during house-burning. For these reasons, this analysis begins with the most important task – the technical-typological analysis of the Mega-structure. The full analysis is available online (https://doi.org/10.5284/1047599 Section 5.1.5). Here, we summarise the general implications of the results for an appreciation of the way that Trypillia builders created their own world of clay.

4.9.2 Methods and Materials A preliminary study of the building materials from House A9 (excavated in 2009) enabled the design of a broad research framework for the analysis of building samples from the Mega-structure. Laboratory methods were designed to fulfill three goals: (1) the establishment of the technical  – typological characteristics of the samples, including their physical-chemical properties, their composition, their structure, their context and their recipes; (2) the determination of the function of the different building materials; and (3) the classification of building materials based on Goal (1). A combination of macro- and micro-level methods was used, including morphological analyses, qualitative analyses (composition), quantitative analyses (ratios between different elements in clay mixture), metric analysis (fractions of content, thickness of layers of building material) and textural-structural analyses. A total of 155 samples was collected, comprising building material (n = 91 samples), pottery (n = 22 samples) and grinding stones and unworked stone (n = 15 samples), as well as materials from clay sources (n = 16 samples) and rock sources (n = 11 samples).67 A total of 89 photographs was taken of these samples68. A total of the

67  The full list of samples appears in https://doi.org/10.5284/1047599 Section 5.1.5.3. 68  For a selection of key photographs, see https://doi.org/10.5284/1047599 Section 5.1.5.2.

258 

 Site Studies

80 most characteristic samples was selected for further laboratory analysis. Visual inspection by magnifying glass, supplemented by stereo-microscopic investigation, confirmed the preliminary field classifications of morphology, mineral content and stratigraphic structure. Thin-sections were made for polarising microscopy to recover more specific information on mineral contents, structural traits, ratios of elements in the clay mixture and mineral temper. Further, micro-probe analysis was used for the clay minerals in some of the samples. Polished sections (Russian: Anschliff) were studied in reflected side light to gain further stratigraphic information.

4.9.3 Clay Materials Used in Building Five different clays were used at Nebelivka. During preliminary fieldwork in the village territory, the project team was able to identify four sources of clay, indicating that the distances required to bring clay to the megasite rarely exceeded 2km. The five clay types could be distinguished initially by colour and, later, by physical-chemical analysis. The colour of the greenish clays was due to ferrous oxide (Fe20), while red clays have a strong component of ferric oxide (Fe203). Colour variation in the clays is due to the ratio of ferrous oxide to ferric oxide. It was the iron elements that were predominant in determining firing temperatures. At temperatures of 7000C in an oxidising atmosphere, Fe20 is oxidised to Fe2O3, with consequent colour changes (e.g., the range ‘greenish-brown–pale yellow’ changes to the range ‘bright orange–red– dark reddish-brown’) without visible changes towards vitrification. These colours show the temperature range of 7000C–9000C (Fig. 4.64/6). The building materials at Nebelivka were often made by mixing different clays, sometimes forming lumpy mixtures of uneven colour and many cracks and pores, caused by uneven shrinking with drying. By contrast, high-quality mixtures led to uniform surface colours and consistent textures, with an even distribution of cracks. Three types of daub mixture were observed: (a) chaff-free clay made of a mixture of two clays (Podium, Platforms); (b) red clay or marl with chaff, turning red or yellow after firing (smoothing layers); and (c) a mixture of two clays and chaff, turning pink after firing (smoothing layer). The daubs contained natural minerals, including (a) clastic quartz-feldspar; (b) calcium carbonate, making the clay more plastic; and (c) ferrous minerals. The main artificial temper was chaff (40–60%), leading to different size of porosities in the daub as burnt-out chaff. However, gaps between minerals formed in drying and expansion of minerals through high firing and subsequent cooling also led to porosities in the daub. All of these types of porosity can be found in one daub sample. It was striking that sand-grade temper was absent; it may have been removed in clay purification.



Analyses of Building Materials 

 259

Figure 4.64: Building materials analysis: (1) finger action; (2) squared timber-impression; (3) withyimpression; (4) finger-impression; (5) vitrified daub; (6) thermal plot, daub, Context 117; (7) crosssection of daub showing clay layers, Nebelivka (by N. Shevchenko).

260 

 Site Studies

4.9.4 Daub and Daub Impressions Impressions of the following materials were found on daub samples: (1) wood; (2) wood shavings; (3) straw; (4) chaff; (5) grain; (6) steppe grasses such as Stipa; (7) woven stems (? twine); and (8) leaves. There were no obvious examples of reed impressions69. The commonest impressions on the underside of daub were regularlycut, flat wooden beams with fibres (Fig. 4.64/2). There were fewer samples with smooth impressions of thin, circular rods of diameter 7–10cm (perhaps hazel withies), probably because of the filing of timber or the process of de-barking (Fig. 4.64/3). Even more rare are impressions of branches of diameter 3–5cm. Some multi-layered impressions were found, some with finger impressions producing ‘barbotine’-like roughening, especially on examples of highly plastic clays (Fig. 4.64/1 & 4). Cereal grain, chaff and straw impressions were very common throughout the daub samples. The wide range of impressions underlines the Trypillia practice of incorporating many aspects of their everyday landscape into their Assembly Houses (for the incorporation of fragmentary objects into house daub, see below, p. 414).

4.9.5 Building the Mega-Structure Walls Samples were investigated from two parts of the Mega-structure walls – the South wall and the South-Eastern wall. One daub block from the South wall was formed by the application of the three layers of different clays, each with 60% chaff temper by volume and showing a red–white–pink cross-section (cf. Fig. 4.64/7). These clays show a firing temperature of 850–9000C. A second sample from the South wall shows the beginning of the vitrification, at 10000C, of poorly mixed polymineral clays (cf. Fig. 4.64/5). A similar practice – mixing two different clays with contrasting properties – was used to make the daub for the South-East wall, which was fired at 8000C. The likelihood of the clays mixed for wall-building deriving from different clay sources may indicate the ‘presencing’ of different parts of the landscape in the same stretch of wall.

4.9.6 Constructing the Podium and the Platforms The most significant features inside the Mega-structure were made of different clays with different tempering practices in comparison with the wall daub. No chaff was used in the application of clay to the horizontal surfaces of Platforms; these were dense clays, stronger than chaff-tempered daub. The surface smoothing layer

69  This observation does not wholly fit with the molluscan evidence presented in Section 4.1.2.



Analyses of Building Materials 

 261

was usually applied in several coats, with chaff temper in the lower part only. The uppermost layer was a light, carbonate-rich clay layer applied in a 1-mm-thick layer as a basis for painted decoration and with red ochre used as paint. The clay materials of the podium and the lower level of platform 257 were not only the same but they were fired at similar temperatures (800–8500C). However, different clay materials were used to construct the different layers of Platform 257, suggesting either different times of making, the practice of repairing the feature or its use for different functions.

4.9.7 The Production of a Storage-Jar A single sample of the thick-walled storage vessel termed a ‘pithos’ showed that the vessel had been built up using two layers of clay – a lower level of faster-drying olivritic clay marl with 50% chaff and an upper level consisting of a mixture of two clays – a polymineral clay and a marly clay with 30% chaff to reduce shrinkage/ cracking during drying. A slip was applied to smooth the outside surface, perhaps to achieve a decorative effect. The vessel was fired at 800–8500C.

4.9.8 The Destruction of the Mega-Structure: The Evidence from Daub Firing Temperatures There were two problems for those engaged in the deliberate burning of the Megastructure at the end of its active life: the extreme size of the built part of the Megastructure – at 36m x 20m the largest known structure in the Trypillia world – and the large central open area which would have hindered the spread of the fire from either end to the other side. For this reason, daub samples from a wide range of contexts were selected for daub re-firing temperature analysis. Each sample was re-fired until its colour was replicated in the kiln atmosphere, with its increase of 1000C every hour or with five 8-hour firings, each producing a temperature change of 2000C (for details, see https://doi.org/10.5284/1047599 Section 5.1.5) (Fig. 4.64/6). The results showed that the firing temperature of most daub blocks was found to be no more than 800–8500C, with three very low temperatures (200–3000C) and one vitrified sample, fired at over 10000C. All samples showed similar tendencies in colour change based upon increases in firing temperature. Very different firing temperatures were reached in different parts of the Mega-structure. These temperatures can be grouped into low (200–4000C), medium (400–9000C) and high (>9000C). Concentrations of vitrified daub (‘high’) may indicate the starting-point for the fires, given the optimal draught conditions; concentrations of unburnt daub (‘low’) may have been the most remote from the fire or were ‘protected’ under collapsed walls and/or as lowest parts of the design (e.g., the podium and the Platforms). In the case of four contexts (the platform Context 6; the podium Context 29); and two daub

262 

 Site Studies

scatters Contexts 107 and 112), contrasting daub temperatures were recorded for different samples from the same context. The higher (highest) of the temperatures is taken to represent the principal impact of the fire, while the lower (lowest) indicates areas protected from the main conflagration. ‘High’ temperatures were found in only three Contexts (the platform Context 6 and the daub scatters Contexts 107 and 112). ‘Medium’ temperatures were found in four contexts (the podium Context 29; the platform Context 257; the West wall Context 118 and the pithos (Context 149) above podium Context 29). Finally, ‘low’ temperatures were found in five contexts (the fired clay wall slots Context 159; the North wall Context 173; and daub scatters Contexts 45, 110 and 120). The spatial distribution of the daub firing temperature evidence by Phase (Fig. 4.41) shows markedly different distributions in Phase 3 Upper and Lower. In Phase 3 Upper, the principal clustering of vitrified daub was in the South-West corner, with a spread along the podium and the South wall; there were hardly any clusters of vitrified daub in the Eastern rooms. However, in Phase 3 Lower, much more vitrified daub was found in the Eastern rooms, with less in the Western wall and less still in the South-West corner. The combined picture suggests at least two starting-points for the fires that were lit to burn down the Mega-structure: a dominant cluster in the SouthWest corner and a less marked concentration in the North-East corner, both taking advantage of good draught conditions for burning. Such a conclusion makes sense in terms of the two problems for those burning the Mega-structure – its size and the central open area which resisted the spread of the fire.

Bisserka Gaydarska & John Chapman 4.10 Summary The site investigations at Nebelivka followed the multi-disciplinary scatter-gun approach now standard in many modern projects, by which a multiplicity of approaches and methods is used to understand a site through the integration of huge quantities of disparate, specialist information70. The diversity of approaches selected is not simply a function of the total number of available approaches, since some approaches were considered and rejected71. Rather, the variety of approaches was

70  Consider Alasdair Whittle and István Zalai-Gaál’s exemplary Ecsegfalva 23 project, in which 31 specialist reports show the variety of approaches to a single, small Early Neolithic Körös settlement (Whittle 2007). 71  The study of soil aDNA was considered and rejected for an understanding of the central open space at Nebelivka because there was no way in which to identify the depth of the occupation surface of such a large open space.

Summary 

 263

related to the diversity of research questions posed about Nebelivka in particular and megasites in general (for Project aims and objectives, see above, Chapter 2.2). The central problem for the Project team was the difficulty of researching such a massive site of 238ha. We were fully aware that our Ukrainian colleagues put the strategy of the excavation of burnt houses at the heart of their decades-long investigations at the two nearby megasites of Taljanki and Majdanetske (Kruts et al. 2001, 2005, 2008, 2010, 2011, 2013; Shmaglij & Videiko 2001–2). While the excavation-centred approach has produced remarkable insights into the architecture of megasite houses and their burning, as well as an unparalleled series of ‘burnthouse assemblages’, only one of the Project objectives (Objective 6, on Trypillia house architecture) could have been answered by a small-scale emulation of this approach. The Project response was to develop a targeted series of complementary investigations to answer the two overarching aims and each objective. It soon became apparent that geophysical prospection was to play a fundamental role in the Nebelivka study. The geophysical plan of the 238ha settlement at Nebelivka was the first such complete plan for any Trypillia megasite. Moreover, the advances in the new geophysical devices meant an improvement in the quality of the plot, with a greater differentiation of anomalies and hence the detection of new kinds of anomalies, as well as the potential to define new combinations of features new and old (Chapman et al. 2014b). These breakthroughs led to the setting of a new agenda for Trypillia megasite studies, which we estimate will engage archaeologists and kindred specialists for at least two decades. A vital first step in tackling the new research agenda was the ground-truthing of a wide range of new and old features through trial and larger-scale excavations. The Project excavated five types of new feature, including the largest Assembly House known in the Trypillia world (the so-called ‘Mega-structure’), an unburnt house, the largest form of pit, an industrial feature (perhaps a ‘pottery kiln’) and the perimeter ditch. The most complex operation was the excavation of the Mega-structure, in which a high proportion of the finds was recorded by Total Station and almost all daub features were drawn, scanned and integrated into Phase plans. While the primary goal of test pit excavation was the recovery of stratified organic samples for AMS dating, the large number of test pits, though small in size at 2–4% of a house area, produced a cumulatively impressive quantity of architectural information, enabling spatial comparisons of house architecture that have not been possible before. The interpretation of excavated remains relies heavily on a good understanding of taphonomic processes, which have been neglected in past investigations of Trypillia sites. The Project focussed on two types of feature – burnt structures and pits. Natalia Shevchenko’s innovative analysis of the Mega-structure building materials provided a framework for the construction of the walls and internal features as well as a daub firing temperature-based consideration of its deliberate destruction by fire. The late introduction of an experimental house-building and -burning programme led by Stuart Johnston was augmented by the excavation of the experimental burnt house

264 

 Site Studies

in 2017. The production of a genuine 'ploshchadka' for the second time, and vitrified daub for the first time, in Cucuteni-Trypillia house-burning experiments underlines the importance not only of deliberate burning but also the unsuspected quantities of firewood required to burn a house – perhaps as much as 10 times the timber required to build the house. The predominantly dry conditions of the South Ukrainian forest-steppe made it particularly difficult to understand the landscape in which the megasite was created. Bruce Albert and Konstantin Krementski were able to locate a small alluvial basin within 250m of the edge of the megasite, from which they recovered a sediment core dating to before, during and after the megasite occupation. Two fundamental conclusions emerged from this research: the discovery of a pre-megasite agricultural signal, including a massive fire event, indicating an Early Trypillia population in the area which has not been detected through fieldwalking; and the extraordinarily low level of human impact found in the sediments coeval with the megasite, emphasising the limited number of people dwelling at Nebelivka at any one time in this massive site. Both of these findings are crucial for our understanding of the growth and development of Nebelivka. The only objective which the Project failed to meet was the derivation of an internal sequencing for the 1,445 structures at Nebelivka. The pioneering use of a large number of 2m × 1m test pits placed in the geophysical anomalies interpreted as dwelling houses or Assembly Houses72 enabled the collection of over 500 AMS samples, from which over 100 were submitted for dating in the ORADS (Oxford) and Poznań labs. Despite the successful production of 83 AMS dates, a wiggle on the radiocarbon calibration curve has complicated even the most basic chronological differentiation – whether the Outer Circuit was built in different decades from the Inner Circuit and how the Inner Radial Streets related in time to the two Circuits. This stumbling-block opens up the interpretation of the Nebelivka plan to a variety of models, from the coeval occupation of all houses to the sequential occupation of a fraction of houses in any generation and several others (for an evaluation of the three developmental models, see Chapter 6.1). But whichever the preferred scenario, the model must conform to the limitations of the basic settlement footprint of 1,445 houses, including 1,077 burnt houses, to be built and burnt within seven or eight generations or 200 years, without any peaks in human impact as shown in deforestation or charcoal peaks.

72  The strategy of widely spaced test pits for the collection of AMS samples has now been adopted by other research teams in Ukraine (e.g., Müller et al. 2017; A. Diachenko, pers. comm.; A. KorvinPiotrovskiy, pers. comm).

Bisserka Gaydarska, Marco Nebbia, John Chapman, Edward Caswell, Sophia Arbeiter, Eduard Ovchinnikov, Dmytro Gaskevych, Cătălin Lazăr, Theodor Ignat, Adrian Boyce, Amanda Dolan, Jason Newton, Dmytro Kiosak, Mykola Belenko, Oliver E. Craig, Harry K. Robson, Matthew von Tersch, Alexandre Lucquin, Zsuzsanna Tóth, Alice Choyke, David Orton, James Nottingham, Giselle Rainsford-Betts, Kim Hosking, Andrew Millard & Galyna Pashkevych

5 The Finds

In this chapter, we present comparative analyses of artifacts and ecofacts from most of the Project excavations at Nebelivka (House A9, the Mega-structure, the Pit in Sondazh 1 and the test pits). The finds from the Ukrainian excavations of Houses B17 and B18, the ‘industrial feature’ and their respective pits are published elsewhere. All classes of finds were subject to the same taphonomic protocols before comparative analysis between the excavation units at Nebelivka and comparisons with other Trypillia sites and megasites. A team of pottery specialists considered alternatives to the Ryzhov pottery system, using the sherd rather than the whole vessel as the unit for 14 different comparative analyses. Dmytro Kiosak examined the small lithic assemblage, identifying a major decrease in lithic deposition after the large Early Trypillia samples. The special finds analysis considered the sample of almost 100 figurines, fired clay tokens and the only gold ornament known so far from the Trypillia group. David Orton and colleagues have written the first modern faunal report of a Trypillia assemblage, paying attention to inter-analyst variability and contextual variability. The small botanical assemblage, discussed by Galyna Pashkevych, was the result of the first water-sieving operation conducted on a Trypillia excavation and confirmed her views, counterfactual for megasites, on Trypillia arable farming as low in production and efficiency.

266 

 The Finds

Edward Caswell, Sophia Arbeiter, Eduard Ovchinnikov, Bisserka Gaydarska, Marco Nebbia & John Chapman 5.1 Pottery 5.1.1 Introduction Pottery constitutes one of the three important classes of material finds in the Trypillia ‘Big Other’. Cucuteni-Trypillia decorated pottery comprised both fine wares (painted in the Western part; incised in the East) and coarse wares (mostly incised and/ or impressed) (Tsvek 1996; Tsvek & Rassamakin 2005). The shapes and decorative motifs of painted wares have been used to classify and date Trypillia phases, subphases and regional groups in a complex, interlocking typological scheme (Ryzhov 1993, 2012, 2012a). Pottery dominated the ‘grave goods’ (Chapman 2015) deposited in house-burning ceremonies, where the painted wares could easily be imagined as a prestige good in their own right. It could then be argued that the way that several households contributed pottery or, more frequently, decorated sherds to a houseburning ceremony was a kind of potlatch, in which fragments and whole vessels were placed in the house before burning and/or placed on top of the burnt mass of daub after the fire had died down. The pottery component of the Cucuteni-Trypillia Big Other was indeed both generic and ambiguous – offering the potential for varied renderings of ceramic forms (Ryzhov 2012), while simultaneously providing the chance for varied readings of these forms (Tkachuk 2005). Part of their success was the combination of the individual and dividual identities that vessels embodied. A vessel was conjointly an individual object with specific meanings and a dividual part of a class of entities, its meaning negotiated in relation to the wider whole. Another key element of the Big Other was its reliance on ancestral values, materialized in long-term pottery traditions. Such ancestral values were nested in a communitarian manner, emphasizing the settlement over the Neighbourhood, the Neighbourhood over the household and the household over the person. Within such general parameters, we can begin to consider the pottery excavated at Nebelivka and its place in megasite pottery studies and the wider cultural context of Trypillia ceramic traditions.

5.1.1.1 Sampling and Comparative Method The Project has developed three underlying premises for our pottery studies: (1) a pottery assemblage cannot be understood without first developing a model of pottery deposition for the context in question; (2) although the form and decoration of ceramics changed through time, time was NOT the reason for these changes – there were social, functional, technological and ritual reasons for such changes, which happened in a temporal setting which was itself initially neutral to change but on

Pottery 

 267

which people could draw for their own reasons; and (3) the best way of exploring these changes is the comparative method, using different classes of deposit to highlight differences. Here, we compare three assemblages from different Excavation Units – the largest Assembly House on the megasite (the ‘Mega-structure’), a normal dwelling house (House A9) and the Sondazh 1 pit, as well as pottery from the test pits excavated into over 80 dwelling houses and Assembly Houses (for excavation reports, see sections in Chapter 4). The Project has made a serious attempt to utilize the pottery typo-chronological system commonly in use in Trypillia site reports, as created by Sergei Ryzhov (1993, 2012). An attempt at correlation of the different systems is made below (Table 5.1). Table 5.1: Correlation of the numbered shape types used in the Ryzhov/Ovchinnikov and Nebelivka systems (see Fig. 5.1) (by B. Gaydarska). TYPES USED IN RYZHOV/ OVCHINNIKOV SYSTEMS

SUB-TYPES OF OVCHINNIKOV SYSTEM

NEBELIVKA TYPES

Miski

Zrizano-konichni (1)

Everted-rim dish/plate

Napivsferichni (2)

Rounded dish

Zakriti (4)

Hole-mouth jar

Posudini Sferokonichni (5)

Amphora

Posudini Bikonichni (6)

Amphora

Gorshtiki (7)

Necked bowl

Krateri (8)

Everted-rim dish/flaring-rim dish/ necked dish

Kubki

Mali (9)

Flask/necked bowl/small amphora

Veliki (10)

One-handled amphora

Kubkopodibni (11)

Amphora or storage-jar

Amfori (12)

Necked flask/amphora

Grushopodibni (13)

Hole-mouth jar (piriform)

Pokrishki (14)

Lid

Unikalni (15)

Rare types (high-handled dish/lugged dish/footed vessel)

Miniaturni (16)

Miniature vessels

Binoklepodibni (17)

Binocular vessels

268 

 The Finds

Before further analysis, we should confront the problems of comparing the Ryzhov and Ovchinnikov systems (Ovchinnikov 2014), based upon whole vessels, with that used at Nebelivka, based upon the sherd as the unit of analysis. The weakness of the inclusive Nebelivka system is the presence of many necked forms and bases unattributable to any specific type. The solution has been to categorise rims into ‘fine-necked’, ‘medium-necked’ and ‘thick-necked’ and bases into ‘fine bases’, ‘medium bases’ and ‘thick bases’. The Ryzhov/Ovchinnikov systems share four weaknesses: an over-reliance on amphorae types; a lack of differentiation of the ‘kubka’ type; a failure to distinguish ‘dishes’ from ‘plates’ in the ‘miska’ type; and a neglect of bowls (found in two types – ‘miski’ and ‘gorshtiki’). Whatever the respective merits of the different systems, this makes it hard to compare the results of the Nebelivka typology with Ryzhov- or Ovchinnikov-based analyses from other sites. The Ukrainian pottery specialist, Dr. Eduard Ovchinnikov, worked with us in the Nebelevka project and analyzed the assemblage from House A9 and the Megastructure (Ovchinnikov 2012, 2015). His approach is an alternative to Ryzhov’s system, in which an initial division into fine painted wares, coarse wares and burnished wares formed the basis for a further sub-division into fabrics, based upon colour and temper (e.g., Ovchinnikov 2014). The next stage was the comparison of vessel shapes and decoration with wares and fabrics; the clearest shape typology is found in Ovchinnikov’s study of the Kaniv group (Ovchinnikov 2014, p. 80 & Rozdil 3 – here, Fig. 5.1). These stages fit well with the system used by the Project, based upon the Mont Beuvray system. The ‘Mont Beuvray system’ is the product of decades of pottery research, leading to a standardized and highly effective system for recording pottery at the Late Iron Age defended urban complex of Mont Beuvray, Central France (Paunier et al. 1994; Barral & Luginbühl 1995). The basis is a chronological system of shape types, each of which has been dated with reference to previous excavation contexts. The Fabric series and the decoration types are overlain on the dated vessel shapes. Since the starting-point of the Mont Beuvray system – dated shape types – was missing from Trypillia pottery research, we had to omit this stage for the Nebelivka assemblage, instead using the fabric types based upon colour as the framework for analysis. In transposing this system to the Trypillia context, three key assumptions were made: (a) the basic unit of analysis is the sherd, with each sherd – no matter how small  – having a ‘voice’; (b) the ideal recording method is the 3-dimensional recording of each sherd on a GIS platform, although this was possible only for the Mega-structure; and (c) the same level of detail is recorded for each sherd. The basic variables recorded included Weight, Pot part, Fabric, Surface Colour (exterior and interior), Temper, Decorative Style and Motif(s), Wear traces and Burning. For rim and base sherds, the rim diameter and the proportion of rim surviving is recorded and the profile was drawn. Photographs were made of a high proportion of decorated sherds and significant undecorated sherds. This rigorous data collection stage required

Pottery 

 269

much training and post-excavation time, amounting to 250 person-days. The corpus of Nebelivka profile reconstructions and decorated pottery photographs is available in the Project Archive (DOI: https://doi.org/10.5284/1047599 Section 5 by excavation unit).

Figure 5.1: Pottery types used in the Ovchinnikov ceramic system (by Ovchinnikov 2014, 80; see our Table 5.1).

270 

 The Finds

Three of the four ceramic assemblages represented a different kind of feature – an Assembly House (the Mega-structure: see Section 4.5.1), dwelling house A9 (see Section 4.6.2) and the Pit in Sondazh 1 (see Section 4.5.3). Our initial expectation was that three pottery assemblages which were created in such different depositional conditions would have shown strong contrasts in many aspects of their basic characteristics. The fourth pottery ‘assemblage’ is a aggregation of all the pottery recovered from the 80+ test pits excavated in dwelling houses and Assembly Houses (see Section 4.5.2). The advantage of the test pit sample – the widespread distribution of the test pits across the megasite – is offset by the small fraction of any house excavated – generally no more than 3%. This sample provides the ceramic equivalent of ‘background noise’ – average values for ceramic deposition across the whole megasite. Comparisons of the three other assemblages with this average ceramic deposition will demonstrate the extent of their specificities. It is also important to understand the effect of the placement within the house of a 2m × 1m test pit on the size and character of its pottery sample (see above, p. 221). In this circumstance, the nul hypothesis is that the placement of the test pit has no effect on its pottery sample. A X2 test on sherd counts vs. test pit location (house corner zones, wall zones and the middle of the house) showed a p-value of 0.525458, which is not significant at 0.01% – i.e., the Null hypothesis is confirmed. This result justifies comparison of test pit pottery samples with each other where informative.

5.1.2 Taphonomy An essential component of the pottery study concerns the taphonomy of the four excavation units. In the architectural analysis, we alluded to the basic taphonomic residues studied by Schiffer (1976) and Kuna (2015) (see above, p. 53). Here, we integrate that discussion with taphonomic information from the pottery assemblages, starting with burnt houses, continuing with unburnt houses and Assembly Houses and concluding with pits.

5.1.2.1 Burnt Houses in Test Pits We developed a standardised stratigraphic sequence of five contexts for all the test pits. While pre-house artifacts were found in some test pits, very few pre-house features were discovered (e.g., the small pit in Test Pit 15/1). The number of sherds found in Context 5 ranges from one (Test Pit 31/3) to 71 (Test Pit 24/4). In test pits with relatively high Context 5 sherd densities, a case could be made out for pre-house practices which may have influenced the building of a house at that place. The question of whether pottery can be securely placed in Context 4 living floors or Context 3 destruction daub remains difficult to answer, since, even if sherds are

Pottery 

 271

found lying horizontally on a living floor, they may have fallen from a wall shelf into such a secondary position. We have allocated most sherds mixed in with destruction daub to Context 3. Sherds lying on the top of the ploshchadka have usually been allocated to Context 3 to show the link between pottery deposition and the burning of the house. However, in a number of test pits, pottery has been found in Contexts 2 and 1, well above the top of the burnt clay mass, presumably through rodent activity (i.e., the formation of krotovini) or ploughing, despite the absence of ploughmarks in such small excavated areas. The study of post-depositional impact on sherds showed that almost all vitrified sherds were found in the test pits (e.g., Fig. 5.2/6), together with the vast majority of eroded sherds (e.g., Fig. 5.2/3 & 8), sherds with calcareous deposit73 (Fig. 5.2/4–5) and half of the sherds with secondary burning (Fig. 5.2/1–2). The variety of houses under investigation makes it difficult to draw general conclusions from such high rates of post-depositional change. The exception to this sequence of five Contexts in a normal burnt house came in the Ukrainian-led excavation of burnt house A9. The finds were recorded with the usual Ukrainian system of location within a 2 × 2m grid, with no vertical differentiation between pre-house, living floor, destruction level and post-destruction material. This finds recording system meant that the House A9 assemblage represented an ‘average’ assemblage for all phases of the house’s life cycle. This circumstance may explain some of the differences between the House A9 assemblage and the other finds assemblages. Given the extent of post-depositional effects on the test pit pottery, it is intriguing to note that not a single sherd from House A9 had been subject to postdepositional impacts such as burning, vitrification or calcareous deposition.

5.1.2.2 Assembly Houses Assembly houses were normally characterised by the presence of an in situ platform, with very few finds (in contrast to unburnt houses) and no ploshchadka. This distinction was not found, however, in the largest assembly house in Nebelivka – the fully excavated Mega-structure, which combined destruction deposits usually typical of burnt houses in burnt areas and features more likely to be have been found in the open areas of Assembly Houses (see Chapter 4.5.1).

73  The surfaces of most of the excavated pottery from Nebelivka was covered in a calcareous crust, which was regularly removed during post-excavation processing with acetic acid. Whatever calcareous crust remained on the vessel after this cleaning meant that the sherd had suffered from stronger-than-usual soil processes.

272 

 The Finds

Figure 5.2: Taphonomy: (1) burnt sherd, Test Pit 20/1; (2) burnt sherd, Test Pit 33/1; (3) sherd with heavy wear, Test Pit 21/2; (4) sherd with heavy deposition, Test Pit 22/4; (5) sherd with moderate deposit, Test Pit 26/5; (6) vitrified sherd, Test Pit 24/3; (7) vitrified sherd, Mega-structure Context 208; (8) wear on base of sherd, Test Pit 1/3 (by K. Harding).

Pottery 

 273

The study of post-depositional impacts (Fig. 5.2) showed that vitrification has been observed on only one sherd deposited in the Mega-structure (Fig. 5.2/7), although many reconstructed vessels74 showed signs of burning. By contrast, half of the sherds suffering from heavy erosion, all sherds with surface marbling effects and 12% of sherds with a calcareous deposit on their surface were deposited in the Mega-structure. The presence of heavily eroded sherds in the Megastructure only partially relates to deposition of sherds in open areas less susceptible to burning and more likely to suffer erosion; there are almost as many eroded sherds in walled areas. But they may equally support the conclusion of a temporal interval between the abandonment of the building and its burning, during which the sherds may have been left protected from the ensuing fire but exposed to erosive forces. Sherds with calcareous deposit are mainly concentrated in walled areas, suggesting local chemical processes.

5.1.2.3 Pit, Sondazh 1 The two main classes of deposit – the Episodes with finds concentrations and the intervening fill layers – were themselves grouped into five ‘Stratigraphic Units’ (henceforth ‘SU’s). One of the primary questions to be explored in the data analysis is the extent to which the SU assemblages differed from each other. The study of postdepositional impacts showed that one third of the sherds with burning were found in the Pit. The rarity of burnt layers in the Pit suggest that these sherds derived from burnt house residues but the absence of erosion on the Pit sherds indicates that the sherds had not been curated long before re-deposition (Kuna’s ‘tertiary refuse’). The occurrence of two-thirds of sherds with heavy calcareous crusting in the Pit suggests the action of specific local soil processes.

5.1.2.4 Summary These taphonomic observations give us a clearer idea than before of how the four main pottery assemblages were formed. The following report will be divided into four sections: information on pottery production; the internal study of the four Nebelivka assemblages, focussing on their similarities and differences; a comparison of the Nebelivka assemblages with other recently excavated pottery assemblages from other megasites; and a more general comparison of the Nebelivka assemblage with Cucuteni-Trypillia pottery as a whole. The preliminary results of the internal Nebelivka analyses were presented to the Kirovograd International Conference (April 2015) and published as Caswell et al. (2016).

74  There are also many burnt animal bones deposited in the Mega-structure (see Orton, D. et al., below, pp. 405–406).

274 

 The Finds

5.1.3 Pottery Production, Consumption, Refitting and Post-Depositional Evidence A total of 1,150 sherds (or 6% of the total sample) showed evidence pertaining to one or more of these processes (Table 5.2). The observations of post-depositional impacts on sherds have been included in the discussion on the taphonomy of the excavation units (see above). A Χ2 test showed that the distribution of ‘taphonomically-affected’ sherds was significantly different (p-value 80kg in House A9. The main activity of sherd deposition in the Pit in Sondazh 1 occurred in Stratigraphic Units 3 and 4 , with the lowest deposition rates in SU 1 and increasing rates until a fall in SU 5. Equally, there was a decline with time in mean sherd weight, with the exception of a slight rise in SU 2, underlining a preferential deposition of large sherds at the base of the pit. The largest pottery sample in the Mega-structure came from the Destruction phase – 10 times more by number than on the living floor and three times more by weight. However, the highest mean sherd weight derived from the living floor – three times that of the Destruction phase. This shows a preferential deposit of large sherds

76  Enchainment is the process of creating social links between people and places through material means (Chapman 2000). It is a fundamental constitutive process of making relationships in the past and present (A. Jones 2012). 77  The excluded assemblages derived from Houses B17 and B18, the pits near B17 & B18 and the industrial feature and its nearby pit.

280 

 The Finds

Table 5.3: Types of analyses of the Nebelivka pottery (by J. Chapman). No. Analysis

Excavation Units

Question(s) addressed

1

Sherd Numbers and Weights, All samples; Quarters; Pit What are the basic parameters of with Mean Sherd Weight (incl. Stratigraphic Units; Mega- the four assemblages and how do cumulative frequencies of structure Phases they change through time? weights) by all sherds, fine vs. coarse wares and decorated vs. undecorated sherds

2

Pot parts (rims, body sherds, lugs/handles and bases)

All samples; Quarters; Pit Were there variations in space and Stratigraphic Units; Mega- time in the deposition of different structure Phases vessel parts?

3

Fabrics

All samples; Quarters; Pit To what extent can we recognize Stratigraphic Units; Mega- distinctive Fabrics in time and structure Phases space?

4

Correlation between vessel type & fabric

House A9, Pit Sondazh 1 and Test Pits

5

Vessel type

All samples; Quarters; Pit Were there variations in space and Stratigraphic Units; Mega- time in the deposition of vessels structure Phases of various types?

6

General classes of vessel (open vs. closed forms)

All samples; Quarters; Pit Were there variations in space and Stratigraphic Units; Mega- time in the deposition of open and structure Phases closed classes of pottery?

7

Minimum Number of Vessel (MINV) estimates

Test Pits with sherd samples of 101 or more

8

Vessel sizes (in general and by All samples; Megavessel type) structure Phases

Were there variations in space and time in the deposition of vessels of varying sizes?

9

GIS distribution of vessel types Mega-structure by Phase and Area

Were there variations in the distribution of vessel types in time and space?

10

Variation in decorative style by Pit Sondazh 1 vessel type

Was there variation through time in the relationship between vessel types and decorative style?

11

Presence/absence and All samples; Quarters; Pit Were there variations in space location of decorative motifs of Stratigraphic Units; Mega- and time in the deposition of all decorative styles structure Phases decorated sherds with varying styles and motifs?

12

Distribution of combinations All samples; Pit Were there variations in space and of decorative motifs (1, 2, 3 or Stratigraphic Units; Mega- time in the number of motifs in 4 motifs) structure Phases decorative combinations?

13

Distribution of decorative motifs

Test Pit groups

Were there variations in space and time in the decoration of selected pottery types and their fabrics?

Is it possible to define the MINV in selected test pits?

To what extent were sherds with specific decorative motifs deposited in different houses?

Pottery 

 281

Figure 5.5: (1) number, (2) weight (kg) and (3) mean sherd weight (g) of pottery groups by Excavation Unit (by J. Chapman).

282 

 The Finds

on the living floor. A variety of sample sizes derived from different Test Pits, with larger samples from Quarters B, G and M and smaller groups from Quarters D–F and J (Fig. 5.6). This pattern was repeated at the level of the Neighbourhood, with large samples found in Neighbourhoods 13, 27, 59, 64–65, 75, 79, 104 and 124. The comparison of vessel sample size from test pits located in Zone 9 (the centre of the structure) showed similar variation at the level of the Test Pit (Fig. 5.7), while the only Neighbourhood with great internal variability was Neighbourhood 124, with samples of all sizes in the five different houses. In a comparison of fine vs. coarse wares, the former predominated in all excavation units and all phases by number and weight (Fig. 5.5). The lowest proportion of deposited coarse ware sherds came from House A9 (4.5% by number, 6% by weight), while other excavation units clustered around 10% by number, with more variation by weight. Higher mean sherd weights were recorded for coarse wares in each unit, with the highest in the Mega-structure and the test pits. Fine wares dominated all SUs in the Pit by number (always over 90%) and by weight (82–90%), with a decline over time of mean sherd weights for both fine and coarse wares, with the exception of larger fine ware sherds in SU 2. Most Quarters showed a similar pattern, except in Quarter M, where 25% of all sherds were coarse wares. The dominance of fine wares over coarse wares in all phases of the Mega-structure was total (>96% by number, >94% by weight). In a comparison of decorated vs. undecorated sherds, the latter was dominant in all excavation units, with ~30% of decorated sherds by number/~42% by weight (Fig. 5.5). Decorated sherds ranged from 23.5% in House A9 to 35% in the test pits by number, whereas the range by weight was 31% in House A9 to 55% in the test pits. Higher mean sherd weights were recorded for decorated sherds in each unit, with the highest in the Mega-structure, most of the Quarters and the test pits. The decorated – undecorated sherd ratio varied by depth in the Pit, Sondazh 1, with unusually more decorated sherds in SU 1 and almost as many in SUs 3 (by number & weight) and 4 (by weight). Larger decorated than undecorated sherds were typical for all SUs, with both showing a decline in mean sherd weight with time (except for SU 2). The living floor has the highest proportion of decorated sherds of all Mega-structure units (40% by number; 55% by weight, with the highest mean sherd weight), in contrast to the small, generally undecorated sherds placed on the destruction daub. A comparison of all these data shows that House A9 and the test pit data lie at opposite ends of the spectrum, with the Mega-structure and the Pit in intermediate positions. House A9 shows the lowest decorated sherd proportion, the lowest ratio of coarse ware sherds and a lower mean sherd weight; by contrast, the test pits show the highest proportion of decorated sherds, and coarse wares and a higher mean sherd weight. There is a strong diachronic trend in the Pit, with largest sherds at the base, decreasing with time, except for SU 2, in which a peak in sherd size and numbers shows a different depositional practice from the usual trend. The Megastructure phases show equally strong depositional contrasts, with fewer but larger,

Pottery 

Figure 5.6: Density of pottery samples by weight (g), Test Pits (by M. Nebbia).

 283

284 

 The Finds

Figure 5.7: Density of pottery samples from Test Pits placed in the centre of burnt houses (Zone 9) (by M. Nebbia).

Pottery 

 285

and more often decorated, sherds on the living floor, more sherds but less decorated in the Destruction phase and many small, mostly decorated sherds placed on the ploshchadka. These data frame an ongoing discussion of how many, and what type of, sherds were placed in different parts of the megasite and how these cumulatively strategic choices were built up of a large number of small-scale tactical depositional decisions.

5.1.4.2 Pot parts (Analysis 2) The comparison of the percentages of rims, body sherds, handles & lugs and bases in all excavation units except one shows, in every case, a clear predominance of body sherds (>80%). Rims are usually at 11–12%, while the range of bases stands at 4–7%. The sole exception – SU 2 in the Pit – showed unusual values of sherd numbers and weights: here again, the lowest body sherd representation, at 77%, is matched by the highest proportion of rim sherds – at 19% – of any excavation unit. With this one exception, these consistent results show minimal difference in the representation of body parts in each excavation unit. 5.1.4.3 Fabrics (Analysis 3) In the Mont Beuvray system of pottery analysis, pottery fabrics play an important role. Their creation stems from an integration of surface colour and types of ware (including temper). Equally, at Nebelivka, it has proved possible to combine the coding for surface colour and fine vs. coarse ware to produce a series of 18 fabrics, in which seven colours were used to make both fine and coarse wares, three colours were confined to fine wares and one colour was restricted to coarse wares (Fig. 5.8). The series of Nebelivka fabrics is as follows (Table 5.4). There is a distinct pattern of fabric preference (Fig. 5.9/1–4), with three of the excavation units showing a clear to strong preference for Fabric B (red-grey fine ware) with a second choice of Fabric A (pink fine ware). This was also found in the pottery of most Quarters, but other Fabrics proved the most popular in four Quarters – Fabrics I, K and Q. The fourth unit – House A9 – has very few Fabric B sherds, though Fabric A is the second choice at 23%. Here, the predominant Fabric is C (red fine ware), with two other Fabrics – E (light brown fine ware) and O (kaolin white fine ware) – of lesser significance but not found to be important in any of the other excavation units. The distinctive profile of Fabric preference in House A9 shows the significance of the Nebelivka fabrics in the underpinning of household identity, while preferences for fabrics different from the norm may also have contributed to the identity of people living in different Quarters.

286 

 The Finds

Figure 5.8: Fabric colours: (1) Fabric C; (2) Fabric A; (3) Fabric B; (4) Fabrics I–J; (5) Fabrics E–F; (6) Fabrics G–H; (7) Fabric D; (8) Fabrics O–P; (9) Fabrics K–L; (10) Fabrics M–N; (11) Fabrics Q–R (by K. Harding).

Pottery 

 287

Table 5.4: Nebelivka pottery fabrics (see Fig. 5.8) (by J. Chapman). Fabric

Designation

Surface Colour

A

Pink Fine Ware

2

B

Red-Grey Fine Ware with sand temper

3

C

Red Fine Ware

1

D

Grey Coarse Ware with shell temper

9

E

Light Brown Fine Ware

5

F

Light Brown Coarse Ware

5

G

Dark Grey Fine Ware

6

H

Dark Grey Coarse Ware with shell temper

6

I

Dark Brown Fine Ware

4

J

Dark Brown Coarse Ware

4

K

Dark Red Fine Ware

8

L

Dark Red Coarse Ware

8

M

Grey-Brown Fine Ware

7

N

Grey-Brown Coarse Ware

7

O

Kaolin White Fine Ware

10

P

Kaolin White Coarse Ware

10

Q

Orange Fine Ware

11

R

Orange Coarse Ware

11

There is only minor change through the Stratigraphic Units of the Pit in Sondazh 1. All SUs are dominated by Fabric B (red-grey fine ware) but without a diachronic trend. The only trend with time is in the greater importance of Fabric A (pink fine ware), with the exception of SU 2. The number of Fabrics selected at above the 10% threshold is broadly similar across the excavation units, with the highest number (n = 10) found in the sample drawn from the widest spatial range – the test pits. Only two Fabrics (A and C) are found at above the 10% threshold in all units, while four coarse ware Fabrics (L – dark Red; N – grey-brown; P – kaolin white; and R – orange) never cross this threshold. Fabric H (dark grey coarse ware) is found only in the Pit, Sondazh 1, while Fabric M (grey-brown fine ware) occurs above the threshold only in House A9. The identity of those contributing vessels or sherds to house or pit deposition is reproduced in the combinations of the Fabrics. The difference in the numbers of fabrics selected by Quarter shows a difference in various parts of the megasite, with the utilisation of most fabrics in the Eastern area, medium numbers in the Western area and the lowest fabric numbers in the Northern area. The fabric preferences of specific houses within three Quarters – B, G and M – can be compared with the overall Quarter statistic. While the four houses in Quarter

288 

 The Finds

B and the three houses in Quarter M show broadly similar fabric preferences, there was far more variability in the four houses of Quarter G, with the most popular fabric different in three of the houses (Fig. 5.9/5–8). This selection of different clay sources, perhaps through time, suggests the formation of different identities, perhaps related to different groups of visitors dwelling in different Quarters.

5.1.4.4 Fabrics vs. Form (Analysis 4) There are four forms found with sufficient frequency to make comparisons between excavation units in respect of their fabric preferences: bowls, carinated forms or amphorae, dishes and plates. In House A9, 11 fabrics were chosen to make these vessel forms (Fig. 5.10/1–4). Five fabrics dominated the four forms – Fabrics A (pink fine ware), D (grey coarse ware), G (dark grey fine ware), I (dark brown fine ware) and K (dark red fine ware). While the dishes and plates shared exactly the same fabric preferences (Fabric D > K > I), the bowls and carinated shapes showed variations between Fabrics A and D for the carinated, K and C for the bowls. In the Pit in Sondazh 1, there was far less variability in choice of fabrics, with all vessel forms dominated by Fabric B (the red-grey fine wares varied between 50% and 67% of all sherds). This was not a fabric that was at all prominent in House A9. In contrast to House A9, the secondary preferences for dishes and plates were Fabrics A (pink fine ware) and K (dark red fine ware), while the secondary preferences for bowls and amphorae (Fabrics A, G and I) also differed from those in House A9. The general conclusion is that selection of fabrics for deposition in House A9 and the Pit was based upon contrasting principles which contributed to the formation of local identities.

5.1.4.5 Vessel Form (Analysis 5)78 It has already been explained why the form of the Nebelivka vessels has been studied by an alternative system to the Ryzhov system commonly used in Trypillia ceramic research (see above, pp. 266–268). The more detailed typology is based upon eight types, which include a broad category of ‘necked form’ for small rims and a ‘Rare’ category which includes such types as hole-mouth rims, binocular vessels and storage jars (Fig. 5.11). The more general typology recognises open types (dishes and plates) and closed types (bowls, amphorae and flasks), as well as an ‘Other’ category (necked forms, miniature vessels and bases). The abundance of otherwise unclassifiable bases in all excavation units – a total of 30% of all sherds – is managed through presenting two versions of both typologies by the inclusion and exclusion of bases from the statistics (Figs. 5.10/5–8 & 5.11/1).

78  See https://doi.org/10.5284/1047599 Sections 5.1.2.4; 5.2.1.2.4; 5.4.2.4; 5.3 for individual Test Pits.

Pottery 

 289

Figure 5.9: Fabric distribution for (1) House A9; (2) Mega-structure; (3) Test Pits; and (4) Pit, Sondazh 1; Fabric by sherd number for (5) Quarter G; (6) Test Pit 24/3; (7) Test Pit 25/1; and (8) Test Pit 25/2 (by J. Chapman).

290 

 The Finds

Figure 5.10: Surface colour vs. vessel form for (1) bowls; (2) carinated vessels; (3) plates; and (4) dishes, House A9; distribution of rim types without Bases for (5) Pit, Sondazh 1; (6) Test Pits; (7) Mega-structure; (8) House A9 (by J. Chapman).

Pottery 

 291

Figure 5.11: (1) distribution of shape types, all Units; distribution of open & closed categories without Bases: (2) all Units; (3) Pit, Sondazh 1; (4) Test Pits; (5) Mega-structure; (6) House A9; (7) Pit, Sondazh 1 SU2; and (8) Pit, Sondazh 1 SU4 (by J. Chapman).

292 

 The Finds

Inclusion of bases in the detailed typology has a variable impact on the results, especially for House A9, where 56% of the sherds were bases, but rather less for the other units, with 25% of bases or less. The over-representation of bases in House A9 emphasises that these house assemblages cannot be considered as ‘living assemblages’ (Schiffer’s ‘primary refuse’). Dishes were the most common form overall (at 28%) and also in three units, at 32–34%, with far fewer in House A9 (11%). Plates were represented in a similar range (10–16%) in all units. The distribution of necked forms followed that of dishes – a similar range in three units (10–11%) but far fewer in House A9. Removing bases from the distribution shows that, while all eight types were found in each unit, dishes now became the most frequent form at 40%, with plates at 18% and necked forms at 14% (Fig. 5.11/1). All other types were found at frequencies of lower than 10%, with the Rare type least common at 2%. In all units except House A9, dishes were found more frequently than plates – by a factor of three in the Pit and a factor of two in the Mega-structure and the test pits. Necked forms were found in one vessel in eight or nine across the site, while there was a special concentration of miniature vessels in the test pits. The same pattern of a preference for dishes and plates over amphorae, bowls and flasks was found in all Quarters. An unusual occurrence was the concentration of 14 miniature vessels in five houses in Quarter H – perhaps related to the special contents of miniature vessels as suggested by lipid analysis of the Mega-structure group of small pots (see below, Chapter 5.2.3.4).

5.1.4.6 Comparisons Between Open and Closed Forms (Analysis 6) The general typology excluding bases shows a strong preference for open over closed forms (Fig. 5.11/2), with the highest in certain Quarters (from 75% in Quarter H to 90% in Quarter N) and a preference for open forms in the Mega-structure (67%) (Fig. 5.11/5). Closed forms peak in the Pit at 24%, while the lowest total occurs in the Mega-structure (12%). The peak of ‘Other’ forms occurs in the test pits (Fig. 5.11/4). Inclusion of bases in the general typology biases the results towards the ‘Other’ category, which peaks at 75% in House A9 with its high number of bases. As in the typology without bases, the highest proportion of open forms occurs in some Quarters and then the Mega-structure, the lowest in the test pits. There are major variations in the open: closed ratios in the different Stratigraphic Units of the Pit, with open forms usually dominant and increasing with time. The often variant values in SU 2 are found in the unusually high frequency of closed forms – 15–22% higher than in other SUs (Fig. 5.11/7–8). By contrast, there is a strong preference for open forms in all Phases of the Mega-structure (range: 60–70%). The use of large plates for communal eating of collectively prepared food is a particular sign of the Mega-structure, which suggests that feasting was an important practice in this building (Fig. 5.14/2). The general preference for plates and dishes is particularly characteristic of the Nebelivka assemblage. What is interesting is the

Pottery 

 293

overall rarity of storage vessels of any kinds, although the so-called ‘pithos’ form is known from the podium in the Mega-structure. Given the coverage of all parts of the house in 80+ test pits, it may be expected that some storage facilities and/or vessels would have been discovered – but only two cases have been identified – in Test Pits 22/1 and 22/3, both placed near the end-wall of the house. The null hypothesis is as unsatisfactory as it is currently untestable – that communal storage facilities and large numbers of storage vessels were located in parts of the megasite that have not yet been investigated.

5.1.4.7 Estimation of the Minimum Number of Vessels (MINV) (Analysis 7) In addition to MINV estimates for the assemblage of each excavation unit, MINV estimates have been made for more detailed components of two of the excavation units – the test pits and the depositional episodes in the Pit in Sondazh 1. However, the depositional context of these two units is strongly contrasting. The test pit samples derive from a small (2–4%) part of the house, usually from the destruction of the house. The episodes in the Pit represent concentrations of deposited sherds in more or less continuous depositional action, when more sherds than usual were gathered from other contexts of primary or secondary deposition and then thrown into, or placed in, the Pit to produce an ‘episode’ of deposition. The sherd clusters were usually found as a complete unit in one zone of the Pit, rarely covering more than 1m × 1m in area. A total of 73 Test Pit assemblages was studied for MINV estimates. The mean MINV was 5.2, with a standard deviation of 4.8. Sixteen Test Pits with samples of 101 or more sherds were selected for more detailed MINV estimates. The question of sherd refits complicates such estimates in two cases. In Test Pit 28/2, 76 refitting sherds (total weight – 0.975kg.) yielded a single decorated amphora, while, in Test Pit 25/1, two complete dishes were reconstructed from 51 sherds (total weight – 0.449kg.) Otherwise, the mean number of sherds per vessel ranged from seven (Test Pit 23/2) to 29 (Test Pit 1/2) and the MINV from five (Test Pit 15/1) to 15 (Test Pit 23/2) vessels, with no statistical relationship between the two values. The number of types in the vessel groups ranged from one to seven types, with two-thirds of groups dominated by dishes and with plates, miniature vessels and amphorae also occasionally most frequent. Considering that the test pits represented no more than 2–4% of complete house floors, it is remarkable that such high MINV frequencies have been recovered. But it is also important to note that almost all of the vessels were recovered as fragments – often quite small fragments – of the vessels. This reinforces the case for the deposition of not only complete vessels but also sherds in the death assemblages of houses at Nebelivka.

294 

 The Finds

Figure 5.12: (1) Minimum Number of Vessel estimates for Episodes and deposits outside Episodes, Pit, Sondazh 1; (2) Vessel sizes by excavation unit (by J. Chapman).

A total of 29 episodes of deposition in the Pit can be used for MINV estimates.79 A poor fit was found between the number of sherds in an Episode and the MINV, indicating preferential deposition of rims in some Episodes and body sherds in others. There was an even wider range of vessels than in the test pit MINVs – from body sherds only in Episode 19 to an estimated total of 43 vessels in Episode 12 (Fig. 5.12/1). By

79 Photographs of the Pit 1 Episodes, with a full list, can be found in https://doi.org/10.5284/1047599 Sections 5.4.2.1 & 5.4.4.3.

Pottery 

 295

any standards, this latter was a major depositional event – an important ceremony bringing together several households or an entire Neighbourhood – while other Episodes were clearly smaller-scale and more intimate in nature. Ten of the Episodes comprised an estimated 20 or more vessels, prompting the question ‘was there a standard ceramic group for deposition in Episodes’? The variability in these 10 Episodes (Fig. 5.12/1) shows that the answer is ‘no’ but the regular deposition of between four to six of the six commonest types shows a polythetic pattern of discard. This regularity was reinforced by the preference for dishes in nine out of the 10 Episodes with large samples; bowls predominated in the only exception (Episode 8). This suggests that the same practice of depositing dishes used for personal or small group consumption was used in both units. A second regularity was the practice of depositing two to four plates in each of the larger Episodes, suggesting that collective preparation and consumption of food was also important before deposition in Episodic mode. The range of sherd numbers in the Episodes makes an interesting comparison with sherd numbers for deposition outside the Episodes, whether spatially or temporally. Sherds deposited in Episodes comprised three times the number of sherds placed outside the Episodes. The mean number of sherds placed in Episodes exceeded the mean for sherds placed outside Episodes (152 cf. 90 sherds), with each type of deposit showing wide variance in sherd numbers (within Episodes: 2–394 sherds; outside Episodes: 4–184 sherds). Such variability in sherd deposition supports the notion that deposition within and outside Episodes were essentially similar practices but with a more concentrated, perhaps formalised mode of discard in the Episodes. This conclusion finds further support from the MINVs estimated for deposition outside the Episodes (Fig. 5.12/1), with values of up to 20 vessels in two areas, and the finding of a similar spread of vessel types in these pot groups.

5.1.4.8 Vessel Size (Analysis 8) Given the rarity of complete vessels, the proxy measure used for vessel size is the rim radius, which has been divided into three size classes: small (2.5–7.5cm), medium (8–15cm) and large (15.5cm and above). The complete sample of rim sherds shows a size distribution of 35% small vessels, 40% medium-sized vessels and 25% large vessels (Fig. 5.12/2).80 Three of the excavation units show similar profiles to the whole sample; it is only in the test pit data that we found a small variation, with rather more mediumsized vessels and fewer small vessels than usual. This overall similarity in size profiles suggests that all of the assemblages under study derived from the same general pool of vessels produced for generalised rather than specialised household utilization.

80 Illustrations of pottery by size (e.g., for the Mega-structure) can be found in https://doi. org/10.5284/1047599 Section 5.1.2.4; see finds from other Excavation Units in Section 5.

296 

 The Finds

Figure 5.13: Upper: distribution of decorated vs. undecorated sherds; lower: distribution of rims by weight, Mega-structure (by M. Nebbia).

Pottery 

 297

5.1.4.9 Distribution of Vessel Types in the Mega-Structure (Analysis 9) The Total Station recording of sherds from the Mega-structure allows a GIS-based distribution of pottery types and decorative styles (for decoration, see below, p. 299ff). Six types of distribution can be recognised (Fig. 5.14): a) all-over heavy scatter, found for body sherds, painted fine wares and all vessels with decoration (Fig. 5.13 upper); b) all-over medium scatter, found for rim sherds (Fig. 5.13 lower); c) medium scatter in most areas, found for sherds with impressed decoration (Fig. 5.14/1); d) all-over thin scatter, found for bases, handles & lugs, coarse wares, plates, necked bowls, dishes, necked dishes (Fig. 5.14/2), fine and medium necked forms; e) specific clusters or areas, found for bowls (Fig. 5.14/3) and one cluster of miniature vessels; and f) singletons, found for complete vessels, amphorae and other miniature vessels. Given that the all-over heavy scatters merge sherd deposits from all Phases, the Phase distributions give more precision to the spatial analysis. Unless not recognised and excavated, signs of pre-Mega-structure activity were limited to activity under the place where the podium would be built. The ten or more sherd clusters placed on the living floor, mostly in the North-East rooms or near the walls (Fig. 5.14/4), suggest depositional events or episodes involving a wide range of social groups – whether households, Neighbourhoods or even Quarters – all contributing the sign of their identities to what was the largest Assembly House at Nebelivka. Equally, the six or more sherd clusters found in the open central area or outside to the West of the Mega-structure betoken similar episodes, even though we cannot date them precisely. What the sherds discarded during the Destruction phase underlines is the massive collective scale of deposition in many different episodes by many different social units – amounting to over 2,500 sherds and almost 60kg of ceramic. This was, in many ways, the defining collective depositional process of the Mega-structure, created by contributions from occupants all over the megasite at the time of the Mega-structure’s burning. These vessels or sherds could not have been placed on the Mega-structure living floor, since they would have been discovered in a Living Floor context! This means that this large number of vessels and sherds were left in places from which they fell (e.g., shelves, hooks, wall niches) with walls and ceilings to form the destruction deposit. The distributions of the individual pottery types present us with an interesting absence – the lack of any functionally coherent pottery groups. By contrast, we find overall thin scatters for the majority of types, including the most common types – the dishes and plates. This means that we are not looking at a collection of ‘living assemblages’ sensu Schiffer (1976) but, instead, a long series of collective depositions placed according to a Nebelivkan logic of quotidian practice which was certainly related to collective consumption and probably feasting.

298 

 The Finds

Figure 5.14: Distribution of (1) impressed sherds by weight; (2) plates, dishes and necked dishes; (3) bowls by weight; (4) pottery found on Living Floor (Phase 2) by weight, Mega-structure (by M. Nebbia).

Pottery 

 299

5.1.4.10 Decorative Style by Vessel Type (Analysis 10) This analysis considers the relationship of the vessel form derived from the detailed typology to decorative style, using the Pit assemblage in Sondazh 1. This analysis shows strong trends in this relationship, suggesting well-developed rules governing the decoration of vessels. Plates are the only type to which no impressed, incised or grooved decoration is applied, with all decoration being painted. There is a major component of painted decoration in the amphorae, dishes and flasks, although a small proportion of non-painted motifs is found on these forms as well. Impressed motifs are almost as common on necked forms as are painted motifs, while the only type with more impressed than painted decoration is the bowl.

5.1.4.11 Distribution and Placement of Decorative Motifs by Excavation Unit (Analysis 11) The typology of the decorative motifs at Nebelivka is based upon an extension to the test pits and House A9 of the research of Ms. Sophia Arbeiter, who created a detailed typology of the motifs found in the Mega-structure and the Pit in Sondazh 1 (Arbeiter n.d.; Caswell et al. 2016). The unit of analysis – the individual motif – was selected because of the small size of most sherds. A total of 169 decorative motifs was defined in three overall groups: 41 nonpainted motifs, 72 motifs painted on vessel exteriors and 56 motifs painted on vessel interiors (Figs. 5.15–5.18). The non-painted motifs and the exterior painted motifs were used more frequently than the interior painted motifs. A comparison of the excavation units where decorative motifs occurred shows the greatest variation in use of interior painted motifs, ranging from 34% of all motifs in the test pits to 60% in both the Mega-structure and the Pit, Sondazh 1. Non-painted motifs ranged from 51% of all possible motifs in the test pits to 72% in House A9, while the narrowest variation in use occurred with exterior painted motifs – a range of 62–72% across all excavation units. There are three trends in the developing choice of motifs in the Stratigraphic Units in the Pit. An increase in nonpainted motifs with time was matched by a decline in exterior painted motifs, with a greater choice of interior painted motifs in the middle part of the Pit. These results indicate that the choice of decorative motifs made a contribution to the identity of local groups at Nebelivka.

300 

 The Finds

Figure 5.15: Coarse ware decorative motifs. Numbers (e.g., 3.1) refer to Motif Numbers. Key: T – Test Pits; M – Mega-structure; P – Pit, Sondazh 1; A – House A9 (L. Woodard).

Pottery 

Figure 5.16: Fine ware exterior painted motifs (by L. Woodard).

 301

302 

 The Finds

Figure 5.17: Fine ware interior painted motifs (by L. Woodard).

Pottery 

 303

Figure 5.18: Fine ware exterior (rows 1–4) and interior (rows 5–6) painted motifs (by L. Woodard).

304 

 The Finds

A more nuanced picture emerges from the analysis of Quarters, where the samples are much smaller than in the full excavation units. There, differences are observed in the selection of non-painted motifs, interior painted motifs and exterior-and-interior painted motifs in different parts of the megasite. Non-painted motifs show consistently moderate values except in Quarters C and F. Interior painted motifs were absent in the Northern part (Quarters G and H), infrequent in the South-East area (Quarters L and N) and most frequent in the Eastern and North-Eastern parts (Quarters B, C and F). Exterior-and-interior motifs were not found at all in the Northern and NorthEastern areas but used most frequently in the South-Eastern area (Quarters L and N). These differences in motif placement show that people living in different parts of the megasite – perhaps also in different decades – were making choices as much at the Quarter level as at the household and Neighbourhood level.

5.1.4.12 Decorative Motif Combinations (Analysis 12) In view of the overall mean sherd weight of 23.7g for the Nebelivka assemblage, it is perhaps not surprising that there is but one motif on 75% of all decorated sherds. The incidence of sherds with one motif only rises to 96% of painted sherds. Most of the non-painted motifs and those sherds with combined motifs on the exterior and interior surfaces were deposited in House A9, while exterior and interior painted motifs were preferentially placed in the Mega-structure. The frequency of decorated sherds with multiple motifs decreases as the number of motifs increase. All decorational locations displayed combinations of both two and three motifs but four-motif combinations were found only on non-painted motifs and exterior painted sherds. Two-motif combinations showed a varied distribution in excavation units, with non-painted motifs equally represented in House A9 and the Mega-structure, exterior painted motifs most frequent in the Pit, exterior-and-interior painting most often in House A9 and interior painting evenly spread. These patterns reinforce the previous conclusion of identity-formation through choice of decorated sherds for final deposition. These performances are especially striking in the burning of House A9 and the Mega-structure, but the multiplicity of smaller performances in pit deposition should not be overlooked.

5.1.4.13 Motif Linkage (Analysis 13) One of the most detailed analyses of decorative motifs concerns the distribution of sherds with specific motifs in the 80+ test pits. From a total of 169 motifs, only 22 were found in three or more test pits. Eighteen cases comprised exterior painted motifs, with only one interior painted motif and three non-painted (impressed) motifs. The rarity of interior motifs is curious in the light of the high frequency of open forms in the test pits. Most of the 22 motifs were found in between three and seven Test Pits, while only three motifs linked more than 10 Test Pits (see below, Fig. 5.19).

Pottery 

 305

Two more detailed plots of motif linkage have been prepared: the four most popular motifs across the entire megasite (Fig. 5.19) and all 11 motifs in the seven test pits excavated in Quarter B (Fig. 5.20). The four most popular motifs were found as follows: Impressed Motif 3.1 (12 test pits) and Exterior Painted Motifs 4.8 (25 test pits), 4.1.2 (15 test pits) and 4.16 (10 test pits) (Fig. 5.19). It should be noted, however, that two of these motifs were relatively simple and had possibly been included in many motif combinations. Only one widespread motif (4.16) was more complex, with inclusion in many combinations unlikely. The distribution patterns of these four motifs show complex relations across the megasite, first related to the overall presence of motifs, then to single motifs, then to motif combinations and fourthly to zonal variations (Table 5.5). Table 5.5: Location of single motifs and motif combinations (Figs. 5.15–5.18) by zone, megasite (by J. Chapman). MOTIFS

EAST SIDE

NORTH SIDE

WEST SIDE

SOUTH SIDE

PRESENCE (SOLO OR IN COMBINATION)

11 Test Pits

14 Test Pits

7 Test Pits

9 Test Pits

3.1

OC/IRS

OC/IC/IRS

OC

IC/IRS

4.8

OC/IC/IRS

OC/IC/IRS

OC/IC/IRS

OC/IC/IRS

4.1

OC/IC/IRS

IC/IRS

OC/IC

OC/IC

4.16

IC/IRS

OC/IC/IRS

-

OC/IC/IRS

3.1

OC

IRS

-

IC/IRS

4.8

OC/IRS

IC

IC/IRS

OC

4.1

OC/IRS

-

IC

OC

4.16

-

IC

-

-

3.1 + 4.8

-

-

OC

-

3.1 + 4.1

OC/IRS

-

-

-

3.1 + 4.8 + 4.1

-

IC

-

IC

3.1 + 4.8 + 4.16

-

OC

-

OC

4.8 + 4.1

IC

IRS

OC

-

4.8 + 4.16

IC/IRS

IRS

-

OC/IRS

4.8 + 4.1 + 4.16

IRS

IRS

-

-

4.1 + 4.16

IC

-

-

-

SOLE MOTIFS

COMBINATIONS

Key: OC – Outer Circuit; IC – Inner Circuit; IRS – Inner Radial Streets

306 

 The Finds

We begin with the overall distribution of single motifs, whether found on their own or in combination. Some motifs, such as 4.8, occurred widely in every zone of all four sides of the megasite, while others occurred in most sides of the megasite but in different zones (e.g., Motif 4.1, found in all zones on the East side but absent from the Inner Radial Streets on the West and South sides and from the Outer Circuit on the North side). Very few motifs were absent from all zones of one side of the megasite, as was the case with Motif 4.16, missing from the West side completely. A different picture derives from those test pits where only one motif of the four was found. With such motifs, the coverage is much more patchy, with no examples of a solo motif found in all zones of a megasite side. Most solo motifs were deposited in only one zone, with only Motif 4.8 found in each zone of every side. Motif 4.16 continues to show the narrowest distribution, found as a solo motif only in the Inner Circuit of the South side. Motif combinations also show a varied linkage pattern, with some combinations occurring in only one zone (e.g., Motifs 3.1 + 4.8 in the Outer Circuit on the North side only; Motifs 4.1 + 4.16 in the Inner Circuit on the South side only), most others found in more than one zone on more than one side and not a single combination found on all sides. Another way of understanding these motif linkages is to consider motifs from houses in the same Neighbourhood. Of the eight Neighbourhoods with two or more motifs represented, not a single Neighbourhood has two houses with identical motif combinations. The opposite case – two houses with mutually exclusive motifs – was found in Neighbourhood 45, while two Neighbourhoods opposite each other in the Outer and Inner Circuits (Neighbourhoods 75 and 76) shared similar motif combinations. Otherwise, difference dominated the Neighbourhood patterning, with the suggestion of the primacy of household identity in all cases. A more spatially intuitive way to understand these complex linkages is the examination of each side in turn. This will lead to a more spatially nuanced story for each side (e.g., ‘West side story’). The West side story has a negative character – the absence of Motif 4.16 is the only absence from all zones of a side in the entire mapped corpus. This suggests that people making painted wares with this complex motif had a distant relationship to all of the seven test-pitted houses on the West side. But other motifs (especially Motif 4.8) linked the West side to all other sides, even though few motif combinations occurred on the West side and, when found, only in the Outer Circuit. The main differentiating elements of the North side story were that it was the only side with Motif 3.1 in all zones and the only side in which Motif 4.16 appeared solo. In contrast to the West side, this side reveals many motif combinations, as well as the deposition of Motifs 3.1, 4.8 and 4.26 in all zones. The general picture is one of greater integration through motif linkage than on the West side.

Pottery 

Figure 5.19: Motif linkage plans for four most common motifs, megasite (by M. Nebbia).

 307

308 

 The Finds

The East side motif distribution shares the high frequency of motif combinations with the North side, although this side has the fewest combinations in the Outer Circuit. Motifs 4.8 and 4.1 were found in all zones of this side, as were they as solo Motifs. These data suggest lower integration than in the North side but more than the West side. The final story – the South side – shows more similarities to the West side than the other sides, with few motif combinations and those found mostly in the Inner Radial Streets. Motifs 4.8 and 4.26 were deposited in all zones on this side, but especially in the Outer Circuit and the Inner Radial Streets. These data show a picture of moderate integration through motif linkage. To summarise these complex patterns of motif linkage, potters created designs on vessels which were used and ultimately deposited as complete, or often broken before deposition, in the house where they were made or in other houses. The closest links between houses can be recovered by solo motif distributions, while the opposite trend – the absence of a motif from an entire side – attests to minimal relationships with other households. Between these extremes lies the vast majority of motifs, which demonstrate links between household at various scales – in the same Neighbourhood, in the same zone or on the same side. The motif linkage data shines a light on the real complexity of inter-household relations on a site as large and complex as Nebelivka. Further research on these distributions would undoubtedly provide most interesting results. The distribution of all 11 motifs in the seven Quarter B test pits (one in the Outer Circuit, four in the Inner Circuit and two in the Inner Radial Streets) shows remarkable diversity, whether at the individual test pit level, the Neighbourhood level, within and between zones. No two test pits shared the same range of motifs; for example, all four houses in Neighbourhood 13 showed different motif combinations. The only test pit in the Outer Circuit was the only place in the entire Quarter containing Motif 13.1. This variability continued in the Inner Radial Streets, where each house had different motifs, but there are more similarities in the Inner Circuit (Motif 4.8 found in three out of four test pits; Motifs 4.15 and 13.2 found in two out of four test pits). When comparing motifs across zones, only two of the 11 motifs were found in each zone (Motifs 4.15 and 4.23), while two more motifs were found in two of three zones (Motif 4.17 in both Outer and Inner Circuits; Motif 4.26 in the Inner Radial Streets and the Inner Circuit). This complex distribution suggests a basic differentiation at the household level, with multiple but highly variable links between households in the Quarter. If most of the pottery deposited in burnt houses consisted of placed deposits, this overlapping motif linkage suggests that people from other houses were making offerings to an about-tobe-burnt house by sharing their favourite motifs. In this sense, motif linkage tells us something about communal practices and house-burning rituals.

Pottery 

Figure 5.20: Motif linkage plan, Quarter B, megasite (by M. Nebbia).

 309

310 

 The Finds

5.1.4.14 Discussion of the Pottery Analyses The key aspect of the Trypillia ceramic Big Other is that, within an overall similarity of technology, form and decoration, there remains the possibility for local difference. This enabled local communities and households to adhere to the general principles of the Big Other without denying themselves the potential for varying practices best suited to the social milieu and the point of time in the overall megasite sequence in which they found themselves. The starting point of the discussion will, therefore, be the multiple similarities between the pottery assemblages from the four excavation units at Nebelivka, before we turn to the equally numerous differences. The first overall similarity concerns the similarity in the fabric profiles at Nebelivka. While there is certainly variation in the less frequent parts of the profile (see below, p. 313), the most popular Fabrics showed similar frequencies in three of the four units. This implies a basic strategy of clay and temper procurement from the Nebelivka micro-region and similar practices in clay preparation. Similarities can also be seen in the shapes and sizes of the discarded vessels. Although the frequencies of use to which the vessel types were put did indeed vary, the same range of shape types can be seen in every excavation unit. This indicates that the basis for vessel use was indeed widely shared, whether the ubiquity of bowls and necked forms for food preparation, the plates and dishes for larger- and smaller-scale consumption and the ubiquitous rarity of the deposition of storage jars. In terms of vessel size, the balance between small, medium-sized and large vessels, despite one minor variation in the test pits, shows the discard of a generalised assemblage in all excavation units, with rare indications of specialisation in form (one example might be the set of 21 miniature vessels that probably fell off a shelf in the Mega-structure, a second the concentration of miniature vessels in houses in Quarter H). This suggests that those depositing the vessels and sherds were making those selections from a stable, enduring pool of vessels created through similar production in many parts of the site. Another similarity concerns the use of the same decorative styles in all parts of the site. The predominance of the painted style in comparison with impressed, incised, grooved and plastic styles can be regarded the hallmark of Trypillia pottery. But it is important that decoration is regarded as a mainstay of the pottery Big Other, even though variations in decorative motifs, and in the placement of motifs, may characterise the different parts of the megasite. A further similarity concerns the shared rules for the application of certain decorative styles to specific vessel forms. The Trypillia practice, which lies somewhere between the principle of mutually exclusive categorisation and more flexible, cross-cutting categorisation (Keightley 1987; Chapman & Gaydarska 2007, Chapter 2), is widely shared across the megasite. The last two similarities concern the choice of the nature and size of the pot part to be deposited. The distribution of pot parts is remarkably similar in all excavation units, showing a cumulative convergence of depositional practices across the megasite. This convergence extends to sherd sizes as plotted on almost identical cumulative frequency graphs. We can envisage similar fragmentation techniques applied to the

Pottery 

 311

full range of vessel forms, which in turn led to the similarities in potpart distributions. Thus, in important stages of the life of many Nebelivka vessels, there was convergence towards the similarities which were vital to the continued working of the Big Other at the local level, played out in month after month, year after year and generation after generation in the maintenance of a pottery tradition. The people using vessels engaged with a broadly similar range of statements about how a pot should be as those who made the vessels  – thereby forming a stable material world throughout the megasite sequence. In the Neighbourhoods and Quarters of Nebelivka, household members were most likely to have used the pottery and then deposited or discarded the vessels in their own house or another household (cf. Wengrow 2001). However, the other side of the Big Other was the local variability which was possible without posing a threat to the global concepts. This variability was more often the greater or lesser preference for a particular ceramic trait rather than its presence or absence. More nuanced than a presence/absence dichotomy, this graded variability is a tacit reinforcement of the ceramic Big Other and can be seen in many of the analyses of the Nebelivka pottery. An important question was whether the Nebelivka residents were making identities more through the use of cross-cutting variability (e.g., the tensions between household, Neighbourhood and Quarter identities) or with polar opposites (e.g., male–female, first settler–latecomer; cf. Chapman & Gaydarska 2007, Chapter 2). To explore this question, we turn to the results for each excavation unit. The test pit sample is the megasite sample in which temporality – mediated by the social – could have had the strongest effect on individual test pit pottery; after all, the difference in the date of dwelling in these houses may have been as much as five generations, or 200 years. Although the Project has been unable to produce a detailed inner chronology for Nebelivka, we should be aware that mediated temporal factors may have been significant in ceramic differences between test pits. The test pit sample shows depositional practices more focussed on decorated sherds than the other units; the decorated sherds in the Test Pits are larger than in the Pit and House A9 but the same size as sherds from the Mega-structure. Counterfactually, the test pit decorated sherds made use of the lowest proportion of interior painted and non-painted motifs of all units, although the high number of non-painted ware two-motif combinations was matched only by the Mega-structure. The distribution of decorative motifs found in three or more test pits showed an even spread with one motif (4.8), a preponderance in the Northern half in two cases (Motifs 3.1 and 4.1) and a predominance in the North-Eastern half in the remaining motif (4.16). These distributions of common motifs underlined the links between all parts of the megasite but also suggested stronger linkage in the Northern part. The distribution of test pit fabrics differed from all other units except the destruction phase of the Mega-structure in a preference for two fabrics and an even distribution of many other fabrics. This distribution is consistent with the diversity of test pits in the total sample. The vessel size profile in the test pits differs from the other three similar profiles, with fewer small and more medium-sized vessels. The

312 

 The Finds

only distinguishing feature of vessel form in the test pits is the high concentration of miniature vessels – another point of resemblance to the Mega-structure. However, the predominance of dishes in the test pit MINV series showed a tendency towards the deposition of food preparation ceramics. In summary, the test pit sample showed the kind of diversity that was expected from the combination of many small house samples but with an added emphasis on those decorated sherds marking local identities. Despite the tendency for the deposition of food preparation ceramics, the test pit sample showed closest links to that of the Mega-structure. The analysis of all the test pit pottery found in a specific Quarter led to questions being asked of the pottery at a different spatial scale – intermediate between the megasite as a whole and the individual test pits. Differences between these smaller ceramic assemblages can be seen at three spatial scales: parts of the megasite, the Quarter itself and houses within some Quarters. Area differences were observed in the choice of motif placement in three out of the four general decorative categories – nonpainted motifs, interior painted motifs and exterior-and-interior painted motifs. They were also found in variations in the number of fabrics used in a Quarter. There are several cases where most Quarters conform to a pattern, with one or two exceptions. One of the most striking cases is the preference for open rather than closed vessel forms in most Quarters. However, in Quarter G, the lowest ratio of dishes to plates in all the Quarters characterised a pottery assembly with the greatest degree of inter-household fabric variation. One interpretation of these observations is that communal feasting was particularly important as a way of integrating people from a wider-than-usual range of small sites in the Nebelivka catchment. The variety of household choices made for vessel fabrics in the four houses in Quarter G contrasts with the relatively homogenous household choices in Quarters B and M. This may be a sign of different clay source preferences, different constituents of the households in a given Quarter (more variable in Quarter G, less variable in the other Quarters) and may also involve different dates of dwelling. Bayesian modelling of dates for houses within each Quarter suggest few chronological differences but, in Quarter G, Test pit 25/3 post-dates all other samples excepting that from 25/1 which itself predates most other samples. Equally, in Quarter E, Test Pit 20/1 is earlier than Test Pit 20/3 and 35/1. In general, we are still finding it hard to make progress with fine-grained dating of houses, Neighbourhoods and Quarters at Nebelivka (see Section 4.8). The Mega-structure sample showed similarities and contrasts both within the various Phases and also with the other units. There was strongly preferential deposition of large, decorated sherds on the living floor before the burning of the Mega-structure, a far greater number of sherds occurring in the destruction phase and a large number of small, undecorated sherds found after the burning of the structure. The preference for decoration was emphasised by the highest number of combinations of both interior and exterior painted motifs and the common use of non-painted motifs, as in House A9. There was a shared preference for three Fabrics

Pottery 

 313

in all phases but the choice of many Fabrics in sherds deposited in the destruction phase suggests a diversity of contributors to the burning event – supposedly from other Neighbourhoods and even Quarters as well as local households. The emphasis on communal consumption – perhaps feasting – is shown in the Mega-structure sample by the highest proportion of open forms in any unit. In summary, the Megastructure sample shows a greater tendency towards communal consumption and large decorated sherd deposition than in any other unit – both choices related to the performance of burning the Mega-structure. There are no greater similarities with any of the three other samples – just traits overlapping with each unit. The House A9 sample shows a wider range of differences from all of the other three samples but, even here, the differences are graded rather than absolute. Thus, the House A9 sample is the only sample where plates outnumbered dishes – a contrast emphasized in the use of more different Fabrics for each form than in the Pit and in the variations in choice of secondary Fabrics too. This assemblage shows the highest proportion of non-painted ware motifs chosen than in any other unit yet, counterfactually, combines this with the lowest proportion of coarse wares deposited. The choice of a different Fabric profile from the other units and the lowest proportion of decorated sherds of all units serves to underline the differences between the House A9 sample and the other three units. It would be interesting to see if other complete house assemblages (such as House B17) showed such differences – as if to emphasise their individual identities in contradistinction to the other units. The strong preference for plates shows a tendency for deposition related to communal consumption on the occasion of burning the house. The dispersed deposition of vessels and sherds across mostly Room 1 shows that the material was carefully and deliberately placed before the burning of the house, so as to reproduce extant inter-household relations (see below, p. 326). Finally, the assemblage from the Pit in Sondazh 1 showed diachronic trends in its stratigraphic units (SUs) that were more significant than similarities and differences with other units. The earliest deposit (SU 1) included the largest fine and coarse ware sherds and even more decorated than undecorated sherds – a most unusual event in all samples. Although this initial deposit lacked special animal deposits (cf. at Majdanetske: Müller & Videiko 2016, pp. 79 & 86), the special emphasis on large and decorated sherds marks this out as an important communal event. Whatever happened next, in the SU 2 deposits, marked a contrast not only to SU1 but also with the later SUs. SU 2 reversed the diachronic trend towards smaller fine and coarse ware sherds and the increased use of Fabric A (pink fine ware). The choice of the highest proportion of closed forms in the entire Pit – itself selecting more closed forms than any other unit – shows that Pit deposits and especially the SU 2 deposits were hardly the result of feasting or communal consumption but rather of smaller-scale food preparation and consumption ceramics. This conclusion is borne out by the far higher ratio than usual of dishes to plates. The preference for Fabric B (red-grey fine ware) is comparable to other units but the preference for closed forms distinguishes the Pit

314 

 The Finds

from all other units. This suggests that, after an initial depositional event emphasising large and decorated sherds, the materials deposited in the Pit derived from domestic food preparation events, presumably from different adjacent houses. Nonetheless, we should not forget that parts of over 40 vessels were deposited in some of the Pit Episodes, suggesting large-scale deposition by several households not necessarily matched by communal consumption. The consideration of the pottery samples from the four different samples allows us to answer the question posed above: was the use of cross-cutting variability more important than polar opposites in the construction of megasite identities? The clear preference for graded differences rather than presence/absence variation indicates that a relational strategy was preferred, with the slow build-up of the remains of depositional events creating and maintaining the identities that related persons to all of their nested social contexts – households, Neighbourhoods, Quarters and ultimately the megasite itself. In the next section, we shall compare pottery developments at other Trypillia sites and, in a more detailed consideration, seek to identify whether or not similar strategies of pottery discard were utilised at the neighbouring megasites of Taljanki and Majdanetske.

5.1.5 Comparisons with Other Trypillia Pottery Assemblages The vast majority of Trypillia pottery studies81 is concerned with the typo-chronology of the assemblage in question (e.g., Ryzhov 2012a) – an issue which we feel is better addressed through AMS dating (see Chapter 4.7). No other site studies include a complete suite of the Nebelivka pottery analyses – indeed some analyses are not found in any other report (analyses 2, 8, 11 and 12), making inter-site comparisons somewhat limited. Ovchinnikov’s (2014) report on the Kaniv group considers the relationships between fabric and form, and form and decorative style, in a qualitative way. His shape typology (2014, p. 80, here Fig. 5.1) is used in preference to the Ryzhov system but Ovchinnikov does not focus on house or pit assemblages. Indeed, the number of reports with pottery presented as a series of ‘house assemblages’ or ‘pit assemblages’ is even smaller; in this comparison, we shall focus on the best example – the megasite of Taljanki (see annual reports) – together with the more limited recent UkrainianGerman excavations at Majdanetske (Müller et al. 2017). Additionally, in an excellent example of scholarly analysis, the study of Trypillia painted signs by T. M. Tkachuk (2005) provides house-by-house comparisons for each type of sign.

81  We do not consider all ‘pottery studies’ to be site ‘pottery reports’; the latter should include as a minimum a general consideration of the total assemblage(s) and the presentation of some form of pottery catalogue.

