In the Footsteps of the Etruscans: Changing Landscapes around Tuscania from Prehistory to Modernity (British School at Rome Studies) [New ed.] 9781009230025, 9781009230018, 1009230026

Citation preview

In the Footsteps of the Etruscans br i t ish s c hool at rom e st u di e s

Changing Landscapes around Tuscania from Prehistory to Modernity

Graeme Barker and Tom Rasmussen

IN THE FOOTSTEPS OF THE ETRUSCANS In the Footsteps of the Etruscans describes the archaeology of the countryside within a 10 km radius of the small town of Tuscania near Rome, throwing light on the unrecorded lives of the generations of farmers and shepherds who have lived there. What was the character of prehistoric settlement prior to Etruscan urbanization? How did urbanization shape the lives of the ‘ordinary Etruscans’ working the land, hardly ever addressed in Etruscan archaeology? What was the impact on these people of being absorbed into the expanding Roman empire and its globalized economic structures? How did the empire’s collapse and the subsequent emergence of the nucleated Medieval village affect Tuscania’s rural population? The project’s 7500-year ‘archaeological history’, from the first farmers to those grappling with globalization today, contributes eloquently to our understanding of how Mediterranean peoples have constantly shaped their landscape, and been shaped by it. GRAEME BARKER is Disney Professor of Archaeology Emeritus and a Senior Research Fellow at the McDonald Institute for Archaeological Research at the University of Cambridge, and a Professorial Fellow at St John’s College. He has investigated the archaeology of human landscapes in semi-arid (Italy), arid (Libya, Jordan, Iraq) and tropical rainforest (Borneo) environments. He has published over twenty-five books and edited books and over 350 research papers. Holding Fellowships of the British Academy, Royal Geographical Society and Society of Antiquaries, he was awarded the Dan David Prize for the Past Dimension in 2005 and appointed CBE in the Queen’s New Year Honours 2014 for services to archaeology. TOM RASMUSSEN studied Classics at Cambridge where, after a Fellowship at the British Institute of Archaeology at Ankara, he also gained his PhD. He is a Fellow of the Society of Antiquaries and was Head of Art History at Manchester University, where he is now Honorary Research Fellow. He has extensive fieldwork experience in Italy and elsewhere, and has published widely on Etruscan, Greek and Roman archaeology and art. He was area editor for The Macmillan Dictionary of Art (now Grove Dictionary), and other publications include Bucchero Pottery from Southern Etruria (1979, reissued 2006), Looking at Greek Vases (1991, Greek translation 1997, co-edited with Nigel Spivey) and The Etruscans (1998, Italian edition Gli Etruschi 2006, co-authored with Graeme Barker).

BRITISH SCHOOL AT ROME STUDIES Series editors Barbara Borg Chair of Publications of the British School at Rome Rosamond McKitterick Chair of the Faculty of Archaeology, History and Letters and member of the Council of the British School at Rome Abigail Brundin Director of the British School at Rome British School at Rome Studies builds on the prestigious and long-standing Monographs series of the British School at Rome. It publishes volumes on topics that cover the full range of the history, archaeology and art history of the western Mediterranean both by the staff of the BSR and its present and former members, and by members of the academic community engaged in top-quality research in any of these fields. In the Footsteps of the Etruscans: Changing Landscapes around Tuscania from Prehistory to Modernity Graeme Barker and Tom Rasmussen Architecture in Ancient Central Italy: Connections in Etruscan and Early Roman Building Edited by Charlotte R. Potts Roman Port Societies: The Evidence of Inscriptions Edited by Pascal Arnaud and Simon Keay The Basilica of St John Lateran to 1600 Edited by Lex Bosman, Ian Haynes and Paolo Liverani Rome in the Eighth Century: A History in Art John Osborne Rome, Pollution and Propriety: Dirt, Disease and Hygiene in the Eternal City from Antiquity to Modernity Edited by Mark Bradley, with Kenneth Stow Old Saint Peter’s, Rome Edited by Rosamond McKitterick, John Osborne, Carol M. Richardson and Joanna Story The Punic Mediterranean: Identities and Identification from Phoenician Settlement to Roman Rule Edited by Josephine Crawley Quinn and Nicholas C. Vella Turin and the British in the Age of the Grand Tour Edited by Paola Bianchi and Karin Wolfe

In the Fo otsteps of the Etruscans Changing Landscapes around Tuscania from Prehistory to Modernity G r a eme Ba r k er University of Cambridge

T om R a smussen University of Manchester

with contributions by Antony Brown, Clare Ellis, Francesco di Gennaro, Annie Grant, Alison MacDonald, Helen Patterson†, Marco Rendeli†, Tim Reynolds, Edward Rhodes, Jeremy Taylor and Nicoletta Vullo

Shaftesbury Road, Cambridge CB2 8EA, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467 Cambridge University Press is part of Cambridge University Press & Assessment, a department of the University of Cambridge. We share the University’s mission to contribute to society through the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781009230025 DOI: 10.1017/9781009230018 © The British School at Rome 2023 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press & Assessment. First published 2023 A catalogue record for this publication is available from the British Library. Library of Congress Cataloging-in-Publication Data Names: Barker, Graeme, author. | Rasmussen, Tom, author. Title: In the footsteps of the Etruscans : changing landscapes around Tuscania from prehistory to modernity / Graeme Barker, Tom Rasmussen. Description: Cambridge, United Kingdom ; New York, NY : Cambridge University Press, 2023. | Series: British School at Rome studies | Includes bibliographical references and index. Identifiers: LCCN 2022043273 | ISBN 9781009230025 (hardback) | ISBN 9781009230018 (ebook) Subjects: LCSH: Landscape changes – Italy – Etruria – Tuscania – History. | Landscape archaeology – Italy – Etruria – Tuscania – History. | Etruscans. | TuscaniaEtruria (Italy) – History. Classification: LCC GF587.T868 B37 2023 | DDC 304.209456/25–dc23/eng20221121 LC record available at https://lccn.loc.gov/2022043273 ISBN 978-1-009-23002-5 Hardback Cambridge University Press & Assessment has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

In the Footsteps of the Etruscans is dedicated to three scholars of the landscape archaeology and history of Italy who shaped our careers and paved the way for this project: John Ward-Perkins, Anthony Luttrell and Tim Potter; and to our friends and collaborators Marco Rendeli and Helen Patterson, both of whom died tragically early during the final months of bringing the project to publication.

Contents

List of Figures  page xii List of Tables  xvi List of Contributors  xviii Preface and Acknowledgements  xix

1 The Tuscania Archaeological Survey: Rationale, Aims and Objectives  Graeme Barker and Tom Rasmussen 1 Introduction 1 Landscapes and Taskscapes 4 Mediterranean Plough-Zone Archaeology 6 Research Issues 9 Etruscan Urbanization 9 ‘Romanization’ and Roman Imperialism 11 Medieval Settlement and Incastellamento 13 The Selection of Tuscania 14 Tuscania’s Settlement Archaeology and History 14 Tuscania’s Neighbours 21 The Physical Landscape 23 Project Planning and Development 25 Conclusion 26 2 Methodologies  Graeme Barker, Tom Rasmussen, Alison MacDonald, Annie Grant and Nicoletta Vullo 28 Introduction 28 Defining the Study Area 28 Sampling Strategy 30 Site and Off-site/Non-site Archaeology 38 Field-Walking 43 Collecting 46 Recording 49 Classifying the Finds 55 Analysing the Finds: Interpretative Issues 56

vii

contents

Defining ‘Sites’ 56 Settlement Densities 58 Continuity or Discontinuity? 59 Conclusion 59 3 The Natural Landscape and Its Evolution  Tony Brown, Clare Ellis and Edward Rhodes 61 Introduction 61 The Structural Components of the Landscape 61 The Regional Palaeoenvironmental Record 65 The Alluvial Stratigraphy of the Marta Valley 68 Dating and Alluvial Chronologies 75 76 Radiocarbon Dating Palaeomagnetic Dating 76 Luminescence Dating 79 The Regional Alluvial Record: Climate, People or Both? 79 The Upper Marta: Natural or Artificial? 82 Conclusion 83 4 Prehistoric Landscapes  Graeme Barker, Francesco di Gennaro and Tim Reynolds 85 Introduction 85 The Chipped Stone Collections 85 Raw Materials 87 Edge Conditions 87 Typology 88 Technology 92 The Pottery 94 Pre-Neolithic Activity (‘Period 0’) 101 Transitions to Farming, c. 5500–3500 bc (c. 7500–5500 bp) 103 Earlier Neolithic Settlement, c. 5500–4500 bc 104 Later Neolithic Settlement, c. 4500–3500 bc 108 Chalcolithic, c. 3500–2200 bc 111 Bronze Age, c. 2200–950 bc 118 Earlier Bronze Age Settlement, c. 2200–1400 bc 120 Later Bronze Age Settlement, c. 1400–950 bc 124 Early Iron Age, 950–700 bc 126

viii

contents

5 Etruscan Urbanization, c. 700–300 bc  Tom Rasmussen, Marco Rendeli and Graeme Barker 130 Introduction 130 Etruscan Urbanization and Its Landscape Impacts 132 Establishing Chronologies 134 Pottery 134 Tile 140 Settlement Densities 141 Settlement Distributions 144 Tombs and Necropolises 151 Landscape Histories 154 Seventh Century bc 154 Sixth Century bc 158 Fifth Century bc 161 Fourth Century bc (and Later) 163 Landscape Engineering: Road Cuttings and Cuniculi 165 Tuscania’s ‘Suburb’: Site R14:13 167 An Etruscan Farm at Guidocinto (R24:19) 168 Conclusion 173 6 ‘Romanization’: The Roman Republican Period, c. 300–30 bc  Alison MacDonald, Jeremy Taylor and Annie Grant 176 Introduction 176 Methodologies 178 Recording and Classifying the Roman Pottery 178 Defining Roman ‘Sites’ and ‘Non-sites’ 179 Interpreting Densities 181 Settlement Forms 182 Grid Surveys 187 Geophysical Surveys 191 Regional Comparisons 193 Settlement Patterns 195 Settlement Trends 202 Town and Territory 205 The Socio-economic Landscape 208 Conclusion 212

ix

CONTENTS

7 The Roman Imperial and Late Antique Periods, c. 30 bc–c. ad 700  Alison MacDonald and Annie Grant Introduction Rural Settlement in the Early and Mid Empire (c. 30 bc–c. ad 260) Rural Settlement in the Late Empire and Late Antiquity (c. ad 260–c. 700) Site Histories Stability or Disruption in the Countryside? Town, Territory and Trade Changing Socio-economic Landscapes Conclusion

226 229 233 237 241 244

8 Incastellamento and Its Aftermath: Medieval and Modern Landscapes, c. ad 700 to the Present  Helen Patterson, Graeme Barker and Tom Rasmussen Introduction Historical Background Tuscania The Territory Classifying the Post-Classical Survey Materials Dating Generic Medieval (Eighth/Ninth to Fourteenth Centuries) High Medieval (Eleventh to Early Thirteenth Centuries) Late Medieval (Thirteenth to Fourteenth Centuries) Post-Medieval (Fifteenth to Seventeenth Centuries) Modern (Eighteenth Century Onwards) Defining ‘Sites’ Changing Landscapes Generic Medieval (Eighth/Ninth to Fourteenth Centuries) High Medieval (Eleventh to Early Thirteenth Centuries) Late Medieval (Thirteenth to Fourteenth Centuries) Post-Medieval (Fifteenth to Seventeenth Centuries) Modern (Eighteenth Century Onwards) From Incastellamento to Agriturismo

246 246 246 248 250 256 256 259 259 260 260 260 260 263 263 266 268 271 271 274

9 A Mediterranean Landscape from Prehistory to Modernity  Graeme Barker, Tom Rasmussen and Nicoletta Vullo Introduction Tuscania’s Landscape History

278 278 280

x

214 214 216

CONTENTS

Reflections on Methodologies Regional Comparisons The Tuscania Archaeological Survey and the Mediterranean Alluviation Debate The Tuscania Archaeological Survey and Mediterranean Landscape History Appendix I  The Tuscania Archaeological Survey Etruscan Coarse Wares  Tom Rasmussen and Marco Rendeli  296 Appendix II  The Tuscania Archaeological Survey Gazetteer  Tom Rasmussen, Graeme Barker, Alison MacDonald, Annie Grant and Nicoletta Vullo  308 Bibliography  350 Index  376

xi

286 290 292 294

Figures

1.1 Tuscania: the walled town page 1 1.2 Tuscania in its geographical setting in central Italy 2 1.3 The British School at Rome’s South Etruria Survey 6 1.4 Tuscania’s Colle San Pietro acropolis 14 1.5 The historic centre (centro storico) of Tuscania after the 1971 earthquake 15 1.6 Tuscania: plan of Colle San Pietro and the later 16 Medieval town 1.7 Restored tomb interior with sarcophagi from 17 Tuscania, displayed at Pall Mall, London, in 1837 1.8 Etruscan sarcophagi on the walls of Tuscania’s 18 Piazza del Comune 1.9 The Ara del Tufo Etruscan necropolis near Tuscania 19 1.10 A fragment of the Via Clodia Roman road at Tuscania 21 1.11 Looking north across the Tuscania Archaeological Survey area 23 1.12 The gorge of the Marta valley near Tuscania 24 1.13 Recent land use in South Etruria 25 1.14 Tufo stone quarry north of Tuscania 26 1.15 Deep-ploughing technology used in the 27 countryside around Tuscania 2.1 The location of the Tuscania Archaeological Survey area on the Istituto Geografico Militare 30 (IGM) 1:25,000 maps 2 2.2 The 354 km study area defined for the Tuscania Archaeological Survey: topography, watercourses and settlements 31 2.3 Geology of the Tuscania Archaeological Survey area 32 2.4 The three landscape sampling strategies used by the Tuscania Archaeological Survey 35 2.5 The km2 unit identifiers of the Tuscania Archaeological Survey 37 2.6 The composite sample of 97 km2, showing the areas actually walked 39