Pottery 

 315

Sherd numbers, weights and Minimum Numbers of Vessel (MINV) estimates are provided for the earlier Taljanki excavations (up to 2008). A total of 60,000 sherds was recovered from 39 complete house excavations plus other investigations, an unstated proportion reconstructed to 800 whole vessels or complete profiles (Ryzhov 2008, p. 134). A total of 23 reconstructed vessels came from House 32, and 56 vessels from House 33, from a combined sample of 776 sherds out of a total of 3,390 sherds in the two houses. Likewise, in the earlier Majdanetske excavations (1984–9), over 100,000 sherds, reconstructed to 2,000 vessels, came from 25 houses and 15 pits (Shmaglij & Videiko 2001–2, p. 89). The excavators estimate that between 1,000 and 5,000 sherds came from any single house, meaning that household living assemblages comprised between 20 and 130 vessels per house. This may be compared with the excavation of a claimed two-storey house – House 44 – at Majdanetske, where 1,735 sherds weighing 61.5kg were found, estimated to derive from a total of 37 vessels. Irrespective of the fact that a set of 130 vessels in one house would hardly leave any space for the inhabitants to walk around the house, let alone work, have sex, rest or sleep, many of the Majdanetske house totals far exceed the ethnographic data on household pottery assemblages, with mean and S.D. of 25±27 vessels in coeval use per house (Varien & Mills 1997). The only direct parallel to the Majdanetske data from Nebelivka showed 3,500 sherds from House A9, with a weight of over 80kg and a MINV (minimum number of vessels) of 192 (excluding bases). This indicates that, although smaller than the number of sherds estimated per house at Majdanetske, the House A9 assemblage exceeded the sherd weight of Majdanetske House 44 by 30% and far exceeded the MINV of any Majdanetske house. These data confirm that House A9 was a ‘death assemblage’ (an assemblage of vessels or sherds placed in a house before it was burnt: pace Schiffer 1976, ‘secondary refuse’) rather than a living house assemblage (Schiffer’s ‘primary refuse’). For comparison, the Nebelivka Mega-structure statistics show a total of 6,162 sherds, weighing 163kg, estimated to derive from a MINV estimate of 332 (excluding bases). As with House A9, the increase in estimated MINV in the Mega-structure is far greater than the increase in sherd number and weight at Majdanetske. It is possible to compare the MINVs from the 2 × 1m test pits at Majdanetske 2013 and Nebelivka (2013–14). In the nine test pits at the former, MINV estimates ranged from zero to 13, with a mean and S.D. of 4.3 ± 4.1 vessels, while the mean and S.D. at 73 test pits at Nebelivka was slightly higher, at 5.2 ± 4.8, with a range of zero to 23 vessels. While there were no systematic statistics on pottery discard from the older excavations at Majdanetske, two pits excavated in 2013 showed contrasting results. In Pit 50 (up to 1.2m in depth), 809 sherds weighing 20kg were estimated to derive from 39 vessels. There was considerable variation in pottery density in this pit, with highest levels reaching 3.5kg/m3, as well as large quantities of daub (581kg). In the deeper Pit 60 (1.5m in depth), far fewer sherds were found – 451, weighing 10kg and derived from an estimated 24 vessels – but they were accompanied by far more daub (1,332kg.) than in Pit 50. By comparison, the much larger Nebelivka Pit (3.5m in depth)

316 

 The Finds

in Sondazh 1 contained 6,948 sherds, weighing 122kg, with an estimated MINV of 640 vessels (excluding bases) but with far less daub. We have some way to go before understanding the pits on megasites in terms of their widely varying functions and contents. However, a common feature was large-scale special deposition82, implying inter-household performances – whether the 640 vessels in the Nebelivka Pit (Sondazh 1) or the more than one tonne of daub in Majdanetske Pit 60, aptly described as the burial of a house (Müller et al. 2017). There are few available statistics on the quantities of fine ware vs. coarse ware, or decorated vs. undecorated sherds, per site and, even more so, excavation unit. The Nebelivka pattern is very clear (over 88% fine wares; ~30% (by number) or ~40% (by weight) of decorated wares by excavation unit). This is replicated in the recent excavations at Taljanki (houses 32–33, 40–47), where fine wares remained above 90% by number, with exceptions in a few houses in the earlier excavations (e.g., House 2, with 83% fine wares). No data on fine wares or decorated wares are currently available from Majdanetske. While the analysis of vessel fabrics is presented as a description of visual studies at Majdanetske and Taljanki, there is technological support for the three fabric types found in the Kaniv group (Shevchenko, pp. 76–79, 92–94 & 96–97; in Ovchinnikov 2014). Fabric I predominates in the Kaniv sites (70–90% by number) and comprises a naturally sandy clay with occasional additions of crushed sherd, organic temper, haematite or limestone, fired to a range of orange hues in an oxidising atmosphere. The mixing of ferruginous clay with kaolinite produced a brick-red surface colour, rare in the Kaniv group but more common in the Nebelivska group. Fabric II comprised the so-called ‘Cucuteni C’ coarse wares – a greasy clay with shell or occasional organic temper, fired in a reducing atmosphere to produce grey hues but with uneven firing often giving variegated colours. Fabric III was a medium fabric, with no added temper, fired in a reducing atmosphere to produce contrasting colours on the interior (browns and red-browns) and exterior (dark brown). We can identify general similarities to the Kaniv fabrics in the Nebelivka assemblage, with Nebelivka Fabrics B and C closest to Kaniv Fabric I, the Nebelivka kaolinite Fabric O matching Kaniv kaolinite clay and Nebelivka Fabric H matching the Kaniv Fabric II. At Nebelivka, the combined frequency of Fabrics B and C ranges from 27% to 60%, with Fabric O (kaolinite) varying from 3% to 12% and Fabric H (coarse wares) up to 10%. What is rare at Nebelivka is the frequency of crushed sherd temper added mostly to kaolinite in the Kaniv sites. There is no parallel for our analysis of the fabrics of four vessel forms – bowls, amphorae, dishes and plates at any other site. The closest conjoint analysis of vessel shapes and fabrics was conducted for the various stages of the Kaniv group (Ovchinnikov 2014, pp. 143–5, 149). In the BII phase, Fabric II vessel shapes are restricted to necked

82  Large-scale depositional events are also implied by the often large numbers of figurines and the intense but thin animal bone scatters found in pits.

Pottery 

 317

bowls and it is only in late CI that dishes are also made in this Fabric. Similarly with Fabric III, the limited repertoire of flaring-rim dishes and lids in Phase BII is broadened in Phase CI with the addition of dishes, flasks and necked bowls. By contrast, there is considerable continuity in the Fabric I forms throughout the group, with additions of biconical amphorae in stage 3 and the narrowing of the repertoire in stages 4 and 5. Since Nebelivka has been dated to Phase BII only, the comparison can be only with Kaniv stage 1. The impression is that there is a much more variable relationship between vessel form and fabric at Nebelivka than in the Kaniv sites. In the following comparative section, we have used the Nebelivka equivalent of the Ryzhov/Ovchinnikov system (see Fig. 5.1 and Table 5.1). The comparison of the recently excavated Majdanetske units (House 44, Pits 50 and 60) with 10 recently excavated Taljanki houses shows the variability of pottery deposited in ‘burnt house assemblages’ (Fig. 5.21/1). While amphorae predominate in all three Majdanetske units, they are the commonest form in only two of the 10 Taljanki houses, which show a far higher proportion of dishes and plates in seven houses. The under-representation of bowls in the Ryzhov system is reflected in their rarity at Taljanki compared to their frequency in the Majdanetske units. Flasks (‘kubki’) show a variable presence in both megasites. The Nebelivka units strongly contrast with the Majdanetske and Taljanki groups, with far fewer flasks and amphorae, many more bowls and a similar frequency of dishes and plates to that of Taljanki but far more than at Majdanetske. These variations betoken differences in daily practices between the three sites which prompt a debate about the similarities in depositional strategies existing between the three sites. A ceramic indicator for intra- or inter-site differences in practice is the ratio of open to closed vessels. It is interesting to note a much higher ratio of open forms in the 10 Taljanki houses in comparison with the three Majdanetske units (Fig. 5.21/2). The pattern at Nebelivka is closer to that of Taljanki, with over 50% of open forms in all units; the Nebelivka Pit (Sondazh 1) has substantially more open forms than either of the Majdanetske Pits 50 and 60. To the extent that these figures conceal an emphasis on plates rather than dishes, the Nebelivka and Taljanki deposition suggests removal and re-deposition of vessels or sherds from contexts of collective rather than individual consumption to a far greater degree than occurred at Majdanetske, despite the possible interpretation of feasting for Pit 50 (Müller et al. 2017, p. 56). The GIS-based distribution of vessels in house units has but recently developed in Trypillian archaeology, whether at Nebelivka (2009, 2012) or Majdanetske (2014). The only published analysis from Majdanetske – House 44 – is complicated by the debate over a one- or two-storey house (see above, Chapter 4.1) and missing information83. The interpretation of a non-overlapping distribution of what has been claimed to be an upper-floor and a ground-floor distribution does, in fact, make good sense as a

83  Unfortunately, looters destroyed ca. one-third of the house surface (Müller et al. 2017, p. 34).

318 

 The Finds

Figure 5.21: (1) Composition of burnt house assemblages by vessel shape; (2) ratio of open: closed vessels; Taljanki (T) and Majdanetske (M); (3) regression analysis of painted signs vs. sample size, Bug-Dnieper Interfluve sites, based upon Tkachuk 2005 (by J. Chapman).

Pottery 

 319

complementary distribution on a single-storey house. The main pottery cluster of flasks, small amphorae, fine ware and storage vessels lay between the platform (aka ‘altar’) and the hearth, with dishes placed on the podium or in a corner, complementing the cluster of storage vessels in the centre. The sum total comprises one large, or a series of small, depositional events in which people from House 44 and from other houses placed vessels or sherds before the burning of House 44. By comparison, the one-storey House A9 pottery distributions showed a strikingly dispersed group of vessels and sherds, with not a single grid square containing more than one pottery cluster (Fig. 5.25 lower). Since we maintain that the House A9 assemblage of a minimum number of 192 vessels is far too large to represent a ‘living assemblage’ (primary refuse pace Schiffer 1976), this finding suggests placement of different types of vessel in different parts of the house before burning, as a staged ‘event’ which may have mimicked daily household practices but which, equally, may have conveyed another message about the way that pottery was used or the relations between occupants and visitors who contributed offerings. Many contrasts were observed in these distributions, the most general of which was the paucity of sherd clusters in in the Southern room (only two small clusters of bowls and three small clusters of impressed sherds) as compared to the six large clusters in various parts of the larger Northern room and the four large sherd clusters placed outside the house away from the three pits (Grid Squares E10, V1, V4 and ZH 3) – bowls, coarse wares and sherds with impressed decoration (Fig. 5.23 & 5.24 upper). More detailed contrasts include the placing of clusters of fine ware sherds near the East wall (Fig. 5.22 upper) with coarse ware clusters near the West wall (Fig. 5.23 upper); painted sherd clusters along the Southern part of the East wall (Fig. 5.22 lower) with impressed sherds in the North-East corner (Fig. 5.23 lower); bowls West of the Northern room platform (Fig. 5.24 upper) compared to dishes North of the platform (Fig. 5.24 lower); and bowls and dishes inside the house (Fig. 5.24) as contrasted with plates deposited outside the house (Fig. 5.25 upper). None of these clusters was exclusive in the sense that all bowls were found in the large and small clusters but they nonetheless indicate places of concentrated deposition, often of 10 or more vessels. These clusters can hardly be accidental, nor do they consist of functionally coherent assemblages  – they simply represent concentrations of specific vessels linked to particular food preparation or consumption practices. It is interesting, therefore, that vessels of individual or small-group consumption were deposited in different places within the house, while the plates used in communal consumption were placed outside the house. This last contrast suggests a difference in the location of the two styles of consumption – larger-scale outside the house and smaller-scale inside the house. The pottery distributions in the Nebelivka Mega-structure equally demonstrated the lack of any functionally coherent pottery groups, instead showing overall thin scatters for the majority of types. We interpret these distributions as a long series of collective depositions related to collective consumption and probably feasting.

320 

 The Finds

Figure 5.22: Distribution of (upper) fine wares; (lower) painted wares, House A9 (by M. Nebbia).

Pottery 

 321

Figure 5.23: Distribution of (upper) coarse wares; (lower) sherds with impressed decoration, House A9 (by M. Nebbia).

322 

 The Finds

Figure 5.24: Distribution of (upper) bowls; (lower) dishes, House A9 (by M. Nebbia).

Pottery 

 323

Figure 5.25: Distribution of (upper) plates; (lower) summary diagram, House A9 (by M. Nebbia).

324 

 The Finds

The relationship between vessel forms and decorative styles has been explored for the Kaniv group (Ovchinnikov 2014, pp. 143–5, 149). Similar rules of the mixture of decorative styles are found in the Kaniv and the Nebelivka assemblages, with both impressed and painted styles on flasks, necked bowls, everted-rim dishes and amphorae. We cannot compare the Nebelivka finding of the exclusive use of painted motifs on plates or the preference for impressed decoration over painting on bowls, since neither of these types is distinguished in the Kaniv group. The intra-site distribution of decorative motifs has not been attempted for any other megasite but there is a comparable analysis from Majdanetske for painted signs which Tkachuk (2005) considers to form part of the Trypillia sign-system or sacred pictographic script84. Tkachuk has studied the painted signs and sign-combinations for 10 sites in the Southern Bug-Dnieper Interfluve in Phases BII and CI. It is interesting to note that a regression analysis of the number of painted signs vs. the total number of complete vessels from each site showed a very close fit (Fig. 5.21/3). This shows that the variability in the number of signs is not related to differences in site size or hierarchical rank but, rather, to sample size. The only sites where Tkachuk compares the deposition of vessels with signs in individual houses are Taljanki (Tkachuk 2005, Vol. II, pp. 153–178) and Majdanetske (Vol. II, pp. 126–152 & Ris. 18). At the latter, comparison of the sign assemblages from 18 fully excavated houses built in parallel in a circuit over 120m (Shmaglij & Videiko 2001–2, Ris. 13) shows that almost all houses were linked to at least a dozen other houses in a complex, dense network of interaction (2005, p. 152). However, if the spatial analysis is limited to signs deposited in a minimum of six houses up to the maximum of 18, three different patterns are revealed (Fig. 5.26/1): (a) little engagement in the network in several houses in the Western and Central groups; (b) a number of links within the Eastern group and between the Eastern and Central groups; and (c) the highest number of links between the Western and the Eastern groups. Five houses in particular showed preferential links between each other, indicating targeted deposition of these significant vessels and, in turn, the probable differentiation of ritual deposition in these houses. A second analysis made possible by Tkachuk’s systematic database of signs by site is a comparison of signs from the Southern Bug-Dnieper Interfluve (SBD) sites with the 128 painted motifs from Nebelivka. A total of 20 painted motifs (or 15% of the total) can be compared with Tkachuk’s signs  – six motifs with close parallels, two with both close and general parallels and 12 motifs with a general similarity (Fig. 5.26/2).

84  Cf. Hudson & Milisauskas’ (2017) characterisation of the Trypillia sign system as ‘a series of signs that are syntactically structured by a linguistic or cultural grammar and housed in the group’s mental lexicon’.

Pottery 

 325

Figure 5.26: (1) distribution of painted signs shared between houses, Majdanetske; (2) comparison of Nebelivka painted motifs with painted signs on other Trypillia sites; (3) comparison of painted signs on Bug-Dnieper Interfluve sites (by J. Chapman, based upon information in T. Tkachuk 2005).

326 

 The Finds

While over half of the Nebelivka motifs were found at one to three SBD sites, two motifs with close similarities to the painted signs – although among the simplest – were found on all of the SBD sites in the study. The eight sites varied in the number of signs comparable to the Nebelivka motifs, with the three largest totals coming from the three closest sites – Volodymyrivka, Taljanki and Majdanetske (Fig. 5.26/3). This suggests that the painted sign-system was reinforcing the long-term network linkage in the SBD, thereby contributing to the stability of the Big Other. In summary, the comparison of pottery analyses conducted at Nebelivka with other Trypillia sites shows patchy results, with four key points: 1. Inter-site similarities between Nebelivka, the Taljanki and Majdanetske megasites and the Kaniv group sites emphasise the way that all of these sites readily conform to the overall canons of the Trypillia ceramic Big Other, whether in vessel forms, fabrics, fine ware preference, decorative styles and their relationship to vessel forms and engagement with the Trypillia sign-system. 2. The inter-site differences between these sites can be subsumed within the inherent regional variability of the Trypillia Big Other, which is expected to offer the possibilities of such differences between sites especially as they occurred at an intra-site level (e.g., the more variable relationship between vessel form and fabric at Nebelivka in comparison to the Kaniv group). 3. Variations in the number of vessels comprising the burnt house assemblages at the megasites show that, while some smaller assemblages may have been selected from living assemblages, the larger groups (e.g., Nebelivka House A9) show contributions from more than one household – perhaps several active houses in the Neighbourhood. 4. Variations in the ratio of open to closed vessels in megasite burnt house assemblages suggest that the vessels were collected from different kinds of discard deposits, in turn indicating variability in the disposal of the ceramic remains of daily practices such as cooking and feasting.

Bisserka Gaydarska, John Chapman, Marco Nebbia, Dmytro Gaskevych, Cătălin Lazăr, Theodor Ignat, Adrian Boyce, Amanda Dolan, Jason Newton, Oliver E. Craig, Harry K. Robson, Matthew von Tersch & Alexandre Lucquin 5.2 Special Finds We have divided the Special Finds from the excavations at Nebelivka into eight categories on the basis of material and type: (1) fired clay figurines; (2) fired clay tokens (aka ‘counters’); (3) the group of miniature vessels from the Mega-structure; (4) the other miniature vessels; (5) chipped stone; (6) ground stone; (7) bone tools; and (8) other Special Finds (which include other fired clay finds, unusual vessels and a single gold hair ornament). The detailed, fully illustrated catalogue of these finds is



 327

Special Finds 

available by excavation unit elsewhere (see https://doi.org/10.5284/1047599 Section 5). In this section, in addition to the specialist reports on the Special Finds, which include comparisons with other megasites and the wider context, we contribute a synthesis of the overall significance of the Special Finds at Nebelivka – in particular, what makes them ‘Special’.

Bisserka Gaydarska, John Chapman & Marco Nebbia 5.2.1 Figurines

Figurines have already been cited as one of the three cornerstones of the Trypillia Big Other (see above, p. 37) – the vital long-term framework for daily practice in this vast time-space network. It is therefore important to see how the Nebelivka people and potentially visitors to the megasite used figurines in their depositional practices. The sample of images from the Nebelivka fieldwork comprises a total of 143 fragments (Table 5.6). This total means that, following Ţerna’s (2017) metric, the mean number of figurines per 100m2 of excavation amounts to 5.0  – matching the mean number for smaller Trypillia sites rather than the mean of 4.0 for megasites (Gaydarska 2019). We were able to make a typological, contextual and fragmentation study of 74 images, comprising 78 fragments, with two additional possible figurines. Descriptions of each image are available in the ADS Archive (https://doi.org/10.5284/1047599 Section 5). Table 5.6: Image fragment count by excavation unit (studied images in BOLD) (by J. Chapman). MEGASTR.

H. A9

H. B17

H. B18

PIT NEAR B17

PIT NEAR OVEN B18

PIT NEAR PIT S1 TEST OVEN PITS

23

20

4

2

17

4

29

1

25

10

SURFACE 8

5.2.1.1 Making In accordance with local pottery-making, the images have been made in three fabrics: fine wares of oxidising colours with painted decoration on a well-smoothed surface (6 fragments); a rare dark burnished ware (2 fragments); and the commonest fabric – a variety of oxidising colours in medium fine ware, from yellow to dark brown. The heavy organic temper discussed for many Cucuteni-Trypillia images (Monah 1997, p. 219) was absent at Nebelivka. Otherwise, the clays used for the images were comparable to those used for pottery (see Section 5.1.3).

328 

 The Finds

Figure 5.27: Alternative pathways to making Trypillia figurines (by L. Woodard).

5.2.1.2 Types The images have been made into three forms from the same initial oval lump of clay: (a) broad shapes (termed ‘statuettes’: Monah 1997, p. 220); (b) thin, cylindrical shapes (termed ‘figurines’: 1997, p. 220); and (c) zoomorphs. A simple, 4- or 5-stage châine opératoire was used to make all three types, with choices of the form of the arms and the legs made at successive stages (Fig. 5.27). A single statuette was modelled in a seated position (Pit, Sondazh 1: SF 67). The few statuettes with preserved bases were made to be free-standing, while the figurines were pointed, to be inserted into a soft matrix (e.g., sand or a foodstuff) or simply lying down. Each of the three types was deposited in broadly similar proportions in all of the main excavation units studied here, with the exception of the absence of zoomorphs in the Mega-structure. The decoration of statuettes and zoomorphs was rare at Nebelivka (Figs. 5.28 & 5.29/2): six statuettes with painted decoration in dark on light paint, and a single example of



Special Finds 

 329

a domestic sheep with incised fleece (Fig. 5.31 upper). The publication of some of the otherwise unstudied images showed that the faces of two statuettes were modelled in a realistic manner (Burdo 2015, Ris. 2/1 & 2/3).

Figure 5.28: Decorated anthropomorphic figurines from Pit, Sondazh 1: (1) SF 23 and (2) SF 28 + 43: (a) back, (b) front, Nebelivka (by K. Harding).

5.2.1.3 Gender There is a complex relationship between sex and gender, with the former more inherent and the latter more performed. Gender was a primary characteristic of Cucuteni-Trypillia images. The maker could choose whether or not to depict gender and in which way (the addition of a tiny clay penis; applied pellet breasts; the incision of a pubic triangle). Subsequently, a gendered image could lose its gender through fragmentation, with the gendered part removed somewhere else and the non-gendered part deposited. There is thus an important difference between a deliberately non-gendered image and an image that has lost its gender information. Almost half of the Nebelivka images were deliberately non-gendered, with a quarter given female characteristics and very few rendered male (Figs. 5.29/1 & 5.30/5). The others  – just under 25%  – presented no remaining gender information. The majority of images with gender characteristics had lost their heads and/ or feet – exactly those parts of the body that normally did not bear gender information.

330 

 The Finds

Figure 5.29: Anthropomorphic figurines: (1) male, SF 3230; (2) non-gendered, Grid F12; Megastructure, Nebelivka (by V. Pankowski).



Special Finds 

 331

5.2.1.4 Fragmentation It cannot be coincidental that no examples of complete images have yet been found at Nebelivka. The body parts were classified according to the system which we developed for the Late Copper Age figurine assemblage at tell Dolnoslav (Chapman & Gaydarska 2007, Chapter 6). Twelve different body parts were found at Nebelivka – heads, torsos, bottoms, legs and feet, and combinations of the basic five parts (Fig. 5.30/1–4). The greatest variety of body parts was found in the Pit in Sondazh 1 – 11/12 parts, with the least (n = 3) found in the Test Pit sample. Torso-to-leg fragments were outnumbered by other types of fragments in the Test Pits (Fig. 5.46). Although sample size undoubtedly affects this statistic, the sample size of House A9 was almost the same as in the Pit but with the deposition of half as many body parts. This may mean that a wider range of people or households contributed to the deposition of images in the Pit or that different body parts were placed in different Pit episodes. The preferred body part in all units except the Test Pits was the torso-to-leg fragment (or ‘TL’). In all units, there was a deficit of heads – indeed only six heads were found in the sample of 74 individuals (viz., 8%; cf. 10% of heads in the Dolnoslav sample: Chapman & Gaydarska 2007, Table 6.2). This statistic provides strong support for the notion that the images were fragmented before their deposition. Two further arguments derive from the four cases of two re-fitting fragments from the same image. There was a single example of two fragments from the same statuette deposited in different stratigraphic units in the Pit Sondazh 1 (No. 28 in SU 4, No. 43 in SU 5) (Fig 5.28/2). This indicates that the statuette was broken, with one part deposited while another part was curated for an unknown period of time before itself being deposited. A second re-fitting example depicts a very unusual break of a statuette into a front part and a back part (Mega-structure Grid Square F11)  – a fracture that would have been impossible to achieve by accident. In general, the Nebelivka fragmentation data support the notion of deliberate fragmentation before deposition, with some evidence for curation of fragments and much evidence for removal of body parts (especially heads) to an as yet unidentified place or places.

5.2.1.5 Context of Deposition In their study of the Majdanetske images, Shmaglij & Videiko (2001–2) noted the preferential deposition of images in pits rather than houses – a finding confirmed by the more recent work at Nebelivka (Burdo & Videiko 2016). However, variable deposition rates have been found at completely excavated Nebelivka houses – 22 in House A9 and only four in House B17. Only one Assembly House has been excavated – the Mega-structure  – but, perhaps surprisingly, it contained no more images than House A9. The finding of only one image in each of 10 Test Pits and none in the remaining 78 shows that the strategy of test-pitting is not the best way to increase the figurine sample!

332 

 The Finds

Figure 5.30: Figurine body parts by excavation unit: (1) House A9; (2) Mega-structure; (3) Pit, Sondazh 1; and (4) Test Pits; Key to figurine parts: H – head; HT – head-torso; HB – head-buttock; HL – head-leg; T – torso; TL – torso-leg; TF – torso-foot; B – buttock; BL – buttock-leg; L – leg; LF – leg-foot; F – foot; TZOO – torso of zoomorph. (5) gender characteristics of figurines; (6) condition of figurines by excavation unit (by J. Chapman).



Special Finds 

 333

In terms of more detailed distributions, the Mega-structure images derived mostly from the destruction Phase 3, with a preference for deposition near the outer walls, with two Phase 2 images found near Platform 89. The marked absence of a concentration of images suggests episodic deposition and/or the lack of clear rules for deposition. Deposition in the Pit in Sondazh 1 consisted mainly of ‘episodes’, in which pottery and images were placed in shallow re-cut scoops. Because the fill of the scoops was the same soil as the matrix, it was difficult to be sure of the outlines of all the scoops (see Fig. 4.53) but it is highly probable that all the figurines were discarded in episodic deposition. The vast majority was discarded where there were most episodes – in the middle Stratigraphic Units (SUs 3 and 4)  – with only two deposited in the lowest SU.

5.2.1.6 Condition The condition of the Nebelivka images varied from a pristine surface with no erosion or wear to small, highly degraded fragments termed ‘unrecognisable fired clay lumps’. The overall pattern was the worse the surface wear, the fewer the images (Fig. 5.30/6). The only excavation unit with all stages of wear was the Mega-structure, where 92% of the images were worn; fewer images were worn in the other excavation units (mean = 70%). The variable wear in the Mega-structure may have resulted from differing firing conditions in the Assembly House burning as well as varying lengths of time that the images were curated before deposition. The general conclusion from the analysis of image condition was that curation of whole and fragmentary images was a widespread practice at Nebelivka.

5.2.1.7 Comparisons with Other Assemblages The images of the Cucuteni-Trypillia group have been well studied for many decades (Pogoševa 1985; Monah 1997, 2016; Burdo 2008). These general studies show that, in stylistic terms, the Nebelivka images resemble the Phase BII figurines from other sites such as Volodymyrivka (Passek 1949), Kolomiishchina II (Pogoševa 1985) and Voroshilivka (Gusev 1995). All three types identified at Nebelivka are also well known from the nearby Phase CI megasites of Taljanki and Majdanetske, where long-running excavations have recovered many figurines. At the latter, 340 figurines were found in the excavation of 25 houses and 15 pits from 1986–1991 (Shmaglij & Videiko 2001–2). Close parallels are often found, such as the incised fleece on caprines at Nebelivka (Test Pit 16/ 1: here Fig. 5.31 upper) and Taljanki House 44 (Kruts et al. 2010).

334 

 The Finds

Figure 5.31: Upper: zoomomorphic figurine, Test Pit 16/1; lower: types of fired clay tokens (upper by K. Harding; lower by L. Woodard).

To put the Nebelivka sample into long-term context, Gaydarska (2019) has extended Ţerna’s (2017) data on figurine density by size of excavated area to show that the site density of figurines stabilises in Phase B and declines in Phase C, with no increase in figurine density in larger sites (e.g., megasites). The same is true of realistic figurines. We interpret this as a sign of stability in the Trypillia Big Other irrespective of major changes in site agglomeration (Phases BII–CI) or population dispersion (Phase CII). The continuity in aggregate figurine use between smaller sites and megasites is perhaps surprising and requires an explanation (see Chapter 6). With reference to the density of figurine discard in houses and pits on the three megasites, the picture is more complicated than a simple preference for higher figurine discard in pits. The average number of images deposited in or near 20 houses at Taljanki and two from Majdanetske (2013 excavations) is fewer than three85. However, in exceptional houses such as Majdanetske House Π, a total of 40 images was found in what was presumably the house of a ritual leader or a centre of communal ritual. Whether the 20 images discarded in Nebelivka House A9 can be similarly interpreted remains an open question. A similar variability of figurine discard is found in pits.

85  The mean and standard deviation is 2.7 ± 1.4.



Special Finds 

 335

While no images at all were found in the 2013 excavations of three pits at Majdanetske (Müller et al. 2017), large numbers of both anthropomorphic images and zoomorphs were placed in the pits near House E (Shmaglij & Videiko 2001–2). It is clearly no longer acceptable to assume that figurine discard in pits is always greater than house discard, though this was often the case (cf. the deposition at the Cucuteni A–B site of Traian: Bem 2007). The overwhelming majority of images at both Majdanetske and Taljanki was discarded as fragments, echoing the Nebelivka practices. However, there was one example of breakage followed by deposition of fragments from the same image in two different contexts. Fragments from the same realistic head were found in pits under adjacent Houses Π and Y at Majdanetske (Shmaglij & Videiko 2001–2, Fig. 49/6). This example of enchainment through images is particularly interesting because it targets ‘places-to-be’ – areas marked out by the digging of a pit prior to house construction. Comparable inter-context re-fittings are found in Taljanki for other types of Special Finds.

5.2.1.8 Summary The Nebelivka images consisted of zoomorphs, statuettes and figurines (terminology following Monah 1997), distributed in all the main excavation units. They represented one part of the Trypillia Big Other but their low frequency conforms to the overall paucity of figurines on other megasites. The images had been made with simple operational chains from clays visually similar to those used for pottery-making and with few attempts at decoration. All the images were fragmentary, with limited re-fitting evidence and fragmentation evidence indicating deliberate breakage – a conclusion amply supported by the high number of missing body parts – especially heads  – from completely excavated units. In addition to being broken, over 2/3rds of images in all excavated units were worn, rising to 92% in the Mega-structure. The varied wear was caused by a combination of differing firing conditions in house fires and the length of time the images were curated before deposition. Many of the ceremonies involving images also included deliberate fragmentation and/or curation. The repeated practice of fragmentation indicates that the large numbers of missing parts enchained the houses and pits to other places, where the missing parts were discarded. Trypillia figurines reflexively contributed to the Big Other in two ways: as part of significant ceremonies (Burdo 2008) at the household, Neighbourhood and possibly even the Quarter level and also through a quotidian role in the enchainment of houses to each other and possibly to other places further away.

336 

 The Finds

Bisserka Gaydarska, John Chapman & Marco Nebbia 5.2.2 Tokens (Counters)

The small fired clay objects found occasionally on Trypillia sites have been termed ‘tokens’ or ‘counters’ in the Nebelivka field reports (see https://doi.org/10.5284/1047599 Section 5). Their defining characteristics are their small size (rarely more than 3cm in diameter and 2cm in height), their lack of decoration on the flat surface (with a few exceptions), the absence of a handle and their expedient manufacture. In the standard summary of fired clay stamp seals (or ‘pintaderas’) in the Balkan Neolithic and Copper Age, Makkay (1984, 2005) publishes a small number of undecorated conical fired clay objects86 without distinguishing them from the often elaborately decorated and much more frequent pintaderas. There is therefore a minimal overlap between pintaderas and tokens, which have been interpreted as children’s toys, gaming pieces or economic aids for counting (Shatilo 2015). The term ‘token’ in the Near East is more appropriate for stylistic comparison with the Trypillia objects. Plain tokens are first known from the 8th millennium BC, coeval with the development of domesticated plants and animals at sites such as Mureybit (Schmandt-Besserat 2010, p. 27). Schmandt-Besserat (2010, p. 32) suggests that these early tokens related to ‘casual daily life items’ which were not especially standardised in shape or size. Tokens first became much more differentiated, and subsequently more standardised, in the Uruk period, with the growth of administrative practices on urban sites (2010, pp. 28–32). A total of 31 counters has been found at Nebelivka, the vast majority being deposited in the Mega-structure, one found in each of five houses in test pits and a single example in the Pit in Sondazh 1. The tokens in the Mega-structure were concentrated in the East rooms and the West area, with occasional examples deposited near the Southern and Northern walls and in the unrooved central part but with a cluster placed outside the Mega-structure to the West (Fig. 5.48 upper). The crumbly nature of most of the tokens shows that these objects were formed from a single lump of red or brown clay and fired at a lower temperature than the images and a much lower temperature than the painted pottery. When well-preserved, the flat surface of the tokens is either round or oval. There is considerable variation in the shape of the cross-section, with six broad types in evidence (Fig. 5.31 lower & Table 5.7).

86  The Cucuteni-Trypillia tokens include four from Luka Vrublevetskaya (Makkay 1984: No. 127) and eight from Frumuşica (Makkay 1984, Nos. 68–75); others derived from Copper Age sites such as Sultana, Ruse and Ezerovo as well as Late Neolithic Turdaş (Makkay 1984, Nos. 255–262) and Liubcova (Makkay 2005, No. 72), while there is even an Early Neolithic Körös example (Makkay 1984, No. 99).



Special Finds 

 337

Table 5.7: Frequency of token types, Nebelivka (see Fig. 5.31 lower). Type

1

2

3

4

5

6

Form

domed

sub rectangular

irregular domed

nicked rounded

trapezoidal

irregular

Frequency

9

5

7

1

4

5

There was decoration on the flat surface of only three of the tokens: one Type 3 token with ribbed decoration and one Type 1 and one Type 2 tokens with incised ridges. No parallels for these motifs were found in Makkay’s pintadera catalogues. However, the poor preservation of the tokens, contingent upon their low firing, means that some decorational information may not have survived the destruction of the Mega-structure or the houses. Indeed, only just over half the tokens were found complete, with the edges of ten tokens (or 32%) eroded and seven tokens (or 23%) preserved in a fragmentary state. The possibility of enchainment through differential deposition of token fragments is lowered because of the poor preservation of the objects. The comparison of the Nebelivka tokens with those from Majdanetske is based upon a rather minimal publication from the latter (Shmaglij & Videiko 2001–2, Ris. 52). If Figure 52 represents the full range of formal variation, then even greater variation in shape was seen in a smaller assemblage – seven basic types with variations (presence or absence of vertical perforation). The cross-sections of several of the Majdanetske tokens closely matched those from Nebelivka. Since no Assembly House was excavated in the 1986–91 campaigns at Majdanetske, these tokens would have been deposited in pits or houses. In summary, the poor quality of the clay, the low firing temperature and the crumbly surface appearance of most of the Nebelivka tokens indicates expedient production, making it improbable that these objects were used in a ritual or administrative role in a complex society. The variability of the tokens’ shapes is at odds with the standardisation of tokens used in early urban contexts at Ur (SchmandtBesserat 2010). However, the key concentration of 80% of the tokens in the Megastructure is suggestive of a more formal role for tokens than simply children’s toys; the most likely use was as gaming pieces made for ceremonial games or divination in Assembly Houses (Shatilo 2015). Two fragments of a possible gaming board have also been discovered in the Mega-structure (see below, p. 378).

338 

 The Finds

John Chapman, Bisserka Gaydarska, Dmytro Gaskevych, Cătălin Lazăr, Theodor Ignat, Adrian Boyce, Amanda Dolan, Jason Newton, Oliver E. Craig, Harry K. Robson, Matthew von Tersch & Alexandre Lucquin 5.2.3 The Group of Miniature Vessels from the Mega-Structure

Miniature vessels have been part of the Nebelivka story ever since Shmaglij’s first test pits there in 1981 (Shmaglij & Videiko 1992). In the current Project, a total of 84 such vessels has been found, some in each excavation unit. The largest group derives from the Mega-structure (n = 33) mainly because of the find of a group of 21 miniature vessels in a destruction context (https://doi.org/10.5284/1047599 Section 5.1.2). Theirs is a story of two parts – the exterior and the interior. It was only when the calcareous crust was removed from these vessels that the true significance of their exterior was revealed. Six vessels were decorated in a graphite-based wash, while graphite-painted motifs were found on three vessels and possibly a fourth. Although this is the first time when graphite-painted decoration has been found on Trypillia pottery, it is not the first use of graphite in pottery production in the prehistory of the Ukraine. The following summaries of the specialist reports explore the early use of graphite, the stylistic parallels of the Nebelivka graphite-painted vessels with East Balkan fine wares and preliminary characterisation studies.

Dmytro Gaskevych

5.2.3.1 Graphite in the Production of Pottery in the Ukrainian Para-Neolithic87 The use of graphite as an admixture in a ceramic paste of Neolithic pottery is the subject of current research (cf. Gaskevych 2017). While there are several graphite sources in Ukraine large enough for industrial use (e.g., one of the largest deposits of graphite in Europe in the Hayvoron district of Kirovograd County), there are over 300 small sources of graphite in the Ukrainian Granite massif, so much analytical work remains to be done (for a start, see below, Section 5.2.3.3). There are five main clusters of sites with pottery in which graphite has been mixed with the clay (Fig. 5.32): the Dnister group, the Southern Buh group, the Middle Dnipro group, the Dnipro Rapids group and the Azov Sea group. Of particular interest is the Southern Buh group, which lies closest to Nebelivka and also close to many early Trypillia sites (Hayvoron, Sabatynivka II, Hrebenniukiv Yar, Hrenivka, etc.).

87  The term ‘para-Neolithic’ refers to the fact that the overall rarity of remains of domesticated species – whether plants or animals – in these sites. Kotova (2003, Chapter 1) refers to such sites as ‘Neolithic’.



Special Finds 

 339

Figure 5.32: Map of Ukrainian para-Neolithic sites with graphite-tempered pottery: Symbols: I – present-day industrial source of graphite; II – para-Neolithic site; III – LBK site of Kamiane-Zavallia; IV – Trypillia culture site of Nebelivka; Buh-Dnister culture: 1–Tătărăuca Nouă XV, 2–Soroka I (level 1a), 3–Soroka V, 4–Pechera I, 5–Samchyntsi I, 6–Samchyntsi II, 7–Shymanovske II, 8–Bazkiv Ostriv, 9–Shumyliv-Cherniatka, 10–Hayvoron-Polizhok, 11–Zavallia, 12–Zhakchyk, 13–Melnychna Krucha, 14–Dobrianka 3, 15–Mykolyna Broiaka, 16–Kompaniiska Skelia, 17–Hrushivskyi Ostriv, 18–Semenivka, 19–Ustia Korabelnoi, 20–Puhach 1, 21–Puhach 2, 22–Klepana Balka, 23–Tashlyk 2, 24–Tashlyk 3, 25–Gard, 26–Gard 3, 27–Gard 4, 28–Lidyna Balka, 29–Novorozanivka; Kyiv-Cherkasy culture: 30–Buzky I, 31–Lysychyi Horb, 32–Uspenka 2; Surskyi culture: 33–Strilcha Skelia, 34– Kizlevyi V, 35–Vovchok; Azov-Dnipro culture: 36–Mykilske 2; Surskyi or Azov-Dnipro culture: 37– Kamiana Mohyla 1 (by D. Gaskevych).

340 

 The Finds

The Southern Buh group in turn comprises three site clusters, which differed in the frequency of use of a graphite admixture and its abundance in the ceramic paste. The first cluster, lying within 20km of the main Hayvoron source, includes several sites of the Buh-Dnister culture (BDC) and the LBK site of Kamyane-Zavallia (Kiosak 2017). The quantity of pottery with graphite is relatively small, but the concentration of graphite in the clay is medium to high. A cluster of sites to the North of the source includes several BDC sites on the Southern Buh River and the BDC site of Dobrianka 3 on the Tikych River. Vessels with a graphite admixture are rare. The concentration of graphite in its paste is low to medium. The third cluster to the South of the source includes BDC sites on the River Southern Buh and its left tributaries (Mykolyna Broiaka on the Chornyi Tashlyk River and Novorozanivka on the Inhul River). The amount of pottery with graphite and its typological diversity are the largest in the Southern Buh area. The concentration of graphite varies, but it is often very high. The surface of such vessels glitters like silver or lead. Nowadays, there are 69 published dates obtained from 20 BDC sites. Fifty-two of the dates correspond to a wide range between 6500 and 4700 BC and can be related to the East European para-Neolithic. The overwhelming majority of them were measured on samples of bone and antler from sites of the Southern Buh River region at the Kyiv radiocarbon laboratory in 1998–2005. The study of archaeological contexts from which the samples of these dates come showed that none of them is related to a feature. Most of the dated sites are characterized by the absence of a clear stratigraphic position of the para-Neolithic materials, as well as by possible non-homogeneity of cultural layers (Gaskevych 2014). While the absolute chronology of the BDC graphite pottery is currently uncertain, some typological characteristics and methods of decoration indicate its contemporaneity with the LBK. This is also evidenced by “imports” of LBK pottery from the sites of Bazkiv Ostriv and Gard, as well as one vessel from the site of Bazkiv Ostriv made of paste with graphite and decorated with a very rare example of brown painting of probable Middle Danube origin (Gaskevych 2017a). It seems that a chronological gap between the latest para-Neolithic pottery which contained graphite in its paste and the discovery of graphite-painting at Nebelivka is filled by some of the latest Trypillia A pottery synchronous with Cucuteni A sites to the West. Thus, Zbenovich (1989, pp. 90 and 93) mentions “a few small fragments of fine vessels with graphitized surface” from sites such as Sabatynivka II, Luka-Vrublivetska and Hrenivka. Tovkailo (2005, pp. 34–35) has also written about his discovery of rather numerous pottery with graphite admixture at the same Early Trypillian phase on the multi-layered sites of Puhach 1, Puhach 2, Gard, Gard 3 and Gard 4 in the steppe zone of the Buh River catchment. However, there remains the likelihood that graphite painting in the Trypillia group was related to exchange relations with the East Balkan Copper Age communities.



Special Finds 

 341

Cătălin Lazăr & Theodor Ignat

5.2.3.2 Graphite Painted Ware Analogies in the East Balkans Pottery painted with graphite to make silver designs represented a technological innovation for the Neolithic communities of the Balkans. The first evidence of the use of this new technique dates from the end of the 6th millennium BC in Eastern Macedonia and South-West Bulgaria, where graphite was a component in the slip used to cover the pots (Aslanis 1989; Bailey 2000; Leshtakov 2005). The emergence of graphite-decorated pottery in the 5th millennium BC is closely related to the development of gold and copper metallurgy (Bailey 2000), with diffusion of that decorative technique from South to North to coincide with the supply route of this kind of raw material (Leshtakov 2005). During the East Balkan Early Copper Age (ca. 5000–4600 cal. BC), graphite decoration is found on the vessels of the Maritsa, Sava, Gradeshnica, Polyanitsa and Final Boian communities (Bojadjiev et al. 1993; Bailey 2000; Voinea 2005; Leshtakov 2005). Many authors claim a substantial increase (even described as an explosion) in the usage of graphite-decorated pottery in the second half of the 5th millennium BC (Bojadjiev et al. 1993; Bailey 2000; Voinea 2005; Leshtakov 2005; Dănilă 2014), which has led to the term ‘the East Balkan graphite pottery complex’ (Tasić, N. 1989; Petrescu-Dâmbovița 2001; Voinea 2005) (here Fig. 5.33). However, this assertion is only partially correct, since graphitedecorated pottery, although specific for the Kodzhadermen-Gumelniţa-Karanovo VI (KGK) and Krivodol-Sălcuţa-Bubanj Hum (KSB) communities, represented a reduced percentage of the total ceramic assemblages in other coeval groups. Thus, the analysis performed on a series of well-published ceramic assemblages indicates that graphitedecorated pottery represented a proportion between 2% and 15% 88. By comparison, ceramics decorated with graphite were very common in the KGK VI group (Todorova & Matsanova 2000; Petrova 2007, 2011; Popova, M. 2012; Popova, M. and Kostov 2017). On the other hand, in Romania, we note that there are tell settlements of the KGK VI group with only a few fragments of graphite-painted sherds89 and conversely (graphite-decorated vessels occur mostly on large tells). Sometimes the graphitepainted pots were associated only with a particular building90. For KSB communities, the graphite-decorated pottery is more rare, ranging from 1% and 3% (Pătroi 2011).

88  The percentage of graphite-painted wares – Lîga = 2% (Randsborg and Merkyte 2005); Pietrele = 2.9% (Toderaș et al. 2009); Hârșova = 5% (Voinea 2005); Căscioarele = 5.9% (Voinea 2005); Vinitsa = 15% (Popova, M. 2012); Sultana-Malu Roșu = 14.9%. 89  e.g., Bucșani, Carcaliu, Măriuța, Șeinoiu and Vitănești (Bem 2001; Parnic and Chiriac 2001; Șimon and Parnic 2001; Andreescu et al. 2003; Burens et al. 2010). 90  e.g., Pietrele – Central House, and Sultana-Malu Roșu – House no. 2 (Andreescu and Lazăr 2008; Reingruber 2012).

342 

 The Finds

Figure 5.33: Map of East Balkan graphite painted ware complex: M: Maliq; N: Nebelivka (by B. Gaydarska).

The analysis of graphite decoration motifs from the KGK VI group reveals an evolution through time, with a shift from simple and rectilinear geometric motifs at the beginning to more complex and curvilinear ones in the middle and final stages. Positive-negative ornamentation is characteristic for all three phases, while strictly negative ornamentation only occurs at the end of the second phase. Elements and motifs with chronological value are relatively few but enough to offer some dating possibilities. The round compositions, depicting elements and motifs structured in one or four levels are relatively rare at the beginning of the KGK VI group and are more common towards its end (Voinea 2005; Petrova 2011). It is generally considered that graphite-decorated pottery decreases or even almost disappears in the last phase of KGK VI (Leshtakov 2005; Voinea 2005). Although we are not denying this assumption, the findings at the Sultana-Malu Roșu tell settlement shows that, in the Gumelnița B1 phase, there is an increase in the percentage of the painted pottery in general and graphite painting in particular (23.5% in B1 levels versus 7.1% in A2 levels). Also, the pottery assemblages at Pietrele (Toderaș et al. 2009), and Gumelnița (Dumitrescu and Marinescu-Bîlcu 2001) present numerous examples of graphite-painted vessels in the final levels of those tell settlements. The complete miniature vessel painted with graphite from Nebelivka has a typical shape for Cucuteni-Trypillia ceramics. Even the motifs painted on the vessel body



Special Finds 

 343

are specific to these communities. There are no specific parallels with the graphitedecorated pottery from KGK VI and KSB communities. By contrast, the rim sherd (Fig. 5.34/1) comes from a typical shape for the 5th millennium BC in the Balkans – a dish with internally-thickened rim with many analogies in the settlements of the KGK VI and KSB communities (Fig. 5.34/4)91. Furthermore, the decoration of the internallythickened rim, with oblique parallel lines in graphite, represent a specific decorative motif for KGK VI and KSB communities (Fig. 5.34/2–3)92. There are no known parallels in the KGK VI and KSB repertoire for the decoration of house daub with a graphite wash. Contacts and exchanges between KGK VI and Cucuteni-Trypillia communities are well documented for the 5th millennium BC (Palaguta 2007, Fig. 96). Thus, in many Gumelniţa settlements North of the Danube, discoveries of imported Cucuteni-Trypillia vessels were made93 (Voinea 2005). Also, the mixed group Stoicani-Aldeni (BolgradAldeni II) in the contact area between the KGK VI and Cucuteni-Trypillia groups in Romania, Moldova and Ukraine proves the coexistence of both communities, as supported by the mutual transfer of material culture elements including pot shapes and decorative motifs. All this proves a long history of exchange and interaction between these communities and, in this context, the discovery of the graphitedecorated pottery from Nebelivka should not surprise us. If the complete miniature vessel painted with graphite represents a local product, specific to Cucuteni-Trypillia communities, only the decoration technique was adopted  – a sign of knowledge transfer. However, the rim sherd was part of a dish which was most probably an imported vessel from the latest communities of the KGK VI group. The dating of Nebelivka to 3970–3770 BC is not an impediment, because many KGK communities survived into the early centuries of the 4th millennium BC according to radiocarbon data available for the final levels of tell settlements such as Sultana-Malu Roșu94.

91  e.g., Azmak, Devetashkata Peshtera, Galabovo, Gumelnița, Kozareva Mogila, Krivodol, Lîga, Ostrovul Corbului, Pietrele, Sadievo, Sălcuța, Sultana-Malu Roșu, Tangâru, Varna, Vinitsa and Yunatsite (Todorova & Matsanova 2000; Dumitrescu & Marinescu-Bîlcu 2001; Randsborg & Merkyte 2005; Voinea 2005; Petrova 2007, 2011; Andreescu & Lazăr 2008; Toderaș et al. 2009; Pătroi 2011; Georgieva 2012; Popova 2012; Reingruber 2012; Popova & Kostov 2017). 92  e.g., Gumelniţa – here, Fig. 1, Pietrele – here, Fig. 2, Tangâru – here, Fig. 3, Kozareva Mogila – here, Fig. 4, Devetashkata Peshtera – here, Fig. 5 (Dumitrescu 1925; Voinea 2005; Hansen et al. 2007; Georgieva 2012; Popova & Kostov 2017). 93  e.g., Brăiliţa, Căscioarele, Gumelniţa, Hârşova and Vidra. 94  As shown by the 14C date of 5140 ± 35 BP (4039–3804 cal BC) (Poz-52551) (Lazăr et al. 2018); 5230 ± 50 BP (4174–3961 cal BC) (Poz-52542); and 5250 ± 40 BP (4230–3973 cal BC) (Poz-52550) (Lazăr et al. 2016).

344 

 The Finds

Figure 5.34: Graphite painted analogies for the Nebelivka internally thickened rim dish (1): (2) Pietrele (after Hansen et al. 2007); (3) Tangâru (after Voinea 2005); and (4) tell Gumelniţa (after Dumitrescu 1925) (by T. Ignat & C. Lazăr).



Adrian Boyce, Amanda Dolan & Jason Newton

Special Finds 

 345

5.2.3.3 Sourcing the Nebelivka Graphite There has, as yet, been no study of the effects of firing graphite in pottery on the chemical composition of the graphite itself. Thus, if isotopically heavier carbon in graphite paint was a result of isotopic fractionation during firing, with the lighter 12 C being driven off to a greater extent, this would leave the residue enriched in the heavier isotope, making it difficult to match the object graphite to the graphite source. Further research is necessary on the chemical transformations of graphite in the process of production. Moreover, since there has been no comprehensive study of the graphite sources in Ukraine and the East Balkans, there is currently no database for the characteristics of graphite from different sources. What we know is that there are several major sources of graphite and over 300 small sources in the Ukrainian Granite Shield alone (Gaskevych 2017). For the East Balkans, while only one graphite source is known from Southern Romania (near the town of Targovishte), multiple sources have been identified in Bulgaria (Leshtakov 2004, 2005). As yet, the graphite from these sources has not yet been analysed for comparison with prehistoric graphite decoration. This means that the present study can contribute only a relative differentiation of the chemical constituents of graphite which, although precise, cannot be positively linked to existing sources. The analysis of the graphite in six Nebelivka objects, as well as two samples of graphite from modern graphite mines in the Ukraine and two standards, was undertaken in the East Kilbride Lab. Each sample was analysed by two carbon isotopic techniques – closed tube combustion with dual source isotope ratio mass spectrometry and an on-line elemental analyser in continuous flow with an IRMS. The comparison between the two techniques delivered reproducible results in all but one sample, for which there was a ready explanation (too little material was loaded into the tube for analysis). The results (Table 5.8) provide four valuable preliminary results: (1) there is a good differentiation between the two graphite sources from Kirovograd County; (2) there is a good differentiation between the graphite in the dish with an internally-thickened rim (ITR dish) and the remaining objects from Nebelivka, which form a convincing cluster; (3) the two contrasting values for the graphite from the ITR dish are too low for a derivation from either of the two Kirovograd County sources. However, the ITR dish results may be the product of changes in the graphite during firing (see above); and (4) the Nebelivka cluster of values is not closely matched with either of the two sources.

346 

 The Finds

Table 5.8: Results of the Carbon isotopic analyses of graphite (by A. Boyce). Sample

Closed Tube (δ13CV-PDB ‰)

Continuous MS (δ13CV-PDB ‰)

Petrovo graphite source, Kirovograd County

-34.232

-34.17

Hayvoron graphite source, Kirovograd County -27.768

-28.94

Internally-thickened-rim dish, Test Pit 24/4

-14.07

-20.47

Internally-thickened-rim dish, Test Pit 24/4

-24.249

-24.38

Comments

Closed tube result suspect: sample too small

Graphite-coated platform daub, Test Pit 18/2 -30.9 Graphite-coated platform daub, Test Pit 18/2 -29.7 Miniature Vessel A3290

-30.7

Miniature Vessel A3299

-30.6

Miniature Vessel A3289

-30.2

Miniature Vessel A3295

-31.62

-32.1

STANDARD IAEA-CH7

-16.102

-16.05

STANDARD USGS 24

-16.102

-16.05

In summary, insofar as any broader conclusions can be drawn from this preliminary study, it may be proposed that the differentiation between the ITR dish and the remaining Nebelivka objects matches the conclusions of the stylistic analysis of the vessels – namely that the ITR dish was probably an import from the East Balkans, while the miniature vessels and the decorated structural daub were probably locally produced.

Oliver E. Craig, Harry K. Robson, Matthew von Tersch, Alexandre Lucquin & John Chapman 5.2.3.4 The Interior of the Miniature Vessels An organic residue analysis of 45 vessels, the contents of two vessels and five soil samples was conducted in order to make sense of the use of pottery at Nebelivka. In addition to an intensive study of all 21 miniature vessels from the Mega-structure, we studied a variety of medium-sized bowls of differing shapes – open bowls, carinated bowls and an inverted-rim bowl – as well as rim sherds from vessels of unknown type (n = 15), and nine other miniature vessels from outside the main concentration



Special Finds 

 347

in the Mega-structure, in order to provide a comparative picture of vessel use at the megasite.

5.2.3.4.1 Materials and Methods Small samples (0.5–1.0 g) were removed from the interior walls of the miniature vessels with a scalpel so as to minimise any damage to these complete vessels. Lipids were extracted from three classes of vessel: 30 miniature vessels, nine fine wares and six coarse wares. The latter were taken from seven test pits made in one of the burnt houses95. The soil samples were collected from the 2013 soil pit. Given the small quantity of material, lipid extraction and methylation were conducted in onestep according to an established method (Craig et al. 2013; Papakosta et al. 2015), which is known to maximise recovery, particularly of fatty acids (Correa-Ascencio and Evershed 2014). In short, methanol was added to the powdered samples (pottery sherd powder: 4 ml to 1g; foodcrust: 1 ml to 10–30 mg) and the mixture was sonicated for 15 minutes followed by acidification with concentrated sulphuric acid (800µl and 200µl, respectively). The sealed acidified samples were heated at 70°C for four hours, then cooled to room temperature. Lipids were extracted from centrifuged samples with n-hexane (3 × 2ml) and directly analyzed by GC-FID and GC-MS. GC-MS analysis was undertaken using an Agilent 7890A series chromatograph attached to an Agilent 5975C Inert XL mass-selective detector with a quadrupole mass analyser (Agilent technologies, Cheadle, Cheshire, UK). A splitless injector was used and kept at 300°C. The GC column was inserted into the ion source of the mass spectrometer directly. Helium was used as the carrier gas and inlet/column headpressure was constant. The ionisation energy of the MS was 70eV and spectra were obtained by scanning between m/z 50 and 800. Compounds were separated on a DB-5ms (30m × 0.250mm × 0.25µm; J&W Scientific, Folsom, CA, USA). The temperature was 50oC for 2 min then 10oC/min to 325oC and then held for 15 min. 5.2.3.4.2 Results Analysis of the miniature reveals an unusual distribution of fatty acids that is typical of a degraded plant oil. First, in all cases, the C16:0 fatty acid is much more abundant than C18:0. Also a range of unsaturated fatty acids are preserved, including C16:1, C18:1, C18:2, C20:1 and C22:1 (Fig. 5.35). Of note is the particularly high abundance of C22:1 a fatty acid (Erucic or Brassidic acid). This acid is rarely encountered in organic residues and is a major compound found in Brassicaceae seed oils. The oxidation products of this acid (vicinal dihydroxy acids and dicarboxylic acids) have been found in Ancient

95  Lipid samples were collected from the following Test Pits: 2012/Test Pit 1, 1/1, 1/4, 1/5, 24/3, 25/2, 25/3, 26/2, 26/6 and 33/1.

348 

 The Finds

Figure 5.35: Gas chromatograms of lipid extracts from Nebelivka miniature vessels: (1) Test Pit 2012/3; (2) Mega-structure MP 16; and (3) Mega-structure MP 29 (by O. Craig).



Special Finds 

 349

Egyptian oil lamps (Copley et al. 2005), Islamic Egyptian shells, and Roman vessels (Colombini et al. 2005). In these cases, the oils were most likely used as an illuminant, which explains the presence of degradation products and the absence of the parent unsaturated acids, which would be readily oxidised during burning. The most likely oil for this purpose based on historic documents is radish oil. However, in the case of the Trypillia miniature vessels, the original unsaturated acids are preserved, with little evidence of degradation products, including an absence of dicarboxylic acids, suggesting exceptional preservation and probably ruling out the use of the vessels as lamps. One concern, given the extremely well preserved nature of the lipid profiles, was to rule out contamination from modern vegetable oils or lipids present in the burial environment. To address this, we analysed a range of other ceramic samples from similar contexts and associated as well as background soil samples. The results showed that similar lipid profiles with a high relative abundance of C22:1 were also found on three out of eight coarse ware bowls. The soil samples showed a profile atypically dominated by a range of saturated and unsaturated fatty acids, with C22:1 representing one of the main fatty acids present. A logical conclusion from the analytical work is that the soils and artifacts from Nebelivka are heavily contaminated with vegetable oil, possibly rapeseed oil, given that the fatty acid content of this source contains ca. 50% erucic acid (cis-13docosenoic). Knowledge of the cultivation history of area would be useful to possibly confirm this source. How the lipids have migrated to a depth of ca. 50cm from the modern land surface is not easily explained. Another possibility is contamination of the samples with vegetable oil during or after excavation. Laboratory contamination is less likely, considering that the blank controls analysed synchronously with the samples produced no lipids. Either way, the extensive testing of soils and controls makes it questionable that the organic residues found on the pots are from a endogenous and ancient origin.

5.2.3.5 Summary Miniature vessels formed an important, if minor, part of everyday living at Nebelivka, with dishes and, less often, bowls deposited in burnt houses and pits, while a group of 21 flasks were found together as a special deposit in the Mega-structure. Organic residue analysis of the flasks and other miniature vessels showed that recent contamination was most probably responsible for the concentration of Brassica oils in most of the analysed vessels. Thus, the use of miniature vessels cannot yet be explained, with containers for specialised oils or pigments still a possible use. The Mega-structure group included six vessels with graphite wash and three or four with painted decoration. Additionally, graphite-painting occurred on an internallythickened-rim dish and fragment of platform daub, both from burnt houses. The results of stylistic and characterisation analyses of these vessels and daub matched

350 

 The Finds

well; there was good differentiation of the carbon isotopic profiles of the ITR dish, stylistically identified as an import from the East Balkans, from the remaining Nebelivka objects, which were stylistically identified as local products. However, no match has yet been found between any Nebelivka object and either of the two sources of Ukrainian graphite tested in this analysis.

John Chapman, Marco Nebbia & Bisserka Gaydarska 5.2.4 Other Miniature Vessels

While the Mega-structure group comprises the most compelling find of all the miniature vessels, a further 63 such pots – almost all of them complete – have been found at Nebelivka. A total of 12 vessels was found in the Mega-structure outside of the main cluster, while 22 such pots were found in the Pit in Sondazh 1, together with smaller numbers in houses (six in House A9, and 23 more in nine different houses). A second major concentration of 20 such vessels was found in Quarters G–H in six different houses, within 320m of each other. Three houses were located in the Inner Circuit and three in the Outer Circuit (including eight miniature vessels in Test Pit 23/2), with 75% of vessels found along the long or short walls, 20% in the corners and only one in the middle Zone. The miniature vessels were copies of larger pots of ‘normal’ size. Two principal shapes were found – dishes (in Ryzhov’s system ‘miski’) (Fig. 5.36/1–2) and flasks (‘kubki’) (Fig. 5.36/3–6), with smaller numbers of bowls and plates. The majority of flasks was found in the Mega-structure, while dishes were more evenly distributed in houses and the Pit as well as the Mega-structure. One of the flasks found in House A9 (SF 17) (see https://doi.org/10.5284/1047599/ section 5.1.2.3.1) bore a striking resemblance to one flask in the main Mega-structure concentration (SF F5 MP17) (see https://doi.org/10.5284/1047599/ section 5.1.2.3.2). Dishes comprised the vast majority of miniature pots (75%) in the Quarter H concentration. The form of the flasks in the Mega-structure group was relatively homogenous, with close shape parallels in only three other flasks at Nebelivka – all in Test Pit houses. Rarely, zoomorphic terminals were added to make these vessels even more special – as in the polypod plate from Test Pit 1/5 and a single dish from the Pit in Sondazh 1. An equally distinctive perforation just below the rim would have enabled the suspension of the vessel from the owner’s neck or from a domestic fitting such as a shelf, chair or table (Fig. 5.36/2).



Special Finds 

 351

Figure 5.36: Types of miniature vessel from outside the Mega-structure cluster: dishes – (1) Test Pit 16/1; (2) Test Pit 31/2 (by K. Harding); flasks – (3) Test Pit 1/4; (4) 33/1; (5) Test Pit 1/4; and (6) 1/1 (by B. Gaydarska).

352 

 The Finds

Dmytro Kiosak, Mykola M. Belenko & John Chapman 5.2.5 Chipped Stone

5.2.5.1 Introduction While the lithic assemblages of Trypillia settlements have often been studied in detail (e.g., Bibikov 1953; Korobkova 1987; Sorokin 1991), few reports have as yet been published on assemblages from the largest megasites (for an exception, see Pichkur 2008). In this chapter, an analysis is offered of what should be considered as a small lithic assemblage. In view of the combined results of four seasons of excavation, including one Assembly House, two dwelling houses, three pits and 88 test pits sampling both Assembly and dwelling houses, there was a remarkably low total lithic discard rate, with just under 150 lithic items recovered. It should be noted that many lithic pieces were recovered from the dry-sieving and bucket flotation that was standard practice for all of the excavation units.

5.2.5.2 Raw Materials Description of the raw material is based upon a code developed by Pawlikowski for Balkan sites (Pawlikowski 1992). The first part of code characterizes the country of samples origin (here ‘UA’), the second part is an abbreviation of site where samples were recovered from (Nebelivka is abbreviated to ‘Neb’), the third part is a letter of the material denomination (flint – F, chert – Ch) and finally an order number of a macroscopic group of raw material. The raw materials of the pieces recovered at Nebelivka in the 2012–2014 seasons are as follows: – Ua–neb–f1 – grey and dark grey, plastic, high quality flint with some quite notable white inclusions. It is transparent when thin. It has a chalky primary cortex (abbreviated term – FUA-1) – Ua–neb–f2 – honey-coloured and yellow flint with multiple white inclusions. It is transparent when thin. It has a reddish, smooth “pebble” cortex (abbreviated term – FUA-2). – Ua–neb–f3 – grey, non-transparent flint with white inclusions (abbreviated term – FUA-3) – Ua–neb–ch1 – greyish-red, large-grained siliceous rock (abbreviated term – CUA1). – UA-NEB-RC1 – rock crystal (abbreviated term – RCUA-1) The raw materials of lithics found in the 2009 season were identified using a different system, in which lithics from two known flint sources were identified – a flint quarry at Korobchino, Novomirgorodski region, with a production centre probably located at Korobchino – Rubanii mist (Tsvek & Movchan 2005) – probably the same as type FUA-2. The second source concerns flint similar to that of Korobchino but more probably



Special Finds 

 353

from Volhynia – i.e., type FUA-1. The other pieces cannot be precisely matched to the 2012–2014 raw material types.

5.2.5.3 Technological Modes On the basis of Kiosak’s analysis of the 2014 lithic assemblage, two modes of production sensu Domborócki et al. (2010) can be identified, which fit the whole assemblage: – Mode No. 1 was is probably a result of expedient knapping done by member of a household for satisfying the immediate needs of the household itself. The raw material for mode No. 1 was mostly acquired by individual expeditions to nearby outcrops of moderately suitable flint and chert. The relevant raw material types are FUA-2, FUA-3, CUA-1 and RCUA-1. – Mode No. 2 was based on raw material obtained and pre-prepared elsewhere, sometime in very distant locations, up to 300km away. Raw material type FUA-1 was turned into blades, which acted as universal blanks, bifacial points and partially polished flint axes by flint-knapping experts who may have been parttime specialists.

5.2.5.4 The 2009 Assemblage The total number of lithic items found in the summer 2009 season was ca. 40. The majority of lithics was found in the excavation of House A9 (Figs. 5.39/13–15 & 5.40/17–18), with the remainder (n = 17) deriving from the gridded surface collection of the 15-hectare geophysics grid (see Roe, Chapter 3.3.2; Roe, n.d., Fig. 24) (here Fig. 5.40/12–16). House A9: five pieces are illustrated to represent the good-quality material of this assemblage: a projectile point (Fig. 5.39/13), a double end- and side-scraper (Fig. 5.39/14), a retouched blade which may have been a sickle insert (Fig. 5.39/15), a burin on a backed blade (Fig. 5.40/17) and a distal blade segment converted into a projectile point (Fig. 5.40/18). Intra-site gridded surface material (see Chapter 3.2.2): this small group included a diverse collection – two complex scrapers on flakes: an end- and side-scraper (Fig. 5.40/12) and an end- and double side-scraper (Fig. 5.40/13); a secondary decortification flake with some cortex left remaining on the dorsal face (Fig. 5.40/14); a distal blade segment which may have been a sickle insert (Fig. 5.40/15) and another multiple scraper (Fig. 5.40/16). Two of the multiple scrapers seem to have been transformed from Volhynian flint core rejuvenation flakes (Figs. 5.40/12–13). According to a preliminary classification, most of the material comprises retouched tools and their fragments. Some pieces have clear signs of wear-traces. There are no traces of local production. This suggests that the pieces arrived on the site as ready-made tools/tool blanks, mostly laminar and with some probable flake blanks.

354 

 The Finds

The surface flints were distributed across the entire gridded area, with no particular patterning and signs of only one minor cluster – four pieces in Square K13 (Roe, n.d., Fig. 24). Most lithics were discarded in grid squares also containing burnt daub – viz., the remains of burnt houses.

5.2.5.5 The 2012 Assemblage (Table 5.9) A small sample of 25 pieces was recovered from the excavation of the Mega-structure – the largest Assembly House in Nebelivka – as well as two flints from the 2012 test pits. Since most of the Mega-structure pieces were recorded by Total Station, the possibility of spatial analysis of this sample is utilised. The two most frequent raw materials represented in this small sample come from the FUA-1 and FUA-3 groups – both at one-third of the sample. FUA-1 is a highquality imported flint - well suited to the production of long regular blades. However, a similar number of pieces was discard from the much lower-quality local flint FUA-3. Some 20% of finds are so burnt that it was impossible to define their macroscopic raw material group. One other variety of lithic raw material is represented by a few pieces – the local chert CUA-1. An interesting single piece of greenstone – perhaps a volcanic tuff – shows that the re-sharpening of polished stone tools was practiced within the Mega-structure. The finds deposited in the Mega-structure were used in different ways according to their different raw material groups. No formal cores were discarded in the Megastructure. Chunks were found as both retouched and unretouched forms in flint FUA-3 and chert CUA-1, as were primary decortification flakes. Cortex was found on the surface of many pieces – clearly not an impediment to their use; a good example is the end-scraper on a flake which retained cortex on 80% of the dorsal surface. It is important to note that core rejuvenation flakes, indicating local working, were made on imported flint FUA-1 as well as local chert CUA-1. Waste flakes showing local knapping were found in two flint types – the imported FUA-1 and the local FUA-3 (Fig. 5.37/11, 15) – as well as on burnt flint. Although flakes were occasionally produced and retouched (Fig. 5.37/7), the basic goal of the production sequence consisted of blades which were often snapped into blade segments  – more often proximal than distal; there were no complete blades in the sample. Scrapers were made on both FUA-1 and FUA-2 flint, while two of the three projectile points were made on Volhynian flint. This low proportion of formal tools is not typical of Trypillia BII assemblages (Sorokin 1991). Although no formal tools were made on other raw materials, retouched edges were made on each raw material type, even the lowestquality chert CUA-1. Every piece of imported FUA-1 flint was retouched, while less than half of the FUA-3 flint pieces were retouched. Some flakes and chunks bore fine marginal retouch (Fig. 5.37/14). Blade segments showed semi-abrupt retouch along both edges or a single edge. Retouch is semi-abrupt and high, with long parallel facets (Fig. 5.37/5).



Special Finds 

 355

Figure 5.37: Lithics: Mega-structure: 5–7, 9–10, 14–15, 17; Pit, Sondazh 1: 2–4, 12–13; Test Pits: 1, 8, 11, 16 (by M. Gurova).

356 

 The Finds

Figure 5.38: Lithics: Pit, Sondazh 1: 1–2, 4–5; Test Pits: 3, 7; Fieldwalking: 6; scale 1:1 (by L. Woodard).



Special Finds 

Figure 5.39: Lithics: 2014 season: 1–12 (by D. Kiosak); 2009 season: 13–15 (by L. Woodard).

 357

358 

 The Finds

The three projectile points used different knapping strategies to reach the same end: abrupt retouch followed by parallel invasive retouch over all the cortex-free surface (Fig. 5.37/8), steep retouch on both edges to form the point, with thinning retouch on the ventral proximal side on both edges (Fig. 5.37/9) and fine retouch on both edges with thinning invasive retouch bifacially at the proximal end (Fig. 5.37/10). This suggests three different knappers were producing projectile points at Nebelivka. While bulbs of percussion were removed on projectile points, striking platforms were retained in other pieces, such as the retouched flake (Fig. 5.37/6) as well as retouched and unretouched blade segments. Cortex was used as backing for blade segments and retouched flakes. Notches, both retouched and unretouched, were found on blade segments (Fig. 5.37/5) and retouched flakes (Fig. 5.37/7). Table 5.9: Blanks and Tools, 2012 assemblage (by D. Kiosak).

Type

CUA-1 chert

FUA-1 flint

FUA-2 FUA-3 flint flint

Burnt flint

Total

%

Primary 1 retouched decortification flake

-

-

1 unretouched

2 retouched

4

16

Chunk

1 retouched

-

-

3: 1 retouched, 1 retouched 2 unretouched

5

20

Core rejuvenation flake

1 unretouched -

-

-

-

1

4

CRF with scraper retouch

-

1

-

-

-

1

4

Flake

-

1 retouched

-

2: 1 retouched, 1 unretouched 4 1 unretouched

16

Micro-blade

-

1 retouched

-

-

4

Blade segment

-

1 unretouched

2: 1 retouched, 1 unretouched 4 1 unretouched

16

End-scraper on blade segment

-

1

-

-

-

1

4

End-scraper on flake

-

-

1

-

-

1

4

Projectile point

-

3

-

-

-

3

12

Total

3

8

1

8

5

25

100

-

1



Special Finds 

 359

Figure 5.40: Lithics: 2014 season: 1–11 (by D. Kiosak); intra-site gridded fieldwalking: 12–16; House A9: 17–18 (by L. Woodard).

360 

 The Finds

The most interesting aspect of this small but varied assemblage is the way in which both production modes were deposited in the Mega-structure. The spatial patterning of this modal difference indicates a contrasting distribution of Modes No. 1 and 2 lithics. There were many more Mode No. 1 ‘local’ flints discarded outside the Megastructure, while none was deposited in the central open area and some near the podium and in the Eastern rooms. Conversely, Mode No. 2 ‘imported’ flint was rarely found outside the Mega-structure, with several in the open area and in the Eastern rooms. Some local re-working of the imported flint shows that not all of the preparatory knapping was completed at some remote workshop. The tools made on the flint (projectile points, blades with semi-abrupt retouch, end-scrapers) were part of a Mode No. 2, curated tool-kit, whereas the local chert was deposited as a Mode No. 1, expedient tool-kit. We may perhaps be talking about people with different skill sets (e.g., experienced vs. novice knappers). But it is intriguing that the products of both modes were deposited in such a prestigious building as the largest Assembly House on the megasite.

5.2.5.6 The 2013 Assemblage (Table 5.10) A small sample of 32 pieces was recovered from the excavation of a total of 41 test pits (only seven pieces) and the large Pit in Sondazh 1 (25 pieces, of which six came from flotation). The two most frequent raw materials represented in this small sample come from the FUA-1 and FUA-3 groups, with half of the sample made of the much higher-quality imported FUA-1 and over a quarter of local FUA-3. Just over 10% of the sample was made on the local chert CUA-1, with two pieces burnt and a single piece made of rock crystal RCUA-1. There is a contrast in scale between the finds deposited in the test pits and those in the large pit in Sondazh 1. Although a smaller assemblage (n = 7), the test pit finds comprised one example each of seven different techno-types, with FUA-1 flint twice as common as FUA-3 flint and the only example of rock crystal. However, it was only in the large Pit that CUA-1 chert and burnt flint was deposited, as well as the only examples of primary decortification flakes, débitage, a core rejuvenation flake with secondary scraper retouch, the only perforator and the sole burin. The principal contrast between the test pit houses and the large pit lay in the types discarded: six out of the seven pieces in or near the test pit houses were retouched examples of Mode No. 2 production, while the vast majority of production debris in Mode No. 1 was, as perhaps expected, placed in the large pit. No formal cores were discarded in either type for context. Chunks were found as unretouched forms in flint FUA-3, although with one example of an unretouched notch. Primary decortification flakes were found in not only the local CUA-1 but also, importantly, in the Volhynian flint FUA-1 – clear evidence of on-site working of this



Special Finds 

 361

imported material. This notion is supported by the two groups of FUA-1 débitage, one group with cortex and one group without, found during flotation. Cortex was found on the surface of a quarter of the pieces, not least on all three retouched projectile points (10%, 30% and 40% cortex found) (Fig. 5.38/2, 4). Only one core rejuvenation flake was found in this assemblage, made on local FUA-3 flint and showing secondary use as a scraper. Although flakes were occasionally produced and retouched, the basic goal of the production sequence consisted of blades which were often snapped into blade segments – more often proximal than distal; there were no complete blades in the sample. Most scrapers (four out of five examples) were made on FUA-1 flint, with the other on FUA-3 flint, while projectile points were made on both Volhynian and local flint (Figs. 5.37/1 & 5.38/4). The sole examples of a burin was made on FUA-1 flint, while the only perforator was made on FUA-3 flint (Fig. 5.38/5). The low proportion of formal tools (12 pieces, or 38%) is not typical of Trypillia BII assemblages (Sorokin 1991). Formal tools were made on either the imported FUA-1 flint or the local FUA-3 flint. However, only ten of the 16 pieces (or 63%) of imported FUA-1 flint were retouched, with other pieces showing early stages of the production sequence – only slightly higher than the proportion of FUA-3 flint pieces with retouch (5 out of 9 pieces, or 55%). Some pieces bore traces of a single form of retouch, whether fine marginal retouch on flakes or blade segments (Fig. 5.38/6), semi-abrupt retouch along both edges or a single edge of a blade segment (Fig. 5.37/13) or abrupt retouch on a flake. However, there were signs of intensive use of the flints, especially the imported FUA-1 material. One piece had scraper retouch at one end and notches on the right side, opposite an unretouched back, another scraper had fine retouch on one end as well as invasive retouch on the other edge (Fig. 5.37/16), while others had scraper retouch on one end and both sides in FUA-3 as well as FUA-1 material (Fig. 5.37/2, 3). Other pieces showed combinations of retouch, whether on a single side of a flake, or the fine retouch on two sides of a rock crystal flake and a notch on one side, or the invasive semi-abrupt retouch on one side of a thick flake with thinning retouch on the striking platform and a retouched notch on the other edge (Fig. 5.38/1). The intensity of usage is convincingly shown by the re-use of a FUA-3 core rejuvenation flake as a notched piece and as a scraper (Fig. 5.39/9). Each of the three projectile points used different knapping strategies to reach the same end: abrupt retouch on one side and semi-abrupt retouch on the other (Fig. 5.37/1), fine retouch on one side and a notch on the other (Fig. 5.38/2) and a notch on one side and retouch across the cortex on the other (Fig. 5.38/1). It may be observed that the three projectile points retrieved from the Mega-structure (2012 season) were made in yet three more different ways (e.g., Fig. 5.37/17)! The working end of the perforator made in local FUA-3 flint was fashioned in yet another way – a retouched notch on one side and semi-abrupt retouch on the other. This reinforces the likelihood of several different knappers in action at Nebelivka.

362 

 The Finds

Table 5.10: Blanks and Tools, 2013 assemblage (italics – deposited in Test Pits: all others from Pit, Sondazh 1) (by D. Kiosak). Type

CUA-1 chert

FUA-1 flint

FUA-3 flint

Rock crystal

Burnt flint

Total %

Primary 1 decortification flake

1

-

-

-

2

6

Chunk

2

-

3: 1 retouched, 2 unretouched

-

-

5

16

CRF with scraper retouch

-

-

1

-

-

1

3

Flake

1 unretouched

4: 2 retouched; 2 unretouched

2 unretouched

1 retouched

8

25

Débitage

-

2: 1 with cortex; 1 without cortex

-

-

1

3

9

Blade segment

-

1 retouched

-

-

1 2 unretouched

6

End-scraper on blade segment

-

1

-

-

-

1

3

End-scraper on flake

-

2 (1)

-

-

-

2

6

Side-scraper on blade segment

-

1

-

-

-

1

3

Side- and end-scraper on blade segment

-

1

1

-

-

2

6

Projectile point

-

2 (1)

1

-

-

3

9

Perforator

-

-

1

-

-

1

3

Burin

-

1

-

-

-

1

3

Total

4

16

9

1

2

32



Special Finds 

 363

While bulbs of percussion were thinned or removed on projectile points, striking platforms were retained in other pieces, such as on proximal blade segments (trapezoidal or sub-rectangular) (Fig. 37/2, 16) as well as unretouched (trapezoidal) and retouched (triangular) flakes (Fig. 5.37/12), chunks (semi-circular) and even one primary decortification flake (triangular). The variability in size and shape of surviving striking platforms suggests several hands at work in the knapping of these pieces. This small but varied assemblage is convincing proof that not all of the preparatory knapping of the imported flint was necessarily completed at some remote workshop. Similarly, the collection and local working of CUA-1 chert and FUA-3 flint is attested. The contrast is still found between Modes No. 1 and No. 2, with Mode No. 2 intensively exploited, curated tools made on the high-quality flint (projectile points, blades with semi-abrupt retouch, four kinds of scrapers, burin) and the Mode No. 1 expedient tool-kit based upon lower-quality flint (FUA-3) and chert (CUA-1), although the FUA-3 perforator shows intensive usage of local flint as well. The pieces in the 2013 assemblage showed a greater intensity of use than in the 2012 or 2009 groups, while also indicating variability in knapping techniques that suggest several knappers were at work. This conclusion supports the notion of ‘local’ flint-knappers associated with different houses in the production of this assemblage.

5.2.5.7 The 2014 Assemblage (Table 5.11) The collection consists of 49 chipped stone artefacts, originating from three sets of contexts. The first set is the most homogeneous, comprising pieces found in Pit 2 near the industrial feature. Sixteen items were found in the test pits scattered over the settlement area. This sample is not homogeneous and it cannot be treated as a single entity. In 2014, several contexts were sampled for water-sieving, producing 17 lithic items. One flake was made on a small red “Carpathian pebble”. The FUA-1 Volhynian flint was represented by 53% of all finds by number. Around a quarter of finds were so burnt or patinated that it was impossible to define their macroscopic raw material group. Other varieties of raw material were represented by only a few pieces. Different raw material groups were put to various uses on the megasite. The majority of blade tools – perforators and end-scrapers – were made on FUA-1 flint, while all projectile points were produced from FUA-2 material. Both raw materials represent Mode No. 2 production. By contrast, other raw material groups yielded no formal retouched tools, conforming to Mode No. 1 production. The only core found is a secondary core for flakes. It has a single platform and was made on a large, thick flake. It is not related to the usual chipped stone tool production at Nebelivka. The majority of the tools was made on FUA-1 raw material. Artefacts often bear some areas of primary cortex on their dorsal surface. It is unlikely that they were produced in the course of on-site decortification activities. It is more probable that the partially corticated dorsal surfaces were not treated as an obstacle

364 

 The Finds

for use of a tool. So, corticated blades could be brought as prepared blanks for tools on the site. The debitage group consists of 10 flakelets and a chunk. Two flakelets have a morphology that is indicative of retouching chips: flat oblique butts, a visible ventral lip, a curved profile and “feather” end, and negative scars of the previous retouching rows in the proximal part of the flakelet’s dorsal surface (Fig. 5.40/7). They can come Table 5.11: Blanks and Tools, 2014 assemblage (by D. Kiosak). Description

Pit S1

secondary core on a flake flake

Trenches

Water-Sieving

1

%

1

2.04

3

2

7

14.29

corticated flake

2

1

3

6.12

blade

1

2

3

6.12

chip

8

8

16.33

retouching flakelet

2

2

4.08

1

2.04

1

1

2.04

chunk

2

Total

1

burin spall retouched flake

2

2

4

8.16

retouched blade with semi-abrupt bilateral convergent retouch with convergent semi-abrupt and fine retouch with semi-abrupt retouch on an edge

3 –1 –1 –1

4 –3 –1

7

12.24

splintered piece

1

1

2

4.08

end-scraper, on a retouched blade on a retouched flake

2 –1 –1

2

6.12

1

2.04

4

8.16

2

4.08

1

2.04

49

100.00

tool with burin detachment

1

perforators

3

geometric microliths lunate? rhomboid point

1

bifacial arrowhead

1

Total

16

1 1 –1

–1

16

17



Special Finds 

 365

from faceting of core’s platform; however, both morphology and context are highly suggestive of their role in the knapping sequence as facconage debris. Flakes outnumber non-retouched blades. But the inclusion of retouched items means an equal proportion of blades and flakes, with most tools made on blades. There are no complete blades in the collection. Some parts of blades can be analyzed from a technological viewpoint. Mostly they are fragments of not too wide and not too thin blades, with a pattern of dorsal negatives “212”, which is characteristic for serial detachment of blades from flat and wide fronts of cores (Léa 2003). The striking platforms (butts) were flat and elliptical, with an area of 9–12 mm2. Sometimes the detachment of a blade was prepared by reduction of overhang, done by means of abrasion “on the flaking surface”. Dorsal negative ridges are wavy and uneven. Angles of detachment were recorded twice – 78o and 85o. Consequently, the morphology of laminar products does not correspond to the use of the lever technique. We would hypothesize that a more widespread, less specialised technique was used for blade production, such as the punch technique or the soft organic percussion technique. Defined types comprise almost a half of the collection, at 49%. This high percentage of retouched items is typical for Trypillia Phase BII–C sites (Sorokin 1991). Some flakes bear fine marginal retouch (Figs. 5.39/6–7 & 5.40/2, 5). One flake has flat alternative retouch (Fig. 5.39/3). Retouched blades are represented exclusively by medial parts of blades with semi-abrupt retouch along a single edge or both edges. Retouch is semi-abrupt and high, with long parallel facets (Figs. 5.39/8 & 5.40/3–4). Several perforators are made on blades in a similar technique – by high, parallel, semi-abrupt retouch on both edges that converges to the end of a blank, becoming flat at the very tip. There are sectors with fine or flat ventral retouch, single flat detachments in the area near the pointed tip. Although the point is often broken, it is quite evident that the retouched blade was pointed by convergent semi-abrupt retouch (Figs. 5.39/10–12 & 5.40/10). A blade was retouched along both edges with semi-abrupt retouch while its tip was rounded by flat retouch into a “scraper” front with an acute section (Fig. 5.40/9). There are small ventral facets opposite this working edge. They probably are macro-traces of use. This tool is close morphologically to the group of perforators which Telegin proposed to call “knives with scraper-like termination” (Telegin 1976, p. 27). Medial parts of blades with semi-abrupt retouch could be fragments of perforators or “knives”. End-scrapers are not really typical. An end-scraper on a flake has an angular front. The proximal end of an end-scraper on a retouched blade was destroyed by a facconage (façonnage) flake (Fig. 5.39/6). Projectiles are made exclusively from honey-coloured or yellowish flint FUA-2. A small (2.5 × 1.5cm) bifacial arrowhead bears a deep notch on its base and has isosceles straight edges (Fig. 5.39/2). It is produced by parallel flat retouch, which covered most of both dorsal and ventral surfaces. A rhomboid point was made on a blade by semi-abrupt and abrupt retouch. The distal truncation is acute (40o) while the proximal one is convex (Fig. 5.39/1). A “lunate”-like tool is a small (1.4 × 0.5cm)

366 

 The Finds

fragment of a blade with abrupt retouch, cutting an arched back in the blank (Fig. 5.40/6). Although it cannot be stated categorically that it is a fragment of a lunate, this fragment should belong to a tool of a similar type – for example to a point with an arched back, a traditional form of tool more typical of the Early Trypillia period. Other tools are represented by two splintered pieces. One of them is made on a retouched blade (Fig. 5.40/11). One edge is retouched finely, while the other has semiabrupt retouch. There are several negatives of flat detachments on the dorsal surface, and a large flat negative on the ventral surface. There is also a corticated flake with large flat ventral negative of burin spall in the collection (Fig. 5.40/8).

5.2.5.8 Discussion We suggest that the observed typological variability can be explained mainly through the use of two or three reduction sequences. The first one is related to the highquality “imported” Volhynian flint FUA-1. It arrived at the settlement in shape of blanks, maybe already retouched, that served as a universal tool, often re-sharpened or re-shaped. Similar items can have a multitude of functions defined by use-wear analysis (Korobkova 1987; Sorokin 1991). Re-sharpening is proved by retrieval of retouching flakelets from the water-sieving. This fact alone demonstrates the effectiveness of water-sieving in Trypillian site excavations. Such tools (perforators, blades with semi-abrupt retouch, end-scrapers, “knives”) were part of a Mode No. 2, curated tool-kit. It is interesting that a majority of flint items in the Test Pits whose source could be identified were imports (Fig. 5.47). The second context is related to the flint FUA-2 and is represented by projectiles of various shapes. In a local anomaly, this ‘local’ flint conforms to Mode No. 2 production. While the bifacial arrowhead with a convex base has numerous analogies on Late Trypillia sites, the rhomboid point made in local FUA-2 flint is rather related to the Early Trypillia technology of knapped stone. It is an obvious outsider in the Trypillia BII context. Its presence in the pit near the industrial feature needs an explanation. One suggestion is that the Phase A tradition of rhomboid arrowheads was retained locally as a connection to the ancestral roots of the Nebelivka settlement. A second idea is that visitors to the megasite brought this geometric form of arrowhead from an area in which this ancestral form was deliberately retained. A third, more straightforward idea is that the projectile point was found in the site area and re-utilised in a later arrow, in turn discarded in a BII megasite. Whichever reason is most plausible, the rhomboid projectile point touches on the interesting question of relations with ancestral Trypillia traditions. A similar link with the past is raised by the discard of an FUA-2 ‘lunate’. The third context illustrates knapping procedures, represented first of all by a core and a thick flake made of a coarse-grained siliceous rock. This raw material was unusual in the tool-kit of a Late Trypillia site. Similar finds are usually interpreted as traces of children learning knapping skills (Shea 2006).



Special Finds 

 367

The low discard of lithics, the high percentage of retouched tools  – although reduced at Nebelivka  – and their probable connection with short (viz., two- or three-stage) reduction sequences are quite characteristic for Late Trypillia sites. A similar model of chipped stone production and utilization is widespread in the Balkan Neolithic and Chalcolithic (Mateva 2011; Hansen et al. 2012; Sirakov 2002; Manolakakis 2005, 2008). It is probably connected with an established system of logistical supply of cores and blanks to Neolithic residents in order to sustain a settled way of life (Zimmermann 1995). According to Bibikov (1970, pp. 3–6), there were two types of production in the Trypillia group – domestic and communal. Communal production was exemplified by pottery, flint-collection and – processing and metal production. The raw material was collected and the objects were produced for exchange, with the flint-knappers receiving some objects in return. If Bibikov’s exchange model is correct, there should be central sources and centres of lithic production to serve a settlement as large as Nebelivka. However, no such centres have been as yet identified in the Nebelivka micro-region. According to Tsvek, flint-extraction pits are considered to indicate flint sources. Such working pits have been found in the Novomirgorod region, near knapping sites. One such BII production site – Rubany Mist (Tsvek & Movchan 2005; Tsvek et al. 2012)  – may have been contemporary with Nebelivka. Given the relatively close distance of Nebelivka to the Korobchino quarry, and the overlap in date, it is possible that this quarry served the Nebelivka site. This notion is supported by the discovery of utilised Korobchino flint nodules at Nebelivka in the 2009 season. However, it is also possible that the Nebelivka people were using the lithic production centres near Volodymyrivka because this was their traditional flint source. The results of these investigations raise questions about the procurement, distribution and further processing of lithics in the Eneolithic period. In the Early Trypillia period (phases A and BI), large lithic assemblages have been recovered, indicating large-scale deposition of mostly local flint with a few imported flint pieces from the Dobrogea and Volhynia, as at Bernishivka (Zbenovich 1989; cf. D. Chernovol’s recent, still unpublished investigations which have yielded thousands of lithics: Shydlovsky & Slesariev 2015). It might be supposed that megasite populations would have needed large quantities of lithics for different purposes, whether scrapers or denticulate sickles and that, consequently, such procurement and processing could have been very important in the region. However, in phase BII, there was a dramatic, and as yet unexplained, decline in the discard practices of Trypillia settlers, with greatly reduced on-site lithic deposition. Despite their size, the megasites also conformed to this new model of minimal lithic deposition. An explanation for this major change in lithic discard is urgently required. In the Late Trypillia phase, long-distance importing of flint was very important, even if not in great quantities. At the CI site of Dobrovodi (Pichkur 2005, p. 117),

368 

 The Finds

there were lithic imports of both light and dark brown varieties of Volhynian flint. Similarly, in the CI/phase 3 at Majdanetske, Volhynian flint imports were also found. According to Petrun (2001–02, p. 139), this flint derived from Turonian and Senonian (Cenomanian) epochs and bears a visual similarity to flint from outcrops in the Ternopil and Rivne regions. Characteristic examples of ‘local’ flint from the Synukha basin were not found in the lithic raw materials at Majdanetske. In the CI horizon at Taljanki, tools were made from imported Volhynian flint (Pichkur 2008, p. 158), although there are some ‘local’ tools made from the so-called Bug flint. Thus, following the marked reduction in lithic discard, there is a change in the relationship of local and imported flint at the BII – CI transition, with imported flint outnumbering local flint at Taljanki, Dobrovodi and Majdanetske. This could show different traditions of lithic procurement, exchange networks and the types of tools used in the CI Tomashevka group. According to Skakun (2004, pp. 74–5 & Fig. 14, 2012; Skakun et al. 2014), Trypillia society imported ready-made tools rather than raw materials of Volhynian flint. She draws parallels between lithics from the Bodaki flint-knapping workshop in the Ternopil region and those deposited at Majdanetske and Taljanki. The character of the Nebelivka lithics made of Volhynian flint suggests that this process of importing ready-made tools had already begun in the BII stage. This conclusion has implications for the dating of Volhynian flint exploitation and the use of the Bodaki workshop. In this way, this investigation of the materials from Nebelivka shed light on the local use of flint, the importing of flint and its processing.

5.2.5.9 Conclusions Trypillian lithic collections shed light on the complex social organization of flintworking. On the level of empirical data, almost any Trypillian lithic collection is composed of two production modes: Mode No. 1, with sets of objects made from “local” flint, often of poor quality, knapped in basic ways, without complex preparations with extensive use of waste products as blanks for retouched tools; and Mode No. 2, with sets of objects produced from good-quality, often imported flint, with a notable input of expertise and skill, oriented towards blade production and their utilization as blanks for a wide variety of tools. Recent finds indicate that Mode No. 2 lithic production appeared as early as Trypillia A III (Precucuteni A3), first of all in the Dniester valley at rich flint outcrops (Kiosak 2016). Volhynian flint was utilized in such a way from the Trypillia BII stage onwards. A small lithic assemblage, consisting of fewer than 150 pieces, has been retrieved from four seasons of excavations at the Nebelivka megasite. Approximately half of the lithics identified to raw material were imported, Volhynian flint (FUA-1). The megasite shares the dual system of lithic production in Modes No. 1 and 2 but with some exceptions. For example, the use of ‘local’ FUA-2 flint for projectile points combines a Mode No. 1 material with a Mode No. 2 technology. If the two modes reflect work by different flint-knappers, such an occurrence would betoken the sharing of knowledge



Special Finds 

 369

between knappers. It is also noteworthy that five types – projectile points, three scraper types and perforators – were made from two raw materials, showing a crossover between Mode No. 1 and Mode No. 2 production, perhaps through imitation of FUA-1 products in local materials. A high proportion of the Volhynian flint (FUA-1) which arrived at Nebelivka came in the form of blade blanks, most often to be converted into blade segments for multiple uses (scrapers, sickle inserts, projectile points and burins). A proportion of these blanks still contained cortex, especially the scrapers. Some of these pieces were repaired on site, leaving FUA-1 debitage. However, there is some evidence for local knapping of FUA-1 flint in the form of three primary decortification flakes and a core rejuvenation flake. Local Mode No. 1 production is more widely attested, whether by primary or secondary decortification flakes, waste flakes, retouched and unretouched chunks and the sole core found at Nebelivka – a CUA-1 secondary flake core on chert. The overall structure of the Nebelivka assemblage indicates a lower proportion of formal tools than in other Late Trypillia lithic groups. While this could indicate a stronger focus on production at Nebelivka, the use of flotation and sieving in excavation recovery may also lead to a higher proportion of production debris than in trench hand-recovery. Finally, Bibikov’s exchange model is not the only possible explanation for such supra-communal production efforts. Some authors link early forms of craft specialization to elite political development (Brumfiel & Earle 1987). However, others believe that the Neolithic-Eneolithic craftsmen acted within a complex kinship-based systems providing access to specialised products to any member of large, kinshiprelated groups of people (Kienlin 2012). The lithic assemblage at Nebelivka shows the typical combination of Mode No. 1 and Mode No. 2 production, with cross-overs between the two modes. Since the greater part of the megasite has not been excavated, one can never rule out the existence of specialist lithic production zones. However, the current evidence suggests import of a small number of exotic lithics from Volhynia and local, small-scale production on a household level. The picture is more or less the same for most Trypillian settlements from the BI period onwards. Flint-knapping is conducted in very limited parts of sites and, quite often, almost all flint items enter the settlement as ready to use and are only rejuvenated and re-sharpened on site. Moreover, the working areas seem to be located away from the dwellings – in particular the activity of working fresh hides. The scrapers required in this practice constitute up to 60% of retouched tools in “rich” lithic assemblages of Trypillian sites and are less numerous in “poor” lithic complexes recovered from most Phase BI–CII sites. One author (John Chapman) concludes that this scale of production is hardly consistent with the ‘maximalist’ position of large-scale, long-term permanent occupation of the megasite. Instead, the lithic results support the use of Nebelivka on a lower-intensity, shorter, perhaps seasonal occupation. The alternative is that the scale of production contradicts the concept of megasites as “large villages”, since at least some of the typical site activities, including flint-knapping, took place outside the foci of megasite social space.

370 

 The Finds

John Chapman

5.2.6 Ground Stone The study of the Nebelivka ground and polished stonework was undertaken by Mr. Tom Bergquist as an undergraduate dissertation in the University of Durham (Bergquist n.d.).

Zsuzsanna Tóth & Alice Choyke 5.2.7 Worked Bone

5.2.7.1 Introduction There are only a very few pieces of worked osseous material available for study from the megasite at Nebelivka. In fact, there are only eight objects made from bone, antler and tooth coming from four seasons of excavations that took place at this site. Due to the small size of the assemblage, it is difficult to draw any far-reaching conclusions. Despite the small number of pieces, however, rich information can be achieved by careful and detailed examination of the individual ornaments and tools. Optical microscopic study of the manufacturing and use wear was carried out on the objects to exploit the maximum information from them. The site itself was largely explored through various kinds of remote sensing techniques and these artefacts come from excavation work on selected house features. Still, despite the limited nature of the excavation, the numbers of bone tools seem very low (cf. almost 150 lithic items), given that the soil was subjected to careful screening. Prehistoric sites of this period and even later in the Bronze Age usually have a greater density of worked osseous materials on them connected to the detritus of everyday life and activities. 5.2.7.2 Description Three of the objects belong to the group of tooth pendants. One of the objects is a bead/pendant made from a red deer (Cervus elaphus L. 1758) mandibular incisor (Fig. 5.41/1). This pendant is barely modified. The tip of the root is smoothed down, probably with abrasion, although marks of this activity are almost completely obliterated by use-wear. The root is also covered with intense scraping made by a flint tool (Fig. 5.41/1b). The first step in manufacturing was probably to create a nice and even-looking surface. Approximately 5 mm from the end, there is a notch created by sawing which could have served for attachment (Fig. 5.41/1c). Otherwise, there are none of the other modifications typical for tooth pendant/beads. The dental crown was left in its natural state, although intense use formed moderate rounding and polish can be observed on it. Similar use-wear, rounding and polish could be observed at the tip of the root as well. The walls of the notch are rounded. The degree of use wear and rounding of edges suggests that this object was used for an extended period spanning years of use.



Special Finds 

 371

Figure 5.41: Worked bone tools: (1) tooth bead-pendant, Pit, S. 1 SF 47, with close-ups (b) and (c); (2) tooth bead-pendant, Mega-structure TsT 1461, Context 64; (3) bone imitation of tooth pendant, Mega-structure TsT 1827, Context 142 (by K. Harding based on photographs by Zs. Tóth).

Two other tooth pendants were found on the surface. One of them was probably made from an actual red deer canine (Fig. 5.41/2). The root, typically, was perforated and served to fix or suspend the ornament. It seems to have been damaged in ancient times since another perforation was made at the other end. This secondary hole looks quite worn as well, indicating that the pendant was used over a rather long period. The other pendant bead (Fig. 5.41/3) appears to be a bone imitation of a red deer canine bead/pendant closely following the form and size of actual red deer canines (for example see Spatz 1999). This bead was made from cortical bone taken near an epiphysis of a horse-cattle size animal long bone diaphysis and later perforated at the narrower end, although the perforation itself lies outside the central long axis

372 

 The Finds

of the object. The positioning of the hole may be connected with the way the object was meant to hang. The cortical bone around the hole was first scraped away in a disk shape before the hole was actually drilled. Although the NISP for red deer bone was very low, the incisor and true canine bead still could easily have been taken from locally killed game. The fourth object is a slender point made of fishbone (Fig. 5.42/1). Unfortunately, the object is broken, with only the active end and middle section remaining intact. It has a delicate point shaped with longitudinal flint scraping. The surface is entirely covered with intense manufacturing (scraping) marks. Only the rounding and polish of use obscures them slightly (Fig. 5.42/1b). There are multiple impact fractures at the tip (Fig. 5.42/1c). Thus, it appears the tool continued to be used even after the tip became slightly damaged. A fragile tool made of a relatively easily chipped raw material, this object must have been used in a task not requiring much force, probably connected to processing soft materials such as thin hide and/or textiles but, in any case, a task that required only light force or pressure but a fine sharp point.

Figure 5.42: Worked bone tools: (1) fishbone point, Pit, S. 1, SF 48, with close-ups in (b) and (c); (2) broken bone awl with copper staining, Mega-structure, Grid Square D10 (by K. Harding based on photographs by Zs. Tóth).



Special Finds 

 373

The fifth artefact is a tiny ad hoc point made of small ungulate long bone shaft splinter (Fig. 5.43/1). The bone is barely modified at all although some scraping can be observed towards the active tip. However, besides this, the craftsperson took advantage of the bone splinter’s natural shape with only the most necessary modification at the active end. This is the only tool which displays clear traces of taphonomic changes in the form of root etching at several spots over the whole surface. All the edges and especially the tip show traces of use wear including rounding and polish (Fig. 5.43/1b). The sixth object is part of a broken awl made from a long splinter of small ungulate (caprine) long bone diaphysis, most probably a metapodial (Fig. 5.42/2). Unfortunately, it is broken and neither the active end (tip) nor the base remain, preventing identification of the exact typological group. However, it very much looks like the remains of one of the finely made small metapodial points so often found on Late Neolithic and Chalcolithic sites. The final shape was produced with scraping with a chipped stone tool. The traces are still visible on the tool’s surface, along with a green staining which may be from contact with a copper object.

Figure 5.43: Worked bone tools: (1) ad hoc bone point, Test Pit 19/2, Context 3, SF 6, with close-up in (b); (2) red deer antler hoe model, Pit, S. 1 SF 71; (3) possible bone tool, Mega-structure Grid Square E6 (by K. Harding based on photographs by Zs. Tóth).

374 

 The Finds

Only one worked osseous object from the site was made from red deer antler. The tool, made from a tine tip, is a small bevel-ended ‘pick’ with a small hafting hole at one end (Fig. 5.43/2). The sharp lateral edges and even polish of the bevelling at the active end strongly suggest that this bevelling is artificial and not a natural shape caused by the buck rubbing his antlers against trees as the velvet is being shed. The tine was cut half way up its length. The compact tissue at the wider end was thinned on both sides of the tool before boring the round hole from both sides. Meanwhile, the tine tip was lightly worked into a narrow, delicate bevel-form. Otherwise, the ridges of the natural surface were left intact. The active end reveals signs of use in the form of rounding and polish. Thus, the tool was surely used in a light activity such as bark-removal or wood-working (where the wood interior is burned first before removal) which did not result in serious, marked damage(s) to the active end. There is no apparent evidence of any curation work on the active end. Finally, there is an object which may or may not be an actual tool (Fig. 5.43/3). If indeed it is a tool, then it was produced on a longitudinal long bone fragment from a pig-caprine – size animal. It appears to have a bevelled active end but the surface is much eroded, making it impossible to determine whether it was worked or simply used.

5.2.7.3 The Manufacturing Continuum These few objects cover the whole range of planning and intensity of modification along the manufacturing continuum (Choyke 1997). There are three planned utensils (Class I): a lightly worked pointed tool with carefully chosen but unusual raw material (fish bone), a typical utensil of the Late Neolithic toolkit made from an intensely modified small ungulate metapodial bone and an antler tine hafted ‘pick’ that has been moderately worked. A fourth specimen represents an object made from a piece of refuse bone that broke accidently into a useful shape and was used with a minimum of alteration. If this latter object is indeed a tool, then it would fall into the ad hoc, barely modified Class II end of the manufacturing continuum. This final possible object may be a badly eroded bevel-ended, ad hoc tool made from the longitudinal splinter of a pig-caprine size animal. However, due to the poor surface preservation it is impossible to say whether the shape was produced naturally or was the result of manufacturing activity. The three pendants made from a red deer incisor, an actual red deer canine and the bone imitation of a red deer canine, as well as the small ungulate awl, all belong at the Class I end of the manufacturing continuum because of the distinctive choice of the skeletal element mostly derived from species not as readily available as local domesticates to local craftspeople. The manufacturing chain for the real tooth pendant beads is not very complicated: the shaping was done by relative simple techniques requiring a few steps. The imitation red deer bead required more knowhow to produce. The small fishbone point falls rather in the Class I–II category in the



Special Finds 

 375

middle of the manufacturing continuum. We suggest that the rules for raw material choice for tools was not very strict and the natural slenderness of the fishbone made it a good candidate for producing a slender, sharp point. Since the object is broken, only the final steps of shaping are apparent, that is, the scraping marks on the surface, but it seems the manufacturing sequence was not particularly complicated. The third object is a real Class II tool, a small ad hoc point. Neither its raw material or skeletal element was strictly chosen and the manufacturing process is simple. If the final bevel-ended tool based on a long bone diaphysis fragment is really a tool, then it too would fall in the Class II (ad hoc) end of the manufacturing continuum.

5.2.7.4 Evaluation Personal adornments, like pendants or beads forming part of complex, composite ornaments such as necklaces, hair pieces or girdles occur regularly in many different periods. Objects made of animal teeth, possibly from animals that possessed a special iconographic or symbolic meaning for the wearer or receiver of the ornament, are very common in site assemblages throughout prehistory and even in later time periods as well. Teeth are objects of display for both the animals they come from as well as the humans who take up their use later. Wearing these objects may be strongly connected to gender as with perforated and shaped boar tusks or perforated red deer canines (both tusks and red deer canines derive from male animals). These teeth may have come from species that have special symbolic significance to the audience for whom they are displayed. Thus, there is also an amuletic, apotropaic aspect to their use, so that the bead comes to represent the way the animal species eats and protects itself (Choyke 2010) – as well as the power(s) ascribed to the living animal in particular cultural contexts. More typical red deer tooth beads from the final Neolithic and Early Chalcolithic in this region are perforated red deer canines, mostly coming from male deer (canine teeth in does are smaller if they are even present), found collected in hoards or as parts of necklaces and scattered finds on sites. The final Neolithic in Hungary and Germany (Spatz 1999) is also represented by imitations of red deer canines that either take the form of nested hour-glass beads that fit into each other cross-wise leaving only the rounded ends exposed. These imitation necklaces and ornaments often contain a single genuine red deer canine. Given the differential wear on the beads, it has been suggested that such composite, fractal ornaments were produced at the time of burial and given to females as grave goods at the site of Polgár 6 (Choyke 2001; Siklósi 2013) and may well represent a kind of materialized social enchainment between the living members of a family group and the dead. Further to the West, in Germany, there are burials in cemeteries, more or less contemporary with Polgár 6, with imitation red deer canine beads, closely following the biological form of the teeth but these are given to men in burials (Spatz 1999). In any case, it seems clear that red deer had a special ideological meaning connected at the very least to gender identity for various

376 

 The Finds

groups of people living across a broad swathe of Central and Eastern Europe. In this sense, their relative abundance in the worked sample compared to the presence of hunted animals in the faunal assemblage is at least worthy of note. Pendants made of perforated or notched animal teeth can be found in the territory of the Trypillia Culture as well. Unlike this specimen, the other known examples are usually made either from predators, mainly canid (wolf/dog/fox) canines such as the pieces from the cemetery at Vinogradnyj (Rassamakin 2004, Abb. 52:18, Abb. 72:10) or Ol’šanka Kurgan 3 (Rassamakin 2004, Abb. 109:5). Red deer canine pendants came to light in Tudorovo 1 Kurgan 1 (Dergačev 1991, Taf. 59), while a perforated elk (Alces alces L. 1758) tooth was used at Luka Vrublevetskaya (Zbenović 1996, Taf. 10:17). Imitations made of animal bones are common as well, such as at the cemetery of Vinogradnyj (Rassamakin 2004, Abb. 52: 16–17) or Giurgiuleşti (Rassamakin 2004, Abb. 72:11). The incisor from Nebelivka is not perforated nor is it a canine, but it was fixed with the help of a notch. Nevertheless, it clearly belongs to the category of pendant/beads within personal ornaments made of animal teeth. The choice of deer canine as a bead is surprising considering the scarcity of red deer in the environs of the site. However, since these beads were used over a very long period, probably involving multiple generations, it may be equally likely they were produced in an area with plenty of red deer and the raw material choice and form was related to other, currently unknowable parts of the Trypillia period belief system. It seems reasonable to suggest that the tooth was chosen because the animal it came from – red deer – was a generally iconic, special animal in the region and for this period. The three small perforators made from small ruminant metapodial, bone shaft and a fishbone are extremely common prehistoric types. Awls made from a variety of skeletal elements from many different species usually represent the most numerous tool type at most prehistoric sites although fish bone is not such a common raw material for producing bone tools of any kind, including points. Slender points made from a variety of osseous raw materials can be found both at settlements, such as the megasite at Taljanki (Kruts et al. 2008, Fig. 22: 6–8) or Bernashivka (Zbenović 1996, Taf. 10: 1–3, Taf. 11: 1–2, 9), Luka Vrublevetskaya (Zbenović 1996, Taf. 11: 6–7), or Okopy (Zbenović 1996, Taf. 10:5, Taf. 11: 10) as well as in burials in cemeteries such as Nikol’skoe, Kurgan 1, Grave 7; Staronižesteblivskaja, Grave 30; Aleksandrija, Grave 22 and Vinogradnyj, Grave 44 (Rassamakin 2004, p. 90). While Late Neolithic-Early Chalcolithic hafted antler tine picks tend to be cut from the beam and used intact, such small pick-like tools are much less common.

5.2.7.5 Conclusions The seven (possibly eight) bone, tooth and antler objects found during excavations at the megasite of Nebelivka are too few to draw any far-reaching conclusions although their unexpected paucity raises interesting ideas about the way these dwellings were lived in and abandoned. Such small sample sizes can result in extremely biased



Special Finds 

 377

results. Nevertheless, the unusual circumstance that three out of seven (eight) objects were red deer tooth beads or their imitation certainly raises questions about activities at the settlement and/or the people dwelling in the structures. The objects themselves span the entire manufacturing continuum: the relatively well worn and intensely used red deer incisor bead which can be considered a Class I planned object; the fishbone point where the raw material choice seems unplanned but where the manufacturing activity is relatively intense that occupies the Class I–II part of the continuum; and the point made from randomly produced scrap bone and barely modified at the Class II end of the manufacturing spectrum. Use wear studies suggest that that the red deer incisor bead was used over a prolonged period of several generations and is connected to other red deer tooth beads, although these beads overwhelmingly come from the canines of bucks. The fish bone point was used in a delicate activity on a contact material that was soft and did not offer much resistance. It was used even after small chips were accidently removed from the tip during use. The small ad hoc point was affected by root etching, precluding the possibility of identifying use wear. The objects themselves do not stand out from the kind of objects generally found at other Trypillia sites in the region. It is hoped that an accumulation of detailed studies will lead to a more comprehensive picture of modified osseous materials at megasites of this period. Based on the faunal evidence, there seems to be some focus on raw material from red deer including antler and teeth. This focus was not reflected in the animal bone material and one cannot discount the possibility that these objects (the beads and the pick) were brought to the site as finished objects.

John Chapman, Marco Nebbia & Bisserka Gaydarska 5.2.8 Other Special Finds

The following Special Finds constitute a diverse collection of rare forms, in all cases but one made of fired clay. This amounts to a total of 29 special finds, with a majority found in the Mega-structure and very few in House A9 (DOI: https://doi. org/10.5284/1047599 Section 5). - Zoomorphic vessel Three fine ware legs of (?) zoomorphic vessels were found in the Pit in Sondazh 1. Blurred painted motifs were found on SF 16. - Vessel with zoomorphic terminal Two such vessels were found in the Mega-structure - both with triangular animal heads attached to wall sherds (MS F19 & F20). The heads resemble the heads of zoomorphic figurines found on the megasite (DOI: https://doi.org/10.5284/1047599 Section 5).

378 

 The Finds

- Polypod vessel A single largely complete dish with four worn feet was found in Test Pit 1/5. The interior of the dish had badly worn fine painted lines in a criss-cross pattern, while the slightly thickened rim had oblique parallel lines, as was found in graphite painting on one of the miniature vessels from the Mega-structure. - Strainer A single basal fragment of a strainer was found in the Pit in Sondazh 1; six perforations, ca. 5mm in diameter, had been made in the base. - House model fragments Eight small fragments are known – four from each of the Mega-structure and the Pit in Sondazh 1. While small, each fragment is identifiable as to its part of the house – a trait typical for the fragmentation practices of significant objects (cf. Chapman 2000 for Adriatic salt pots). The most striking fragment is a painted roof and upper wall fragment (Pit SF 5727), with the parallel painted lines on the flat roof evoking roof timbers (Fig. 5.44/1). The closest parallel comes from the Voroshilivka roof fragment (No. 17 in Shatilo’s (2005) thorough catalogue) and also dating to Phase BII. Other structural parts include three angles of the ground floor and wall, one angle of the upper floor and wall, two wall fragments and the corner of two walls. The concentration of house model fragments in an Assembly House and a pit shows the importance of these places for deposits embodying a key aspect of the Trypillia Big Other – the house itself. - Sledge model One fired clay fragment showing the angle of a vertical and a horizontal edge has a narrower cross-section than the house model fragments. It may have been a fragmentary sledge model (cf. the many examples found at Majdanetske and Taljanki: Müller & Pollock 2016, p. 284). - Incised sign The only sign found at Nebelivka was incised on the exterior of a single greybrown medium coarse ware body sherd. There are no close parallels for this incised sign in the corpus of painted Trypillia signs (Tkachuk 2005). - Gaming board Two fragments of a fired clay plate decorated with incised concentric circles and with semi-perforations in two circles were deposited in the Mega-structure (Fig. 5.44/2–3). The concentric incisions are reminiscent of the fine incised decoration of platform daub. The small diameter of the semi-perforations (up to 10 mm.) make them too small for the insertion of any of the Nebelivka tokens but other sticks or plant stems could have been inserted (cf. Shatilo 2015). A general parallel for such ‘gaming boards’ was found at Taljanki (Shatilo 2015, Fig. 2).



Special Finds 

 379

Figure 5.44: Special Finds: (1) fragment of house model, Pit, Sondazh 1, SF 5727; (2)–(3) two fragments of (?) fired clay gaming board, unstratified; (4) fired clay ring, House A9; and (5) gold hair ornament, SF 1181, Mega-structure (by K. Harding).

380 

 The Finds

- Spindle-whorl Two fired clay spindle-whorls were found at Nebelivka – one in House A9 and one in the Mega-structure. The A9 item was larger (4cm diameter, with perforation diameter of 1.1cm), of coarser clay and chipped, while the whorl from the Megastructure was in mint condition, of finer clay and smaller (2.5cm diameter, with perforation diameter of 0.4cm). Spindle-whorls were rarely discarded in the houses at Majdanetske and Taljanki. - Clay ring A simple fired clay ring, of outer diameter 3.2cm and inner diameter 1.35cm was found in House A9 (Fig. 5.44/4). The red clay surface was smoothed but undecorated. - Perforated fired clay object A fragment of a crudely shaped, oval clay object with a partially surviving perforation was found in Test Pit 24/3. Its function remains unknown. - Clay ball Three examples of fired clay balls have been found at Nebelivka – two in Test Pits and one in the Mega-structure. All objects share a similar coarse fabric with the majority of tokens (see above, pp. 336–337), with one ball being spherical (diameter 3.5cm) and two being oval and with similar dimensions (2.8 × 2.5cm; 3 × 2.6cm). The spherical ball was burnt on one face; none had any sign of decoration. It is possible that they could have been used as sling bullets or for throwing at stray animals to bring them back into line. - Unidentifiable fired clay lump Three unidentifiable fired clay lumps were found at Nebelivka – one in the Megastructure, one in House A9 and one in Test Pit 25/3. Minor shaping, such as a groove or a pinched area, give these irregular items more shape. However, their function remains unknown. - Gold spiral ornament The first gold ornament known from the Trypillia group was deposited in one of the East rooms of the Mega-structure. The item is 15mm in length and 4.5mm in diameter and consists of a thin gold rod which has been heated to allow the wrapping of the rod around a solid (? timber) core six times (cf. Leusch et al. 2014, Fig. 9a) to produce a bead or a small hair ornament (Fig. 5.44/5). Leusch et al. (2014, p. 183) consider two possible uses for such a spiral bead – as an ornament in its own right or as a semi-finished product in the preparation of gold disc beads. The find is so precious that no analysis of the gold has yet taken place; the possibilities include the Caucasus Mountains, the alluvial gold sources of Eastern Bulgaria or the rich gold sources of the Munţii Metaliferici of Central Transylvania. Whichever was the source of the Nebelivka gold ornament, the implication is some form of long-distance exchange into the heartland of a loessbased, gold-free landscape.



Special Finds 

 381

The closest parallels to such a gold ornament derives from the Varna I cemetery, which dates 500–600 years earlier than Nebelivka (e.g., the gold spiral ornament from Grave 97: Leusch et al. 2014, Figs. 3b & 9b). Similar dates apply to the sheet copper spiral ornament from the Chapli cemetery (Chernyh 2010, Ris 3/19), as well as the copper spiral bead with six twists from the Cucuteni site of Traian (Mareş 2002, Pl. 59/14). This ornament was the only example of a undoubted prestige good found so far at Nebelivka. Its deposition in the Mega-structure underlines the significance of that building to the entire megasite.

John Chapman 5.2.9 Summary

These Special Finds can shed light on five closely interlinked aspects of the Trypillia world: the all-encompassing Big Other, inter-regional and regional exchange networks, local production practices, deposition and, fifthly, personhood. The pattern of material engagement described above is a clear indication of the way that those dwelling in, or visiting, the megasite resisted, for the most part, the deposition of polished stone and metal objects in favour of the key element of the Trypillian Big Other – clay. The discard of two polished stone axe fragments (DOI: https://doi.org/10.5284/1047599 Section 5) and one gold spiral ornament, in addition to the green (? copper) staining of a single bone awl, indicates the limitations of stone and metal deposition at the megasite. The exception to the discard of stone objects concerns the chipped and ground stone fragments which were part and parcel of daily lifeways. This is not to claim that Trypillia farmers did not fell trees with stone axes or that Trypillia leaders did not wear costume enriched with copper or gold ornaments – only that such items were as a rule excluded from settlement deposition. The positive side of this coin is that a great variety of fired clay objects discarded on site materialised the Big Other, as they did at countless other Trypillia sites, big or small. The raw materials and objects flowing across the networks which sustained the Big Other indicate inter-regional exchange and use of local resources for Nebelivka. While none of the Nebelivka metal or pigments has been located to a single source, the gold spiral ornament, the copper that produced the green stain on the bone awl and the manganese pigment used on painted pottery must have come from an exotic source several hundred kms from the megasite, while it is very probable that the graphite-painted dish was a direct import from the East Balkans. The same is true of the high-quality Volhynian flint whose sources lay more than 200km to the West. Local resources included lower-quality flint, sandstone for grindstones and possibly graphite for the decoration of miniature vessels and house platforms. However, with the exception of salt exchange (pp. 471–472), most of this exchange and procurement was low-bulk and probably episodic and/or seasonal, integrated with the Nebelivka

382 

 The Finds

calendar of ceremonies and house-burning. The persons engaged most directly in such exchanges – especially if going on long-distance voyages (Helms 1988) – would have become different kinds of persons, respected but also feared for their contact with the ‘Other’. The small scale of lithic and worked bone discard and prestige goods deposition was a source of surprise to the Nebelivka team, who were expecting major depositional practices consonant with a long-term, socially-differentiated urban site with a large population. Instead, painted fine wares, whether as complete vessels or in fragments, comprised by far the greatest bulk of discard in all excavation units (see Chapter 5.1). This scale of production of chipped and ground stone tools is consonant with household production – even the Volhynian flint, which did not always come in ready-for-use forms. There is also a wide range of quality of products, expressed for worked bone as the ‘manufacturing continuum’ (Choyke 1997) and also found in chipped stone (Modes 1 and 2). While high-quality objects produced by skilled practitioners (? specialists) were found (e.g., the gold spiral ornament, the miniature vessels, the imitation in bone of a red deer canine bead/ pendant), most of the clay figurines and tokens showed simple shaping skills, no decoration and low firing temperatures. These findings, again, suggest household production and local use until the objects were worn, followed by fragmentation and local deposition of parts and wholes. There are two clear patterns of deposition of Special Finds at Nebelivka: a concentration of most categories of special find in the Mega-structure (the obvious example being the group of 21 miniature vessels) (Fig. 5.48 lower), with dispersed distribution in the Test Pits (Fig. 5.45) and a generalised distribution within the Megastructure with a notable lack of finds concentrations (Fig. 5.48 upper). The former underlines the significance of the Mega-structure to the entire megasite. There was also greater diversity of Special Finds in the Southern part of the megasite (Quarters L and M) (Fig. 5.45). The latter suggests episodic deposition of figurines, tokens (and gaming boards), Mode 1 and 2 lithics and ground stone. Those participating in the exchanges and ceremonies would have extended and differentiated their personhood over and above those remaining on the social periphery. The Special Finds distribution in Neighbourhoods showed great variability, with only one case of two houses with identical Special Finds – both in Neighbourhood 13. There were no Neighbourhoods where figurines were found in more than a single house – perhaps a hint of ritual variability within Neighbourhoods? It was also in Neighbourhood 13 that we found the only instance of a house with deposition of both exotic and local flint.



Figure 5.45: Special Finds distribution, megasite (by M. Nebbia).

Special Finds 

 383

384 

 The Finds

Figure 5.46: Distribution of figurines, megasite (by M. Nebbia).



Figure 5.47: Distribution of lithics, megasite (by M. Nebbia).

Special Finds 

 385

386 

 The Finds

Figure 5.48: Special Finds distribution: (upper) figurines and tokens; (lower) lithics and Other finds, Mega-structure (by M. Nebbia).



Special Finds 

 387

There are many aspects of the Nebelivka Special Finds that shed light on personhood – the way that persons developed an (in)dividual identity through their lives. Engagement in exotic voyages and participation in ceremonial life have already been mentioned. Two other ways of producing personhood were through the development of productive skills and the representation of the human form. Many household members would have possessed skills sufficient to make tools from local flint or sandstone, spin yarn or make ad hoc bone tools (Chapman & Gaydarska 2011). However, the skill of carving a red deer canine into a fine ring/pendant would have marked out the craftsperson as someone special, whose identity may have become as extended in time as the highly curated ring/pendant itself – on the basis of experimental work estimated to be two or three generations. Equally, the use of an ‘archaic’ style for a rhomboid point and a lunate suggests a form of lithic curation linking the persons dwelling at Nebelivka with persons and their skills from times past. Contemporary individualised production and perhaps use are suggested by the six flint projectile points, each made in a different way into varied, visually distinctive forms and perhaps with different hunting results (cf. Wiessner 1983). The only representation of the human form at Nebelivka – the fired clay figurines and statuettes  – emphasised that individual and dividual aspects of personhood were often in tension. While image fragmentation focussed on the dividual aspect of personhood, with heads frequently missing, the rare use of statuette heads to define ‘realistic’ persons reminded users of the individual side of personhood. In parallel to such representation, the visual aspects of costume were important for personhood. Whoever wore the Nebelivka gold spiral ring must have had a special identity, even if only for annual ceremonies, and perhaps different from the identity of the wearer of the fired clay ring. The persons dwelling in, or visiting, Nebelivka created a visually differentiated, colourful world of houses, pottery, figurines and Special Finds, all of which related to the Trypillia Big Other. To the extent that persons, households, Neighbourhoods, Quarters and the entire megasite were engaged with Special Finds, these objects contributed to the formation of new and continuing identities that, in turn, helped to shape their own social world. This section is a testimony to the diversity of the objects produced in, or for, Nebelivka.

388 

 The Finds

David Orton, James Nottingham, Giselle Rainsford-Betts, Kim Hosking & Andrew Millard 5.3 Animal Bones This report is dedicated to the memory of Charles Schwartz, who contributed a significant portion of the primary research on which it is based, and who very sadly died a few weeks before it was completed.

5.3.1 Introduction Over the four joint Kyiv-Durham field seasons at Nebelivka (2009, 2012–2014) a large quantity of animal bones was recovered from many of the excavated features. These were recorded and analysed variously by Olena Sekerskaya (Archaeological Museum, Odessa), Charles Schwartz (independent consultant, Los Angeles), Zsuzsanna Tóth (Eötvös Loránd University, Budapest), Carrie Armstrong and Louisa Gidney (both Durham University), but only the first-named's work has resulted in a published report (Sekerskaya 2017). Accordingly, the present chapter aims to draw together the work produced by these analysts, along with a last few bones analysed in York by the authors, in order to give an overview of the animal bone assemblages from the site. An additional comment by Andrew Millard refers to the isotopic dietary information for the animals as derived from the AMS dating procedure. Apart from broad questions of subsistence, set in the wider context of Trypillia economy, particular attention will be paid to differences in faunal composition between different features and feature types, and especially to understanding the nature of bone deposition in the Megastructure. Necessarily given the history of study, we shall also assess potential interobserver bias. Some key questions for this report are as follows: – What was the relative contribution of hunted vs. herded animals at Nebelivka, and how does this fit into wider trends noted for the Trypillia period? – Are there any detectable differences in animal use (or at least bone deposition) between areas of the site and/or between different context types, e.g., houses and their associated pits? – What was the nature of bone deposition in the Mega-structure?

5.3.2 Areas and Assemblages Table 5.12 shows the total counts of bones included in this chapter, by site area and analyst.



Animal Bones 

 389

Table 5.12: Counts of diagnostic and non-diagnostic bones by excavation area/feature and analyst (by D. Orton). Area

Analyst(s)

Diagnostic

Non-diagnostic

Sondazh 1 – Pit

Louisa Gidney and Carrie Armstrong

25

77

Olena Sekerskaya

345

489

 

Present authors (wet sieved)

 

6

Ditches

Louisa Gidney and Carrie Armstrong

1

2

Present authors (wet sieved)

3

134

House A9

Olena Sekerskaya

179

277

House B17

Louisa Gidney and Carrie Armstrong

2

3

Olena Sekerskaya

201

283

Louisa Gidney and Carrie Armstrong

4

2

Olena Sekerskaya

416

519

 

Present authors (wet sieved)

 

42

House B18

Louisa Gidney and Carrie Armstrong

1

3

Olena Sekerskaya

14

5

House B18 – Pit

Olena Sekerskaya

15

34

 

Present authors (wet sieved)

5

96

Kiln

Louisa Gidney and Carrie Armstrong

4

9

Present authors (wet sieved)

5

72

Kiln – Pit

Louisa Gidney and Carrie Armstrong

5

1

 

Present authors (wet sieved)

9

43

Barrow

Olena Sekerskaya

2

7

House B17 – Pit

Present authors (wet sieved)

4

Mega-structure

Charles Schwartz & Zsuzsanna Tóth

220

1570

 

Olena Sekerskaya

132

216

Test pits

Louisa Gidney and Carrie Armstrong

106

551

Olena Sekerskaya

11

12

Present authors (wet sieved)

16

167

 

1721

4624

Grand Total

390 

 The Finds

- House A9 Excavated in 2009 by the Ukrainian team, with the UK team conducting dry sieving, flotation and environmental sampling. All of the bones from this burnt structure were studied and published by Olena Sekerskaya (2017). - Mega-structure This building was excavated over an eight-week period in 2012, the first and the last weeks by the Ukrainian side alone and weeks 2–7 by the joint Kyiv-Durham team. Bones from the joint excavations were recorded by Charles Schwartz and Zsuzsanna Tóth; those from the final week subsequently by Olga Sekerskaya. A 10-litre earth sample from each Context was subject to dry-sieving and flotation. This material forms the largest single bone assemblage from Nebelivka, at 2,138 fragments, though not the largest identified sample (n=352). - Houses B17 and B18 These two burnt houses were excavated in 2013 by the Kyiv team, along with adjacent pits interpreted as being associated with the respective houses. House B17 and its pit, in particular, produced significant bone assemblages of 203 and 430 diagnostic specimens respectively (489 and 983 fragments in total). Only small parts of House B18 and the adjacent pit were excavated. This sample was recovered by hand-excavation with no dry-sieving or flotation. - Sondazh 1 – Pit This was a similar feature to the B17 and B18 pits, but in this case the pit was excavated by the Durham team and the associated house was not excavated. A sample of 20 litres from each level (2013 excavations) and each context (2014 excavations) was subject to dry-sieving and flotation. - Test pits A total of 82 small test pits was excavated in 2013 and 2014 across numerous sectors of the site, primarily in order to obtain samples for AMS radiocarbon dating. The majority of bones from these test pits were studied by Gidney and Armstrong with a view to identifying suitable radiocarbon samples, hence only diagnostic specimens were consistently recorded, although total bone counts were also available. The present authors revisited this material and confirmed that, with the exception of some small mammal and fish bones that have been updated accordingly in the database, the unrecorded specimens are indeed nondiagnostic. For simplicity, all this material is nonetheless listed under Gidney in Table 5.12. A 10% sample of each test pit deposit was also subject to dry-sieving and flotation. - Barrow This sample derived from the 2013 cleaning of a robber trench within a barrow located in the Northern part of the megasite. None of the deposits was subject to dry-sieving or flotation.



Animal Bones 

 391

- Kiln/Cooking feature This feature and its associated pit were excavated in 2014 by the Kyiv team, and have been described in detail by Burdo & Videiko (2016). The nature of the feature remains contentious, with the Kyiv team interpreting it as a kiln and the Durham team as a communal cooking feature. A small number of bones was recovered by hand and recorded by Louisa Gidney. - Ditches Three trenches (Sondazh 2, 4 and 10) were dug to investigate the perimeter ditch in various parts of the site. Very few bones were recorded from these, by Louisa Gidney, and they are subsumed together here and included in Table 5.12 for completeness. - Flotation residues (various areas) A final batch of bones from flotation during the 2014 season was recorded in York in 2017 by the present authors. These derive mainly from the test pits, the kiln/ cooking feature, and its associated pit.

5.3.3 Methodology Given that the underlying data was produced by multiple analysts with differing methodologies, it was necessary to adopt what may be termed a ‘lowest common denominator’ approach, limiting the detail that can be presented here. Pre-existing data from ten separate data sheets were combined into a single master-database, decoded as far as possible  – using keys provided by the analysts where available plus-cross referencing with Sekerskaya (2017) – and the terminology standardised. Inevitably, there were details that were either incommensurate or could not be decoded.

5.3.3.1 Diagnostic and Non-Diagnostic Specimens Ideally, a consistent approach to defining ‘diagnostic’ specimens would be applied, based on objective criteria regarding elements and portions present, in order to minimise identification biases between taxa and between analysts (see e.g., Russell & Martin 2005). Since this is not possible when working with existing data, our primary criterion for a specimen to be ‘diagnostic’ is simply whether or not it was identified to a particular taxon (usually genus-level or below). It was evident, however, that major differences in identification protocol between analysts needed to be addressed: most notably, Olena Sekerskaya routinely identified ribs and vertebrae to genus or species level – not the practice of the other analysts. To avoid introducing a substantial interanalyst bias, it was thus necessary to treat all of these elements (excluding atlas, axis, and sacrum) as ‘non-diagnostic’. This should be borne in mind when reviewing Table

392 

 The Finds

5.12: many of the specimens listed as ‘non-diagnostic’ were originally identified to taxon by Olena Sekerskaya, hence the discrepancy between the numbers presented here and those in the original report (Sekerskaya 2017, p. 18). Despite this adjustment, it is apparent from Table 5.12 that the ratio of diagnostic to non-diagnostic bones varies considerably across the overall assemblage. This could relate to any combination of three explanations: (a) differing levels of fragmentation, (b) differing approaches to identification and particularly recording of small, indeterminate fragments, or (c) differential recovery linked to excavation strategies. To explore this, Table 5.13 compares identification rates by excavation team, area, and faunal analyst, for all subdivisions with greater than 300 fragments recorded. The results suggest that inter-analyst differences are the primary factor in identification rates at Nebelivka, with Sekerskaya consistently recording ca. 40% of specimens to taxon (excluding ribs and vertebrae) regardless of excavation team or site area, while the other analysts reported a significantly higher proportion of nondiagnostic specimens – albeit without a comparison from the solely Kyiv-excavated areas.

Table 5.13: Identification rates by excavation area and analyst. Numbers represent proportion of bones identified to taxon, out of 1. NB this excludes wet-sieved material (by D. Orton). Team

Area/Feature

Charles Schwartz & Zsuzsanna Tóth

Louisa Gidney

Olena Sekerskaya

Joint

House A9

 

 

0.39

 

Mega-structure (wks 2–7)

0.13

 

 

Durham

Sondazh 1 – Pit

 

 

0.41

 

Test pits

 

0.16

 

Kyiv

House B17

0.41

House B17 – Pit

0.45

  NB this excludes wet-sieved material

Mega-Structure (wks 1 & 8)

 

 

0.38



Animal Bones 

 393

5.3.3.2 Quantification The sole quantification method used here is Number of Identified Specimens (NISP, aka fragment count), since reliable calculations were not possible for Minimum Numbers of Individuals (MNI) and related measures when working from previously recorded data. MNI is, in our view, of limited use at a settlement scale in any case, but its impossibility here is regrettable from the point of view of comparisons with other sites in the region (see below). The inability to calculate minimum numbers for specific elements, meanwhile, severely limits the potential to examine anatomical representation.

5.3.3.3 Burning and Taphonomic Modification Different analysts used slightly different terminology for burning and other surface modifications, making it difficult to compare results. In order to track frequencies of burning across certain areas of the site, we therefore simply collapsed the range of descriptions to ‘burnt’ or ‘unburnt’.

5.3.3.4 Measurements Few metrical data are available for the Nebelivka fauna. Only Gidney provided metrics along with standard von den Driesch (1976) codes, but this amounted to just fifteen specimens, the majority of which are cattle. Sekerskaya (2017) reports a further small number of cattle measurements, with generic descriptions, that can probably be equated with von den Driesch codes. No key was available for the coding system used by Schwartz and Tóth. Log Size Index (LSI) values were calculated for cattle following Meadow (1981), using the Ullerslev Cow (Degerbøl & Fredskild 1970) as the standard.

5.3.3.5 Age Data While Gidney provided information on proximal and distal fusion explicitly, Sekerskaya and Schwartz and Tóth only gave relative age classes (‘juvenile’, ‘subadult’, etc.). Coupled with portions present, it was often possible to reconstruct fusion data from these relative ages, but this was not consistent and hence the results are not deemed systematic enough to be reliable. Only Gidney provided details of dental wear stages, resulting in samples of fewer than ten mandibles per species that could be assigned to mandibular age stages following Payne (1973). Accordingly, no analysis of age-at-death is conducted here, although the results reported by Sekerskaya (2017) can be consulted.

394 

 The Finds

5.3.4 Taxonomic Frequencies Table 5.14 shows taxonomic frequencies by excavation area, with the overall assemblage summarised in Figure 5.49/upper. The vast majority of specimens identified were large mammals, dominated by the main Neolithic domesticates (cattle, pig, sheep, goat, and dog – collectively making up 93.8% of identified fragments) and especially cattle (62.0% alone). The range of presumably hunted taxa is rather small, including red and roe deer, aurochs, equids, hare, turtle but no wild carnivores with the possible exception of wolf and apparently no wild pigs (see below). Fish and bird bones were very rarely recovered, even in wet-sieved material, while the small number of rodent specimens recorded by the present authors includes at least one vole, hamster (Cricetus cricetus) and ground squirrel (Spermophilus sp.) – the latter identified with the aid of images in L. Popova (2016). Hamsters and ground squirrels are burrowing taxa that are likely to be intrusive, though not necessarily significantly post-dating occupation.

5.3.4.1 Identification Issues In general, the shared use of Linnaean taxonomy makes comparison of different analysts’ taxonomic identifications straightforward. Potential complications arise, however, with (a) sheep versus goat identification; (b) the potential presence of wild and domestic forms of cattle (Bos), pigs (Sus), and Canis; and (c) equids. Sheep and goat identifications are reported as given in the first instance, but are combined into a general ‘Ovis/Capra’ category for subsequent analysis due to widely differing identification rates and likely asymmetry in the identifiability of sheep and of goat depending on criteria used (Zeder & Lapham 2010; see also reasoning in Orton, D. et al. 2016, p. 5). The separation of wild and domestic specimens is also problematic. While Schwartz and Tóth explicitly recorded all Bos and Sus specimens as belonging to the domestic form (with the sole exception of a vertebra recorded only as ‘cattle’, which is in any case treated here as non-diagnostic–see above), Gidney and Armstrong recorded specimens simply as ‘Bos’ or ‘Sus’. Sekerskaya, meanwhile, recorded the vast majority of cattle, and all pigs, as ‘domestic’, with only a handful of aurochs separated out. No analyst definitively identified any pig specimen as wild. Cattle from Nebelivka were probably overwhelmingly domestic, given the Sekerskaya and Schwartz and Tóth results and the generally low contribution of aurochs at Trypillia sites (see e.g., Zhuravlev 2008). Further limited support for this is provided by the small number of available LSI measurements for cattle (Fig. 5.49/ lower): the majority of measured specimens are clearly smaller than the standard – a small Danish female aurochs – with two only slightly larger. Some additional



Animal Bones 

 395

Figure 5.49: Upper: overall taxonomic distribution of faunal remains (NISP); lower: Distribution of Log Standard Index (LSI) values for measurements on cattle bones (by D. Orton).

1  

1  

220

Canis familiaris 6

 

3

 

 

 

Bos taurus

Equus sp.

Equus caballus

Equus hydruntinus

Bos primigenius

Cervidae

Cervus elaphus 5

Dog

Horse family

Horse

Wild ass

Wild cattle

Deer family

Red deer

 

1

1

 

 

2

144

5

3

 

2

143

28

4

2

 

5

305

50

 

 

1

13

2

 

 

1

12

 

 

 

4

 

 

 

8

26

 

1

2

1

1

150

88

 

 

 

 

2

63

10

 

 

29

Cattle

23

3

4

73

63

1

Sus scrofa domesticus

3

Pig

1

5

2

30

1

Capra hircus

8

24

Goat

1

12

18

7

Ovis aries

1

Sheep

MegaTest structure pits

39

Kiln Pit

Ovis/Capra

House House House House House Kiln A9 B17 B17 - Pit B18 B18 - Pit

Sheep/ Goat

Ditches

Sondazh 1 - Pit

Latin name

English name

Table 5.14: Taxonomic frequencies by excavation area (NISP) (by D. Orton).

43

1

5

1

5

1

21

1063

262

6

61

195

Total NISP

2.5

0.1

0.3

0.1

0.3

0.1

1.2

62.0

15.3

0.4

3.6

11.4

%NISP

396   The Finds

9

1

2

1

 

 

369

Lepus

Emys orbicularis

Aves

Cricetus cricetus

Spermophilus cf. suslicus

Arvicolinae

Rodentia (large)  

 

Capreolus capreolus

Rodentia (small)

 

 

Roe deer

Hare

Turtle

Birds

Hamster

Ground squirrel

Vole family

Large rodent

Small rodent

Fish

 

 

 

Sondazh 1 - Pit

Latin name

English name

3

 

 

Ditches

179

 

 

202

 

 

1

420

 

 

15

 

 

20

 

 

9

 

2

 

House House House House House Kiln A9 B17 B17 - Pit B18 B18 - Pit

Table 5.14: Taxonomic frequencies by excavation area (NISP) (by D. Orton).

Continued

14

2

1

2

 

Kiln Pit

352

 

 

2

1

131

1

 

11

2

6

2

1

 

1

2

MegaTest structure pits

1714

3

3

13

2

6

2

2

2

4

13

Total NISP

100.0

0.2

0.2

0.8

0.1

0.4

0.1

0.1

0.1

0.2

0.8

%NISP

 Animal Bones   397

398 

 The Finds

aurochsen may be hidden within the indeterminate ‘Bos’ specimens, but these are unlikely to change the overall picture considerably. The same case can be made for pigs, which were overwhelmingly recorded as definitively domestic, but here the absence of definitive wild specimens is more surprising since they are present in almost all comparable assemblages, often in greater numbers than their domestic counterparts. The majority of Canis specimens was recorded by Sekerskaya, who identified them as domestic dog, Canis familiaris. Schwartz and Tóth, and Gidney, recorded one Canis specimen each, without indicating domestication status. In the absence of any positive identifications of wolf, it seems reasonable to assume that all, or nearly all, canid specimens represent domestic dog. Finally, equids were recorded variously as horse (Equus caballus), wild ass (Equus hydruntinus), and indeterminate equid. Given the recent demonstration that even experienced researchers cannot reliably distinguish these taxa, even using teeth (Twiss et al. 2017), they are all treated as Equus sp. from here on. The domestication status of horses at Nebelivka is unclear (Sekerskaya 2017, p. 21).

5.3.4.2 Regional Comparisons Zooarchaeological data from Ukrainian Trypillia sites have previously been collated by Zbenović (1996), Kruts (2002), Videiko & Burdo (2004), and Zhuravlev (2008). Of these, only the last-named lists raw data in NISP form that be used for direct comparison with the Nebelivka results. Additional data for Romanian Cucuteni sites come from Bejenaru and colleagues (Bejenaru & Stanc 2011, 2012; Bejenaru et al. 2011; Oleniuc & Bejenaru 2011). Figure 5.50/upper shows the sites used for comparison here, after applying a minimum NISP cut-off of 300, while Figure 5.51 summarises these data by broad chronological phases: (1–2) Trypillia A and BI (3–4) Trypillia BII–CI (i.e. the period of the megasites), and (5–6) Trypillia CII, including sites listed as “CI–CII”. A trend from hunted taxa to the major domesticates over the course of the Trypillia period has been previously noted (Kruts 2002; Kirleis & Dal Corso 2016) but is not clearly apparent in this dataset, with variation in the wild:domestic ratio appearing to be geographical as much as temporal (NB equids have been separated here due to ambiguities over their domestication status: see also Sekerskaya 2017, pp. 20–21). Hunting generally seems to play a bigger role in Western areas than in the East, with domesticates particularly dominant in the core megasite region around Uman, where Nebelivka’s ca. 94% domestic fauna is consistent with neighbouring sites.96 A low relative contribution of wild meat might be expected at such large sites a priori, simply

96 Dal Corso et al. (2019, p. 7) report 98% domestic species, with ten times the bone weight of cattle in comparison to either caprines or pigs, in a sample of 1,334 bone fragments from Majdanetske.



Animal Bones 

 399

Figure 5.50: Upper: Trypillia and Cucuteni sites with raw NISP available and used for comparison here. 1. Berezivka, 2. Bilshivtsi, 3. Cucuteni, 4. Draguşeni, 5. Feteşti, 6. Ghelăieşti, 7. Maidanetske, 8. Hoiseşti, 9. Ignatenkova Gora, 10. Konovka, 11. Kosenivka, 12. Liveni, 13. Grebenyukov Yar, 14. Mitoc, 15. Poduri-Dealul Ghindaru, 16. Santana de Mureş B, 17. Sarata-Monteoru, 18. Sverdlikove, 19. Taljanki, 20. Târpeşti, 21. Truşeşti, 22. Valea Lupului, 23. Vasylivka, 24. Velyka Slobidka, 25. Vesely Kut, 26. Zhvanets-Shovb, 27. Zhvanets, 28. Nebelivka; lower: main taxa identified at Nebelivka by excavation area (%NISP) (by D. Orton).

400 

 The Finds

Figure 5.51: Contributions of wild versus domestic taxa (1, 3 & 5) and breakdown of the main domesticates (2, 4 & 6) for Trypillia and Cucuteni sites in the Early (1–2), Middle (3–4), and Late (5–6) Phases as defined in the text (by D. Orton).



Animal Bones 

 401

due to limits on the amount of game that could be caught within a practical distance from the site. That said, this broad regional pattern appears to hold both before the megasites and in the period of their decline, albeit based on limited data. Breaking the domestic component down into the three main taxa (cattle, pigs, and sheep/goat), regional trends again appear at least as prominent as temporal ones. Cattle are the dominant taxon at most sites in all periods. The exceptions in the early period are two sites in North-Eastern Romania, Hoiseşti and Truşeşti, which have – respectively – a dominance of pig and a fairly even split in the domestic fauna (although this would still mean that cattle provided the most meat in all cases). In the middle period, the four South-Westernmost Cucuteni sites in the dataset stand out as having similar numbers of sheep and of cattle specimens, while sites across the rest of the region have fairly uniform cattle-dominated assemblages. The few assemblages in the dataset from the late period are uniformly cattle-dominated, with no sign of the shift towards sheep and goats observed elsewhere (e.g., Kruts 2002) and argued to represent the development of more extensive pastoralism (Diachenko 2016a). To summarise, the overall results from Nebelivka are consistent with the broader regional picture in terms of both the wild:domestic ratios and of the balance of domestic taxa.

5.3.4.3 Intra-Site Comparisons Figure 5.50/lower plots the main taxa identified at Nebelivka by excavation area, excluding wet-sieved material and the smallest samples. Some variation is seen across the site, with some areas – notably House A9 – being heavily dominated by cattle, and others – particularly the Mega-structure and some of the pits – having much higher percentages of the smaller domesticates. The assemblages from House B17 and its associated pit are very similar, perhaps confirming the association between these two features, although the route by which bones became included in the burnt house remains is unclear – particularly in the absence of data on burning – so this conclusion remains tentative. The samples from House B18 and its pit are really too small to be reliable, but are included for comparison. Even given the small sample sizes, the absence of pigs (which make up 15.3% of the overall site assemblage) in both features is notable. Given the variable identification rates noted above, it is necessary to assess possible inter-analyst and inter-excavation-team differences. To this end, Table 5.15 compares (a) two assemblages from the Mega-structure, excavated by different teams and recorded by different analysts; and (b) Sondazh 1 and the House B17 pit – two similar features excavated by different teams but both recorded by Sekerskaya. Figure 5.52/1 presents the same comparisons visually, for the main taxa only. Unfortunately, there was no case in which excavation team could be held constant across large samples recorded by different analysts.

402 

 The Finds

Table 5.15: Comparison of taxonomic frequencies between areas and analysts (NISP) (by D. Orton). (a).

Mega-structure

Team

Joint

Analyst(s) Ovis/Capra

(b).

Sondazh 1 – Pit

House B17 – Pit

Kyiv

Durham

Kyiv

Schwartz & Tóth

Sekerskaya

Sekerskaya

Sekerskaya

66

7

31

24

Ovis aries

1

3

18

30

Capra hircus

3

Sus scrofa dom.

65

23

61

49

Bos taurus

53

97

205

302

Canis familiaris

1

6

5

Equus sp.

4

3

Cervus elaphus

24

2

Bos primigenius

2 2

5

Capreolus capreolus 1

9

Lepus europaeus

1

2

Emys orbicularis

2

Aves Total

4

1 220

132

344

416

Two main observations can be made here: first, the number of taxa reported from the jointly or Durham-excavated areas is considerably higher than from the Kyivexcavated areas, even where the analyst is the same and the Kyiv sample is larger. Since the additional taxa are mostly fairly small, this might reflect differences in bone recovery. Secondly, there is a marked difference in taxonomic composition between the two samples from the Mega-structure, with cattle dominating in the Kyiv (week 8) portion studied by Sekerskaya, but trailing pigs and caprines in the jointly excavated portion (weeks 2–7) studied by Schwartz and Tóth. It is impossible to say for sure to what extent this relates to excavation methodology or to recording protocols and inter-analyst variation, but the similarity in results obtained by Sekerskaya for the two pits points to the importance of the latter. In order to remove this inter-observer effect, Figure 5.52/2 compares excavation areas using only data recorded by Sekerskaya. The immediate impression is of similarity, with cattle dominating across the board, though in fact the contribution of caprines ranges widely, from 5% in House A9 to 15% in the Sondazh 1 Pit. This variation might point to functional differentiation, spatial variation in preferences for or access to resources, or perhaps temporal changes within the period of occupation.



Animal Bones 

 403

Figure 5.52: (1) comparison of frequencies of major taxa between (a) analysts for the Mega-structure and (b) areas recorded by Sekerskaya (%NISP); (2) comparison of frequencies of major taxa in areas studied by Sekerskaya (%NISP); (3) findspots of bones assigned to different phases within the Mega-structure. NB. Each dot shows a total station record that can represent a single or multiple bone fragments (by D. Orton).

404 

 The Finds

However, any such interpretation must be tempered by the possibility of taphonomic differences, especially between house and pit contexts. In particular, it is not clear how bones found within building remains relate to the use and/or abandonment of those buildings – a topic that is explored below in the context of the Mega-structure – particularly since Sekerskaya did not record evidence for burning. Without more taphonomic data, we cannot rule out, for example, the possibility that some of the bones recovered from the rubble of houses had actually been spread from the associated pits by ploughing or other disturbance. Whether due to spatial differences in the deposition of animal remains or to inter-observer factors, the amount of variation seen within the recorded assemblage from Nebelivka should be taken as a cautionary tale when comparing data between similar sites. Had only the Mega-structure been excavated, for example, Nebelivka would have stood out from neighbouring sites in terms of the balance of domesticated species (Fig. 5.51/4).

5.3.5 The Mega-Structure During the excavation of the Mega-structure, considerable effort was dedicated to understanding the building’s abandonment and eventual destruction by fire, and the relationship of finds to this process. Importantly, there was probably a period of abandonment prior to the final burning and collapse, as indicated by the formation of a thin chernozem between the living surface and the burnt building debris in a number of different areas within the Mega-structure (Chapman et al. 2014). Bones beneath this layer are assumed to represent activity within the building at or immediately after the point of abandonment (Phase 2); those within the burnt debris, or at the interface between chernozem and debris, are assigned to the destruction event (Phase 3). The authors also suggest a final phase (4) of post-destruction deposition. Table 5.16 shows counts of bone fragments assigned to each of these phases, with findspot locations plotted visually in Figure 5.52/3. The number of bones assigned to the destruction phase is somewhat surprising, but at face value might be taken to indicate significant deposition within the abandoned structure shortly before the burning event. Alternatively, some of these bones may belong to phase 4, perhaps becoming mixed into the destruction debris by subsequent disturbance and recent ploughing. Bones beneath the chernozem may have been protected from burning to some extent, especially in less intensely burnt parts of the structure, but those that were on the surface within the building at the point of destruction can reasonably be expected mostly to be visibly burnt. Where burning is not evident, bones are likely to represent subsequent activity, i.e. Phase 4. To this end, Table 5.17 compares burning rates and



Animal Bones 

 405

Table 5.16: Frequencies of diagnostic and non-diagnostic bone fragments assigned to each phase of the Mega-structure (by D. Orton).  

Diagnostic

Non-diagnostic

Total

Destruction

123

844

967

Living Floor

24

423

447

Pre-Mega-structure

2

2

4

Unattributable

71

296

367

Total

220

1565

1785

distributions between phases of the Mega-structure (excluding bone recorded by Sekerskaya, who did not report burning), and other site areas with reasonable sample sizes. Burning rates are low throughout the structure: 4% amongst unattributed specimens (mostly from the unburnt portion of the structure); 2% on the living floor; and 14% even in the destruction phase. Much higher rates are seen elsewhere on the site, though this comparison should be treated with caution due to differences in recovery strategy and analyst. The low burning rate within the building suggests either that a substantial portion of the bone was deposited after the burning event or, perhaps more likely, that bones within the building were somehow protected from the effects of the fire. Within the burnt portion of the building, burnt fragments are generally present in proportion to the total density of bones, with the surprising exception of the South-West corner (Fig. 5.53/1). This matches the distribution of hightemperature vitrified daub as measured by Shevchenko (see above, Chapter 4.9 and Fig. 4.41).

Table 5.17: Comparison of burning rates within the Mega-structure and with other areas (by D. Orton).  

 

Burnt

Unburnt

Burning rate

146

1643

0.09

Living Floor

9

438

0.02

Destruction

121

845

0.14

Unattributable

13

354

0.04

Sondazh 1 - Pit

6

96

0.06

Test pits

120

537

0.22

Mega-structure

 

406 

 The Finds

Figure 5.53: (1) findspots of burnt bone within the Mega-structure. As above, single dots may represent multiple specimens; (2) spatial distribution of major taxa within the Mega-structure; (3) foetal/neonatal bones recovered from the living floor and destruction phases of the Mega-structure. Diameter of markers is proportional to number of specimens. Length of burnt part of Mega-structure – 36m. (by D. Orton).



Animal Bones 

 407

The spatial distribution of taxa within the Mega-structure remains does not show any obvious patterning (Fig. 5.53/2); cattle are possibly over-represented within the unburnt area, but this is based on a small sample size. Nor is there any discernible pattern of body part representation. There is an interesting concentration of foetal/ neonatal bones in the South-West corner of the building, however, including remains from at least one pig and at least two sheep or goats  – all unburnt and associated with the living floor (Fig. 5.53/3). All diagnostic perinatal bones within the structure are mandibles, apart from an isolated sheep/goat maxilla in the destruction phase and a distal pig humerus within the cluster – a pattern that may simply reflect limited preservation since these are some of the most robust bone portions in the body. Without more detailed information on the taphonomy, treatment, and precise situation of these bones, it is hard to suggest a firm explanation, though the presence of two different species of perinatal bones in the same exact location is unlikely to be coincidental, and may point to deliberate deposition associated with the abandonment of the structure.

Andrew Millard

5.3.6 Isotopic Dietary Information Supplementary dietary information produced through AMS dating of over 80 animal bone samples led to the plotting of carbon and nitrogen isotopic values (Fig. 5.54). These data indicate that the animal fodder for the domestic suite of caprines, cattle and pigs primarily consisted of a range of C3 plants, which include most naturallyoccurring plants in the region as well as domesticates such as wheat and barley. The 3.5‰ range of δ13C suggests variability in fodder plants consumed, including some C4 plant consumption, which might include naturally occurring C4 grasses. This contrasts with the Early Neolithic South Romanian site of Măgura-Boldul lui Moş Ivănuş, where δ13C values did not exceed – 19.9 ‰ (Balasse et al. 2013). As many bones from Nebelivka were not identifiable to species, both wild and domestic animals may be included. The one outlier is a cattle bone clearly indicating maize consumption, that was dated to the 1950s or 1970s AD (OxA-31730).

408 

 The Finds

Figure 5.54: Isotopic collagen values, Nebelivka mammals (by A. Millard).

David Orton

5.3.7 Summary The various zooarchaeological datasets from Nebelivka have been combined, though the secondary nature of this report inevitably limits the depth of conclusions. Nonetheless, some general observations can be made. First, while differences both in excavation strategy and in analytical protocols appear to have had an effect upon the results, there do also appear to be genuine differences in bone deposition between areas. Cattle remains are extremely abundant in the remains of House A9, for example, while the pit in Sondage 1 contained more of the smaller domestic taxa. These intra-site differences have obvious implications for inter-site comparisons, highlighting the risks in assuming that bones recovered from large and complex settlements, especially from small-scale excavation, are necessarily representative of those sites as a whole. Likewise, the results from the Mega-structure highlight the dangers of variation between analysts and excavation teams. The processes by which the bone assemblage from the Mega-structure formed remain enigmatic, with the low rate of burning being particularly hard to explain.



Plant Remains 

 409

While the small number of bones found on the living floor – including a curious cluster of perinatal lamb and piglet bones  – may have been protected from fire by accumulation of sediment during the period between abandonment and destruction, the large number of unburnt bones found within the destruction layer is harder to explain. One possibility is that a significant quantity of these bones was actually deposited on the house after the destruction event, becoming included in the daub layer by subsequent disturbance. This would be a very interesting phenomenon, if correct, but does not accord well with stratigraphic observations in the field. If the combined recorded fauna from Nebelivka are taken at face value, they indicate that the settlement relied heavily on domestic animals and particularly cattle, in keeping with other nearby Trypillia sites.

John Chapman, Galyna Pashkevych & Dan Miller 5.4 Plant Remains John Chapman

5.4.1 The 2009 Season A wet-sieving operation led by Mr. Ronan O’Donnell was able to process key deposits from House A9. The method used has been developed as a standard for the water-sieving of Ukrainian samples for archaeo-botanical research by Dr. Galyna Pashkevych: a sample of one bucket of standard size was divided into six parts, with each part washed in another bucket five or six times and the light fraction collected before the heavy fraction was retained. For time reasons, some of the samples were washed only three or four times. A total of 11 samples was processed from sealed contexts inside the daub layers and the remains were air-dried in the field base. With the exception of one small grain of Triticum sp., no plant remains were identified. This charred seed was AMS-dated in Poznań (Poz-32552), with the date of 5030 ± 40 BP showing that it was indeed coeval with the house. This indicates that at least House A9 was kept extremely clean during their occupation. There are two principal candidates for the disposal of the plant remains: (a) the pits which were often dug within 5m of the house; and (b) the incorporation of plant remains into the house daub during its manufacture.

Galyna Pashkevych

5.4.2 The Mega-Structure (2012) The palaeo-ethnobotanical analysis was conducted on soil samples collected during the field season 2012. Wet sieving was carried out on the soil systematically selected from a range of different site contexts. Botanical macro-remains, charcoal

410 

 The Finds

and different organic materials were separated from soil using the flotation tank constructed by Mr. Mykhailo Videiko Jnr. The samples contained rootlets of modern plants, pieces of charcoal, small fragments of ceramics and daub and rare carbonized grains of cultivated plants. The analysis of samples, identification and measurement of grains was carried out according to a standard laboratory procedure based on the use of a standard lab. microscope (Pashkevych 2014). The composition of the samples was very diverse. In the course of the microscope study, it was revealed that grains of cultivated plants comprised a very small quantity. The majority of the samples contained practically no grains of cultivated plants. In most cases, the plant remains have been destroyed or damaged during flotation. Sometimes this damage was so serious that there was no possibility of identifying the samples to either species or genus. Individual grains of cultivated plants were found in the samples. Several samples contained only a number of small fragments of grains which cannot be identified. Grains and seeds of the following cultivated plants were discovered: – Cereals: emmer (Triticum dicoccon), einkorn (Triticum monococcum), hulled barley (Hordeum vulgare) – Pulses: lentil (Lens culinaris), pea (Pisum sativum), bitter vetch (Vicia ervilia). Grains of emmer and einkorn are the most frequent among the finds. Palaeoethnobotanical materials from more than a hundred Trypillia sites have yielded thousands of pottery fragments with impressions of plant remains, hundreds of kg of clay daub with similar impressions and carbonized grains and seeds. This large data set makes it possible to determine the assortment of plants cultivated by the Trypillia groups (Yanushevich 1989; Pashkevitch & Videiko 2006; Kirleis & Dal Corso 2016; Dal Corso et al. 2019). This assortment consisted of hulled wheat: emmer, einkorn and spelt, as well as pulses – pea, lentil, bitter vetch. Thus, the assortment of cultivated plants revealed in the samples from Nebelivka is typical for Trypillia cultivation practices. In contrast to the restricted finding of grains and pulses, the seeds of weeds and wild plants were present in many samples and in a well-preserved condition. The seeds of white goosefoot (Chenopodium album) and yellow foxtail (Setaria glauca) prevail among the weeds. Other weed species also discovered included: fumitory (Fumaria sp.), lady’s bedstraw (Galium aparine), and small seeds of Cruciferae and Brassicaceae of indeterminable genus. Since the good preservation of these seeds suggested recent deposition in the soil, a group of weed seeds was AMS-dated in Oxford, with the result that they were indeed modern in date. The abundant impressions of plant remains on daub were more informative than the weed seeds. The impressions included the grains of emmer wheat, einkorn wheat and hulled barley, with occasional impressions of well-preserved ears.



Plant Remains 

 411

Thus, all these data show that, included in the crops grown by the inhabitants of the settlement of Nebelivka, there was hulled wheat, barley and pulses – peas, lentil and bitter vetch.

Dan Miller

5.4.3 The 2013 and 2014 Seasons The 2014 season included an extensive palaeo-environmental testing program, with over 285 samples collected and ca. 4,050 litres of deposit processed in the field. This produced a very limited archaeo-botanical (charcoal) assemblage, and the first molluscan (shell) evidence from the site. The extremely low levels of wood charcoal across all features except pits raises a number of unanswered questions about the total amount, life-use, destruction, and taphonomy of charred timber on the site. In 2014, the author’s own version of bucket-flotation/sieving was used, derived from North-West European traditions of wet-sieving and screen-processing. These techniques are especially suited to deposits with low charcoal content, often wet or moist. The method can be extremely efficient, with very high recovery rates, especially of semi-buoyant items, such as sediment-infiltrated charcoal and shells. It can therefore be stated with certainty that many of the Nebelivka deposits, especially those associated directly with houses, are virtually charcoal-free, including micro-charcoal in silt/clay fractions. Isolated fragments (eg 150m2, with estimates of 600–700 person-days or a month for 20–25 people or two months for 10–12 people. Building a larger house meant access to a larger pool of labour than was usual. The different construction methods of Assembly Houses involved far less labour, with the possibility of portable structural elements from unroofed structures moved to different locations during their use-lives. Each house made a statement about the whole of the megasite landscape, for all parts of the landscape were built into a Trypillia house. Construction would have included less-skilled work, such as the gathering of reeds for thatched rooves or chaff for mixing with the clay, heavy work such as carrying wood, clay and water to the building site and more skilled work, such as carpentry and thatching. Provisioning of the building team also involved resources, including feasting at the end of the project. This implies that building was a co-operative effort by men, women and children from more than just the household itself. Building a house meant building social relationships in the Neighbourhood or further (Hofmann & Smyth 2013). Each house also made a spatial statement concerned not only about proximity to antecedent structures but also about the impact of the house on visibility patterns in, and movement around, the Neighbourhood. Many Visibility Graph Analysis (VGA) plots showed an extension of private or low-permeability space outwards from individual houses, showing the constraints on movement close to a house (Fig. 4.19– 4.25). The visual impact of houses in their immediate area extended over an entire Neighbourhood and often over much of a Quarter. The predominance of 2-storey buildings over 1-storey houses by a factor of five would have altered the Nebelivka skyline, providing a highly visible form of house differentiation which may have been reinforced by the diversification of household practices. Just as at a higher, site level, dwelling on the Nebelivka promontory converted a space into a place, so the building of a house on a certain area gave that spot the cultural value of a place, partly through its visual impact (see Section 4.3.2) but also related



The Nebelivka Megasite 

 419

to the identities of the residents. The length and complexity of the house biography would augment that place-value until the decision to end its life (Tringham 2005). At that point, the household would encounter one of the Trypillia group’s strongest principles – viz., never to build a second house above the burnt or unburnt remains of an earlier structure. This meant that the cultural memory of a burnt or unburnt house survived in the cognitive maps of the local residents, unless the burning produced what we have termed a ‘memory mound’ of burnt debris (see below, p. 421). In this journey of a house’s life, a small area of Nebelivka began as undifferentiated public ‘space’, became a domestic ‘place’ and then returned to public space but enriched by ancestral connotations. Although prehistorians make the assumption that household production was dominant in the Neolithic and Copper Age, direct evidence of household production is rare at most sites – as is the case at Nebelivka. The piecing-together of the use-lives of Nebelivka houses from scattered evidence (mostly test pits but also two excavations of entire houses) creates a varied picture of maintenance activities (Alarcón García & Sánchez Romero 2010). Soil micromorphological analyses show that houses were kept clean of coarse organic and cultural residues. Practices such as lithic toolmaking and -repairing were only occasionally represented: curated tools tended to be found in houses, whereas débitage from short-term tools was found in pits. Pottery in burnt house ‘death assemblages’ was dominated by vessels for food preparation and consumption. However, there was little materialization of food storage, textile production (either spinning or weaving) or cooking. Moreover, low discard rates are attested for lithics, polished stone tools, worked bone tools, ornaments, figurines, fired clay counters, whole vessels, complete animal bones, plant processing residues and prestige metal goods. Deposition and discard were dominated by fragments of pottery and fragments of animal bones, although we also encountered ‘complete’ vessel profiles with or without sherds missing. The deposition of whole vessels was a special act of reverence by a household or person. The marked absence of what we may call ‘functionally coherent’ pottery assemblages reinforces the conclusion that we are not, for the most part, dealing with ‘living assemblages’ but, rather, staged deposits to mark quotidian or special ritual events. An example of the locus of such events concerns the unusually large number of platforms in the Mega-structure, where eight such ritual foci were distributed across the building, some inside rooms and others in the open central courtyard (Fig. 4.31 & 4.35 upper). Turning to consumption, the most dramatic ritual event on a Trypillia megasite was the burning of the Mega-structure. It greatly exceeded the visual impact of the burning of a dwelling-house, engaging a wide array of senses which included smell, sound and sight. Fires were started in two different parts of the Mega-structure – the South-West corner and the Eastern rooms  – with little evidence for burning in the central open space or the Eastern courtyard. Extrapolation from the volume of fuel required to burn the two-storey Nebelivka experimental house indicates ca. 292m3 for the Eastern rooms and 64m3 for the Western rooms  – a major communal effort

420 

 Discussion

of fuel collection. This collection is approximately double the fuel required to burn a ‘normal-sized’ two-storey 15m × 5m dwelling house (an estimated 187.5m3). These figures emphasise the huge collective undertaking needed to achieve a successful, and obviously deliberate, burning of the Mega-structure! The house ‘death assemblages’ contained sherds as well as rare complete vessels, indicating the deliberate fragmentation of pots before deposition and the placing of sherds from the same vessel in different deposits, as was occasionally found in the Mega-structure. The ceramic groups in the Mega-structure’s destruction layers amounted to almost 60kg of pottery, comprising over 2,500 sherds – an indication of multiple depositional events. The sherds probably derived from the entire megasite, with a low closed: open ratio showing an emphasis on communal consumption vessels. Preferential deposition of large sherds, often from open vessels, typified the living floor of the Mega-structure. The placing of foetal/neonatal bones of at least one pig and two caprines on the living floor in the South-West corner suggests another form of special deposition. The large concentration of miniature vessels had probably fallen from a shelf in one of the East rooms after storage for ceremonial use. The very special graphite-painted decoration underlines that these were precious parts of the Mega-structure possessions, not readily exchangeable for other vessels. A similar pattern of deposition of large sherds was found in the basal layer of the large clay pit in Sondazh 1, whose secondary use included widespread discard of objects in a short period of perhaps years rather than decades. The episodes in the Pit in Sondazh 1 comprised mostly small assemblages, perhaps placed by members of a single household, but with occasional larger assemblages of parts of over 40 vessels derived from several houses or perhaps a whole Neighbourhood. However, the high closed: open ratio of vessels in most Pit episodes shows that food consumption was more likely to be small-scale than large, communal events. Variations between excavation units in the fabrics used for various vessel types revealed different contributions by different people to the depositional assemblages of the Pit in Sondazh 1 and House A9 – perhaps an indication of links between the local place and contributing groups of neighbours. The faunal assemblages from different excavation units showed variations between a preference for domestic cattle in the Mega-structure and House A9 and a more even consumption of domestic species – caprines, cattle and pigs – in House B17 and its pit, as well as the Pit in Sondazh 1. Given the large quantities of meat yielded by butchering a bovid, these local dietary preferences may be related either to larger-scale consumption in the Mega-structure and House A9 or to the choice of animals from different herds and flocks in the other contexts. The construction of the ‘industrial feature’ for communal cooking (? baking or roasting) is consistent with the faunal evidence for large-scale meat consumption. The identities of household residents were focussed on family membership and consolidated by choice of house location and the process of house-building itself. Expansion of identities beyond the house was related to resource acquisition and



The Nebelivka Megasite 

 421

consumption choices. A rare example of personal identity comes from the recognition of multiple ways of producing projectile points on high-quality imported flint – here interpreted as different knappers acquiring exotic flint to make their own points for hunting. The recognition of only two figurines made in a realistic style, each with their own distinctive faces, hints at the ways in which persons could be differentiated at Nebelivka, although this was a rare occurrence. Each miniature vessel in the group stored in the Mega-structure was different, suggesting an element of personalisation in this group. The household member who deposited a sherd stylistically linked to the ancestral settlement of Volodomyrivka, or an Early Trypillia-style lunate or rhomboid point, was making a personal point about cultural memory – about their links to a valued past out of which the Nebelivka megasite emerged. It is also possible that the household preference for beef over mutton or pork was a traditional form of diet maintained by the latest members of the lineage in that house. The final question about houses at Nebelivka concerns their chronology. Although the overall chronology of the dwelling of Nebelivka is not in doubt, a wiggle on the calibration curve has prevented us from producing a definitive intra-site sequence to enable us to place houses and Neighbourhoods precisely within the period 3970–3770 BC. However, it was possible to model the proportion of houses that were coeval depending upon the mean length of house usage (10, 25 or 50 years) (see below, pp. 426–427). In summary, houses were familiar parts of the Trypillia Big Other, known from all the home communities who came to live at Nebelivka. In the absence of functionally coherent pottery assemblages at Nebelivka, most pottery was placed as a ‘death assemblage’ in the house before it was burned and the locale abandoned. In a minority of cases, the burnt mass of house remains formed a low memory mound to mark the place of an ancestral dwelling.