2.7 Modern rural buildings in the Tuscania Archaeological Survey area 40 2.8 A dense scatter of pottery and tile fragments in the ploughsoil (RS27:10), marking the location of a partly ploughed out Roman-period farm 41 2.9 Twentieth-century archaeology: a longabandoned car near Tuscania 42 2.10 Percentages of land-use types searched by the survey teams 44 2.11 Field-walking in progress during the Tuscania Archaeological Survey 45 2.12 Collecting surface material on a grid at R14:13 47 2.13 Collecting surface material on a grid, and weighing the finds, at T24:8 48 2.14 The effects of erosion on surface archaeology 50 2.15 Giant furrows and ridges produced by modern deep ploughing 51 2.16 The effects of light on the visibility of surface artefacts 52 2.17 Ratios of Etruscan and Roman pottery to land use, and land use to Etruscan and Roman pottery, using presence/absence 53 2.18 Ratios of Roman pottery to land use, and land use to Roman pottery, using sherd frequencies 54 3.1 The Marta catchment, the archaeological survey area, and the region, with sites outside the Marta mentioned in Chapter 3 62 3.2 The geology of the region 63 3.3 Proxy palaeoclimatic records for the area 67 3.4 The archaeological survey area, the Marta catchment and other sites described in Chapter 3 69 3.5 The exposure at the Corvena sluice 71 3.6 Generalized stratigraphies of Marta valley alluvial sites 72 3.7 Site Q1 terrace exposure 73 3.8 Site Q1 terrace stratigraphy and height relations 74

xii

list of Figures

3.9 The excavation into the lower Marta floodplain (site M5) 75 3.10 Palaeomagnetic curves and Marta samples 78 4.1 Etruria, showing the principal prehistoric sites outside the Tuscania Archaeological Survey area mentioned in Chapter 4 86 4.2 Stone tools from the Tuscania Archaeological Survey 89 4.3 Stone tools from the Tuscania Archaeological Survey 90 4.4 Stone tools and utilized pieces from the Tuscania Archaeological Survey 91 4.5 Cores from the Tuscania Archaeological Survey 93 4.6 The Tuscania Archaeological Survey: distribution of Neolithic units 105 4.7 Some Earlier Neolithic pottery from the Tuscania Archaeological Survey 106 4.8 The location of Earlier Neolithic site R24:18 107 4.9 Later Neolithic material from R34:22 109 4.10 Later Neolithic material from T88:11 110 4.11 The hilltop location of Later Neolithic site R18:3 111 4.12 The Tuscania Archaeological Survey: distribution of Copper Age units 114 4.13 Copper Age pottery from unit R34:9 115 4.14 View over part of the RS18 Copper Age habitation zone 117 4.15 Earlier Bronze Age pottery from unit CP94:7 119 4.16 Later Bronze Age pottery from unit R14:13 119 4.17 The Tuscania Archaeological Survey: distribution of Bronze Age units 122 4.18 RS1:10, the location of a likely significant Bronze Age habitation area 123 4.19 The Tuscania Archaeological Survey: distribution of Iron Age units 127 5.1 Etruria and northern Latium, showing the principal regions and sites mentioned in Chapter 5 131 5.2 Bucchero from R14:13 138 5.3 Bucchero from R14:13 139 5.4 Etrusco-Corinthian ware from T85:7 140 5.5 Gridded site R14:13: tile weights 141 5.6 Architectural terracotta from site R54:2 143

5.7 Distribution of Roman and Etruscan pottery at site RS18:13 145 5.8 (a)–(e) Distributions of ‘definite’ and ‘probable’ Etruscan sites by period 146 5.9 Distribution of Etruscan necropolises, tombs and tumuli 152 5.10 Rock-cut tomb at site T54:11 154 5.11 Italogeometric painted sherd from T85:7 155 5.12 Early impressed and incised pottery from T85:7 156 5.13 Bucchero and fine impasto fragments from T85:7 156 5.14 Site T89:3 showing results of grid collection 157 5.15 Attic black-figure fragment (rim of volute krater) from J2:12 159 5.16 Etruscan sites RS22:7 and T54:15 160 5.17 Site T88:4 showing results of grid collection 161 5.18 Terracotta figurine of human head (SF 303) from R14:14 164 5.19 Bronze figurine of Herkle/Herakles from C97:14 164 5.20 Road cutting at J12:6 likely to be of Etruscan or Roman origin 166 5.21 Inscribed bucchero sherd from site R14:13 168 5.22 Attic black-figure cup (kotyle) fragment from R14:13 168 5.23 Attic red-figure cup (tondo) from R14:13 169 5.24 View of Guidocinto* terrace, the location of the excavated Etruscan farm (R24:19) 170 5.25 Plan and photograph of the main Guidocinto (R24:19) excavations 171 5.26 (a) pithos (storage jar), (b) mortar and (c) slab with grinding wear, on the floor of the Guidocinto building (R24:19) 172 5.27 Loomweight from the Guidocinto Etruscan farm (R24:19) 172 5.28 Bucchero from the Guidocinto Etruscan farm (R24:19) 173 5.29 Handle fragment of Attic black-glaze kylix from the Guidocinto Etruscan farm (R24:19) 174 6.1 Roman Tuscania and its Republican regional context 177 6.2 Roman settlement around Tuscania: distribution of ‘site’ and ‘non-site’ units 186 6.3 Distribution of Roman pottery sherds at the gridded unit R14:13 188 *  For the spelling, see p.137.

xiii

LIST OF Figures

6.4 Distribution of Roman pottery sherds at the gridded unit CP84:18 6.5 Distribution of Roman pottery sherds at the gridded unit R44:9 6.6 Geophysical survey results at site T34:21 6.7 Geophysical survey results at site RS1:14 6.8 Early Republican settlement around Tuscania 6.9 Late Republican settlement around Tuscania 6.10 The distribution of Republican-period sites (including tombs and rock chambers) south of Tuscania 6.11 The distribution of Republican-period sites (including tombs and rock chambers) north of Tuscania 6.12 The distribution of principal Republican-period sites outside the areas shown in Figures 6.10 and 6.11 6.13 Stamped terra sigillata or Arretine sherds from site RS30:54 6.14 Fired clay loomweight from site T34:21 6.15 Deviations from the mean for selected Roman Republican ceramic fabric assemblages 7.1 Roman Tuscania and its Imperial regional context 7.2 Early Imperial settlement around Tuscania 7.3 Mid Imperial settlement around Tuscania 7.4 Material culture at R14:13 7.5 Rural settlement around Tuscania in the Early and Mid Empire according to the Forma Italiae survey 7.6 Site 159 identified in the Forma Italiae survey 7.7 Trends in the rural free population of Roman Italy between the second century bc and the first century ad 7.8 Late Imperial settlement around Tuscania 7.9 Late Antique settlement around Tuscania 7.10 Deviations from the mean in Roman Imperial pottery 8.1 Etruria and central Italy, showing the principal regions and sites mentioned in Chapter 8 8.2 The basilica of San Pietro 8.3 The typical promontory location of an Early Medieval castle: San Savino

8.4 Plan of the Early Medieval castle and abbey at San Savino 8.5 San Savino abbey 8.6 The Cistercian monastery of San Giusto 8.7 Examples of High and Late Medieval painted pottery collected by the Tuscania Archaeological Survey 8.8 Copper coin of Innocent X minted around 1650, from RS12:11 8.9 The distribution of Generic Medieval sites located by the Tuscania Archaeological Survey 8.10 RS18:18 Medieval habitation and burial area 8.11 Millstone/quernstone fragment from CP84:18 8.12 The distribution of High Medieval sites located by the Tuscania Archaeological Survey 8.13 The rich High Medieval site C99:6 8.14 R44:10, a High Medieval site: building stone and tile 8.15 R44:10, a High Medieval site: millstone fragment and human bones 8.16 Late Medieval carinated bowl from R14:5 8.17 The distribution of Late Medieval sites located by the Tuscania Archaeological Survey 8.18 The distribution of Post-Medieval sites located by the Tuscania Archaeological Survey 8.19 RS22:15, ruins of a rural building of PostMedieval date 9.1 Tuscania and its regional context, showing locations and sites mentioned in Chapter 9 9.2 GIS maps predicting different probabilities of locating Etruscan sites using the data from the Transect Sample, Random Sample and Judgement Sample 9.3 GIS maps predicting different probabilities of locating Republican Roman sites using the data from the Transect Sample, Random Sample and Judgement Sample A1.1 Etruscan coarseware typology: jars A1.2 Etruscan coarseware typology: jars A1.3 Etruscan coarseware typology: jars A1.4 Etruscan coarseware typology: jars A1.5 Etruscan coarseware typology: jars and bowls A1.6 Etruscan coarseware typology: bowls

189 190 192 192 196 198

199

199

201 202 209 212 215 217 221 222

223 224

225 227 228 240 247 248 252

xiv

253 254 255

259 260 264 265 266 267 268 269 270 271 272 273 275 279

287

289 297 298 300 301 302 304

LIST OF Figures

A1.7 Etruscan coarseware typology: feet/lids and basins A1.8 Etruscan coarseware typology: basins A2.1 Location map accompanying Gazetteer A2.2 Survey units: North Transect A2.3 Survey units: East Transect A2.4 Survey units: South Transect

A2.5 Survey units: West Transect A2.6 Survey units: Random Squares, group 1 A2.7 Survey units: Random Squares, group 2 A2.8 Survey units: Random Squares, group 3 A2.9 Survey units, Judgement Squares J1–J12 A2.10 Survey units, Judgement Squares J13–J20

305 306 309 310 311 312

xv

313 314 315 316 317 318

Tables

5.5 Etruscan bucchero shapes and their percentages 140 5.6 Tile fabric series 142 5.7 Etruscan site and necropolis numbers, and presence of bucchero 144 5.8 Numbers of Etruscan sites by century 155 6.1 The pottery phasing used in the analysis of the Roman pottery from the Tuscania Archaeological Survey 179 6.2 Summary of the Roman pottery assemblage from the Tuscania Archaeological Survey 179 according to ware categories 6.3 Summary of the Roman pottery assemblage from the Tuscania Archaeological Survey 179 according to area of production 6.4 Summary of the Roman pottery assemblage from the Tuscania Archaeological Survey 180 according to chronological groupings 6.5 Hierarchy of Roman sites developed from the materials and unit data collected in the 180 Tuscania Archaeological Survey 6.6 Classification of ‘sites’, ‘probable’ sites and 181 ‘non-sites’ according to size categories 6.7 Classification of ‘sites’, ‘probable’ sites and 184 ‘non-sites’ according to status categories 6.8 Classification of ‘sites’, ‘probable’ sites and ‘non-sites’ according to size and status categories 184 6.9 Gridded sites showing number of grid squares 187 per phase 6.10 Continuity and non-continuity of units 203 (Etruscan to late Republican) 6.11 Density of Roman units at different distances 206 from Tuscania 6.12 ‘Value’ of sherds of selected fabrics per phase 211 7.1 Summary of the classification of the Imperial and Late Antique units according to size categories 218 7.2 Summary of the classification of the Imperial and Late Antique units according to status categories 218

2.1 The periodization established for the Tuscania Archaeological Survey page 56 2.2 Classification system used for defining Etruscan and Roman sites in the Tuscania Archaeological Survey 58 3.1 The section at site P1b on the Fosso Pantacciano 68 3.2 Summary of two adjacent sections (sites C1 and C2) logged on the Fosso Capecchio 70 3.3 The alluvial stratigraphy of the exposure at 70 site M10 3.4 A summary of the exposures at site M3, 72 downstream from the Corvena sluice 3.5 Section log of site Q1, a terrace section exposed 75 in a quarry at Guado della Spina 3.6 Radiocarbon dates from Marta valley alluvial sections 77 3.7 OSL dates from Marta valley alluvial sections 79 4.1 Density of locations with prehistoric pottery 87 4.2 Chipped stone tools recovered by the Tuscania Archaeological Survey 88 4.3 The fabrics of the Tuscania Archaeological 94 Survey prehistoric pottery and their frequencies 4.4 Frequencies of the major and minor pottery 95 fabrics in the principal prehistoric units 4.5 Classification of Tuscania Archaeological 96 Survey units with prehistoric material 4.6 Lithics from Copper Age unit R34:9 115 4.7 Existing and new units with prehistoric pottery, 116 phase by phase 4.8 Lithic assemblages from Bronze Age units J3:38 and J18:2 120 5.1 Etruscan coarseware fabrics: descriptions and percentages 135 5.2 Etruscan coarseware shapes and their percentages 136 5.3 Etruscan coarseware: numbers of sherds per 136 fabric per shape/type 5.4 Etruscan bucchero fabrics: descriptions and percentages 137

xvi

list of Tables

7.3 Summary of the classification of the Imperial and Late Antique units according to size and status categories 7.4 Chronological trends in Tuscania Archaeological Survey site numbers 7.5 Gridded sites showing ‘values’ of sherds per phase 7.6 Grid squares at site R14:13 showing ‘values’ of sherds per phase 7.7 Grid squares at site CP84:18 showing ‘values’ of sherds per phase 7.8 Grid squares at site R44:9 showing ‘values’ of sherds per phase 7.9 Continuity and discontinuity in Roman-period settlement around Tuscania

7.10 Density of sites at different distances from Tuscania 237 8.1 Catalogue of Tuscania Archaeological Survey post-classical pottery 257 8.2 Period and site classification of Tuscania Archaeological Survey units with post-classical material 261 9.1 The pros and cons of field survey 280 9.2 Tuscania Archaeological Survey: trends in site numbers 281 A1.1 Etruscan coarseware shapes/types and their principal findspots 296 A2.1 Guidance notes to the Survey Gazetteer 308 A2.2 Survey Gazetteer 319