6.1.2.2 The Neighbourhood The concept of a ‘Neighbourhood’ provides a multi-dimensional, social significance for the otherwise neutral term ‘house group’ (cf. Rassmann et al. 2014; Ohlrau 2015). The definition of a Neighbourhood is a group of at least three houses with no internal space between them and gaps at either end separating the group from the next Neighbourhoods (see Section 4.2; Chapman & Gaydarska 2016). In terms of twodimensional planning, there were two space-time characteristics of a Neighbourhood – linearity and length. Linearity assumes spatial direction – building started at one end rather than in the middle and continued with the addition of one house to the current end of the line and not a building some distance away on the projected line. Linearity also assumes temporal direction – the second house in the line is built after the first house, the third after the second and so on. We are still unaware of the duration of the intervals between the expansion of the line with a new house. If all houses in a Neighbourhood were not built at the same time, the length of a Neighbourhood line also reflected a temporal dimension, with the assumption that longer Neighbourhood

422 

 Discussion

lines meant a longer dwelling period – perhaps more successful Neighbourhoods, if ‘success’ is seen as enduring. Although improbable, we cannot currently refute the notion that a single family lived in an entire Neighbourhood, with successive dwellings replacing the last burnt house until the final house was built and abandoned. The number of houses in a Neighbourhood ranged from three to 27, but, interestingly, over half of the Neighbourhoods were small, comprising between three and seven houses. Whichever model for Nebelivka is preferred (see Section 6.1.3), there is a shared assumption of people from different communities coming to live at/ visit Nebelivka and build houses in groups perpetuating those ‘home’ communities (cf. Kruts 1989). The construction of seven houses in a single building season would have required a labour input of ca. 2,000 person-days, or 40 days for an incoming group of 50 people from the same home community, whose seven families could then inhabit those seven houses. Larger Neighbourhoods may well have not been built in one season, with the addition of houses to the longest Neighbourhoods over time. The total of 150 Neighbourhoods identified at Nebelivka indicates that temporal differences must have existed between Neighbourhoods. But the key factor in Neighbourhood creation was the communal labour required to build the houses, whether deployed in a series of building teams from the same home communities or involving co-operation between people from different home communities. While most Neighbourhoods at Nebelivka were linear in layout, there was a small number of exceptions which we have called ‘Squares’ – all six of which were found inside the Inner House Circuit. This special kind of arrangement of houses most probably derived from small home communities outside Nebelivka from which the people came to the megasite. We have named the most regular of the squares ‘Nebelivka Square’ (Neighbourhood 137, Quarter N) in recognition of its special layout; this square features prominently in the Pilgrimage Model as the residence of the Nebelivka ‘Guardians’ (see below, Section 6.1.3). One important feature of Neighbourhoods was the considerable variation in house size. In addition to the visual differentiation of 1-storey and 2-storey houses, many Neighbourhoods contained one or two houses in the highest size class as well as much smaller dwellings. It is suggested that house size differentiation was a standard feature of many Neighbourhoods, although whether these differences were inherited from home communities or developed at Nebelivka remains unclear98. Alongside the regular dwelling houses were smaller structures, which may be considered sheds, workshops or storerooms  – often located outside the Outer Circuit. A last feature found in the OOC was a special kind of weak geophysical anomaly which suggests garden features – perhaps paths or garden beds.

98  Functional differentiation between houses based upon size and internal distribution of objects has been claimed for Majdanetske (Müller & Videiko 2016; Müller et al. 2017).



The Nebelivka Megasite 

 423

An important point about Neighbourhoods concerned their curtailment or abandonment. The decision whether to burn a house or abandon it with little or no burning (conventionally termed ‘unburnt’) was closely related to the location of the house on a major House Circuit, an Inner Radial Street or other less prominent areas of the megasite. Much higher proportions of houses on the Circuits were burnt in comparison with those inside or outside the Circuits. But there were many Neighbourhoods in which houses were treated in both ways  – up to 80% burnt in Quarters B, C and L, in comparison with 40% burnt in Quarters F, I and N (Fig. 4.5). The proportion of mixed house firing practices is a good indicator of inter-household differentiation in respect of small-group agency, with implications for access to labour and fuel for house-burning. The discussion of variability in the deposition of objects at the Neighbourhood level is constrained by the previously-mentioned issue of sample size from small Test Pits but, nonetheless, some interesting patterns have emerged, confirming what we should expect – variable discard in a basic social unit in which people spent most of their time at Nebelivka. There was a tendency for different amounts of pottery deposition in different houses in the same Neighbourhood, which was especially marked in Nebelivka Square. By contrast, and with one exception (NBH 124), similar amounts of pottery were recovered from the central part of test-pitted houses (Zone 9). The decorative motif linkage study showed that, although almost all households selected their own distinctive motif combinations, half of these appeared in more than one Neighbourhood, with some motifs deposited in four Neighbourhoods. This shows that some linkages were manifested at the Neighbourhood rather than at the household level. The only Special Finds link between Neighbourhoods concerns lithics, found in four different Neighbourhoods; all other Special Finds are variously distributed, each in a single Neighbourhood (Fig. 5.45–5.47). The overall conclusion from the modelling of the AMS dates is that there is no statistical difference between the dates of houses within a Neighbourhood whenever there are two or more dated houses. This does not, however, mean that the Neighbourhoods were not occupied for different timespans. In summary, the importance of Neighbourhoods lay in mutual support in a new social environment – one created on an unprecedented scale, with no person familiar with the architectural totality. While the initial challenge for the megasite was the creation of a site-wide identity – the ‘megasite identity’ – the local identities of home communities could be replicated or developed within the Neighbourhood, if not also the Quarter. 6.1.2.3 Quarters The key point about the identities of those living in different Quarters was an outside-to-inside relationship that was stronger than the lateral relations existing across the boundaries of the Quarters. The diachronic development of these outside-

424 

 Discussion

to-inside relationships across zones, even in the five Quarters with fewer than ten Neighbourhoods, was achieved largely through reciprocal movement from the outside to the open inner area, often channelled through the Inner Radial Streets. In all but two Quarters (Quarter M in the Inner Circuit, Quarter E in the Outer Circuit), there were gaps between Neighbourhoods which allowed easy movement between the outer and the inner open spaces. We should not, however, overlook the potential for lateral movement in all three spaces – opportunities within as much as between Quarters. The second important characteristic of the Quarters was the major increase in scale over and above the next dwelling unit  – the Neighbourhood. Since few Neighbourhoods covered more than 1ha, there was a regular tenfold increase in the scale of dwelling in a Quarter. Allowing for the near certainty that not all houses, or all Neighbourhoods, in a Quarter were in coeval occupation, this scalar change would have led to between five and ten times the number of people living in a Quarter, with all the attendant opportunities and risks. Each of the alternative models for Nebelivka (see below, Section 6.1.3) makes the same point that the scale of dwelling in a Quarter could have encompassed, if not exceeded, the scale of dwelling in a small Trypillia site (see also Section 6.2 for the Grebeni site). This meant that, whichever way Nebelivka developed as a megasite, the Quarters offered a structural equivalent to a home community – an important way of coping with scalar stress through reliance on prior social relations. The size range of Quarters – from 5.3ha (Quarter E) to 21.8ha (Quarter B)  – suggests that home communities of varying sizes contributed to the Quarters and/or to their subsequent growth over different periods of time. Assembly Houses constituted a key feature of Quarters, with distinctively open access. Whatever the details, the VGA results emphasise the significance of Assembly Houses through their high visual accessibility to corridors of movement. Assembly Houses were located in the most integrated and public zones of the Quarters, with these locations being a key feature of their temporal development. The variations in the number of Assembly Houses per Quarter can best be explained by the temporal development of the megasite (e.g., in the Pilgrimage Model), with the inner Assembly House of a pair built outside the only existing House Circuit and the second Assembly House built after the construction of the later, Outer Circuit. It could be argued that, in the four Quarters with only a single Assembly House, that building continued to act as a focus for both House Circuits, whereas the sole Quarter with three Assembly House locations attests to the greater mobility of the building in response to the increased need for formalised space. Given the frequent house size differentiation within Neighbourhoods, we may expect similar variability between Quarters. Indeed, we have found results supporting the general trend in two different analyses of house size by Quarter, but these two analyses differed in detail. The analysis of the variation in classes of house size by Quarter (the so-called ‘mean breadth scores’) shows marked variability between Quarters (Fig. 4.16–4.18/1), meaning that some Quarters showed much wider ranges



The Nebelivka Megasite 

 425

of house size than others. Eight Quarters comprised the group with lower mean breadth scores – G (lowest) – C & A–I–F–E–H & N. The other six Quarters comprised the group with higher scores – L–K–B –D & J–M (highest). Potential inter-household social differentiation was also found in a GINI coefficient analysis of house sizes by Quarter, in which there was a good spread of values, with many of the Quarters scoring close to the mean value but with four Quarters (C, D, E & G) markedly lower and three (A, K & M) higher than the mean value (Fig. 4.18/3). There is a tendency for greater house size diversity in the South half of the megasite in comparison with the Northern half. Was this finding corroborated by the Visibility Graph Analyses (VGA) or in the distribution of objects? The results of the VGAs for both entire Quarters and for the successive stages of two of the alternative models for the development of Nebelivka show that each Quarter would have had similar structuring of visibility and movement across the entirety of the site and across significant periods of time. There is a widespread similarity in visual accessibility measurements for most Quarters, except for a group of Quarters located in the Southern half of the site. Buchanan advances two possible reasons for the minor differences in spatial arrangements from their neighbours to the North: temporal (these areas were developing at different times) or ‘kinship’ (different social groups inhabited or used these areas). It is interesting that these Southerly Quarters were the same as those with higher GINI co-efficient scores. There are several examples of the differential deposition of objects across Quarters. The most general case concerns motif linkage, with sherds with different multiple motifs indicating identities at the Quarter level in most areas of the megasite. This feature is reinforced by differences between Quarters in the location of the same motifs – whether on the inside or the outside of the vessel (Fig. 5.19). More specific traits include the largest concentration of miniature vessels outside the Mega-structure – found in several houses in Quarter H. The interpretation of this concentration depends on one’s views on the function of miniature vessels  – whether containing special foodstuffs or ritual substances – but the concentration differentiates Quarter H from the others. A fourth difference between Quarters concerns the various ratios of open vessels (dishes: plates), with the highest ratio of communal eating plates in Quarter G – also the Quarter with the greatest degree of ceramic fabric variation in all Quarters. Taken together, these findings suggest an emphasis on communal consumption for a wider variety of persons than was usual at Nebelivka. This same Quarter G was one of the Quarters with a lower-than-average GINI score, while at the same time scoring the lowest out of all the 14 Quarters on the house mean breadth analysis! Special Finds distributions indicate a threefold division per Quarters: a group of Quarters with many SFs (Quarters B, G and L–N), a group with no or few SFs (Quarters A, C–F and H) and two Quarters (I and J) with middling numbers. Exotic flints (Fig. 5.47) and figurines (Fig. 5.46) were found in all five Quarters in the first group, whereas they were rare in the other two groups (Fig. 5.47). These data offer a crude measure of inter-Quarter differentiation, with exotic flints providing a rare example

426 

 Discussion

of the dispersion (? re-distribution) of an exchanged object within the megasite. There is a match between the higher frequencies of Special Finds in the Southern half of the megasite and the tendency to higher GINI values in this half. A further comment concerns the AMS dating of the different Quarters (Fig. 4.63/2). While there was no statistically significant difference in the start-dates and end-dates of most Quarters at Nebelivka, Quarter E may well have started earlier, and finished later, than other Quarters, with Quarter G also ending later than most other Quarters. The combination of temporal specificity and the possibly higher level of social differentiation in Quarter G singles out this Quarter from the others at Nebelivka – perhaps as a place where longer dwelling times led to social diversification. However, there is as yet no other evidence to confirm the temporal singularity of Quarter E. To summarise, comparison of the Quarters with distinctive pottery deposition and high Special Finds deposition with those with high scores on one or other of house size analyses and those Quarters which were outliers in the VGA shows a difference between the Southern and Northern halves of the megasite. There could have been two pathways to this modest social differentiation: a Nebelivka pattern where local variations in exchange led to local architectural differentiation; and the replication of pre-existing variations in home communities in the various Quarters. At present, we have no conclusive evidence to differentiate these two processes, both of which contradict the notion of intra-site hierarchy.

6.1.2.4 The Whole Site The highest spatial level of consideration  – the whole site  – also takes into comparative account the two spatial divisions cutting across the Quarters – the main Zones, from outside the Outer Circuit (OOC) to inside the Inner Circuit (IIC), as well as the four Areas (North, East, South and West: see Fig. 4.5). Given that the scale of construction  – whether the number of houses, the timber and other resources required or the co-ordination of builders and helpers – inevitably means that there is a temporality to the megasite, we begin with a consideration of the AMS dates before moving to the plan and the distribution of objects. The results of the modelling of the duration of Nebelivka show two peaks of high probability  – 100 years (3970–3870 BC) and 200 years (3970–3770 BC)  – with a higher probability of 200 years (Fig. 4.62). The initial modelling of the number of coeval houses was based upon both megasite durations. The high-input version of a 100-year occupation and the building of 30 houses per annum, each of which lasted 50 years, shows that a maximalist model would have required both high building rates and a long house duration. The model produced an increase of 300 houses per decade up to a peak of 1,500 houses for a maximum of 10 years (the 5th decade), with a symmetrical fall-off of 300 houses until the 9th decade. This trajectory does not fit with Müller et al’s (2016) three-stage model for megasite developments, with a maximum number of coeval houses at the end of the occupation and would have required the burning of 30 houses per annum, or one per week in the dry season –



The Nebelivka Megasite 

 427

the effects of which are not attested in the Nebelivka pollen diagram. For the 200year duration, the 10-year house life implies abandonment so soon after construction that there was little continuity in the dwelling plan. Thus, the middle-range models of houses with a use-life of 25 years produces the most realistic number of coeval houses – between 210 and 250 houses, with modest building rates of between 9 (210 coeval houses) and 20 houses (250 coeval houses) per annum and burning rates rarely if ever exceeding 12 per annum. The goodness of fit of this model with the three alternative models is discussed below (Section 6.1.3). The severe effect of the wiggle on the calibration curve on modelling the Nebelivka internal chronology prevents clear sequencing of even the major planning elements. Three statements about sequencing may be made: (a) the end of the occupation of the Outer Circuit post-dates the end of the occupations of the Inner Circuit and the Inner Radial Streets; (b) there is a clockwise order in construction in the Inner Circuit in Test Pit Group 25 and (c) there is an inner to outer sequence for two out of seven Inner Radial Streets (Test Pit Groups 20/35; Test Pit Group 26). The later abandonment of the Outer Circuit may have been consequent upon its later construction but also from closer relations with the Assembly Houses built in the OOC. The clockwise order of construction is interesting but does not affect any of the three models. The early date for a house well inside the IIC suggests that this may have been a dwelling for one of the Nebelivka Guardians who first occupied the site and planned its skeleton outline. Alternatively, this date may relate to early signs of local dwelling recorded in the Nebelivka sediment core – dwelling related to the early stage of the Distributed Governance occupation of the megasite. The essence of the development of the megasite plan is the creation of new spatial arenas for novel discourses 99, to which may be added at an entirely unprecedented spatial scale. Indeed, as Pauketat (n.d.) reminds us, every (re-)building project (re-) created the site anew, with fresh political opportunities for expanding social power and the potential transformation of the social order. In each of the three Nebelivka models, there are key decisions about how the megasite would next develop – and it was at those critical transformational junctures that the megasite was re-created anew, with revitalised opportunities for political negotiations. The most obvious junctures were the decision to start building a second house circuit and the plan to construct inner radial streets. Both elements offered new opportunities for households and Neighbourhoods located near the places of planned expansion. What differentiated megasites from other contemporary sites was the key initial feature of all megasites  – the inner open area for large-scale assembly. This space started off at Nebelivka as a space of 105ha – bigger than any small coeval Trypillia site and 2/3rds as big as the largest megasite known before Nebelivka. It is important

99  This idea was discussed by Doonan & Hommel (n.d.) in relation to the invention of new objects but is equally relevant for the creation of new settlement forms.

428 

 Discussion

to note that the inner open space was intentionally defined from the very foundation of the site, rather than a space that emerged organically as the site developed or that was wedged-in over time. Monica Smith (2008) has discussed the importance of open spaces in settlements of different sizes, emphasizing that empty spaces were not neutral spaces but were controlled  – indeed, in the case of urban sites, formally sustained to support a wide range of cultural memories at different temporalities and to allow a variety of uses and denoted meanings. Smith underlines the possibilities for conflict as well as consensus over open spaces, reminding us that open spaces can be used against ‘official expectations’. An excellent example of Smith’s ideas concerns the alternating uses of the Igbo village arena, at the physical and socio-cultural centre of village life, between a sacred space for hosting initiation rites and religious rituals and a profane space for meetings and ceremonies (Ugwuanyi & Schofield 2018). The inner open space at Nebelivka can also be conceptualised as the physical and socio-cultural centre of megasite life, with alternating sacred and profane uses. However, it was not a question of copying the principal plan elements from an earlier ‘blueprint’ inherited from the Trypillia A and BI Phases, as has long been assumed (Videiko 2013). Critical analysis of the plans of large sites prior to Phase BII shows that some of the main plan elements were found at earlier sites but never on the same site. This means that the construction of Phase BII megasites such as Nebelivka was more a creative synthesis – or bricolage – of earlier plans than a copying of a wellknown plan (for elaboration, see Section 6.1.3). This conclusion makes bottom-up heterogeneity in Neighbourhoods and Quarters all the more likely from the viewpoint of emergent planning. But such heterogeneity was also a property of the fusion of groups from many different communities in a single large site. Architectural variability is apparent in every aspect of megasite planning, of which 17 can be noted here (Table 6.1). Indeed, it is quite startling to find that the main planning elements at Nebelivka continued to be operational in the face of such heterogeneity! How was this possible? Table 6.1: Types of planning and architectural variability, Nebelivka megasite (by J. Chapman). Type of variability

Variations observed

length of causeways in the perimeter ditch

20–55m

length of uninterrupted ditch lengths between causeways

32–640–720m

width of the outer open space between the perimeter ditch and the Outer Circuit

40–70m

presence or absence of dwellings, or even Neighbourhoods, in the outer open space

Outside OC – total of 78; from 0 (Quarters C, M, N) to 16 (Quarter K)



The Nebelivka Megasite 

 429

Table 6.1: Types of planning and architectural variability, Nebelivka megasite (by J. Chapman).

Continued

Type of variability

Variations observed

presence or absence of pits inside or outside houses in either Circuit or Inner Radial Street and the relationship of the pits to those houses

OC NBH 76 & 102 – no pits; OC NBH 9 & 10 – pits outside houses; OC NBH 47 & 48 – pits inside some houses; OC NBH 62 – pits inside all houses; IC NBH 26 & 34 – no pits; IC NBH 118 – pits outside some houses ; IC NBH 3 & 4 – pits inside houses; IRS 20 & 28 – no pits; IRS 42 & 83 – pits on one side; IRS 95 & 96 – irregular location of pits vis-à-vis houses

presence or absence of an Assembly House in the outer or middle open space

9 outside the outer circuit; 14 between the circuits

number of houses in a Neighbourhood (excluding Squares)

3–22

Number of houses in Squares

15–27

alignment or otherwise of gaps between Neighbourhoods and causeways in the perimeter ditch and between Neighbourhoods in both Circuits

No alignment: Gap 2, Quarter F; Gap 6, Quarter H; Gap 7, Quarter I; Gap 10, Quarter J; Gaps 12 & 13, Quarter L; Alignment: Gap 8, West entrance; Gap 14, Quarter L

presence or absence of kinks in Inner or Outer Circuits

Inner: NBHS 64 & 79 / border of Quarters G & H; NBHS 112 & 117A / border of Quarters K & L; NBHS 134 & 135 / Quarter N; Outer: NBHS 48 & 58 / border of Quarters F & G; NBHS 75 & 76 / Quarter H

alignment or otherwise of kinks in both Circuits

NBHs 104 & 110 (Inner) + 103 & 106 (Outer) – Palaeo-channel at border of Quarters J & K

width of the middle open space between the two 60–160m Circuits length of offsets (kinks)

25–40m

presence or absence of Inner Radial Streets inside the Inner Circuit

Highest in E–H, lowest in I–K; 52 RS – houses from 2 to 26

presence of squares in the IIC

In Quarters B, C and N

presence or absence of blocking streets (streets cutting off the further development of Inner Radial Streets)

NBH 18, Quarter B; NBH 73, Quarter G; NBH 146, Quarter N

number of Neighbourhoods in a Quarter

6–18

number of Assembly Houses in a Quarter

A, C, E, M – 1; B, D, F, G, I, J, K, L – 2; H – 3

430 

 Discussion

Each of the alternative Nebelivka models does, in fact, propose a different solution to the problem of integrating such diversity (see Section 6.1.3). But the general point is that alongside the ‘local’ identities of each community whose members dwelt at Nebelivka was a ‘global’ Nebelivka identity which had to be created, nurtured and maintained from the very origins of the megasite and which provided sufficient common ground for all the various residents/visitors for the project to continue and thrive. This growth required time for the slow build-up of depositional practices which led to the Nebelivka global identity without undermining local identities as well. This was an unprecedented problem for Trypillia communities, with success and enduring dwelling for those megasites who found new ways to reconcile heterogeneity. None of the lines of evidence discussed in this chapter has supported the notion of a megasite hierarchy functioning to impose top-down decisions about the layout and character of the megasite. The absence of materialised hierarchy is one of the most surprising aspects of a site as large as Nebelivka and points to the significance of heterarchy, the habitus and the Trypillia Big Other as alternatives to hierarchy. Alongside the daily household practices which characterised the habitus, the Big Other also played an overarching, integrative role. It is worth recalling Žižek’s (2007) point that the Big Other was not only something which was sufficiently general and significant to attract the support of most members of society but was, at the same time, sufficiently ambiguous to allow the kinds of localized alternative interpretations that avoid constant schismatic behavior. These localised interpretations became materialised in different types of painted pottery, different kinds of figurines and houses of different shapes and sizes. The general conclusion from the multiple analyses of buildings and objects is that material culture was used to build up a relational picture of social differences, with the broader use of graded differences between objects (greater frequencies of plates and dishes) and houses (bigger or smaller structures) in preference to oppositional strategies (the presence or absence of figurines in houses). Such an overall preference for the use of material culture supports hetararchical relations at the megasite, as well as being consistent with the Big Other and the habitus.

6.1.2.5 Production, Distribution and Consumption The question of production, distribution and consumption at the overall site level is best addressed through an examination of the evidence for specialised and quotidian practices. We can identify at least six Limited Interest Groups (or LIGs: see above, p. 31) at Nebelivka  – builders, potters, flint-knappers, bone toolmakers, figurinemakers and metallurgists. T. Taylor refers to persons united by a common skill set in the production of similar structures or objects. Although not necessarily implying the existence of ‘specialists’ or indeed a formal structure (such as ‘craft guilds’), LIGs indicate a local identity which is likely to have underpinned connections between craftspeople at more than one site. An obvious advantage of belonging to a LIG is the



The Nebelivka Megasite 

 431

way that members may have shared production information and innovations between settlements. One caveat over the diagnosis of specialised production concerns site formation processes. The interpretation that most of the Nebelivka deposits were ‘death assemblages’ rather than a direct reflection of lifeways complicates the assessment of the scale of production. Does the small number of bone tools mean few were produced or few were discarded? Lacking a grasp of the scale of production hinders our discussion of productive specialisation. Moreover, all of the three alternative models feature many visitors to Nebelivka, who may have brought objects with them made in their home community and also taken objects made at Nebelivka away with them. This issue is of particular concern with painted pottery. These issues leave us with an approach based on the skill levels which can be inferred from the objects themselves. Skilled production in LIGs can be seen in the individual knapping of different projectile points, worked bone tools with the fashioning of various animal tooth ornaments or metalwork in the creation of the gold hair-ornament. It is likely that the paucity of bone, stone and metal objects was related to depositional strategies on what to in/exclude in a house ‘death assemblage’. There were only two examples of specialised production at Nebelivka  – the construction of public buildings and the production of painted pottery. The public buildings known as Assembly Houses were made, used and burnt in contradistinction to the usual dwelling houses, perhaps as moveable structures that could be set up again in other places. These structures were, in essence, public buildings with specialised purposes, made by builders with a higher level of skill than those building dwelling houses. The specialised production of painted pottery at Nebelivka does not depend only on the disputed interpretation of the ‘industrial feature’ – whether a pottery kiln or a communal cooking installation (Section 4.7.4). Hale’s rejection of a number of high-intensity magnetic anomalies at Nebelivka as ‘kilns’ (see above, Section 4.2) differentiates Nebelivka from the Phase CI megasites of Majdanetske and Taljanki, with their multiple examples of excavated pottery kilns (see Section 6.2). What is important about the Nebelivka painted ware assemblage is the scale of production (higher than for any other material), the complexity of the new châine opératoire and the high level of skills shown in the preparation of the clay, the painting and the firing. It is not coincidental that the two examples of specialised production at Nebelivka were often brought together in some of the most spectacular performances made at the megasite  – the burning of Assembly Houses, including the Megastructure. It is difficult to infer the form or intensity of exchange practices from the scale of deposition of local and exotic, imported flint on the megasite, although there may have been some integration of West-to-East exchanges in three materials – Volhynian flint, manganese pigments for pot-painting and Transylvanian copper. What was important in Phase BII was the regular exchange of low-bulk high-value manganese

432 

 Discussion

pigment to produce trichrome painted wares. Central places such as megasites would have been natural attractors for such exotic materials. The household, or at most Neighbourhood, level seems appropriate for local flint acquisition, with small-scale re-distribution within the Neighbourhood, while Neighbourhoods or Quarters may have acquired black pigments for re-distribution to local potters. The range of consumption practices discussed for Nebelivka covers individual events by one or two persons or a single household (e.g., small-scale Episodic deposition in the Pit, Sondazh 1) up to large-scale depositional events involving many households – perhaps an entire Quarter – when it came to the burning of the Megastructure. It is hard to describe any of these depositional events as ‘specialised’  – rather, the cumulative scale of the practice could have involved many people in the materialisation of the Big Other or it was confined to the iteration and reiteration of the local household ritual habitus. Only the largest-scale depositional practices would have constituted novel ritual practices not found on smaller sites and which could have acted as focal points for the development of ‘Nebelivka identities’. Confirmation of the Nebelivka ‘industrial feature’ as a large-scale cooking facility kiln would be consistent with the ceramic and pit depositional evidence for collective consumption events including feasting. By contrast, the consumption of the dwelling houses at Nebelivka – over 1,000 burnt at temperatures high enough to create a ploshchadka – was almost universally a large-scale practice but hardly a specialised one. The communal effort of collecting enough firewood to fill the house and then burn it down unified the residents and their closest kin, friends and neighbours in a ritual practice that not only tied that individual house into the millennial Trypillia tradition of house-burning but also cemented the household into Nebelivka’s specific ancestral relations. The appearance of a memory mound in one house-burning out of nine located the household for ever in the Nebelivka landscape, which slowly evolved from a centre for the living into a combination of ancestral site and living place. These points lead to the conclusion that limited interest groups did pursue specialised production at Nebelivka in two fields – the construction of public buildings and the production of painted pottery – but that the scale of production in other areas, such as lithics, bone tools and metallurgy, cannot yet be established. This should not lead us to the incorrect conclusion that Nebelivka was only a small Trypillia rural village writ large. Nebelivka was so much larger than the usual Trypillia site that it was relationally different in so many ways, leading us to consider its status as an urban place (see Section 6.3). The question of scale was central to the significance of the Nebelivka megasite.



The Nebelivka Megasite 

 433

6.1.3 Modelling the Growth of Nebelivka We have just noted the size and scale of Trypillia megasites. The question of the actual size of their population has produced wildly fluctuating answers for the nearby megasite of Majdanetske (Fig. 6.1). We have characterised the views of those colleagues who have placed estimates of between 12,000 and 46,000 people at the megasite of Majdanetske as ‘maximalists’ – views which have long typified the debate on Trypillia megasites (see Chapter 2.1). This was a view which we shared at the outset of the Project and indeed for some years later – but which we have felt compelled to abandon. In the same manner as Kuhn’s (1970) hypothesis of changing paradigms in science, the paradigm accepted as useful for doing ‘normal science’ was never able to fit all of the empirical data but these anomalous data were conveniently ignored until they could no longer be sidelined. At this moment, there was the possibility for a new paradigm. We began to accumulate discrepant observations on Trypillia megasites until it became clear that the ‘maximalist view’ was no longer tenable. We have called this process reaching the ‘tipping-point’, when the maximalist view could be challenged on no fewer than nine different lines of evidence (Chapman 2017; Gaydarska & Chapman 2016), to which we have subsequently added a tenth (Table 6.2).

Figure 6.1: Population estimates for the Majdanetske megasite. Horizontal bars show range of population estimates by author(s) (by J. Chapman).

434 

 Discussion

Table 6.2: The tipping-point for the understanding of the Nebelivka megasite (by J. Chapman). Topic

Problem

Monograph Reference

ADS Archive Reference

Reference

Megasite planning

There is so much variability in the Nebelivka plan at all levels that this does not fit a top-down model; many plan features show a temporality that does not fit with maximum coeval occupation.

4.2

4.2–4.6

Hale et al. 2016; Chapman & Gaydarska 2016

Assembly Houses

The heterogeneity in their size, their spacing and the number of AHs per Quarter indicate bottom-up development; the lack of the materialisation of hierarchy in the Mega-structure.

4.2

4.6

Chapman & Gaydarska 2016

Scalar stress

A large number of coeval houses leads to the problem of four or five levels of decision-making – something not remotely possible in a Neolithic or Copper Age society.

Communal cooking

The interpretation of the ‘industrial feature’ as a large-scale cooking place rather than a pottery kiln is more in line with Neighbourhood communal consumption than Quarterwide feasting or large-scale pottery production.

Chapman 2017

4.7.4

5.5

Videiko & Burdo 2016

The The absence of fieldwalking scatters 3.3, 4.1.1 absence of indicating small sites within the 5-km hinterland radius of Nebelivka and the presence of only a few coeval sites within a 25-km radius shows that Nebelivka had no hinterland – a serious objection to a hierarchical central place, with no small sites to provide logistical support.

2.2, 3.4

Chapman 2017; Albert et al. 2019

The The absence of manuring scatters 5.4 agricultural within a 5-km radius conforms to basis an absence of arable intensification shown in the low-yielding cereals found during flotation. All these data support Pashkevych’s model of inefficient, lowyielding Trypillian agriculture – hardly the basis for huge populations.

3.5

Chapman 2017; Albert et al. 2019



The Nebelivka Megasite 

 435

Table 6.2: The tipping-point for the understanding of the Nebelivka megasite (by J. Chapman).

Continued

Topic

Problem

Provision of salt

Modelling of salt requirements for large megasite populations led to estimates far beyond the capabilities of Neolithic – Copper Age transport technology, given the distance of 200–400km from the two probable sources to the megasites.

Resources for housebuilding & -burning

The assumption of thousands of coeval houses in the closing decades of a megasite would have placed an impossible burden on the forest-steppe for firewood needed to burn these houses: 20 million trees for the burning of 1,000 houses.

4.4

6

Scale of deposition (lithics, bone tools, Special Finds)

The minimal scale of deposition of a wide range of finds raises the question of the intensity of the megasite occupation, with very few flint and bone tools, in particular, deposited in comparison with the proposed number of houses and people.

5

5

4.1.1

3.4

The The expectation of a major human Nebelivka impact on the local environment, as pollen core recorded by a sediment core less than 250m from the edge of the megasite, was not met for five measures: deforestation, peaks in agro-pastoral indicators, burning, soil erosion and water quality. The modest human impact signals were dated to before or after the megasite occupation rather than during the dwelling phase.

Monograph Reference

ADS Archive Reference

Reference Chapman & Gaydarska 2003

Johnston et al. 2019; Johnston et al. 2018

Albert et al. 2020

Reaching the tipping-point led to the urgent need for alternative accounts of the megasite phenomenon. It was this process that has led us to three alternative formulations  – the Distributed Governance Model, the Assembly Model and the Pilgrimage Model. The choice of three models rather than one, two or five has been explained above (Section 2.1.3). A full account of each model has already been published. In this section, we present a summary of the models and make an assessment of each model in terms of

436 

 Discussion

the goodness-of-fit to four criteria: the modelled length of the duration of Nebelivka – 200 years; the number of houses built – 1,445; the number of houses burnt down – 1,077; and the absence of a major human impact on the Nebelivka landscape and its forest steppe.

Figure 6.2: The Distributed Governance Model (by C. Unwin).



The Nebelivka Megasite 

 437

The Distributed Governance model (Gaydarska, submitted) works on the premise of a permanent but substantially scaled-down contemporary occupation of 400 houses (Fig. 6.2). Members of ten extended kinship groups (clans) occupied 40 houses each at the site, while other members of the same clans were living in smaller sites within the Nebelivka 100-km catchment area. Ten groups meant that there was a 10-year cycle during which each clan was consecutively in power for one year. The group responsible for running the megasite provided food, water and fuel, waste management, the organization of feasts and ceremonies and conflict resolution. The organising clan was strongly supported in this enterprise by their counterparts living in the wider catchment area, namely by supplying regular provisions of grain, timber, salt, meat and milk. The rewards for such responsibilities included the power to make decisions on behalf of the group, to impose policies or sanction free-loaders, as well as contribute to the formation of alliances. But all of the clan members would have benefitted from the greater opportunities for interaction, more frequent access to a wider range of traded goods, meeting relatives married to women in a distant village or just enjoy gossip. Power and responsibilities were not concentrated at one place or vested in one person but shared across a social network based on common descent. The building and burning of seven houses per annum would have produced the footprint of 1,450 structures in 150 years that fits well with the current evidence. The Assembly model is based upon a month-long seasonal aggregation of increasing numbers of visitors at a centre maintained through the year by a small number of permanent occupants (Nebbia et al. 2018) (Fig. 6.3). The interlinked processes of house-building and house-burning were modelled over five 30-year generations in two iterations – Models A and B – with the latter showing the highest number of houses occupied in the fourth generation and a steep decline in the last generation. The continuous increase of the index showing the proportion of houses occupied over two or more generations underlined the importance of the local ‘built heritage’ at Nebelivka, while also suggesting that visitors continued to come from the same or related small sites throughout the life of the megasite. There was a tension between the ‘local’ identities of visitors from many small sites and the ‘central’ or ‘Nebelivka’ identity which was dominant at the time of the assembly and which sustained the development of a regional political unit to create and run the seasonal assembly. The assembly model created a place of such scale that, in relational terms, dwarfed all other Trypillia settlements, leading to its seasonal functioning as a local city.

438 

 Discussion

Figure 6.3: The Assembly Model (by C. Unwin).



The Nebelivka Megasite 

 439

The Pilgrimage Model considers a concept hitherto rarely developed in prehistory  – the megasite as a pilgrimage centre (Chapman & Gaydarska 2019) (Fig. 6.4). This model is based upon extensive pre-existing social networks linking sites across regions, as well as on the ubiquitous shared symbolic order of the ‘Trypillia Big Other’. Following on from the assemblies of the earliest megasites (Phase BI), pilgrimage centres were selected for a range of different reasons by ritual leaders who became ‘site guardians’. It was these guardians who prepared the ground, organised the large-scale woodland management necessary for initial house-building and negotiated with other settlements for major contributions to the construction of the site. The life of the pilgrimage centre was divided into two stages – an initial stage in which the skeleton of a site structure was constructed in two years, and a later stage, in which house-building and -burning proceeded at a much slower rate, with variable numbers of pilgrims visiting from sites within a 100-km radius for one month within a pilgrimage season of eight months. The pilgrimage centre was controlled by the site guardians, who were initially Nebelivkans but who may have been gradually replaced by non-locals in later generations. The pilgrimage model claims to be capable of explaining many of the key features of the megasite plan.

6.1.3.1 Assessment of the Models In assessing the three alternative models, we should bear in mind that no Trypillia groups had ever created such a large site before, or integrated all of the pre-existing planning elements into a single, coherent settlement plan. It is hard to factor the significance of improvisation and flexibility of planning into these models but this must have been an important characteristic of the successful outgrowth of any model. It is important to note that the successful maintenance of any of the models would have created an alliance between the residential centre and those living on other sites outside Nebelivka – potentially at the regional level but possibly stretching even further. The location of Nebelivka at the heart of such a regional alliance would have increased the megasite’s reputation and made it a political centre of some significance. While two of the models – the Distributed Governance and Pilgrimage Models – were found to fit the three requisite criteria in their first iteration, the Assembly Model required a second variant before a good fit was found. In Model A, a multi-focal, lowgrowth version starting off in only one Quarter and with a low percentage of new houses and a low house abandonment rate failed to reach the requisite number of 1,445 houses in the Nebelivka footprint. This led to the rejection of Model A. However, building in a faster growth rate into Model B was achieved by settlement starting in four Quarters and with double the percentage of new houses than in Model A. This Model reached the desired number of new houses by the 4th generation, thus meeting all three criteria (Nebbia et al. 2018).

440 

 Discussion

Figure 6.4: The Pilgrimage Model (by C. Unwin).



The Nebelivka Megasite 

 441

Table 6.3: Model comparisons (by J. Chapman & B. Gaydarska). VARIABLE

DISTRIBUTED GOVERNANCE MODEL

ASSEMBLY MODEL

PILGRIMAGE MODEL

Spatial development

Inner Circuit built over 5 years in blocks – one block per Clan; later building of Outer Circuit & Inner Radial Streets

Quarter-based: GEN1 + 5 Quarters, GEN2 + 5 new Quarters; GEN3 + 4 new Quarters

Inner Circuit built first in 2 years; slow growth of Outer Circuit from GEN1 onwards; slow growth of Inner Radial Streets from GEN2 onwards.

No. of permanent ‘residents’

2,400–3,200, living in 400 houses and divided into 10 Clans

100 in GEN1, 200 in GEN2 and 300 in GEN4

100 site guardians, some replaced each generation

No. of ‘visitors’ Food and resources brought in to Nebelivka from Clan territory / settlements

900 in GEN1; 2,400 in GEN2; 3,300 in GEN4.

1,500 builder-pilgrims in Years 1–2; thereafter, maximum capacity of 1,920 pilgrims, varying by month and year

Timing of ‘visits’

All through the year, whenever supplies needed

1-month assembly season

After initial construction phase (builder-pilgrims for 8 months p.a.), visits of 1 month for 8 months p.a. for groups of 20–100 pilgrims from any Pilgrim Home Community (PHC)

Trajectory of house-burning

Flat: 7–10 houses burnt (and built) every year, 1 for each Clan; 400 unburnt houses left standing at end

None in GEN1; rises to a peak of 371 burnt houses in GEN4, falling off again in GEN5 to 300

Low in GEN1, rising to just under 300 burnt houses in GEN2, remaining at that level until GEN5; 361 unburnt houses left standing at end

Main threats to No particular threat after human impact initial deforestation and their timing

GEN3, with 452 built houses and 171 burnt houses (woodland requirement of 6.5km2 of forest steppe)

Years 1–2, with 381 houses to build; thereafter, minimal because the building and burning was spread over the 8-month pilgrimage season

Extent of woodland management

Vital hazel coppicing and general management of timber before the start, with general management thereafter

Vital hazel coppicing and general management of timber before the start, with general management thereafter

Vital hazel coppicing and general management of timber before the start, with general management thereafter

442 

 Discussion

Table 6.3: Model comparisons (by J. Chapman & B. Gaydarska).

Continued

VARIABLE

DISTRIBUTED GOVERNANCE MODEL

ASSEMBLY MODEL

PILGRIMAGE MODEL

Discard / deposition

Special deposition / discard reached peak at special ‘Change of Clan’ ceremonies, with other seasonal events

Special discard, especially of exotics, and intensive discard of fine ware ceramics during the 1-month Assembly period; otherwise, settlement discard

Ritual objects (especially figurines), exotics and intensive fine ware ceramics discard during periodic ceremonies in each month and one major festival p. a.

Tension between the overall ‘Nebelivka Identity’ and the Identities of the home communities; overall plan organised from top down, but Neighbourhoods and Quarters organised from bottom up; political body controlled only the Assembly time, so heterarchical

Key group of ‘site guardians’ who made most important decisions about setting up the site and running its large, quick building programme; otherwise, Neighbourhood and Quarter leaders, some of whom would have become ‘site guardians’ in time.

Extent of Each Clan managed specialisation / the site for one year, heterarchy drawing on supplies from 10–13 small sites in Clan territory

Key: GEN – generation

Although all of the models met the three criteria, each model has advantages and disadvantages in comparison with other two models (Table 6.3). The Distributed Governance Model (DGM) has three strong advantages over the others – its dwelling permanence, its creation of a reasonably coherent plan at an early stage and its relatively low, steady consumption of timber. The solidity and spacious layouts of the Nebelivka houses would appear to favour permanent dwelling, since it is not clear why such solid houses would have been built for seasonal occupation. The development of the Inner House Circuit as the major structural element in the DGM produced a robust plan but this achievement took five years – longer than the rapid development of the Inner Circuit and the Perimeter Ditch in the Pilgrimage Model. Thirdly, the building and burning of seven houses per annum would have consumed an estimated 1.4 million trees from the forest-steppe within a 2km radius of the megasite (assuming the megasite occupied the innermost km ring). Spread over the eight snow-free months, this building and burning programme would scarcely have produced a large impact on the surrounding landscape.



The Nebelivka Megasite 

 443

However, there were also challenges for the DGM in terms of the five-year build-up of the initial population, the means of maintaining the clan logistics of supply, the question of how to deal with freeloaders and the operationalisation of a model of social order never tried before. The initial population estimate leading to between 2,400 and 3,200 after five years remains high, requiring major co-ordination between clans. The organisation of a clan whose members were split between their home community and the megasite – with a distance of perhaps 100km – raises questions of leadership and the chain of command, while the continuity of supplies from the home community would have been threatened in times of bad harvest. A robust mechanism by leading members  – for example, a Nebelivka Council  – involving the principles of the Big Other would be needed at clan level to deter small groups of freeloaders who might take advantage of other clans’ logistical supply but who refused to make their own fair contribution. It is accepted that the DGM was, with high probability, an innovative social practice which could have led to local difficulties, if not rejections, especially in the early stages. The concomitant ability to negotiate with a range of community leaders would have strengthened the position of the Nebelivka Council in the DGM. The terms of the Assembly Model (ASM) stipulated a one-month assembly season and 11 months of small-scale settlement at the megasite by the Nebelivka Guardians. The greatest attraction of the ASM was obviously the one-month assembly which would have formed the highlight of many Trypillia communities’ social calendars. Maximum social interaction of all kinds provided the stimulus for visitors to return year after year to Nebelivka, leaving early if meetings went badly and staying if they went well. Feasting, marital exchanges, trading, resolution of inter-village (or intervillage groups) disputes and religious rituals were all potential collective activities that were prioritised in the assembly period. The 11-month period after the assembly time enabled the recovery of both liver and landscape after the hectic, resourceconsuming assembly month. In fact, the second advantage of the ASM was that it required the lowest contribution of timber for construction and destruction of all the models – utilising an estimated 1.3 million trees in the forest-steppe, available within 1.7km radius of Nebelivka and closer to home with a 30-year forest regeneration period and if building took place in the remaining seven snow-free months. This remained true even with the peak of house-building in the 3rd generation and the high peak of house-burning in the 4th generation. However, two significant weaknesses were built into the ASM  – a planning issue and the question of maintenance of the megasite over 11 months of the year. The decision to use Quarters as the main planning element in the ASM left the development of a coherent plan until the 3rd generation – far later than in the other two models. It is not altogether clear how decisions were made on which gaps between settled Quarters would next be filled  – perhaps this was part of the improvisation and flexibility needed to make the ASM work? The short period of one month meant that visitors to Nebelivka spent ten times more time in their home community than at Nebelivka. This left the intensity of the assembly month as the main time to

444 

 Discussion

create a ‘Nebelivka identity’. The Nebelivka Guardians fulfilled the important role of megasite maintenance, including small-scale agriculture and pastoralism for eating and drinking, house repairs and some degree of building as well as the gathering of materials in readiness for the assembly month. Finally, the Pilgrimage Model (PIM) developed three important advantages to offset its main weaknesses: a rapid route to a coherent settlement plan, a long pilgrimage season enabling the spread of activities over this period and a relatively low consumption of timber for both building and burning after the critical initial construction boom. The decision to create an entire House Circuit and dig the entire Perimeter Ditch provided the level of intensity of ritual action to spread the fame of Nebelivka far and wide (cf. comments on British Neolithic pilgrimage sites: Loveday 2015). Moreover, the rapid construction of two of the main planning elements of the megasite enabled the development of a coherent plan for Nebelivka  – something which had never been accomplished before on Trypillia sites. The eight-month-long pilgrimage season meant that far more activities could be completed than in the ASM, with a concomitant gain in strengthening the ‘Nebelivka identity’ alongside ‘local community identities’. Perhaps the ritual ramifications of the pilgrimage season strengthened the Nebelivka pilgrimage identity more than the bonds of the assembly identity. The much slower, more gradual expansion over almost 150 years of the Nebelivka plan to include a second House Circuit and all of the Inner Radial Streets meant relatively little impact on the forest-steppe, with ten houses burnt per annum at an estimated annual cost of 2 million trees, collected during each of eight months from 1.2km radius of the megasite. From the prehistorian’s viewpoint, there is much to support the PIM which alone can explain many aspects of megasite planning focussed on the practice of processions, including the building of pairs of Assembly Houses, with one AH outside each House Circuit in most Quarters. The biggest challenge of the PIM concerned the initial two-year phase, with its very high demands on both labour and construction resources. The timber alone would have required clearance of a radius of 2.2km of forest-steppe in the first year and up to 2.5km radius in the second year, with no time for forest regeneration. The organisation of 1,500 pilgrim-builders was also no small task, with the co-ordination of pilgrims from 13 home communities a major social task of persuasion and encouragement. However, a successful two-year construction phase would have created one of the most intensive pilgrimage experiences that any Trypillia group could have experienced. A second issue for the PIM concerns the number of new houses built by pilgrims visiting for one month but then left unoccupied until the home community’s next visit. The regular establishment of agreements between different home communities to share unoccupied houses would probably prevent the PIM from reaching the total of 1,445 houses required in the model.



The Nebelivka Megasite 

 445

Another question for the PIM concerns the abandonment of the pilgrimage centre, given that McCorriston (2011) emphasises the great stability of pilgrimage practices. This question will be discussed below (Section 6.1.4) but schismatic behaviour and the availability of rival pilgrimage centres would both have undermined the Nebelivka centre. It is thus by no means easy to discriminate between the three models and choose one with obviously greater advantages and fewer disadvantages. At this stage, we cannot reject any of the three models and thus our proposal is to leave all three models in play until we can discount any one of them. It is, however, clear that, on some level, while sharing some characteristics, the three models are mutually exclusive and so a decision is ultimately required, if, that is, that ambiguity cannot be tolerated. We now move to a key topic in megasite discussions and the Project’s second fundamental aim – their origins. Just as the three models presented here show striking differences, so it is that a single model of megasite origins cannot fit all three models.

6.1.4 The Origins of the Megasites There has been a long history of traditional explanations for megasite origins, which have centred on migration and internal or external warfare (see critique in Section 2.1.2). However, an account of the origins of megasites needs to move beyond these traditional factors to examine in more detail the social, settlement and exchange networks into which emerging megasites were embedded. In a recent paper (Chapman et al. 2019), we discussed the origins of megasites in terms of smaller, long-term occupations or seasonal assembly places, creating a settlement rather than military perspective on origins. Shukurov et al. (2015)’s model of Trypillia arable land use demonstrates that subsistence stresses begin when site size exceeded 35ha. Over half of the sites dated to the Trypillia BI stage – the stage before the first megasites  – were larger than 35ha, suggesting that some form of buffering involving exchange of goods for food was in operation. There were two settlement responses to buffering  – clustering of sites with enhanced intersite exchange networks and the creation of megasites. The trend to increased site clustering can be seen over a period of more than a millennium, from Phase BI to CI (Nebbia, Section 3.4; here Fig. 6.5). During this period, megasites emerged not as an alternative to site clusters but within site clusters, leading to the question of why megasites were created in such social groupings. We can summarise the picture of Trypillia settlement at Phase BI/II, before the emergence of the first planned megasites, in the following way. The underlying factor is the pre-existing Trypillia Big Other, which enabled the expansion of networks while being strengthened by those same networks. The three key material traits of the Big Other – the house, the pottery and the figurines – were all demonstrably part of the initial agro-pastoral expansion East of the Dniester valley, proving to be the

446 

 Discussion

Figure 6.5: Distribution maps of (a) Forest Neolithic and Trypillia (b) Phase A; (c) Phase BI; and (d) BII (by M. Nebbia).

Phase

A

A

A/BI

BI

BI

BI

BI

BI

Site name

Mogylna 2

Mogylna 3

Stepanivka

Trifaneşti

Putineşti

Cobani

Ivanovka

BrynzeniOstrov

6.2

4.7

4.5

3.1

2.7

15

14.5

2

Size (ha)

Magnetic

Magnetic

New magnetics

Magnetic

Magnetic

No plan

Magnetic

Magnetic

Type of Investigation

3–5 nests

Nest in inner area

Present

Present

Nest(s) of houses

K. ris 5.2, 5.3

6 or 7 nests

K. ris 5.16 Central nest + 1 other ‘block’

Rassmann Absent et al. 2016 Fig. 16

K. ris 5.8

K. ris 5.7

K. ris 2.7, 2.8

K.ris 2.3, 2.5

Reference

Table 6.4: Early Trypillia settlement plans (by J. Chapman).

Absent

Absent

Absent

Absent

3 weakly concentric circuits

present

Absent

Concentric circuits around nests

Absent

2 weakly concentric circuits

Absent

1 outer circuit

Absent

Absent

Absent

Free-standing concentric circuits

Absent

Absent

Absent

Absent

Absent

Absent

Absent

present

Absent

Absent

Absent

Absent

Absent

Absent

Inner radial Division streets into sectors

Absent

Absent

Absent

Absent

Absent

Absent

Absent

Empty inner space

Absent

Absent

Absent

Absent

Absent

Absent

Absent

Concentric nests AND circuits

 The Nebelivka Megasite   447

25

30

60–80

100

100

150

BI/II

BI/II

BI/II

BI/II

BI/II

BI/II

BI/II

Singerei

Veremya I

Onopryvka

Kharkivka

Vil’hovets II BI/II

BI/II

Tzvizhyn

Kolomiytsiv BI/II Yar

BI/II

Chyzhivka

Vesely Kut

Myropillya

Trypillia

Magnetic

Magnetic

No plan

No plan

No plan

No plan

New magnetics

No plan

No plan

Type of Investigation

No plan

Key to references: K – Koshelev 2004.

200

100–200 No plan

22

15

20–30

Phase

Site name

Size (ha)

Nest(s) of houses

Concentric circuits around nests

Videiko 2012, ris. 27

Videiko 2004, 99 Several nests

Several nests

Absent

Absent

Rassmann Multi-focal Absent et al. plan 2016 Figs. 5–6.

Reference

Table 6.4: Early Trypillia settlement plans (by J. Chapman).

Continued

Absent

Absent

Absent

Free-standing concentric circuits

Absent

Absent

Absent

Absent

Absent

Absent

Inner radial Division streets into sectors

Absent

Absent

Concentric nests AND circuits

Small empty Absent area

Absent

Absent

Empty inner space

448   Discussion



The Nebelivka Megasite 

 449

most attractive elements of Trypillia communities to the Forest Neolithic groups, who produced a limited range of fine wares but lacked figurines and rectangular houses. In comparison to the earlier Phases, figurines in Phase BI/II showed tendencies towards more realistic modelling, with even some portraiture, as well as a greater incidence of painted decoration. The BI/II network brought modest amounts of copper and Volhynian flint from the Western CT area into settlements which, by Phase BI/II, had grown to 100ha and, in the case of Vesely Kut, 150ha. All three large sites were found in the same site cluster, located in the small Northern tributaries of the Southern Bug but with access to the rich chernozem soils of the interfluves. While some elements of what would become central elements of Phase BII megasite planning (concentric house circuits, sectoral growth, Assembly Houses and open inner areas) had already developed by Phase BI/II, they were not apparent on the very large sites and no site showed more than a single ‘advanced’ planning element (Table 6.4). What sociocultural changes stimulated the growth of the planned megasites? A key aspect of megasite origins concerns their temporality  – whether they are conceived of as permanent settlements or seasonal assembly sites. This debate concerns Neolithic and Chalcolithic sites in many parts of Europe – not just Ukraine (for the Balkans, see papers in Bailey et al. 2005). The expansion of remote sensing research in Central and Eastern Europe since the 1990s has brought to the fore previously unknown classes of enclosed sites, especially the Rondel, which can be seen as a seasonally occupied ritual assembly site (Bertok & Gáti 2014) (here Fig. 6.6). It is thus hardly controversial to suggest that some Trypillia sites may have been seasonal assembly places rather than long-term permanently occupied settlements. However, the three alternative models that we have developed for the Nebelivka megasite include both a model with permanent occupation and two models focussed on seasonal assembly places. The vast majority of these assembly sites was laid out around an open inner space that acted as the ritual core of the site and which was intentionally created from the outset. Each alternative model would need to account for the significance of the inner open area that was central to the Nebelivka megasite, whether it is a model with permanent occupation or a model focussed on a seasonal assembly place. Any account of megasite origins must take into account both forms of temporality. We begin with the permanent, Distributed Governance model. The idea to develop a megasite in a special place was predicated upon the occurrence of prior gatherings in that place, which drew not only on a widespread settlement network but also on pre-existing site clusters. For Distributed Governance, the decision to make a megasite as a more permanent arrangement in such a place was an agreement made by the whole network, as mediated by representatives of the many home communities who would settle at the megasite, and was based upon the greater potential for meetings, exchanges and ceremonies than was available in the home communities. This act had five implications. The first was to consolidate alliances between those clans participating in the megasite dwelling, bringing those groups closer to each other than to other neighbouring clans. Secondly, the permanent

450 

 Discussion

arrangement led to a more formalised site plan which, in turn, supported the idea of a community identity. Thirdly, a system of what may loosely be described as ‘tribute’ needed to be put in place to provide subsistence resources for the megasite in the social context of heterarchical organisation. The early megasite community derived several benefits from these new lifeways: increased place-value accrued to the locus of the megasite as a result of recurrent visits to what gradually became an ancestral place, leading to the rising importance of places where large gatherings were held and the flexibility of making decisions at the most relevant and effective stage of the annual cycle, rather than being postponed or limited to the season of occupation of a less permanent settlement. Fourthly, the unprecedented scale of exchange occurring on such early megasites led to cumulative social advantages for those dwelling on such sites. And fifthly, the combination of the traditions embedded in the Trypillia Big Other with innovations in planning and pottery production led to highly desirable lifeways that many other home communities would wish to emulate. The sum total of these advantages led to the attraction of the megasite lifeways to a wider pool of people living in the extended Nebelivka network of 100km radius. The origins in the other two, seasonal models  – the Assembly Model and the Pilgrimage Model  – incorporates elements from Kopytoff’s (1987) African frontier model of colonisation. In this model, Kopytoff sought to develop an alternative to the linear model of North American East-West colonisation  – one that fitted the mosaic character of African settlement. Thus, he emphasised the social primacy of the first settlers in evolving lineage relations over secondary settlers and their relative superiority over still later settlers. This relational model seems highly relevant to the spatial development of early megasites. In both of the seasonal models for Nebelivka, the organisational role for the megasite Guardians – people who represented and acted for their home communities in the evolution of early megasites  – was more important than in the Distributed Governance model. This group of up to 100 people founded the megasite, chose the promontory for settlement, laid out the basic outline of the plan, began the programme of woodland management that enabled building at the requisite large scale, maintained a distinctive form of Trypillia Big Other and negotiated with the members of surrounding small Trypillia settlements to contribute to the take-off of the Nebelivka centre. While the Guardians formed a corporate group composed of members of different lineages and various Trypillia home communities, their identity as a founding group would also have provided individual Guardians with opportunities for preferentially different kinds of interactions with visitors, whether in exchange or ritual practices. These interactions would have led to the formation of alliances between Nebelivka and the most regular visitors and their home communities. In this way, there was the potential for heterarchical differentiation between the Guardians, as a group and as individuals, and early visitors to Nebelivka who settled there in the 1st Generation. The residential area for the Guardians was likely to have stood apart, or been differentiated, from other early houses in the Neighbourhoods or Quarters



The Nebelivka Megasite 

 451

and it would have been to the Guardians’s area that leaders from other Trypillia home communities would have arrived to discuss how to join the Nebelivka centre. Thus, the social network of the Guardians would have been wider than any later group’s network. The overall advantage from these opportunities would have motivated the site Guardians to make a success of the megasite.

Figure 6.6: Lengyel enclosures and Rondels, South-West Hungary (after G. Bertok & Cs. Gáti 2014, Fig. II.89).

452 

 Discussion

However, there were limits to the extent of power differences between Guardians and later settlers, whether because the Guardians became numerically less important with each successive generation, because of the traditional ideology of limiting material differentiation or because excessive differences would have discouraged new home communities from joining the Nebelivka centre. As in Kopytoff’s model, the first generation of settlers at Nebelivka was therefore in a subordinate lineage position to the Guardians but were closer to the ancestors and the Guardians than the settlers of Generation 2, and so on. Every successive generation of settlers was equal but some were more equal than others. Complex heterarchical relations of precedence and subordination were developed and played out across the megasite, partly through the locations of households in Neighbourhoods and Quarters and partly through supra-household practices such as feasting and pit deposition. One visual marker of distinction was the memory mound that appeared at one out of nine house-burnings, which materialised the role of key ancestors in local mental maps in the different sectors of the megasite. Another important sense which changed properties in time and space was sound. Whatever chants, songs, music or rhetorical performances comprised the assembly events and processions, their audial execution and perception perhaps changed very little from generation to generation. We now return to a discussion of the key innovations in production, distribution or consumption that shaped the early development of the megasite for each of the three models. The key innovation concerned the introduction of painted pottery, with changes in the importance of animal husbandry also found. It is important to note that, while major changes were apparent in one aspect of the Trypillia Big Other (pottery), there was relative stability in the other two elements of houses and figurines. The painted pottery that gradually replaced incised fine wares in the Southern Bug-Dnieper Interfluve introduced one of the biggest changes in the Trypillia Big Other in Phase BII. Not only did this mean a new mode of decoration but also a new and much more extensive range of decorative motifs applied to a different suite of vessel forms in which painted decoration was often the only decorative style for many shapes. This meant that painted vessels would come to dominate the visual world of the household (Figs. 5.4, 5.16–5.18). But there were also implications for the entire châine opératoire of making painted vessels, starting with the kinds of clay which could be used for the new, finer vessel surfaces but also the procurement of pigments for the red and black motifs and the requirement of better controlled and higher firing temperatures. These unintended consequences of the adoption of painted wares as the fine ware of choice led to profound changes in Trypillia lifeways, themselves leading to entangled organisational changes (cf. Hodder 2012) with yet further ramifications in Trypillia settlements. Thus the acceptance of the new painted ware style had a transformational effect on the megasite and its relations with other, smaller sites. Four key inter-linked changes can be noted. First, the higher quality of painted wares led to a differentiation in the traditional method of household production, with skill differentials in forming



The Nebelivka Megasite 

 453

vessels – especially very large plates and storage-jars (Fig. 5.1) – and the design and realisation of large painted scenes much larger, more complex and less repetitious than on incised vessels (Ryzhov 2000). The emergence of potters with higher skills in forming and painting led to specialisation – not necessarily full-time – but certainly leading to the creation of a Limited Interest Group with similar skills and connections between Trypillia sites. Secondly, while local iron oxides were available for red and yellow pigments, black pigments were made from exotic manganese, probably from the Eastern Carpathians but possibly from other, remote sources  – requiring an expansion of the exchange networks already delivering Volhynian flint and salt. Any small Trypillia settlement desiring trichrome painted vessels would have needed to participate in this extended network and it may well have been easier for such sites to procure manganese from a well-known local central site rather than negotiate a place with strangers in an extended network. The use of manganese provided the potential for local site differentiation in network linkage, with the greatest opportunities most probably falling to the emerging megasites. Thirdly, as Ellis observed long ago (1984), the increased size of storage-jars enabled a new scale of food storage which offered subsistence buffering for large sites. But, just as importantly, the parallel innovation in large plates expanded the potential for communal food consumption over the opportunities supported by the small to medium-sized bowls and dishes of the Phase BI fine wares (Palaguta 2007, Fig. 7). This in turn opened up new possibilities for feasting at both household and wider scales, with significant impacts on animal keeping. Fourthly, the quality of the new painted wares created a different class of material  – so plentiful that it could not be classed as ‘elite’ or ‘prestige goods’ but distinctive enough to be considered as ‘fine products’ which developed special roles in many social practices. This distinction enabled painted pottery to develop opportunities for identity-formation in depositional practices such as house-burning and pit deposition. In summary, changes in pottery making, use and deposition offered opportunities in differential, specialised production, access to widespread exchange networks and new consumption patterns and depositional practices which were by no means inevitable but, if they developed anywhere, would have been most likely to have appeared at megasites. Elsewhere (Chapman et al. 2019), we have referred to these developments in settlement planning and pottery production as examples of bricolage  – an anthropological term signifying the construction or creation of a work from a diverse range of things that happen to be available (Levi-Strauss 1962; Derrida 1970). We consider the integration of the different plan elements into a coherent, overall plan to parallel the assembly of different elements of pottery-making to produce the distinctive Phase BII painted style as two critical bricolage-led contributions to the emergence of megasites. It was the sea-change in the integration of practices at megasites – both in site planning and in pottery-making – that enabled scalar transformations in the

454 

 Discussion

quantity of people involved, the quantity of material involved and the quantity of house-building and -burning involved. A further change may well have developed from the bottom-up  – changes in animal keeping. Orton (see above, Chapter 5.3) has proposed that the preference for domestic over wild animals was as much a spatial as a temporal trend, with Eastern (Trypillian) sites often selecting more than 90% domestic animals by Phase BII. This greater control over animal keeping would have developed at the household level, with preferences for cattle over caprines and pigs for larger sites with widerscale feasting opportunities. The 100kg of meat available from a single sheep could have been consumed in a small site’s feast for 200 people, whereas slaughtering a bull to produce 350kg of meat would have been (literally) overkill for a small site! However, visitors to Nebelivka could have readily brought live animals to the megasite (Halstead’s (1982) ‘storage on the hoof’) for contributions to feasting on a far wider scale of potentially hundreds of people. The deposition of more cattle than caprine bones at megasites makes sense in terms of feasting logic. One of the greatest attractions of a megasite was the increased scale of interaction, not least at feasting events. Yet quantitative scaling-up was not universal in the new world of Trypillia Phase BII megasites, where continuity in many social practices can be observed. There is no sign of change in the scale and efficiency of arable production at this time. House design and size hardly changed at all in Phase BII. Equally, the frequency of figurine production, once corrected for excavated area and settlement size, remained just the same in the megasites as in earlier and smaller sites (Gaydarska 2019), although most figurines were now made from the same paste as painted wares, sometimes with realistic painted faces, body parts, clothing and ornaments and occasionally portraits (Monah 2016) (here Fig. 2.3). It is likely that the increased similarity of most Phase BII figurines is related to the creation of a Limited Interest Group for figurine-makers (cf. Orphanides 2010). The Trypillia Big Other may have changed through the introduction of painted pottery but other aspects remained broadly the same. Moreover, declining deposition marked Phase BII in the two areas of lithic and copper production. Kiosak (Chapter 5.2.5) emphasises the small size of lithic assemblages found on Phase BII– CI sites, including Nebelivka and Taljanki, relative to the large number of lithics deposited in Phases A and BI. Two traits characterise the Nebelivka assemblage – an efficient and often skilled production of lithic items, including re-sharpening and re-working, and a diversity of knapping styles for the same lithic type which does not fit greater specialisation or centralisation of production. Once again, the household basis for lithic production could well explain both of these patterns, with members of a Limited Interest Group of flint-knappers making projectile points in a variety of styles (cf. Wiessner’s (1983) San points) and repairing flint tools as often as possible. Ryndina (1998) envisages copper metallurgy at two scales – clan and household, with a chronological development. In Trypillia Phase A & BI, clan specialists moved between various settlements while, from Phase BI/II onwards, specialists settled in



The Nebelivka Megasite 

 455

specific villages, developing greater skills in the Cucuteni zone. Again, the notion of a Limited Interest Group for copper metallurgists would cover both phases of the Ryndina sequence. The small number of copper objects deposited on Phase BII and CI megasites (including one axe at Majdanetske (Ryndina 1998, Fig. 66/6), one awl at Taljanki (Ryndina 1998, Ris. 66/12) and no copper finds at all at Nebelivka), as well as on other, smaller sites, is also related to continued working of increasingly small copper items (Greeves 1975) but perhaps more to Taylor’s notion of lateral re-cycling by which damaged copper items are melted down and re-fashioned rather than being deposited in special places (as in the Phase A hoard of Karbuna). This practice reminds one of Annette Weiner’s (1992) famous phrase used to describe Melanesian exchange  – ‘keeping-while-giving’, in which the most important objects are not exchanged but kept in the household as a key element of its ritual identity. In the megasite case, copper items were perhaps not even brought for display to Nebelivka but retained in the smaller home communities as ancestral items. Whatever the exact mechanism for preventing the deposition of copper goods, the pattern is diametrically opposed to the lavish metal deposition occurring in cemeteries such as Varna and Durankulak 500 years earlier (Higham et al. 2018). We have seen how a series of structural changes in Phase BII Trypillia society could have benefited the original megasite settlers more than later arrivals, to the extent of their strong commitment to megasite success. But there is still a residual concern that these structural changes were necessary but insufficient factors in the emergence of these extraordinary sites. It is hard to envisage the scale of social interaction at an early megasite, with visitors meeting people from 30–50 home communities whose previous face-to-face engagement had been limited. In return for commitment to corporate projects (ditch- and pit-digging, the gathering of materials and housebuilding), early residents participated in an unprecedented range of special events, from ‘local’ Neighbourhood pit deposition and feasting to annual ‘global’ celebrations of the megasite itself. There was an element of success feeding success, with tales of the events, their scale and magnificence, spreading through the Trypillia network and attracting more and more people to visit the megasite for themselves. The stimulus of the megasite community for the creation of alliances made Nebelivka and other early megasites such special centres. Thus, the megasite developed an internal network of special places, including houses, Assembly Houses and pit clusters, where memories of past events fuelled repetition, elaboration and differentiation. It was this upward trend of alliance-formation and diversification in meeting that was the spark leading to the emergence of megasites.

6.1.5 The Demise of the Megasites Turning to the demise of the megasites, the first point to note is that we can observe at least three cycles of megasite foundation and decline – in Phases BI–BI/II, BII and CI.

456 

 Discussion

Any explanation of megasite demise should pay attention to the revival of megasites except in the final cycle. The second point to note is that the collapse of the largest Trypillia sites was not the same process as the demise of the Trypillia group tout court (Diachenko 2016a). Current AMS dating suggests that the latest Trypillia communities survived the fall of the megasites by many centuries – perhaps 600 years. In an article entitled ‘Small is beautiful’, Diachenko (2016a) is surely correct to characterise the latest (post-megasite) settlements as much smaller than the megasites, even though sites larger than 70ha were occasionally found even in the CII phase (e.g., Kosenivka and Kocherzhyntsi: Videiko 2013, Tab. 4). The third point is that, although the dating of the latest Trypillia sites shows contemporaneity with the earliest barrow construction on the North Pontic steppe (Ivanova 2014), even the earliest barrows were not coeval with the latest megasites. These were two separate demises and cannot necessarily be explained by the same processes of change. We have already considered the traditional reasons proposed for the demise of the Trypillia megasites  – external or internal invasion, environmental collapse and socio-economic changes (see critique, pp. 25–26). How can the demise of the Trypillia megasites be explained in the context of the three alternative scenarios of less intensively occupied megasites with lower, perhaps seasonal populations and far lower tendencies to reach the local carrying capacity? We can identify three key factors in any account of the demise: personal or small-group agency; shifts in social networks; and the availability of alternative assembly places. Small-scale agency at the individual or small-group level was vital for the success of a permanent DGM, an assembly site with meetings in one month per annum or a pilgrimage centre with a longer season of eight months per annum. A negative scenario can be proposed for each model. For the DGM, the inability of a home community to sustain the additional demands of megasite provisioning, perhaps owing to a succession of poor harvests and negative family growth, would threaten the model. For the PIM, a crisis of legitimacy causing families or whole communities to abandon the pilgrimage centre could threaten the PIM. For the ASM, a succession of assemblies dominated by tensions rather than good times could undermine the model. Each of these negative scenarios shared two key features – they started off small and spread to wider social groups and different communities and their effect took several years/ seasons to provoke change. Even the modus operandi of smaller megasites than those envisaged by Müller & Rassmann (2016) were subject to vulnerabilities which could not be overcome by social formations comprising many communities, depending upon widespread consent, supported by the Big Other and with little or no permanent hierarchy. This form of explanation is potentially applicable to all megasite cycles. The social networks underpinning megasites as assembly places of one kind or another were in as much need of successful interactions at the central sites as the central sites needed strong and stable exchange networks. One material which was not so dependent upon exchange was metal (mostly copper and occasionally gold), because of its potential for what T. Taylor (1999) calls ‘lateral cycling’ – the melting and



The Nebelivka Megasite 

 457

re-shaping of existing copper objects into ‘new’ items. The discovery and exploitation of local lithic sources meant that although, in theory, there was no need for exotic Volhynian flint, in practice there would have been a high status attached to the acquisition and use of exotic lithics. Any major fluctuations in supply of exotic lithics may have created an issue but variable provision of manganese for pot-painting would have created a much more serious problem – leading to major depositional changes and perhaps the switch to incised fine wares, so widespread in the Eastern Trypillia group. Such fluctuations could well have impacted on the stability of the Big Other, which in turn may have affected the functioning of exchange networks. Both of these mutually amplifying threats would have posed problems at the major assembly sites where exchange and ritual deposition were vital for social reproduction. However, such fluctuations were not necessarily permanent, with recovery within a generation, but could have caused the abandonment of megasites until a new cycle of megasite construction. The final decline of the exchange network would have been so serious as to threaten the stability of the latest megasites. The third factor militating against the stability of any single megasite to contain its problems was the availability of alternative assembly places to which people could have moved. The AMS dates for the start of dwelling at Majdanetske and Taljanki estimate 3850 BC (at 68% probability) for the former and 3830 BC for the latter (Fig. 4.63/4). This means that both of these megasites started at least a century after the start of Nebelivka and overlapped with Nebelivka for 80 years (at 68% probability: Majdanetske) and 60 years (Taljanki). Thus there were two alternative megasites to attract dissatisfied or dissident members away from the Nebelivka community. Such possibilities opened up new alliances for people prepared to risk the unknown positive and negative consequences. The severity of the problems besetting the Nebelivka centre, whether small-scale local decisions and the wider-scale problem of discontinuities in exchange networks or indeed the undermining of the Big Other, determined a response from the Nebelivka network. One response was the attempt to solve ‘global’ and ‘local’ difficulties because there were no other megasites to move to. But if there were alternative assembly places, the principle that ‘the grass is always greener on the other side’ may have come into focus and prompted small groups to try out either Majdanetske or Taljanki. A cumulative consequence to this scenario would have led to the weakening and ultimate collapse of the alliances on which the Nebelivka megasite depended over a period of decades. Although leaving open the question of how and why the last-surviving CI megasites collapsed or declined (see Section 6.2), this proposal could explain the demise of megasites in earlier cycles. In summary, the traditional explanations for the demise of the megasites remain unconvincing, not only because of their reliance on highly questionable maximalist assumptions, but also because there has been little detailed palaeo-environmental modelling on which to base arguments about the unsustainability of the megasites. Social explanations have begun to gain traction, with the alternative modelling of

458 

 Discussion

the megasites as smaller and sometimes less permanent and seasonal offering more appropriate accounts of megasite decline and fall.

6.2 Megasites – a Comparative Approach 6.2.1 Introduction The investigation of Trypillia megasites has been patchy in the extreme, with detailed, inter-disciplinary investigations at the sites of Taljanki and Majdanetske and a much lower research input at all other megasites. For that reason, most of the comparative attention in this section will be devoted to Taljanki and Majdanetske, for whom data has been published in contrasting ways. The Taljanki excavations have been published in a series of annual reports which, despite the variable level of details from volume to volume, provide a good start for assessment of the data (Kruts et al., 2001, 2005, 2010, 2011, 2013). By contrast, the excavations at Majdanetske have been published in a series of general overviews, the most comprehensive of which is Shmaglij and Videiko (2001–2)  – an excavation report offering a more holistic insight into the functioning of a megasite. Both ways of presenting the data have their advantages and disadvantages; the best approach might have been a combination of detailed annual reports, complemented by summaries and syntheses. More recently, both sites have been highlighted in the EAA volume on Trypillia megasites (Müller et al. 2016), where there are excavation reports on the Ukrainian-German research at Majdanetske and other chapters referring to the geophysical plan and excavation of kilns at Taljanki.100 A comparison of the ceramic finds from Taljanki and Majdanetske with those of Nebelivka has already been made (see Section 5.1.4). The main findings were the differences in daily practices that resulted in contrasting burnt house (‘death’) assemblages at the three sites – all of which comprised pottery deliberately placed (? staged) before the burning of the house. This finding is reinforced by the contrasting open to closed vessel ratios found at the three sites, with collective consumption materialised more often in Nebelivka and Taljanki than in Majdanetske. Both Nebelivka and Majdanetske showed varied pit deposition but shared some large-scale events. We now turn to megasite planning before a consideration of the resource base of the megasites, whether daily resources, including plants, animals and salt, or housebuilding and -burning resources.

100 We note some different views on the modelling of Majdanetske in a recent paper from the Kiel team (Dal Corso et al. 2019), which move closer to our position on some points on house construction and -burning but not others.



Megasites – a Comparative Approach 

 459

6.2.2 Megasite Plans Both Taljanki and Majdanetske have been partially surveyed by cesium magnetometer, with claims that 70% of the settled area in Taljanki has been surveyed, while, for Majdanetske, this figure varies from 65% (Ohlrau 2015, Table 2) to 79% (Rassmann et al. 2016, p. 39).

6.2.2.1 Taljanki There are some uncertainties about the size of Taljanki (320 or 340ha) and the number of burnt houses (1,335 or 1,370), which probably reflect the variability of the boundary between settled and unsettled area, as well as a clear cut-off point of what is/ is not a burnt house (Fig. 6.7). In any case, the burnt houses in Taljanki were more numerous than at Nebelivka in both absolute numbers and as a percentage of all dwellings. However, given that Taljanki was almost 50% larger than Nebelivka, such figures should be considered from a different angle. If the un-surveyed area is taken into account, it is possible that the total number of Taljanki houses will rise to 2,000–2,200 (Ohlrau et al. 2016, p. 207). This will be again ca. 50% more houses than at Nebelivka, suggesting a consistent pattern of keeping house density between 5 and 6 per ha and pointing to overall continuity in settlement planning between the two megasites. Another line of evidence showing the two megasites’ similarity is the location of most radial streets in the Northern part of each site, with a square, converging streets, streets ‘running into’ circuits and blocking streets seen at Taljanki (Ohlrau 2015, Abb. 34). The incompleteness of the Taljanki plan does not allow the adoption of the nine criteria used in Nebelivka for the separation of Quarters but a formal computation has identified large numbers of house groups that broadly correspond to the category of ‘Neighbourhood’ in Nebelivka, whether 200 (Rassmann et al. 2016, Fig. 15) or 302 (Ohlrau 2015, Table 5 and Abb. 47). The range of houses in the Taljanki house groups was broadly comparable with the Nebelivka range – two to 30 at the former, three to 27 at the latter. Another similarity between Nebelivka and Taljanki is that the large houses in the latter were primarily found in the Outer Circuit, with house sizes in the Outer Circuit in Nebelivka peaking at 66–75m2, comparable to a mean house size of 72m2 at Taljanki. Thus, there are several lines confirming the shared planning aspects of the habitus between Taljanki and Nebelivka. A number of tracks has been identified at Taljanki, which, together with the claimed eight entrances, suggests high accessibility and refutes the possibility for a defensive function (cf. above, p. 127). Rather, the huge open (public) space points to high permeability and an emphasis on assembly, with a variety of sacred and profane uses (Smith, M. L. 2008).

460 

 Discussion

The minimum distance between houses at Taljanki was 3m, while the maximum was 305 m. This once more confirms that, within an overall integrated planning, there was also a great degree of variability that was a result of bottom-up growth rather than the replication of ready-made modular plans. Although 3m left enough space for movement, houses built at odd angles to standing structures would have blocked views and constrained various outdoor activities. It would have made much more architectural, visual and phenomenological sense if structures at odd orientation to each other were built at different times, whereby the ancestral space of where once a house stood was respected and not over-built. Interestingly, this was exactly the opposite on tells, where over-building emphasised the incorporation of ancestral identity (Chapman & Gaydarska 2018a). Converging streets and blocking of radial streets undermined the premise of careful planning that has been often cited as evidence for contemporaneity. Thus, either there was not careful planning – which is doubtful given the experience of constructing large sites accumulated to this point  – or there was a certain temporality involved in the Taljanki planning. However, claims for coeval use of all burnt houses continue unabated (Ohlrau et al. 2016, p. 207). A significant difference from the ancestral pattern in Nebelivka was the almost complete lack of Assembly Houses at Taljanki. Two larger than usual houses were located opposite gaps (entrances) in the outer circuit in the Northern and Southern part of the site (Ohlrau 2015, Abb. 42) – locations echoing the position of the Megastructure at Nebelivka at an entrance but within the inner circuit. So although there were certainly differences in the architecture and location of these large houses, there was also a memory, albeit distant, of an ancestral pattern, showing the combination of tradition and innovation. This structural difference indicates a very different form of social and spatial organisation at Taljanki, with either Neighbourhoods, or Quarters not founded on Assembly Houses, playing more important roles than at Nebelivka or Majdanetske. There are 74 anomalies on the Taljanki plan that have been interpreted as kilns (Korvin-Piotrovskiy et al. 2016, Fig. 5). Until this is confirmed by excavations, such a pattern remains hypothetical. The implication of such a pattern would be an unprecedented scale of pottery production, whose effects on other aspects of Taljanki lifeways would have been expected to be great. However, forty years of excavations at Taljanki have not revealed a distinctly different way of life on this megasite. If we assume that even half of the anomalies were cooking facilities similar to that found in Nebelivka, the number of production units would be still high but comparable with the estimated number of 20 kilns at Majdanetske. This would also open up the possibility for communal cooking and feasting that has been argued for Nebelivka.



Megasites – a Comparative Approach 

 461

Figure 6.7: Magnetometer plan of Taljanki megasite (by R. Ohlrau in K. Rassmann et al. 2014, Fig. 9a).

6.2.2.2 Majdanetske This megasite was the smallest of the three most studied settlements but still impressive at 200ha, of which 165ha have been surveyed. Even a simple visual inspection of the plan (e.g., Ohlrau 2015, Abb. 44) shows that the contemporary use of house circuits and radial streets would have led to access problems, whether radial streets blocking movement along the circuits or the circuits limiting the expansion of radial streets (Fig. 6.8). This would have resulted in radial streets of varied length, very few of which would have been open-ended, and multiple cul-de-sacs resembling

462 

 Discussion

a rabbit warren, which contradicts the idea of careful, single-phase planning. The house circuits added outside the principal ‘Outer Circuit’ – varying from one in the North-East part to four in the South-West part – appear to represent an alternative to the addition of houses inside the Inner Circuit and suggest populations from different home communities settling at Majdanetske. A structural analysis of a Northern segment of the Majdanetske plan (Ohlrau 2015, Abb. 44A: here, Figs. 6.9–6.10) reveals a complex layout constructed from seven circuits or segments of circuits, seven routes into the central open space from the outside and six different groups of inner radial streets. The proposed reconstruction of the building sequence takes into account the multiple blocking of radial streets by the concentric circuits which are themselves limited by the routeways (Fig. 6.9). Considerable local diversity is attested, with between none and five radial streets beginning from a single Neighbourhood (Fig. 6.10: cf. NBH 5 with NBH 3). The minimum number of phases to allow the evolution of this building plan is five (Fig. 6.10). In the first phase, the inner circuit was established; in the second phase, the construction of the Outer group of IRSs could have been contemporary with the building of the first and second Inner Circuits. In the third phase, building the third group of ICs could have been coeval with that of the Inner group of IRSs, while the construction of the fifth group of IRSs could have been contemporary with all of the Innermost group of ICs except IRS 18. The final phase shows that IRS 18 must have been built before the sixth and seventh ICs. The duration of Majdanetske at 250 years (Müller et al. 2018, p. 253) suggests a period of 50 years for each of the building phases proposed here.101 In any case, it is quite impossible for a contemporary use of space in this Northern part of the Majdanetske megasite. This is tacitly acknowledged by Rassmann et al.’s (2016, p. 34) sensible statement that contradictions in the planning layout ‘... might indicate two settlement phases’ (but cf. Ohlrau et al. 2016, p. 207). The method of Kernel Density Estimation (KDE) was used for a formal ‘calculation’ of the number of Neighbourhoods in Majdanetske, producing the high figure of 533 (Ohlrau 2015, Table 5 and Abb. 46) as opposed to 282 (Rassmann et al. 2016, p. 44). The number of buildings in a house group varies between two and 41, with the maximum dwelling in one Neighbourhood well above that at Nebelivka. Further modelling taking the presence of Assembly Houses into specific account has been undertaken to identify larger special units. Nine such units, each ca. 9ha in size, were recognized consisting of an average of 185 houses in 66 clusters (Neighbourhoods), three to five

101 This multi-phase proposal is confirmed by the four-phase scheme for Majdanetske – a scheme first provided with AMS dates during the proof-stage of this monograph. According to the Bayesian modelling, Phase 1 lasted 55 years, Phase 2 135 years, Phase 3 100 years and Phase 4 55 years (Dal Corso et al. 2019, Fig. 2).



Megasites – a Comparative Approach 

 463

Figure 6.8: Magnetometer plan of Majdanetske megasite with Ohlrau’s Quarter boundaries; numbers refer to Ohlrau Quarters; A – inset to be found in Fig. 6.9 (by L. Woodard, based on Müller & Videiko 2016, Fig. 2).

464 

 Discussion

Figure 6.9: Detail of Northern part of magnetometer plan, Majdanetske megasite (Area A in Fig. 6.8). Key – large numbers refer to Ohlrau’s Quarters; small numbers refer to individual paths, house circuits and radial streets (by L. Woodard, based on Ohlrau 2015, Abb. 44A).



Megasites – a Comparative Approach 

 465

Figure 6.10: Building sequence of Northern part of Majdanetske magnetometer plan (as shown as Box A in Fig. 6.9) (by L. Woodard).

466 

 Discussion

economic units consisting of a large building and a pottery kiln, and one Assembly House with a large pit (Ohlrau 2015, p. 62 & Abb. 51). It is not clear how this division relates to the previously identified house groups. The layout of these units is by no means similar (Ohlrau 2015, Abb. 52) and contradicts the idea of regularity that is obviously sought in this modelling by assuming that the constituent components of the units – notably ten Quarters, nine Assembly Houses and 20 kilns – should and would be the same. The final configuration of the empty space in the middle of the Majdanetske megasite was far smaller than in the other two megasites – a mere 26ha – although it would have started out as a much bigger space of over 100ha. The close links between the foundation of Majdanetske and the ancestral pattern in Nebelivka and Taljanki would have reinforced the importance of a much larger empty space than was found. The most obvious candidate to have enclosed a larger space is the 5th circuit (counting inwards), providing space for assembly, ceremonies and rituals. The later encroachment of that space, thus limiting the size of the assembly, may well have caused the demise of Majdanetske, especially given the multi-functional centrality of the inner open space in all megasites (cf. Smith, M. L. 2008: see above, p. 428). The average house size in Majdanetske was 67m2 – less than the 72m2 in Taljanki but similar to the range of Nebelivka house sizes. The minimum distance between houses at Majdanetske is 3m, just as at Taljanki, while the maximum is 165m. There is a lot of uncertainty about the number of unburnt and/or eroded houses in all three sites, and comparison between them is further hampered by the incomplete plans of Majdanetske and Taljanki. Nonetheless, some general observations are possible. If we assume that the rate of unburnt houses is the same in the unsurveyed as in the surveyed areas, then Majdanetske would reach the Nebelivka figure of about one-third weakly burnt/unburnt dwellings, in comparison with a quarter of Taljanki houses from all recorded houses. Given the amount of natural and human resources needed to burn a house, Taljanki seems to have been unusually adept at mobilizing these resources. The declining forest component in the forest steppe at Majdanetske may also have contributed to the higher number of unburnt houses there. Once again, a temporal dimension to construction is implied in the relationship of building to the opening up of the landscape. In contrast to the Taljanki decision to abandon Assembly Houses, the ‘typical’ Assembly House was a strong presence at Majdanetske and has been used as a major component in modelling socio-economic units (see above). The location of most Majdanetske Assembly Houses was a combination between the pattern observed in Nebelivka – at 90 degrees to a house circuit – and that found at Taljanki – opposite gaps (entrances) in the outermost circuits. Three Assembly Houses were exceptions to this pattern by being embedded within Neighbourhoods in IRSs. The apparent neutrality of the Nebelivka Assembly Houses was partly vitiated at Majdanetske by showing a deliberate preference for one Neighbourhood over others.



Megasites – a Comparative Approach 

 467

In summary, apart from the evident similarities between all three megasites in planning elements and their nested patterns of Quarters, Neighbourhoods and households, there are far more similarities between Nebelivka and Taljanki than between the other site pairs, not least in the concentration of Inner Radial Streets at the North end (although this may have been an early feature at Majdanetske before the infilling of the North end), the overall house densities, the distribution of larger houses in the outer circuit and the range of houses in Neighbourhoods. But the differences between Nebelivka and Taljanki were perhaps even more significant than the similarities – the virtual absence of Assembly Houses and the presence of kilns and trackways at Taljanki. Given the centrality of Assembly Houses in the Nebelivka Quarters, the shift away from these communal structures at Taljanki suggests a weakening of the structuring effects of the Quarters in the site organisation. The demise of the Assembly Houses and the start of kiln production of painted ware may not have been coincidental. The partial replacement of the Assembly Houses by limited interest groups formed by potters using the new kilns led to the introduction of a system of ceramic production based on the collaboration of several households in the collection of raw materials (clay, temper, red and yellow paints and kiln fuel), the pooling of manganese from different exchange operations and the construction and maintenance of the kiln. The new Limited Interest Groups were task-based, flexible and temporary and may have vitiated the need for Assembly Houses because of the growing importance of fine painted pottery. The non-overlapping distribution of kilns in relation to Quarters suggests that Neighbourhoods sought to control the operations of the LIGs rather than to allow the Quarters to take full control of the new productive system. The distribution of the kiln products to the Neighbourhoods would have consolidated daily relations with the LIGs, as well as strengthening the Big Other at the intermediate scale. By the time that the early circuits had been laid out at Majdanetske, there were two coeval alternative models for the social organisation of megasites  – one with Assembly Houses but no kilns and one without Assembly Houses but with many kilns operated by Limited Interest Groups. The Majdanetske community leaders will certainly have visited both other megasites and gained insights into the strengths and weaknesses of each model. The resultant plan was a mixture of both models, with one Quarter with a kiln but no Assembly House, five Quarters with an Assembly House but no kiln, four Quarters with both and one kiln placed between three Quarters (Ohlrau 2015, Abb. 52). Moreover, Majdanetske is the only megasite known where Assembly Houses were incorporated into normal Neighbourhoods as well as being located ‘neutral’ to the Neighbourhoods (e.g., in Quarters 2 and 5 and possible 5 and 7). This shows a dispersion of social power into smaller units than the Quarter, where Assembly Houses were traditionally rooted, as at Nebelivka. The diachronic infilling of the North end at Majdanetske shows no difference in the type of Quarter-ly arrangement, with two different configurations (two cases with

468 

 Discussion

only Assembly Houses; one case with both Assembly House and kiln) found in the inner Quarters. This pattern strongly suggests the presence of people at Majdanetske with a knowledge of their own planning ‘heritage’ and the possibilities to install it at the new centre. Not surprisingly, the co-existence of these different formations would have led to considerable tension between the Quarters, one response to which would have been to build their way out of trouble. This is where we become aware of the full significance of the building sequence at the North end of Majdanetske (see above, p. 462), which includes parts of five of Ohlrau’s Quarters (here Quarters 1, 3–6 on Fig. 6.8). The five-stage building sequence not only demonstrates diachronic development – the notion that buildings in stage 3 were built before those of stage 2 contravenes basic megasite planning principles – but also embodies local politics on the ground. One of the important megasite planning principles which allowed the inward expansion of Quarters was the unbroken development of Inner Radial Streets (Fig. 6.10). This is evident at Majdanetske in the SW and NE parts of Quarter 1, where IRS 3 - 4 (NE) and 6–9 (SW) and their continuations 5 (now in Quarter 5) and 12–17 (SW) were blocked only by the 6th Inner Circuit. The summary of Majdanetske (Müller et al. 2018) implies that IRSs were part of the development of Quarters. However, no IRS development was possible in Quarters 3/4 and 6/4 because inner circuits were laid out instead. The summary of social organisation explicitly linked inner circuits to different lineages (Müller et al. 2018), although another way of blocking IRSs is seen in the central Part of Quarter 1 through the building of a nest of houses – the most traditional form of house grouping. The continued blocking of IRSs in the SW part of Quarter 1 allowed the construction of the 7th Inner Circuit which incorporated a large house and an Assembly House (Fig. 6.10). Quarter 1 best exemplifies the tensions between different organisational principles, specifically between the Quarter and the Lineage, with the blocking of IRS development in the central part and its full inward extension in the NE and SW parts (Fig. 6.10). But what was important was that the vast majority of the new Inner Circuits were not laid out as part of a site-wide planning initiative but as an aspect of the local development of a Quarter; here is an example of an important planning principle being taken over by the people of the Quarter, in a way that was not seen at either Nebelivka or Taljanki. Quite how this clash of socio-spatial units – the Quarter and the Lineage – is resolved is never clarified (Müller et al. 2018). These planning decisions were made by settlers who were staking out their building heritage in a bricolage of planning contrasts, being pulled in different directions by competing identity groups. These tensions were perhaps related to the inclusion of both an Assembly House and a kiln in this Quarter. By contrast, the Quarters with the strongest attachment to Inner Circuits rather than IRSs (Qus. 3 and 4) included only an Assembly House, although Quarter 6, with its multiple Inner Circuits, included both an Assembly House and a kiln (Fig. 6.9). Again, there is no discussion of the overlap between supra-household economic units, Quarters and Lineages for this part of the Majdanetske plan (Müller et al. 2018).



Megasites – a Comparative Approach 

 469

Moving away from the area of detailed analysis, Majdanetske shows profound differences from the other two megasites (Fig. 6.8). First, people in Quarter 9 built the only case of Outer Radial Streets known to us. Secondly, outside what is normally regarded as the ‘Outer Circuit’, there were several Quarters (e.g., Qus. 1–3) in which further circuits were added to increase the fragmentation of planning at Majdanetske. As in some Inner Circuits, these additional circuits were often the width of a Quarter, indicating that they were not subject to site-wide planning principles but added ad hoc as part the development of several Quarters. In view of all of these variable planning decisions, it is perhaps not surprising that there was no neat correlation between Quarter-ly organisation and the preference for either IRSs or Inner Circuits. Instead, people living in Quarters made active use of their own settlement backgrounds and their affiliations to other Majdanetske settlers to make choices over their local layouts. The tensions between a preference for IRSs over Inner Circuits can be added to the disputes over how to organise a megasite Quarter – all of which playing out over two centuries and leading to the most complex plan yet seen on any megasite. Although Majdanetske may well have begun its life as a classic megasite plan, closely resembling Nebelivka and Taljanki, the decisions of later residents to differentiate the internal layout and expand outwards (as well as inwards) suggests that, if the megasite had started life as a pilgrimage or an assembly centre, its principal function had changed at some point during its development – perhaps into more of an exchange centre or as a political centre with larger populations. It seems highly improbable that the functions of Majdanetske remained unchanged for the whole of its life. What it evolved into is a most interesting question.

6.2.3 Daily Resources The daily functioning and temporal continuity of every settlement required solutions to the logistical issues of food provision, including salt. We consider how the inhabitants of Majdanetske and Taljanki faced these large-scale issues, beginning with food production. The German group has attempted much palaeo-economic modelling for Majdanetske and, to a lesser extent, for Taljanki. Kirleis and Dal Corso’s (2016) research tacks between the remains from a single house at Majdanetske (House 44) and a general comparison of plant species found at all hitherto known Trypillia sites.102 They make it clear that, despite the very disparate list of exploited plant species deriving from the Majdanetske house, it was still possible to reach the (to the authors surprising) conclusion that there was very little change over time in arable

102 A new study by Dal Corso et al. (2019) broadens the range of activities for which wood was widely required.

470 

 Discussion

practices, even with the appearance of the megasites. If anything, there is a slight relative decline in the frequency of the only ‘advanced’ (viz., bread) wheat found over time. This is why Kirleis & Dal Corso consider Trypillia to operate a broadly ‘Neolithic’ economy. The availability of productive subsistence resources discussed and mapped by Ohlrau et al. (2016, Figs. 2–5) is separated into periods of 100–150 years. What neither the maps nor the discussion addresses is the possibility of bad years and crop failures over this rather long period, while there is an underlying assumption that soil fertility remained constant. However, this contradicts one of the major reasons put forward for the abandonment of the megasites, namely resource depletion (Diachenko 2016a). The lack of a sediment core near Majdanetske and the necessary use of site data limits Kirleis & Dreibrodt’s (2016) reconstruction of the palaeo-environment to the site itself. While the generic environmental account lacks any temporal references to the stages of the archaeological settlement, there is a clear account of the site’s soil development from an Early–Middle Holocene cambisol to the early stages of a chernozem at the same time as the start of dwelling. The authors state (2016, p. 177) that “the location of the excavated houses explicitly excludes a greater age for the Chernozem compared with the Trypillia settlement” and “only thin A-horizons correspond to the level of the Trypillia surface”. We would reconstruct this as the clearance of forest cover over the site area to allow the start of construction, with subsequent development of the chernozem during the Trypillia dwelling. However, Vysloužilova et al.’s (2016, p. 93) conclusion that “chernozem is a result of the process that has been going on for thousands of years” surely means that a “thin A-horizon” cannot be considered as a fully developed chernozem on the area of the settlement. Thus, the chronology of the proposed pedogenesis caused by agricultural steppe development (Kirleis & Dreibrodt 2016, p. 178) is poorly related to the duration of the Majdanetske occupation between 3900 and 3650 cal BC (Müller et al. 2018, p. 253). More importantly, this proxy soil record does not give us information about landscape changes in the Majdanetske micro-region, which we assume to have been covered in a soil mosaic typical of a forest steppe environment. The statement that the initially wooded area of Majdanetske was followed by an open environment (Müller et al. 2017, pp. 75–79) is not supported by any proxy palynological evidence, although the cumulative effects of a hypothetical massive coeval occupation with intensive utilization of resourses would have had a major impact on the environment (for discussion of the Nebelivka core, see Section 4.1). It is important to note that chernozem formation has also been documented for forest steppe conditions in Central and Eastern Europe (Eckmeier et al. 2007; Ehwald et al. 1999). By the same token, the modelling for the megasite of Taljanki offered in the same article (Ohlrau et al. 2016) should be revised by factoring in cumulative environmental impact. The available radiocarbon and botanical data for this site are so problematic as to vitiate any meaningful discussion about flexible subsistence strategies, sustainability and resource management. However, some general observations are



Megasites – a Comparative Approach 

 471

Table 6.5: Estimates of salt requirements for various population estimates for Trypillia megasites (by J. Chapman & B. Gaydarska). Site

Human population + requisite herd size

Low annual salt intake (kg.)

High annual salt intake (kg.)

Small-site model

100 people + 30 cattle + 150 sheep

450

>1,000

Jatranovka Phase B

1,500 people + animals

7,335

22,005

Nebelivka Phase BII

2,000 people + animals

9,780

29,340

Majdanetske Phase CI

7,000 people + animals (Ohlrau 2015)

31,675

95,025

Majdanetske Phase CI

12,000 people + animals (Müller et al. 2017)

54,300

162,900

Majdanetske Phase CI

24,000 people + animals (Shmaglij 1982)

108,600

325,800

Majdanetske Phase CI

46,000 people + animals (Rassmann 208,150 et al. 2014)

624,450

still possible. The site is the largest of the currently known megasites, it has the largest empty middle area and its house density is closer to that of Nebelivka than to Majdanetske. By implication, the middle space may have been variably utilized, including small-scale gardening/horticulture with rotation of crops and fallow years. A residential density close to that of Nebelivka may indicate a continuation of a known pattern delivering stable resource consumption backed by experience in working with large numbers of households. Another vital resource for any human and animal population was salt (Chapman & Gaydarska 2003). The sources of Trypillia megasite salt is still under discussion. The two prime candidates – the multiple Eastern Carpathians brine sources and the limans of the North Pontic – both have arguments for and against. While the Eastern Carpathian sources have been associated with exploitation by local groups from the Early Neolithic onwards and have unambiguous signs of Cucuteni exploitation (Weller & Dumitroaia 2005; Weller et al. 2008; Chapman & Monah 2007), it is a long distance to the Southern Bug-Dnieper interfluve, across four major rivers and their interfluves. While salt from the North Pontic limans could have been brought by river transport to the Trypillia megasites core area, there is as yet no evidence for 5th or 4th millennium BC exploitation of these potential sources (Mircea & Alexianu 2007). Modelling of the salt requirements of a range of Trypillia sites, as well as the Uman area, indicates high cumulative values based even on the lowest salt requirement of 2g per person per day (pppd), with 10g pppd leading to a major increase in demand (Chapman & Gaydarska 2003, Table 6; here, Table 6.5).

472 

 Discussion

The salt requirements of 10–30 tons for a ‘minimalist’ Nebelivka population still represented a major logistical challenge, whether or not the sledge had been developed for land transport. But even the lowest population estimates for Majdanetske would have posed a massive logistical problem of moving 30–95 tons of salt across country. Moving up the scale to the higher Majdanetske estimates of 24,000 creates an unrealistic annual salt demand of between 100 and 325 tons. The dilemma was to cut down on salt intake, with threats to human and animal health and the quality of dairy products such as cheese or spend increasing amounts of time and effort in bringing salt to the megasites. The ‘minimalist’ solutions to this dilemma would seem to have been an obvious way out.

6.2.4 Building and Burning Resources The Nebelivka house-building experiment enabled estimates of the quantities of several different materials all essential to house construction – timber, thatch, chaff for daub production and hazel withies. The scale of house construction implies that the collection of each of these materials posed a logistical challenge. The only way in which megasites could have faced these challenges was through the organisation of communal labour. Here, the estimates for timber, thatch, chaff, withies and fuel for burning are compared for Nebelivka, Taljanki and Majdanetske. There are several issues with the most recent timber estimates for a range of sites (Ohlrau et al. 2016, p. 208 & List 1; see also Section 4.4), despite the allowances made for forest regeneration: (1) the type of house (1- or 2-storey) is not specified; (2) the architectural style of the house  – log cabin or timber-framed or combination  – is not mentioned; (3) firewood requirements are excluded; (4) broad-leaved deciduous forest timber densities rather than forest steppe densities are used to calculate the area of forest exploited; and (5) most seriously, the quantity of timber required to burn the houses is omitted. It is also self-evident that timber was not the only resource required to build a house; consequently, we include estimates for the use of reeds for thatching, withies for wattle construction and chaff for daub-making (Table 6.6). The calculations begin with Nebelivka, where 28,890m3 of timber collected from 246ha of forest steppe103 was estimated to build a total of 1,445 houses with a mean requirement of 20m3 per house104. We have estimated the area of forest steppe required for timber to burn a house at 2ha, or 2,140ha given the total of 1,077 burnt houses.

103  The question of the development of the forest steppe in the Nebelivka landscape is discussed above (see Section 4.1). 104  The figure of 20m3 per house (Ohlrau et al. 2016, p. 208) is close enough to Stuart Johnston’s figure of 17m3 for a two-storey house of average size to be used in these calculations.



Megasites – a Comparative Approach 

 473

For Taljanki, the comparable figures are 2,200 coeval houses, built from 52,800m3 of timber from an area of 352ha of forest steppe (or 176ha of deciduous forest: Ohlrau et al. 2016, List 1). In using the higher number of houses – an estimated 2,700 – Kruts et al. (2001) argued that the exploitation of the forest steppe for house-building and everyday firewood for cooking and heating the houses was unsustainable, forcing abandonment of the site after 50 years. But Kruts did not even factor in the addition of timber for burning 1,553 houses, which translates Ohlrau et al.’s estimate into the figure of 328,680m3 of fuel, collected from 3,106ha of forest steppe, or Kruts’ figures into 405,000m3 of timber collected from 3,820ha of forest steppe. Each of these maximalist estimates would have led to a severe resource crisis, which would not necessarily have occurred using alternative models for the megasites. Comparable figures to those of Kruts et al. (2001) for Taljanki are provided for Majdanetske  – 3,000 coeval houses built from 78,200m3 of timber collected from 260ha of forest steppe – figures which translate into a requirement for burning 2,300 houses of 486,680m3 of fuel from 4,600ha of forest steppe. The same conclusion can be drawn that the high population estimates for megasites were unsustainable for something as basic as construction building and fuel for successful house-burning. This conclusion, incidentally, does not take into account all of the other resources required to build a house. Johnston (Section 4.4) has estimated the quantities of three key building materials required for house construction at Nebelivka, with extrapolations for the other megasites. Large quantities of reeds would have been required for thatching the houses, with collection from the wetlands and valleys nearest the megasites. The higher level of water quality shown in the Nebelivka core during the megasite occupation than before or after the site means that reeds would have been widely available (Albert et al., https://doi.org/10.5284/1047599 Section 3.4). The extent of past productive wetland near Taljanki or Majdanetske has not yet been quantified; modern wetland exists on three sides of Taljanki and on the East and South sides of Majdanetske (Müller & Videiko 2016, Fig. 1). Johnston’s estimates for the quantities of chaff required to make daub for housebuilding, as observed by Kirleis & Dal Corso (2016) (see Section 4.4), indicates a potentially serious logistical problem for storing the chaff, whose requirements vary between almost 3m3 of chaff for a 1-storey house to 5.7m3 of chaff for a two-storey house. Without other storage containers such as linen bags or wooden boxes, this may have been a serious issue in view of the rarity of large storage-jars of 2m3 volume at Nebelivka. Finally, Johnston’s estimates for the quantity of hazel withies needed to construct wattle-and-daub walling emphasise the forward planning essential for any Trypillia building project – let alone projects on the scale of the megasites. The coppicing of hazel bushes for long, straight rods (withies) for use as wattle takes forest management over a five-year period – an important consideration at the start of the creation of a megasite.

474 

 Discussion

Table 6.6: Estimates for the collection, storage and use of building timber, reeds, chaff, hazel withies and fuel for burning at megasites (based on Ohlrau et al. 2016 for timber and S. Johnston for remaining materials: https://doi.org/10.5284/1047599 Section 6.5.2).

BUILDING MATERIAL

NEBELIVKA

TALJANKI

MAJDANETSKE

Total no. of houses

1,445 houses

2,200 houses

3,000 houses

Timber for building

28,890m3

52,800m3

78,200m3

Volume of reeds for roofing

22,013m3

33,393m3

45,688m3

Time for thatching

108,338 person-days

164,340 person-days

224,845 person-days

Number of houses (15 × 5 × 4m) required for winter storage of reeds

73 houses

112 houses

152 houses

Volume of chaff for daub- 6,237m3 making

9,460m3

12,943m3

Time for mixing daub (chaff + clay)

76,558 person-days

116,134 person-days

158,890 person-days

Number of granaries/ houses (15 × 5 × 4m) required for winter storage of chaff

22 granaries/houses

31 granaries/houses

43 granaries/houses

Volume of withies required for wattle production

2,578m3

3,912m3

5,350m3

Time for weaving wattle

76,558 person-days

116,134 person-days

158,890 person-days

13 houses

19 houses

328,680m3

449,690m3

Number of houses (15 9 houses × 5 × 4m) required for winter storage of withies BURNING MATERIALS Fuel for burning (timber)

219,048m3

These estimates suggest large-scale advance planning was necessary not only for the collection of such quantities of each material but mainly for its storage, which may have required the construction of special structures, such as granaries for chaff



Megasites – a Comparative Approach 

 475

storage. However, it is to be doubted that storage for such large quantities of building materials could have been provided in any megasite, which could have amounted to over 210 additional storage buildings at Majdanetske. In other words, the scale of resource provision seems at odds with the capacity of the megasite itself. The important implication is that the maximalist figures for megasite populations are far in excess of the maximum possible from resource and labour inputs derived from the complete châine opératoire of house-building and -burning. There is one possible escape from this fundamental conclusion, viz., the coeval use of all houses at the megasites does not necessarily imply a massive burst of coeval construction. The materials estimates suggest the possibility that the building of megasite houses was spread over a considerable time – on a decadal timescale and perhaps over a century. Modelling the labour requirements for 100 years of construction by a Quarter of 200 people means an annual contribution of 10 person-days per Quarter at Nebelivka, 20 person-days at Taljanki and 26 person-days at Majdanetske. Acknowledgement of this construction scenario could bring the maximalist view closer to the alternative views, at least for house-building. But was such a scenario possible given the number of houses built at Taljanki and Majdanetske in the space of two centuries respectively? Millard’s modelling of the ca. 1,500 Nebelivka houses built over 200 years introduced three values for the variable of house-duration: 10 years, 25 years and 50 years (for details, see Section 4.8: summary  – Table 6.7). This produced three estimates for the number of coeval houses, with the maximum of 1/3rd occupancy. Comparable models for Taljanki and Majdanetske produced similar results of 1/3 maximum occupancy if the duration of the sites was 200 years. Increasing the site duration to 300 years would lead to an even lower occupancy. In other words, unless strong AMS evidence is produced that the Taljanki and Majdanetske sites lasted less than 200 years or the houses themselves lasted 200 years, there is indeed a very low probability that there was ever a coeval occupation of all the houses at either site.

Table 6.7: Modelling of the number of coeval houses at megasites (by A. Millard). Site duration

House use-life

Number of coeval houses/total no. of houses, Nebelivka

Number of coeval houses/total no. of houses, Taljanki

Number of coeval houses/total no. of houses, Majdanetske

200 years

10 years

80/1500

115/2200

157/3000

200 years

25 years

210/1500

314/2200

428/3000

200 years

50 years

500/1500

733/2200

1000/3000

476 

 Discussion

Even with minimalist house estimates at these three sites, we are still contemplating a building ‘industry’, supported by provisioning through families or kinship partners, mobilised labour on a large scale and long-term forest management practices which ensured a regular flow of construction materials to the right part of the megasite105. The key point to emphasise here is the overriding need for communal labour at every megasite, with the requirement for recruitment and deployment of skilled workers (carpenters, roofers, wattle-makers), unskilled workers (to gather materials and prepare them for use) and supporting groups (especially cooks) becoming ever more important with the rising scale of construction. The continuity in keeping the communal labour force mobilised and ready to work on the next project meant that every completed task – whether a single house, a Neighbourhood complex, an Assembly House or an entire Quarter – would have been the occasion for celebration of the achievement through feasting and special depositional events. By the same token, communal activity fostered a sense of attachment to co-workers and support groups, to the local place, with its own local identity, and to the overall ‘Nebelivka identity’. All communal activities based upon huge labour inputs would have been underpinned by alliances between Nebelivka and the home communities providing the labour. Given the serious logistical issues at all megasites, it is all the more surprising that the supposed 5-km hinterlands of the three megasites were almost empty of sites, with no sites in the Nebelivka 5-km micro-region, two sites of uncertain chronology in the Taljanki micro-region (Moshurov 2 and 3) and two sites of equally uncertain chronology within 5km of Majdanetske (Talnoe 2 and 3). None of these four sites has produced AMS dates, so the general attribution to the same phase as their nearby megasites by pottery typo-chronology, as well as similar ceramic imports from other regions in Talnoe and Majdanetske (Shmaglij & Videiko 2001–2), does not document coeval dwelling of the kind that would suggest small sites providing logistical support to the megasites. The absence of fieldwalking in the Taljanki or Majdanetske microregions precludes comparisons with the Nebelivka micro-regional survey.

6.2.5 Testing the Alternative Models at Taljanki and Majdanetske It is important to see how the three alternative models developed to account for the building, burning and environmental footprint at Nebelivka work for the other two megasites. While house-building and -burning estimates are available, there is one major obstacle to this procedure – the lack of a fine-grained palaeo-environmental record for either Taljanki or Majdanetske to act as a control over the pace of building

105  Accessing the right part of a megasite is of some significance, given that, for example, Taljanki measured 2.5km from the Northern to the Southern outer rings.



Megasites – a Comparative Approach 

 477

and burning. The alternative has been the use of a building châine opératoire for all three megasites to provide an overall control on labour inputs (see above, pp. 467–468). The incompleteness of the Taljanki plan meant that modelling of only the Pilgrimage and the Distributed Governance Models was possible, while Ohlrau’s (2015) excellent combination of the earlier and the recent geophysical data allows the testing of all three models at Majdanetske.

6.2.5.1 The Assembly Model at Majdanetske The chronological component of the Assembly B Model was five generations, each of 30 years. The division of the Northern part of Majdanetske into 10 Quarters (see Fig. 6.8, based upon Ohlrau 2015, Abb. 52) was modified to combine Quarters 1 and 2 into one, and 9 and 10 into another. The range of all houses except the Assembly Houses (Sonderbauen) in these Quarters was 176–341, substantially more than in the Nebelivka Quarters. This left a total of 924 houses in the remainder of the plan, which were divided into six smaller Quarters (n = 151–159 houses), since there were fewer IRSs in the rest of the site. In each Generation, up to the 3rd, new Quarters were built – five in each of Generation 1 and 2 and the final four in Generation 3. The sequence of Quarter creation was partly dependent on the layout of Ohlrau’s (2015) Quarters. The number of houses in each of the 14 Majdanetske Quarters was scaled up from the house statistics for the 14 Nebelivka Quarters (Fig. 4.5) to produce an equivalent model for Majdanetske. The difference in scale at Majdanetske was that the peak of new building reached 1,079 houses in Generation 3, the maximum number of structures in coeval use (in the same Generation) was 1,316–1,383 and the peak of house burning was reached in Generation 4 at 585 houses. If future palaeo-environmental research in the Majdanetske hinterland produces a fine-grained pollen diagram with the same minimal human impact factors as in Nebelivka core, then Assembly Model B would fail because of the peaks in building and burning. Even if human impacts were shown to be higher at Majdanetske than at Nebelivka, the footprint of over 1,000 new houses in 30 years would have required 6,500 person-days per annum to process reeds, chaff and withies, not to mention the collection of timber and its shaping into a house structure. It is highly probable that the creation of open landscape recognised in the later stage of the Majdanetske sequence (Kirleis & Dal Corso 2016) was caused by the building programme, which must have placed extreme stress on the landscape of the Majdanetske hinterland. If such a degree of stress is deemed acceptable, it is possible that the trajectory of Assembly Model B would have been logistically possible at Majdanetske. However, one problematic aspect of the Majdanetske plan concerned the whittling away of the inner open space from a potential 114ha, as measured from the dominant Outer Circuit, to 78ha, as measured from the dominant Inner Circuit, to a mere 26ha after factoring in all of the IRSs and Inner Circuits (Fig. 6.8). The performative aspect of the inner open space centred on a dramatic and impressive open space bigger than

478 

 Discussion

any of the participants had ever seen before. The effects of this infilling on the central meeting area within 60 years – or living memory – of the start of Majdanetske would have been to diminish the core area of the megasite, perhaps even posing a threat to the central function of the site. It would appear that Majdanetske became a casualty of its own success! The creation of additional outer circuits outside the dominant Outer Circuit may well have been a way of coping with problems caused by the infilling of the inner open space. The Pilgrimage Model in the form developed at Nebelivka required a large burst of building activity in the first two years, as well as a major ditch-digging programme. After this intensive phase, the pace of construction dramatically slowed, with often no more than 200 houses built per generation – coincidentally the length of time for the regeneration of deciduous woodland. Thus, the main threat to the Pilgrimage Model concerned the potential over-exploitation of forest resources in the intensive building phase – a threat that the model suggested was contained. What would have been the consequences at Majdanetske for an intensive building and digging programme in the first two years of the megasite?

6.2.5.2 The Pilgrimage Model at Taljanki and Majdanetske The incompleteness of both the Taljanki and the Majdanetske plans led to the estimation of the total number of houses in each main Inner Circuit. At Taljanki, 392 houses were found in an estimated 60% of the Inner Circuit (Ohlrau 2015, Abb. 49), leading to an estimated total of 650 houses in the whole circuit. An initial construction time of two seasons would have meant labour input of 85,000 person-days  – a heavy demand even for a megasite. It is thus suggested that three years of initial construction, with an input of 57,000 person-days per season, would be a more reasonable model. The building of up to 40 houses per month would have necessitated close collaboration between two or more pilgrim home communities. The principal elements of the Taljanki plan would have fitted the Pilgrimage Model just as well as the Nebelivka layout, with two exceptions. The absence of a perimeter ditch would have removed from the pilgrimage centre a critical defining characteristic – what separated the centre from its outside world. In effect, the house circuit at Taljanki became the perimeter line and the first construction feature, probably limiting the main procession to inside the house circuit. The second absence at Taljanki concerned the Assembly Houses, which at Nebelivka were considered to be important elements of processional practices, even though not built until the second generation. The implication is that the organisation of the processions was carried at the site level or at the Quarter level. Neither of these absences would have prevented the development of a pilgrimage centre at Taljanki in what should be considered as a rival establishment to Nebelivka. The estimation of the number of houses in the principal inner house circuit at Majdanetske is complicated not only by the incompleteness of the geophysical



Megasites – a Comparative Approach 

 479

plan but also by Ohlrau’s (2015, Abb. 27) incorporation of Dudkin’s earlier and less precise geophysical results. The counting of both Rassmann’s and Dudkin’s housesized anomalies produces a total of 267 houses for the inner circuit but this is almost certainly an under-estimate. Scaling-up the total of recent houses to a complete plan suggests a higher total of 350 houses, which is the estimate used in this model. The total of 350 houses required to be built in the initial intensive construction phase means an estimated total of over 90,000 person-days in an eight-month season. This task is certainly excessive even for a megasite, leaving the choice of a two-year period, with over 45,000 person-days and up to 22 houses built per month, or over 30,000 person-days for each of three years, with up to 11 houses per month. Both of these options seem possible but the three-year option may be preferred, given that additional labour input would have been needed for the perimeter ditch. An estimate of the length of the many separate perimeter ditch segments can be made, using the 47% of the ditch present (Ohlrau 2015, Abb. 27), with the result that digging a total length of 273m, or 300m3 of earth, was required. Much of the clay excavated from the ditch segments could have been used immediately for daub-making for the houses in the house circuit. The two principal issues with the Majdanetske plan for the Pilgrimage Model concern its evolution into a very complex layout with many cross-streets (inner circuits) blocking the processional ways of the IRSs, and the infilling of the inner open area which was central to the main pilgrimage ceremonies (see above, pp. 444–445).

6.2.5.3 The Distributed Governance Model at Taljanki and Majdanetske The DGM is well suited to both Taljanki and Majdanetske in terms of a permanent, if scaled-down population and a planned layout incorporating elements of earlier megasites. The model assumes a more or less static number of houses, with a constant rate of house-building and -burning throughout the duration of the site. Testing the model involves setting the initial building rate and the number of builds and burns per annum against the estimated duration of the site and the likely capacity of the site for achieving the initial build. Given the total number of houses at Taljanki as 2,200, two values were tested for the initial build – 400 houses and 500 houses – and two values were used for the number of houses built and burnt each year – eight and ten. The four calculations produce variations of site duration between 170 and 225 years – all of which are compatible with the modelled AMS date range (see above, Millard, Section 4.8). The construction of 400 or 500 houses in one major constructional phase would undoubtedly have tested the builders of Taljanki to the limits unless the constructional phase was spread over five years. The much larger number of houses at Majdanetske, at 3,000 structures, led to an extremely high initial build of 800 houses and two values for per annum housebuilding and -burning. The building and burning of 15 houses per annum led to a

480 

 Discussion

modelled duration of 150 years, while a decrease to 10 houses per annum extended the duration to 220 years, with both durations being consistent with Millard’s range of modelled AMS dates. However, an extension of the site duration to 300 years meant a lower initial build of 600 houses, with eight houses built and burnt per annum.106 Alternatively, ten houses could have been built and burnt per annum if the site duration was 240 years. The initial build of 800 houses is problematic insofar as it represents the highest total of houses to be built in a short time of any of the variants of the three models. The reduced initial build of 600 houses, perhaps spread over three years, means that the DG Model could have worked at Majdanetske.

6.2.6 Summary It is a platitude to note both similarities and differences between the three largest megasites or that there were both continuities and innovations in a temporal sequence. The effect of the new AMS modelling of the three sites (Dal Corso et al. 2019; see above, Millard, Section 4.8) is that, with high probability, the three sites were in coeval occupation for some decades. This leads to the possibility of a more dynamic interpretation of these sites, not least in planning terms. Before we attempt that, it is worth turning to the comparison of basic daily resources and resources for house-building and -burning. Despite differences in animal bone recovery, the faunal spectra of all three megasites show broadly similar patterns, with a low percentage of wild animals and a dominance of domestic cattle over caprines107 and few pigs. Equally, the botanical reports for Nebelivka and Majdanetske show a similar range of wheats and barleys, with some pulses and wild fruits of the forest. What remains surprising is the lack of extensification or intensification of arable farming at either of the megasites but rather a continuation of the same low-efficiency agriculture found from Phase A onwards (Pashkevych 2012). The low representation of the only six-row bread wheat found at the megasites – T. aestivum – is particularly striking in terms of the neglect of an opportunity for increased productivity. Only fieldwalking in the Taljanki and Majdanetske micro-regions would confirm or deny the absence of manuring scatters as established at Nebelivka. Low-efficiency agriculture would seem to be a counter-intuitive strategy for some of the largest settlements in Eurasia. A major difference between Nebelivka and Majdanetske was the development of chernozems before the megasite occupation at the former, while a human-induced chernozem is claimed for the late stages of the latter, representing a change from fertile loams in parallel with forest clearance of the forest steppe to a more open landscape.

106 The 300-year duration of Majdanetske has been confirmed by the most recent Bayesian modelling (Dal Corso et al. 2019, 3). 107  Orton et al. (2016) (Section 5.3) show inter-house differences in preferences for cattle or caprines.



Megasites – a Comparative Approach 

 481

The last, essential food resource presented in a comparative manner was salt. The estimates for salt procurement for varying scales of population size at Majdanetske show the need to transport massive amounts over hundreds of km, using basic technology (cart or sledge). Such estimates are far too high to be feasible for the megasites, leaving an insoluble logistical conundrum. The same type of insoluble conundrum has been posited for all four of the main building materials for houses  – timber, reeds for roofing, chaff for temper in daub and hazel withies for wattle-making. Moreover, Kruts et al. (2001) concluded that exhaustion of the forest for cooking and heating fuel would have limited the megasite occupation to only 50 years. Neither Kruts’ nor Ohlrau et al’s (2016) estimates for house-building have ever taken into account the much larger quantity of timber needed to burn the houses to create a ploshchadka, as demonstrated in the Nebelivka house experiment. The huge scale of collecting reeds, chaff and withies in a single mass building programme, not to mention the importance of their winter storage, was far too great for even a megasite. Even taking the Nebelivka minimalist models into account, we are still talking about a large-scale building industry, combining specialist joiners and roofers and long-term woodland management to ensure a supply of all the key resources when necessary. But the overall conclusion from the resource study is that populations of more than 10,000 people on a megasite created insoluble logistical problems for basic resources. How do these conclusions relate to the comparative study of megasite planning? The partially overlapping occupations of the three largest megasites create an opportunity for interpretation in terms of major structuring and planning principles. The earliest megasite of the three – Nebelivka – shows a combination of acceptance of the main planning principles together with a bottom-up growth at all levels of aggregation – the typical nested pattern of households, Neighbourhoods and Quarters (for the cross-cultural significance of the nested spatial levels, see below, pp. 501– 502). Assembly Houses were well integrated into the Quarters and the only ‘industrial’ feature has been interpreted as a large-scale cooking facility for communal feasting. We propose that the combination of an almost total absence of Assembly Houses with a large number of kilns at Taljanki meant the increased importance of Limited Interest Group potters who built, managed and supplied the kilns for production of fine painted pottery. It is suggested that the LIGs led to the mobilisation of Neighbourhood labour in such a way that this by-passed the Assembly House structure. The people who started settling at Majdanetske, therefore, had a choice of two models of Quarter-ly organisation – the Nebelivka model with Assembly Houses but no kilns, and the Taljanki model with LIGs to organise kiln production and which replaced the Assembly Houses. The intriguing aspect of the Majdanetske plan is the occurrence of both forms of model in the same site, suggesting that people with knowledge of both social models settled at Majdanetske. Many aspects of the Majdanetske plan indicated tensions between the two Quarter-ly models, while other architectural responses to tensions in settlement layout are evident. If Taljanki took the classic megasite plan

482 

 Discussion

of Nebelivka and enlarged it, then Majdanetske was an example of the fragmentation of the classic model through the local politics of planning and building what is the most complex megasite plan yet known. Without fine-grained palaeo-environmental data from near Taljanki and Majdanetske, it is impossible to make a direct test of the three alternative models formulated for Nebelivka. While all three models could conceivably conform to the footprint of the respective layouts, progressive infilling of the inner open space at Majdanetske threatened the Assembly and Pilgrimage models at that site but not necessarily the Distributed Governance Model. There is no reason to suppose that Taljanki and Majdanetske followed the Nebelivka model of large-scale communal practices but the scale of the housebuilding and -burning activities makes the maximalist model highly improbable, if not downright impossible. This conclusion leaves the megasites as something of a paradox  – massive planned layouts in the landscape but far fewer people living there than had once been accepted. What does this paradox imply for the notion of megasites as the world’s first cities?

6.3 Low-Density Urbanism – a Global Approach The first of the Project’s fundamental questions concerns the urban status of Trypillia megasites such as Nebelivka (Chapman & Gaydarska 2016a). In this section, we develop a fresh, relational approach to the question of what constitutes ‘urbanism’ by adopting a methodology from the social sciences and recently summarised by Cartwright and Runhardt (2014). But before we can turn to the characterisation of megasites using this methodology, we must answer the key question  – ‘urbanism relative to what?’ through a comparative examination of the typical small Trypillia site of Grebeni. The scalar differences pinpointed between Grebeni and Nebelivka are then further discussed for the Trypillia megasites of Taljanki and Majdanetske. Only then do we turn to the question of the place of Trypillia megasites in global perspective, comparing them to other huge anomalous sites and grounding them in the perspective of low-density urbanism.

6.3.1 An Example of a Small Trypillia Settlement If Trypillia megasites were a little later or had been in an area with a tradition of urban development, it would not have been difficult to argue that such agglomerations were inter-regional centres (Ryndina 1998), whose function in hindsight modern scholars call ‘urban’ in character. Their space-time location and lack of legacy are the megasites’ worst enemies. More efforts are spent to portray them as ‘normal, only bigger’ and to undermine the implications of what their size entails rather than to see them for what they are – the earliest low-density massive occupations of urban status.



Low-Density Urbanism – a Global Approach 

 483

So far, we have been arguing against an idealised, essentialised view of urbanism and insisting on a relational view (Gaydarska 2016, 2017). Here we start with a different approach. We ask the question ‘how are the lifeways on a megasite consistent with Neolithic lifeways? At first sight, such a question falls straight into a Childean Urban– Neolithic opposition. However, as pointed out earlier, his two revolutions are often conflated in the theoretical stances of Trypillia specialists. This is because, strictly speaking in Childean terms, the megasites are neither entirely ‘Neolithic’ nor are they entirely ‘Urban’. They are also an exception in R. Fletcher’s (1995) global model of settlement growth. At this point, the discussion could have taken one of two routes. The first assesses the nature of the various classificatory systems and how and why the Trypillia megasites fall between their cracks. In the second, preferred pathway, we compare the life on a small, 4.5ha site (e.g., Grebeni: Kolesnikov 1993) at the beginning of the 4th mill. BC in Ukraine which may be considered as an archetypal Neolithic site, with life on a 238ha megasite of the same period. We are currently uncertain about the overall area of the Grebeni site, as well as about the total number of dwellings (Fig. 6.11). If we take the maximum figures of 4.5ha and 38 dwellings, that will mean a residential density of ca. 9 houses per ha on a larger than usual dwelling place, given that 55% of sites with size information in the Trypillia Encyclopaedia (Videiko 2004) fall within the size range 0.3h–1.0ha (using a sample of 499 sites: Nebbia 2016). The circuit layout may account for this occupational density, which may have been a deliberate choice. Such a residential density leaves room for small gardens and/or pens near each of the houses, with additional fields for cultivation and grazing beyond the settlement boundaries. Even the most distant arable areas would not have been more than half an hour’s walk, with communally organised pasture for domestic animals. The household scale of animal-keeping of five caprines and two cattle would have meant that the flocks of fewer than 200 caprines and ca. 80 head of cattle would have been a responsibility of two small groups of experienced herders. It was unlikely that a total of 38 dwellings would have been divided into Quarters but five to ten Neighbourhoods were probable (Fig. 6.11). Coeval construction of all the buildings in one building project would have taken two months of intensive building by almost all the site inhabitants. Alternatively, the settlement was constructed over a longer period or with the additional help of skilled workers from related communities. In either case, even the most intensive operation would not have involved more than 420 people  – a large enough but manageable number considering that no more than 10–12 people would have participated in a single house construction. Given the social undesirability of living on a settlement with strangers, such an operation would have strengthened the already existing ties between neighbours and future occupants. Members from each Neighbourhood would have seen each other daily, while inter-Neighbourhood encounters were, if not a daily, then a weekly event. And when a member of this community passed away, a gathering of mourners would have been

484 

 Discussion

Figure 6.11: Geophysical plan of Grebeni site. Key: 1: excavated buildings; 2: test pits; 3: geomagnetic anomalies (by B. Gaydarska, based on Koshelev 2004, Ris. 4.14).

materialised by the deposition of feasting remains and other objects and materials in a pit. If the deceased was an important member of the community, this may have been accompanied by a spectacular symbolic burial through a house-burning performance. This was a largely self-sufficient community, with occasional encounters with relatives living at other settlements, with travellers/traders passing by their site en route to an exchange centre and with members from the wider Trypillia community at regional gatherings. In other words, this was an archetypal Neolithic way of life (cf. Childe 1958). Now imagine that the site of Grebeni is smaller than the smallest of the Nebelivka Quarters (Quarter E, at 5.3ha). Scalar contrasts of this level make us appreciate the stark differences between megasites and 'normal' small Trypillia settlements. 6.3.2 An Analytical Construct for ‘Urban’ Sites If there is a common feature between the various definitions and understandings of what is urban, it is that this form of dwelling is somehow different from other and/or



Low-Density Urbanism – a Global Approach 

 485

contemporary forms of dwelling. The nature of this difference (e.g., between Neolithic and Urban in Childean terms, and between city and village more generally) is what causes endless definitions, re-definitions and discussions. If this difference is not to be essentialised, as it often is, its characteristics could be very useful indicators for the local variants of the modern analytical construct of urban. In 2016, the authors of two articles both argued against the fixed concept of cities (Gaydarska 2016; Smith, M. E. 2016). Both echoed Cowgill’s (2004, p. 543) inspirational plea to think about alternatives to typologies of cities by introducing measurable axes/ variables along which specific cases can be arranged in multivariate space. While Gaydarska cautiously offered a non-prescribed range of variables and insisting on the development of the methodology of measurement in archaeology, Smith expounded his analytical construct using 21 traits measured in one of three ways  – presence/ absence, according to scale and in absolute numbers (Smith, M. E. 2016, p. 159 & Table 10.1). Smith calls his approach ‘attribute-based’ and although he warns that this is a provisional list of attributes, it is a list, nonetheless, that together with the cut-off point of 12 traits could easily resurrect the ghost of check-list definitions. Gaydarska calls her approach ‘relational’, putting the emphasis on local contexts. At first sight, the measurement methodology she seeks is covered by Smith’s attribute-based table and discussion. Indeed, there is some overlap between Smith’s change of emphasis when asking questions about ancient cities (2016, p. 160) and Gaydarska’s conclusion that the definition of cities depends on the research question asked (2017, p. 182). However, we feel that a better framework for answering these questions is provided by the Cartwright and Runhardt (2014) approach to measurement advocated by Gaydarska, as it allows for both a sensible degree of fuzziness and the social construction of concepts. Nonetheless, both Smith’s attribute-based approach and the Cartwright & Runhardt approach make use of the polythetic nature of urban data sets. The outlines of the Cartwright & Runhardt approach are presented above (see Section 2.2). Here, the actual measurement – just as in Smith – includes presence/ absence, according to scale and in absolute numbers but for all three components (characterization, representation, procedures) rather than just one or two. Table 6.8 summarizes the three components for the category ‘urban’ in the Trypilian context. If the characterization (and hence representation and procedures) is extended or shrunk, it still will be socially constructed but may become too general, or create boundaries. Provided that these two caveats are avoided, there remains the key question ‘is the revised characterization useful for its purpose?’ In other words, there is no prescribed number of factors for characterization, as long as each is consistent with the four principles of utility, social construction, a lack of rigid boundaries and a lack of generality. Ideally, the entire Trypillia dataset should be assessed according to Table 6.8 in order to see which and how many sites would fall into the category ‘urban’. At present, there are too many gaps and inaccuracies in our data set to allow such an endeavour. Instead, we hope to provide an example of the process.

486 

 Discussion

Table 6.8: Constituent parts of the ‘urban’ category in the Trypillia context. Characterization

Representation

Procedure

Territory/area to which a site is central

More links in a network analysis

Formal network analysis

Distribution of contemporary sites

Field survey, remote sensing, ground-truthing, spatial analyses, Moran test (whether megasites are outliers)

Formalized space where interaction geophysics could take place Size (although in absolute Settlement planning numbers, scalar in meaning- after which point size becomes an obstacle/ issue of logistics and infrastructure)

Geophysics

Population number (like Number of contemporary houses size – absolute in number, scalar in meaning)

AMS dates, palaeoenvironmental analysis

Population heterogeneity

Various and overlapping identities (e.g. different clans, permanent dwellers vs visitors/traders, members of the same guild/trade – potters, metalworkers, etc.)

Excavations of wide range of features; depositional analyses; detailed artefact and ecofact analyses

Concentration of skilled labour and management

Limited Interest Groups (e.g. potters, Field survey, geophysics, metalworkers), workshops, kilns, excavations, palaeospecial buildings, precision of environmental analysis, settlement layout, planned and depositional analyses, detailed managed resource utilization, waste artefact and ecofact analyses. management

Built environment/ Assembly houses, fortification, Geophysics, excavations, formalized space with non- ditches, large cooking and/or palaeo-environmental analysis, quotidian function storage facilities, large delineated depositional analyses space for assemblies, pilgrimage or other social gatherings Scale of subsistence (self- Agricultural fields, animal bones, sufficient or combined logistics and infrastructure (self-produce plus tribute/ sponsorship))

Field survey, geophysics, palaeo-environmental analysis, osteological analysis

Affordance to be a node and re-distribution centre of far-reaching exchange networks

Excavation, detailed artefact and ecofact analyses. technological analyses

Wide range of exotic (rare) and imported (bulk) objects and raw materials, non-local mode of production

Social framework that Intra-site: nested social structure of Theoretical justification, spatial allows such sites to function household, Neighbourhood, quarter. analyses, depositional analyses, Inter-site: shared symbolic fiction detailed artefact and ecofact (Big Other) analyses



Low-Density Urbanism – a Global Approach 

 487

The representations and procedures are often overlapping (e.g., a formal toolkit made of exotic raw material may represent both a wide-ranging exchange network and skilful labour; excavation and geophysics feature in almost all procedures). Not all of the procedures were followed when measuring the data from Nebelivka (e.g., network analysis). Also, not all measurements were possible (e.g., the affordance for wide-ranging exchange networks), since such measurements are scalar and the comparison data are only selectively available. These are all tasks for the future. Still, there is enough data allowing the measurement of the category ‘urban’ in the Trypillian context at Nebelivka. We now turn to a discussion of each of the nine characterizations. Territory/area to which a site is central Nebbia’s analyses (Chapter 3) of the wider settlement pattern established the social catchment area of Nebelivka at 100km. Since centrality is not measured by outdated Thiessen polygons but by the intensity of interaction possibilities, a future network analysis may confirm Nebelivka as a central place. At present, such a status is argued on the basis of two points: (a) a LISA (Local Indicators of Spatial Association) test supported the hypothesis that megasites are outliers and provided further confirmation that it is highly likely that these are outliers of high values within a 100km Neighbourhood of low values, indicating an overall even spacing with regional regularities of settlements attached to each megasite; and (b) specifically for Nebelivka, the creation of an initial 105ha central empty space implies a scale consistent with being able to accommodate large numbers of people. A LISA test would give no sense at all of a site the size of Grebeni as being central to a settlement network. Size By the time Nebelivka emerged, there was perhaps a social memory and practical knowledge of up to five previous 100–150ha sites, although the only site with a geophysical plan (Vesely Kut) lacked a large inner open area. The full extent of Nebelivka reached 238ha, while, through time, the open inner space shrank to 65ha. It is very important here to remember Barrett’s cautionary note (1994) that the final footprint of a site does not necessarily reflect its initial appearance. The Nebelivka perimeter ditch outlined the maximum extent of the site, perhaps at an early or even initial stage, thus declaring an intention to scale up previous large gatherings by 50% to 80%. In absolute terms, Nebelivka was the largest site in the Trypillia world at that time. If Nebelivka was permanently occupied, its daily interactions would have encountered serious communication and logistical issues, especially if the habitus with which residents were familiar was formed at settlements that were 200 times smaller. If the occupation was seasonal, broadly similar interaction issues would have arisen, if the normal scale of the most distant neighbour was 500–600m, rather than 1.5km, away. Thus, in relational terms the size of Nebelivka provided a very different

488 

 Discussion

experience for its inhabitants and visitors, as compared to life on a small Trypillia site such as Grebeni, with a probable site size of 200m × 200m and therefore furthest neighbour distances of no more than 200m. Population numbers How many people lived on a megasite is and will remain a contentious issue. It is clear that the potential of large sites to accommodate hundreds of people is many times greater than any of the smaller sites. It is unlikely that such potential was not used for some or all of the time. After all, megasites were probably created to host more people than usual. And, just as with size, living with 100 people is different from living with 500 people, and even more different still from living with 3,000. Colombijn (1994, 1) observes that, in urban sites, the larger the number of people, the more complex is the use of space in comparison with villages. The plans of Nebelivka and Grebeni confirm this observation, showing the complex nested spatial order which characterises Nebelivka in comparison with a single house circuit and houses both inside and outside the circuit accounts for the structures at Grebeni. At present, the radiocarbon dates from Nebelivka are inconclusive about the contemporaneity of houses but modelling of the overall duration of the site (150 to 200 years) with various house durations (10, 25, 50 years) suggests that if the houses lasted for 50 years, the maximum number of coeval dwellings would be 500 for periods of 10 years. The three models for Nebelivka operate with a range of 400 to 700 contemporary houses that is broadly comparable with such estimates, given the inherent probability of variations in house duration. In any case, both the social modelling and the radiocarbon modelling are in good agreement with the palaeoenvironmental data for low to moderate human impact and the lack of massive fires to account for the contemporary burning of hundreds of houses. We have accepted that six to eight people occupied each house (for demographic considerations, see Gaydarska, submitted). The structure of a megasite based upon Quarters and Neighbourhoods helped to reproduce ‘Grebeni-type interactions’ at the smallest settlement scale but the multiple opportunities for meeting people from other home communities (‘strangers’) meant much stronger links between individuals from different communities than could ever have occurred on smaller sites. The regular scale of face-to-face contacts on settlements rarely exceeded 450–500 people (Forge 1972). While everyone could potentially have seen or met everyone else, people’s interactions were probably channelled so that there were some more regular and other less regular meetings, introducing heterogeneity into the range of social contacts. Population heterogeneity Who were the people living or visiting Nebelivka? It is unlikely that they all belonged to one extended social group, whether a clan or a lineage, since a massive congregation of people of common descent at one place could have created a threat to other clans that may even have escalated to conflict. In any case (see above, pp. 25–26),



Low-Density Urbanism – a Global Approach 

 489

the evidence for warfare is minimal. Thus, our assumption is that varied groups of people from many home communities came to live at, or visit, Nebelivka. The complex relationship between material culture and group identities can be teased out at Nebelivka using our analyses of architectural variability, modes of deposition, consumption of animals and the stylistic analyses of pottery, figurines and lithic tools. The counterpoint to the overall Nebelivka plan, with its grand planning principles of concentric house circuits, radial streets, a perimeter ditch and a vast inner open area, was local variability in the size and shape of houses and Assembly Houses as well as the layout and size of Neighbourhoods and Quarters (Table 6.1). The most obvious explanation for this ubiquitous variability was the creation of local versions – hardly copies – of home community architecture and layouts as a way of adjusting to a settlement scheme on a far greater scale. It was the juxtaposition of so many different house groupings at the megasite that differentiated it from a single home community such as Grebeni, which itself may have made a small contribution to Nebelivka’s architectural heritage. A similar point can be made concerning deposition which, at Nebelivka, showed scalar differences from the household deposition in ‘events’ at a single pit, characterised by mostly small consumption units, through a range of medium-sized deposition events in which most or all of the houses in a Neighbourhood offered goods to a death assemblage before the burning of a house, up to the contribution of many houses, if not all houses in use at the time, to the mega-event of the burning of the Mega-structure, in which an estimated 322 vessels, or parts of vessels, including a group of 21 complete miniature vessels, were placed in rooms and outdoor areas in a multi-phase destruction process. These scalar differences betokened a nested set of depositional contrasts, moving up from single house events to whole-site events. Once again, it is clear that sites of the size of Grebeni could have contributed only to the lower part of the depositional scale. The Nebelivka faunal assemblage comprised sub-assemblages from different houses, one pit and the Mega-structure. The bone remains show a pattern of different meat preferences, with beef strongly selected in House A9 in contrast to more lamb/ mutton in the Mega-structure and some of the Pits. These contrasts may refer to the latest set of meals in the house contexts and different emphases in feasting remains in House A9, some Pits and the Mega-structure. In each case, the size of the animals and the quantities of meat available for consumption tell a different story about various contexts in the megasite. However, we should note that such variations in household consumption may also have occurred at smaller sites such as Grebeni. Pottery is the commonest material relating to the identity of the people and potentially demonstrating whether Trypillia sites were exclusively inhabited by separate descent groups. However, Trypillia pottery studies are currently dominated by the typo-chronological approach and one is well advised to stay clear from relating such time-space entities to people (Shennan S. J. 1989). The positive aspect of such pottery studies is that they demonstrate that certain ceramic styles are shared between

490 

 Discussion

sites, rather than exclusively related to one, with varying preferences of ceramic styles selected at different sites. Instead of using the similarities and variations only to build typo-chronology, we can view them as choices that people make from a wide range of available forms and decorative schemes that suit best their aesthetic and personal preferences. Only in that sense can we accept that variations in ceramic production bore traces of group identity. It is unfortunate that intra-site variability on Trypillia sites is difficult to prove or disprove due to selective publication of shapes and decorations in typo-chronological schemes rather than assemblages according to excavated units (e.g., household assemblages). As the pottery analyses at Nebelivka have shown, even incomplete ceramic assemblages from a small part of an excavated house (up to 3%) can provide a good basis for intra-site comparison. The differences in forms, fabric and decorative motifs may have varied through the passage of time but these variations had social referents which require explication. This is one further facet of the heterogeneity characterizing material culture at Nebelivka – one that was highly improbable at a small site such as Grebeni. The variations in the small lithic assemblage were primarily related to three variables - the use of higher-quality exotic flint or lower-quality local flint, different châines opératoires and the presence of various individual knapping strategies for the production of projectile points. All of these causes of variability can be related to household lithic knapping rather than centralised production, with the acquisition of exotic Volhynian flint by some households only leading to different cultural narratives connecting Nebelivka to the outside world than for houses with access to local flint only. By contrast, the fragmentary figurines at Nebelivka showed little variability in form, breakage and condition, with the exception of changes in the number of fragments deposited in pits and houses. This finding supports Orphanides’ (2010) contention that figurines of similar form mediated community agreements on important concepts of being or place in the social order. It is also important to note that the density of figurines per site as a whole, and the sub-group of realistic figurines, hardly varied with the size of the site or its date (Gaydarska 2019). This finding reinforces the notion of the Trypillia Big Other as a stable entity functioning in similar ways on megasites and on small sites such as Grebeni. This tells us something important about the Big Other: the replication of practices deriving from acceptance of this framework for life on all Trypillia sites confirmed the Big Other as a strong force for social integration which transcended even the differentiation of practices on small and large settlements. Concentration of skilled labour and management An important aspect of personhood is the range of personal skills and abilities that individuals develop over their life-course (see above, Section 5.2.1: Chapman & Gaydarska 2011). Very different persons can be defined for their closeness to one of the extremes – a person with high skills in one area whom we may call a ‘specialist’ or a person with a broad range of low-level skills in a range of different areas (roofing,



Low-Density Urbanism – a Global Approach 

 491

flint-knapping, grinding), such that we may term them a ‘generalist’. Generalists would have been more typical, and more useful, in small sites such as Grebeni, whereas megasites with such large populations may well have included specialists in many households. Although, at first sight, the absence of kilns and other workshops at Nebelivka would suggest few concentrations of skilled labour, there is a lot of indirect evidence for skilful production and careful management. We have already suggested the presence of Limited Interest Groups for several types of production. First, nearly 1,500 households would have required an enormous amount of pottery. The local production of at least some of pottery, as shown by raw material analyses (Chapter 4.9), reveals that Nebelivka is midway between a small Trypillia site with its household-based pottery production and Taljanki and Majdanetske, with their much larger scale of kiln-based production. The graphite painting of typical decoration motifs on three special miniature vessels certainly betokens highly skilled work by a non-local. This may indicate movement of skilled labour attracted by the reputation of Nebelivka which would hardly have been possible at the Grebeni site. Secondly, the variety of styles for producing projectile points shows a range of highly-skilled flintknappers at the megasite. The higher degree of weapons standardization as a good predictor for greater incidence of warfare (Brill 2016) confirms the low significance of warfare in Trypillia communities. Thirdly, the construction of Assembly Houses, that, at least until very recently, is typical only for megasites, is an unique example of either off-site modular building and on-site assemblage or a still unidentified method of on-site building and burning that leaves a very distinctive footprint. Larger than usual houses are known from the inception of the Trypillia-Cucuteni network (e.g., Baia: Hofmann et al 2016; or Olexandrivka: Videiko 2005). They are rare and their position within the overall settlement planning is unclear but there is a tendency towards a central location. Among the pre-BII large houses, so far there is only one example at Adâncata – Dealul Lipovanului in Moldova (Hofmann et al 2016, Fig. 16) which displayed the construction features of future Assembly houses but not their specific location in relation to the house circuits. Hence, this is an additional plan element which, through bricolage, was later included in planned BII megasites such as Nebelivka. A vital inter-personal skill for the megasite was the ability to plan and manage the people and the resources on which the site was dependent. Many facets of woodland management came into play at Nebelivka, including an understanding of forest regeneration, the coppicing of hazel, possibly planting and most certainly maintenance of trees and reeds. Procurement of chaff and clay (for pottery as well as for house-building) and drying of wood/withies would have required planning and storage. This ability also concerned the temporal scheduling of tasks completed by different groups (e.g., the coppicing of hazel should have been started five years before the start of house-building) and planning for the availability of tools and facilities before a task was started (e.g., the making of quantities of antler picks and scapula shovels before the digging of a stretch of the perimeter ditch). These skills

492 

 Discussion

were no small matter, in view of the staggering number of people needed to build the site, collect the fuel for daily needs over 200 years and assemble the timber to burn over 1,000 houses (Table 6.6). None of these skills would have been necessary for the development, building and running of Grebeni with its 38 buildings, provided that the builders had access to a sufficiently large area of forest to search for hazel withies. If Grebeni was close to other settlements, woodland management may already have been in place, enabling relatively easy exploitation. The advantage of planning woodland utilization two or three years in advance for even a small new settlement would have been the provision of materials close to the site, with a marked reduction in transport requirements (S. Johnston, pers. comm.). Finally, we should not forget that the agglomeration of so many people and animals would have generated a large quantity of daily waste that also needed to be managed, especially to conserve local water quality in the Nebelivka wetlands. The same problems of waste management would hardly have existed at small sites such as Grebeni. The built environment/formalized spaces with special functions One of the most skilful operations in Nebelivka was the vision and precision needed to achieve the layout of a regular site plan, especially in the face of the variety of grass-root opinions. The Nebelivka promontory offered a gently sloping area of almost 300ha. The intervisibility of most parts of the megasite from the central zone facilitated the layout of a site over 1.5km in length from North to South. Although concentric arrangements and large settlements were known before the emergence of Nebelivka, they were rare and their plan elements were not juxtaposed at such a scale. The irregular nature of the settlement planning data of the only earlier 100+ ha site with a geophysical plan (Vesely Kut) justifies the claim that Nebelivka represents a novel social and spatial concept that combines dwelling areas with a place for massive assemblies. Palaeo-environmental data indicating rising levels of cereal pollen (Section 4.1) suggests that the place where the future site would appear was settled before and may even have hosted gatherings. Formalizing the inner open area through enclosure with a house circuit was a conceptual breakthrough enabling the integration of an assembly function with a novel scale of dwelling. This is not to deny that rituals, ceremonies or structured deposition took place on small sites like Grebeni – rather to suggest a limited scale of feasting and exchange of exotic items and low-bulk, high-value materials. Megasite assembly for up to 3,200 people would have been a major event in the regional calendar, where people from many sites could meet each other and celebrate their belonging to the Trypillia way of life in inclusive performances. The creation of a distinctive form of large, public structure  – the Assembly House – also differentiated megasites from most, if not all, small Trypillia settlements. Although the excavated Nebelivka example  – the so-called ‘Mega-structure’  – was



Low-Density Urbanism – a Global Approach 

 493

not a place for extravagant deposition (see Section 4.5.1), finds such as the gold hairornament and the fragments of the gaming board point to special practices found rarely on megasites, if at all on smaller sites. Moreover, the formal roles and distinctive locations of the Mega-structure and other Assembly Houses at Nebelivka were quite different from the sole example of a pre-Phase BII Assembly House. The scale of subsistence This aspect of characterization depends entirely on the number of inhabitants on the site and on the type of occupation  – whether seasonal or permanent. In any case, the organization of subsistence practices at a megasite would be different from that of Grebeni. Starting with the Pilgrimage Model, growing crops and keeping animals for the small group of site guardians seemingly reproduced the small-scale economy of a standard Neolithic site. However, key pilgrimage practices of ritual and feasting necessitated larger-scale production of food and drink, that in turn would have been replenished by pilgrims’ gifts. Such a gift-oriented arrangement led to risks of overproduction (superfluous giving) or under-supply (supply issues), with experience gained over the years mitigating these risks with forward planning and storage. In that sense, the apparent lack of storage vessels at Nebelivka is intriguing but we should not forget that containers could be made of perishable materials. More importantly, the ceramic assemblages at Nebelivka were not ‘living assemblages’ but, for the most part, the results of deliberate deposition. The Assembly Model would have had a similar arrangement – small-scale mixed farming for the permanent occupants, topped up with provisions and donations by the visitors catering for their own consumption but also planning for trade, exchange and gifts. The Distributed Governance Model relied on a different principle of a social pact to which the supply of off-megasite resources was essential. A wide support network assured the constant or seasonal flow of foodstuffs to Nebelivka, as the burden was shifted annually among the participating social groups. There were enough contemporary smaller sites in the 100km catchment of Nebelivka to have sustained both themselves and also the permanent occupants of Nebelivka (maximum 3,200 people) over a 10-year cycle (Gaydarska, submitted). A key element in each of the Models was the provision of live animals to Nebelivka as a contribution to subsistence and/or feasting. Here is a good example of the agency of animals, whereby one 400kg bull or a 300kg cow would cover the contribution of a home community to a megasite pilgrimage for one season. It has now been recognised that the slaughtering of animals was a weakness in any household economic model which emphasised separate household identities, simply because the quantity of meat produced could not have been consumed by household members in a short time (Halstead 2011; Mlekuž 2005). Even on a small site such as Grebeni, 400kg of meat, at 500g of meat per person, would have been sufficient for a feast for its entire population of 240 people and many hundred visitors as well, with surplus meat preserved by

494 

 Discussion

imported salt. This calculation suggests that slaughtering a bovid at a small site was not a common occurrence, with caprines more likely than bovids at feasting on small sites. By contrast, a grand feast for the entire population of 3,200 people in the DGM would have required the slaughter of four mature bulls. In summary, the scale of subsistence is where the allowance for fuzziness in the characterization is particularly useful as what is produced locally and what is brought as gift/tribute depends on which of the three models is preferred. Once again, whichever model is preferred is far removed from Grebeni subsistence practices, with the most distant fields located half an hour’s walk from the settlement and two communal flocks of 80 cattle and 200 caprines pastured near the settlement by a handful of shepherds. Affordance to be a node and re-distribution centre of wide-reaching exchange networks That Trypillian communities were parts of different-sized exchange networks is evident from the metal supplies, initially related to the Balkan copper sources but which shifted to Transylvanian sources in Phase BII (Ryndina 1998). Such a largescale network was the basis of the introduction of Balkan copper to Hvalynsk in the Volga Basin via North Pontic or South Trypillian communities (Chernykh 1992). Medium- and small-scale networks would have involved the exchange and/or trade of salt, flint, pigments including manganese and graphite, pottery and animals. Some of these low-bulk goods could have been distributed in a ‘down-the-line’ manner, whereby adjacent sites such as Grebeni and its neighbours took what they needed and passed on the remainder to the next settlement. Alternatively, they may have been a by-product of high-bulk (e.g., salt) exchange activities that were re-distributed in certain nodes of the network within a social practice that we now associate with assemblies. Megasites are obvious candidates for such centres, because of the open space allowing the storage or exchange of high-bulk (e.g., animal) goods and access to a wide range of consumers – pilgrims, or assembly visitors, or permanent occupants some of whom may have been specializing in trade and exchange. Direct, although limited, evidence for exchange is the golden hair ornament found in Nebelivka – a rare and exotic find in the Trypillia world; other important items include the imported graphite-decorated dish from South-East Romania and Volhynian flint from the Prut-Dniester valleys. The scarcity of copper on megasites and small sites related principally to the lateral recycling of copper into other objects (cf. Taylor, T. 1999). Thus Grebeni-type sites could have occasionally acquired exotic objects directly via down-the-line exchange or by sending representatives to assembly sites such as Nebelivka to benefit from redistribution108. It was only in large sites such as Nebelivka

108  An example of such a high-value item occurring on a small Cucuteni site is the pre-Balkan Platform flint blade from North-East Bulgaria found at Silişte-Prohozeşti, near Moineşti, Moldavia, presumably as the result of redistribution from the nearby tell of Poduri (Chapman & Monah 2007).



Low-Density Urbanism – a Global Approach 

 495

that redistribution was feasible because of the scale of assembly or dwelling, with 10 times more potential exchange partners than on sites such as Grebeni. The social framework for megasites The perceptive reader will immediately have noticed that the persistently discussed concept of the ‘Trypillia Big Other’ features in this section as a form of representation rather than an element of characterization. This is deliberate and in accordance with the principle of characterization not to create boundaries. The analogous problem would be that we listed ‘hierarchy’ or ‘heterarchy’ as a characterization rather than considering them as alternative representations of the characteristics of ‘social framework’. On current evidence, the social framework of the vast Trypillia network is more heterarchical than hierarchical in nature, as there were no proxies for the latter, unless the status symbols were subject to destruction or poor preservation, such as the feather cloaks, tattooing, wood carving and boat-building of East Polynesian stratified societies (Kirch 1989). The idea of heterarchy would also better fit the widely accepted egalitarian tendencies visible in Trypillan sites of all sizes, Grebeni included. There is an important caveat for such an equation. The term ‘egalitarian’ is often conflated with communities which are socially undifferentiated. But the kind of inter-personal differentiation (personal ranking) which may have arisen in persons who had, for example, mastered certain high-level skills did not necessarily imply structural ranking on a wider social scale – more a greater distinction for the household of that skilled person or their LIG. It seems almost trite to reiterate that, on a megasite as vast as Nebelivka, the spatial recapitulated the social. In each of the three alternative models, residents from between ten and 40 home communities settled at, or visited, Nebelivka, constructing well-built, sometimes two-storied, houses whether they lived there permanently (the Distributed Governance Model) or seasonally for a month each year (the Assembly and Pilgrimage Models). The long-term tension between the ‘global’ Nebelivka Identity and the ‘local’ identities of the home communities was spatially replicated by the contrast between top-down planning (the global identity) and bottom-up planning and architectural heterogeneity (local identities). Just as each Neighbourhood was rooted in and reproduced the spatiality of its own home community, so the leaders of each Quarter sought to integrate several different home communities into their understandings of the overall Nebelivka planning schema. The differing numbers of Neighbourhoods in each Quarter influenced the way that Quarter leaders tried to reduce their planning tensions, leading to the different shapes and sizes of each Quarter as well as varying widths and shapes of the vital intra-concentric circuit space. The intriguing result of the VGA (Section 4.3.2) was that, despite all of the variability, the Quarters produced a consistently similar spatial order over the long term and across the megasite.

496 

 Discussion

At the more local scale, there were endless spatial ways of differentiating a specific group from another group, whether laying out a separate ‘square’ in the area inside the inner house circuit, the use of a ‘kink’ by those laying out the next Neighbourhood to separate their houses from the previous house group, or the decision to ‘block’ the continuation of an Inner Radial Street with a segment of a concentric street at right angles to the IRS. The most extreme example of such ‘political’ planning was the Northern area of Majdanetske, where the tensions between IRSs and concentric streets erupted into planning chaos (see Section 6.2). But such differentiation of Neighbourhoods did not prevent interaction between the house groups, with grander ceremonies such as Assembly House-burning attracting participants from several or many Neighbourhoods. For the Assembly and Pilgrimage Models, the long-term tension between ‘global’ and ‘local’ identities could be sublimated in a one-month visit but what happened for the permanent residents of the ten clans in the Distributed Governance model? The simplest scenario was that each clan built one or two Quarters and lived there permanently, with Neighbourhoods reproducing local clan variations whether in one or several home communities. In this way, those clan leaders in the 100km Nebelivka macro-territory who played important roles at the central site would have become Quarter leaders and, in some cases, members of the Nebelivka Council. Behind all of the changes in the evolving megasite layout was the group who had initially made the decision to begin a centre at Nebelivka – for the Assembly and Pilgrimage Models the site Guardians, for the DG Model the Council. How did this group not accrue such status or worldly goods as to be translated into a ‘Royal Lineage’ or some such exalted title? The strong possibility of such a development means that any constraints on household or small-group accumulation would have been established in the deep structure of the Trypillia group, for example in the Trypillia Big Other (for the disadvantages of small-group competition on urban sites, see Jennings & Earle 2016). An egalitarian ethos was fundamental to many of the Balkan Neolithic and Copper Age groups from which the Cucuteni-Trypillia group emerged (e.g., LBK, Boian and Vinča: Chapman, in prep.) and such a strong principle would have guided many of the Trypillia social practices. The founding group obviously held an important position - the founding group with lineage primacy if this were based upon a descent group system. But if members of the founding group decided to extend their kinship authority or ritual primacy into a more general social dominance, one would have expected strong resistance against such an inegalitarian policy which ran counter to the Big Other. In a similar way, it is unlikely that households with members showing signs of distinction, whether as particularly skilled flint knappers, carpenters or figurine-makers, would have desired, let alone achieved, the structural prominence of their household on a permanent basis. They may also not have desired to become the leaders of their particular Limited Interest Group. However, this did not preclude the expression of different types of personhood as denoted by figurine styles within the Trypillian Big Other (Gaydarska 2019).



Low-Density Urbanism – a Global Approach 

 497

In the megasites of Taljanki and Majdanetske, the transformation of pottery production from a largely or wholly household affair, as in Nebelivka, into a system of dispersed, kiln-based production created the first serious caesura in the communal mode of production. Importantly, the new system changed the basis of the relationship between households, Neighbourhoods and Quarters, with specialized workshop production underpinned by Limited Interest Groups of skilled potters, who fostered communal co-operation with individual households and Neighbourhoods. Thus, a specialized production practice combining the labour of many persons was itself integrated into pottery production – one of the key aspects of Trypillia lifeways. The fact that pottery kilns were widely dispersed across two megasites suggests the importance of the Quarter as a heterarchical organisational and decision-making unit, in contradistinction to a centralised model of a single, specialised pottery-producing Quarter organised at site level in a hierarchical manner. It is clear that such a complex social structure, based largely on the reproduction of prior socio-spatial relationships in home communities onto the megasite planning frame, would have been entirely superfluous in a Trypillia site as small as Grebeni. Indeed, if the reader were searching for a single, decisive differentiating trait between urban and smaller settlements, it would come in the form of social complexity at the former and its complete absence at the latter.

6.3.3 Summary In summary, a clear distinction between a megasite and a smaller Trypillia settlement can be drawn for every single characterization discussed above. The relational analysis can thus be said to have reached a successful conclusion – we have demonstrated that there is no possibility that the Nebelivka megasite was simply a very large example of a typical small rural settlement. Such an equation would be a categorical mistake, of the kind which suggests that aircraft carriers are simply very large examples of yachts. This result indicates that, for the 4th millennium BC on the forest steppe of Ukraine, there was a class of sites which was so different from typical small settlements that it merits the title of ‘city’. The claim is that Trypillia megasites exhibited the same order of qualitative and quantitative differences from the typical small Trypillia settlement as the city of Uruk did from small tells in the Fertile Crescent, or Roman London from the villas of South-East England. The fact that this class of megasites can be dated to the earliest part of the 4th millennium BC  – several centuries earlier than what had been supposed to be the earliest cities in the world  – is a further ground for considering their significance in world prehistory in a new light. Before considering this finding in a wider context, it is worth pausing to consider the two recent comparisons between what has been for long believed to be the earliest city in the world – the Late Chalcolithic 5 phase at Uruk – and the Trypillia megasites which we now claim as the earliest cities. In the first account (Müller &

498 

 Discussion

Pollock 2016), large sites in the South Mesopotamian alluvial lands were inhabited for millennia, while small sites were more volatile. The city of Uruk in Late Chalcolithic Phases 2–4 was said to be too large to be self-sufficient, requiring tribute from people living outside it. Increasing centralization was seen in Uruk Phase LC5, with a wide range of objects associated with administration. However, it is conceded that Uruk urbanism was ‘exploratory’ in the sense that many architectural innovations were not perpetuated. By contrast, although large, the absence of a tributary system and an elaborated urban management at the Trypillia megasites showed that they were a social ‘experiment’ in nucleated living. The parallel notions of ‘exploratory urbanism’ (Uruk) and ‘social experiments’ (megasites) show that there were common elements between the two regional cases. In the second account (Wengrow 2015), the case is made that there were two examples of cities before the state in the 4th millennium BC – in South Mesopotamia and in the Ukrainian forest-steppe. There is an emphasis on the diversity of 4th millennium BC social organisation, with significant autonomy of smaller households and kin groups in agrarian and craft production – including local irrigation systems – alongside the large-scale centralised institutions. This diversity is matched in decision-making bodies, which included civic assemblies, tribal councils and other bottom-up groupings in addition to the centralised temple institutions. Wengrow does not deny the obvious differences between Uruk and Trypillian megasites in the capacity of the exchange networks in the ‘Uruk Expansion’ and the standardisation of Mesopotamian pottery in comparison with the elaborate variety of Trypillia painted wares. But he does emphasise that the differences between the political backgrounds of the two cases were not as extreme as Müller and Pollock suggest. The key point from these two comparative studies is the idea that cities before the state could develop in two such different ways in the two regions of South Mesopotamia and the Ukrainian forest-steppe. We now turn to the further application of the relational approach to two other Trypilla megasites.

6.3.4 Comparisons with Other Trypillia Megasites It has been possible to complete only a partial comparative assessment of the characterization of the megasites of Taljanki and Majdanetske in the way attempted for Nebelivka, since the respective projects were not formulated with such questions in mind. Similar comments as were applicable to Nebelivka may be assumed for the other two megasites in respect of the territory/area to which the megasites were central, the heterogeneity of their population, the built environment with special areas of formalized space, the scale of subsistence, their affordance as re-distribution centres in wide-ranging exchange networks and the social order. Concerning site size, scaling-up from Nebelivka to Taljanki by an additional 50% would have introduced logistical pressure, even for people with experience of living on large sites. Although



Low-Density Urbanism 

 499

the site clearly thrived for one or two centuries, it is probably not a coincidence that such a size was never attempted again. Majdanetske reverted to the more manageable size of 200ha, with the much more crowded layout representing the major difference from Nebelivka. The difficulty of testing any of the three alternative models on the incomplete site plans of Taljanki and Majdanetske precludes discussion of alternative population estimates which would be more reasonable than the maximalist views. However, the identification of trackways and pottery kilns indicates a greater concentration of skilled labour and management than in Nebelivka – something that is highly improbable at small sites such as Grebeni. In other words, it is worth assuming that the more expansive discussion of the characterization of the Nebelivka megasite can stand for the megasites of Taljanki and Majdanetske. In all three cases, the structural differences and variations in possible social practices on megasites and small sites show that there is a relational gulf between the two site types. As proposed above (p. 482ff.), this result is very important in terms of the relational model of urbanism, for it suggests that we can identify in the Trypillia megasites the earliest known case of European urban settlements. How do the Trypillia megasites relate to other urban sites of the type known as low-density urban sites?