220 226 230 231 231 232 234

xvii

CONTRIBUTORS

Graeme Barker: McDonald Institute for Archaeological Research, University of Cambridge, UK

Helen Patterson†: former Assistant Director, British School at Rome, Italy

Antony Brown: Geography and Environmental Science, University of Southampton, UK, and Natural Sciences, University Museum, Arctic University of Norway, Tromsø, Norway

Tom Rasmussen: School of Arts, Languages and Cultures, University of Manchester, UK

Clare Ellis: Argyll Archaeology, Campbeltown, Argyll, UK

Tim Reynolds: Department of History, Classics and Archaeology, Birkbeck College London, UK

Francesco di Gennaro: former Director of the Soprintendenza al Museo Nazionale Preistorico Etnografico ‘Luigi Pigorini’, Italy

Edward Rhodes: Department of Geography, University of Sheffield, UK

Annie Grant: School of Art, Media and American Studies, University of East Anglia, UK Alison MacDonald: Department for Continuing Education, University of Oxford, UK

Marco Rendeli†: Dipartimento di Storia, Scienze e della Formazione, Università degli Studi di Sassari, Italy

Jeremy Taylor: School of Archaeology and Ancient History, University of Leicester, UK Nicoletta Vullo: Director of Account Management, Prometric, Italy

xviii

PREFACE AND ACKNOWLED GEMENTS

The field project that is the subject of this book investigated the archaeology of the countryside within a 10 km radius of the small town of Tuscania some 80 km northwest of Rome. The town is best known in the tourist guides for its two beautiful Early Medieval basilicas, San Pietro and Santa Maria Maggiore, but in fact it has a more-orless unbroken occupation history from pre-Etruscan times to the present day, a time span of some 3000 years, while the archaeological evidence of people living in the surrounding countryside extends Tuscania’s history many thousands of years earlier still. As we describe in Chapter 1, the project was devised to combine several aims, some historical, others methodological, but its overall objective was to contribute to present understanding of the processes that have shaped the development of the modern Mediterranean landscape as a physical and cultural construct. Our particular focus was the changing nature of the relationship between town and countryside, taking Tuscania and its environs as our exemplar. The fieldwork was undertaken between 1986 and 1990. Although a number of papers were published promptly on emerging results for particular periods or approaches (Barker 1988; Barker and Rasmussen 1988, 1998; Barker et al. 1993a, 1993b; Brown and Ellis 1996; MacDonald 1995; Rasmussen 1991; Rendeli 1993a) and much of the text for this monograph was first drafted by 2000, for a variety of academic, bureaucratic and personal reasons – another book in itself! – it has taken another two decades to bring it to completion. The project was therefore planned and executed within an intellectual context in many respects different from some current archaeological interests, and in a very different technological age, but as the subsequent chapters discuss, the questions and our findings remain as pertinent today as then. We were able to collect a quality of field data that would be extremely difficult, indeed impossible, to collect today in the study area, as in many other regions not just of Italy but elsewhere in

the Mediterranean basin, because of changes in land use, landownership and political and administrative structures. We believe that the deep ‘archaeological history’ of the countryside around this small Italian town, for all its historical contingency, contributes not insignificantly to our wider understanding of Mediterranean landscape history. To write such an archaeological history is the work of many hands (and feet in our case in the fieldwork), requires a great deal of logistical organization and depends on the support, commitment and goodwill of many institutions and individuals. We first wish to acknowledge the formal support of the regional office of the state archaeological authorities, the Soprintendenza Archeologica per l’Etruria Meridionale, for our request for the fieldwork permit, and the personal support and advice of the then Soprintendenti Paola Pelagatti and Giovanni Scichilone and of the Soprintendenza’s Ispettrici for Tuscania Dott. ssa Anna Maria Sgubini Moretti and Dott.ssa Laura Ricciardi. In the same vein we thank the British School at Rome for formally preparing and promoting the project’s permit application, but also for providing excellent logistical support throughout the project, and we are extremely grateful to the staff of the School for their help, especially the administrative secretary Maria Pia Malvezzi when Graeme Barker was director, and we are also extremely grateful for the commitment to the project’s success of his successor, Professor Richard Hodges. The financial support for the project was provided by the British Academy, the British School at Rome, the Higginbotham Trust, the Royal Society, the Society of Antiquaries, the University of Leicester and the University of Manchester, and the University of Bologna provided scholarship funding for Nicoletta Vullo to undertake her research stay at the University of Leicester. The on-the-ground support of the comune of Tuscania and mayors Paolo Pantalei, Antonio Marconi

xix

Preface and Acknowledgements

and Domenico Staccini was fundamental to the project’s success, especially in the provision of the Madonna del Cerro school as the main project base. In addition to the Soprintendenza permit to undertake the excavation of the Etruscan farm at Guidocinto (see note on ­spelling, p. 137) reported in Chapter 5, we are grateful for the permission of the landowner La Marchese Ferrari and the tenant Felice Mariotti. We benefited from the advice of Luigi Salvatori of Tuscania’s branch of the Gruppo Archeologico Romano. Among the many kindnesses we received from the citizens of Tuscania we thank in particular those of the late artist Rudolf Kortokraks, our most enthusiastic supporter throughout the project, who arranged an exhibition of the project’s work and was always extremely generous with his time, local knowledge, contacts and hospitality. We are very grateful to the substantial pottery cataloguing undertaken by Stefano Coccia, John Patterson and Nick Whitehead in the early phase of the project, to Giuseppina Battaglia for her assistance in the study of the prehistoric pottery and to Paul Arthur and J. P. Morel for their specialist advice on aspects of the Roman material. Alison MacDonald’s many acknowledgements for the help and advice she received in the study of the prolific Roman material are detailed fully in her Oxford DPhil thesis, but we would like to thank especially her supervisor, the late John Lloyd, for proposing she use the Tuscania material as the main case study for her thesis and for guiding her through it until his tragic and premature death in 2000. Dr Jan van Dalen and Professor Peter Fisher (University of Leicester) provided technical support and guidance for Nicoletta Vullo’s GIS project and her probability models were computed in GRASS using a script compiled by Dr van Dalen. We also thank Alessandro Launaro for his

detailed comments on Chapters 6 and 7, Richard Hodges and Emmanuele Vaccaro for commenting on Chapter 8, and the helpful comments of the British School at Rome’s two anonymous reviewers. In the production of the volume, we would like to acknowledge the professional support of Vicki Herring for her excellent illustrations, Ed Moss for scanning slides and Naomi Rasmussen for her invaluable assistance with the development of the Survey Gazetteer. And finally we offer our heartfelt thanks to the almost seventy archaeologists who participated in the fieldwork, mostly under broiling heat: Giuseppina Battaglia, Paul Beavitt, Rachel Bellamy, David Best, Stefania Bevastro, Elizabeth Cloud, Robert Coates-Stephens, Emma Coleman, Diane Collier, Lisa Cooke, Bibi Cordtz, Jules Cox, Sally Cupitt, Clare Dales, Simon Dobinson, Antonia Douthwaite, Melanie Down, Karen Elder, Jean Gilbert, Kevin Glowacki, Annie Grant, Simone Grosse-Brauckmann, Anna Hamilton, Sanne Hansen, Jonathan Hayes, Peter Hinge, Andrew Hoaen, Ben Hobbs, Stephen Hoyes, Christopher Hunt, Jo Jones, Alex Layman, Helen Loney, Cecilia Luttrell, Alison MacDonald, Stefano Mammini, Federico Marazzi, Isobel McDonald, Ian McGuire, Mikolai Melnyczek, Eric Milne, Sheila Mitchiner, Kate Morton, Diane Moss, Mandy Munro, Alessandro Naso, Catherine Nightingale, Kim Nissan, Marsha Okun, Helen Patterson, Matt Ponting, Eve Pugh, Marco Rendeli, Laura Ricciardi, Jeremy Robinson, Christina Rushe, Sarah Ryder, David Sankey, Kyla Scott, Emma Skipper, Jon Snoxhall, Nigel Spivey, Nigel Thew, Philip Tye, Rachel Tyson, Andrew Upton, Jake Waters, Ross Whitehead, Bruce Whitmee, Helen Wilson and Andrea Zifferero. In the most literal sense In the Footsteps of the Etruscans owes everything to them. Graeme Barker and Tom Rasmussen

xx

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY: RATIONALE, AIMS AND OBJECTIVES Graeme Barker and Tom Rasmussen introduction

of the town. The project was devised to combine several aims, some historical, others methodological, but it was driven first and foremost by a wish to learn more about the historical processes that have shaped the development of the Mediterranean landscape. In particular, we focused on the changing nature of the relationship between town and countryside by taking as our exemplar the territory of a small town in central Italy that had been continuously occupied since Etruscan times nearly 3000 years ago.

Tuscania is a small town some 80  km north-west of Rome in central Italy, in the modern Italian province of Viterbo and administrative region of Lazio (Figs. 1.1 and 1.2). Located at 42°41’86 N x 11°87’03 E, the town is about 150 m above sea level and today has a population of about 8,500 people. The Tuscania Archaeological Survey, the field project that is the subject of this book, investigated the archaeology of the countryside within a 10 km radius

figure 1.1  Tuscania: the walled town. (Photograph: Tom Rasmussen.)

1

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

Fiesole

APE

Florence

o

Arn

Pisa

NNI NE MOU NTA Tiber

Chiana

Murlo/Poggio Civitate

Colline Metallifere

IN S

Volterra

Lake Trasimene

Perugia

Lago dell’Accesa

Mount Amiata Cinigiano, Pievina

on

e br

Podere Tartuchino

Om

Roselle Grosseto

Paglia Saturnia

gn

a

Bolsena

be

Heba

Fiora

Tuscania Pian Fasciano Ancarano

TYRRHENIAN SEA

Blera

Lake Vico Monte Forco

e on

n

ig

M

Viterbo

SABINA

Cimini Mountains

San Giovenale

M ar

Gravisca (Graviscae)

Acquarossa Castel d’Asso

Norchia

Tarquinia (Tarquinii)

ta

Settefinestre Giardino Vecchio

Arro

Vulci

Cosa

Musarna

Castro ne

A

M

Giglio Island

Poggio Buco

EM

AR

M

Mount Argentario

Lake Bolsena

Bisenzio

Al

Doganella

Orvieto (Volsinii)

Volsini Mountains

Tolfa Mountains

Lake Bracciano

Cerveteri (Caere)

Veii

Tiber

Populonia

AN GN M PA RO AM C

Rome

A

50

0 km

Land over 200 m

Ostia

figure 1.2  Tuscania in its geographical setting in Etruria (western central Italy), showing the principal locations and sites in Etruria mentioned in this chapter. Some of the ancient names are shown in brackets; Tarquinia (Etruscan Tarch(u)na and Roman Tarquinii) was known for most of its history as Corneto and only ‘renamed’ Tarquinia in 1922.

2

introduction

The changing relationship between town and countryside over the timescale of Tuscania’s existence has been one of the most important threads running through Mediterranean history (Horden and Purcell 2000). The first half of the first millennium bc was the period of initial urbanization in the Mediterranean, in the aftermath of state formation in the Near East and Egypt (Broodbank 2013). Mediterranean urbanization at this time was characterized by city states, systems of small more-or-less independent polities. Although the focus of most scholarship has traditionally been the city states of classical Greece as the home – Athens in particular – of the literature regarded as one of the foundations of modern Western culture, somewhat comparable political institutions also developed in the central and western Mediterranean. In Italy, urbanization began in Etruria (the western side of the Italian peninsula between the Arno and Tiber rivers, broadly the area between the modern cities of Pisa, Florence and Rome: Fig. 1.2), where the Etruscan city states became the dominant political power in the central Mediterranean until they yielded to the expanding power of Rome in the fourth and third centuries bc (Cornell 1995; Smith 2005, 2014). By the beginning of the Christian era, Rome’s empire encompassed the entire Mediterranean basin. Existing cities and towns had greatly expanded in size, new urban settlements flourished and the countryside was densely settled and intensively farmed to provide for these burgeoning urban populations, especially the c. 1 million inhabitants of Rome itself. The decline and contraction of the Roman empire by the middle of the first millennium ad brought profound changes to both town and countryside, with urban life all but extinguished in much of the western and central Mediterranean and the countryside greatly denuded of population (Christie 2006, 2010; Wickham 2005). By the end of the first millennium ad, urban life began to flourish here once more and rural populations to increase, the principal focus of settlement for the latter being the nucleated hilltop villages, the settlement form that is still the dominant feature of the Mediterranean landscape today (Brogiolo et al. 2000; Francovich and Hodges 2003). In the past fifty

years, though, most such villages have contracted again: towns and cities have exploded in size and the countryside has been progressively denuded of population, as people whose forbears traditionally worked on the land have moved to jobs in the expanding sectors of industry, services and tourism (e.g. Gaggio 2017). Most history has been written by literate elites, and it has often been said that ordinary people to large measure have been denied their history, in the sense of either being ignored by contemporary writers or being written about rather than being able to write about their lives themselves. In the past, as today, such elites have often owned estates in the countryside, rural idylls away from the pace of city life, but from the beginnings of urbanism in the Mediterranean the primary focus for most political activity and elite social intercourse has been the city and town. Hence although most Mediterranean peoples before the modern era lived in the countryside, the history of the Mediterranean landscape, and in particular the changing relationship between town and countryside, has been written mainly from the urban perspective, looking outwards as it were from the city walls to the countryside beyond (Horden and Purcell 2000: 90–92). Archaeology is commonly defined as the study of past societies through their material remains. Classical and Medieval archaeology in the Mediterranean region has traditionally been dominated by the study of the lives of the rich and powerful – great cities, great monuments, great art – but one of the great strengths of archaeology is that it is also extremely good at revealing the lives of ordinary people as well as the rich and powerful. All societies, and all levels of society, create archaeology: everybody, literate or illiterate, uses material culture, and some of it survives in the ground for archaeologists to recover and study. Like historical documents, though, archaeological data pose profound challenges of bias to scholars in their interpretation: archaeologists have to try to understand why particular types of evidence have survived, how they have been biased not just by physical conditions of survival but also by the discard activities of the people who once used them (artefacts might have been lost, for example, or thrown away as rubbish, or carefully buried

3

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

in a ritual context), and how such activities may relate to wider issues of cultural behaviour. Nevertheless, in recent decades archaeologists have demonstrated that they have a considerable contribution to make to the ­writing – rewriting in fact – of Mediterranean landscape history, including the changing relations between town and countryside, through the application of the techniques of landscape archaeology.

of the 1960s. Past human societies, the New Archaeology proposed, needed to be studied not in terms of the culture history that had dominated previous decades but as interacting sub-systems – technological, social, economic, ideological and so on (e.g. Binford 1964, 1965). Archaeologists needed to understand the processes by which such systems developed and changed over the long term (hence the term ‘Processual Archaeology’ came to be used instead of New Archaeology). Social and economic systems could be understood especially as adaptations to particular environmental, technological or demographic circumstances, with changes in the latter being the most likely stimuli of changes in the former. For prehistory, a major focus of Processual Archaeology was on how ecological and subsistence systems interacted. To investigate these relationships, it was argued, archaeologists needed to apply scientific ways of thinking, in the form of hypothesis testing and model building, and use scientific methods so that high quality data were collected systematically and analysed rigorously. The interest in the explanation of diachronic change in social and economic systems favoured the systematic collection of data at the regional scale, and regional field survey was explicitly advocated as an important technique (Binford 1964; Flannery 1976; Plog et al. 1978). Through the 1980s and 1990s, there was a strong reaction by ‘post-processual’ archaeologists led by Ian Hodder against these concerns with environment, system and process (e.g. Hodder 1982a, 1982b, 1986), with parallel trends in geography (e.g. Cosgrove 1984; Hirsch and O’Hanlon 1995). The argument was that a focus on process dehumanized the past by demoting the role of individual agency (Gosden 1995). The focus on topography, technology and land use, on what people did to the land and how it aided or constrained them, was likely to be at the expense of experience and meaning, of how people thought or felt about it (Knapp and Ashmore 1999: 7). The Western notion of landscape that implicitly or explicitly underpinned much landscape research, it was argued, drew upon the Enlightenment vision of the land viewed by a seemingly disengaged observer, but the archaeologist or historical geographer could not have the detached

landscapes and taskscapes People use the term ‘landscape’ in a wide variety of senses. It may be used as a gloss to describe a locale or region; to describe the physical environment of a place, shaped by climate and geography; as the physical space, including the built environment, that participates in the structuration of daily life; and to refer to the paintings, photographs and texts that ‘capture’ a place as a cultural image, ‘a pictorial way of representing, structuring or symbolising surroundings’ (Daniels and Cosgrove 1988: 1). For archaeologists the multiple senses and meanings of the term landscape, and its ability to encompass both the physical and the conceptual (what Gosden and Head [1994] termed its ‘useful ambiguity’), have given rise to an increasingly diverse landscape archaeology, or rather landscape archaeologies, encompassing very different theoretical agendas and technical approaches. In his 1925 essay ‘The morphology of landscape’, the geographer Carl Sauer proposed the concept of the ‘cultural landscape’ as a means to bring anthropology and geography together. In some ways W. G. Hoskins’ The Makings of the English Landscape (1955), a survey of the historical development of rural England since AngloSaxon times, though very different in scope and method in its integration of documentary records, maps, place names and the limited archaeological evidence available to him, stemmed from a similar tradition in historical geography. However, it was Gordon Willey’s pioneering archaeological survey of the Viru Valley in Peru (1953) that provided the best exemplar of regionally based multiperiod (diachronic) settlement studies that were one of the most enduring outcomes of the ‘New Archaeology’

4

landscapes and taskscapes

gaze of the landscape painter because past landscapes were not like painted landscapes, fixed in time: they were created and constantly refashioned through engagement and occupation, sustaining multiple identities (Layton and Ucko 1998a; Thomas 1993). The ‘Western Gaze’ – elite, usually male, commonly colonialist – had invariably privileged those at the top of the hierarchy and masked and dehumanized those at the bottom (Bender 1993a). The focus should therefore be on the subjective and socially constructed nature of landscape, of landscape situated in ideology and being-in-the-world (e.g. Bender 1993a, 1993b; Cosgrove 1984; Layton and Ucko 1998b). Tilley (1994) in particular advocated a phenomenological approach to try to understand past landscapes as they were perceived and experienced by their various inhabitants, perceptions and experiences that would differ between different individuals and social groups – the ‘multivocality’ of the past.

in short, studying chronological sequences of what he termed ‘taskscapes’. The Tuscania Archaeological Survey was conceived and executed in the years straddling the processual and post-processual debates about the ‘proper concerns’ of landscape archaeology. Its overriding focus of interest was in ‘the lives and works of past generations’, in Ingold’s telling phrase, in our case the people who had lived in the particular terrain demarcated by the 10 km radius from a small Italian town with origins going back 3000 years. Given our interests in long-term landscape histories and societies at very different levels of complexity and scales, we endeavoured to steer between the more extreme divisions of the processual/post-processual debate characterized by Ingold (1993: 172) as ‘the “scientific” study of an atemporalized nature’, on the one hand, and ‘the “humanistic” study of a dematerialized history’ on the other. In his classic study of Mediterranean history that laid the foundations for the Annales school of historical geography, the French historian Fernand Braudel characterized history as the interplay between short-term, medium-term and long-term processes (Braudel 1949, 1972). The former (événements) he envisaged as the events of political and military history. Medium-term processes (conjonctures) were the kinds of changes in society operating, say, at the scale of one or two generations. Long-term processes included factors such as the constraints of a particular technology, or the natural characteristics of a particular kind of landscape, on how people could live in it (the longue durée). Shaping all of these were the mentalités, the world-views of particular societies. Building on the experiences of one of us in the Biferno Valley Survey (Barker 1995a, 1995b), we set out to bring a similarly holistic perspective to the Tuscania Archaeological Survey. We were interested in how different kinds of societies and social groups in the past had shaped or created different kinds of landscapes – natural, social, economic, ideological – the interactions between these landscapes, and the interplay between external and internal factors operating at different timescales in shaping the trajectories of landscape change from prehistoric times to the present day.

Landscape has to be contextualised. The way in which people – anywhere, everywhere – understand and engage with their worlds will depend upon the specific time and place and historical conditions. It will depend upon their gender, age, class, caste and on their social and economic situation. People’s landscapes will operate on very different spatial scales, whether horizontally across the surface of the world, or vertically – up to the heaven, down to the depths. They will operate on very different temporal scales, engaging with the past and the future in many different ways … Each individual holds many landscapes in tension. (Bender 1993b: 22)

An influential paper from this time that has influenced many landscape archaeologists ever since was ‘The temporality of landscape’ by the anthropologist Tim Ingold (1993). In it he sought to find a way forward between what he called the ‘sterile opposition between the naturalistic view of landscape as a neutral, external backdrop to human activities, and the culturalistic view that every landscape is a particular cognitive or symbolic ordering of space’ (Ingold 1993: 152). The landscape was better imagined, he suggested, as ‘an enduring record of – and testimony to – the lives and works of past generations who have dwelt within it and in so doing have left there something of themselves’. Landscape archaeologists were,

5

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

mediterranean plough-zone archaeology

employed by the Tuscania Archaeological Survey. The two major pioneering applications of this technique were the University of Minnesota’s Messenia Expedition in the 1960s, which set out to reconstruct settlement patterning around the second-millennium bc Mycenaean palace of Pylos in the Greek Peloponnese (McDonald and Rapp 1972), and the British School at Rome’s South Etruria Survey in the 1950s and 1960s, a study of changing settlement patterns in the territory of the ancient city of Veii, and adjacent areas, north of Rome (Potter 1979; Ward-Perkins et al. 1986; Fig. 1.3). The South Etruria Survey was particularly relevant for our own project because, as described in the following section of this chapter, its results provided the principal ­starting point for our investigation.

Via Amerin

a

The techniques developed by landscape archaeologists for mapping human activity include air photography, satellite imagery, and a variety of systems of geophysical survey for investigating the nature of buried structures (Campana 2018; Pasquinucci and Trément 2000). In the Mediterranean, probably the most important weapon in the landscape archaeologist’s armoury is what is generally termed ‘field survey’ or ‘field-walking’: the systematic searching for and collection of archaeological artefacts such as stone tools and potsherds visible on the ground surface, especially in ploughsoil (Alcock and Cherry 2004a; Francovich et al. 2000). This was the main methodology Lake Vico

Sutrium

Via

Ager Faliscus

nia

mi

Fla

Lake Monterosi

Porciano Mazzano

Ager Capenas

Narce

Mola di Monte Gelato

ia

ass

C Via

Lake Bracciano Cl

Monte Forco

Vallelunga

Lucus Feroniae

Baccano

od

ia

Crocicchie

Quattro Fontanili

Tib er

Ager Veientanus S. Cornelia Veii

land over 300 m pre-Roman Roman Medieval

S. Rufina

pollen sample

0

Rome

10 km

figure 1.3  South Etruria, showing the location of the British School at Rome survey projects of the 1950s and 1960s. (Adapted from Potter 1979: fig. 1.)

6

mediterranean plough-zone archaeology

The South Etruria Survey was coordinated by the School’s then director, John Ward-Perkins. It was developed in the context of the increasing damage to the archaeological record that was visible throughout South Etruria in the form of ploughed-up remains of previously buried ancient structures, as farmers changed from their traditional ox-pulled ploughs, that had ploughed only a few centimetres deep, to tractor-pulled ploughs that cut down 30–50 cm. Ward-Perkins organized teams of archaeologists to walk over freshly ploughed fields. They mapped the locations of concentrations of artefacts lying on the ploughsoil surface that were the traces of buried or destroyed archaeological sites and collected samples of these artefacts as a means of dating when the sites had been occupied. The teams found hundreds of prehistoric, Etruscan, Roman and Medieval sites, the Etruscan and Roman periods being particularly well represented (Duncan 1958; Jones 1962, 1963; Kahane et al. 1968; WardPerkins 1961, 1962, 1964; see also Cascino et al. 2012; Patterson 2004; Patterson et al. 2020). Over the years Ward-Perkins and his collaborators also excavated a number of sites in the survey area including a Bronze and Iron Age settlement, an Iron Age village and cemetery, Roman rural sites and Early Medieval settlements and churches, as well as parts of Etruscan and Roman Veii (e.g. Christie 1991; Potter 1972, 1976a; WardPerkins 1961). These excavations produced stratified collections of pottery that were vital to help with the dating of the mixed pottery from the ploughsoil collections, and they also yielded important information about the likely characteristics of the buried structures represented by surface artefacts. For example, excavated Roman remains suggested that artefact collections could be interpreted as the residues of either villas or poorer farmsteads on the evidence of differences in pottery types and the presence or absence of wealth indicators such as mosaic tesserae, pieces of statuary and wall plaster. Ward-Perkins also encouraged palynologists to reconstruct vegetation history from fossil pollen preserved in lake sediments, and geomorphologists to reconstruct changing river regimes from alluvial sediments, their sequences often having implications for the effects on the

landscape not just of climatic change but also of human activities such as forest clearance for agriculture. The result of this remarkable multidisciplinary programme of survey, excavation and environmental science, as brilliantly summarized by Potter (1979), was an archaeological history of landscape change from the centuries preceding Etruscan state formation to the emergence of the modern landscape of nucleated hill villages at the end of the first millennium ad. In the ensuing decades, regional survey projects were undertaken in almost all parts of the Mediterranean, building on the examples of the Messenia and South Etruria Surveys. As described in a number of edited volumes summarizing much of this work (e.g. Alcock and Cherry 2004; Barker and Lloyd 1991; Favory and Fiches 1994; Keller and Rupp 1983) and individual project publications (e.g. in Spain: Carreté et al. 1995; southern France: Trément 1999; Italy: Attema 1993; Attema et al. 2000; Barker 1995a, 1995b; Carandini and Cambi 2002; Coccia and Mattingly 1992, 1996; Cucini 1985; Delano-Smith et al. 1986; Hayes and Martini 1994; Lock and Faustoferri 2008; Moreland 1986, 1987; Percorsi et al. 2006; Yntema 1993a, 1993b; Cyprus: Given et al. 1999; Given et al. 2013; Dalmatia: Chapman et al. 1996; Gaffney et al. 1997; Greece: van Andel and Runnels 1987; Cherry et al. 1991; Hayden 2005; Mee and Forbes 1996; Renfrew and Wagstaff 1982; Watrow et al. 2012; Wright et al. 1990), probably the most important achievement of these regional field-walking projects was their demonstration of the complexity of rural settlement in classical times – what John Lloyd (1991) termed ‘the busy countryside’. The classical landscape, it became clear, was characterized by an abundance and diversity of settlement forms entirely unsuspected from the written sources (Launaro 2011). Collaboration between archaeologists and geographers has been a feature of many of these regional landscape studies and has demonstrated the same sort of complexity regarding the development of the natural landscape and of people’s impact on it (e.g. Hunt et al. 1992; van der Leeuw 1995; Leveau et al. 2000; Lewin et al. 1995). Classical farmers in particular seem to have caused deforestation and accelerated erosion in many regions, but