6.4 Low-Density Urbanism 6.4.1 The Low-Density Urbanism Model This section is not the place for a wide-ranging and detailed analysis of the many and varied trajectories towards low-density urbanism (henceforth LDU) that can be identified across the world today (but see Fletcher & Kim, submitted). Instead, we can accomplish a lesser goal – to compare and contrast Trypillia megasites with the characteristics of other megasites which have become part of the global debate on low-density urbanism (Chapman & Gaydarska 2016a). Roland Fletcher (2009, n.d.) has characterised LDU as a different kind of urbanism from the classic nucleated form – one in which the urban centre moved outward and colonised the rural areas to produce a patchwork of monumental foci, residential houses and agro-pastoral areas. This insight led to the search for other large, LDU sites which did not, by definition, fit well the traditional definitions of a city. The global distribution of low-density urban sites is currently patchy, owing to variable research but covers the temperate zone in Europe (Trypillia, Iron Age oppida), North America (Cahokia, Chaco Canyon) and East Asia (Longshan group, China), the savannah zone in Africa (Great Zimbabwe), as well as the tropical zone in Africa (Yoruba towns), East Asia (Co Loa, Angkor Wat) and South America (Amazonia) (Fig. 6.11). Current areas without obvious LDUs include much of the Near East, North Africa and South Asia.

500 

 Discussion

For a decade, these low-density urban sites comprised a group without a satisfactory name. Our proposal was to give the generic name of ‘megasites’ to these embodiments of LDU (see above, p. 22). These sites were settled in immensely varied landscapes, demonstrating varied forms of architectural and social complexity as well as trajectories of development, peak and decline. Yet there are seven polythetic traits shared by most, if rarely all, forms of LDU. First, the importance of major building projects, which transformed the LDU site at each new phase. There was a sense, at major monumental foci such as Cahokia (Pauketat 2009, n.d.), that the world was created anew each time a new temple or monument was constructed. The foundation of each new stone compound by a different Shona king at Great Zimbabwe (Pikirayi, n.d.) also re-formulated royal interactions, creating new foci of elite attention for the entire population. If we think, together with McAnany (n.d.), of building projects as ‘experiments’ in agglomeration – in sociality – we can avoid the dangers of anachronism and teleological reasoning. Secondly, there was a strong modular component of kinship-base, house-oriented planning practices in LDU. This element was especially significant in the emergence of Yoruba towns, where, from a starting position of houses linked to mega-houses ca. AD 500–800, with with a corporate head, under an overall ruler, the mega-houses were dismantled and replaced (AD 11th–12th) by an urban component of multiple corporate groups under a centralised government and a divine king (Ogundiran 2013, n.d.). Thirdly, many of the smaller LDUs – sites of 50–70km2 rather than 500–1000km2 – had seasonal components in their populations. The most extreme example may be the Solomonic Empire of Ethiopia (AD 1270–1529), which moved its low-density capital of some 40,000 residents, with a standing army of a further 40,000 soldiers, every few weeks (Marcus 1994; Fletcher 1998). White suggests that many of these smaller LDUs may have had a seasonal component which had the perhaps unintended consequence of weakening household ties (White, in prep.). Fourthly, most cases of LDU downplayed the mortuary domain. Although the Early Shang period of China was an exception, many other examples of LDU showed very little burial evidence (Cahokia, many European oppida, Amazonian megasites, Yoruba towns and Great Zimbabwe). The concentration of ideological power in the hands of the ruling elites would have created tensions between households and the centre, especially at the time of deaths in the family. Fifthly, there was rarely an obvious successor to the LDU site, whether on the same place or in the same region. Regional capitals such as post-Han Co Loa were abandoned ca. 150 BC, not to be re-settled again (Nam 2013, 2015). After the great house-burning events at Cahokia, depopulation took the form of sub-group outmigration with no monumental centres left (Pauketat n.d.). Pauketat (n.d.) has drawn an ontological contrast between mud-brick cities, which lasted millennia, and wattleand-daub cities, which lasted centuries if not mere decades.



Low-Density Urbanism 

 501

The sixth feature was the transformation from higher-density to lower-density large sites in the creation of LDUs. Fletcher (1995) postulates that a common urban trajectory began with high-density cities with a size of up to 100ha and a population density of 300–600 people per ha, but that these cities morphed into increasingly large but also increasingly low-density settlements of 1–40km2. While the most noteworthy modern example is Megalopolis on the East Coast of North America, there are also many ancient examples which have been transformed from higher-density precursors into huge low-density centres (e.g., Angkor, Cahokia and Co Loa: Fletcher & Kim, submitted). The time taken from the origins of agriculture to the formation of urban communities is the seventh trait which differentiates high-density from low-density centres (Feinman & Nicholas 2016, Fig. 13.2; Fletcher, n.d.). The large, compact, highdensity examples took far longer to develop after the initial domestication of plants and animals – up to 4,000 years – but then lasted far longer. By contrast, megasites took a far shorter time to develop since the regional emergence of agro-pastoralism and generally lasted far shorter than the high-density capitals. There are many other interesting aspects of the global pool of LDUs but these seven characteristics provide a picture of dynamic, changing settlements whose population may have been more mobile than we may have expected and whose building projects created a succession of ‘new worlds’ in which successive generations of residents made their lives. How do the Trypillia megasites compare and contrast with this heterogeneous bunch of ‘misfits’? Or do the megasites accord better with Gabriel Cooney & Eoin Grogan’s (1999, p. 232) characterization of the Neolithic as ‘local worlds linked by exotic elements’? Were Trypillia megasites ‘Neolithic’ in subsistence, with their inefficient, extensive mixed farming and scarce prestige goods, but ‘urban’ in appearance, with their massive size and organised planning? It is surely not coincidental that the Trypillia megasites meet all seven shared characteristics for global megasites. The major building projects required for the creation of the megasite plan provided the kind of inter-generational temporality at the planning scale which Kujit (n.d.) deemed necessary for households. One of the key elements in the plan – the Quarters – was found in other megasites such as the Swahili cities (termed ‘wards’: Kusimba et al. 2006), while focal Assembly Houses of the Nebelivka type were important in the house clusters of the Moundville complex as much as in those of Tahitian villages (Kahn n.d.). Moreover, there is a structural parallel for the Quarters/Assembly Houses in the development of 30 seasonal ceremonial centres each with a mega-mound in the 3rd millennium BC in the Norte Chico group of Northern Peru (Piscitelli n.d.). The range of housing density on all well-studied Trypillia megasites is compatible with the low density of other newlydesignated megasites (Fletcher 2009).

502 

 Discussion

The next characteristic of megasites  – the modular component of kinship-based, house-oriented planning – is fundamental to Trypillia megasites. While the modular unit of the household has been recognised as basic to all Trypillia sites, the largescale planning which prioritised the location of each house relative to its neighbours was restricted to the Trypillia megasites. The significance of house groups from the same home community arriving together at the start of the Nebelivka megasite is reminiscent of the importance of extended households in the origins of the Central and West European Iron Age oppida (Moore 2017). In two of the three models proposed for Nebelivka  – the Pilgrimage and the Assembly Models – there is a strong seasonal component: for the former, an eightmonth-long pilgrimage season when the pilgrimage centre was open for visits from Trypillia pilgrims, while, for the latter, a one-month Assembly period in the summer for a high-intensity period of gathering, ceremony and consumption. One aspect of the Trypillia megasites which differs from other urban practices is the way that links between the centre and the home communities were never broken – if anything they were even strengthened through seasonal movement or, in the Distributed Governance Model, by regular provisioning of the megasite. This contrasts strongly with Underhill’s (n.d.) description of urban centres deliberately ‘forgetting’ the ties that bound people to local places in the urban hinterland, and especially to the ancestors, in the search for a stronger urban identity. From the outset ca. 5000 BC, the entire Cucuteni-Trypillia group developed a much stronger domestic arena than a mortuary domain, with no sign of formal cemeteries until centuries after the end of the megasites. The fact that there was a pre-existing dearth of extra- and intra-mural burials in the Trypillia group may have facilitated the creation of a distinctive megasite identity, which could have been threatened by local ancestral ties. Fifthly, in the post-megasite part of the Trypillia group’s development, small settlements began to dominate the landscape, followed by a strong reaction to the predominance of the domestic arena in the barrow-building period (very late Eneolithic–Early Bronze Age), when monumental barrows gained visual and material control of the landscape. There was thus no obvious successor to Trypillia lowdensity urbanism, whose wattle-and-daub architecture provided limited potential for material survivorhood. The only apparent exception in the suite of shared LDU characteristics was the decreasing density of Trypillia megasites with increasing size in comparison with earlier sites. We cannot yet confirm this sequence for the maximalist model of Trypillia megasites, perhaps because so few complete Early Trypillia settlement plans have been published using modern geophysics (but cf. Mogylna III and Nebelivka, Fig. 6.12). However, the comparison of occupational densities for any one of the three alternative models for Nebelivka with earlier Trypillia sites shows a steep decline in the number of structures per ha, with fewer than four buildings per ha for each



Low-Density Urbanism 

 503

model109 – well below any of the occupational densities published for Early Trypillia settlements. The short period of time of less than one millennium between agricultural origins and megasite formation was also confirmed for the Trypillia settlements. Since it is hard to discern traces of domesticated plants and animals in the ‘Forest Neolithic’ which preceded the spread of Trypillia pottery-using groups Eastwards from the Prut valley, the current view is that the Eastward spread of agriculture coincided with the expansion of maximal settlement size in the forest steppe zone through the 5th millennium BC, with planned megasites dating from the beginning of the 4th millennium BC. These comparisons show that the Trypillia megasites have as much justification for inclusion in the list of LDU sites as any other regional class of sites, sharing all of the seven most widely distributed characteristics in the global distribution of lowdensity urban sites. While the appearance of all seven LDU characteristics does not prove with absolute certainty the claim that the Trypillia megasites were part of the global LDU Phenomenon, their presence in 4th millennium BC Ukraine makes a strong case for the inclusion of these extraordinary sites in the emerging group of low-density urban sites. This claim should not obscure the point that there are both urban and non-urban sites in both categories of occupational density – high and low (Fig. 6.12). We have deliberately chosen the small number of sites in Figure 6.12 lower to emphasise the range of sizes/densities in play. As a total contrast, the tiny North-East Bulgarian tells of Ovcharovo and Poljanitsa - by no stretch of anyone’s imagination a pair of cities - have some of the highest residential densities in Eurasia, while the residential density of 7.5 persons/ha applies to one of the world’s largest sites  – the 1,000km2 LDU of Greater Angkor. In between these extremes lie most known archaeological sites. The estimated populations of the non-urban settlements from South-East and Eastern Europe fall between 59 and 180 people/ha, with sizes of between 1ha and 16ha. Two urban sites of comparable size – Late Chalcolithic 5 Uruk and Nebelivka – show population densities of 100–200 for the former and 10–42 for the latter. There can be no doubt that living on any of these exemplary settlements would have engendered a very different experience from living on most other sites. The contrasting residential densities of Late Chalcolithic 5 Uruk and Nebelivka make the strongest point about the kind of lifeways experienced on these two sites of similar size.

109  The highest occupational density estimates for the three models show 1.6 structures/ha for the Distributed Governance Model (all generations); 2.7 structures/ha for the Assembly Model (Generation 4); and 3.3 structures/ha for the Pilgrimage Model (Generation 3).

504 

 Discussion

Figure 6.12: Densities of people and structures per ha, urban and non-urban sites of low- and highdensity (upper by L. Woodard; lower by J. Chapman).

Summary 

 505

6.4.2 Summary The comparison of Trypillia megasites with the broad range of low-density urban sites found elsewhere in the world produces a convincing demonstration that the Trypillia megasites are currently the earliest known examples anywhere on earth not only of urban settlements but also of low-density urban settlements. Moreover, the acceptance of two of the three alternative models of megasite development means, in effect, that not only do we have low-density urban centres whose occupations may well have been seasonal in nature but their political economy was also partly restricted by the small volume of exotics crossing the Ukrainian forest steppe. This is perhaps not the conclusion that supporters of classic Childean urbanism would have been happy to contemplate. But it is a conclusion that places the Trypillia megasites closer to the worlds of the Neolithic than to those of the Near Eastern Late Chalcolithic or Early Bronze Age, in the sense of Cooney and Grogan’s (1999) characterization of the Neolithic as a series of ‘local worlds linked by exotic elements’. What we are seeing is a local world fitted within a 100km radius of a megasite and developing close network links with many of the sites in that macro-territory  – a combination of Cooney and Grogan’s insight with the megasite characteristics of low-density urbanism. The second point is to underline that the very low quantities of exotics so far found at any of the three most widely explored megasites fits the second part of the description rather well.

6.5 Summary This chapter has focussed on a complex, comparative approach which has covered a huge amount of ground, from the lowest spatial scale of the household up to global comparisons of low-density urban sites. There are seven conclusions which are of the greatest significance for our study of Trypillia megasites. First, the spatial and artifactual analyses at Nebelivka confirm the importance of the four nested levels that we have proposed for the megasite. The performances of house-burning were reinforced by the ‘death assemblages’ of pottery and other objects, which rarely reflected the living assemblages of the time but instead showed the selection of objects which conveyed the relationships of other groups to the burnt house and its residents. The variability of content and house size in the second nested level of Neighbourhoods showed how different ‘home communities’ were represented at Nebelivka – a practice that reduced the size of the entire megasite to more manageable, ‘local’ proportions. The VGA analysis of Quarter space confirmed that, despite local variability in size and arrangement, the Quarters respected similar rules of spatial patterning across the whole megasite. In view of the surprising lack

506 

 Discussion

of materialisation of a hierarchical social order, the principle of heterarchy was important at the megasite level for each of the three alternative models. Secondly, three models have been developed as alternatives to the rejected ‘Maximalist’ model of large, permanently occupied, all-year megasites with massive populations. Each of the three models  – the Distributed Governance Model, the Assembly Model and the Pilgrimage Model  – is consistent with the four principal constraints of the Nebelivka occupation  – the modelled length of the duration of Nebelivka at 200 years; the number of houses built at 1,445; the number of houses burnt down at 1,077; and the absence of a major human impact on the Nebelivka landscape and its forest steppe. While each model has its own specific strengths and weaknesses, on current evidence we cannot eliminate any model. This remains a research priority for the future. Preliminary testing of the models at Majdanetske and Taljanki gives promising results. Thirdly, the origins of the megasites were founded on the site clustering that began as a long-term settlement trend in Phase BI and their emergence as nodes in wide-ranging exchange networks in this Phase. The first combination of the key planning principles of inner open areas, concentric house circuits and inner radial streets dated to Phase BII, with the establishment, at sites such as Nebelivka, of a new combination of inner ritual space and outer residential space. The key potential for alliance formation differentiated megasites from other, smaller sites. Fourthly, any successful account of the demise of the megasites must explain the three cycles of decline and recovery, from Phase BI onwards. There is weak current evidence to support the traditional explanations of the demise – warfare and unsustainable subsistence. The alternatives concern fluctuations in long-distance exchange networks (flint, copper, pigments) and the competition between the megasites of Nebelivka, Majdanetske and Taljanki, whose partially coeval occupation has been demonstrated by new AMS modelling. Fifthly, the new approach of ‘relational urbanism’ has yielded a positive result in terms of the comparison between the Nebelivka megasite and a typical, small Trypillia settlement such as Grebeni. The comparison took the form of nine different characterisations, on all of which there were substantial differences between the megasite and the small settlement110. This means that the experiences of people living on megasites and small settlements were so different that the megasites could not be considered as simply ‘very large settlements’ but, in relational terms, were different types of site – urban settlements. Sixthly, a global comparison of Roland Fletcher’s category of ‘low-density’ urban sites – an alternative to the classic high-density urban sites such as Rome, London

110  The only exception to this pattern of difference can be found in one aspect of the characterisation of ‘population heterogeneity’, in which the Trypillia Big Other, identified through the proxy of the density of figurines per site size was similar on large and small sites (Gaydarska 2019).

Summary 

 507

and Constantinople – shows that the seven principal characteristics of low-density sites were also found in the Trypillia megasites. This means that, finally, the Trypillia megasites can be considered not only as urban centres but also as the earliest urban centres in the world; not only as lowdensity urban sites but also as the earliest low-density urban sites in the world. Furthermore, future acceptance of either the Assembly or the Pilgrimage Models for megasite growth would mean that the earliest urban centres in the world were not only low-density but also featured temporary occupations.

John Chapman and Bisserka Gaydarska

7 Conclusions

In the final chapter, we present the twelve principal conclusions of the monograph, discussing the major implications of each finding for Trypillia megasite archaeology. We then move away from a traditional account of our Project findings to a more intimate, person-focussed narrative about the Nebelivka megasite from the perspectives of ten different people – a Nebelivka Guardian, a house-builder, a clan leader, a visitor to the Assembly, an organizer of the Assembly, a pilgrim, an adolescent visitor, a Nebelivka ritual leader, a trader and one of the last generations of those living at Nebelivka. While these narratives are fictional, we hope that they convey something of the sense of being part of the extraordinary phenomenon that was a Trypillia megasite. We then return to a more traditional theme by considering what would form part of a future Trypillia megasite research agenda. We conclude the conclusion with a personal ‘Endword’.



Principal Results 

 509

7.1 Principal Results 1. There is a strong case that Trypillia megasites were not only urban sites in relational terms, not only examples of low-density urbanism sharing all of the seven principal traits of megasites in general, but also the earliest known examples of LDU in the world, dating to ca. 3900 BC onwards. This result is perhaps surprising to the majority of archaeologists and certainly to all of the general public. It challenges the longheld view of the primacy of Near Eastern urban origins in the Fertile Crescent in the Late Uruk period (3400–3000 BC) and puts the Trypillia megasites at the heart of the debate over the emergence of urbanism in prehistoric Eurasia. Perhaps now it is time for colleagues to put themselves in Benedict Anderson’s (1991) mindset and imagine the possibility of urban developments in Eastern Europe before those in the Near East? 2. The basis for the emergence of megasites was the Trypillia Big Other  – that supra-regional consensus about the importance of houses, figurines and pottery for Trypillia lifeways – which reproduced itself as much through everyday practices (the habitus) as through the extensive social networks, including exchange, stretching from the Cucuteni group Eastwards to the Dnieper valley. The Big Other provided a common interactional framework for communal projects at a supra-settlement scale – a common set of shared values materialised in regionally specific ways. Without the Big Other, it seems highly unlikely that communal projects on the scale of the creation of megasites would have been possible. 3. The planning of Nebelivka was not a question of inheriting and reproducing a pre-existing and well-known form with established planning elements (concentric house circuits, inner radial streets, a vast inner open area, Assembly Houses) but rather a process of bricolage whereby the planning elements appearing for the most part individually on earlier, 5th millennium BC large sites were integrated over time into a single unified layout. The creation of an unprecedented settlement plan required a key group of early residents to imagine the possibility of a megasite – in itself a remarkable achievement. The ‘final plan’ of Nebelivka was not the same as the initial plan, which started as a minimal version of either (a) a house circuit with perimeter ditch (the Distributed Governance and Pilgrimage Models) or (b) several Quarters with later infilling of additional Quarters (the Assembly Model). The new geophysical investigations of the 2000s and 2010s have not only confirmed the well-known, basic elements of the megasite plan but also identified several novel elements and combinations of elements in the plan. Individual elements include the Assembly House, the unburnt house, different sizes of pit, industrial features including kilns, and pathways. Groups of elements include Neighbourhoods, Quarters and pit lines or groups, with single pits often close to houses. The precision of these geophysical plans has opened up a new world of interpretational possibilities limited only by the elaboration of theoretical support for the interpretations and a fine-grained chronological inner sequencing of megasites. In this project, we have made every effort to provide the former, while the latter is

510 

 Conclusions

so far lacking on every megasite where such an effort has been made. The absence of a fine-grained inner sequence has led to two approaches to spatial analysis: the (highly questionable) assumption that all structures at a megasite were in coeval use (e.g., Ohlrau 2015; cf. some of the Visual Graphic Analyses assumed coeval use of all structures in a Quarter: Section 4.3.2), and the integration of the modelling of inner sequences with spatial analyses (other VGA, Section 4.3.2). The overall conclusion of the spatial analyses is that, nested within an overall site framework agreed or imposed from the top-down at the start of megasite dwelling, there is a huge amount of ‘local’ variability in the layout of every single planning element – whether the length of ditch segments, the size or shape of dwelling houses or Assembly Houses, the size and temporal duration of Neighbourhoods, the size and layout of Quarters, the constitution of inner and outer circuits, inner radial streets, blocking streets and squares and the size and shape of the three main open areas. This heterogeneity strongly suggests that the megasite was formed by a variety of home communities, each of whom seeking to maintain the identities of their own home sites in the face of tendencies to adopt an overall Nebelivka identity. In addition, the diversity of house sizes in any single Neighbourhood suggests ‘local’ competition in the construction of homes. Perhaps the key finding of the visual graph analyses of both entire Quarters and temporal groupings within Quarters was that, despite the detail of architectural variability, each Quarter would have had similar structuring of visibility and movement across the entirety of the site and over long periods of time. This finding shows an important link across scales of inhabitation, with megasite spatial order emerging as a monumental part of the Trypillia habitus. 4. An interpretation difference was soon to open up between the Durham group and our colleagues from Ukraine and Germany in respect of the very nature of a megasite and the populations at these sites. This difference is introduced at this juncture because it underlies the remainder of our conclusions. We have used the shorthand of ‘Maximalists’ and ‘Minimalists’ for these two opposed positions. We suggest that there were two kinds of Maximalist-Minimalist relationship: the Ukrainian-German teams are demographic maximalists (site population estimates) and empirical minimalists (estimates for subsistence and building resources), while the Durham group consists of empirical maximalists (estimates for subsistence and building resources) and demographic minimalists (site population estimates). We claim that we have demonstrated that the Maximalists have consistently underestimated the resources and labour required to build megasites by using our experimental programme of house-building, -burning and burnt house excavation to produce realistic estimates of resources and labour for these tasks. 5. The divergence in views about megasites summarised in point 4 did not exist at the start of our research project but came to a head with time at the so-called ‘tipping-point’, when we recognised as many as nine (later 10) lines of evidence that cast doubt on, or simply contradicted, the standard position of all-year-round, permanently occupied megasites with populations in their tens of thousands. One of



Principal Results 

 511

the most important lines of evidence concerned the Nebelivka sediment core, located 250m East of the perimeter ditch of the megasite and in which we had expected to find proxies for a massive level of human impact on the local chernozem-dominated forest steppe. It was a great surprise to find that the level of human impact on the environment was modest - indeed, no greater during the megasite occupation than it was before and after it. This result and the other nine lines of evidence prompted a re-conceptualisation of megasites as long-term centres but with much smaller populations (in their thousands rather than tens of thousands) and with either seasonal or permanent occupations. This led us to develop three models for Nebelivka: the Distributed Governance Model, the Assembly Model and the Pilgrimage Model. The comparative evaluation of the three models shows that each model has interpretative advantages and disadvantages; it is not possible at this juncture to reject any model or to give a final preference for the most likely scenario. One of the primary reasons for the ambiguity in our assessment of the models was our failure to create an internal phasing of the various building units in the Nebelivka plan. The coincidence of our 80+ AMS dates with a wiggle on the calibration curve meant that the most useful dating result was the estimate of a 200-year duration for Nebelivka, most probably between 3970 and 3770 BC but no fine-tuned discrimination between the durations of inner or outer house circuits or inner radial streets. 6. The fieldwalking and remote sensing programmes in the Nebelivka microregion showed that there were no other Trypillia ‘sites’ (i.e., dense clusters of surface pottery and house daub) within 8km of Nebelivka; indeed, there was only one Trypillia sherd deposited within the 5km radius of the megasite. This absence of a ‘hinterland’ distanced Nebelivka from the classic ‘urban - rural’ relationship in favour of a concentration of people at the megasite coming from sites in a wider region. Following intensive data cleaning, the spatial analysis of the site database from the ‘Trypillia Encyclopaedia’ (Videiko 2004), amounting to 499 dated Trypillia sites of known location and area, showed that Trypillia megasites were outliers in the general size distribution of Trypillia settlements at a scale of 100km, which acted as the operational limit of their social territories. This means that a large group of small settlements could have acted as ‘home communities’ for either visitors to, or residents at, the Nebelivka centre. This result was important in establishing baselines for the three alternative Nebelivka models, each of which conform to the parameters of overall megasite house numbers and the lack of major human impact on the forest steppe environment. 7. The Distributed Governance Model starts from the premise that there was small-scale settlement in the Nebelivka area before the megasite was founded and that settlement was linked into a wider network of sites who would co-operate in a long-term communal centre. Once the decision was made to start the centre on the Nebelivka promontory, building of the first house circuit defined the overall shape of the site over a decade, with a group of ten clans each sending settlers to build and live there permanently. Each year, a different clan took over all aspects of the running

512 

 Conclusions

of the megasite and contributed to its provisioning by bringing resources from their home site. The population level varied between 2,400 and 3,200 people, living in ca. 400 houses, with a steady rate of house-building and -burning which emphasised the success of clan control. Festivals and ceremonies peaked at the end of the year with the ‘Change of Clan’ ceremony but with smaller-scale rituals throughout the year. In summary, this model shows the possibilities of a much smaller population than had been previously considered living permanently, yet sustainably, on a megasite without creating major human impacts on the local landscape and without overstretching local resources by the provision of food, drink, salt and other resources from outside the megasite. Any small-scale threat to regular contributions would have been dealt with at clan level, while the centralising tendencies of the Trypillia Big Other would have been invoked to ensure support from all ten clans. 8. The Assemblage Model invokes a seasonal settlement of the megasite, with a major annual one-month assembly period supported by a small population of megasite Guardians. The Guardians have settled on or near the Nebelivka promontory before the start of the megasite to lay out the outline of the site, plan for the assembly and organise the supply of building materials, not least five years of coppicing hazel rods for wattle-construction. The megasite was developed through the layout of four or five Quarters in each of the first three 30-year generations, with an estimated fewer than a 1,000 visitors in the first generation, rising to a maximum of 3,300 visitors in Generation 4. This expansion brings the major resource challenge of a peak of housebuilding in Generation 3 and maximum house-burning in Generation 4. However, spread over 30 years, the rate of building and burning would not have produced more than minor peaks in human impact. The visitors would have supplied their own food and drink resources during the assembly period, while the small permanent resident population would have been engaged in small-scale mixed farming all-year-round. The Assembly visitors would have benefited from an intensive social life, meeting people from more home communities than would otherwise have been possible and creating alliances through the ceremonies of the start and the end of the assembly month and the frequent special events (deposition, house-burning) throughout the month. There was a tension between the overall ‘Nebelivka Identity’ and the identities of the visitors’ home communities which was never fully resolved because of the short time of the assembly period. The heterarchical social structure varied seasonally, with the group of Nebelivka Guardians controlling the overall organisation of the Assembly period from the top down but with Neighbourhoods and Quarters organised from the bottom up, having much more freedom to make their own ‘local’ decisions as regards house-building, -burning and other ceremonial events during the assembly period. 9. The Pilgrimage Model built on the pre-existing social networks linking communities across the entire Cucuteni-Trypillia world. In earlier periods, local site clustering brought together settlements in seasonal interaction, often through important ritual aspects of the Trypillia Big Other. This model extends the scale of this seasonal interaction through the decision of a group of site Guardians (cf. the Assembly



Principal Results 

 513

Model) to found a larger pilgrimage centre on the Nebelivka promontory. The biggest challenge in this model was the large-scale construction effort in the first two years of the megasite to dig the complete perimeter ditch and use the clay materials to build an entire house circuit. This ambitious decision was based on the premise that a massive communal effort would produce a particularly impressive, monumental pilgrimage centre, whose form and fame would attract many visitors and leave a lasting spiritual impression on the thousand or so builder-pilgrims. Once the pilgrimage centre’s layout was established, later construction proceeded at a gentle pace, with the start of a second house circuit in Generation 1 and the Inner Radial Streets from Generation 2. The pilgrimage season lasted for the eight snow-free months of the year, with a month’s visit from 20–100 pilgrims from as many home communities as could be accommodated up to just under 2,000 pilgrims per month. Ceremonies of arrival and departure, with appropriate visual and sound effects, were therefore regular events, as well as the larger annual ceremonies celebrating the Nebelivka identity, which were focussed on the Mega-structure and the other Assembly Houses. The importance of processions, whether executed in silence and contemplation, or with chanting and rejoicing, can be linked to the evolution of many specific megasite planning features, including the space between the two concentric house circuits, the frequent pairing of Assembly Houses and the cumulative increase in house memory mounds. The death of a Nebelivka Guardian would lead to a replacement from Quarter or Neighbourhood leaders, increasingly from home communities outside Nebelivka. 10. The importance of house architecture in Trypillia archaeology can hardly be overstated. Most excavations focus largely or completely on burnt house remains, while, before the Project, eight different experiments had been conducted on the building and burning of eleven Cucuteni-Trypillia houses. Trypillia archaeologists were amongst the earliest in Europe, if not the earliest, to recognise the deliberate burning of domestic houses. So how could an integrated study of small excavated samples of burnt and unburnt house remains and the results of an experimental programme of house-building, -burning and the excavation of the burnt remains two years after the conflagration contribute to this long-running debate? There were four major issues to which the Nebelivka Project’s experimental research has made a useful contribution: (a) the comparative interpretation of features, fittings and objects in the experimental house and Trypillia burnt houses; (b) whether the burning of experimental one- and two-storey houses left traces that would be recognisable in excavations of Trypillia ploshchadki; (c) the quantity of fuel needed for a successful house-burning; and (d) whether house-burning was a deliberate social practice. Many of the excavation features found in our Test Pits were replicated in the burnt house, including general features (the burnt mass of house daub (ploshchadka), vitrified daub) and specific construction details (wall panels, sandwich layers of two fallen walls). Wall daub could readily be differentiated from floor daub in the experimental excavation.

514 

 Conclusions

The key finding as regards the indication of one- or two-storey houses was the state of household features such as hearths, podia and platforms. The appearance of dispersed middle-floor fragments and fragments of features (especially platform fragments) in excavation is good evidence of a two-storey house, while intact or fragmented but in situ features indicate the high probability of a one-storey building. The finding of a high ratio of two-storey to one-storey houses in the test pits (5:1) shows that such building differentiation was common at Nebelivka. Because several former Trypillia house-burning experiments had failed to achieve complete combustion, a large quantity of firewood (30m3, equivalent to 420 trees 0.15m in diameter and 4m in length) was fitted into the two-storey house. It is important to note that almost 10 times the amount of wood was needed to burn a twostorey house as was used to build it. This conclusion has important implications for the question of deliberate house-burning. There are several reasons that make it highly improbable that a complete combustion of a timber-framed, wattle-and-daub house leading to the creation of both a ploshchadka and vitrified daub would have been possible through an accidental fire or even a military attack. The full implications of this important conclusion have yet to be digested in European prehistory. The implications are perhaps strongest for Balkan prehistory, in which the concept of a ‘Burnt House Horizon’ (Tringham & Krstić 1990a) has been debated since the 1980s. An example concerns the recent papers produced by the ‘Time of the Their Lives’ Project, in which there has been a consistent assumption that burnt houses were the product of attacks or accidents. The sequence of burnt and unburnt houses at tell Uivar was used as a major element in the interpretation of the whole site and of wider Vinča developments (Draşovean et al. 2017). However, the internal site changes from deliberate burning on one horizon and the decision not to burn down houses on a later horizon require other, ‘local’ explanations in terms of site dynamics which have not yet been considered. The high probability of deliberate burning of most burnt houses raises many questions in a wide variety of prehistoric contexts. 11. One of the many attractions of working with Trypillia material is the quantity and quality of the finds associated with the burnt houses which are the main focus of excavation. The spectacular painted pottery and fired clay figurines have formed the centrepiece of recent international exhibitions about the Cucuteni-Trypillia group, which have led to the recent popularization of this group. Yet the aesthetic qualities of the finds have seduced most Trypillia specialists into a reflectionist attitude to the material: viz., the finds constitute a direct reflection of the daily lifeways of Trypillia households and communities – what in Schiffer’s (1976) terms was ‘primary refuse’. We are reminded of Hayden & Cannon’s (1983) observation that “Artifact distributions in sedentary contexts provide the least reliable, most ambiguous indicators of specific activity areas, but are nevertheless the indicators most widely used” (see discussion in Chapman & Gaydarska 2007, Chapter 4). This quotation summarises the typical approach of Trypillia specialists to household finds assemblages. It is deeply ironic



Principal Results 

 515

that while these colleagues were the earliest to recognise deliberate house-burning, they did not extend this insight to how Trypillia households and their relatives and affines contributed to the house-burning ceremony with their material offerings. In other words, most Trypillia house assemblages were ‘death assemblages’ created for the deliberate burning event rather than reflections of living household practices. Likewise, the large quantities of finds found in pits derived more often from specific depositional ‘events’ than from generalised household refuse. As an alternative to reflectionism, we see the archaeology of Nebelivka as an archaeology of selective fragmentation and episodic discard/depositional practices, mediated by principles which we can glimpse but which are rarely in clear focus. A little more accessible to spatial analysis is the recognition of the varying spatial scales at which deposition took place, ranging from the individual act of placing an old and worn red deer incisor pendant in the Mega-structure before it was destroyed to the communal deposition of an estimated 322 vessels or vessel parts in that same Mega-structure. These various spatial scales of deposition were, by the same token, proxies for the inclusivity of social identities signalled by these deposited objects, as exemplified by the distribution of several painted decorative motifs across the full range of the megasite to signify the ubiquity of personal interactions on the Nebelivka promontory. 12. It is fundamental to our understanding of the Nebelivka megasite to paint a clearer picture of the social structure governing daily practices and wider social networks at the centre. The following six principles are sufficiently general to be applicable to each of the three alternative models for Nebelivka yet specific enough to make detailed proposals for future critical evaluation. The social order must be inextricably linked to the Trypillia Big Other in a reflexive relationship where the materialisation (houses, pots, figurines, etc.) should be demonstrably symbolic of that wider Trypillia social order. The social order needs to be a horizontally open kind, able to accommodate a wide variety of people from many different home communities – meaning also a lot of people – and yet create the possibility for a megasite identity to which all can build loyalty and affection. Such a social order would have privileged consensus-building over exclusionary strategies. The obvious possibilities are sodalities based upon relations between non-kin groups covering many settlements in a region or some form of descent group, such as lineages, whether segmentary or not. The social order needs to be heterarchical, avoiding more than two levels of vertical differentiation, so as to control any individual or household tendencies towards aggrandizement and/or accumulation. This principle excludes the possibility of some variant on the ‘House Society’ and should mitigate scalar stress. In terms of the corporate-exclusionary continuum, the social order is clearly closer to the corporate end. The social order needs to be nested in accordance with the subsidiarity principle (e.g., different occasions for deposition or feasting would have occurred at each nested level or at more than one nested level where appropriate). This principle would fit the

516 

 Conclusions

nested spatial order of household  – Neighbourhood  – Quarter – megasite, thereby mitigating scalar stress. The social order needs to be simple and flexible enough to cope with changing demands and challenges yet robust enough to support experimentation with problems that no-one in the Trypillia world had ever had to face before. The social order needs to include ‘safety valves’ which allow for failures, mistakes, unforeseen problems and conflicts between the same community as well as, more probably, between members of different home communities. The proxies for such social safety valves included low-density living, with many places in a huge site to escape to in case of trouble or locationally marginal houses or Neighbourhoods. The social experiment of creating a planned Trypillia megasite would have needed to integrate such principles by bricolage from previous experience, past smallerscale lifeways and on-the-ground improvisation. Whether or not these principles fit the prehistorian’s preconceptions of the form of an early, unprecedented urban social order, we submit that the bricolage of these principles would have allowed the development of a social order which lasted, on different megasites and in different places, for close to 600 years.

7.2 What People Thought About the Nebelivka Megasite We have discussed the growth of the Nebelivka megasite from what came before, the early dwelling, the development of the megasite plan, the emergence of Quarters and Neighbourhoods, the importance of Assembly Houses, all of the basic components of the plan, the transformation of the site from living houses to living- and dead-houses and the abandonment of Nebelivka. We wish to conclude this monograph with a more intimate, person-focussed narrative about the Nebelivka megasite – a narrative which is not at odds with the structural conclusions presented above but which highlights what it may have been like for people to organise the coppicing of hazel on a big scale, move onto the site and live at, or come on a monthly visit to, a place in the company of hundreds or even thousands of people from different home communities. In this section, we present the imaginative viewpoints of ten different kinds of people who engaged with the megasite in contrasting ways.

7.2.1 The Viewpoint of a Nebelivka Guardian It was always going to be difficult to persuade all of the other settlement leaders to agree on one specific place for a future centre, even though everyone agreed that it was an important task. Leaders with roots in the Southern Bug valley to the SouthWest knew dozens of promontories framed by small streams  – places replete with fertile soils and good pasture between the stands of trees. The problem was which of



What People Thought About the Nebelivka Megasite 

 517

the many places to choose. But, one Midsummer’s Eve, we were sitting round the fire in the small site South of the Nebelivka promontory and we saw shooting stars fall right on the promontory. This dazzling event made a big impression on the elders, who, there and then, decided that this was the most auspicious place for their centre. As I said, there was already an agreement between many site leaders to create something big and impressive which would attract interesting people and special objects. Now we had a site, the work of planning began. The first task was to plan the stock-piling of tools for the house-building and for digging the big ditch round the site. Builders and diggers should come with their own tools but hard work wears the tools out within a month and we needed to have more tools in reserve. Local flint was readily available from within a day’s walk but good stone for axe-making was not to be found in our local valleys. So exchange was needed, which meant objects for exchange in return for the axes had to be collected from the leaders. The keeping-back of scapulae from every bovid that was to be slaughtered henceforth would provide for shovels, while there were still enough mature red deer in the woodlands to provide antlers for picking tools. So teams of hunters started the job of killing the deer, while groups of antler-collectors collected the shed antler in late spring. Those who knew the wetlands of our valleys were recruited to identify the areas of the best-quality reeds for house-roofing, while those with potting experience dug pits to test the quality of the clay on the promontory and marked out ditch sections so as to exploit this heavy material. Elsewhere, woods-men and -women found the closest hazel stands and started the work of coppicing immediately, for it would take five years to produce a good harvest of hazel rods for wattle-making. Everyone in the small sites near the promontory was engaged in the preparations for the site and it took a lot of talking, eating and drinking to keep everyone at their tasks. Which meant a lot of food and drink production to oil the logs111 of the project. The negotiations on the size and layout of the new centre were long and hard, because all views had some merit – those wanting a smaller site would have an easier task for marking out the perimeter ditch and the areas for building, those in the middle sought a fairly impressive site with a fairly heavy planning load and those ambitious types who thought nothing of marking out a 7km perimeter with a site length of over 2km in order to make a huge impression on the clans around our valley. In the end, the maximalists won and, under protest, we agreed to create a megasite to exceed the size of any earlier site. The planning team walked the promontory many times to get to know the building site, its points of intervisibility, the breaking-points where slopes became too steep for construction and, in particular, the best clay patches which would define the line of the ditch. In the end, the areas for early house-building and the line of the ditch were marked out with vertical timbers hammered into the soil and tied with coloured threads woven on our household looms.

111  In the absence of wheels, the appropriate metaphor here is ‘logs’.

518 

 Conclusions

7.2.2 The Viewpoint of an Early House-Builder I suppose I was a good choice for a builder on the new site. With my close connections to a reliable supplier of stone axes and flint tools, I had had experience of building houses before and knew the kinds of good-quality timber, hazel rods and reeds to make a solid house. And I was also heavily indebted to our site leader for providing food and drink for the wedding of our eldest daughter to a poor young lad from the next settlement. I started off for the new site with five other builders from my settlement, carrying our tools and food and drink. We stayed overnight with relatives on the way, which took care of the food and drink we had brought. So by the time we reached the new site in the afternoon of the second day, we were ready for our first real meal of the day. Slowly that day, and for the next few days, other carpenters, wattle-makers and roofers arrived. All of us were expected, since the local group had prepared stocks of food and drink for us. The residents had also collected lots of clay, reeds and hazel withies in piles around the areas marked out for house-building. What was not ready – and this was our first job – was the construction timber. There were dense but small stands of trees in the centre of the site which became the source of the house timber. The sound of a old oak tree crashing to the ground, narrowly missing a pile of hazel rods, filled the promontory with the sad noise of growth destroyed. Many trees came down in those first few days. Once the trees were de-barked, a new noise took over – the rhythmic cutting of mature wood into beams for building. Many stone axes and chisels were destroyed in the cutting but, again, the site guardians seemed wellprepared with replacement tools. There was a strong motivation to build the houses, since there were no others for the building team to rest or sleep in. After a week of little sleep but long hours of work, the first house was ready for occupation. The building season seemed never-ending and it must have felt the same for the ditch-diggers, who exhausted their own digging tools before working through the centre’s tool reserves. The cattle consumed in the feasting provided a steady supply of scapula shovels but hunting parties had to be organised to bring back more deer antler for picks. Once, the tools were so slow in coming that a return visit was possible to my home community – then back to the job. But, as I approached what was a building site from the Southern ridge for the first time since I had arrived three months ago, I was deeply impressed with the view – a huge arc of new houses in groups and lines inside a big ditch separated by intermittent causeways. And the first part of the site was still only half-finished! To have helped to build the biggest site in the region – in our local world – this was a story to tell my children and grandchildren.



What People Thought About the Nebelivka Megasite 

 519

7.2.3 The Viewpoint of a Clan Leader It is now halfway through the year when our clan took responsibility for the provisioning of the now-complete megasite. Why are there so many new people arriving from other clan settlements? Perhaps this is not so surprising, since the word is that our site is the biggest in the whole of the Trypillia world. But the new numbers put pressure on our clan settlements to bring the food and drink to Nebelivka every fortnight. It is not so much the growing of such quantities of additional grain – after all, we have had nine years to store grain in readiness for the 10th year of clan provisioning – what we call the ‘year of commitment’. Neither is it the ‘sacrifice’ of prime beef and good sheep and goats for their walk to the megasite – this trip just needs careful guards to protect the stock against brigands. No – it is the carrying of what seems like the additional tons of grain over 60km to Nebelivka, as well as the tons of salt which we had to acquire through exchange of other clan valuables. One thing is certain  – transport would be even harder if it had not been for the invention of the sledge, which works reasonably well over grass in the snow-free months. I suppose it is inevitable that there will be protests about the extra hard work involved in the year of commitment. One household with access to less fertile land than most of the community suffered two bad crop years in a row and one of them was the year of commitment. So, of course, there were moans and groans about the unfairness, the hard life ... There were even complaints from the others about the clan support offered to the unfortunate household, even though the family knew they had to make retribution in times of future good harvests. But these protests amounted to small-scale bickering, with no sign ever of a clan-wide revolt against the year of the commitment. This is because all the clans accepted that the benefits of the system outweighed the costs – that for one hard year, you gained nine years of care from the other clans, who supported your presence at Nebelivka. I am surprised that the clan system has worked out so well at the megasite  – initially, there were fears that freeloaders would bring down the whole enterprise and that it would end in disaster. Another common concern was living with so many different clans with whom our own clan had never been on the best of terms. But we had enjoyed poor hostile relations with other clans because we had never lived, worked, co-operated and partied with them. Once you formed part of a community of over 2,500 people, you realised that other clans were not so very different from your own group and that clan symbols were simply that  – painted pottery which could have been painted in another way, with four vertical lines inside a lozenge instead of two diagonal lines in a circle. Megasite living brought a new tolerance to our clan members – a recognition of the superficiality of minor differences and the importance of genuine opportunities for interaction.

520 

 Conclusions

7.2.4 The Viewpoint of a Visitor to the Assembly As a mother of three, I live in a typical small Trypillia farming settlement of 200 or so people, where we live in 30 houses and meet with relatives and affines from two or three other similar communities a few times each year. Talking to my parents and their grandparents, this is how life was when they were growing up. As long as the fertile soil delivered its promise of a good harvest, there were few stresses in our lives, while births, marriages, serious illnesses and deaths provided the surprises and the peaks of excitement, enjoyment or sorrow of our lives. Perhaps the greatest moment in community life was a house-burning; everybody in the community helped to gather the fuel for a successful burning, which provided a spectacular event for the day. So when news filtered through from the next community of the building of a huge new Assembly site only 40km from our settlement, there was a palpable sense of anticipation for new experiences, especially meeting new people and maybe seeing new kinds of objects made by skilled people of the kind that we did not have in our midst. The offer of my parents to look after the children released me to go with our community group of 30 people to make our community’s first visit to the assembly site and report back. We were aware that our contribution to the new centre was to build a series of four new houses in our month’s visit, so we took our own tools as well as food and drink, two cows and gifts to the residents of Nebelivka. The first surprise on our two-day journey was the number of other people following the stream-side tracks Southwards to Nebelivka. Even if no-one from our group knew the whereabouts of the promontory (in fact, we did have one such person), we could not possibly have got lost – we just needed to follow the crowd. It did not take long to find affines who had visited our settlement in the past – people whom we could co-operate with on the megasite and whose presence removed any sense of fear or apprehension about whom we may find at the assembly. Our arrival at Nebelivka was preceded by views of the site from the Northern side, which showed us the full width of the promontory site, with its dominant circuit of 70 or 80 new houses, often two-storied, and including some houses of a size I had never seen before. The Nebelivka folk were seemingly well-prepared for so many visitors, for there were people to guide us round the Northern house circuit towards the West entrance. We were sent to a big triangular area near a stream with a source seemingly in the centre of the megasite and asked to settle down until the time for an evening meal. This triangular area was many times larger than our settlement, yet it made up only one small part of the assembly site. It was the place where our group, and those from nearby settlements, would live and build their houses.



What People Thought About the Nebelivka Megasite 

 521

7.2.5 The Viewpoint of an Organizer of the Assembly One month to go before the assembly begins! Will it be better organized than last year’s Assembly? I certainly hope so – to run out of food and drink as well as building resources half way through the festival because of the surge of visitors was, frankly, an embarrassment ... It is great that the Assembly has proven to be such a success but not so good that we were overwhelmed by sheer numbers. Luckily, the potential for frustration among the visitors, which could have boiled over into violence, was contained and the basically good-humoured assembly crowd were remarkably tolerant of our failings. We could have countered that visitors were supposed to bring more resources of their own but I doubt that visitors would have appreciated such a reply. Better to turn our attention to the following year and see what we could improve on. The site guardians will need to get help to bring more forest resources (building timber, withies) and other supplies (clay, reeds) in advance to the actual building plots where the new houses will be erected. Even if they do not bring all the materials for each house, this work would make a huge difference to the assembly’s building programme of perhaps 40 houses in one month if a lot more material is brought to the site next year. We shall also need to stockpile more sledges, lithic tools and stone axes for transport and construction. It will also be important to improve ways of guiding new arrivals to the parts of the megasite where they need to build their houses. We thought we could count on second-time visitors being able to show new arrivals around the megasite and ensure they ended up in their Neighbourhood. This did not happen. Since we cannot rely upon experienced visitors, we shall have to mobilise more assembly helpers to meet and greet new arrivals and escort them to their building site, where some of their affines and relatives should already be in place. Most importantly, we need to manage the supply of food and drink much better than last year. It is very bad to have a rumour circulating that Nebelivka ran out of food and drink for the main events of their assembly for the second year running. I think that the home communities will not object overmuch to bringing extra food on the hoof to the assembly, by way of one additional bull and three extra caprines per community; an extra animal from each household of visitors would, however, probably be too much to ask. The site guardians will need to invest in the building of extra communal baking facilities for the obvious higher demand for unleavened bread at the main ceremonies. If we can resolve these issues, I am confident that the Nebelivka assembly will become well established in the region as the key event of our annual social calendar.

522 

 Conclusions

7.2.6 The Viewpoint of a Pilgrim I am a figurine-maker from a home community almost 80km from Nebelivka and, as such, am fully committed to the system of religious beliefs in which figurines play an important role in ceremonies112. My sister and I joined a small pilgrim group of 20 souls from our settlement, partly to intercede in a healing ritual for her cranial disfigurement, partly as an act of devotion that would strengthen our belief and practices. Although we were not able to build houses, we could prepare food and drink for our brethren. The scale of the pilgrimage was such that neither my sister nor I had ever experienced before. In the summer month of July when we visited, there were almost 2,000 pilgrims at the centre. Two massive ceremonies of arrival and departure framed the visit, using the huge inner open space which occupied an area 10 times the size of our home community. In addition, there were weekly ceremonies based upon processional routes and Assembly Houses. The processions wound their way along the perimeter ditch, then outside the outer house circuit, then between the two house circuits and then between two inner radial streets to reach the inner open area. More intimate rituals for smaller numbers of important ritual leaders were held in the Assembly Houses, which were located at intervals along the processional route. There were also periodic rituals based on the houses where the pilgrims dwelt, including healing rituals where pilgrims were invited to deposit figurines or figurine parts representing the ill pilgrim. As a result of my placing a realistic image of my sister in our dwelling house, my sister’s health improved and she had fully recovered by the time we had returned to our home. The sense of solidarity with other pilgrims, the scale of ceremonial shared with so many other co-believers and the sheer size of the centre itself – they say that it is the largest pilgrimage centre in the Trypillia world – these were what made the Nebelivka pilgrimage so special to me and my sister. We may never return to the pilgrimage centre again but it has had a profound effect on my life as a pilgrim and an explicitly beneficial effect on the health of my sister. I encourage every person who is committed to the faith to visit the Nebelivka centre.

7.2.7 The Viewpoint of an Adolescent Visiting Nebelivka for the First Time My parents had been to visit the Nebelivka assembly several times, always leaving me behind with my grandparents, whom I loved deeply but who were a poor substitute for a vast assembly. Finally, once I had reached my 13th birthday, I was allowed to accompany my parents to the assembly, together with my best girl-friend from the

112  The system which the Project refers to as the’Trypillia Big Other’.



What People Thought About the Nebelivka Megasite 

 523

settlement. This would be the first long trip that we had ever made outside our settlement. We reached the assembly site after a long, dry and dusty 20-km walk in which the settlement group set what seemed to us to be far too quick a pace but which was designed to bring us to the site by the evening of a single day. An unbelievable sight greeted us  – a vast central site many times bigger than our settlement, completely illuminated by hundreds of campfires lit up across the site. Hundreds of people were present and yet not present – in the shadows, their faces invisible but playing their part in the creation of an atmosphere that was quite new to me and utterly fascinating. The talking and the singing – the playing of pipes, each with a different tune, at many places on the site – together produced not so much a cacophony as a welcoming wall of sound which I had never heard before. So this was an assembly site. The biggest difference from our home settlement was that something different happened every day, with the biggest impact being that I met different young people from new villages every day. Some of them dressed in completely different ways from the people in my community, others spoke with the same language but used odd words, often with a slightly different accent, and still others looked different, with faces the like of which I had never seen before, with different hairstyles and personal ornaments. But because this was an assembly site, it was safe to meet these different kinds of people, talk to them and get to know them in a way that my parents would never have permitted back in our settlement. Some of the meetings were open and in the daytime, others were surreptitious, at night, covered by the shadows of the houses cast by the bonfires. It is hardly surprising that mutual boy  – girl, boy  – boy and girl – girl attractions started up in such a magical place. Love-gifts were exchanged in the last evening, to be hidden on the walk back and treasured in secret at home until another assembly. Would my special friend from over the river come back? Would I ever be allowed to pay another visit?

7.2.8 The Viewpoint of a Nebelivka Ritual Leader I was a ritual leader in a small settlement before moving permanently to the Nebelivka centre as one of the site guardians, where I helped to create the context for a far wider range of ceremonies than any ritual specialist had ever participated in before. We managed to do this by drawing on a small number of basic ritual sequences, which could then be elaborated or re-combined so as to work at very different scales of participation, whether a single household, an Assembly House, a Neighbourhood or a Quarter. Many of the rituals which we led were staged in the Mega-structure, where the many different platforms created stages for different presentations, observed by different groups. I used the ritual board game in the Mega-structure with several tokens brought in by visitors to the megasite; at the end of the game, the players took

524 

 Conclusions

their tokens away and I ritually smashed the board. One of the special decorations which I would wear for such ritual games was the gold hair-ornament which I lost in the Mega-structure during a performance. By far the largest ceremony centred on the burning down of the Mega-structure, which involved collecting sherds or vessels from all over the megasite to create a massive offering before starting two fires – one in the Eastern rooms and one in the South-West corner. We transferred the same form of rituals to many other megasite contexts so as to enable the residents or visitors to familiarise themselves with the ritual for themselves. We made this particularly effective in house-burning ceremonies, which acted as a ritual magnet for an entire Neighbourhood or Quarter and helped to integrate the residents through a highly emotional performance. Another form of ritual which we adopted for many different contexts concerned the rituals of arrival and departure which framed group visits to Nebelivka. These often local rituals made most sense through depositional events involving pit-digging and re-filling. Many different groups participated in such rituals, with each group using the materials they had at hand, often markedly different from the materials for a nearby pit deposition-event. One of the most specific rituals which we created for the pilgrim visitors to Nebelivka involved the staging of offerings in Assembly Houses during important processions through the megasite. The ritual took the form of blessing the objects which the pilgrims had brought so that they could take part of the Nebelivka blessing back home with them.

7.2.9 The Viewpoint of a Trader Visiting Nebelivka at Assembly Time I am quite good with people and languages and this has helped me to facilitate trade and exchange between people from different backgrounds in the widespread CucuteniTrypillia networks which linked hundreds of communities. My mobile life has taken me to many places, from the Black Sea to the peaks of the Eastern Carpathians and the Dnieper Rapids. I have seen many sights, including salt mountains and salt lagoons, many strange species, including lions on the Black Sea shore and strange red birds in the Black Sea, and many settlements, including great dwelling mounds and unconquerable hilltop fortifications, but I have never experienced anything quite like Nebelivka. It is so obvious to me: since it is people who do exchange, the megasite is the greatest centre for my livelihood – there may be 2,000 people on site in the fortnight that I visited. Unlike many people from small communities, I actually enjoy meeting strangers – it is one of the pleasures of a mobile life – so Nebelivka is a paradise for a trader, an unrivalled opportunity to make contacts and set up future exchanges on an undreamt-of scale. You can imagine why I don’t carry much ‘stock-in-trade’ with me: it’s too dangerous for a solo trader, who can be attacked and robbed anywhere along the



What People Thought About the Nebelivka Megasite 

 525

forest track. So I carry just a few shiny, colourful, attractive but especially light objects to start a trading relationship. I usually have in my bag some copper ornaments, a red deer canine pendant or two, several flint cores from the Prut valley, a few marble or limestone beads and, occasionally (they are so rare nowadays) a marine shell bracelet. I used to carry gold ornaments too but word got around and, after two muggings, I abandoned the idea. But what I do have attracts the attention of Trypillia people and then I can bring the new partner what they specially desire on a return trip. Here in the Nebelivka assembly, I often need to exchange a copper item for a good dinner but there are regular exchange partners at the megasite at the same time who will invite me to their feast to keep my attention. Until now, I used to go from community to small community, sleeping at each site for a few nights, exchanging stories, spreading gossip about families living on nearby sites and trading ornaments and tools. This was good for obtaining a diversity of things for exchange but it is a hard life. The Nebelivka assembly makes me think of a new trading plan – stocking up with more exchange goods and visiting the assembly every year, making it my main exchange event. I would never ask the guardians to let me settle down at Nebelivka – for now, I still prefer life on the road – but that centre, and others which may develop, would allow me to travel less and meet more people. I’ll continue site  – to  – site trading for another year or two and then make up my mind.

7.2.10 The Viewpoint of One of the Last Generations of Residents at Nebelivka Everyone as old as me – and I’m over 50 now – harps on about the ‘good old days’ – how things were better then than now (better beef goulash, tastier borsch, bigger portions of venison, more beer). And also how you could get firewood in a 15-minute walk, whereas now you’d be lucky to find some in an hour. It’s not surprising that there are fewer house-burning ceremonies nowadays – no-one can find enough fuel for a proper fire and there’s rarely a spare bull that can be sacrificed for a good funeral feast. The last time the Eagle Clan had their ‘year of commitment’, there was an unprecedented disaster, with several other clans having to bring their own food and drink from their home communities. No-one really seemed to know the cause of it but the rumours were that half of the Eagle Clan communities were in favour of paying their dues while the other half were opposed to the idea. If a clan is split down the middle, it is hard to find a compromise. So the supporters had to work extra hard to supply the megasite and, of course, they couldn’t manage the additional load. There were too many objectors (I would call them ‘backsliders’) for the supporters to impose any sanctions and the rule of the Big Other is not what it was. Something else has changed in these valleys recently. Once Nebelivka was the only major centre for clan meetings but now one or two new centres have emerged, and less than two days’ walk from Nebelivka at that. The leaders of the new centres

526 

 Conclusions

are, of course, opposed to each other’s success but this is nothing compared to the rivalries they provoke with the established centre at Nebelivka. There are no doubt many inducements to persuade clan leaders to abandon the old centre and join a new megasite. New arrangements for house-building with construction timber on site rather than 10km away; a lower proportion of food and drink for the clan to supply for the first 10-year cycle; new metal status symbols for clan leaders to take home and display in their own domestic ceremonies, etc., etc. The new leaders are doubtless also skilful in dredging up old memories of less-than-happy times at Nebelivka, as well as in exploiting those divisions in the clans which have already threatened the social order of almost two centuries. It is easy to forget that many 10-year cycles had already been successfully completed at Nebelivka until the first, recent disaster. But, as the old Trypillia saying goes, the grass is always greener on the other side of the river. As someone who has lived all my life at Nebelivka in the family of a site guardian, my own fear is that the megasite may soon disappear. And then what would I do? While investigating a site as large and complex as that of Nebelivka, it is easy to overlook the people who lived there permanently or paid seasonal visits to the centre. We hope that our small cast of ten ‘representatives’113 has brought more life to the Project findings by showing the human face of meeting at such a centre.

7.3 A Future Research Agenda Every interesting research project raises more questions than it can answer. The questions which we have posed and still not answered vary from detailed questions about individual objects or events (see the list of eight ‘inexplicable’ occurrences at the start of Section 5.5) to general issues affecting all Trypillia settlements (e.g., the problem of the wiggle on the calibration curve at the start of the 4th millennium BC). We have narrowed down these questions for the future to five issues for the Trypillia group in general and three issues pertaining specifically to Nebelivka.

7.3.1 Issues for Trypillia Studies First, in our 2014 article (Chapman et al. 2014b, p. 398), we predicted that the new generation of high-precision geophysical investigations would create a new research agenda for field investigations that would last two decades. We are delighted that this prediction is already being fulfilled, with new excavations of Assembly Houses at two more megasites – Majdanetske and Dobrovodi (Müller et al. 2018; Hofmann et al. 2019).

113  All of the characters are imaginary and any resemblance to a living individual is purely fortuitous.



A Future Research Agenda 

 527

Careful stratigraphic excavation of large pits has also been built into the Majdanetske project (Müller & Videiko 2016). The excavation of kilns has made perhaps the greatest progress in the last five years (Korvin-Piotrovskiy et al. 2016), while the method of test-pitting to recover a series of samples for AMS dating has also been adopted at Majdanetske (Müller et al. 2017) and is urgently needed for Taljanki. We are still poorly informed about the form and content of the smaller pits and, especially, the pit groups. And there is much to be learned about the entrances to megasites (the breaks between ditch sections) and, especially, the ditches themselves. Long ditch exposures of the kind favoured at Rondels such as Svodín, Slovakia (Němejcová-Pavúková 1995) or causewayed enclosures such as Etton, England (French & Pryor 2005) would yield vital information about the practices of ditch-digging, re-cutting episodes and the deposition of finds in ditches. The second gap in our understanding of Trypillia megasites concerns the settlement patterns in the megasite hinterlands. The current project was the first to include intensive, systematic as well as targeted fieldwalking in their research design, with vital results for our understanding of the Nebelivka settlement. A positive sign is that further fieldwalking has already begun in the Vinnitsa region (V. Rud, pers. comm.). Given the large size of megasite territories, a huge effort will be required for this task. Thirdly, while we have a broader range of general vegetational histories for the Ukraine than we had a decade ago (e.g., Harper 2016, 2019; Pashkevych 2012; Shumilovskikh et al. 2017), targetted palaeo-environmental investigations near megasites are still very rare. While samples collected from archaeological contexts can provide useful data (e.g., Kirleis & Dal Corso 2016; Kirleis & Dreibrodt 2016), it is essential to recover long, well-dated sediment cores from wetlands close to megasites. It is not an exaggeration to say that no other information will confirm or disprove the alternative hypotheses about the size of megasite populations and their alleged effects on the local forest steppe. Fourthly, following a central decision by the directorate of NAS Institute of Archaeology, the vast majority of financial resources devoted to the understanding of Trypillia megasites has been channelled into intensive excavations at only two megasites  – Taljanki and Majdanetske. The third megasite under intensive investigation is Nebelivka. It is fundamental to explaining the origins of megasites in the Trypillia Phases BI and BI/II that complex inter-disciplinary investigations are targeted at smaller and earlier Trypillia sites in the Southern Bug-Dnieper Interfluve (e.g., the Mogylna sites, Onopriivka and Vesely Kut). The fifth point concerns the whole of world archaeology, not only the Trypillia group, and is related to the promised improvements to the radiocarbon calibration curve (Alex Bayliss, pers. comm.). These improvements should reduce the deleterious effects of the wiggles, which in our specific case, may help to overcome the problems of early 4th millennium BC chronology and help achieve the Project’s only unfulfilled objective so far – the modelling of a tight internal Nebelivka chronology.

528 

 Conclusions

7.3.2 Issues for Nebelivka First, we have proposed three alternative models of smaller-scale, sustainable megasite dwelling – namely the Distributed Governance Model, the Assembly Model and the Pilgrimage Model. In the current state of research, we do not find it possible to discard any of these three models. The models are only to a certain degree overlapping, so it is unsustainable to continue to support all three models. However, this part of the Nebelivka research agenda requires further elaboration in future years. The second point relates to the experimental house-building programme. The comparative element of the programme was vitiated for good village political reasons. However, were the political situation in Nebelivka to change, any future opportunity to burn the one-storey house and excavate its burnt remains would be beneficial in allowing the comparative examination of both houses. The third issue concerns artifact studies. Very few characterization studies have yet been performed on the Nebelivka samples of pottery, figurines and ground stone, not to mention the only gold object currently known from the Trypillia group. Characterization studies would strengthen our understanding of the exchange networks which linked Nebelivka to the rest of the Trypillia world. A single detail about exchange concerns the discovery of graphite-painted decoration on pre-Trypillia pottery in the forest steppe zone and on some of the Trypillia pottery at Nebelivka. There is an urgent need for a programme characterizing the most important graphite sources in Southern Ukraine and the related pottery, both Neolithic and Trypillia.

7.4 Endwords To conclude the conclusion, the Project would like to repeat its thanks to all of the individuals, groups and institutions who have made this research so productive over the last decade. If the development of new and ‘unacceptable’ (i.e., challenging) ideas and hypotheses have provoked strong reactions and led to the breaking of partnerships or friendships, we can only state that, although we had hoped that there would not be a choice, the honest pursuit of a better understanding of the past is more important than personal relations. We hope that any intellectual failings published here are treated with respect.

References Adams, R. (1965). Land Behind Baghdad: A History of Settlement on the Diyala Plains. Chicago: The University of Chicago Press. Agapiou, A., Hadjimitsis, D., Sarris, A. et al. (2013). Optimum Temporal and Spectral Window for Monitoring Crop Marks over Archaeological Remains in the Mediterranean Region. Journal of Archaeological Science, 40 (3), 1479–92. Retrieved doi:10.1016/j.jas.2012.10.036. Agnoletti, M. & Emanueli, F. (2016). Biocultural Diversity in Europe. Environmental History 5. Berlin: Springer. Agnew, J. (2011). Space and place. In J. Agnew & D. Livingstone (Eds.), Handbook of Geographical Knowledge (pp. 316–331). London: Sage. Albert, B., Innes, J., Kremenetskiy, K. et al. (2020). What was the ecological impact of a Trypillia mega-site occupation?: multi-proxy palaeo-environmental investigations at Nebelivka, Ukraine. Vegetation History and Archaeobotany, 29, 15–34. Retrieved doi: https://doi.org/10.1007/ s00334-019- 00730-9. Albert, B. & Pokorný, P. (2012). Pollen taphonomy at Vranský potok versus Zahájí alluvial pollen sites: Methodological implications for prehistoric Cultural Landscape Reconstruction in the Peruc Sandstone area, Czech Republic. Interdisciplinaria Archaeologica, 3, 27–44. Alcock, S. & Cherry, J. (Eds.) (2004). Side-by-Side Survey: Comparative Regional Studies in the Mediterranean World. Oxford: Oxbow Books. Allison, P. (Ed.) (1999). The archaeology of household activities. London: Routledge. Al-Sayed, K. (2014). Space syntax methodology. London: Bartlett School of Architecture. Anderson, B. (1991). Imagined Communities. Reflections on the origins and spread of nationalism. London & New York: Verso, 2nd edition. Andreescu, R. & Lazăr, C. (2008). Valea Mostiştei. Aşezarea gumelniţeană de la Sultana-Malu Roşu. Cercetări Arheologice, XIV–XV, 55–76. Andreescu, R., Mirea, P. & Apope, S. (2003). Cultura Gumelniţa în vestul Munteniei. Aşezarea de la Vităneşti, Judeţul Teleorman. Cercetări Arheologice, XII, 71–87. Anselin, L. (1995). Local Indicators of Spatial Association — LISA. Geographical Analysis, 27 (2), 93–115. Retrieved doi:10.1111/j.1538-4632.1995.tb00338.x. Anthony, D. (2007). The horse, the wheel and language. How Bronze-Age riders from the Eurasian steppes shaped the modern world. Princeton: Princeton University Press. Aqdus, S., Hanson, W. & Drummond. J. (2012). The Potential of Hyperspectral and Multi-Spectral Imagery to Enhance Archaeological Cropmark Detection: A Comparative Study. Journal of Archaeological Science, 39 (7), 1915–24. Retrieved doi:10.1016/j.jas.2012.01.034. Arbeiter, S. (n.d.). Analysis of Tripillia pottery decoration from the site of Nebelivka, Ukraine. Unpub. B. A. Dissertation, Durham University (2014). Aslan, C. (2006). Individual, household, and community space in early Bronze Age western Anatolia and the nearby islands. In E. C. Robertston, J. D. Seibert, D. C. Fernandez & M. Zender (Eds.), Space and Spatial Analysis in Archaeology (pp. 133–142). Calgary: University of Calgary Press. Aslanis, I. (1989). Das Neolithikum und das Chalcolithikum im nordgriechischen Raum. In S. Hiller (Ed.), Tell Karanovo und das Balkan Neolithikum (pp. 65–81). Salzburg: Institut fur Klassische Archäologie der Universität Salzburg. Ashmore, P. (1999). Radiocarbon dating: avoiding errors by avoiding mixed samples. Antiquity, 73, 124–130. Attwell, M. & Fletcher, M. (1987). An Analytical Technique for Investigating Spatial Relationships. Journal of Archaeological Science, 14 (1), 1–11. Retrieved doi:10.1016/S0305-4403(87)80002-X. Bailey, D. (2000). Balkan Prehistory. Exclusion, Incorporation and Identity. London: Routledge.

530 

 References

Bailey, D. (2010). The figurines of Old Europe. In Anthony, D. W. with Chi, J. Y. (Eds.), The lost world of Old Europe. The Danube Valley, 5000 - 3500 BC (pp. 112–127). New York: The Institute for the Study of the Ancient World. Bailey, D., Whittle, A., & Cummings, V. (Eds.) (2005). (Un)settling the Neolithic. Oxford: Oxbow Books. Bailey, T. & Gatrell. A. (1995). Interactive Spatial Data Analysis. Harlow: Longman Scientific & Technical. Balasse, M., Bălăşescu, A., Janzen, A. et al. (2013). Early herding at Măgura-Boldul lui Moş Ivănuş (early sixth millennium BC, Romania): environments and seasonality from stable isotope analysis. European Journal of Archaeology, 16/1, 1–26. Bánffy, E., Osztás, A., Oross, K. et al. (2016). The Alsónyék story: towards the history of a persistent place. Bericht der Römisch-Germanischen Kommission, 94, 283–318. Barral, P. & Luginbühl, T. (1995). Typologie des formes de céramique régionale de Bibracte. Version 1.1. Glux-en-Glenne: Centre Européen d’Archéologie du Mont Beuvray, Université de Lausanne & Université de Franche-Comté. Barrett, J. (1994). Fragments from antiquity. An archaeology of social life in Britain, 2900 - 1200 BC. Oxford: Blackwell. Bayliss, A. (2013). Bayesian chronological modelling. In J. Hines & A. Bayliss (Eds.), Anglo-Saxon graves and grave-goods of the 6th and 7th centuries AD: a chronological framework (pp. 73–87). London: Society for Medieval Archaeology. Bejenaru, L. & Stanc, S. (2011). New archaeozoological data concerning the Cucuteni A settlement of Poduri-Dealul Ghindaru (Bacău county, Romania). Faculty of Biology, Universităţii “Al. I. Cuza” Iaşi. Bejenaru, L. & Stanc, S. (2012) Animal Resources in the Chalcolithic subsistence economy: settlements of Cucuteni culture in Eastern Romania. Seventh World Archaeological Congress: “Al. I. Cuza” Iaşi. Bejenaru, L., Stanc, S., & Cavaleriu, R. (2011). Preliminary Analysis of an Archaeozoological Assemblage Discovered in the Cucuteni B Settlement from Sarata-Monteoru (Buzau County). Universităţii “Al. I. Cuza” Iaşi. Bem, C. (2001). Consideraţii generale asupra cercetărilor arheologice pluridisciplinare de pe valea Neajlovului. In S. Marinescu-Bîlcu (Ed.), O civilizaţie “necunoscută”: Gumelniţa. cd-rom. Bucharest: cIMeC. Bem, C. (2007). Traian Dealul Fântânilor. Fenomenul Cucuteni A-B. Monografii V. Bucureşti: Muzeul Naţional de Istorie a României. Benedikt, M. (1979). To take hold of space: Isovists and isovist fields. Environment and Planning B Planning and Design, 6, 47–65. Bentley, A. & Maschner, H. (2003). Complex Systems and Archaeology: Empirical and Theoretical Applications. Salt Lake City: The University of Utah Press. Berezanskaya, S. (Ed.) (1971). Arheologia Ukrainskoi RSR. Kyiv: Naukova Dumka. Bernardini, W. & Schachner, G. (2018). Comparing Near Eastern Neolithic megasites and Southwestern pueblos: population size, exceptionalism and historical trajectories. Cambridge Archaeological Journal, 28 (4), 647–663. Bertok, G. & Gáti, Cs. (2014). Old times – new methods. Budapest: Archaeolingua. Bevan, A., Crema, E., Li, X., & Palmisano, A. (2013). Intensities, Interactions, and Uncertainties: Some New Approaches to Archaeological Distributions. In A. Bevan & M. Lake (Eds.), Computational Approaches to Archaeological Spaces (pp. 27–54). London: Institute of Archaeology Publications. Bevan, A., Jobbová, E., Helmke, C., & Awe, J. (2013a). Directional Layouts in Central Lowland Maya Settlement. Journal of Archaeological Science, 40 (5), 2373–83. Retrieved doi:10.1016/j. jas.2013.01.011.

References 

 531

Bibikov, S. (1953). Poselenie Luka-Vrublevetskaya na Dnestre. MIA, 38. Moskva. Bibikov, S. (1965). Hoziajstvenno-Ekonomicheskij Kompleksa Razvitogo Tripolja. Sovetskaya Archeologiya, 1, 48–62. Bibikov, S. (1970). Domashnie promisli i remeslo. Tezisi rasshirenogo zasedaniya sektora Srednei Azii i Kavkaza. LOIA, 3–6. Bickle, P. (2013). Of time and the house: The Early Neolithic communities of the Paris Basin and their domestic architecture. In D. Hofmann & J. Smyth (Eds.), Tracking the Neolithic house in Europe (pp. 117–149). London: Springer. Binford, L. (1964). A Consideration of Archaeological Research Design. American Antiquity, 29 (4), 425–41. Bintliff, J. (2000). The Concepts of ’Site’ and ’off-Site’ Archaeology in Surface Artefact Survey. In M. Pasquinucci & F. Trément (Eds.), Non-Destructive Techniques Applied to Landscape Archaeology (pp. 200–215). Oxford: Oxbow Books. Bintliff, J. L. & Snodgrass, A. (1988). Off-Site Pottery Distributions: A Regional and Interregional Perspective. Current Anthropology, 29 (3), 506–513. Retrieved doi:10.1086/203668. Bojadjiev, J., Dimov, T., & Todorova, H. (1993). Les Balkans orientaux. In J. Kozłowski (Ed.), Atlas du Néolithique européen. Vol. 1: L’Europe orientale (pp. 81–82). Liège, ERAUL. Bourdieu, P. (1977). Outline of a theory of practice. Cambridge: Cambridge University Press. Bourdieu, P. (1985). The genesis of the concepts of habitus and of field. Sociocriticism, 2 (2), 11–24. Bourdieu, P. (1988). Social space and symbolic power. Sociological theory, 7 (1), 14–25. Bowen, H. (1962). Ancient Fields: A Tentative Analysis of Vanishing Earthworks and Landscapes. London: The British Association for the Advancement of Science. Bowles, S., Smith, E., & Mulder, M. (2010). The Emergence and Persistence of Inequality in Premodern Societies. Current Anthropology, 51 (1), 7–17. Retrieved doi:10.1086/649206. Bowser, B. & Patton, J. (2004). Domestic spaces as public places: An ethnoarchaeological case study of houses, gender, and politics in the Ecuadorian Amazon. Journal of Archaeological Method and Theory, 11, 157–181. Braaksma, J. & Cook, W. (1980). Human orientation in transportation terminals. Journal of Transportation Engineering, 106, 189–203. Braasch, O. (1995). 50 Jahre Verloren. In O. Braasch (Ed.), Luftbildarchäologie in Ost- und Mitteleuropa. Forschungen zur Archäologie im Land Brandenburg, 3, 109–22. Bradley, R. (2005). Ritual and domestic life in prehistoric Europe. London: Routledge. Braudel, F. (1975). The Mediterranean and the Mediterranean world in the age of Philip II. (translated by S. Reynolds). Collins: Fontana. Brill, J. B. (2016). The technology of violence and cultural evolution in the Santa Barbara Channel area. In M. W. Allen & T. L. Jones (Eds.), Violence and warfare among hunter-gatherers (pp. 314–332). London: Routledge. Brock, F., Higham, T., Ditchfield, P., & Ramsey, C. B. (2016). Current Pretreatment Methods for AMS Radiocarbon Dating at the Oxford Radiocarbon Accelerator Unit (ORAU). Radiocarbon, 52 (1), 103–112. Retrieved doi: 10.1017/S0033822200045069 Bronk Ramsey, C. (2008). Deposition models for chronological records. Quaternary Science Reviews, 27 (1–2), 42–60. Bronk Ramsey, C. (2009). Bayesian analysis of radiocarbon dates. Radiocarbon, 51, 337–360. Bronk Ramsey, C. (2009a). Dealing with outliers and offsets in radiocarbon dating. Radiocarbon, 51, 1023–1045. Bronk Ramsey, C. & Lee, S. (2013). Recent and Planned Developments of the Program OxCal. Radiocarbon, 55, 720–730. Brookes, S., Tente, C., & Prata, S. (2017). Interpreting Rock-Cut Grave Cemeteries: The Early Medieval Necropolis and Enclosure of São Gens, Portugal. Medieval Archaeology, 61, 215–238. Retrieved doi.org/10.1080/00766097.2017.1374093

532 

 References

Brughmans, T., Keay, S., & Earl, G. (2015). Understanding Inter-settlement Visibility in Iron Age and Roman Southern Spain with Exponential Random Graph Models for Visibility Networks. Journal of Archaeological Method and Theory, 22, 58–143. Retrieved doi.org/10.1007/s10816014-9231-x Brumfiel, E. & Earle, T. (1987). Specialization, Exchange and Complex Societies: an Introduction. In E. M. Brumfiel & T. Earle (Eds.), Specialization, Exchange and Complex Societies (pp. 64–75). Cambridge: Cambridge University Press. Buchanan, B. (2015). Households, Settlements, and Landscapes in Iron Age, Roman, and Early Medieval Northumbria: A Spatial Analysis of North-East England, c. 100 BC-AD 800. Unpub; Ph.D., Durham University. Buchanan, B. (2017). Space, Place and Transition: An Introduction to the Use of Visibility Graph Analysis for Settlement Analysis in North-East England c. AD 350-800. In S. Semple, C. Orsini & S. Mui (Eds.), Life on the Edge: Social, Religious and Political Frontiers in Early Medieval Europe (pp. 263–270). Neue Studien zur Sachsenforschung. Braunschweigisches Landesmuseum in connection with the Internationales Sachsensymposion. Burdo, N. (2008). Sakral’nii svit Tripil’skoi tsivilizatsii. Kyiv: Nash Chas. Burdo, N. (2011). Rekonstruktsii domov i poselenij tripolskoj kultury: ot risunkov V. L. Hvoiki k 3D modelyam. Archaeology and Ancient History of Ukraine, 5, 38–52. Burdo, N. (2015). Antropomorfna plastika z Nebelivki. In M. Videiko, J. Chapman, I. Kozir & V. Sobchuk (Eds.), At the Eastern frontiers of Old Europe. Conference programme and abstracts, Kirovograd International Conference, 12th - 14th May 2015 (pp. 29–32). Kirovograd: Kirovograd County Council. Burdo, N. & Videiko, M. (2016). Nebelivka: From Magnetic Prospection to New Features of Megasites. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 95–116). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Burdo, N., Videiko, M., Chapman, J., & Gaydarska, B. (2013). Houses in the archaeology of the Tripillia-Cucuteni groups. In D. Hofmann & J. Smyth (Eds.), Tracking the Neolithic house in Europe. Sedentism, architecture and practice (pp. 95–116). New York: Springer. Burens, A., Ailincai, S., Micu, C., et al. (2010). Premières observations sur les techniques de façonnage et de finition de la céramique chalcolithique Gumelniţa A2 du site de Carcaliu (Dobroudja, Roumanie). Studii de Preistorie, 7, 95–123. Buzgar, N., Apopei, A. I., & Buzatu, A. (2013). Characterization and source of Cucuteni black pigment (Romania): vibrational spectrometry and XRD study. Journal of Archaeological Science, 40, 2128–2135. Buzgar, N., Bodi, G., Buzatu, A.I., et al. (2010). Raman and XRD studies of black pigment from Cucuteni ceramics. Analele Stiinţifice Ale Universităţii “Al. I. Cuza” Iaşi: Geologie, 56, 95–108. Campbell, R. (2009). Toward a Networks and Boundaries Approach to Early Complex Polities: The Late Shang Case. Current Anthropology, 50 (6), 821–48. Cartwright, N. & Runhardt, R. (2014). Measurement. In N. Cartwright & E. Montuschi (Eds.), Philosophy of Social Science: A New Introduction (pp. 265–287). Oxford: Oxford University Press. Caswell E., Arbeiter S., Ovchinnikov E., Ponroy C., & Chapman J. (2016). The Speaking Fragments: what Trypillia Sherds can Tell., Arkheolohiya i davnya istoriya Ukrayiny (Institute of Archaeology, Kyiv) 1/ 18, 5–14. Chabaniuk, V. (2008). New experimental data on the Tripolye house construction techniques. In A. Korvin-Piotrovskiy & F. Menotti (Eds.), Tripolye culture in Ukraine: the giant-settlement of Talianki (pp. 211–226). Kiev: Institute of Archaeology.

References 

 533

Chapman, J. (1988). Putting pressure on population: social alternatives to Malthus and Boserup. In J. B. Bintliff, D. A. Davidson & E. G. Grant (Eds.), Conceptual Issues in Environmental Archaeology (pp. 291–310). Edinburgh: Edinburgh University Press. Chapman, J. (1988a). From "space" to "place": a model of dispersed settlement and Neolithic society. In C. Burgess, P. Topping & D. Mordant (Eds.), Enclosures and defences in the Neolithic of western Europe (pp. 21–46). International Series 403. Oxford: BAR. Chapman, J. (1990). The Neolithic in the Morava-Danube confluence area: a regional assessment of settlement pattern. In R. Tringham & D. Krstić (Eds.), Selevac. A Neolithic village in Yugoslavia (pp. 13–44). Monumenta Archaeologica 15, Los Angeles: University of California Press. Chapman, J. (1990a). Social inequality on Bulgarian tells and the Varna problem. In R. Samson (Ed.), The social archaeology of houses (pp. 49–98). Edinburgh: Edinburgh University Press. Chapman, J. (1992) The Creation of Arenas of Social Power in Serbian Prehistory. Zbornik Narodnog Muzeja Beograda, 14, 305–17. Chapman, J. (2000). Fragmentation in archaeology. London: Routledge. Chapman, J. (2002). Domesticating the exotic: the context of Cucuteni-Trypillia exchange with steppe and forest-steppe communities. In K. Boyle, C., Renfrew & M. Levine (Eds.), Ancient interactions: east and west in Eurasia (pp. 75–92). Cambridge: McDonald Institute for Archaeological Research. Chapman, J. (2009). The Danube and Settlement Prehistory - 80 Years On. European Journal of Archaeology, 12, 145–156. Chapman, J. (2013). From Varna to Brittany via Csőszhalom — Was There a Varna Effect? In A. Anders & V. Kulcsár (Eds.), Moments in time. Papers Presented to Pál Raczky on His 60th Birthday (pp. 323–335). Budapest: L’Harmattan. Chapman, J. (2015). Burn or bury? Mortuary alternatives in the Neolithic and Chalcolithic of Central and Eastern Europe. In A. Diachenko, F. Menotti, S., Ryzhov, K. et al. (Eds.), The CucuteniTrypillia cultural complex and its neighbours. Essays in memory of Vlodomyr Kruts (pp. 259–278). Lviv: Astrolyabia. Chapman, J. (2017). The Standard Model, the Maximalists and the Minimalists: New Interpretations of Trypillia. Journal of World Prehistory, 30 (3), 221–37. Retrieved doi:10.1007/s10963-0179106-7. Chapman, J. (in prep.). Forging identities in Balkan prehistory: dividuals, individuals and communities, 7000 – 3000 BC. Leiden: Sidestone Press. Chapman, J., & Dolukhanov, P. M. (1993). Cultural transformations and interactions in Eastern Europe: theory and terminology. In J. Chapman & P. M. Dolukhanov (Eds.), Cultural transformations and interactions in Eastern Europe (pp. 1–36). Aldershot: Avebury. Chapman, J. & Gaydarska, B. (2003). The Provision of Salt to Tripolye Megasites In A. KorvinPiotrovskiy, V. Kruts & S. Ryzhov (Eds.), Tripolian Settlement-Giants: The International Symposium Materials (pp. 203–210). Kiev: Korvin-press. Chapman, J. & Gaydarska, B. (2007). Parts and wholes: fragmentation in prehistoric context. Oxford: Oxbow Books. Chapman, J. & Gaydarska, B. (2011). Can we reconcile individualisation with relational personhood. A case study from the Early Neolithic. Documenta Praehistorica, XXXVIII, 21–43. Chapman, J. & Gaydarska, B. (2015). Fragmentation in archaeology: sorting through the pieces. Stratum Plus, II, 85–110 (in Russian). Chapman, J. & Gaydarska, B. (2016). Low-density urbanism: the case of the Trypillia group of Ukraine. In M. Fernández-Götz & D. Krausse (Eds.), Eurasia at the Dawn of History. Urbanisation and social change (pp. 81–105). Cambridge: Cambridge University Press. Chapman, J. & Gaydarska, B. (2016a). Low-density agrarian-based cities: A principle of the Past and Present. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia-Megasites and European

534 

 References

Prehistory, 4100-3400 BCE (pp. 181–193). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Chapman, J. & Gaydarska, B. (2018). The Cucuteni - Trypillia ‘Big Other’ - reflections on the making of millennial cultural traditions. In D. Werra & M. Woźny (Eds.), Between History and Archaeology. Papers in honour of Jacek Lech (pp. 267–277). Oxford: Archaeopress Archaeology. Chapman, J. & Gaydarska, B. (2018a). Concepts of time and history on Chalcolithic tell settlements and Trypillia megasites. In S. Souvatzi, A. Baysal & E. M. Baysal (Eds.), Time and history in prehistory (pp. 147–171). London: Routledge. Chapman, J. & Gaydarska, B. (2019). The pilgrimage model for Trypillia mega-sites: the case of Nebelivka, Ukraine. In V. Stîrbu & A. Comşa (Eds.), Digging in the past of Old Europe. Studies in honor of Cristian Schuster at his 60th anniversary (pp. 73–102). Târgu Jiu - Brăila: Editura Istros a Muzeului Brăilei "Carol 1". Chapman, J., Gaydarska, B. & Hale, D. (2016). Nebelivka: Assembly Houses, Ditches and Social Structure. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 117–131). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Chapman, J., Gaydarska, B., & Nebbia, M. (2019). The origins of Trypillia mega-sites. Frontiers in Digital Humanities, section Digital Archaeology: special issue on “Where Do Cities Come From and Where Are They Going To? Modelling Past and Present Agglomerations to Understand Urban Ways of Life”. Retrieved doi: 10.3389/fdigh.2019.00010. Chapman, J., Gillings, M., Magyari, E. et al. (2010). The Upper Tisza Project: Studies in Hungarian Landscape Archaeology. Book 2: Settlement Patterns in the Bodrogköz Block. Oxford: BAR International Series 2087. Chapman, J., & Monah, D. (2007). A seasonal Cucuteni occupation at Silişte- Prohozeşti, Moldavia. In D. Monah, G. Dumitroaia, O. Weller & J. Chapman (Eds.), L’exploitation du sel à travers le temps (pp. 71–88). Piatra Neamţ: Centre International de Recherche de la culture Cucuteni. Chapman, J. & Videiko, M. (2011). The Tripillia culture megasite near Nebelivka: summer 2009 season. Praehistorica (Prague), XXIX, 79–94 Chapman, J., Videiko, M., Gaydarska, B. et al. (2014). An Answer to Roland Fletcher’s Conundrum?: Preliminary Report on the Excavation of a Trypillia Mega-structure at Nebelivka, Ukraine. Journal of Neolithic Archaeology, 16, 135–157 (URL: http://www.jungsteinsite.de/). Chapman, J., Videiko, M., Gaydarska, B. et al. (2015). Nebelivka: in search of the history and meaning of a Tripillye Copper Age megasite. Budapest proceedings. In S. Hansen, P. Raczky, A. Anders & A. Reingruber, A. (Eds.), Neolithic and Copper Age between the Carpathians and the Black Sea (pp. 419–436). Archäologie in Eurasien 31. Bonn: Habelt. Chapman, J., Videiko, M., Gaydarska, B. et al. (2014a). The Planning of the Earliest European Proto-Towns: A New Geophysical Plan of the Trypillia Megasite of Nebelivka, Kirovograd Domain, Ukraine. Antiquity Gallery, Volume 88, Issue 339, online gallery: http://antiquity.ac.uk/projgall/chapman339/ Chapman, J., Videiko, M., Hale, D. et al. (2014b). The Second Phase of the Trypillia Megasite Methodological Revolution: A New Research Agenda. European Journal of Archaeology, 17 (3), 369–406. Retrieved doi:10.1179/1461957114Y.0000000062 Chatford Clark, D. (2007). Viewing the liturgy: a space syntax study of changing visibility and accessibility in the development of the Byzantine church in Jordan. World Archaeology, 39, 84–104. Chernykh, E. N. (1992). Ancient metallurgy in the USSR: the Early Metal Age (translated by S. Wright). Cambridge: Cambridge University Press. Chernykh, E. N. (2010). Med iz Hvalynskih mogilnikov i ee paralleli. In S. Agapov (Ed.), Khvalynskie eneoliticheskie mogilniki i khvalynskaya eneoliticheskaya kultura (pp. 219–233). Povolzhe: Samara.

References 

 535

Chernysh, E. (1977). Formirovanie lokalnih variantov tripolskoi kulturi. In Vses. Konferentsiya Noveishie dostizhenie sovetskih arheologov (pp. 18–21). Moskva: AS USSR (in Russian). Chernovol, D. (2012). Houses of the Tomashovskaya Local Group. In F. Menotti & A. KorvinPiotrovskiy (Eds.), The Tripolye Culture Giant-Settlements in Ukraine: Formation, Development and Decline (pp. 182–209). Oxford: Oxbow Books. Chernovol, D. & Radomskiy, I. (2015). Kremjani vistria z trypilskogo poselennia Ozheve-Ostriv. In A. Diachenko, F. Menotti, S. Ryzhov et al. (Eds.), The Cucuteni-Trypillia cultural complex and its neighbours. Essays in memory of Vlodomyr Kruts (pp. 367–384). Lviv: Astrolyabia. Childe, G. (1929). The Danube in prehistory. Oxford: Clarendon Press. Childe, V. G. (1945). Review of Tripilska kultura by T.S. Passek. Antiquity, 19, 203–6. Childe, V. G. (1949). The sociology of knowledge. The Modern Quarterly, 4, 302–309. Childe, V. G. (1950). The Urban Revolution. The Town Planning Review, 21 (1), 3–17. Childe, V. G. (1951). Review of Periodizatsiya Tripolskikh Poselenii by T.S. Passek. Soviet Studies, 2 (4), 386–389. Childe, V. G. (1958). The prehistory of European society. London: Cassell. Choyke, A. (1997). The bone tool manufacturing continuum. Anthropozoologica, 25–26, 65–72. Choyke, A. (2001). Late Neolithic Red Deer Canines and Their Imitations. In A. Choyke and L. Bartosiewicz (Eds.), Crafting Bone: Skeletal Technologies through Time and Space. Proceedings of the 2nd meeting of the (ICAZ) Worked Bone Research Group Budapest, 31 August – 5 September 1999 (pp. 251–266). Oxford: BAR International Series 937. Choyke, A. (2010). The bone is the beast: Animal amulets and ornaments in power and magic. In D. Campana, P. Crabtree, S. DeFrance et al. (Eds.), Anthropological Approaches to Zooarchaeology: Colonisation, Complexity and Animal Transformations (pp. 197–209). Oxford: Oxbow Books. Christophersen, A. (2015). Performing towns. Steps towards an understanding of medieval urban communities as social practice. Archaeological Dialogues, 22 (2), 109–170. Clark, G. (1969). World Prehistory. A new outline. Cambridge: Cambridge University Press. Coe, M. (2003). Angkor and Khmer Civilisation. London: Thames and Hudson. Colombijn, F. (1994). Patches of Padang. The history of an Indonesian town in the twentieth century and the use of urban space. Leiden: The Netherlands Research School CNWS. Colombini, M., Giachi, G., Modugno, F., & Ribechini, E. (2005). Characterisation of organic residues in pottery vessels of the Roman age from Antinoe (Egypt). Microchemical Journal, 79, 83–90. Conolly, J. & Lake, M. (2006). Geographical information systems in archaeology. Cambridge: Cambridge University Press. Cooney, G. & Grogan, E. (1999). Irish prehistory: a social perspective. Dublin: Wordwell. Copley, M., Bland, H., Rose, P. et al. (2005). Gas chromatographic, mass spectrometric and stable carbon isotopic investigations of organic residues of plant oils and animal fats employed as illuminants in archaeological lamps from Egypt. The Analyst, 130 (6), 860–871. Correa-Ascencio, M. & Evershed, R. (2014). High throughput screening of organic residues in archaeological potsherds using direct acidified methanol extraction. Analytical Methods, 6 (5), 1330–1340. Cotiugă, V. (2009). Experimental archaeology: the burning of the Chalcolithic dwellings. In V. Cotiugă, F. Tencariu & G. Bodi (Eds.), Itinera in praehistoria. Studia in honorem magistri Nicolae Ursulescu (pp. 303–342). Iaşi: Editura Universităţii ‘Alexandru Ioan Cuza Iaşi. Courty, M., Goldberg, P., & Macphail, R. (1989). Soils and Micromorphology in Archaeology. Cambridge Manuals in Archaeology. Cambridge: Cambridge University Press Cowgill, G. (2004). Origins and Development of Urbanism: Archaeological Perspectives. Annual Review of Anthropology, 33, 525–549. Craig, O., Saul, H., Lucquin, A. et al. (2013). Earliest evidence for the use of pottery. Nature, 496, 351–354.

536 

 References

Creamer, W. & Haas, J. (1985). Tribe versus Chiefdom in Lower Central America. American Antiquity, 50 (4), 738–54. Crîşmaru, A. (1977). Drăguşeni. Contribuţii la o monografie arheologică. Botoşani: Muzeul Judeţean de Istorie. Curta, F. (2014). Ethnic identity and archaeology. In C. Smith (Ed.), Encyclopaedia of global archaeology (pp. 2507–2514). New York: Springer. Dal Corso, M., Hamer, W., Hofmann, R. et al. (2019). Modelling landscape transformation at the Chalcolithic Tripolye mega-site of Majdanetske (Ukraine): wood demand and availability. The Holocene. Retrieved DOI: 10.1177/0959683619857229. Dănilă, C. (2014). Pictura cu grafit. O abordare experimentala. Buletinul Muzeului Judeţean Teleorman. Seria Arheologie, 6, 151–199. de Arruda Campos, M. B. & Fong, P. S. P. (2003). Bending the axial line: Smoothly continuous road centre-line segments as a basis for road network analysis. In Proceedings of the 4th International Symposium on Space Syntax (pp. 35.1–35.10). UCL, London, UK. Deetz, J. (1996). In small things forgotten: An archaeology of early American life. New York: Anchor Books. Degerbøl, M. & Fredskild, B. (1970). The Urus (Bos primigenius Bojanus) and Neolithic domesticated cattle (Bos taurus domesticus Linne.) in Denmark. Det Kongelige Dansk Videnskabernes Selskab, Biologiski Skrifter, 17, 1–177. Delle, J. A. (1998). An archaeology of social space: analyzing coffee plantations in Jamaica’s Blue Mountains. New York: Springer Science & Business Media. DeNiro, M. (1985). Postmortem preservation and alteration of in vivo bone collagen isotope ratios in relation to palaeodietary reconstruction. Nature, 317, 806–809. Dergačev, V. (1991). Bestattungskomplexe der späten Tripolje-Kultur. Materialien zur Allgemeinen und Verlgleichenden Archäologie, Band 45. Mainz am Rhein: Verlag Philipp von Zabern. Dergachev, V. (1998). Karbunskij klad. Chisinau: Institut de Arheologii. Dergachev, V. (2002). Two studies in defence of the migration concept. In K. Boyle, C. Renfrew & M. Levine (Eds.), Ancient interactions: east and west in Eurasia (pp. 93–112). Cambridge: McDonald Institute for Archaeological Research. Derrida, J. (1970). Structure, sign and play. In R. Macksey and E. Donato (Eds.), The Languages of Criticism and the Sciences of Man. Baltimore: Johns Hopkins Press. de Vries, J. (1984). European Urbanization 1500–1800. London: Routledge. Diachenko, A. (2010). Tripolskoe naselenie Bugo-Dneprovskogo mezhdurechya: prostranstvennovremennoj analiz. Unpub. PhD thesis, Kiev: Institute of Archaeology. Diachenko, A. (2016). Demography Reloaded. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 181–193). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Diachenko, A. (2016a). Small is beautiful: a democratic perspective. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 269–280). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Diachenko, A. & Harat-Strotsen K. (2016). Buildings in the Settlement of the Trypillia Culture in the Light of Daub Analysis. In M. Szmyt (Ed.), Biały Potok. Materials from Józef Kostrzewski’s Podolia Excavations (pp. 69–112). Poznań: Muzeum Archeologiczne w Poznaniu. Diachenko, A. & Menotti, F. (2012). The Gravity Model: Monitoring the Formation and Development of the Tripolye Culture Giant-Settlements in Ukraine. Journal of Archaeological Science, 39 (8), 2810–17. Retrieved doi:10.1016/j.jas.2012.04.025. Diachenko, A. & Menotti, F. (2015). Cucuteni-Tripolye contact networks: cultural transmission and chronology. In A. Diachenko, F. Menotti, S. Ryzhov et al. (Eds.), The Cucuteni-Trypillia cultural complex and its neighbours. Essays in memory of Vlodomyr Kruts (pp. 131–152). Lviv: Astrolyabia.

References 

 537

Diachenko, A. & Menotti, F. (2017). Proto-Cities or Non-Proto-Cities? On the Nature of Cucuteni– Trypillia Megasites. Journal of World Prehistory, 30, 207–219. Retrieved DOI 10.1007/ s10963-017-9105-8 Dobres, M-A. & Robb, J. (Eds.) (2000). Agency in archaeology. London: Routledge. Domboróczki, L., Kaczanowska, M. & Kozłowski, J. (2010). The Neolithic Settlement of TiszaszőlősDomaháza-Puszta and the Question of the Northern Spread of the Körös Culture. Atti della Societa Preistoria a Protostoria Regione Friuli-Venezia Guilia, XVII, 101–155. Doonan, R. & Hommel, P. (n.d.). Between ideas and objects: the doings in invention in pottery and metallurgy. Paper presented to Session 93, Invention as a process: pyrotechnologies in pre-literate societies, SAA 78th Annual Meeting, Honolulu, 3rd - 7th April 2013. Draşovean, F., Schier, W., Bayliss, A. et al. (2017). The Lives of Houses: Duration, Context, and History at Neolithic Uivar, Romania. European Journal of Archaeology, 20 (4), 636–662. Retrieved doi:10.1017/eaa.2017.37 Drennan, R. & Peterson, C. (2008). Centralized Communities, Population, and Social Complexity after Sedentarization. In J-P. Bocquet-Appel & O. Bar-Yosef (Eds.), The Neolithic Demographic Transition and Its Consequences (pp.359–386). New York: Springer. Dudkin, V. (1978). Geofizicheskaya Razvedka Krupnih Tripol’skih Poselenii. In V. Genning (Ed.), Ispol’zovanie Metodov Estestvennih Nauk v Arheologii (pp. 25–45). Kiev: Naukova Dumka. Dudkin, V. & Videiko, M. (2004). Planuvanna poselen trypil’skoi kul’tury. In M. Videiko (Ed.), Entsiklopedia Tripil’skoi tsyvilizatsii, Vol. 1 (pp. 303–314). Kyiv: Ukrpoligrafmedia. Duffy, P. (2015). Site Size Hierarchy in Middle-Range Societies. Journal of Anthropological Archaeology, 37, 85–99. Retrieved doi:10.1016/j.jaa.2014.12.001. Dumitrescu, V. (1925). Fouilles de Gumelniţa. Dacia, II, 29–103. Dumitrescu, V. & Marinescu-Bîlcu, S. (2001). Noi descoperiri de la Gumelnița. Cultură şi Civilizaţie la Dunărea de Jos, XVI–XVII, 114–144. Earle, T. (1987). Chiefdoms in Archaeological and Ethnohistorical Perspectives. Annual Review of Anthropology, 16, 279–308. Eckmeier, E., Gerlach, R., Gehrt, E., & Schmidt, M. W. (2007). Pedogenesis of Chernozems in Central Europe — A review. Geoderma, 139 (3-4), 288–299. Ehwald, E., Jäger, K. D., & Lange, E. (1999). Das Problem Wald - Offenland im zirkumherzynen Trockengebiet vor der neolithischen Besiedlung sowie die Entstehung der zirkumherzynen Schwarzerden. In R. Rolle & F. M. Andraschko (Eds.), Frühe Nutzung pflanzlicher Ressourcen (pp. 12–34). Internationales Symposium, Hamburger Werkstattreihe zur Archäologie. Hamburg: Duderstedt. Ellis, L. (1984). The Cucuteni-Tripolye Culture. A Study in Technology and the Origins of Complex Society. International Series 217. Oxford: BAR. Evans, D., Pottier, C., Fletcher, R. et al. (2007). A new archaeological map of the world’s largest pre-industrial settlement complex at Angkor, Cambodia. Proceedings of the National Academy of Sciences, 104, 14277–82. Fairclough, G. (1992). Meaningful constructions–spatial and functional analysis of medieval buildings. Antiquity, 66, 348–366. Fedorov, S. (2015). Proces restauraciji hrestopodibnoi vimostki z poseleniya Nebelivka. In M. Videiko, J. Chapman, I. Kozir & V. Sobchuk (Eds.), At the Eastern frontiers of Old Europe (pp. 45–47). Conference programme and abstracts, Kirovograd International Conference, 12th - 14th May 2015. Kirovograd: Kirovograd County Council. Feinman, G. M. & Nicholas, L. M. (2016). Framing the rise and variability of past complex societies. In L. F. Fargher & V. Y. Heredia Espinoza (Eds.), Alternative pathways to complexity (pp. 271–293). Boulder, CO: University Press of Colorado. Ferguson, L. (2012). Uncommon ground: archaeology and early African America, 1650-1800. Washington, D. C.: Smithsonian Institution.

538 

 References

Ferguson, T. (1996). Historic Zuni architecture and society: An archaeological application of space syntax. Phoenix AZ: University of Arizona Press. Fisher, K. (2009). Placing social interaction: An integrative approach to analyzing past built environments. Journal of Anthropological Archaeology, 28, 439–457. Flannery, K. (1976). Evolution of Complex Settlement Systems. In K. Flannery (Ed.), The Early Mesoamerican Village (pp. 162–173). New York: Academic Press. Flannery, K. (Ed.) (1976a). The Early Mesoamerican Village. New York: Academic Press. Fleming, A. (1988). The Dartmoor Reaves. London: Batsford. Fleming, A. (2008). The Dartmoor Reaves. 2nd. ed. London: Batsford. Fletcher, R. (1995). The limits to settlement growth. Cambridge: Cambridge University Press. Fletcher, R. (1998). African Urbanism: Scale, Mobility and Transformations, In G. Connah (Ed.), Transformations in Africa (pp. 104–138). London and Washington: Leicester University Press. Fletcher, R. (2009). Low-Density, Agrarian-Based Urbanism: A Comparative View. Insights (Durham University, Institute of Advanced Study), 2 (4), 2–19. Fletcher, R. (2012). Low-density, agrarian-based urbanism: scale, power, and ecology. In M. Smith (Ed.), The Comparative Archaeology of Complex Societies (pp. 285–320). New York: Cambridge University Press. Fletcher, R. (n.d.) Overview. Paper presented at Amerind Foundation Workshop, ‘“Cities or Big villages?”. New approaches to great, anomalous places’, Dragoon, AZ, May 2014. Fletcher, R. & Evans, D. (2012). The dynamics of Angkor and its landscape. In A. Haendel (Ed.), Old Myths and New Approaches: Interpreting Ancient Religious Sites in Southeast Asia (pp. 42–62). Clayton, Victoria: Monash University Publishing. Fletcher, R. & Kim, N. (Eds.) (submitted) Anomalous giants – category problems and trajectory analysis. Tucson, AZ: University of Arizona Press. Fletcher, R. & Pottier, C. (2002). The gossamer city - a new enquiry. Museum International, 54, 23–7. Forge, A. (1972). Normative factors in the settlement size of Neolithic cultivators (New Guinea). In P.J. Ucko, G. W. Dimbleby & R. Tringham (Eds.), Man, settlement and urbanism (pp. 363–376). London: Duckworth. French, C. & Pryor, F. (2005). Archaeology and environment of the Etton landscape. East Anglian Archaeology Report 109. Norwich: Fenland Archaeological Trust. Friedman, J. & Rowlands, M. (1977). Notes towards an Epigenetic Model of the Evolution of ‘Civilization’. In J. Friedman & M. Rowlands (Eds.), The Evolution of Social Systems (pp. 201–278). London: Duckworth. Gaffney, C., Gaffney, V., & Tingle, M. (1985). Settlement, Economy or Behaviour? Micro-Regional Land Use Models and the Interpretation of Surface Artefact Patterns. In C. Haselgrove, M. Millett & I. Smith (Eds.), Archaeology from the Ploughsoil: Studies in the Collection and Interpretation of Field Survey Data (pp. 95–107). Sheffield: Department of Archaeology and Prehistory, University of Sheffield. Galaty, M. (2005). European Regional Studies: A Coming of Age? Journal of Archaeological Research, 13 (4), 291–336. Retrieved doi:10.1007/s10814-005-5110-z. Gallant, T. (1986). ‘Background Noise’ and Site Definition: A Contribution to Survey Methodology. Journal of Field Archaeology, 13 (4), 403–18. Retrieved doi:10.2307/530167. Gamble, C. (2007). Origins and revolutions: human identity in earliest prehistory. Cambridge: Cambridge University Press. García Sanjuán, L., Scarre, C., & Wheatley, D. W. (2017). The megasite of Valencina de la Concepción (Seville, Spain): Debating settlement form, monumentality and aggregation in Southern Iberian Copper Age societies. Journal of World Prehistory, 30 (3), 239–257. Garrow, D., Beadsmoore, E., & Knight, M. (2005). Pit clusters and the temporality of occupation: an earlier Neolithic pit site at Kilverstone. Thetford, Norfolk. Proceedings of the Prehistoric Society, 71, 139–158.

References 

 539

Gaskevych, D. (2014). Problemy radiovuhletsevoho datuvannia buho-dnistrovskoi neolitychnoi kultury. Arkheolohiya (Kyiv), 4, 3–17. Gaskevych, D. (2017). Pottery with Admixture of Graphite and Exchange Networks in Neolithic of Ukraine: Problem Statement. In Y. Morozova & P. Shydlovskyi (Eds.), Wetland Archaeology and Prehistoric Networks in Europe (pp. 41–42). NEENAWA International Scientific Conference, September 15th-18th, 2017. Kyiv–Kaniv. Gaskevych, D. (2017a). Keramika s rospis’yu v neolite Yuzhnogo Buga. Stratum plus, 2, 231–255. Gates, C. (2011). Ancient cities: the archaeology of urban life in the Ancient Near East and Egypt, Greece and Rome. London: Routledge. Gaydarska, B. (2003). Application of GIS in Settlement Archaeology: An Integrated Approach to Prehistoric Subsistence Strategies. In A. Korvin-Piotrovskiy, V. Kruts & S. Ryzhov (Eds.), Tripolian Settlement-Giants: The International Symposium Materials (pp. 212–216). Kiev: Korvin-press. Gaydarska, B. (2007). Landscape, Material Culture and Society in South East Bulgaria. BAR International Series 1618. Oxford: Archaeopress. Gaydarska, B. (2015). Poseleniya-giganty ili proto-goroda – oshibochnoe protivopostavlenie (Giantsettlements or proto-cities: the false opposition). In A. Diachenko, F. Menotti, S., Ryzhov et al. (Eds.), The Cucuteni-Trypillia cultural complex and its neighbours. Essays in memory of Vlodomyr Kruts (pp. 35–56). Lviv: Astrolyabia . Gaydarska, B. (2016). The city is dead – long live the city! Norwegian Archaeological Review, 49, 40–57. Gaydarska, B. (2017). Introduction: European prehistory and urban studies. Special Issue on Urbanism. Journal of World Prehistory, 30 (3), 177–188. Gaydarska, B. (2019). ‘If we want things to stay as they are, things will have to change’: the case of Trypillia. In Kadrow, S. & Müller, J. (Eds.), Habitus? The social dimensions of technology and transformation (pp. 47–69). Leiden: Sidestone Press. Gaydarska, B. (2019a). Trypillia Megasites—the First Cities in Europe? In A. Gyucha (Ed.), Coming together: Comparative approaches to population aggregation and early urbanization (pp. 165–187). Buffalo: SUNY Press. Gaydarska, B. (submitted). Fragmenting Trypillian megasites: a bottom-up approach. (To appear in M. Fernández-Götz and T. Chevral (Eds.), Power from below. Cambridge: Cambridge University Press. Gaydarska, B. & Chapman, J. (2016). Nine Questions for Trypillia megasite research. In B. D. Borisov (Ed.), Studia in honorem Professoris Boris Borisov (pp. 179–197). Veliko Trnovo: IBIS. Gaydarska, B., Nebbia, M. & Chapman, J. (2020). Trypillia megasites in context: independent urban development in Chalcolithic Eastern Europe. Cambridge Archaeological Journal, 30 (1). Retrieved doi: 10.1017/S0959774319000301. Georgieva, P. (2012). The Pottery of the Krivodol-Salcuta Culture. Sofia: University Press St. Kliment Ohridski. Gieryn, T. (2000). A space for place in sociology. Annual Review of Sociology, 26, 463–496. Gillings, M. (2015). Mapping invisibility: GIS approaches to the analysis of hiding and seclusion. Journal of Archaeological Science, 62, 1–14. Retrieved doi.org/10.1016/j.jas.2015.06.015 Gillings, M. (2017). Mapping liminality: Critical frameworks for the GIS-based modelling of visibility. Journal of Archaeological Science, 84, 121–128. Gilman, A. (1981). The Development of Social Stratification in Bronze Age Europe. Current Anthropology, 22 (1), 1–23. Gimbutas, M. (1977). The first wave of Eurasian steppe pastoralists into Copper Age Europe. Journal of Indo-European Studies, 5, 277–338. Gimbutas, M. (1982). Goddesses and Gods of Old Europe, 6500-3500 BC: Myths and Cult Images. London: Thames and Hudson.

540 

 References

Gini, C. (1912). Variabilità e Mutabilità. In E. Pizzetti and T. Salvemini (Eds.), Memorie Di Metodologia Statistica. Rome: Libreria Eredi Virgiliio Veschi. Goodchild, M. & Janelle, D. (2010). Toward critical spatial thinking in the social sciences and humanities. Geojournal, 75, 3–13. Gradmann, R. (1933). Die Steppenheide-Theorie. Geographische Zeitschrift, 39, 265–78. Greeves, T. (1975). The use of copper in the Tripolye-Cucuteni culture of south-east Europe. Proceedings of the Prehistoric Society, 41 (2), 153–166. Guldager Bilde, P., Attema, P., & Winther-Jacobsen, K. (Eds.) (2012). The Dzarylgaz Survey Project. Aarhus: Aarhus University Press. Gusev, O. (1995). Tripilska kultura Srednogo Pobužžja. Vinnitsa: Anteke-Ultd. Hahn, H-P. (2016). Cities between de-territorialization and networking: on the dynamics of urbanization in the global context. In M. Fernández-Götz & D. Krausse (Eds.), Eurasia at the Dawn of History. Urbanisation and social change (pp. 169–82). Cambridge: Cambridge University Press. Hale, D., Chapman, J., Swann, N., et al. (2010). Early Urbanism in Europe? Geophysical Survey at Nebelivka, Ukraine. NSGG conference abstracts. Recent Work in Archaeological Geophysics (pp. 35–36). London: Geological Society. Hale, D. N., Chapman, J., Videiko, M., et al. (2017). Nebelivka, Ukraine: geophysical survey of a complete Trypillia megasite. In C. Jennings, V. Gaffney, T. Sparrow & S. Gaffney (Eds.), AP 2017: 12th International Conference of Archaeological Prospection (12th-16th September 2017, University of Bradford) (pp. 100–102). Oxford: Archaeopress. Hall, E. (1966). The hidden dimension. Garden City. NY: Doubleday. Halstead, P. (1982). The economy has a normal surplus: economic stability and social change among early farming communities of Thessaly, Greece. In P. Halstead & J. O’Shea (Eds.), Bad year economics: cultural responses to risk and uncertainty (pp. 68–80). Cambridge: Cambridge University Press. Halstead, P. (2011). Farming, material culture and ideology: repackaging the Neolithic of Greece (and Europe). In A. Hadjikoumis, E. Robinson & S. Viner (Eds.), Dynamics of Neolithisation in Europe. Studies in honour of Andrew Sherratt (pp. 131–151). Oxford: Oxbow Books. Hansen, S., Toderaș, M., Reingruber, A. et al. (2007). Pietrele, Măgura Gorgana. Ergebnisse der Ausgrabungen im Sommer 2006. Eurasia Antiqua, 13, 43–112. Hansen, S., Toderaș, M., Reingruber, A. et al. (2012). Pietrele an der Unteren Donau. Bericht über die Ausgrabung im Sommer 2011. Eurasia Antiqua, 18, 1–68. Harding, A., Sievers, S. & Venclová, N. (Eds.) (2006). Enclosing the Past: Inside and Outside in Prehistory. Sheffield Archaeological Monographs 15. Sheffield: J R Collis Publications. Harper, T. (2011). Visualizing Agricultural Production during the Eneolithic: A Case Study from the Tripolye Giant-Settlement of Talyanki. Chronika 1, 27–32. Harper, T. (2016). Climate, migration and cities on the Old European periphery: a spatialdemographic approach to understanding the Tripolye giant-settlements. PhD, SUNY-Buffalo. Ann Arbor, MI: Proquest. Harper, T. K. (2019). Demography and climate in Late Eneolithic Ukraine, Moldova and Romania: multiproxy evidence and pollen-based regional corroboration. Journal of Archaeological Science: Reports, 23, 973–982. Haselgrove, C., Millett, M., & Smith, I. (Eds.) (1985). Archaeology from the Ploughsoil: Studies in the Collection and Interpretation of Field Survey Data. Sheffield: Department of Archaeology and Prehistory, University of Sheffield. Hastorf, C. (1991). Gender, space and food in prehistory. In J. Gero & M. Conkey (Eds.), Engendering Archaeology: Women and Prehistory (pp. 132–159). Oxford: Blackwell. Hayden, B. & Cannon, A. (1983). Where the garbage goes? Journal of Anthropological Archaeology, 2, 117–163.

References 

 541

Helms, M. (1988). Ulysses’ sail: an ethnographic odyssey of power, knowledge, and geographical distance. Princeton, N.J: Princeton University Press. Higham, T., Slavchev, V., Gaydarska, B., & Chapman, J. (2018). AMS dating of the Late Copper Age Varna cemetery, Bulgaria. Radiocarbon, 60 (2), 493–516. Hillier, B. (2005). The art of place and the science of space. World Architecture, 185, 96–102. Hillier, B. (2014). Spatial analysis and cultural information: the need for theory as well as method in space syntax analysis. In E. Paliou, U. Lieberwirther & S. Polla (Eds.), Spatial Analysis and Social Spaces: Interdisciplinary Approaches to the Interpretation of Prehistoric and Historic Built Environments (pp.19–48). Berlin: De Gruyter. Hillier, B. & Hanson, J. (1984). The social logic of space. Cambridge: Cambridge University Press. Hillier, B, Penn, A, Hanson, J, Grajewski, T, & Xu, J. (1993). Natural movement: or con- figuration and attraction in urban pedestrian movement. Environment and Planning B: Planning and Design, 20, 29–66. Hirth, K. (2008). Incidental Urbanism: The Structure of the Prehispanic City in Central Mexico. In J. Marcus & J. Sabloff (Eds.), The Ancient City: New Perspectives on Ancient Urbanism (pp. 273–298). Santa Fe: School for Advanced Research Press. Hodder, I. (1990). The domestication of Europe. Oxford: Blackwell. Hodder, I. (2012). Entangled. An Archaeology of the Relationships between Humans and Things. Chichester: Wiley & Sons. Hofmann, D. & Smyth, J. (Eds.) (2013). Tracking the Neolithic house in Europe. London: Springer. Hofmann, R., Müller, J., Shatilo, L. et al. (2019). Governing Tripolye: integrative architecture in Tripolye settlements. Plos One, September 2019, pp. 1–54. DOI: 10.1371/journal.pone.0222243 Hofmann, R., Ţerna, S., Ursu, C.-E. et al. (2016). Spatial organization and population size of small Cucuteni-Tripolye settlements: Results of geomagnetic surveys in Baia and Adâncata, Suceava County, Bucovina, Eastern Romania. Journal of Neolithic Archaeology, 18 (1), 157–189. Hritz, C. (2014). Contributions of GIS and Satellite-Based Remote Sensing to Landscape Archaeology in the Middle East. Journal of Archaeological Research, 22 (3), 229–76. Retrieved doi:10.1007/ s10814-013-9072-2. Hudson, K. M. & Milisauskas, S. (2017). Cognitive landscapes and the archaeology of the extended mind. Reti, saperi, linguaggi, 2, 213–232. Ingold, T. (1993). The temporality of the landscape. World Archaeology, 25, 152–174. Ingold, T. (2009). Against space: Place, movement, knowledge. In P.W. Kirby (Ed.), Boundless Worlds: An Anthropological Approach to Movement (pp. 29–43). Oxford: Berghahn Books. Ivanova, S. (2014). Istoriya naselenia severo-zapadnogo Prichernomor’ya v kontse IV/III. tys. do n.e. Kyiv: Institute of Archaeology. Ivanova, S. (2015). Kurgan na tripilskomu poselenni Nebelivka. In M. Videiko, J. Chapman, I. A. Kozir & V. V. Sobchuk (Eds.), At the Eastern frontiers of Old Europe. Conference programme and abstracts, Kirovograd International Conference, 12th - 14th May 2015 (pp. 47–48). Kirovograd: Kirovograd County Council. Jakab, A. (2011). Téglaégető kemencék a középkori Magyarország területén. Nyíregyházi Jósa András Muzeum Évkönyve, LIII, 131–160. Jarman, M., Bailey, G., & Jarman, H. (Eds.) (1982). Early European Agriculture: Its Foundations and Development. Cambridge: Cambridge University Press. Jennings, J. & Earle, T. (2016). Urbanization, state formation, and cooperation: A reappraisal. Current Anthropology, 57 (4), 474–492. Johnson, G. (1977). Aspects of Regional Analysis in Archaeology. Annual Review of Anthropology, 6, 479–508. Johnson, G. (1982). Organizational Structure and Scalar Stress. In C. Renfrew, M. Rowlands and B. Segraves (Eds.), Theory and Explanation in Archaeology (pp. 389–421). New York: Academic Press.

542 

 References

Johnston, A. (2013). Stanford Encyclopaedia of Philosophy. https://plato.stanford.edu/entries/ lacan/ Johnston, S. (n.d.). Experimental recreation of house-burning in the Tripolye-Cucuteni culture. Unpub. B.Sc. dissertation, Durham University (2015). Johnston, S., Diachenko, A., Gaydarska, B. et al. (2018). The experimental building, burning and excavation of a two-storey Trypillia house. In S. Ţurcanu & C.-E. Ursu (Eds.), Materiality and identity in pre- and proto-historic Europe. Homage to Cornelia-Madga Lazarovici (pp. 397–434). Suceava: Karl Romstorfer. Johnston, S., Litkevych, V., Gaydarska, B. et al. (2019). The Nebelivka experimental houseconstruction and house-burning. Be-Ja, 9 (1), 61–90. Jones, A. (2012). Prehistoric materialities. Oxford: Oxford University Press. Kahn, J. (n.d). Houses of power: community houses and specialized houses as markers of social complexity in the pre-Contact Society Island chiefdoms. Paper presented in Symposium 278, Households and social evolution: comprehensive approaches to social transformations, SAA 2017, Vancouver, BC, 29/III - 2/IV/2017. Kantner, J. (2008). The Archaeology of Regions: From Discrete Analytical Toolkit to Ubiquitous Spatial Perspective. Journal of Archaeological Research, 16 (1), 37–81. Retrieved doi:10.1007/ s10814-007-9017-8. Keightley, D. (1987). Archaeology and mentality: the making of China. Representations, 18, 91–128. Keswani, P. S. (1996). Hierarchies, Heterarchies, and Urbanization Processes: The View from Bronze Age Cyprus. Journal of Mediterranean Archaeology, 9 (2), 211–50. Khvoika, V. (1901). Kamennii vek Srednogo Pridneproviya. In Proceedings of the 11th Archaeological Convention in Kiev in 1899. Vol. 1. (pp. 737–812). Moskow: Moskow Archaeological Society Press. Khvoika, V. (1904). Raskopki 1901 g. v oblasti tripo’lskoi kul’turi. Russian Archaeological Society Proceedings, 5, 1–20. Kienlin, T. (2012). Beyond Elites: An Introduction. In T. Kienlin & A. Zimmermann (Eds.), Beyond Elites. Alternatives to Hierarchical Systems in Modelling Social Formations. International Conference at the Ruhr-Universität Bochum, Germany, October 22-24, 2009. Universitätsforschungen zur prähistorischen Archäologie 215 (pp. 15–32). Bonn: Habelt. King, A. (2000). Thinking with Bourdieu against Bourdieu: a ‘practical’ critique of the habitus. Sociological Theory, 18 (3), 417–433. Kiosak, D. (2016). K interpretatsii kremneobrabotki Precucuteni - Trypillia A. Tyragetia, X (XXV)(1), 147–63. Kiosak, D. (2017). Kamyane-Zavallia, the easternmost Linear Pottery culture settlement ever excavated. Sprawozdania Archeologiczne, 69, 253–70. Kirch, P. V. (1989). The evolution of the Polynesian chiefdoms. Cambridge: Cambridge University Press. Kirleis, W. & Dal Corso, M. (2016). Trypillian subsistence economy: animal and plant exploitation. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE. Themes in Contemporary Archaeology Volume 2 (pp.195–206). Abingdon: Routledge. Kirleis, W. & Dreibrodt, S. (2016). The natural background: forest, forest steppe or steppe environment. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE. Themes in Contemporary Archaeology Volume 2 (pp.171–179). Abingdon: Routledge. Klejn, L. (1982). Archaeological typology. International Series 153. Oxford: BAR. Knappett, C. (2011). An Archaeology of Interaction: Network Perspective on Material Culture and Society. Oxford: Oxford University Press.

References 

 543

Kohler, T., Smith, M., Bogaard, E. et al. (2017). Greater Post-Neolithic Wealth Disparities in Eurasia than in North America and Mesoamerica. Nature, 551, 619–22. Retrieved doi:10.1038/ nature24646. Kohring, S. (2012). A scalar perspective to social complexity: complex relations and complex questions. In T. Kienlin & A. Zimmermann (Eds.), Beyond elites. Alternatives to hierarchical systems in modelling social formations (pp. 327–338). Bonn: Rudolf Habelt. Kokkinidou, D. & Nikolaidou, M. (2010). Dimitra Kokkinidou & Marianna Nikolaidou. In D. Bailey, A. Cochrane & J. Zambelli (Eds.), Unearthed. Norwich: Sainsbury Centre for Visual Arts, 76. Kolesnikov, A. G. (1993). Tripolskoe obshtestvo srednego Podneprovya. Opit sotsialnih rekonstruktsiya v arheologii. Kyiv: Naukova Dumka. Kopytoff, I. (Ed.) (1987). The African frontier: the reproduction of traditional African societies. Bloomington, IND: Indiana University Press. Korobkova, G. (1987). Khoziajztvennye kompleksy rannih zemledelchesko-skotovodcheskih obschestv juga SSSR. Leningrad: Nauka. Korvin-Piotrovskiy, A. G. (2003). Teoreticheskie problemy issledovanij poselinij-gigantov. In A. Korvin-Piotrovskiy, V. Kruts & S. Ryzhov (Eds.), Tripolian Settlement-Giants: The International Symposium Materials (pp. 5–7). Kiev: Korvin-press. Korvin-Piotrovskiy, A. G. (2015). Megastruktura na poseleniyah gigantah Tripolskoi kulturi: variant interpretaciji. In M. Videiko, J. Chapman, I. A. Kozir & V. V. Sobchuk (Eds.), At the Eastern frontiers of Old Europe. Conference programme and abstracts, Kirovograd International Conference, 12th - 14th May 2015 (pp. 19–20). Kirovograd: Kirovograd County Council. Korvin-Piotrovskiy, A.G., Chabaniuk, V., & Shatilo, L. (2012). Tripolian house construction: conceptions and experiments. In F. Menotti & A. Korvin-Piotrovskiy (Eds.), The Tripolye Culture Giant-Settlements in Ukraine: Formation, Development and Decline (pp. 210–29). Oxford: Oxbow Books Korvin-Piotrovskiy, A., Hofmann, R., Rassmann, K. et al. (2016). Pottery kilns in Trypillian Settlements. Tracing the division of labour and the social organization of Copper Age communities. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 221–252). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Korvin-Piotrovskiy, A. & Shatilo, L. (2008). Experimental modelling on Tripolian houses of the Tomashovska local group. In A. Korvin-Piotrovskiy & F. Menotti (Eds.), Tripolye Culture in Ukraine: The Giant-Settlement of Talianki (pp. 199–210). Kiev: Institute of Archaeology of the NAS of Ukraine. Koshelev, I. N. (2004). Pamyatniki tripolskoi kulturi po dannim magnitnoi razvedki. Kiev: Samizdat. Kotova, N. S. (2003). Neolithization in Ukraine. International Series 1109. Oxford: BAR. Kowalewski, S. (2008). Regional Settlement Pattern Studies. Journal of Archaeological Research, 16 (3), 225–85. Retrieved doi:10.1007/s10814-008-9020-8. Kremenetski, C. (1995). Holocene vegetation and climate history of southwestern Ukraine. Review of Palaeobotany and Palynology, 85 (3–4), 289–301. Kremenetski, C. (2003). Steppe and forest-steppe belt of Eurasia: Holocene environmental history. In C. Renfrew & M. Levine (Eds.), Prehistoric Steppe Adaptation and the Horse (pp. 11–27). Cambridge: McDonald Archaeological Monographs. Kruk, J. (1980). The Neolithic settlement of southern Poland. International Series 93. Oxford: BAR. Krukar, J. & Dalton, R. (2013). Spatial Predictors of Eye Movement in a Gallery Setting. In 1st International Workshop on Eye Tracking for Spatial Research, 2nd September 2013, Scarborough. Kruts, V. (1989). K istorii naseleniya tripolskoj kultury v mezhdurechye Yuzhnogo Buga i Dnepra. In S. Berezanskaya (Ed.), Pervobytnaya Archeologiya: Materialy i Issledovaniya (pp. 117–132). Kyiv: Naukova Dumka.

544 

 References

Kruts, V. (1990). Planirovka Poseleniya u s. Taljanki I Nekotorye Voprosy Tripolskogo Domostroitelstva. In V. Zbenovich (Ed.), Rannezemledelcheskie Poseleniya-Giganty Tripolskoj Kultury Na Ukraine. Tezisy Dokladov Pervogo Polevogo Seminara (pp. 43–47). Taljanki; Institute of Archaeology of the AS of the USSR. Kruts, V. (2002). Zhivotnovodstvo v ekonomike tripolskoj kultury. Stratum Plus, 2, 179–186. Kruts, V. (2003). Tripolskie ploshchadki - rezul’tat ritual’nogo sozhzheniya domov. In A. KorvinPiotrovskiy, V. Kruts & S. Ryzhov (Eds.), Tripolian Settlement-Giants: The International Symposium Materials (pp. 74–76). Kiev: Korvin-press. Kruts, V. (2012). Giant-Settlements of Trypolye Culture. In F. Menotti & A. Korvin-Piotrovskiy (Eds.), The Tripolye Culture Giant-Settlements in Ukraine: Formation, Development and Decline (pp. 70–78). Oxford: Oxbow Books. Kruts, V., Korvin-Piotrovskiy, A., Chabanyuk, V., & Shatilo, L. (2013). Tripolskoe poselenie-gigant Taljanki Issledovaniya 2012 g. Kyiv: Institute of Archaeology. Kruts, V., Korvin-Piotrovskiy, A., Menotti, F. et al. (2010). Ploshchadka No.44. In A. Korvin-Piotrovskiy & F. Menotti (Eds.), Tripolskoe poselenie-gigant Taljanki Issledovaniya 2010 g. (pp. 6–19). Kyiv: Institute of Archaeology. Kruts, V., Korvin-Piotrovskiy, A., & Ryzhov, S. (2001). Tripolskoe poselenie-gigant Taljanki Issledovaniya 2001 g. Kyiv: Institute of Archaeology. Kruts, V., Korvin-Piotrovskiy, A., Ryzhov, S. et al. (2005). Issledovanie poselenij-gigantov tripolskoj kultury v 2002 – 2004 gg. Kiev: Korvin-Press. Kruts, V., Korvin-Piotrovskiy, A., Ryzhov, S., & Chernovol, D. (2008). The 2005-2006 research at Talianki. In Tripolje Culture in Ukraine: The Giant-Settlement of Talianki (pp. 71–108). Kiev: Institute of Archaeology. Kruts, V., Korvin-Piotrovskiy, A., Ryzhov, S. et al. (2011). Tripolskoe poselenie-gigant Taljanki Issledovaniya 2011 g. Kyiv: Institute of Archaeology. Kruts, V. & Ryzhov, S. (1985). Fazy Rozvytku Pamjatok Tomashivsko-Sushkivskoi Grupy. Archeologiya, 52, 45–56. Kubiena, W. (1953). The soils of Europe. London: Thomas Murby. Kuhn, T. (1970). The structure of scientific revolutions. Chicago: University of Chicago Press. Kujit, I. (n.d.). Micro-history and macro-evolution: material geographies of multifamily Neolithic households. Paper presented in Symposium 278, Households and social evolution: comprehensive approaches to social transformations, SAA 2017, Vancouver, BC, 29/III - 2/IV/2017. Kulska, O. & Dubitska, N. (1940). Budivelni materialy trypilskoi kultury (khimiko-tehnologichnyj analiz). In M. Yachmeniov, T. Passek and L. Slavin (Eds.), Trypilska kultura. Tom 1 (pp. 325–34). Kyiv: Vydavnytstvo AN URSR. Kuna, M. (2015). Categories of settlement discard. In K. Kristiansen, L. Šmejda and J. Turek (Eds.), Paradigm found. Archaeological theory - present, past and future (pp. 278–292). Oxford: Oxbow Books. Kusimba, C., Kusimba, S., & Agbaje-Williams, B. (2006). Precolonial African Cities: Size and Density. In G. Storey (Ed.), Urbanism in the Preindustrial World: Cross-cultural Approaches (pp. 145–58). Tuscaloosa: University of Alabama Press. Kvamme, K. (1990). One-Sample Tests in Regional Archaeological Analysis: New Possibilities through Computer Technology. American Antiquity, 55 (2), 367–81. Retrieved doi:10.2307/281655. Lacan, J. (1988). Book II: J-A. Miller (Ed.), The Ego in Freud’s Theory and in the Technique of Psychoanalysis, 1954–1955. New York: W.W. Norton and Company. Lawrence, D. L., & Low, S. M. (1990). The built environment and spatial form. Annual Review of Anthropology, 19 (1), 453–505. Lazăr, C., Mărgărit, M. & Bălășescu, A. (2016). Dogs, jaws, and other stories: Two symbolic objects made of dog mandibles from southeastern Europe. Journal of Field Archaeology, 41 (1), 101–117.

References 

 545

Lazăr, C., Mărgărit, M., & Radu, V. (2018). Evidence for the production and use of Lithoglyphus naticoides beads in Europe during the Holocene: The case of Sultana-Malu Roşu site (Romania). Quaternary International, 472 (A), 84–96. Léa, V. (2003). Mesurer, quantifier et croiser. Une méthode d’approche pour les industries lithiques du Chasséen méridional. Histoire & Mesure, XVIII (1–2), 3–39. Leshtakov, P. (2004). Graphite deposits and some aspects of graphite use and distribution in Bulgarian Chalcolithic. In V. Nikolov, K. Băčvarov & P. Kalchev (Eds.), Prehistoric Thrace (pp. 485–498). Sofia: Institute of Archaeology. Leshtakov, P. (2005). The sources and distribution of graphite as a means of decoration in the Bulgarian Chalcolithic. In G. Dumitroaia, J. Chapman, O. Weller et al. (Eds.), Cucuteni 120 years of research. Time to sum up (pp. 293–297). Piatra Neamţ: Centrul de Cercetare a culturii Cucuteni. Leusch, V., Pernicka, E., & Armbruster, B. (2014). Chalcolithic gold from Varna: provenance, processing, circulation and function. In H. Meller, R. Risch & E. Pernicka (Eds.), Metals of power. Early gold and silver (pp. 165–182). Halle:Tagungen des Landesmuseums fur Vorgeschichte Halle 11/1 Levi-Strauss, C. (1962). La pensée sauvage. Paris: Plon. Lichardus, J. & Lichardus-Itten, M. (1985). La Protohistoire de l’Europe. Le Néolithique et le Chalcolithique entre la Méditerranée et la mer Baltique. Paris: Nouvelle Clio I bis, P.U.F. Lillesand, T., Kiefer, R., & Chipman, J. (2014). Remote Sensing and Image Interpretation. New Jersey: John Wiley & Sons, Inc. Liu, L. (1996). Settlement Patterns, Chiefdom Variability, and the Development of Early States in North China. Journal of Anthropological Archaeology, 15 (3), 237–88. Retrieved doi:10.1006/ jaar.1996.0010. Llobera, M. (2003). Extending GIS-based visual analysis: the concept of visualscapes. International Journal of Geographical Information Science, 17, 25–48. Llobera, M., Wheatley, D., Steele, J. et al. (2010). Calculating the inherent visual structure of a landscape (inherent viewshed) using high-throughput computing. Budapest: Archaeolingua. Loveday, R. (2015). Religious routine and pilgrimage in the British Isles. In C. Fowler, J. Harding & D. Hofmann (Eds.), The Oxford Handbook of Neolithic Europe (pp. 463–479). Oxford: Oxford University Press. Macphail, R. & Goldberg, P. (2010). Archaeological materials. In G. Stoops, V. Marcelino & F. Mees (Eds.), Interpretation of Micromorphological Features of Soils and Regoliths (pp. 589–622). Oxford: Elsevier. Madden, G., Karsten, J., Sokhatsky, M., & Heins, S. (2018). Violence at Verteba Cave: cranial trauma as evidence for intergroup conflict in late Neolithic Ukraine. International Journal of Osteoarchaeology, 28 (1), 44–53. Makkay, J. (1984). Early stamp seals in South-East Europe. Budapest: Akadémiai Kiadó. Makkay, J. (2005). Supplement to the Early stamp seals of South-East Europe. Budapest: privately published. Manolakakis, L. (2005). Les industries lithiques énéolithiques de Bulgarie. Rahden: Verlag Marie Leidorf GmbH. Manolakakis, L. (2008). Le mobilier en silex taille des tombes de Varna 1. Acta Musei Varnaensis, VI, 115–136. Manzura, I. (2005). Steps to the Steppe: Or, How the North Pontic Region Was Colonised. Oxford Journal of Archaeology, 24 (4), 313–38. Retrieved doi:10.1111/j.1468-0092.2005.00239.x. Marcus, H. (1994). A history of Ethiopia. Berkeley, CA: University of California Press. Mareş, I. (2002). Metalurgia aramei în neo-eneoliticul României. Suceava: Editura Bucovina Istorică. Martin, L., Edwards, Y., & Garrard, A. (2010). Hunting practices at an Eastern Jordanian Epipalaeolithic aggregation site: the case of Kharaneh IV. Levant, 42 (2), 107–135.

546 

 References

Masson, V. (1990). Tripolskoe obshtestvo i ego sotsialno-ekonomicheske harakteristiki. In V. Zbenovich (Ed.), Rannezemledelcheskie poselenie-giganti tripolskoi kulturi na Ukraine. Tezisi dokladov pervogo polevogo seminara (pp. 8–10). Taljanki: Institut Arheologii AN USSR. Mateva, B. (2011). Exploiting Flint Deposits in Northeastern Bulgaria in the Chalcolithic. In S. Mills & P. Mirea (Eds.), Lower Danube in Prehistory: Landscape Changes And Human-Environment Interactions. Proceedings of the International conference, Alexandria, 3-5 November 2010 (pp. 173–178). Bucuresti: Renaissance. McAnany, P. (n.d.). Discussion paper. Paper presented at Amerind Foundation Workshop, “Cities or Big villages?”. New approaches to great, anomalous places, Dragoon, AZ, May 2014. McCorriston, J. (2011). Pilgrimage and household in the ancient Near East. Cambridge: Cambridge University Press. McIntosh, R. (2005). Ancient Middle Niger: Urbanism and the Self-Organizing Landscape. Cambridge: Cambridge University Press. McIntosh, S. (1995). Pathways to complexity: An African perspective. In S. McIntosh (Ed.), Beyond chiefdoms: Pathways to complexity in Africa (pp. 9–30). Cambridge: Cambridge University Press. Meadow, R. (1981). Early Animal Domestication in South Asia: a First Report of the Faunal Remains from Mehrgarh, Pakistan. In H. Hartel (Ed.), South Asian Archaeology 1979 (pp. 143–179). Berlin: Dietrich Reimer Verlag. Menotti, F. & Korvin-Piotrovskiy, A. (Eds.) (2012). The Tripolye Culture Giant-Settlements in Ukraine: Formation, Development and Decline. Oxford: Oxbow Books. Mercer, R. (2006). The first known enclosures in southern Britain: their nature, function and role, in time and space. In A. Harding, S. Sievers & N. Venclová (Eds.), Enclosing the past: Inside and Outside in Prehistory. Sheffield Archaeological Monographs 15 (pp. 69–76). Sheffield: Equinox. Milesi García, L. (2018). Lugares de agregación en la prehistoria desde una perspektiva etnográfica. Los recintos de fosos como caso de estudio. Unpub. Ph.D., University of Granada. Milisauskas, S. (1978). European prehistory. New York: Academic Press. Milisauskas, S. & Kruk. J. (1984). Settlement Organization and the Appearance of Low Level Hierarchical Societies during the Neolithic in the Bronocice Microregion, Southeastern Poland. Germania, 62 (1), 1–30. Millard, A. (2014). Conventions for reporting radiocarbon determinations. Radiocarbon, 56 (2), 555–559. Retrieved doi: 10.2458/56.17455 Miller, C. E., Conard, N. J., Goldberg, P., & Bema, F. (2010). Dumping, sweeping and trampling: experimental micromorphological analysis of anthropogenically modified combustion features. Palethnologie. Révue francophone en Préhistoire. http:// palethnologie.org. Mircea, M. & Alexianu, M. (2007). A re-examination of salt resources for the Cucuteni - Tripolye area. In D. Monah, Gh. Dumitroaia, O. Weller & J. Chapman (Eds.), L’exploitation du sel à travers le temps. BMA XVIII (pp. 199–220). Piatra Neamţ: Centre de Recherches sur la civilisation Cucuteni. Mlekuž, D. (2005). The ethnography of the Cyclops: Neolithic pastoralists in the East Adriatic. Documenta Praehistorica, XXXII, 15–51. Monah, D. (1997). Plastică antropomorfă a culturii Cucuteni-Tripolie. Piatra Neamţ: Centrul de Cercetare a culturii Cucuteni. Monah, D. (2003). A ghost is haunting Europe: the Neolithic proto-cities. In A. Korvin-Piotrovskiy, V. Kruts & S. Ryzhov (Eds.), Tripolian Settlement-Giants: The International Symposium Materials (pp. 239–243). Kiev: Korvin-press. Monah, D. (2016). Anthropomorphic representations in the Cucuteni-Tripolye culture. Oxford: Archaeopress.

References 

 547

Monah D. & Monah, F. (1997). The last great Chalcolithic civilization of Old Europe. In C-M. Mantu, Gh. Dumitroaia & A. Tsaravopoulos (Eds.), Cucuteni. The last great Chalcolithic civilization of Europe (pp. 15–95). Thessaloniki: Athena Printing House. Moore, T. (2017). Alternatives to urbanism? Reconsidering oppida and the urban question in Late Iron Age Europe. Journal of World Prehistory, 30 (3), 281–300. Moran, P. (1950). Notes on Continuous Stochastic Phenomena. Biometrika, 37, 17–23. Müller, J., Czebreszuk, J., & Kadrow, S. (Eds.) (2017). Majdanetske 2013. New excavations at a Trypillia Megasite. Studien zur Archäologie in Ostmittleuropa Band 16. Bonn: Rudolf Habelt. Müller, J., Hofmann, R., Brandstätter, S. et al. (2016). Chronology and Demography: How Many People Lived in a Megasite? In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 133–70). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Müller, J., Hofmann, R., & Ohlrau, R. (2016a). From Domestic Households to Mega-Structures: ProtoUrbanism? In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 253–68). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Müller, J., Hofmann, R., Ohlrau, R., & Shatilo, M. (2018). The social constitution and political organisation of Tripolye megasites: hierarchy and balance. In H. Meller, D. Gronenborn & R. Risch (Eds.), Surplus without the state - political forms in Prehistory. 10th Archaeological Conference of Central Germany October 19-21, 2017 (pp. 247–262). Halle (Saale): Tagungen des Landesmuseums für Vorgeschichte (Halle) 18. Müller, J. & Pollock, S. (2016). Trypillia and Uruk. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 281–88). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Müller, J. & Rassmann, K. (2016). Introduction. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 1–5). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Müller, J., Rassmann, K., & Videiko, M. (Eds.) (2016b). Trypillia Megasites and European Prehistory 4100-3400 BCE. Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Müller, J. & Videiko, M. (2016). Maidanetske: new facts of a megasite. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 71–94). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Nam, K. (2013). Lasting Monuments and Durable Institutions: Labor, Urbanism, and Statehood in Northern Vietnam and Beyond. Journal of Archaeological Research, 21 (3), 217–67. Nam, K. (2015) The Origins of Ancient Vietnam: An Archaeological History. Oxford: Oxford University Press. Nandris, J. (1978). Some features of Neolithic climax societies. Studia Praehistorica, 1–2, 198–211. Nebbia, M. (2016) Early cities or large villages?: settlement dynamics in the Trypillia group, Ukraine. Unpublished Ph.D. thesis, Durham University. Nebbia, M., Gaydarska, B., Millard, A., & Chapman. J. (2018). The Making of Chalcolithic Assembly Places: Trypillia Megasites as Materialized Consensus among Equal Strangers? World Archaeology, 50 (1), 41–61. Retrieved doi: 10.1080/00438243.2018.1474133 Neiman, F. (1997). Conspicuous Consumption as Wasteful Advertising: A Darwinian Perspective on Spatial Patterns in Classic Maya Terminal Monument Dates. Archeological Papers of the American Anthropological Association, 7 (1), 267–90. Retrieved doi:10.1525/ap3a.1997.7.1.267. Němejcová-Pavúková, V. (1995). Svodín. Zwei Kreisgrabenanlagen der Lengyel-Kultur. Bratislava: Comenius University of Bratislava. Nikolova, A. & Pashkevich, G. (2003). K Voprosu Ob Urovne Razvitiya Zamledeliya Tripolskoj Kultury. In A. Korvin-Piotrovskiy, V. Kruts & S. Ryzhov (Eds.), Tripolian Settlement-Giants: The International Symposium Materials (pp. 89–95). Kiev: Korvin-pres.

548 

 References

Ogundiran, A. (2013). Towns and states in the West African forest belt. In P. Mitchell & P. Lane (Eds.), The Oxford Handbook of African archaeology (pp. 861–73). Oxford: Oxford University Press. Ogundiran, A. (n.d.). Ile Ife, Nigeria. Paper presented at Amerind Foundation Workshop, “Cities or Big villages?”. New approaches to great, anomalous places, Dragoon, AZ, May 2014. Ohlrau, R. (2015). Trypillia Großsiedlungen: Geomagnetische Prospektion und architektursoziologische Perspektiven. Journal of Neolithic Archaeology, 17, 17–99. Ohlrau, R., Dal Corso, M., Kirleis, V., & Müller, J. (2016). Living on the Edge? Carrying Capacities of Trypillian Settlements in the Buh-Dnipro Interfluve. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE (pp. 207–20). Themes in Contemporary Archaeology Volume 2. Abingdon: Routledge. Oleniuc, F. & Bejenaru, L. (2011). Preliminary Data Concerning the Manufacturing of Animal Raw Materials in the Chalcolithic Cucuteni B Settlement of Poduri-Dealul Ghindaru, Romania. In J. Baron & B. Kufel-Diakowska (Eds.), Written in Bones (pp. 263–72). Wrocław: University of Wroclaw. Orphanides, L. (2010). Laura Orphanides. In D. Bailey, A. Cochrane & J. Zambelli (Eds.), Unearthed. Norwich: Sainsbury Centre for Visual Arts, 109. Orton, C. (2000). Sampling in Archaeology. Cambridge: Cambridge University Press. Orton, D., Gaastra, J., & Vander Linden, M. (2016). Between the Danube and the deep blue sea: zooarchaeological meta-analysis reveals variability in the spread and development of Neolithic farming across the western Balkans. Open Quaternary, 2, 6. Ovchinnikov, E. (2012). Keramichni posud Trypilskogo poselennya Nebelivka (za rezultatami rozkopok 2009 r.). Arkheolohiya i davnya istoriya Ukrayiny (Institute of Archaeology, Kyiv) 8, 30–37. Ovchinnikov, E.V. (2014). Trypil's'ka kul'tura Kanivs'koho Podniprov’ya (etapy BII — CI). Kyiv, NASU, Institute of Archaeology. Ovchinnikov, E. (2015). Rozpysna keramika z megastruktury na trypilskomu poselenni Nebelivka. In M. Videiko, J. Chapman, I. Kozir & V. Sobchuk (Eds.), At the Eastern frontiers of Old Europe. Conference programme and abstracts, Kirovograd International Conference, 12th - 14th May 2015 (pp. 12–14). Kirovograd: Kirovograd County Council. Palaguta, I. (2007). Tripolye culture during the beginning of the Middle Period (B1). International Series 1666. Oxford: BAR. Papakosta, V., Smittenberg, R., Gibbs, K. et al. (2015). Extraction and derivatization of absorbed lipid residues from very small and very old samples of ceramic potsherds for molecular analysis by gas chromatography–mass spectrometry (GC–MS) and single compound stable carbon isotope analysis by gas chromatography–combustion–isotope ratio mass spectrometry (GC–C–IRMS). Microchemical Journal, 123, 196–200. Retrieved doi.org/10.1016/j.microc.2015.06.013. Parkinson, W. (2002). Integration, Interaction, and Tribal ‘Cycling’: The Transition to the Copper Age on the Great Hungarian Plain. In W. Parkinson (Ed.), The Archaeology of Tribal Societies (pp. 391–438). Ann Arbor: International Monographs in Prehistory. Parkinson, W. & Galaty, M. (2007). Secondary States in Perspective: An Integrated Approach to State Formation in the Prehistoric Aegean. American Anthropologist, 109, 113–29. Parnic, V. & Chiriac, D. (2001). Aşezarea eneolitică de la Mariuta.Considerăţii preliminare asupra habitatulului. Civilizaţie la Dunărea de Jos, XVI–XVII, 129–206. Pashkevitch, G. (1997). Early Farming in Ukraine. In J. Chapman and P. Dolukhanov (Eds.), Landscapes in flux: Central and Eastern Europe in antiquity (pp. 267–274). Oxford: Oxbow Books. Pashkevitch, G. (2005). Palaeoethnobotanical evidence of the Tripolye culture. In G. Dumitroaia, J. Chapman, O. Weller, et al. (Eds.), Cucuteni 120 years of research, time to sum up (pp. 213–245). Piatra-Neamţ: Centrul de Cercetare a culturii Cucuteni.

References 

 549

Pashkevitch, G. & Videiko, M. (2006). Rilnytstvo plemen Tripilskoi kulturi. Kiev: Institute of Archaeology NANU. Pashkevych, G. (2012). Environment and Economic Activities of Neolithic and Bronze Age Populations of the Northern Pontic Area. Quaternary International, 261, 176–82. Retrieved doi:10.1016/j.quaint.2011.01.024. Pashkevych, G. (2014). Data of the Trypillian Agriculture: New Palaeoethnobotanical Evidence. Abstract. Cucuteni 130 Conference, Piatra Neamţ, 24–27. Passek, T. (1949). Tripolskoe Poselenie Vladimirovka. Kratkie Soobshcheniya Instituta Istorii Materialnoj Kultury, 26, 47–56. Passek, T. (1949a). Periodizatsiya Tripolskih Poselenij III-II Tys. Do N.e. Materialy i issledovakiya po Archeologii SSSR. 10th ed. Kiev. Passek, T. (1961). Rannezemledelchiskie (Tripolskie) Plemena Podnestrovia. Moskow Pătroi, C. (2011). Caracterul gumelniţean al culturii eneolitice Sălcuţa. Buletinul Muzeului Judeţean Teleorman. Seria Arheologie, 3, 143–154. Pauketat, T. R. (2009). Cahokia: America's great city on the Mississippi. New York: Penguin. Pauketat, T. (n.d.). Cahokia, Mississippi Basin, USA. Paper presented at Amerind Foundation Workshop, “Cities or Big villages?”. New approaches to great, anomalous places, Dragoon, AZ, May 2014. Paunier, D., Barral, P., Luginbühl, T., & Paratte, C-A. (1994). Système de déscription et de gestion du mobilier céramique. Glux-en-Glenne: Centre Européen d’archéologie du Mont Beuvray & Université de Lausanne. Pawlikowski, M. (1992). The Origin of Lithic Raw Materials. In J. Kozłowski, H. Laville & B. Ginter (Eds.), Temnata Cave (pp. 241–286). Kraków: Jagiellonian University Kraków. Payne, S. (1973). Kill-off patterns in sheep and goats: the mandibles from Aşvan Kale. Anatolian Studies, 23, 281–308. Peregrine, P. (2004). Cross-Cultural Approaches in Archaeology: Comparative Ethnology, Comparative Archaeology, and Archaeoethnology. Journal of Archaeological Research, 12 (3), 281–309. Peterson, C. E. & Drennan, R. D. (2012) Patterned Variation in Regional Trajectories of Community Growth. In M. E. Smith (Ed.), The Comparative Archaeology of Complex Societies (pp. 88–137). New York: Cambridge University Press. Petrescu-Dâmboviţa, M. (2001). Eneoliticul dezvoltat. In M. Petrescu-Dâmboviţa & A. Vulpe (Eds.), Istoria Românilor, vol. I. Moştenirea timpurilor îndepărtate (pp. 154–168). Bucureşti: Editura Academiei Române. Petrova, V. (2007). Ceramic assemblage of the Late Chalcolithic Karanovo VI culture in Thrace: phases I and II. Acta Terrae Septemcastrensis, VI, 53–72. Petrova, V. (2011). Main characteristics and development of graphite ornamentation during the Late Chalcolithic in Thrace. Studia Praehistorica, 14, 311–332. Petrun, V. (2001–02). Zakljuchenie o rezultatah arheologo-petrograficheskogo izucheniya materialov iz Maidenetskogo. Stratum Plus, 2, 137–140. Philp, B., Parfitt, K., Willson, J. et al. (1991). The Roman villa site at Keston, Kent. First report (excavations 1968-1978). Dover: Kent Archaeological Rescue Unit. Piazza, A., Rendine, S., Minch, E. at al. (1995). Genetics and the Origin of European Languages. Proceedings of the National Academy of Sciences of the United States of America, 92 (13), 5836–40. Retrieved doi:10.1073/pnas.92.13.5836. Pichartz, N., Hofmann, R., Dreibrodt, S. et al. (2019). Deciphering archaeological contexts from the magnetic map: determination of daub distribution and mass of Chalcolithic house remains. The Holocene. Retrieved DOI: 10.177/0959683619857238. Pichkur, E. (2005). Orudiya truda iz kollektsii materialov. poluchennykh pri issledovanii ploshchadki №4 v Dobrovodakh. In V. Kruts, A. Korvin-Piotrovskiy, S. Ryzhov et al. (Eds.), Issledovanie

550 

 References

poselenie-gigantov Tripolskoi kulturi v 2002 - 2004 gg (pp. 119–139). Kyiv: Institute of Archaeology. Pichkur, E. (2008). Kremnevie izdelya s poseleniya Taljanki. In A. Korvin-Piotrovskiy & F. Menotti (Eds.), Tripolye culture in Ukraine: the giant-settlement of Talianki (pp. 152–167). Kyiv: Institute of Archaeology. Pikirayi, I. (n.d.). Great Zimbabwe, Southern Africa. Paper presented at Amerind Foundation Workshop, “Cities or Big villages?”. New approaches to great, anomalous places, Dragoon, AZ, May 2014. Piscitelli, M. (n.d.) Re-evaluating the case for America’s first cities: evidence from the Norte Chico region of Peru. Paper presented in Symposium 76, What to do with ‘megasites’ in prehistory? Further exploring the ‘megasite’ conundrum, SAA 2017, Vancouver, BC, 29/III - 2/IV/2017. Pogoševa, A. (1985). Die Statuetten der Tripolie-Kultur. Beitrage zur allgemeinen und vergleichenden archeologie. Band 7, 95–242. Popova, L. (2016). Occlusal pattern of cheek teeth in extant Spermophilus: a new approach to the identification of species. Journal of Morphology, 277, 814–825. Popova, M. (2012). Possibilities for Establishing Local Groups Based on Graphite Decoration on Ceramics within the Gumelnitsa - Karanovo VI Culture. Results from a Study on the Ceramics from Tell Vinitsa, Shumen District. Pontica, 45, 27–38. Popova, M. & Kostov, R. (2017). Gold and silver-like (graphite) glittering decoration: Symmetry patterns on Chalcolithic (5th Mill. BC). Pottery from Eastern Bulgaria. Symmetry: Culture and Science, 28 (4), 409–419. Pottier, C. (1999). Carte Archéologique de la Région d’Angkor. Zone Sud. Ph.D thesis, 3 vols. Université Paris III - Sorbonne Nouvelle (UFR Orient et Monde Arabe). Quinn, P. (2013). Ceramic Petrography: The Interpretation of Archaeological Pottery and Related Artefacts in Thin Section. Oxford: Archaeopress. Raduntcheva, A. (2003). Kusnoeneolitnoto obstestvo v bulgarskite zemi. Razkopki i prouchvania XXXII. Sofia: AIM BAN. Randsborg, K. & Merkyte, I. (Eds.) (2005). Lîga. Copper Age Strategies in Bulgaria. Acta Archaeologica, 76 (1), 1–194. Rapoport, A. (1982). The meaning of the built environment: A nonverbal communication approach. 2nd ed. Tucson, AZ: University of Arizona Press. Rassamakin, J. (2004). Die nordpontische Steppe in der Kupferzeit. Gräber aus der Mitte des 5. Jts. bis Ende des 4. Jts. v. Chr. Teil I-II. Archäologie in Eurasien, Band 17. Mainz: Verlag Philipp von Zabern. Rassamakin, Y. & Menotti, F. (2011). Chronological Development of the Tripolye Culture GiantSettlement of Talianki (Ukraine): 14C Dating vs. Pottery Typology. Radiocarbon, 53, 645–657. Rassmann, K., Korvin-Piotrovskiy, A., Videiko, M., & Müller, J. (2016). The New Challenge for Site Plans and Geophysics: Revealing the Settlement Structure of Giant Settlements by Means of Geophysical Survey. In J. Müller, K. Rassmann & M. Videiko (Eds.), Trypillia Megasites and European Prehistory 4100-3400 BCE. Themes in Contemporary Archaeology Volume 2 (pp. 29–54). Abingdon: Routledge. Rassmann, K., Ohlrau, R., Hofmann, R. et al. (2014). High Precision Tripolye Settlement Plans, Demographic Estimations and Settlement Organization. Journal of Neolithic Archaeology, 16, 96–134. Retrieved doi:10.12766/jna.2014.3. Redman, C. (1987). Surface Collection, Sampling and Research Design: A Retrospective. American Antiquity, 52 (2), 249–65. Reimer, P. J., Bard, E., Bayliss, A. et al. (2013). IntCal13 and Marine13 Radiocarbon Age Calibration Curves 0-50,000 Years cal BP. Radiocarbon, 55 (4),1869–1887. Reingruber, A. (2012). Copper-Age House Inventories from Pietrele: Preliminary Results from Pottery Analysis. In R. Hofmann, F.-K. Moetz & J. Müller (Eds.), Tells: Social and Environmental Space,

References 

 551

Proceedings of the International Workshop Socio-Environmental Dynamics over the Last 12,000 Years: The Creation of Landscapes II (14th-18th March 2011) in Kiel, Volume 3 (pp. 139–152). Bonn: Verlag Dr. Rudolf Habelt GmbH. Roe, J. (n.d.). Spatial Analysis of Intra-Site Surface Collection Data from a Cucuteni-Tripolye Megasite. Unpublished BSc Dissertation, Durham University (2011). Rollefson, G. (1989). The Aceramic Neolithic of the Southern Levant: the view from ‘Ain Ghazal. Paléorient, 15 (1), 135–140. Ross, S. A., Sobotkova, A., Tzvetkova, J. et al. (2018). The Tundzha Regional Archaeology Project. Oxford: Oxbow Books. Russell, N. & Martin, L. (2005). The Çatalhöyük Mammal Remains. In I. Hodder (Ed.), Inhabiting Çatalhöyük: Reports from the 1995-1999 Seasons (pp. 33–98). Cambridge: McDonald Institute. Ryndina, N. (1998). Drevneishee metaloobrabativajushtee proizvodstvo Jugo-Vostochnoi Evropi. Moskva: URSS. Ryndina, N. & Engovatova, A. (1990). Opyt planigraphicheskogo analiza kremnevyh orudij tripolskogo poseleniya Drutsy I. In V. Zbenovich (Ed.), Rannezemledelcheskie PoseleniyaGiganty Tripolskoj Kultury Na Ukraine. Tezisy Dokladov Pervogo Polevogo Seminara (pp. 108–14). Taljanki: Institute of Archaeology of the AS of the USSR. Ryzhov, S. (1990). Mikrohronologiya Tripolskogo Poseleniya u s. Taljanki. In V. Zbenovich (Ed.), Rannezemledelcheskie Poseleniya-Giganty Tripolskoj Kultury Na Ukraine. Tezisy Dokladov Pervogo Polevogo Seminara (pp. 83–90). Taljanki: Institute of Archaeology of the AS of the USSR. Ryzhov, S. (1993). Nebelivska grupa pamyatov Trypilskoi kultury. Archaeologia, 3, 101–114. Ryzhov, S. (1999). Keramika Poselen Trypilskoi Kultury Bugo-Dniprovskogo Mezhyruzhzhia yak Istorychne Dzerelo. Unpublished PhD Thesis, Institute of Archaeology of the NAS of Ukraine, Kyiv. Ryzhov, S. (2000). Raspisnaya keramika tomashovskoj lokalnoj gruppy Tripolskoj kultury. Stratum Plus, 2, 459–73. Ryzhov, S. (2007). Suchasnyj stan vyvchennia kuturno-istorychnoi spilnosti Cucuteni-Trypillia. In Yu. Rassamakin & S. Ryzhov (Eds.), O. Olzhych. Archeologiya (pp. 437–77). Kiev: Vydavnytstvo im. Oleny Teligy. Ryzhov, S. (2008). Kuhonnaya keramika tripolskogo poseleniya. In A. Korvin-Piotrovskiy & F. Menotti (Eds.), Tripolye culture in Ukraine: the giant-settlement of Talianki (pp. 132–151). Kyiv: Institute of Archaeology. Ryzhov, S. (2012) Relative Chronology of the Giant Settlement Period BII-CI. In F. Menotti & A. KorvinPiotrovskiy (Eds.), The Tripolye Culture Giant-Settlements in Ukraine: Formation, Development and Decline (pp. 79–115). Oxford: Oxbow Books. Ryzhov, S. (2012a). Tripolian Pottery of the Giant-settlements: Characteristics and Typology. In F. Menotti & A. Korvin-Piotrovskiy (Eds.), The Tripolye Culture Giant-Settlements in Ukraine: Formation, Development and Decline (pp. 139–168). Oxford: Oxbow Books. Sabins, F. (2007). Remote Sensing: Principles and Interpretation. 3rd ed. Long Grove III: Waveland Pr. Inc. Sahlins, M. (1958). Social stratification in Polynesia. Seattle: University of Washington Press. Schiffer, M. (1976). Behavioral Archaeology. New York: Academic Press. Schiffer, M. (1987). Formation processes of the archaeological record. Albuquerque, NM: University of New Mexico Press. Schmandt-Besserat, D. (2010). The token system of the ancient Near East: its role in counting, writing, the economy and cognition. In I. Morley & C. Renfrew (Eds.), The archaeology of measurement (pp. 27–34). Cambridge: Cambridge University Press. Schmidt, H. (1932). Cucuteni in der oberen Moldau. Berlin-Leipzig: De Gruyter.

552 

 References

Schofield, A. (Ed.) (1991). Interpreting Artefact Scatters: Contribution to Ploughzone Archaeology. 4th ed. Oxford: Oxbow Monograph. Sekerskaya, O. P. (2017). Analiz arheozoologichnogo kompleksu tripolskogo poseleniya Nebelivka. Arheologia (Kyiv), 2017 (3), 18–23. Semple, S. (2018). Editorial. World Archaeology, 50 (1), 1–6. Service, E. (1962). Primitive Social Organization: An Evolutionary Perspective. New York: Random House. Shatilo, L. (2005). Glinyani modeli budivel kulturi Kukuteni-Trypillia. In V. Kruts, A. KorvinPiotrovskiy, S. Ryzhov et al. (Eds.), Issledovanie poselenie-gigantov Tripolskoi kulturi v 2002 - 2004 gg (pp. 119–139). Kyiv: Institute of Archaeology. Shatilo, L. (2015). Board Games of Early Farmers in South-Eastern Europe: Artifacts and Problems. In C. Virag (Ed.), Neolithic cultural phenomena in the Upper Tisza Basin (pp. 267–288). Satu Mare: Editura Muzeului Sătmărean. Shea, J. (2006). Child’s Play: Reflections on the Invisibility of Children in the Paleolithic Record. Evolutionary Anthropology, 15, 212–216. Shennan, S. (1989). Introduction: archaeological approaches to cultural identity. In S. Shennan (Ed.), Archaeological approaches to cultural identity (pp. 1–32). London: Routledge. Shevchenko, N. (2015). Issledovanie materijalov drevnego hrama. In M. Videiko, J. Chapman, I. Kozir & V. Sobchuk (Eds.), At the Eastern frontiers of Old Europe. Conference programme and abstracts, Kirovograd International Conference, 12th - 14th May 2015 (pp. 40–43). Kirovograd: Kirovograd County Council. Shevchenko, N. & Ovchinnikov, E. (2003). Stroitelnyye materialy i konstruktivnyye elementy v tripolskom domostroitelstve. In A. Korvin-Piotrovskiy, V. Kruts & S. Ryzhov (Eds.), Tripolian Settlement- Giants: The International Symposium Materials (pp. 194–198). Kiev: Korvin-press. Shishkin, V. (1973). Z Praktiki Deshifruvannya Aero-Fotoznimkiv U Arheologichnih Tsiliyah. Arheologiya, 10, 32–41. Shishkin, V. (1985). Planuvannya Tripilskih Poselen Za Danimi Aerofotoziomki. Arheologiya, 52, 72–77. Shmaglij, N. (1980). Krupnye Tripolskie Poseleniya v Mezhdurechje Dnepra i Yuzhnogo Buga. In I. Artemenko (Ed.), Pervobytnaya Archeologiya Poiski I Nahodki (pp. 198–203). Kiev: Naukova dumka. Shmaglij, N. & Videiko, M. (1992). Trypilski poselennia na Cherkashchyni. Archeologiya, 3, 124–130. Shmaglij, N. & Videiko, M. (2001–2). Maidanetskoe – tripolskij proto-gorod. Stratum Plus, 2, 44–140. Shukurov, A., Videiko, M., Sarson, G. et al. 2015. Productivity of Premodern Agriculture in the Cucuteni-Trypillia Area. Human Biology, 87 (3), 235–82. Shumilovskikh, L. S., Novenko, E., & Giesecke, T. (2018). Long-term dynamics of the East European forest-steppe ecotone. Journal of Vegetation Science 29, 416–426. Retrieved DOI: 10.1111/ jvs.12585. Shydlovsky, P. & Slesariev, E. (2015). Technologichni osoblyvosti kremianogo kompleksu ranniotrypil’s’kogo poselennia Bernashivka 1. Kamiana doba Ukrainy, 16, 204–24. Siklósi, Zs. (2013). Traces of Social Inequality during the Late Neolithic in the Eastern Carpathian Basin. Dissertationes Pannonicae Ser. IV. Vol. 3. Budapest: L’Harmattan. Simmons, A. H. (2007) The Neolithic Revolution in the Near East: transforming the human landscape. Tucson, AZ: University of Arizona Press. Şimon, M. & Parnic, V. (2001). Aspecte ale eneoliticului din SE Câmpiei Române. Aşezarea de la Seinoiu. Civilizaţie la Dunărea de Jos, XVI–XVII, 223–233. Sirakov, N. (2002). Flint artefacts in prehistoric grave-good assemblages from the Durankulak necropolis. In H. Todorova (Ed.), Durankulak, Band II. Die Prähistorishen Gräberfelder. Teil 1 (pp. 213–46). Sofia: Deutsches Archäologisches Institut.