7

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

significant episodes of erosion have been noted as well from the Bronze Age to the recent past. Furthermore, different kinds of agricultural processes had different environmental impacts. In the Argolid region of the Greek Peloponnese, for example, erosion seems to have been caused by arable intensification in the Bronze Age, pastoral expansion and terrace abandonment in Hellenistic times, deforestation for arable expansion in the Byzantine period and finally – as throughout the Mediterranean – on a vastly increased scale in recent decades by mechanized deep ploughing (van Andel and Runnels 1987). The Biferno Valley Survey found an equally complex sequence, with a different chronology (Hunt 1995a, 1995b). Climatic change also accelerated erosional trends in the Late Roman and Early Medieval periods, as VitaFinzi (1969) first surmised. Alongside field survey’s remarkable contribution to knowledge of Mediterranean landscape history, however, has been continuous debate among both its critics and its practitioners about its methodologies and overall effectiveness (Campana 2018). Areas of discussion included the relative effectiveness of different techniques for defining survey areas and sub-samples within them, conducting the field-walking and interpreting the materials collected; the effects of soil processes such as alluviation and erosion moving or burying surface material; the effects on artefact discovery of different kinds of land use, ploughsoil conditions, and changing conditions of light and shadow; and biases caused by the variable skills and experiences of field team members. Differential ‘archaeological visibility’ was recognized as likely to be particularly significant: the fact that some components of the archaeological record were inherently likely to be better represented than others in terms of the quantities of what there was to be found, or likely to be visible or both. In Italy, for example, the Roman period was generally characterized by high rural populations living in dispersed farms (Launaro 2011). Potentially, therefore, they built lots of sites for archaeologists to find. These farms, moreover, usually had well-built structures with walls of brick and roofs of tile, both durable materials. The people used well-made pottery (so durable) that was

produced on a large, sometimes almost industrial, scale, and the finest wares tended also to have bright polished surfaces (so likely to be visible in the ploughsoil), and the period of manufacture of many such sherds can also be dated to individual centuries. In the Early Medieval period, by contrast, there was a much smaller population, living in houses that excavations showed were for the most part of wood and thatch (so leaving no durable, easily visible, traces), in small nucleated settlements on hilltops that frequently today are wooded and so effectively inaccessible to systematic field-walking (Francovich and Hodges 2003; Moreland and Pluciennik 1993; Moreland et al. 1993). Also, much of their technology was probably of organic materials that do not survive (wooden bowls, for example), and much of the pottery they used was rather poorly made and friable. The Biferno Valley Survey was typical of many field projects in Italy in finding hundreds of sites for the (approximately) thousand years of the classical period (c. 500 bc–ad 500), but less than a dozen for the ensuing 500 years (Barker 1995a). In Tuscany, 95 per cent of the c. 20,000 archaeological sites located in a 30-year-long programme of landscape research by the University of Siena relate to the time span between the sixth century bc and sixth century ad (Campana 2018: 20). Plough-zone survey has also been criticized for its common delineation of a past landscape as a map of dots (most assumed to represent habitation loci of some kind) separated by white space, with little direct insight into the multifarious tasks that must have characterized most taskscapes beyond the habitations (Campana 2018). (Excavation of activity loci could, of course, provide indirect evidence of the activities beyond them.) Also, the landscape activities of different kinds of societies produce different kinds of signatures, some more visible than others. Ethnoarchaeological studies of hunting and pastoral societies, for example, show that they often move between a series of seasonal camps which may be in the same general location year by year, but the settlement archaeology created can consist of thin spreads of debris extending over hundreds of metres rather than a concentration of occupation materials at a fixed site. Mobile people in the

8

research issues

past would likely have created a much more ephemeral archaeological record than people living in fixed settlements. How people disposed of their rubbish will also have affected the kind of surface archaeology created: for example, people might spread their rubbish as manure on the fields surrounding their settlements or bury it in pits – the latter was commonly the case on Medieval urban sites, including at Tuscania itself (Johns et al. 1973; Ward-Perkins et al. 1972). At the time we were planning the project, therefore, it was clear that Mediterranean landscape archaeology had to confront challenging methodological problems in trying to distinguish absence of settlement evidence from evidence for an absence of settlement, and evidence for dense settlement from evidence for prolific, wellpreserved and conspicuous artefacts. The variability of the field techniques, the context of the POPULUS project (Barker and Mattingly 2000a–e), was severely weakening the potential of landscape archaeology to write regional or in particular Mediterranean-wide landscape histories integrating the results of different regional survey projects (Alcock 2000; Alcock and Cherry 2004; Mattingly 2000). These were all challenges that we hoped to address in the Tuscania Archaeological Survey fieldwalking programme, using the methodologies described in the next chapter.

the role of Greece as a possible source of inspiration. Some historians have tended to see the question in terms of the importation of an already fully developed Greek model c. 700 bc, the beginning of the Orientalizing period (socalled because of Eastern influences discerned in Etruscan art from this time) (Drews 1981; Harris 1989). Genetic studies of modern central Italian populations were taken as evidence for an east Mediterranean/Anatolian origin of the Etruscans (Achilli et al. 2007; Brisighelli et al. 2009). Some studies of ancient DNA in Etruscan skeletons did not find persuasive evidence for significant genetic continuity with later Italian populations (Belle et al. 2006; Ghirotto et al. 2013); others proposed indigenous rather than exotic origins (Tassi et al. 2013). The most recent, using the ancient DNA (aDNA) of around 80 individuals from Etruria spanning from 1000 bc to ad 1000, including around fifteen from the centuries of Etruscan hegemony and independence from Rome (the seventh to the fourth centuries), proposes an Indo-European-associated steppe ancestry for the Etruscans in line with the steppe ancestry that geneticists have also proposed, equally controversially in relation to the archaeological evidence, for the wider European population in later prehistory (Allentoft et al. 2015; Olalde et al. 2015). While acknowledging the profound cultural impacts of Phoenician and Greek settlement and commercial activity, most archaeologists have argued that there is no need to look beyond Italy for the dominant impulse towards the formation of the Etruscan city states, because the seeds of state-level or urbanized societies were present already in the communities of the Villanovan Iron Age in Etruria c. 900–700 bc, and even perhaps before then (Barker and Rasmussen 1998; Broodbank 2013; Fulminante 2014; Guidi 2006; Rasmussen 2005; Riva 2020; Spivey and Stoddart 1990; and see Chapters 4 and 5). While this is the view that we ourselves have favoured (Barker and Rasmussen 1988) and continue to favour, demonstrating a significant increase in social complexity is one thing but explaining it quite another. Advancing understanding about the trajectory of urbanism in Etruria has been greatly hampered by lack of detailed knowledge about the nature of settlement in the centuries before the appearance of Etruscan towns,

research issues The specific research agenda of our project was developed in the light of previous archaeological and historical studies of town and country relations in central Italy, building especially on the work of the South Etruria Survey. Sets of questions were framed focusing especially on the Etruscan, Roman and Medieval landscapes and the transitions between them.

Etruscan Urbanization The first main area of interest related to the origins and character of Etruscan urbanization. Debates over this have centred around the chronology of its emergence and

9

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

and in the initial stages of their development. Had there been a gradual process of nucleation, with people coming together into fewer, larger, settlements in the preceding phases of prehistoric occupation, or had the growth of central sites been rapid? In either case, to what extent was the surrounding countryside depopulated? Etruscan archaeology for generations has been concerned with the controlling elites and with the expensive goods with which they surrounded themselves that fill the world’s great museums today. What has been conspicuously missing has been any focus on the lower end of the settlement hierarchy, most of whom are assumed to have been living on the land (Barker et al. 1993a; Potts and Smith 2021). Knowledge about Etruscan lives had been derived almost exclusively from necropolis archaeology, and D. H. Lawrence, writing in 1932, was hardly exaggerating when he commented that ‘now, we know nothing about the Etruscans except what we find in their tombs … Of first-hand knowledge we have nothing except what the tombs offer’ (1986: 31). It is, of course, because the tombs have offered up so great a wealth of objects, from the great assemblages of the Orientalizing period, such as that from the Regolini-Galassi tomb at Cerveteri to the Hellenistic riches of the Volumnii tomb at Perugia, that the temptation to pillage, and later to excavate, cemeteries has always been extreme. At first, objects were simply pulled out of context and treated in museum displays as objets d’art. Later, they were studied for the creation of coherent ­typologies of artefacts on which the major chronological divisions of Etruscan culture are based: Orientalizing (700–570 bc), Archaic (570–470 bc), classical (470–300 bc) and Hellenistic (300–31 bc). (The Hellenistic phase encompasses the final retreat of Etruscan power in the face of Roman territorial expansion, 31 bc being the date when Augustus, Rome’s first emperor, came to power.) Researchers, especially in more recent times, have also tried to make sense of Etruscan museum collections in social, economic and ideological terms (e.g. Izzet 2010; Riva 2020), but their success has always had to be tempered by the biases in the material itself: that it is mainly funerary, and that it is concerned for the most part with the highest strata of society (Potts and Smith 2021).

Settlement archaeology came late in Etruscan studies and, with the exception of Marzabotto near Bologna on the edge of the Po plain in the north, especially late where investigation of the major urban sites is concerned. Attention on the city sites has, by intention or luck, focused mainly on sanctuary sites. At Cerveteri this is true both of the old excavations of Mengarelli and of more recent initiatives (Cristofani and Nardi 1988; de Grummond and Pieraccini 2016), though a huge deposit of dumped material looks more domestic in nature (Cristofani 1992–1993). It is true, too, of excavations on the acropolis of Populonia, at Volterra, Fiesole and also on the Pian di Civita at Tarquinia (Bonghi Jovino and Chiaramonte Treré 1997). Part of an industrial complex was uncovered in the lower town of Populonia (Cristofani and Martelli 1979), as was a series of Iron Age huts on the Monterozzi ridge at Tarquinia (Linington 1982). At some other sites, investigations have been on a larger scale but have penetrated through to Etruscan levels only at certain points – notably at Roselle to a substantial archaic house (Donati 1994) and to one or two even earlier domestic structures. Rather different are the cases of Doganella (Perkins and Walker 1990) and Veii (Cascino et al. 2012; Guaitoli 1982; Patterson et al. 1999; Tabolli and Cerasuolo 2019; WardPerkins 1961), both large urban sites which have been carefully surveyed and field-walked, but at which only relatively small areas of domestic structures have been excavated. At Cerveteri too, in addition to excavations mentioned above, survey resulted in a series of settlement maps of the urban area from the Early Iron Age to the first century ad (Merlino and Merenda 1990). A few smaller sites – towns and large villages rather than cities – have been investigated with considerable care. The process began with the Swedish excavations at San Giovenale and Acquarossa in the 1960s and 1970s (Wikander and Roos 1986) and continued at Poggio Civitate near Murlo (Phillips 1993). The latter is usually discussed in terms of a large isolated building complex but is more likely to have been part of a larger settlement (there is a necropolis area nearby). Work was conducted also at a group of houses set in the vicinity of Lago dell’Accesa in the Colline Metallifere (‘metal-bearing

10

research issues

hills’) east of Populonia (Camporeale 1997). However, it is buildings and farms out in the countryside that have received least attention by excavators – the humble dwelling of the kind that leaves traces in the form of small surface scatters of artefacts that can be encountered in many parts of Etruria. Some of these have proved on further investigation to have been completely ploughed out, but an example of the high quality of data that can be extracted from a poorly preserved rural site was the Etruscan farm at Podere Tartuchino (Perkins and Attolini 1992), a striking discovery of the Albegna Valley Survey (Carandini and Cambi 2002). Field survey therefore has significant potential to tell us about the nature of rural settlement patterns and densities. Agriculture lay at the heart of the Etruscan socio-economic system and of the conditions necessary for the development of Etruscan urbanism, as elsewhere around the ancient Mediterranean, and we wanted to devise a research programme that would shed light on the exploitation of land resources in a chosen area and on the density of rural settlement both at the beginning and height of Etruscan power and as it was affected by the aggrandizing power of Rome in the last centuries bc. Etruscan society was characterized by clientship, and the economy of Archaic Etruria was pre-monetary in the sense that coinage was not used extensively or systematically, with most economic relations being embedded in networks of social and ritual obligations (Barker and Rasmussen 1998; Izzet 2010; Riva 2020). However, beginning before the advent of coinage, bronze ingots (aes rude) are thought to have acted as some kind of standard medium of exchange and, as they have been found in both towns and farms, there could have been some form of primitive market exchange between urban centres and their surrounding rural populations. Could such relationships be investigated from comparisons of urban and rural material culture? Could we by the same means also gain insights into the nature of rural Etruscan societies, and the extent of their independence from or obligations to urban societies from factors such as wealth indicators from structural remains and artefacts? Were Etruscan cities and towns nucleated centres housing not only elites

and specialist groups such as craftworkers but also people who farmed the surrounding landscape (what are sometimes called ‘agro-towns’)? Or did they function more as administrative centres or markets of some kind for dispersed rural populations? Did some Etruscan elites in fact live in the countryside and not – as always assumed – in the city? Could graves and grave goods in the countryside be used as signatures of the rural population’s participation in or exclusion from the norms of Etruscan ideology? In short, could field survey around a typical Etruscan centre cast new light on how Etruscan towns interacted with their hinterlands in economic, social and ideological terms?

‘Romanization’ and Roman Imperialism The next major group of research questions concerned processes of Roman imperialism in central Italy. Much historical work on this process of ‘Romanization’ in Italy, as in the rest of the Roman empire, has suffered from a Romano-centric, colonialist perspective (Barbara Bender’s ‘Western Gaze’ again), characterized by an assumption of a normative experience of Roman imperialism and an evolutionary cultural paradigm in which under-developed societies succumbed inevitably to Roman power and culture. More recent approaches have tended to emphasize the diversity of people’s engagements with Roman power and cultural norms, and the dialogues and negotiations between colonizers and colonized (Alcock 1993; Ceccarelli 2016; Keay and Terrenato 2001; Mattingly 1997, 2006, 2011; Millett 1990; Webster and Cooper 1996; Witcher 1999). To what extent was Romanization an active policy imposed on subject peoples, or the result of local elites actively emulating Roman ways, and in the latter case was motivation primarily socio-political or economic or ideological (or, more likely, complex combinations of these)? Much of this rethinking has been stimulated not only by the changing paradigm of post-colonial perspectives but also by the results of archaeological research: the complex changing social relations of Romanization manifested themselves in changing material culture in towns, changes in rural settlement

11

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

forms and systems of land use, and changes in town and country relations, all of which were amenable to archaeological investigation. Although there has been a considerable history of research on individual Roman towns in Italy as elsewhere in the Roman empire, most excavation has concentrated either on exceptionally well-preserved and in many respects atypical towns, such as Pompeii and Ostia, or on major public buildings within selected towns. Understanding of Roman urbanism in Italy has been transformed in recent decades by the systematic investigation of different kinds of urban sites in the area of the British School at Rome’s South Etruria Survey using a variety of remote sensing techniques that in favourable conditions are capable of surveying hectares of land per day (e.g. Campana 2018; Carlucci et al. 2007; Gaffney et al. 2014; Hay et al. 2010; Johnson et al. 2004; Keay et al. 2000, 2014; Opitz 2009; Verdonck et al. 2020), enhancing survey archaeologists’ ability to ‘fill in the white space’ of an archaeological landscape (Campana 2018). Modern excavations of rural sites are still remarkably rare in central Italy. An early example was Barri Jones’ excavation of Monte Forco in the Ager Capenas, a small rural site was investigated at Giardino Vecchio in coastal Tuscany (Carandini 1985b: 106–107) and a modest Samnite and Roman villa was excavated at Matrice in the Biferno valley (Lloyd 1995a), but the picture has recently been transformed by the excavation of a series of small rural sites around the villages of Cinigiano and Pievina in inland Tuscany (Bowes 2020; Ghisleni et al. 2011; Vaccaro et al. 2013). The largest-scale study has been of the Settefinestre senatorial villa at the top end of the social and economic spectrum in the territory of the city of Cosa (Carandini 1985a). By collecting new survey data from the countryside around Tuscania, a town we knew was occupied in Roman as well as Etruscan times, we hoped to inform current debates about the nature of Roman imperialism in its first critical phase of expansion north of the Tiber. An obvious question regarded the extent to which the conquest and Romanization of Etruria meant continuity or rupture in rural settlement. Potter (1979) had

concluded from the South Etruria Survey data that in the Ager Faliscus to the north of Veii the survival rate of farms into the Roman period was less than 20 per cent, whereas around Veii itself the ratio was much higher – about two-thirds. Hemphill’s survey of the westernmost part of the BSR survey area, between the Via Cassia and Via Clodia (Hemphill 1975), showed a similar continuity of settlement between the Etruscan and Roman periods to that of the Veii area. Yet there were also clear examples elsewhere in Etruria of disruption to patterns of rural settlement. The rich farmlands of the Maremma coastal lowlands, for example, attracted particular Roman interest once the Roman colony of Cosa was established on the coast in 273 bc. The Albegna Valley Survey had picked up traces of ‘centuriation’ (systems of Roman land division) around this centre, as well as around the colonies of Heba and Saturnia further inland, revealing how a landscape that was quite densely populated in Etruscan times became much reduced in population immediately after the Romans moved in, and once under Roman control the land was increasingly intensively farmed in the last two centuries bc (Attolini et al. 1991; Perkins 1991). On the Adriatic side of the peninsula in Samnite territory, the Biferno Valley Survey had picked up a marked decline in numbers of farms, a drop in fact of around 40 per cent, when this area became part of the Roman state after 80 bc (Lloyd 1995a, 1995b). The reliable demonstration of settlement continuity and discontinuity is one of the most debated areas of field survey methodology (Francovich et al. 2000). One significant problem concerns the difficulties of dating site foundations and abandonments when rich sites with plentiful fine wares can be dated more precisely than the (usually) many more sites with poorer ceramics. Another is judging the significance of mixed sherd assemblages: when do a few sherds of Etruscan pottery in a rich Roman assemblage denote continuity, and even if continuity of occupation has been demonstrated, what did that mean in terms of continuity of settlement forms and social and economic structures – of ways of living? Could we contribute to these methodological issues, and understanding of the Romanization of Etruria, by

12

research issues

and development of the phenomenon called incastellamento whereby populations moved from lowland habitations and established fortified hilltop sites – in the case of Tuscania, the hilltop settlement of Colle San Pietro now crowned by the magnificent Early Medieval church of San Pietro (Fig. 1.4). The nature of these processes, their chronology, and in particular their cause or causes, were and remain much debated (Francovich 2002; Francovich and Hodges 2003; Wickham 1989) and, exactly as in the case of Etruscan nucleation, much of the debate has concerned the nature of rural settlement across the Roman/Early Medieval boundary: was the countryside abandoned, or did lowland farms and hamlets continue to be occupied as vestiges of the Roman landscape? Were the new hilltop villages agro-centres from the outset, or did they function more as seats of feudal power controlling rural populations in their environs? Most modern Italian hill villages have foundation dates around a thousand years ago, giving the impression of a rather static landscape since then of long-lived successful villages surrounded by dispersed farms. However, archaeological surveys such as that in the Biferno valley had shown how some regions of the Italian countryside had witnessed significant episodes of population expansion and contraction over the past thousand years, with ‘lost villages’ from past phases of high population marked today by isolated churches in the countryside, or by concentrations of surface archaeological materials around what are now isolated farms (Hodges and Wickham 1995). Such episodes were not apparent in the evidence for Medieval and Post-Medieval settlement collected by the South Etruria Survey (Potter 1979), but studies of postRoman ceramics had advanced considerably since that project’s completion, making it an open question whether the rather stable post-Roman landscapes indicated by the South Etruria Survey reflected genuine evidence of an absence of episodes of settlement expansion and contraction, or were more a question of archaeological invisibility and insufficiently refined ceramic chronologies. (This was in fact an important focus of the major re-study of the South Etruria Survey material in the late 1990s and early 2000s: Cascino et al. 2012; Harrison et al. 2004; Patterson

a detailed study of the transition between Etruscan and Early Roman (Republican) settlement in the countryside around Tuscania? Formalist or modernist perspectives on the Roman economy emphasize money, markets, long-distance trade in low-value goods, and specialization in craft and agricultural production, whereas substantivist or primitivist models argue that production and exchange were embedded in social relations, in particular the social and economic needs and aspirations of elites. The survey and excavation data from Roman Italy had been cited notably by Andrea Carandini as evidence in support of general historical models broadly within the formalist paradigm, concerning the impact of conquest on agricultural development in Italy, in particular the investment of wealth in land and the growth of large villa estates sustained by the ‘slave mode of production’ at the expense of the small farmer (Carandini 1981, 1985a, 1985b, 1988). Could the surface archaeology of the countryside around Tuscania inform such debates about the nature of the Roman economy, and the changes to the agricultural landscape of central Italy, through the several centuries of the Roman Imperial period? Methodological concerns were critical here. For example, could historically attested settlement forms such as the villa be satisfactorily identified from archaeological survey sites using measurements of size, density and wealth of surface remains? To what extent could negative evidence in an archaeological survey (mindful of our earlier comments about different survey methodologies) be used to support historical models regarding the collapse of the peasantry (Foxhall 1990)?

Medieval Settlement and Incastellamento South Etruria in the Early Medieval period was one of the principal arenas where changing power relationships between the Lombards and Carolingians on the one hand, and Rome and the Church on the other, provided the context for the emergence of the city states of Medieval Italy (Christie 1991, 2006). Our final set of questions concerned the transformations to the landscape in the post-Roman period onwards, in particular the origins

13

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

figure 1.4  Tuscania’s Colle San Pietro acropolis, looking south-east from the Medieval/modern town. (Photograph: Graeme Barker.)

2004.) Could a new survey project give better understanding of the development of post-Roman landscapes in South Etruria?

a markedly late ‘Archaic Etruscan’ manner (Moretti 1984: fig. 6). A few may also be struck by the realization that much of the bleak atmospherics of Pier Paolo Pasolini’s 1966 Uccellacci e Uccellini (Hawks and Sparrows) was created by filming on the Colle San Pietro acropolis around the church itself and the ruined towers. We selected Tuscania for three principal reasons: its known archaeology and history; its geographical location; and its surrounding landscape.

the selection of tuscania Today Tuscania is off the main tourist routes – most travellers crossing Etruria to and from Rome take the coast road past Tarquinia, or the inland route, the Via Cassia, passing through the provincial capital of northern Lazio, Viterbo. Those who come to visit Tuscania for sightseeing do so with two main purposes: to see Etruscan tomb material, and to admire the Medieval architecture of San Pietro and Santa Maria Maggiore on Colle San Pietro, among the finest Romanesque churches in Italy (Figs. 1.4 and 8.2). The discerning among them may even notice a visual connection between the two: the ‘running’ male figure on the left of the façade of San Pietro is carved in

Tuscania’s Settlement Archaeology and History Tuscania seemed likely to be typical of many towns in this part of central Italy in its evidence for more-or-less continuous occupation since Etruscan times, and with some indications of earlier settlement too. The Colle San Pietro hill had produced a sporadic find of a Neolithic axe of dark stone, perhaps 5000–6000 years old (Gianfrotta and

14

the selection of tuscania

Potter 1980: 438). In her survey of the known archaeology around Tuscania for the Forma Italiae series (Figs. 2.6 and 7.5), Stefania Quilici Gigli (1970: 148; figs. 211 and 212) had illustrated a sherd with the characteristic stippled decoration of the Earlier or Apennine Bronze Age (2200– 1400 bc) that had been picked up at a site some 2 km to the south on the right bank of the Marta river. A number of sherds of the Late and Final Bronze Age (1400–900 bc) had been reported from the foot of the hill itself (Colonna 1974: 256, plate 54). In 1974 traces of Early Iron Age (900– 700 bc) huts were found on its north-west slope in an excavation directed by F. Boitani (Sgubini Moretti 1986b: 247, note 7). In February 1971 Tuscania was struck by a devastating earthquake which ruined much of the town (Fig. 1.5) and killed many inhabitants. Before reconstruction of the urban centre began, the British School at Rome (BSR) was asked to help conduct archaeological soundings and investigations on Colle San Pietro and in the centro storico, the historic centre of the town within the Medieval walls (Johns et al. 1973; Ward-Perkins et al. 1972). Immediately after the earthquake, hasty explorations and clearance operations had taken place on Colle San Pietro in an area measuring some 55 m by 35 m. The subsequent investigation of exposed surfaces and structures by an archaeological team revealed a long sequence of occupation from the Early Iron Age to the Late Middle Ages (Gianfrotta and Potter 1980). A detailed study of the Medieval town walls was also undertaken (Andrews 1982). This work had enabled a reasonably clear picture to be drawn of the development of Tuscania, showing that Colle San Pietro had been the focus of pre-Etruscan, Roman and Early Medieval occupation, with settlement then shifting a few hundred metres north-westwards to the Rivellino hill, the site of the later Medieval and present-day centre of habitation (Fig. 1.6). One consequence of the earthquake for our own project was that a small school or asylum for the children orphaned by the earthquake was built outside the town adjacent to the church of the Madonna del Cerro. Long empty at the time of our fieldwork, it was made available by the comune as the project base for the first three seasons of fieldwork.

The first major exhibition of Etruscan objects staged outside Italy, in Pall Mall in London in 1837, was in fact organized by excavators who were themselves based at Tuscania (Swaddling 2018: 45–53). It consisted of a series of ‘walk-in’ reconstructions of chamber tombs, together with their contents, from Tuscania, Tarquinia, Vulci and Bomarzo, and was clearly a great success, with enthusiastic reviews in The Times. A full catalogue accompanied it (Campanari 1837), and graphic designs for the tomb interiors are in the possession of the British Museum. One of these (Fig. 1.7) shows a chamber with several stone sarcophagi, all of which were found in a tomb near Tuscania by the road to Tarquinia. The repercussions of the London exhibition were far-reaching. The British Museum bought up much of the material, including the sarcophagi from

figure 1.5  The historic centre (centro storico) of Tuscania after the 1971 earthquake. (Photograph: Graeme Barker.)

15

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

S. Maria della Rosa

Porta di Montascide

S. Marco

Porta di Poggio

S. Silvestro

S. Biagio

?

S. Paolo

S. Francesco

140m

13

0m

Roman Road

?

Rivellino

chiolo

Rom

an

Roa

d

Fosso Mas

12

0m

S. Maria Maggiore

S. Pietro

Foss

o Fo

ssac

io

Gate

?

? ?

River

Marta

0

200 m

figure 1.6  Tuscania: plan of Colle San Pietro and the later Medieval town; the dashed lines indicate David Andrews’ suggested reconstruction of the town walls, incorporating suggestions of earlier studies by Turriozzi and Campanari (the dashed lines with question marks). (Adapted from Andrews 1982: fig. 3.21.)