References 

 553

Skakun, N. (2004) Predvaritel’nie rezultati izucheniya materialov tripol’skogo poseleniya Bodaki (kremneobrabativajushtie kompleksi). In G. Korobkova (Ed.), Orudiya truda i sistemi zhizneobezpecheniya naseleniya Evrazii (pp. 57–79). Sankt-Petersburg: Evropeyskiy Dom. Skakun, N. (2012). Excavations at Bodaki, Ukraine, and the importance of flint working activities in the eneolithic. In J.-C. Marquet & C. Verjux (Eds.), L’Europe, déjà, à la fin des temps préhistoriques. Actes de la table-ronde internationale, Tours (Indre-et-Loire, France) 7 Septembre 2007 (pp. 91–108). Archea. Skakun, N., Samzun, A., Mateva, B., & Terekhina, V. (2014). Features of flint mining and processing during the Chalcolithic period in the Southeast of Europe (based on the materials of the Tripolian Bodaki settlement, Ukraine). In F. Bostyn & F. Giligny (Eds.), Lithic Raw Material Resources and Procurement in Pre- and Protohistoric Times. BAR International Series 2656 (pp. 93–105). Oxford: Archaeopress. Smith, A. (2003). The political landscape. Constellations of authority in early complex polities. Berkeley and Los Angeles: University of California Press. Smith, M. E. (2007). Form and meaning in the earliest cities. A new approach to ancient urban planning. Journal of Planning History, 6 (1), 3–47. Smith, M. E. (2010). The archaeological study of neighborhoods and districts in ancient cities. Journal of Anthropological Archaeology, 29, 137–154. Smith, M. E. (2016). How Can Archaeologists Identify Early Cities: Definitions, Types, and Attributes. In M. Fernández-Götz & D. Krausse (Eds.), Eurasia at the Dawn of History: Urbnaization & Social Change (pp. 153–168). New York: Cambridge University Press. Smith, M. L. (2003). The social construction of cities. Washington DC: Smithsonian Books. Smith, M. L. (2008). Urban empty spaces. Contentious places for consensus-building. Archaeological Dialogues, 15 (2), 216–231. Retrieved doi:10.1017/S1380203808002687 Sorokin, V. (1991). Orudiia truda i khoziajstvo plemen srednego Tripolia Dnestrovsko-Prutskogo mezhdurechia. Chişinau: Stiinţa. Souvatzi, S. (2008). A Social Archaeology of Households in Neolithic Greece. An Anthropological Approach. Cambridge: Cambridge University Press. Spatz, H. (1999). Das mittelneolithische Gräberfeld von Trebur, Kreis Gros-Gerau. Materialien für Vor- und Frühgeschichte. Wiesbaden: Hessen 19. Steadman, S. (1996). Recent research in the archaeology of architecture: Beyond the foundations. Journal of Archaeological Research, 4, 51–93. Stevanović, M. (1997). The age of clay: the social dynamics of house-construction. Journal of Anthropological Archaeology, 16, 334–395. Tasić, N. N., Marić, M., Penezić, P. et al. (2015). The end of the affair: formal chronological modelling for the top of the Neolithic tell of Vinča-Belo Brdo. Antiquity, 89, 1062–84. Tasić, N. (1989). Karanovo V/VI und das Neolithikum des östlichen teils von Jugoslawien. In S. Hiller (Ed.), Tell Karanovo und das Balkan Neolithikum (pp. 127–138). Salzburg: Institut fur Klassische Archaologie der Universitat Salzburg. Taylor, J. (2000). Cultural depositional processes and post-depositional problems. In R. Francovich, H. Patterson & G. Barker (Eds.), Extracting Meaning from Ploughsoil Assemblages (pp. 16–28). Oxford: Oxbow. Taylor, T. (1999). Envaluing metal: theorizing the Eneolithic ‘hiatus’. In S. Young, A. Pollard, P. Budd and E. Ixer (Eds.), Metals in antiquity. International Series 792 (pp. 22–32). Oxford: BAR. Teklenburg, J. A. F., Timmermans, H. J. P., & van Wagenberg, A. F. (1993). Space synax: Standardised integration measures and some simulations. Environment and Planning B: Planning and Design, 20 (3), 347–357. Telegin, D. (1976). Pro nomenklaturnyh spysok kremianykh vyrobiv doby mezolitu-neolitu. Arkheologija, 8, 21–45.

554 

 References

Ţerna, S. (2017). K voprosu rasprostranenii melkoi antropomorfnoi plastiki na poseleniah kulturi Kukuteni-Tripolie, KSIA, 247, 223–44. Thomas, D. (1975). Nonsite Sampling in Archaeology: Up the Creek Without a Site? In J. Mueller (Ed.), Sampling in Archaeology (pp. 61–81). Tucson: The University of Arizona Press. Thomas, J. (1999). Rethinking the Neolithic. 2nd ed. London: Routledge. Tilley, C. (1994). A phenomenology of landscape: places, paths, and monuments. Oxford: Berg. Tkachuk, T. (2005). Semiotichnii analiz tripil’skoi-kukutenskih znakovih sistem. Vinnitsia: Nova kniga. Tkachuk, T. (2010–11). Naibilsha pamiyatka kul’turi Tripilia-Cucuteni Talianki - poseleniya chi sakralnii tsentr? Ukrainoznavchi studii, 11–12. Tobler, W. (1970). A Computer Movie Simulating Urban Growth in the Detroit Region. Economic Geography, 46, 234–70. Toderaș, M., Hansen, S., Reingruber, A., & Wundderlich, J. (2009). Pietrele-Măgura Gorgana: o așezare eneolitică la Dunărea de Jos, între 4500-4250 î.e.n. Materiale și Cercetări Arheologice (serie nouă), V, 39–90. Todorova, N. & Matsanova, V. (2000). Late Chalcolithic Ceramic Style at Yunatsite Tell (Approach to the Systematization of the Ceramic from the New Excavations Levels). In L. Nikolova (Ed.), Technology, Style and Society. Contributions to the Inovations between the Alps and the Black Sea in Prehistory. International Series 854 (pp. 331–361). Oxford: Archaeopress. Tovkailo, M. (2005). Neolit Stepovoho Pobuzhzhia. Kam’iana doba Ukrainy, 6. Kyiv: Shliakh. Trigger, B. (2003). Understanding early civilizations: a comparative study. Cambridge: Cambridge University Press. Tringham, R. (2005). Weaving house life and death into places: a blueprint for a hypermedia narrative. In D. Bailey, A. Whittle & V. Cummings (Eds.), (Un)settling the Neolithic (pp. 98–111). Oxford: Oxbow Books. Tringham, R. & Krstić, D. (Eds.) (1990). Selevac: A neolithic village in Yugoslavia, Monumenta archaeologica. Los Angeles: UCLA Institute of Archaeology. Tringham, R. & Krstić, D. (1990a). Conclusion: Selevac in the wider context of European Prehistory. In R. Tringham & D. Krstić (Eds.), Selevac. A Neolithic village in Serbia (pp. 567–616). Monographs in Archaeology Vol. 15. Los Angeles: University of California Press Tsvek, E. (1996). Structure of the Eastern Tripolye culture. In Gh. Dumitroaia & D. Monah (Eds.), Cucuteni aujourd’hui (pp. 89–113). Piatra-Neamţ: Bibliotheca Memoriae Antiquitatis. Tsvek, E. & Movchan, I. (2005). Enioliticheskiy proizvodstvennii kompleks po dobiche i obrabotke kremniya na reke Bol’shaya Vis. In Na poshanu Sofii Stanislavivni Berezanskoi (pp. 52–57). Kyiv: Shliakh. Tsvek, E., Movchan, I., & Skakun, N. (2012). Eneolithic centres of flint processing on the territory of Central Ukraine. In J.-C. Marquet & C. Verjux (Eds.), L’Europe, déja, à la fin des temps préhistoriques. Actes de la table-ronde internationale, Tours (Indre-et-Loire, France) 7 Septembre 2007 (pp. 109–119). Archea. Tsvek, E. & Rassamakin, Yu. (2005). The interactions between the Eastern Tripolye culture and the Pontic steppe area: some aspects of the problem. In G. Dumitroaia, J. Chapman, O. Weller et al. (Eds.), Cucuteni 120 years of research, time to sum up (pp. 173–192). Piatra-Neamţ: Centrul de Cercetare a culturii Cucuteni. Tuan, Y.-F. (1977). Space and place: The perspective of experience. London: E. Arnold. Turner, A. (2001). Depthmap: a program to perform visibility graph analysis. In Proceedings of the 3rd International Symposium on Space Syntax (pp. 31.1). Atlanta, GA. Turner, A. (2003). Analysing the visual dynamics of spatial morphology. Environment and Planning B: Planning and Design, 30, 657–676. Turner, A. (2004). Depthmap 4: A researcher’s handbook. London: Bartlett School of Graduate Studies, UCL.

References 

 555

Turner, A., Doxa, M., O’Sullivan, D., & Penn, A. (2001). From isovists to visibility graphs: a methodology for the analysis of architectural space. Environment and Planning B: Planning and Design, 28, 103–121. Turner, A. & Penn, A. (1999). Making isovists syntactic: isovist integration analysis. In Proceedings of the 2nd International Symposium on Space Syntax (pp. 1–9). Brasilia: Universidade de Brasilia. Twiss, K., Wolfhagen, J., Madgwick, R. et al. (2017). Horses, hemiones, hydruntines? Assessing the reliability of dental criteria for assigning species to southwest Asian equid remains. International Journal of Osteoarchaeology, 27, 298–304. Retrieved doi: 10.1002/oa.2524. Ugwuanyi, J. K. & Schofield, J. (2018). Permanance, temporality and the rhythms of life: exploring the significance of the village arena in Igbo culture. World Archaeology, 50 (1) 7–22. Underhill, A. (n.d.). Processes of site abandonment in Ancient China: some proposed new directions for research. Paper presented to Session 65, Living abandonment: the social process of detachment from place, SAA 78th Annual Meeting, Honolulu, 3rd - 7th April 2013. Ur, J. (2003). CORONA Satellite Photography and Ancient Road Networks: A Northern Mesopotamian Case Study. Antiquity, 77 (295), 102–15. Ur, J. (2005). Sennacherib’s Northern Assyrian Canals: New Insights from Satellite Imagery and Aerial Photography. Iraq, 67 (1), 317–45. Ur, J. (2014). Households and the Emergence of Cities in Ancient Mesopotamia. Cambridge Archaeological Journal, 24 (2), 249–68. Retrieved doi:10.1017/S095977431400047X Van Nes, A. (2009). Measuring the degree of street vitality in excavated towns: how can macro and micro spatial analyses tools contribute to understandings on the spatial organization of urban life in Pompeii? In D. Koch, L. Marcus & J. Steen (Eds.), Proceedings of the 7th International Space Syntax Symposium, Stockholm, Sweden (p. 1). Trita-ARK: Forskningspublikationer. Varien, M. & Mills, B. (1997). Accumulations research: Problems and prospects for estimating site occupation span. Journal of Archaeological Method and Theory, 4, 141–191. Videiko, M. (1996). Die Großsiedlungen der Tripol’e-Kultur in der Ukraine. Eurasia Antiqua 1, 46–80. Videiko, M. (Ed.) (2004). Entsiklopedia Tripil’skoi tsyvilizatsii, Vol. 1. Kyiv: Ukrpoligrafmedia. Videiko, M. (2005). Arhitektura tripil’skoi kul’turi VI–III tis do n.e. In M. Videiko, R. Terpilovskiy & V. Petrashenko (Eds.), Davni poselienia Ukraini (pp. 9–80). Kyiv: Akademperiodika. Videiko, M. (2007). Contours and Contents of the Ghost: Trypillia Culture Proto-Cities? Memoria Antiquitatis, 24, 251–76. Videiko, M. (2008). Ukraina: ot Tripolya do Antov. Kyiv: KVITS. Videiko, M. (2012). Kompleksnoe Izuchenie Krupnykh Poselenij Tripolskoj Kultury 1971–2011. Stratum Plus, 2, 225–263. Videiko, M. (2013). Kompleksnoe Izuchenie Krupnykh Poselenij Tripolskoj Kultury V – IV Tys Do N.e. Saarbrücken: Lambert Academic Publishing. Videiko, M. & Burdo, N. (2004). Jivotnovodstvo tripol’skoi kul’turi v period 5400-2750 do n.e. In A. Kosko & M. Szmyt (Eds.), Nomadism and Pastoralism at the confluence of the Wisła and Dnieper (Neolithic, Eneolithic, Bronze Age) (pp. 67–93). Poznań: Wydawnictwo Poznanskie. Voinea, V. (2005). Ceramica gumelniţeană fazele A1 şi A2. Constanța: Ex Ponto. von den Driesch, A. (1976). A Guide to the Measurement of Animal Bones from Archaeological Sites. Cambridge, MA: Harvard University Peabody Museum. von Stern, E. (1907). Die ‘prämykenische Kultur’ in Süd-Russland. Bericht Über Die Ausgrabungen in Petreni 1902 Und 3. Archeologicěskogo Cezda v Jekaterinoslav, 13, 1–95. Vysloužilova, B., Ertlen, D., Schwartz, D., & Šefrna, L. (2016). Chernozem: from concept to classification: a review. AUC Geographica, 51 (1), 85–95. Waagen, J. (2014). Evaluating Background Noise: Assessing Off-Site Data from Field Surveys around the Italic Sanctuary of S. Giovanni in Galdo, Molise, Italy. Journal of Field Archaeology, 39, 417–29. Retrieved doi:10.1179/0093469014Z. 00000000099.

556 

 References

Wallerstein, I. (1974). The Modern World-System. I, Capitalist Agriculture and the Origins of the European World-Economy in the Sixteenth Century. New York and London: Academic Press. Ward, G. & Wilson, S. (1978). Procedures for comparing radiocarbon age determinations: a critique. Archaeometry, 20 (1), 19–31. Weiner, A. (1992). Inalienable possessions: the paradox of keeping-while-giving. Berkeley: University of California Press. Weller, O. & Dumitroaia, Gh. (2005). The earliest salt production in the world: an early Neolithic exploitation in Poiana Slatinei-Lunca, Romania. Antiquity Gallery, 79 (306), online gallery: http://www.antiquity.ac.uk/projgall/weller306 Weller, O., Dumitroaia, Gh., Sordoillet, D. et al. (2008). Première exploitation du sel en Europe. Techniques et gestion de l’exploitation de la source salée de Poiana Slatinei à Lunca (Neamţ, Roumaine). In O. Weller, A. Dufraisse & P. Pétrequin (Eds.), Sel, eau et forêt. Hier et aujourd’hui (pp. 205–230). Besançon: Presses Universitaires de Franche-Comté. Wengrow, D. (2001). The evolution of simplicity: aesthetic labour and social change in the Neolithic Near East. World Archaeology, 33 (2), 168–188. Wengrow, D. (2015). Cities before the State in Early Eurasia. Jack Goody Lecture in Max Planck Institute for Social Anthropology. Department ‘Resilience and Transformation in Eurasia’. Wheatley, D. (2004). Making space for an archaeology of place. Internet Archaeology 15, https://intarch.ac.uk/journal/issue15/wheatley_index.html. Wheatley, D. & Gillings, M. (2002). Spatial technology and archaeology: the archaeological applications of GIS. London: Taylor and Francis. White, K. (in prep.). Low-density Megasites: a global comparative appraisal. (To be submitted as University of Sydney PhD thesis). Whitelaw, T. (2013). Collecting Cities: some Problems and Prospects. In P. Johnson & M. Millett (Eds.), Archaeological survey and the City (pp. 70–106). Oxford: Oxbow. Whittle, A. (Ed.) (2007). The early Neolithic on the Great Hungarian Plain: investigations of the Körös culture site of Ecsegfalva 23, County Békés. Budapest: Institute of Archaeology, Hungarian Academy of Sciences. Whittle, A., Pollard, J., & Grigson, C. (1999). The harmony of symbols: the Windmill Hill causewayed enclosure. Cardiff Studies in Archaeology. Oxford: Oxbow Books. Widell, M., Hritz, C., Ur, J., & Wilkinson. T. (2013). Land Use of the Model Communities. In T. Wilkinson, G. McGuire & M. Widell (Eds.), Models of Mesopotamian Landscapes: How Small-Scale Processes Contributed to the Growth of Early Civilizations. BAR International Series 255, 2nd ed. (pp. 56 –80). Oxford, Archaeopress. Wiessner, P. (1983). Style and social information in Kalahari San projectile points. American Antiquity, 48 (2), 253–276. Wilk, R. & Rathje, W. (1982). Household archaeology. The American Behavioral Scientist, 25, 617–639. Wilkinson, T. (1982). The Definition of Ancient Manured Zones by Means of Extensive Sherd-Sampling Techniques. Journal of Field Archaeology, 9 (3), 323–33. Retrieved doi:10.1179/009346982791504616. Wilkinson, T. (1989). Extensive Sherd Scatters and Land-Use Intensity: Some Recent Results. Journal of Field Archaeology, 16 (1), 31–46. Retrieved doi:10.2307/529879. Wilkinson, T. (2003). Archaeological Landscapes of the Near East. Tucson, AZ: University of Arizona Press. Wilkinson, T., Philip, G., Bradbury, J. et al. (2014). Contextualizing Early Urbanization: Settlement Cores, Early States and Agro-Pastoral Strategies in the Fertile Crescent During the Fourth and Third Millennia BC. Journal of World Prehistory, 27 (1), 43–109. Retrieved doi:10.1007/s10963014-9072-2.

References 

 557

Wilson, G. & Pauketat, T. R. (n.d.). The organizational implications of architecture at Moundville and Cahokia. Paper presented in Symposium 20, Cities, large villages or neither? The conundrum of ‘megasites’ in prehistory, SAA 2017, Vancouver, BC, 29/III - 2/IV/2017. Windler, A., Thiele, R., & Müller, J. (2013). Increasing Inequality in Chalcolithic Southeast Europe: The Case of Durankulak. Journal of Archaeological Science, 40 (1), 204–10. Retrieved doi:10.1016/j. jas.2012.08.017. Wirth, L. (1938). Urbanism as a Way of Life. American Journal of Sociology, 44, 1–24. Wiseman, J. & El-Baz, F. (2007). Remote Sensing in Archaeology. New York: Springer Science and Business Media. Yanushevich, Z. (1989). Early agriculture North of the Black Sea. In D. Harris & G. Hillman (Eds.), Foraging and farming: the evolution of plant exploitation (pp. 607–619). London: Unwin Hyman. Zbenović, V. (1996). Siedlungen der frühen Tripol’e-Kultur zwischen Dnestr und Südlichem Bug. Archäologie in Eurasien, Band 1. Espelkamp: Verlag Marie Leidorf GmbH. Zbenovich, V. (1989). Ranny etap Trypol’skoi kul’tury na territorii Ukrainy. Kyiv: Naukova dumka. Zbenovich, V. (1996). The Tripolye Culture: Centenary of Research. Journal of World Prehistory, 10 (2), 199–241. Retrieved doi:10.1007/BF02221076. Zbenovich, V., Kruts, V., & Shmaglij, M. (1966). Arheologichni pamiatki Ukrainskoi RSR. (Korotkii spisok). Kyiv: Naukova dumka. Zeder, M. & Lapham, H. (2010). Assessing the reliability of criteria used to identify postcranial bones in sheep, Ovis, and goats, Capra. Journal of Archaeological Science, 37, 2887–2905. Zhuravlev, O. (2008). Tvarynnytstvo Ta Myslyvstvo U Trypilskyh Plemen Na Terytorii Ukrainy. Kiev: Shlyah. Zimmermann, A. (1995). Austauschsysteme von Silexartefakten in der Bandkeramik Mitteleuropas. Bonn: Habelt. Zinkovskiy, K. (1975). Do problem trypilskogo zhytlobuduvannya. Archeologiya, 15, 13–22. Žižek, S. (2007). How to read Lacan. New York: W. W. Norton. Žižek, S. (2012). Less than nothing. Hegel and the shadow of dialectical materialism. London: Verso.

558 

 List of Figures

List of Figures Figure 1.1: Map of Cucuteni-Trypillia distribution (by M. Nebbia) — 2 Figure 1.2: Timeline of Cucuteni-Trypillia group — 3 Figure 1.3: Cucuteni A pottery, Drăguşeni (by B. Gaydarska, based upon Crîşmaru 1977, Fig. 20) — 3 Figure 1.4: Trypillia BII – CI pottery, Bug-Dnieper Interfluve (by L. Woodard, based upon Ryzhov 2012, Figs. 6.4–6.5) — 4 Figure 1.5: Cucuteni-Trypillia figurines (by Y. Beadnell, based upon Monah D. 1997) — 6 Figure 1.6: Settlement model for Central and Eastern Europe: key – darker shades show higher densities of a site type, lighter shades lower densities (by C. Unwin) — 7 Figure 1.7: Early remote sensing of Yatranivka: (a) plot of air photograph; (b) geophysical plot (by L. Woodard, based upon Videiko 2013) — 10 Figure 1.8: Geophysical plan of the 2009 season overlain on satellite image of Nebelivka (by M. Nebbia, based on Hale et al. 2010) — 11 Figure 1.9: Two experimental ‘Trypillia’ houses in the process of construction (by S. Johnston) — 14 Figure 2.1: The ‘Maximalist’ model (by C. Unwin) — 33 Figure 2.2: The Trypillia ‘Big Other’ (by C. Unwin) — 38 Figure 2.3: Realistic and stylised figurines (by B. Gaydarska) — 43 Figure 2.4: The Scânteia vessel with multiple symmetries (source: Monah D & M 1997, Fig. 46) — 45 Figure 3.1: Coverage of CORONA imagery for the study area (by M. Nebbia; copyright: The Project) — 63 Figure 3.2: Coverage of the acquired WorldView-2 satellite images (8-band multispectral 1.85 m and panchromatic 0.46 m) for the Nebelivka micro-region (5 km radius) and panchromatic (0.46 m) satellite image for the Nebelivka macro-region (25 km radius) (by M. Nebbia; copyright: The Project) — 64 Figure 3.3: Distribution of anomalies mapped on the WorldView-2 satellite image, Nebelivka microregion (by M. Nebbia; copyright: The Project) — 66 Figure 3.4: Distribution of anomalies interpreted as traces of a palaeo-channel network, the Nebelivka micro-region (by M. Nebbia; copyright: The Project) — 67 Figure 3.5: Anomalies mapped on the WorldView-2 pansharpened multispectral 8-band (0.46 m) satellite image, which have been interpreted as having anthropic origins (by M. Nebbia; copyright: The Project) — 68 Figure 3.6: Shallow depth of the top of the anthropogenic deposit on the North-Eastern part of Nebelivka, allowing a better visibility of the anomalies compared to the rest of the site (by M. Nebbia) — 70 Figure 3.7: Two views of the Trypillia site of Perehonivka (BII): clearly visible in crop-free field conditions (upper) and totally invisible when the field is cultivated (lower): WorldView-2 panchromatic images (0.46m resolution), acquired on April 2008 (upper) and September 2011 (lower) — 71 Figure 3.8: Comparison between two extreme examples of barrows as they appear on the WorldView-2 satellite image and on the ground (by M. Nebbia; copyright: The Project) — 72 Figure 3.9: Distribution of all the anomalies that can be interpreted as burial mounds mapped within the Nebelivka macro-region, with Trypillia sites by Phase (copyright: The Project) — 74 Figure 3.10: Upper: Interpretation of the geomagnetic survey, South-East corner of Nebelivka. Visible are parts of the two concentric circuits of houses (by J. Roe); Lower: sherd densities across the surveyed land units (by M. Nebbia) — 78 Figure 3.11: Interpolated contour plot of daub (upper) and pottery (lower) densities by number of fragments (by J. Roe) — 79



List of Figures 

 559

Figure 3.12: Fields surveyed with a non-site sampling strategy in 2009, 2012 and 2013, Nebelivka hinterland, to assess the definition of ‘sites’ from surface scatters (by M. Nebbia) — 82 Figure 3.13: Areas covered by the field survey in the 2012–2014 seasons (by M. Nebbia) — 84 Figure 3.14: Distribution map of all sites recovered during the 2012–2014 field survey seasons, including the locations of known Trypillia sites in the Nebelivka macro-region (by M. Nebbia) — 85 Figure 3.15: Surface material counts by site sampling transects at the newly discovered Trypillia settlement of Kutsa (20 km North-East of Nebelivka) (by M. Nebbia) — 87 Figure 3.16: Sampling transects on the Trypillia site of Volodymyrivka. The colours indicate the quantity of both pottery and burnt daub collected at each sample. The red line shows the actual limit of the built-up area (by M. Nebbia) — 89 Figure 3.17: Material counts of the 6 transects walked on the BII megasite of Volodymyrivka (by M. Nebbia) — 89 Figure 3.18: Upper: Boxplot of site areas reported in the Encyclopaedia by phase, showing the megasites sizes as outliers; Lower left: plot of GINI coefficients of Trypillia site sizes by phase; Lower right: plot of Trypillia megasites vs. smaller sites by phase (by M. Nebbia) — 91 Figure 3.19: Distribution of the 499 Trypillia sites used in the study, derived from the Encyclopaedia (Videiko 2004) (by M. Nebbia) — 98 Figure 3.20: Distribution map of known Trypillia settlements, highlighting the location of megasites in the Southern Bug-Dnieper interfluve (by M. Nebbia) — 100 Figure 3.21: Histograms displaying Trypillia site sizes by phase. Phase A (N=33 sites); Phase BI (N=46); Phase BII (N=176); Phase CI (N=234); Phase CII (N=85) (by M. Nebbia) — 102 Figure 3.22: Spatial distribution of Trypillia settlements by Phase: upper left – Phase A; upper right – Phase BI; middle left – Phase BII; middle right – Phase CI; lower left – Phase CII (by M. Nebbia) — 104 Figure 3.23: 5-km hinterlands of megasites in the Southern Bug-Dnieper interfluve (by M. Nebbia) — 108 Figure 4.1: Pollen diagram, Nebelivka P1 core (by B. Albert) — 113 Figure 4.2: Human impact proxies, Nebelivka P1 core: pollen zone, pollen zone depth and human impact depth columns show darker shades with increased depth (by C. Unwin) — 116 Figure 4.3: Geophysical plan of Nebelivka showing magnetic gradient data (by J. Watson) — 124 Figure 4.4: Interpretative geophysical plan of Nebelivka (by J. Watson) — 125 Figure 4.5: Interpretative geophysical plan of Nebelivka, showing boundaries of Quarters (by Y. Beadnell on the basis of D. Hale’s geophysical plan) — 126 Figure 4.6: Tightly-spaced and loosely-spaced houses, Nebelivka (by J. Watson) — 131 Figure 4.7: Upper & middle: gaps and kinks in house circuits; lower: converging inner radial streets, Nebelivka (by J. Watson) — 132 Figure 4.8: Assembly Houses 1–11, Nebelivka (by J. Watson) — 134 Figure 4.9: Assembly Houses 12–23, Nebelivka (by J. Watson) — 135 Figure 4.10: Squares and short inner radial streets, Nebelivka (by J. Watson) — 140 Figure 4.11: Upper: megasite entrances and the main palaeo-channel; lower: short inner radial streets and blocking structures, Nebelivka (by J. Watson) — 141 Figure 4.12: Upper: blocking streets; lower: poorly burnt dwelling houses and pits, Nebelivka (by J. Watson) — 143 Figure 4.13: Upper: linear pits and short inner radial streets; lower: strong anomalies possibly representing ‘kilns’, Nebelivka (by J. Watson) — 144 Figure 4.14: Upper: anomalies outside the Northern end of the megasite; lower: main palaeo-channel with kinks in house circuits, Nebelivka: numbers refer to Assembly Houses (by J. Watson) — 147

560 

 List of Figures

Figure 4.15: (1) bivariate plot of house sizes; (2) histogram of all house sizes; (3) house sizes by Outer Zone, Nebelivka (by J. Chapman) — 150 Figure 4.16: (1) house sizes outside the Outer Circuit; (2) house sizes in cross streets; (3) house sizes in Squares, Nebelivka (by J. Chapman) — 152 Figure 4.17: Maximum class sizes by Quarter; (1) Class 3 equal to, or fewer than, other size Classes; (2) Class 3 dominant, with Class 4 but no Class 5; (3) Class 3 dominant, with Classes 4 and 5, Nebelivka (by J. Chapman) — 156 Figure 4.18: (1) mean breadth of all house size classes; (2) spread of house size classes; (3) No. of houses vs. GINI House Size Co-efficient plot by Quarters, Nebelivka ((1) & (2) by J. Chapman; (3) by M. Nebbia) — 157 Figure 4.19:Visibility Graph Analysis of Quarter L: (1) Connectivity; (2) Point First Moment; (3) Point Second Moment; (4) Visual Entropy; (5) Visual Integration: HH; (6) Visual Integration: P-value; (7) Visual Integration: TEK; (8) Visual Mean Depth; (9) Visual Node Count; (10) Visual Relativised Entropy (by B. Buchanan). For explanation of these models, see Table 4.5 — 167 Figure 4.20: Connectivity analysis of all 10 Quarters: (1) B; (2) C; (3) D; (4) G; (5) H; (6) I; (7) L; (8) F; (9) N; (10) M (by B. Buchanan) — 168 Figure 4.21: (1) Average measurements of VGA analysis of entirety of structural evidence; (2) Average VGA measurements of Model A; (3) Average VGA measurements of Model B (by B. Buchanan) — 169 Figure 4.22: VGA Mean Depth analyses of all Stages of both Models, Quarter F, Nebelivka: (1) Distributed Governance Model (A), Stage 1; (2) Model A, Stage 2; (3) Model A, Stage 3; (4) Assembly Model (B), Stage 1; (5) Model B, Stage 2; (6) Model B, Stage 3 (see Table 4.9) (by B. Buchanan) — 176 Figure 4.23: VGA Integration-TEK analyses of all Stages of both Models, Quarter F, Nebelivka: (1) Distributed Governance Model (A), Stage 1; (2) Model A, Stage 2; (3) Model A, Stage 3; (4) Assembly Model (B), Stage 1; (5) Model B, Stage 2; (6) Model B, Stage 3 (see Table 4.10) (by B. Buchanan) — 178 Figure 4.24: VGA Mean Depth analyses of all Stages of both Models, Quarter L, Nebelivka: (1) Distributed Governance Model (A), Stage 1; (2) Model A, Stage 2; (3) Model A, Stage 3; (4) Assembly Model (B), Stage 1; (5) Model B, Stage 2; (6) Model B, Stage 3 (see Table 4.11) (by B. Buchanan) — 179 Figure 4.25: VGA Integration-TEK analyses of all Stages of both Models, Quarter L, Nebelivka: (1) Model (A), Stage 1; (2) Model A, Stage 2; (3) Model A, Stage 3; (4) Assembly Model (B), Stage 1; (5) Model B, Stage 2; (6) Model B, Stage 3 (see Table 4.12) (by B. Buchanan) — 180 Figure 4.26: Upper: Burnt house remains (ploshchadka), Mega-structure Context 55; lower: bone and ceramic scatter, Pit near House B17, Nebelivka (upper by J. Chapman; lower by M. Videiko) — 183 Figure 4.27: Upper: platform, Durham Experiment; lower: withy impressions, Nebelivka Experimental Burnt House Excavation (by S. Johnston) — 185 Figure 4.28: House collapse scenario 1 (1-storey houses) (by L. Woodard) — 187 Figure 4.29: House collapse scenario 2 (2-storey houses) (by L. Woodard) — 188 Figure 4.30: Upper: house panels falling inwards, Nebelivka House Burning Experiment; lower: house panel fallen outwards, Nebelivka Experimental Burnt House Excavation Contexts 425 & 426 (upper by M. Nebbia; lower by J. Chapman) — 189 Figure 4.31: Kite photo of Mega-structure, Nebelivka: North to Right side; burnt area 36m East-West (by M. Houshold) — 196 Figure 4.32: Digitised remains, Nebelivka Mega-structure; Phase 3 Lower (by M. Nebbia) — 197 Figure 4.33: Digitised remains, Nebelivka Mega-structure; Phase 3 Upper (by M. Nebbia) — 198 Figure 4.34: Digitised remains, Nebelivka Mega-structure; Phases 2 and 3 Lower & Upper (by M. Nebbia) — 200



List of Figures 

 561

Figure 4.35: Upper: digitised remains, Phase 2. Key to Context Numbers: Fired Clay Bin – 80; Platforms – 6, 46, 58, 89, 176 & 272; Podium – 29; Threshold – 120; Platform 257 was excavated on the last day of Week 7 and therefore not digitised; it is found in Grid Square C21–22; lower: fired clay slot, Nebelivka Mega-structure (upper by M. Nebbia; lower by J. Chapman) — 201 Figure 4.36: Burnt timber remains, Nebelivka Experimental Burnt House Excavation: upper: burnt timber fallen obliquely, Context 419; lower: timber void, Context 226 (by J. Chapman) — 202 Figure 4.37: Upper: long North-South section (the width of both walls reflects daub tumble out from the walls); lower: three sections across podium, Nebelivka Mega-structure (by C. Unwin) — 203 Figure 4.38: (1) Durham reconstruction; (2)–(5) Bayesian plots of AMS dates, Nebelivka Megastructure (1 by C. Unwin; 2–5 by A. Millard) — 204 Figure 4.39: Upper: general view from East; lower: Stage 4 placement of vessel from North; Fired Clay Bin, Nebelivka Mega-structure (by J. Chapman) — 206 Figure 4.40: Platform 46 from East, Nebelivka Mega-structure (by J. Chapman) — 207 Figure 4.41: Plot of daub firing temperatures and vitrified daub, Nebelivka Mega-structure (by M. Nebbia based upon information from N. Shevchenko); Colour Key for daub firing temperatures: blue: 200–4000C; yellow: 400–9000C; orange: >9000C — 210 Figure 4.42: General view of Nebelivka barrow from South-West (by J. Chapman) — 213 Figure 4.43: Distribution of all excavated areas, Nebelivka (by M. Nebbia) — 215 Figure 4.44: Distribution of burnt houses, unburnt houses and Assembly Houses (by M. Nebbia) — 216 Figure 4.45: Typical Test Pit stratigraphy, Test Pit 26/6, Nebelivka (by C. Unwin) — 218 Figure 4.46: Sections across (1) Fired Clay Bin; (2) Contexts 215 & 310; (3) Platform 46; and (4) Contexts 210 and 310 (by C. Unwin) — 219 Figure 4.47: Depth of burnt houses in excavation units and Test Pits (by C. Unwin) — 220 Figure 4.48: Upper: section of thick pile of destruction daub, Test Pit 13/3; (a) South-facing; (b) East-facing; lower: sections of platform below destruction daub, Test Pit 15/2; (a) East-facing; (b) South-facing (by L. Woodard). Numbers in Figs. 4.48–4.50 refer to general pit stratigraphic sequence in Fig. 4.45 — 222 Figure 4.49: Upper: sections of two well-defined layers of destruction daub, Test Pit 33/1; (a) SSEfacing; (b) WSW-facing; lower: sections of two-storey house with platform above destruction daub, Test Pit 26/5: (a) West-facing; (b) North-facing (by L. Woodard) — 223 Figure 4.50: (1): West-facing section of unburnt house, Test Pit 1/4; (2): North-facing section of burnt house with platform daub with pit under floor, Test Pit 17/1; (3) SW-facing section of platform in Assembly House, Test Pit 27/3; (4) section of mound of burnt house debris, Test Pit 24/4; (a) ESE-facing; (b) NNE-facing (by L. Woodard) — 224 Figure 4.51: Distribution of Test Pits with in-situ vs. dispersed platform daub (by M. Nebbia) — 225 Figure 4.52: Distribution of number of layers of Destruction Daub by Test Pit (by M. Nebbia) — 227 Figure 4.53: Sections of Pit, Sondazh 1: (1) East-facing; (2) West-facing; (3) East-facing; (4) Southfacing; (5) West-facing (by C. Unwin) — 229 Figure 4.54: Upper: plan of base of pit; Pit, Sondazh 1 (by L. Woodard); lower: proportion of burnt and ‘unburnt’ houses by Neighbourhood and Quarter (by J. Chapman) — 232 Figure 4.55: Upper: North Ditch profile from South-East; lower: Triple Ditch profiles 1–3 from East (by L. Woodard) — 235 Figure 4.56: Plan of House A9, Nebelivka (by L. Woodard) — 237 Figure 4.57: Upper: sections of House A9: B–B1 – South-facing; C–C1 – North-facing; A–A1 – Eastfacing (by L. Woodard) — 238 Figure 4.58: General view of Industrial feature, Nebelivka (by M. Videiko) — 242

562 

 List of Figures

Figure 4.59: Hungarian Mediaeval brick kilns as analogies for Nebelivka ‘industrial’ feature; (1) Békéscsaba–Mezőmegyer; (2) Debrecen–Józsa Pláza (by B. Gaydarska, based upon Jakab 2011) — 244 Figure 4.60: Corn-drying ovens, Keston Roman villa (internal width of South Oven – 4.15m) (by B. Gaydarska, adapted from Philp et al. 1991, Fig. 22) — 245 Figure 4.61: Site plan showing Test Pits with(out) AMS dates, Nebelivka (by M. Nebbia) — 249 Figure 4.62: Bayesian plots of AMS dates, Nebelivka: (1) all AMS dates plotted on calibration curve; (2) start and end dates for occupation at Nebelivka; (3) duration of occupation at Nebelivka (by A. Millard) — 251 Figure 4.63: Bayesian plots of AMS dates, Nebelivka: (1) start and end dates for circuits and streets modelled independently; (2) start and end dates for Quarters; (3) AMS dates from the Pit, Sondazh 1; (4) start and end dates for three megasites (by A. Millard) — 254 Figure 4.64: Building materials analysis: (1) finger action; (2) squared timber-impression; (3) withyimpression; (4) finger-impression; (5) vitrified daub; (6) thermal plot, daub, Context 117; (7) cross-section of daub showing clay layers, Nebelivka (by N. Shevchenko) — 259 Figure 5.1: Pottery types used in the Ovchinnikov ceramic system (by Ovchinnikov 2014, 80; see our Table 5.1) — 269 Figure 5.2: Taphonomy: (1) burnt sherd, Test Pit 20/1; (2) burnt sherd, Test Pit 33/1; (3) sherd with heavy wear, Test Pit 21/2; (4) sherd with heavy deposition, Test Pit 22/4; (5) sherd with moderate deposit, Test Pit 26/5; (6) vitrified sherd, Test Pit 24/3; (7) vitrified sherd, Megastructure Context 208; (8) wear on base of sherd, Test Pit 1/3 (by K. Harding) — 272 Figure 5.3: Pottery production: (1) handle pushed in, interior, Test Pit 24/3; (2) handle pushed in, exterior, with potting lines, Test Pit 24/3; (3) sherd with grit temper, Mega-structure Context 35; (4) sherd with possible wheel-marks, Mega-structure, TsT 8958, Context 3; (5) sherd with grooved decoration, House A9; (6) base with mat impression, Mega-structure TsT 1905, Context 151; (7) sherd with shell temper, Mega-structure Context 143; (8) sherd with interior potting lines, barrow (by K. Harding) — 275 Figure 5.4: Sherd re-fits: (1) Type 1 re-fit, Test Pit 1/3; (2) Type 2 re-fit, Mega-structure Context 232; (3) Type 1 and 2 re-fits, Mega-structure Contexts 157 and 180; (4) Type 2 re-fit, Test Pit 1/3; (5) Type 3 re-fit, Mega-structure Contexts 157 and 4 (by K. Harding) — 278 Figure 5.5: (1) number, (2) weight (kg) and (3) mean sherd weight (g) of pottery groups by Excavation Unit (by J. Chapman) — 281 Figure 5.6: Density of pottery samples by weight (g), Test Pits (by M. Nebbia) — 283 Figure 5.7: Density of pottery samples from Test Pits placed in the centre of burnt houses (Zone 9) (by M. Nebbia) — 284 Figure 5.8: Fabric colours: (1) Fabric C; (2) Fabric A; (3) Fabric B; (4) Fabrics I–J; (5) Fabrics E–F; (6) Fabrics G–H; (7) Fabric D; (8) Fabrics O–P; (9) Fabrics K–L; (10) Fabrics M–N; (11) Fabrics Q–R (by K. Harding) — 286 Figure 5.9: Fabric distribution for (1) House A9; (2) Mega-structure; (3) Test Pits; and (4) Pit, Sondazh 1; Fabric by sherd number for (5) Quarter G; (6) Test Pit 24/3; (7) Test Pit 25/1; and (8) Test Pit 25/2 (by J. Chapman) — 289 Figure 5.10: Surface colour vs. vessel form for (1) bowls; (2) carinated vessels; (3) plates; and (4) dishes, House A9; distribution of rim types without Bases for (5) Pit, Sondazh 1; (6) Test Pits; (7) Mega-structure; (8) House A9 (by J. Chapman) — 290 Figure 5.11: (1) distribution of shape types, all Units; distribution of open & closed categories without Bases: (2) all Units; (3) Pit, Sondazh 1; (4) Test Pits; (5) Mega-structure; (6) House A9; (7) Pit, Sondazh 1 SU2; and (8) Pit, Sondazh 1 SU4 (by J. Chapman) — 291 Figure 5.12: (1) Minimum Number of Vessel estimates for Episodes and deposits outside Episodes, Pit, Sondazh 1; (2) Vessel sizes by excavation unit (by J. Chapman) — 294



List of Figures 

 563

Figure 5.13: Upper: distribution of decorated vs. undecorated sherds; lower: distribution of rims by weight, Mega-structure (by M. Nebbia) — 296 Figure 5.14: Distribution of (1) impressed sherds by weight; (2) plates, dishes and necked dishes; (3) bowls by weight; (4) pottery found on Living Floor (Phase 2) by weight, Mega-structure (by M. Nebbia) — 298 Figure 5.15: Coarse ware decorative motifs. Numbers (e.g., 3.1) refer to Motif Numbers. Key: T – Test Pits; M – Mega-structure; P – Pit, Sondazh 1; A – House A9 (by L. Woodard) — 300 Figure 5.16: Fine ware exterior painted motifs (by L. Woodard) — 301 Figure 5.17: Fine ware interior painted motifs (by L. Woodard) — 302 Figure 5.18: Fine ware exterior (rows 1–4) and interior (rows 5–6) painted motifs (by L. Woodard) — 303 Figure 5.19: Motif linkage plans for four most common motifs, megasite (by M. Nebbia) — 307 Figure 5.20: Motif linkage plan, Quarter B, megasite (by M. Nebbia) — 309 Figure 5.21: (1) Composition of burnt house assemblages by vessel shape; (2) ratio of open: closed vessels; Taljanki (T) and Majdanetske (M); (3) regression analysis of painted signs vs. sample size, Bug-Dnieper Interfluve sites, based upon Tkachuk 2005 (by J. Chapman) — 318 Figure 5.22: Distribution of (upper) fine wares; (lower) painted wares, House A9 (by M. Nebbia) — 320 Figure 5.23: Distribution of (upper) coarse wares; (lower) sherds with impressed decoration, House A9 (by M. Nebbia) — 321 Figure 5.24: Distribution of (upper) bowls; (lower) dishes, House A9 (by M. Nebbia) — 322 Figure 5.25: Distribution of (upper) plates; (lower) summary diagram, House A9 (by M. Nebbia) — 323 Figure 5.26: (1) distribution of painted signs shared between houses, Majdanetske; (2) comparison of Nebelivka painted motifs with painted signs on other Trypillia sites; (3) comparison of painted signs on Bug-Dnieper Interfluve sites (by J. Chapman, based upon information in T. Tkachuk 2005) — 325 Figure 5.27: Alternative pathways to making Trypillia figurines (by L. Woodard) — 328 Figure 5.28: Decorated anthropomorphic figurines from Pit, Sondazh 1: (1) SF 23 and (2) SF 28 + 43: (a) back, (b) front, Nebelivka (by K. Harding) — 329 Figure 5.29: Anthropomorphic figurines: (1) male, SF 3230; (2) non-gendered, Grid F12; Megastructure, Nebelivka (by V. Pankowski) — 330 Figure 5.30: Figurine body parts by excavation unit: (1) House A9; (2) Mega-structure; (3) Pit, Sondazh 1; and (4) Test Pits; Key to figurine parts: H – head; HT – head-torso; HB – headbuttock; HL – head-leg; T – torso; TL – torso-leg; TF – torso-foot; B – buttock; BL – buttock-leg; L – leg; LF – leg-foot; F – foot; TZOO – torso of zoomorph. (5) gender characteristics of figurines; (6) condition of figurines by excavation unit (by J. Chapman) — 332 Figure 5.31: Upper: zoomomorphic figurine, Test Pit 16/1; lower: types of fired clay tokens (upper by K. Harding; lower by L. Woodard) — 334 Figure 5.32: Map of Ukrainian para-Neolithic sites with graphite-tempered pottery: Symbols: I – present-day industrial source of graphite; II – para-Neolithic site; III – LBK site of KamianeZavallia; IV – Trypillia culture site of Nebelivka; Buh-Dnister culture: 1–Tătărăuca Nouă XV, 2– Soroka I (level 1a), 3–Soroka V, 4–Pechera I, 5–Samchyntsi I, 6–Samchyntsi II, 7–Shymanovske II, 8–Bazkiv Ostriv, 9–Shumyliv-Cherniatka, 10–Hayvoron-Polizhok, 11–Zavallia, 12–Zhakchyk, 13–Melnychna Krucha, 14–Dobrianka 3, 15–Mykolyna Broiaka, 16–Kompaniiska Skelia, 17–Hrushivskyi Ostriv, 18–Semenivka, 19–Ustia Korabelnoi, 20–Puhach 1, 21–Puhach 2, 22–Klepana Balka, 23–Tashlyk 2, 24–Tashlyk 3, 25–Gard, 26–Gard 3, 27–Gard 4, 28–Lidyna Balka, 29–Novorozanivka; Kyiv-Cherkasy culture: 30–Buzky I, 31–Lysychyi Horb, 32–Uspenka 2; Surskyi culture: 33–Strilcha Skelia, 34–Kizlevyi V, 35–Vovchok; Azov-Dnipro culture: 36– Mykilske 2; Surskyi or Azov-Dnipro culture: 37–Kamiana Mohyla 1 (by D. Gaskevych) — 339

564 

 List of Figures

Figure 5.33: Map of East Balkan graphite painted ware complex: M: Maliq; N: Nebelivka (by B. Gaydarska) — 342 Figure 5.34: Graphite painted analogies for the Nebelivka internally thickened rim dish (1): (2) Pietrele (after Hansen et al. 2007); (3) Tangâru (after Voinea 2005); and (4) tell Gumelniţa (after Dumitrescu 1925) (by T. Ignat & C. Lazăr) — 344 Figure 5.35: Gas chromatograms of lipid extracts from Nebelivka miniature vessels: (1) Test Pit 2012/3; (2) Mega-structure MP 16; and (3) Mega-structure MP 29 (by O. Craig) — 348 Figure 5.36: Types of miniature vessel from outside the Mega-structure cluster: dishes – (1) Test Pit 16/1; (2) Test Pit 31/2 (by K. Harding); flasks – (3) Test Pit 1/4; (4) 33/1; (5) Test Pit 1/4; and (6) 1/1 (by B. Gaydarska) — 351 Figure 5.37: Lithics: Mega-structure: 5–7, 9–10, 14–15, 17; Pit, Sondazh 1: 2–4, 12–13; Test Pits: 1, 8, 11, 16 (by M. Gurova) — 355 Figure 5.38: Lithics: Pit, Sondazh 1: 1–2, 4–5; Test Pits: 3, 7; Fieldwalking: 6; scale 1:1 (by L. Woodard) — 356 Figure 5.39: Lithics: 2014 season: 1–12 (by D. Kiosak); 2009 season: 13–15 (by L. Woodard) — 357 Figure 5.40: Lithics: 2014 season: 1–11 (by D. Kiosak); intra-site gridded fieldwalking: 12–16; House A9: 17–18 (by L. Woodard) — 359 Figure 5.41: Worked bone tools: (1) tooth bead-pendant, Pit, S. 1 SF 47, with close-ups (b) and (c); (2) tooth bead-pendant, Mega-structure TsT 1461, Context 64; (3) bone imitation of tooth pendant, Mega-structure TsT 1827, Context 142 (by K. Harding based on photographs by Zs. Tóth) — 371 Figure 5.42: Worked bone tools: (1) fishbone point, Pit, S. 1, SF 48, with close-ups in (b) and (c); (2) broken bone awl with copper staining, Mega-structure, Grid Square D10 (by K. Harding based on photographs by Zs. Tóth) — 372 Figure 5.43: Worked bone tools: (1) ad hoc bone point, Test Pit 19/2, Context 3, SF 6, with close-up in (b); (2) red deer antler hoe model, Pit, S. 1 SF 71; (3) possible bone tool, Mega-structure Grid Square E6 (by K. Harding based on photographs by Zs. Tóth) — 373 Figure 5.44: Special Finds: (1) fragment of house model, Pit, Sondazh 1, SF 5727; (2)–(3) two fragments of (?) fired clay gaming board, unstratified; (4) fired clay ring, House A9; and (5) gold hair ornament, SF 1181, Mega-structure (by K. Harding) — 379 Figure 5.45: Special Finds distribution, megasite (by M. Nebbia) — 383 Figure 5.46: Distribution of figurines, megasite (by M. Nebbia) — 384 Figure 5.47: Distribution of lithics, megasite (by M. Nebbia) — 385 Figure 5.48: Special Finds distribution: (upper) figurines and tokens; (lower) lithics and Other finds, Mega-structure (by M. Nebbia) — 386 Figure 5.49: Upper: overall taxonomic distribution of faunal remains (NISP); lower: distribution of Log Standard Index (LSI) values for measurements on cattle bones (by D. Orton) — 395 Figure 5.50: Upper: Trypillia and Cucuteni sites with raw NISP available and used for comparison here. 1. Berezivka, 2. Bilshivtsy, 3. Cucuteni, 4. Draguşeni, 5. Feteşti, 6. Ghelăieşti, 7. Majdanetske, 8. Hoiseşti, 9. Ignatenkova Gora, 10. Konovka, 11. Kosenivka, 12. Liveni, 13. Grebenyukov Yar, 14. Mitoc, 15. Poduri-Dealul Ghindaru, 16. Santana de Mureş B, 17. SarataMonteoru, 18. Sverdlikove, 19. Taljanki, 20. Târpeşti, 21. Truşeşti, 22. Valea Lupului, 23. Vasylivka, 24. Velika Slobidka, 25. Vesely Kut, 26. Zhvanets-Shovb, 27. Zhvanets, 28. Nebelivka; lower: main taxa identified at Nebelivka by excavation area (%NISP) (by D. Orton) — 399 Figure 5.51: Contributions of wild versus domestic taxa (1, 3 & 5) and breakdown of the main domesticates (2, 4 & 6) for Trypillia and Cucuteni sites in the Early (1–2), Middle (3–4), and Late (5–6) Phases as defined in the text (by D. Orton) — 400 Figure 5.52: (1) comparison of frequencies of major taxa between (a) analysts for the Mega-structure, and (b) areas recorded by Sekerskaya (%NISP); (2) comparison of frequencies of major taxa in areas studied by Sekerskaya (%NISP); (3) findspots of bones assigned to different phases



List of Figures 

 565

within the Mega-structure. NB. Each dot shows a total station record that can represent a single or multiple bone fragments (by D. Orton) — 403 Figure 5.53: (1) findspots of burnt bone within the Mega-structure. As above, single dots may represent multiple specimens; (2) spatial distribution of major taxa within the Mega-structure; (3) foetal/neonatal bones recovered from the living floor and destruction phases of the Megastructure. Diameter of markers is proportional to number of specimens. Length of burnt part of Mega-structure – 36m. (by D. Orton) — 406 Figure 5.54: Isotopic collagen values, Nebelivka mammals (by A. Millard) — 408 Figure 6.1: Population estimates for the Majdanetske megasite. Horizontal bars show range of population estimates by author(s) (by J. Chapman) — 433 Figure 6.2: The Distributed Governance Model (by C. Unwin) — 436 Figure 6.3: The Assembly Model (by C. Unwin) — 438 Figure 6.4: The Pilgrimage Model (by C. Unwin) — 440 Figure 6.5: Distribution maps of (a) Forest Neolithic and Trypillia (b) Phase A; (c) Phase BI; and (d) BII (by M. Nebbia) — 446 Figure 6.6: Lengyel enclosures and Rondels, South-West Hungary (after G. Bertok & Cs. Gáti 2014, Fig. II.89) — 451 Figure 6.7: Magnetometer plan of Taljanki megasite (by R. Ohlrau in K. Rassmann et al. 2014, Fig. 9a) — 461 Figure 6.8: Magnetometer plan of Majdanetske megasite with Ohlrau’s Quarter boundaries; numbers refer to Ohlrau Quarters; A – inset to be found in Fig. 6.9 (by L. Woodard, based on Müller & Videiko 2016, Fig. 2) — 463 Figure 6.9: Detail of Northern part of magnetometer plan, Majdanetske megasite (Area A in Fig. 6.8). Key – large numbers refer to Ohlrau’s Quarters; small numbers refer to individual paths, house circuits and radial streets (by L. Woodard, based on Ohlrau 2015, Abb. 44A) — 464 Figure 6.10: Building sequence of Northern part of Majdanetske magnetometer plan (as shown as Box A in Fig. 6.9) (by L. Woodard) — 465 Figure 6.11: Geophysical plan of Grebeni site. Key: 1: excavated buildings; 2: test pits; 3: geomagnetic anomalies (by B. Gaydarska, based on Koshelev 2004, Ris. 4.14) — 484 Figure 6.12: Densities of people and structures per ha, urban and non-urban sites of low- and highdensity (upper by L. Woodard; lower by J. Chapman) — 504

566 

 List of Tables

List of Tables Table 2.1: Summary of Project activities — 46 Table 3.1: Main periods and archaeological cultures found during the field survey (by M. Nebbia) — 89 Table 3.2: List of all the fields of the attribute table of the Trypillia database (by M. Nebbia) — 97 Table 4.1: Numbers of all structures (partial/complete houses and Assembly Houses) within each Quarter and in the circuits and ‘spaces’ at Nebelivka (OC – outer circuit, IC – inner circuit) (by D. Hale) — 129 Table 4.2: Summary of all structures (partial/complete houses and Assembly Houses) in the circuits and spaces at Nebelivka (OC – outer circuit, IC – inner circuit) (by D. Hale) — 129 Table 4.3: Assembly Houses at Nebelivka (OC – outer circuit, IC – inner circuit) (by D. Hale) — 137 Table 4.4: GINI Co-efficients for House Size by Quarters (by M. Nebbia) — 153 Table 4.5: Explanation of 10 forms of VGA analysis (see Fig.4.19) used at Nebelivka (by B. Buchanan) — 165 Table 4.6: Average VGA measurements of the entirety of structural evidence (by B. Buchanan) — 171 Table 4.7: Number of houses in each Quarter by Model and Stage (by B. Buchanan) — 172 Table 4.8: Average VGA measurements of Models A and B (by B. Buchanan) — 173 Table 4.9: VGA Mean Depth analyses of all Stages of both Models, Quarter F, Nebelivka (see Figure 4.22) (by B. Buchanan) — 176 Table 4.10: VGA Integration-TEK analyses of all Stages of both Models, Quarter F, Nebelivka (see Fig. 4.23) (by B. Buchanan) — 177 Table 4.11: VGA Mean Depth analyses of all Stages of both Models, Quarter L, Nebelivka (see Figure 4.24) (by B. Buchanan) — 177 Table 4.12: VGA Integration-TEK analyses of all Stages of both Models, Quarter L, Nebelivka (see Fig. 4.25) (by B. Buchanan) — 181 Table 4.13: Estimates for house-building of (a) Nebelivka experimental houses and (b) full-sized Trypillia houses, Majdanetske (by S. Johnston) — 191 Table 4.14: The barrow stratigraphy (by S. Ivanova) — 213 Table 4.15: Stratigraphic division, Pit, Sondazh 1 (by J. Chapman) — 230 Table 5.1: Correlation of the numbered shape types used in the Ryzhov/Ovchinnikov and Nebelivka systems (see Fig. 5.1) (by B. Gaydarska) — 267 Table 5.2: Distribution of ‘interesting’ sherds by excavation unit (by J. Chapman) — 274 Table 5.3: Types of analyses of the Nebelivka pottery (by J. Chapman) — 280 Table 5.4: Nebelivka pottery fabrics (see Fig. 5.8) (by J. Chapman) — 287 Table 5.5: Location of single motifs and motif combinations (Figs. 5.15–5.18) by zone, megasite (by J. Chapman) — 305 Table 5.6: Image fragment count by excavation unit (studied images in BOLD) (by J. Chapman) — 327 Table 5.7: Frequency of token types, Nebelivka (see Fig. 5.31 lower) — 337 Table 5.8: Results of the Carbon isotopic analyses of graphite (by A. Boyce) — 346 Table 5.9: Blanks and Tools, 2012 assemblage (by D. Kiosak) — 358 Table 5.10: Blanks and Tools, 2013 assemblage (italics – deposited in Test Pits: all others from Pit, Sondazh 1) (by D. Kiosak) — 362 Table 5.11: Blanks and Tools, 2014 assemblage (by D. Kiosak) — 364 Table 5.12: Counts of diagnostic and non-diagnostic bones by excavation area/feature and analyst (by D. Orton) — 389 Table 5.13: Identification rates by excavation area and analyst. Numbers represent proportion of bones identified to taxon, out of 1. NB This excludes wet-sieved material (by D. Orton) — 392 Table 5.14: Taxonomic frequencies by excavation area (NISP) (by D. Orton) — 396



List of Tables 

 567

Table 5.15: Comparison of taxonomic frequencies between areas and analysts (NISP) (by D. Orton) — 402 Table 5.16: Frequencies of diagnostic and non-diagnostic bone fragments assigned to each phase of the Mega-structure (by D. Orton) — 405 Table 5.17: Comparison of burning rates within the Mega-structure and with other areas (by D. Orton) — 405 Table 6.1: Types of planning and architectural variability, Nebelivka megasite (by J. Chapman) — 428 Table 6.2: The tipping-point for the understanding of the Nebelivka megasite (by J. Chapman) — 434 Table 6.3: Model comparisons (by J. Chapman & B. Gaydarska) — 441 Table 6.4: Early Trypillia settlement plans (by J. Chapman) — 447 Table 6.5: Estimates of salt requirements for various population estimates for Trypillia megasites (by J. Chapman & B. Gaydarska) — 471 Table 6.6: Estimates for the collection, storage and use of building timber, reeds, chaff, hazel withies and fuel for burning at megasites (based on Ohlrau et al. 2016 for timber and S. Johnston for remaining materials: https://doi.org/10.5284/1047599 Section 6.5.2) — 474 Table 6.7: Modelling of the number of coeval houses at megasites (by A. Millard) — 475 Table 6.8: Constituent parts of the ‘urban’ category in the Trypillia context — 486

Index A Abu Hureyra 21 Adâncata - Dealul Lipovanului 491 adaptive cluster sampling. See also sampling strategy aerial archaeology 9 photography 62 African frontier model 450 age-depth model 49, 52, 114, 117 agency 456 agenda for Trypillia megasite studies 263 aggregation 25, 34, 39, 270, 437, 481 'Ain Ghazal 21 Albert, B. 16, 52, 264 alliance(s) 439, 449, 450, 455, 457, 476 AMS 34 AMS dates 52, 246, 423, 426 Mega-structure 195 pit in Sondazh 1 233 Anderson, B. 36 Angkor 501 animal bones 241, 388, 489 animal ornament symbolism bone tool 375 Anomalous Great Sites 22 Anselin's Local Moran's I 51, 105 Apolianka 10, 15, 47, 73 arable 480 Grebeni 483 Majdanetske 469 arable production 454 Arbeiter, S. 299 assemblies. See also aggregation assembly 34, 456 Assembly House 11, 30, 52, 53, 77, 123, 133, 146, 151, 163, 171, 178, 194, 199, 214, 271, 418, 424, 427, 431, 491, 492 absence at Taljanki 460 Majdanetske 462, 466 Moundville 501 Assembly Model 34, 35, 49, 57, 139, 171, 437, 439, 443, 493, 511, 512, 528 Majdanetske 477 Awat'ovi 32 B Baia 491 barrow 502. See also kurgan Bayesian modelling 52, 246 Bernishivka 367 bone tools 376

Big Data 22 Big Other 42, 44. See also Trypillia Big Other binocular vessel 276 biography fired clay bin 208 birthing-hut 417 board game 412 Bodaki flint workshop 368 bone imitation of a red deer canine bead 371 Bourdieu, P. 38 Bradley, R. 42 bricolage 37, 428, 453, 468, 509, 516 buffering 39, 445 building industry 476, 481 building materials 53 analyses 256 building remains 54 built environment. See also space and place burial Trypillia-Cucuteni 502 Mega-structure. See also taphonomy burnt house 42, 52, 128, 155, 181, 214, 248, 513 Burnt House Horizon 514 burnt seeds 248 C Cahokia 500, 501 calcareous deposit pottery 271, 273 calibration curve 34 cambisol 470 carbon isotope analysis of graphite 345 Carrying capacity 106 Cartwright & Runhardt 485 Caswell, E. 273 Çatalhöyük 21, 32 ceremonies 492 cesium magnetometer 56 cesium magnetometrer. See also geophysical investigations chaff 260, 473 châine opératoire 328, 431, 452, 475, 477, 490 channels 242 Chapli. See also gold spiral ornament characterization 58, 485 characterization studies 528 charcoal 231, 411 Cherkassy 86, 93, 96, 99 chernozem 27, 92, 106, 112, 119, 209, 211, 214, 217, 231, 234, 240, 241, 404, 449, 470, 480 Childe, G. 21, 22, 25, 36, 44

Index  chipped stone 352 chronological modelling 426 chronology 421 cities before the state 498 clan 31, 437, 443, 449, 454, 496, 511 classification of building materials 257 climate 112 cluster 103 clustering 39, 94, 445 coeval structures 255 cognitive congruence model. See also Rapoport coil-building 276 Co Loa 500, 501 communal cooking feature. See also kilns communal food preparation area 243 communal open space 32 communal production 367 comparative method pottery 267 compartmentalisation 276 complex use of space 488 construction burning 55, 194 construction daub 55 container revolution 44 controlled firing. See also construction burning copper 8, 37, 431, 449, 454, 455, 457, 494 coppicing 473, 491 core 363 CORONA 62, 69, 92 CORONA satellite 50, 59 Council 443, 496 counters 326, 336, 419. See also tokens courtyard 211 cropmarks 62, 67 Cucuteni Archaeological Park experiments 193 Cucuteni-Cetăţuia 2 Cucuteni C pottery 316 cultivated plants Mega-structure 410 cultural steppe 121 D death assemblage 293, 315, 419, 420, 431, 489, 515 decorated sherds 282 decorative motif combinations 304 decorative motif linkage 304, 423 decorative motifs 452 distribution & placement 299 decorative style vs. vessel form 299 de facto refuse 53 de-husking. See also chaff deliberate fragmentation 279 deliberate house burning 193, 194 deposition 413, 417, 432, 489, 515

 569

animal bones 388 figurines 331 pottery 266 Dergachev, V. 25 destruction daub 55, 221, 226, 242 Diachenko, A. 28, 93, 456 Distributed Governance Model 34, 35, 49, 57, 171, 427, 437, 439, 442, 493, 511, 528 Majdanetske 479, 482 Taljanki 479 ditch 120, 233. See also perimeter ditch animal bones 391 Dobrovodi 15, 16, 47, 62, 130, 145, 155 lithics 367 Dolnoslav figurines 331 down-the-line exchange 494 Drutsi I 26 dry-sieving 195, 352 Durankulak 455 Dzarylgaz Survey Project 75 E East side decorative motif linkage 308 egalitarian 495 Ellis, L. 9, 15, 50, 51, 276, 453 empty space. See also inner open space enchainment 42, 279 Encyclopaedia of Trypillia Civilization 88, 94 entangled 452 entrances 133 Entropy 165, 174 episodes 233, 273, 420 MINV 295 episodic discard 417, 515 excavation 53, 56 excavation of the experimental burnt house remains. See also experimental programme exchange 8, 15, 37, 39, 343, 381, 431, 492 Grebeni 484 exchange networks 456, 457 exogamous marriage 31, 39 experimental programme 13, 17, 53, 55, 181, 199, 418, 513, 528 external residue 53 F fabrics figurines 327 faunal remains 53 feasting 41, 454, 455, 492, 493 fieldwalking 50, 75, 511, 527 figurine 44, 454 figurine condition 333

570 

 Index

figurines 5, 42, 239, 240, 326, 327, 387, 413, 421, 490, 514 realistic 490 types 328 fired clay ball 380 fired clay bin 208, 209 fired clay ring 380 firing temperature analysis Mega-structure 261 first methodological revolution 122 fishbone point 372 Fleming, A. 12 Fletcher, R. 12, 15, 16, 20, 22, 41, 58, 105, 483, 499 flint 494 flint-knappers 491 flint projectile points. See also personhood flotation 52, 56, 195, 352, 366, 409, 410, 411, 413 animal bones 391 fluxgate gradiometry 47. See also geophysical investigation foetal/neonatal bones 420 Mega-structure 407 Forest Neolithic. See also para-Neolithic forest-steppe 51, 112, 264, 444, 472 formal random sampling. See also sampling strategy fragmentation 42, 412, 417, 420, 515 figurines 331 fuel for house-burning 186, 192 function of building materials 257 G Gamble, C. 44 gaming board 378. See also tokens: counters garden feature 146 Gaskevych, D. 340 Gaydarska, B. 10, 13, 17, 42, 58 gender figurines 329 generalist 491 geodatabase 65 geology 112 geophysical investigation 9, 10, 122 geophysical prospection 263. See also geophysical investigation Gimbutas, M. 5, 25, 26 GINI coefficient 49, 103, 149, 155 GIS 160, 268, 279 pottery in houses 317 Giurgiuleşti 376 Glybochok 247 gold 8, 37 gold spiral ornament 380

GPS 81, 87, 88 grain 260 granite rockhead 137, 195 graphite 414, 494 sources 338 graphite-painted decoration 528 graphite-painted motifs. See also miniature vessels graphite painted ware East Balkans 340 graphite-painting 420 graphite sources 345 graphite temper 340 graphite wash. See also miniature vessels graphitic decoration 56 Greater Angkor. See also Angkor Great Zimbabwe 500 Grebeni 482, 483 Greek amphora 80, 214 green (copper) staining bone tool 373 grey literature 61 ground and polished stonework 370 ground-penetrating radar 56 group identity 490 guardians 493 Guardians 427, 439, 444, 450, 496, 512 H habitus 38, 40, 44, 175, 417, 430, 432, 487, 510 Hahn, H-P. 40 Hale, D. 148, 241 Hall, E. 159 Hanson ,H. 160 hazel withies 260, 473 heterarchy 430, 450, 495, 515 heterogeneity 40, 428, 510 high-precision geophysics. See also geophysical investigation Hillier, B. 160 hinterland 50, 58, 69, 94, 107, 109, 511, 527 horizontally open social order 515 house 30, 42, 48, 50, 52, 53, 423 House A9 236, 238, 257, 270, 271, 277, 420 animal bones 390, 401, 402 decorated sherds 282 decorative motif combinations 304 plant remains 409 pottery 313 pottery decorative motifs 299 pottery distribution 319 pottery fabric & form 288 pottery fabrics 285 pottery form 292 sherds numbers & weight 279

Index  House B17 240, 420 animal bones 390, 401 house-burning 8, 257, 266, 432 Grebeni 484 house circuits 128 Majdanetske 461 house construction 214 house daub graphite wash 343 household 29, 30, 34, 41, 42, 44, 417 household death assemblage 53 house mean breadth analysis 425 house mean breadth scores 425 house model 120, 257, 378 house size 49, 130, 148, 149, 422 human impact 16, 34, 114, 115, 117, 118 Hungarian Mediaeval brick oven 243 huts 149 I identity 421 Igbo village arena 428 imagined community 36, 41 incised fine wares 457 incised sign 378 inclusions pottery 274 Incremental Global Moran's I 51, 105 industrial feature 241, 243, 276, 420, 431, 432 inner circuit 252, 264, 427 Majdanetske 468 inner house circuit 422 inner open area 427 inner radial street 139, 142, 162, 264 Bayesian modelling 253 Majdanetske 468 inner space 127 Integration 164, 175 internal chronology 49 internal conflict 28 internal residue 53, 211 inter-regional networks. See also exchange inter-scalar approach 94 Intra-mega-site gridded collection 76 intra-mural burial 8 Intra-site comparisons 401 animal bones 401 intra-site gridded collection. See also fieldwalking; See also intra-mega-site gridded collection invasion. See also warfare Iron Curtain 23 J Johnston, S. 473

 571

K Kaniv group 56, 326 pottery 314 pottery fabrics 316 vessel form vs. decoration 324 vessel shapes vs. fabrics 316 Karbuna 8, 37 keeping-while-giving 455 Kernel Density Estimation 462 Khareneh IV 21 Khvoika 9, 15. See also Khvojka Khvojka 2 kiln 48, 241, 243, 276, 481, 491, 497 animal bones 391 Majdanetske 460 Taljanki 460 kilns 145, 146 kink 130, 496 Kirovograd 86 Kodzhadermen-Gumelniţa-Karanovo VI group 5 Kolesnikov, A. G. 29 Kolomiishchina 9 Kolomiishchina II figurines 333 Kopytoff, I. 450 Korobchino flint source 352, 367 Korvin-Piotrovskiy, A. G. 55 Kruts, V. 16, 25, 26, 39, 473 Kuhn, T. 8, 433 Kuna, M. 53 kurgan 212. See also barrow L Lacan, J. 37 land erosion 117, 118 landscape 51, 61, 121, 186, 264, 418 Majdanetske 470 lateral cycling 456 lateral recycling 494 Lengyel group 8 Lichardus, J. 5 Limited Interest Groups 31, 430, 453, 454, 455, 481, 491, 496, 497 Majdanetske Taljanki 467 lineage 31, 450, 468, 496 linearity neighbourhood 421 linear pit 145 LISA (Local Indicators of Spatial Association) test 487 lithic assemblage 2009 353 lithic assemblage 2012 354 lithic assemblage 2013 360 lithic assemblage 2014 363

572 

 Index

lithic raw materials 352 lithics 454 live animals 493 living assemblage 417 long bone shaft point 373 Loss on Ignition analysis 52 Loveday, R. 57 low-density urbanism 22, 482, 499 house-based planning 500 major building projects 500 short time after domestication 501 trend to lower densities 501 Trypillia 509 weak mortuary zone 500 Luka-Vrublevetskaja 376 lunate 365, 387, 421 Lysaya Gora 37 M macro-region 69, 86 maintenance activities 419 Majdanetske 8, 9, 10, 15, 16, 24, 27, 28, 29, 30, 32, 34, 47, 49, 54, 56, 62, 73, 106, 119, 123, 130, 145, 146, 155, 190, 243, 255, 263, 276, 313, 314, 315, 316, 317, 324, 326, 331, 334, 335, 368, 418, 422, 433, 455, 457, 469, 498 building and burning resources 472 daily resources 469 figurines 333 kilns 431 lithics 368 MINV 315 open vs. closed forms 317 painted signs 324, 326 plan 496 planning 459 re-fitting of figurine parts 335 sherd numbers & weights 315 sledge model 378 spindle-whorl 380 tokens 337 vessel forms 317 Makkay, J. 336 manganese 274, 431, 453, 494 manuring scatters 107 Manzura, I. 93 Marroquíes Bajos 21 mat-impression 276 maximalist 32, 57, 114, 119, 433, 482, 502, 510 lithics 369 mean breadth score 154 Mean Depth 164, 175 mean house size Majdanetske 466 mean sherd weight 279

mega-cluster 99, 109 megasite 8, 15, 20, 21, 22, 24, 27, 31, 41, 49, 56, 59, 73, 190, 212, 304, 367, 381, 382, 414, 422, 424, 430, 443 megasite demise 455 megasite identity 423, 430, 437, 444, 476, 495 megasites 445. See also Anomalous Big sites Mega-structure 13, 14, 52, 53, 56, 57, 137, 257, 263, 270, 271, 277, 279, 413, 419, 420, 431, 492 animal bones 388, 390, 401 decorative motif combinations 304 distribution of vessels 297 figurine deposition 333 lithics 354 miniature vessels 338, 350 open vs. closed forms 292 plant remains 409 pottery 312 pottery distribution 319 sherd numbers & weight 279 tokens 336 memory mound 221, 419, 421, 432, 452 metapodial awl 373 microcharcoal counting 52 micro-region 50 middle space 127 migration 25, 26 Milesi García, L. 34 Milisauskas, S. 5 Millard, A. 475 miniature antler pick 374 miniature vessels 56, 310, 338, 350, 421, 425, 491 organic residue analysis 346 minimalist 32, 114, 118, 472, 481, 510 Minimum Number of Vessels 293 Minoans 20 mobility 40 mode of production 382, 452, 454, 497 bone tool 374 lithics 353, 366, 368 Mogylna 527 molluscan assemblage 52, 119, 214 molluscan evidence 121 Monah, D. 12, 24 Mont Beuvray system 268 monumental. See also monumentality monumentality 24 Moran’s I index 105 mortuary zone 5 Moshurov 2 and 3 476 motif linkage 308, 425 Müller, J. 17, 497

Index  N Nandris, J. 5 Nebbia, M. 13 Nebelivka P1 core 57, 111, 112, 114, 115 Nebelivka concentration of skilled labour and management 490 Nebelivka population 488 Nebelivka size 487 Nebelivka Square 423 Nebelivka territory 487 neighbourhood 11, 30, 49, 77, 123, 130, 133, 139, 142, 153, 228, 421, 455. See also : sherd numbers & weight Bayesian modelling 253 decorative motif linkage 306 Grebeni 483 Majdanetske 462 new research agenda Trypillia megasite studies 526 non-pollen palynomorph analysis 52 North side decorative motif linkage 306 Novi Ruşeşti 37 nucleation 109. See also aggregation; See also clustering Number of Identified Specimens animal bones 393 O Ohrlau, R. 49, 473 Okopy bone tools 376 Olexandrivka 491 Ol’šanka Kurgan 3 376 one-storey house 186 Onopriivka 52, 527 open central space 40 open vessels 420, 425 open vs. closed forms pottery 292 oppida 500 Optical microscopic study bone tools 370 organic residue analysis miniature vessels 346 Orphanidis' theory of repetition 44 Orton, D. 57, 454 outer circuit 252, 264, 427 Outer Radial Streets Majdanetske 469 outer space 127 Ovcharovo 503 Ovchinnikov, E. 268, 317 overall duration Bayesian modelling 253

 573

Ozheve-Ostriv 88 P painted motifs 452 painted pottery 514 painted signs 314 Majdanetske 324 palaeo-channel 72, 92, 139, 146, 148 palaeo-channel network. See also palaeohydrological network palaeo-environmental investigation 527 palaeo-hydrological network 67 palisade 236 Panskoe I 76 para-Neolithic 338 parkland. See also Steppenheide hypothesis particle size analysis 52 Pashkevitch, G. 119 Pashkevych, G. 57, 409. See also Pashkevich, G. Passek, T. 94 Perdigões 21 Perehonivka 69, 71, 90, 558 performances 417 perimeter ditch 48, 127, 130, 133, 234, 487 personhood 42, 490, 496 Special Finds 387 Petreni 2, 9 photogrammetry 56 Pietrele graphite painted ware 342 Pilgrimage Model 34, 35, 49, 57, 422, 424, 439, 444, 493, 511, 512, 528 Majdanetske Taljanki 478 pintadera 336 pit 52, 53 near House B17 240 near House B18 241 near industrial feature 243 pit group 145 pithos clay materials 261 pit in Sondazh 1 228, 270, 273, 279, 420 animal bones 390, 402 decorated shards 282 decoration vs. vessel form 299 miniature vessels 350 pottery 313 pottery decorative motifs 299 pottery fabric & form 288 sherd numbers & wieght 279 pits 142, 145 place-value 40, 419, 450 plan elements 428, 449, 492, 509 plant remains 409

574 

 Index

plan variability 489 platform 184, 186, 205, 211, 239, 240, 258, 260, 261, 419 daub 221, 226 Platform 205, 208, 209 ploshchadka 9, 17, 33, 42, 50, 182, 185, 193, 199, 209, 211, 240, 256, 271, 432, 481, 513 podium 205, 209, 258, 261 Poduri 494 point pattern analysis 103 polished stone tool re-sharpening 354 political economy 28 political organization 103 Poljanitsa 503 Pollock, S. 498 polypod vessel test pit 378 population pressure 25 portrait heads 42 postholes 239 potparts 285 potters 453 pottery 44, 56, 266 pottery fabrics 285, 287, 310 pottery fabrics vs. form 288 primary refuse 53, 319, 413, 514 Primate Pattern 101 procedures 485, 487 projectile points 358, 361, 421 proxemics 159 Prut-Dniester flint. See also Volhynian flint public building 211 pueblo 32 Q Quarter 31, 49, 123, 139, 140, 146, 149, 158, 162, 163, 170, 178, 228, 423, 443 Bayesian modelling 252 decorative motif linkage 305, 308 Majdanetske 468 miniature vessels 350 Norte Chico Peru 501 open vs. closed forms 292 pottery 312 pottery decorative motifs 304 pottery fabrics 285, 287 Swahili cities 501 R radial streets Majdanetske 461 radiocarbon calibration curve 50, 527 radiocarbon dates House A9. See also AMS dates

Rapoport, A. 159 razvedki. See also fieldwalking red deer canine bone tool 375 red deer tooth pendant 8 redistribution 494 reeds 120 relational 23, 430 relational approach 58, 485, 497 relational framework. See also relational approach remote sensing 51, 61, 511 representation 485, 487 residential density 41 resource crisis 26 rhomboid point 366, 387, 421 rock crystal lithic raw material 360 Roe, J. 73, 76, 228 Rollefson, G. 21 Roman corn-drying ovens 243 Rondel 449 Ryzhov, S. 10, 94, 247, 266, 268, 317, 350 system 288, 314 S safety valves in social order 516 salt 453, 471, 472, 481, 494 East Carpathian sources 471 North Pontic sources 471 salt exchange. See also exchange sampling AMS dating 248 sampling strategy 50, 80 samples for AMS dating 49 scalar stress 27, 41, 424, 516 Schiffer, M. 53 Schmandt-Besserat, D. 336 seasonality 40, 502 secondary burning pottery 271 secondary refuse 53, 211 second methodological revolution 10, 122 second phase of the methodological revolution 93 sedimentological analysis 52 settlement 5, 39 settlement-giants 25, 28 settlement hierarchy. See also settlement settlement legacy 25 settlement pattern 93 shadowmarks 62 shed 146, 418 sheds 149 sherd numbers 279

Index  sherd refits 277 Shevchenko, N. 54, 274 Shukurov, A. 26, 106, 445 Silicon Valley 20 Silişte - Prohozeşti 494 simulations. See also AMS dates site clustering 103, 110 site 'Passport' 95 site planning 36, 41 site scatter. See also fieldwalking site size hierarchies 99, 101, 103 Size hierarchies 94 Skelya 37 sledge 472 sledge model 378 Smith, M. open spaces 428 social complexity 28 social conflict 26 social evolution 20, 23 social structure 25, 28 soil coring 48 soilmarks 62, 67, 69 soil micromorphological analysis 53, 54, 419 soil micromorphological investigation 217, 231, 240 Sonderbauen. See also Assembly House South-eastern wall building materials 260 Southern Bug - Dniester Interfluve 24, 26, 51, 93, 99, 109, 452, 527 painted signs 324 South side decorative motif linkage 308 South wall building materials 260 Space and place 158 space syntax theory 160 spatial autocorrelation 105 special finds 56, 326 specialised production 431 specialist 490 spindle-whorl 380 Spondylus 8, 37 square 142, 422, 496 Squares 145 Sredni Stog 25 status competition 41 Steppenheide hypothesis 121 storage structure 146 storage vessels 293 storeroom 418 storey house one- 13, 34, 221, 422, 514 two- 13, 24, 34, 56, 221, 422, 514

 575

strainer pit in Sondazh 1 378 Stratigraphic Units pit in Sondazh 1 273 subsistence 57 Sultana-Malu Rosu graphite painted ware 342 Synukha 86 T Taljanki 9, 10, 15, 16, 26, 39, 47, 49, 56, 62, 73, 106, 123, 130, 145, 146, 155, 241, 243, 255, 263, 276, 314, 316, 317, 326, 334, 335, 368, 418, 454, 455, 457, 498 bone tools 376 building and burning resources 472 figurines 333 gaming board 378 kilns 431 lithics 368 open cs. closed forms 317 painted signs 324, 326 planning 459 sledge model 378 spindle-whorl 380 vessel forms 317 Talnoe 2 and 3 476 taphonomy 181, 214, 239, 257, 263, 270, 274, 431 animal bones 393 taxonomic frequencies animal bones 394 Taylor, T. 31, 430, 456 technical - typological characteristics 257 tertiary refuse 53 test pits 13, 52, 53, 56, 119, 214, 264, 270, 423 animal bones 390 decorative motif linkage 304 MINV 293 placement in house 270 pottery 311 threshold 205, 240 Mega-structure 199 'Time of the Their Lives' Project 514 tipping-point 17, 433, 435 Tkachuk, T. 314 Tobler’s first Law of Geography 105 tokens 336. See also counters tooth pendant 370, 371 Traian. See also gold spiral ornament figurines 335 trans-scalar approach 61 tribe 29 tribute 450 Trypillia Big Other 36, 38, 42, 266, 310, 334,

576 

 Index

413, 430, 432, 445, 450, 452, 456, 457, 490, 495, 496, 509, 515 clay 381 figurines 327 house model 378 houses 421 Tudorovo 1 Kurgan 1 376 Turner, A. 161 two-storey house 186 U Uivar 514 unburnt house 128, 182, 214 Uruk 20, 22, 58, 497, 503 tokens 336 V Valencina de la Concepción 21 Varna 455 Varna I cemetery 37. See also gold spiral ornament Varvareuvka XII 44 Vesely Kut 36, 449, 487, 492, 527 vessel form 288, 310 sizes 295 with zoomorphic terminal 377 VGA 161, 162, 163. See also Visibility Graphic Analysis Videiko, M. 13, 24, 51 vignette adolescent visitor 522 clan leader 519 early house-builder 518 last generation of residents 525 Nebelivka Guardian 516 pilgrim 522 ritual leader 523 trader 524 visitor to the Assembly 520 Vinča 193 Vinnitsa 96 Vinogradnyj 376 Visibility Graphic Analysis 49, 158, 418, 424 vitrification 209 vitrified 239, 243, 261, 262, 271, 273, 513 vitrified daub 55, 193, 221 Volhynian flint 366, 368, 431, 449, 453, 457. See also exchange Volodymyrivka 9, 73, 86, 90, 367, 421 figurines 333 painted signs 326 Voroshilivka

figurines 333 house model 378 W Wadi Jilat 6 21 Wall plaster 240 warfare 491. See also social conflict waste management 492 weeds of cultivation 411 Weiner, A. 455 Wengrow, D. 17, 25, 44, 498 West side decorative motif linkage 306 Wheatley, D. 160 wheel-marks pottery 276 wiggle 250, 264, 421, 427, 511 wild vs. domestic taxa animal bones 398 Wirth, L. 21, 23 withy impressions 184, 239 wood impressions. See also withy impressions woodland management 190, 439, 450, 481, 491 worked bone 370 workshop 146, 149, 418 WorldView-2 imagery 64 satellite 50 Y Yatranivka 9, 73 Yoruba towns 500 Z Žižek, S. 37 zoomorphic 328 zoomorphic vessel pit in Sondazh 1 377