16

the selection of tuscania

Tuscania, to augment its burgeoning Etruscan collections. One visitor to the show was Elizabeth Caroline Hamilton Gray, who was inspired by it to go out and visit some of the sites of Etruria. These included Tuscania, where she was entertained by the Campanari family who had been responsible for the Pall Mall event. In Vincenzo Campanari’s courtyard-garden she was shown a full-scale reconstruction of the tomb of the Vipinana family, with its several generations of sarcophagi, excavated on the lower slope of the San Pietro hill only a few months before her arrival (Colonna 1978: 93). It was partly in response to her travel book (Hamilton Gray 1841) that George Dennis wrote his classic work The Cities and Cemeteries of Etruria (Dennis 1848) to correct what he considered to be deficiencies in her work. He included a substantial chapter on Tuscania’s antiquities. After these pioneering and momentous activities of the Campanari family, tomb discoveries continued to be made around Tuscania, the Statlane tomb found in 1896 producing even more sarcophagi than the Vipinana tomb. The big find of more recent times was the three tombs of the Curuna family excavated in 1967–1970 at the

Madonna dell’Olivo necropolis 1  km south of the town (Moretti and Sgubini Moretti 1983). The sarcophagi and associated material from these and from other tombs are on display at the Museo Nazionale in the cloister of S. Maria del Riposo. In microcosm, Tuscania presents to the public the kind of Etruscan archaeology that is well known throughout Etruria, one that is very much cemetery-based. The visitor can see this material in the town’s attractive museum, and there are more sarcophagi laid out in the aisles of San Pietro church and dramatically crowning the walls of the Piazza del Comune (Fig. 1.8). The tombs from where all this material comes, however, are not themselves especially dramatic: at Tuscania they tend to be very plain rock-cut chambers, though some of the tombs discovered later do have considerable architectural and sculptural elaboration (Sgubini Moretti 1982, 1986a, 1989). However, these latter have been little publicized and today the tourist will probably only be directed to the very intricate series of chambers of the Tomba della Regina complex (of Hellenistic date) at the Madonna dell’Olivo 1 km south of the town.

figure 1.7  Restored tomb interior with sarcophagi from Tuscania, displayed at Pall Mall, London, in 1837. (After Pryce 1931, fig. 48.)

17

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

figure 1.8  Etruscan sarcophagi on the walls of Tuscania’s Piazza del Comune, with the Late Medieval walls and towers of the centro storico behind. (Photograph: Graeme Barker.)

At only 8.4 ha (hectares), Colle San Pietro was a rather small Etruscan settlement. The five major settlements of South Etruria – Veii, Caere (modern Cerveteri), Tarquinia, Vulci and Volsinii (modern Orvieto) – each measured between 100 ha and 200 ha. The Etruscan remains known from Tuscania before our project were also few. In the rescue excavations on Colle San Pietro after the earthquake, most early Etruscan material, beginning from the later eighth and early seventh centuries bc, was found out of its original context mainly mixed in with later Medieval deposits. Two rough chamber tombs, one with material from the later sixth to early fifth centuries bc, the other with black-glazed pottery from the end of the fourth into the third centuries bc, suggested that the hill itself, or part of the hill, was not a centre of habitation in these periods. Yet the Etruscan cemeteries around Tuscania are very extensive, prompting the question as to whether

the Etruscan centre was in fact somewhere else in the vicinity (Torelli 1993: 227). The Tuscania Archaeological Survey did in fact confirm that the San Pietro hill was the main centre of habitation, the Etruscan levels having been almost completely obliterated by the large-scale Medieval building works, but that settlement extended down the southern slopes to near the Marta river, where we found dense spreads of domestic Etruscan settlement material (see Chapter 5). Presumably the hill functioned primarily as an arx or acropolis for the Etruscan population living on its flanks. Clear evidence of sophisticated urban life in the Archaic Etruscan period (600–500 bc) is provided by figurative architectural terracottas. Some of these have been known for a long time, though their precise findspot is uncertain (Andrén 1940: 73). Others have been excavated, not on Colle San Pietro, but at the Ara del Tufo necropolis to the

18

the selection of tuscania

south (Sgubini Moretti 1982; Winter 2009: 561–562; Fig. 1.9), where they were scattered about some time after the tombs in question went out of use in the sixth century bc. Probably they had decorated one or more funerary shrines (naiskoi) at the necropolis. They show that Tuscania had wide contacts in this period: several types of antefix and revetment plaque from Ara del Tufo are matched by identical terracottas – from the same moulds, indeed – from the Etruscan settlement at Acquarossa north of Viterbo. Settlement at Tuscania in the Hellenistic period is indicated by remains of foundations of Late Republican houses with later black-glazed pottery and associated coarse wares. To the same period, or rather later still, belong the remains of a bath-building in the valley at the bottom of the northern slope of Colle San Pietro, which include walls in opus quadratum and opus reticulatum and traces of black-and-white mosaic (Quilici Gigli 1970:

163–167). A part of it, set against the foot of the Rivellino hill, is still visible. When in 89 bc the whole of Etruria was given Roman citizenship, Tuscania (called Tuscana in Roman times) was included in the Roman tribe Stellatina, along with a number of other towns such as Tarquinia and Graviscae (Harris 1971: 244). In the Roman Imperial period the town had the status of a municipium (Quilici Gigli 1970: 22, note 11; Pliny 3.5.52). Roman buildings covered much of Colle San Pietro, and a street laid with basalt blocks climbed the northern slope, lined with houses embellished with mosaic floors. However, this part of the hill at least was abandoned by the end of the fourth century ad, and there is no evidence of further occupation here until around three centuries later, when houses of wood were constructed along much the same alignment as previously, later to be rebuilt with stone walls.

figure 1.9  Circular tumuli of the Ara del Tufo Etruscan necropolis; looking north, with Tuscania in the distance. (Photograph: Graeme Barker.)

19

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

Yet Tuscania (Medieval Toscanella, a name that persisted until modern times) clearly continued to be a centre of some importance throughout this period, and it was the seat of a bishopric at least from ad 595 until ad 1192 when the latter was transferred to Viterbo, confirming that town’s growing superiority in status in relation to Tuscania (Andrews 1982: 138–139; and see Chapter 8). Around 1200 Colle San Pietro was given a circuit of walls, of which there are today very few traces, though some of the towers which were later built to reinforce it still remain. The great churches of San Pietro (Fig. 8.2) and Santa Maria Maggiore, the latter in the valley to the north-west of the Colle San Pietro hilltop but still within the early wall circuit (Fig. 1.6), seem to have been first built in the eighth or ninth century ad. Colle San Pietro, with its big towers for defence and habitation, together with smaller houses clustered around, appears to have been densely populated until the fourteenth century, by which time the population had begun to expand to the adjacent Rivellino ridge. Gradually Colle San Pietro was abandoned as a place of settlement, while the enhanced status of the Rivellino, its area considerably greater than that of the neighbouring hill, was confirmed by the final completion in the fifteenth century of the circuit of defensive walls around the present centro storico. Today, Tuscania is one of many small towns in the province of Viterbo, the large and bustling provincial capital. At the time of our survey, it had a paper mill (which closed in 2014) and one or two other smallscale industries, but its economic life was much centred on agriculture, an annual highlight of the entertainment calendar being a tractor race around the streets. Another feature of the known archaeological record of Tuscania that was important in persuading us to select the territory of the town for our project was the existing archaeological map of the town and the immediately surrounding area compiled by Stefania Quilici Gigli in the Forma Italiae series, published in 1970 (Figs. 2.6 and 7.5). The Forma Italiae is a long-running mapping project by classical archaeologists of the University of Rome, and the basis of each of its surveys is one of the 1:25,000 maps

of the modern Italian topographic grid. The mapping exercise is usually undertaken by a classical archaeologist working on their own for a research thesis, who compiles the map by a combination of exhaustive bibliographic research, the study of museum collections, discussions with local museum curators and landowners, and followup visits to known or suspected archaeological sites. As a compilation of material collected in different ways by different people at different times, it is akin to what used to be termed the Sites and Monuments Record of an English county. The best of the Forma Italiae maps, of which the Tuscana map is certainly one, are extremely informative, but the reliance on individual fieldwork and on the assembling of known information inevitably means a bias towards the most highly visible components of the archaeological landscape, Etruscan tombs being a prime example. The special focus of Quilici Gigli’s study was on the Etruscan cemeteries and tombs in the vicinity of Tuscania, which were carefully recorded and illustrated. The resulting map of the area was very informative on the position of these, as well as of the visible cuttings and other traces of ancient roads. There were also references in her commentary to ‘scatters of surface material’, though relatively few had been visited and their spatial characteristics recorded. The study was before knowledge of pottery styles had been refined, especially for later Roman pottery, so the description of most of these scatters was very general, with little attempt at periodization. Nevertheless, the Forma Italiae map gave us reasonable confidence that, fifteen or so years after its compilation, such sites were still going to be well enough preserved in the ploughsoil for us to map them. It also provided an ideal opportunity to compare this kind of archaeological map, compiled over the years by chance discoveries and individual researchers interested in particular topics or classes of material, with the data collected by systematic team-based fieldwork, in which material of all periods would be valued equally. Theoretically, the latter should be more effective than the former as a means of ‘writing archaeological history’, but would the reality on the ground bear this out?

20

the selection of tuscania

Tuscania’s Neighbours

here called Visentium. The decline from the seventh century bc onwards and the growth of Tuscania at the same time may not simply be coincidental. The closest ancient settlement of any size was situated to the east of Tuscania, at Musarna. This small (4.5  ha) foundation was a late one, laid out in a planned manner at the end of the fourth century bc (de Casanove and Jolivet 1983). The habitation site has been under excavation by the École Française de Rome, but from the Hellenistic cemeteries there are ambitious stone sarcophagi that have long been known and now fill up much of the Museo Civico at Viterbo. Musarna continued to support a population, even if a diminished one, through into the Late Roman period.

The second reason for selecting Tuscania was its location in terms of other ancient settlements in the region and its likely communication links with them (Fig. 1.2), given that it could be expected that the history of an ancient town’s relationship with its surrounding countryside would need to be studied in the context of its place within changing networks of power at the regional scale. Like many other ancient settlements in South Etruria, Tuscania was situated on a promontory flanked on both sides by streams, which at Tuscania run into the river Marta flowing at its foot. Clearly the advantages of the site lay in its proximity to the river and its valley, which from earliest times must have provided a north–south communications route. But there was another important route that crossed here too and led south-east to other Etruscan settlements such as Blera and Norchia. After the whole area became Roman, this road was systematized and properly laid with basalt blocks, perhaps as early as the third century, but in any case not later than 183 bc. Called then the Via Clodia, it ran from a junction of the Via Cassia, where that road was closest to the city of Veii, to the Roman colony of Saturnia (founded in 183 bc) above and to the west of Lake Bolsena. At Tuscania the Via Clodia crossed the river Marta just downstream from the present road-bridge, ran up the left slope of Colle San Pietro and on up the Rivellino hill; a short excavated section is visible here beyond the ruins of the Roman baths (Fig. 1.10). Etruscan Tuscania lay quite close to a number of other smallish settlements, but these did not all flourish at the same time. To the north was Bisenzio on the edge of Lake Bolsena (Babbi et al. 2019), a site extending to 35 ha over its various phases. Excavations on the hill here had uncovered Late Bronze Age buildings (Delpino 1982), while its cemeteries reveal considerable prosperity between the ninth and early seventh centuries bc, the Early Iron Age and early (‘Orientalizing’) Etruscan period. A decline followed, and there is no indication of settlement at all from the fifth century bc until Roman times when, according to literary and inscriptional evidence, there was a town

figure 1.10  A fragment of the Via Clodia Roman road at Tuscania. Scale: 10 cm. (Photograph: Tom Rasmussen.)

21

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

Still in the east, but further anti-clockwise from Tuscania, lie Castel d’Asso (11  ha) and Norchia (9.5  ha). Like Musarna, they occupy precipitous promontory positions, but on a rather larger scale. The rock-cut necropolises of both have been closely studied, but the settlements themselves have only been surveyed. At Castel d’Asso the town site seems to have been at its maximum extent in the Archaic Etruscan period (Colonna di Paolo and Colonna 1970: 52), but contemporary settlement at Norchia was only small-scale. At both sites, however, the necropolises, as so far explored, belong essentially to the Hellenistic period (after 300 bc), with no evidence of further use after the middle of the first century ad (Colonna di Paolo and Colonna 1978: 412). This fact in the case of Norchia is especially noteworthy, as the Via Clodia ran right across the settlement plateau. To the south and west respectively, Tarquinia (220 ha) and Vulci (190 ha) were major Etruscan cities and important centres of settlement from the Early Iron Age until the Roman occupation of 281–280 bc. After the granting of Roman citizenship in 89 bc, both continued as Roman municipia, though their populations must have been much reduced. At Vulci, many of the visible remains on the site are of Roman Imperial date, including a mithraeum that was in use through the third and fourth ­centuries ad (Carandini 1985b: 73). These two cities are equidistant from Tuscania, but the influence of Tarquinia on the development of Etruscan Tuscania and the territory around it was by far the greater. That is not much of a surprise given the topography of the region: the natural lines of communication run northwards along the Marta valley and the plateaus to either side, stretching up as far as Lake Bolsena, whereas communications eastwards from Vulci to Tuscania are across the grain of the country, hampered by numerous obstacles in the form of deep stream gulleys running north to south. Vulci’s line of influence was especially north along the valleys of the Fiora and its tributaries, again reaching as far as Lake Bolsena but from a different direction (Rendeli 1993a: 171). Significantly, the tomb material from the lakeside site of Bisenzio in the eighth century bc has noticeable stylistic similarities with Vulcentine products (Delpino

1977: 48; Sprenger and Bartoloni 1983: 77), while that from Fiora valley settlements such as Castro and Poggio Buco has affinities with Vulci in the Archaic period. In the Roman period Tarquinia seems to have been the key settlement of the area that includes Tuscania, and its territory, the Ager Tarquiniensis mentioned by Cicero a couple of times, seems to have been very extensive. Pliny (2.95) talks of a Lacus Tarquiniensis, which must refer to Lake Bolsena, for he mentions its two islands, which he describes as floating on the lake. But Tarquinia’s hold on the area was probably strongest in the period from the later fourth to the first centuries bc, when the status of Tuscania would seem to have been very much that of a dependency. Several factors suggest this. The beginning of the period saw the rise of great families such as the Curuna, whose three sarcophagus-filled tombs at Tuscania can be matched with another Curuna tomb at Tarquinia (Pallottino 1937: 515, 525, 544), from where it is likely that the family originated. In style and subject matter, there is often little to distinguish the carving of stone sarcophagi from Tuscania from those from Tarquinia, and no doubt Tarquinian sculptors were employed at both centres (Barker and Rasmussen 1998: 289). Moreover, Etruscan titles of magistracies found in funerary inscriptions not only at Tuscania but also at neighbouring centres such as Musarna (TLE 169–76) and Norchia may refer to offices with jurisdiction not just over these local communities but over the wider area controlled by Tarquinia. The river Marta, the only outlet of Lake Bolsena, reaches the sea just north of Tarquinia Lido. Tuscania, rather closer to the lake than to Etruscan Tarquinia, is the only modern settlement of any substance that lies beside the river. In previous times there were no doubt others, but none were of notable size. Perhaps one of the more important was Ancarano, a promontory site midway between Tuscania and Tarquinia, and around 10 km from both, where there are remains of Medieval defensive walls surrounding an inner citadel. There may also have been an Etruscan settlement here, for there are Etruscan tombs in the neighbourhood (Pallottino 1937: 581). On a spur opposite, across the river, are the ruins of Pian Fasciano, another small defended Medieval site

22

the selection of tuscania

of the thirteenth century and later (Andrews 1981: 324; Maggiore 2012: 230–231).

tuffs (tufo) and ignimbrites. In the survey area, the terrain within a 10 km radius of Tuscania, the plateau shelves gradually from the north-east at about 300 m above sea level to the south-west at around 100 m. Much of the volcanic plateau landscape, including that around Tuscania, looks benignly flat from a distance (Fig. 1.11) but is, in fact, dissected by rivers and streams which at frequent intervals gouge dramatic ravines into the soft tufo, their faces making favoured locations for Etruscan rock-cut tombs (Fig. 1.12). The principal streams in the Tuscania area are the Marta (which flows immediately past Colle San Pietro) and, in the west, the Arrone. The highest areas surround the crater lakes, the highest altitude being east of Viterbo at 1053 m above sea level at Monte Cimino, the summit of the hills forming the northern rim of the Lake Vico crater some 30  km east of Tuscania. Immediately north of the Tuscania Archaeological Survey area, the hills surrounding Lake Bolsena rise to about 600 m above sea level.

The Physical Landscape The third reason for the selection of Tuscania was the character of the surrounding landscape, which was typical of the wider region in its topography and land use, and demonstrably well suited to investigation by field survey. The geology of the area is discussed in detail in Chapter 3, but its principal characteristic is that it is dominated by volcanic formations. South Etruria consists predominantly of a recently formed volcanic landscape, one of the most extensive in the Mediterranean area, forming an undulating plateau punctuated by a series of crater lakes. The most northerly of the latter is Bolsena, which is also the largest, and was the main source of the lava and ash flows that underlie most of the Tuscania countryside as

figure 1.11  Looking north across the Tuscania Archaeological Survey area from near its southern boundary. Tuscania is in the distance, at the centre of the image. The land rises slowly behind it towards the hills edging Lake Bolsena. (Photograph: Graeme Barker.)

23

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

figure 1.12  The Marta valley to the north-east of Tuscania. (Photograph: Tom Rasmussen.)

The main forested area today is the Monti Cimini around Lake Vico. In the later fourth century bc, these mountains were renowned and feared by the Romans for their deep and impenetrable forests (Livy 9.36). Today there are tracts of tangled wood and macchia especially on the plateau to the east and south-east of Tuscania as well as in the stream and dry-valley gorges, but most of the plateau-land in the environs of the town is open and intensively farmed (Fig. 1.13), so we were not hampered by large areas of woodland inaccessible to field-walking though now amenable to investigation by airborne Lidar survey. Crops include cereals, lucerne, olives, vines and tobacco, but there is also some stock-raising of sheep and cattle (bufalo) especially on outcrops of cretaceous limestone and flysch. Farms and field systems are all the time increasing in size, and large-scale agro-industrial ranches were making their appearance at the time of the fieldwork. Ever larger swathes of the terrain are being

farmed for high-output crop management and, as in other parts of central Italy (e.g. Barker 1995a: 306), year by year the landscape is becoming increasingly eroded and prairie-like. The predominantly open character of the land obviously made it very suitable for investigation by field survey. Archaeologists’ access to land has been relatively easy in Italy with appropriate permission from the town authorities, with which field-walking teams can enter unfenced land as long as they do not cause any damage. Although increasing amounts of land around Tuscania were being fenced in by the big estates during the period of our fieldwork (1986–1990), most of the landscape remained accessible, though part of it north of the town has literally been quarried away for its tufo, which is in high demand as a building material (Fig. 1.14). As for the ploughlands, these are being subjected to an ever greater depth of excavation by ploughs pulled by tractors

24

project planning and development

Lake Bolsena

Tuscania

Montefiascone

Viterbo

Vulci

r ive

ne

ro

Ar

R

er

Riv

Lake Vico

a

rt Ma

Tarquinia 15

0 km

1

2

3

4

5

r ive

R

e

non

Mig

figure 1.13  Simplified patterns of recent land use in South Etruria: 1. drained land; 2. arable; 3. polyculture (cereals, olives, vines); 4. woodland; 5. pasture. (Adapted from the Carta dell’Utilizzazione del Suolo d’Italia, 1959: sheet 12.)

on caterpillar tracks, which are bulldozing the bedrock and bringing it to the surface (Fig. 1.15). All these factors together made our survey intervention an especially timely one. Far less land would be available to survey now than when we conducted the fieldwork, and it is highly unlikely that the quality of surface finds today would be nearly as good as it was for us.

interests in Etruscan archaeology, was keen to see a resumption of survey in South Etruria that would cast light on the least-understood component of Etruscan culture, rural settlement. Both of us had in fact been doctoral research students together at the British School at Rome in the early 1970s, kept in friendly check by John Ward-Perkins in his final years as director. Like most students who passed through the School at that time, we had on various occasions either volunteered, or been volunteered, to help with the South Etruria Survey fieldwork. We planned a programme of five seasons of fieldwork, which took place each September from 1986 to 1990, followed by a further month’s study season of the survey finds in 1991. The geomorphological fieldwork described in Chapter 3 and the detailed analysis of the major classes of finds was undertaken during the 1990s (Brown and Ellis 1996; MacDonald 1999; Rendeli 1993a).

project planning and development The project was originally planned as a joint collaboration between Graeme Barker and Tom Rasmussen. GB, as Director of the British School at Rome at that time, was keen to develop a field project in South Etruria to build on the tradition established by the British School’s South Etruria Survey in the 1950s and 1960s and his own Biferno Valley Survey. TR, with primary research

25

1 THE TUSCANIA ARCHAEOLO GICAL SURVEY

figure 1.14  Tufo stone quarry north of Tuscania. (Photograph: Tom Rasmussen.)

Like most archaeological field projects, the Tuscania Archaeological Survey did not proceed to an exact prearranged plan but developed organically as discoveries fed back into theories and methodologies, and as new personnel brought fresh perspectives and research interests. Perhaps the most intellectually stimulating feature of archaeological fieldwork, in fact, is this interaction between theory, method and discovery on the one hand, and between overall project goals and individual research agendas of team members on the other. Annie Grant joined the project after the first two seasons and developed the databases for the field survey and finds cataloguing systems. Chris Hunt initiated the work on the history of the natural environment of the study area with a palynological study, and the palaeoenvironmental programme was then continued and enlarged by a team led by Tony Brown. A variety of specialists was involved in the study of the survey finds. The prehistoric material

was studied by GB and Francesco di Gennaro (pottery) and Tim Reynolds (lithics), the Etruscan material by TR and Marco Rendeli, the prolific Roman material by Nick Whitehead and Phil Perkins in the first two seasons and then the whole data set by Alison MacDonald for her Oxford DPhil (MacDonald 1999), and the Medieval and Post-Medieval material by Helen Patterson. Nicoletta Vullo made an important contribution by bringing GIS (Geographical Information Systems) to bear on the entire data set, as described in Chapters 2 and 9.

conclusion The Tuscania Archaeological Survey was planned as a contribution to Mediterranean landscape history, especially town–country relations, by focusing on the archaeology of the countryside around a small town in South Etruria near Rome. Chance finds and systematic work

26

conclusion

figure 1.15  Deep-ploughing technology used in the countryside around Tuscania at the time of the fieldwork. (Photograph: Graeme Barker.)

on Tuscania’s archaeology previous to our project indicated that the site had probably been continuously occupied since urbanism first began in Italy with the Etruscan city states almost 3000 years ago. The archaeology in the immediate environs of the town, especially Etruscan tombs, had been mapped previously in the Forma Italiae series, a study that also indicated that other kinds of archaeological sites would be found by modern systematic survey, in the process providing us with a good opportunity to compare results. The terrain was well suited to the techniques of field survey, with much land under the plough and accessible to archaeological survey teams. We hoped that the project would not only make use of

current survey methodologies but also contribute to their evaluation and improvement. Our overarching aim was to use the techniques of landscape archaeology to write a long-term archaeological history from a holistic perspective, illuminating major themes such as relations between people and environment, between town and countryside, between Tuscania and the wider world, patterns of settlement, economic structures and the like – and in the process, revealing the successive ‘taskscapes’ that the lives and works (in Ingold’s phrase) of past peoples had created in this landscape, taskscapes that would also have been shaped by the lives and works of past generations just as they are for us today.

27

2 METHOD OLO GIES Graeme Barker, Tom Rasmussen, Alison MacDonald, Annie Grant and Nicoletta Vullo

introduction

that it was the primary focus of Etruscan, Roman and Medieval settlement at Tuscania (Chapter 1). Before the days of motor transport, the distance travelled by most Mediterranean farmers to their fields, either on foot or riding a donkey, mule or horse, was rarely more than 5  km. A classic study of the pattern of radial land use around an Italian hill village found that the orchards and gardens requiring the most time and effort in their cultivation tended to be closest to the settlement and the arable land not more than a few kilometres away, with land further afield used for grazing flocks and herds (Chisholm 1968). Such an ‘agricultural territory’ around Tuscania, however, need not have equated with other spheres of influence. We knew that there were periods in Tuscania’s history when the town was the political centre of a larger territory, as, for example, in the twelfth century (Pierdomenico 1974: 30; and see Chapter 8), or was under the political control of another centre, as was the case with its relations to Tarquinia in much of the Etruscan period, and then to Rome. Any settlement is in fact likely to be located within a series of overlapping territories of different sizes – economic, social, political, ideological. Thus modern Tuscania is an agro-town from which farmers travel to their fields up to about 5 km away; it is a market centre for a larger area, about 10 km; and it belongs to the administrative territory of the provincial capital Viterbo and to Rome as the capital of the Lazio regione, to say nothing of its relationship to Rome in the latter’s role as an international city (now within daily commuting distance of Tuscania), communications hub and seat of the Vatican church. Without prejudging the size of Tuscania’s hinterland, therefore, or the nature of changing relationships between the town and its countryside through time, we decided to make the focus of the investigation the area within

It is often stated that one of the major advantages of survey over excavation is that survey is repeatable, whereas excavation is a once-only destructive process. However, repeat surveys have in fact been surprisingly rare, and those that were undertaken in the years before and in parallel with the Tuscania Archaeological Survey, including experimental studies of fields seeded with mock artefact scatters, had sometimes shown considerable variations in the surface archaeology from year to year (e.g. Ammerman 1985a, 1985b; Ammerman and Feldman 1978; Cherry 1983, 1984; Shennan 1985; Terrenato and Ammerman 1996). The variability in the results obtained by different projects working in similar terrain has prompted enduring debates about the reliability of field survey data and the effectiveness of different methodologies, especially in Mediterranean contexts (e.g. Alcock and Cherry 2004; Barker 1991, 1994; Barker and Lloyd 1991; Barker and Mattingly 2000a–e; Bintliff 2000; Bintliff et al. 2000, 2007, 2017; Cherry et al. 1991; Mattingly 2009; Witcher 2006, 2008). Having described in the previous chapter our aims and objectives in planning the Tuscania Archaeological Survey, in this chapter we discuss the decisions we took about our approach, the methodologies we selected, and our assessment of their strengths and weaknesses.

defining the study area Given our interest in investigating the relations between Tuscania and its hinterland, we needed first to define the size of the territory we would investigate around the town. For this purpose we took the Colle San Pietro acropolis (Figs. 1.4, 1.6) as the centre of the study area, given that chance discoveries and previous excavations had shown

28

defining the study area

10 km of Colle San Pietro. We were mindful in particular of the distances to neighbouring settlements of any size with evidence for ancient occupation (Fig. 1.2). The nearest such settlement to the north is Bisenzio on the edge of Lake Bolsena about 17 km from Tuscania. To the east are Musarna and Castel d’Asso, at 8 km and 12 km respectively. The Etruscan necropolis and settlement of Norchia lie 10 km to the south-east. The Etruscan city of Tarquinia was situated 22 km to the south-west and its rival Vulci 20 km to the west. While we acknowledged that Tuscania might sometimes in the past have been an administrative centre servicing or controlling a larger area, and at other times might have been under the political control of much more powerful centres in its region (both of which were in fact the case), we decided that a 10 km radius should encompass a sufficiently large area to enable us to investigate the nature of relations between the town and its countryside at different times in the past, without either excluding a large part of the countryside that might have been within the town’s ambit, or including too much of the equivalent hinterlands of neighbouring settlements. Rather than setting the boundary precisely at a 10 km radius, creating a circular survey area, for ease of working we set it using the grid of the main topographic map cover in Italy of the IGM (the Istituto Geografico Militare, the equivalent of the Ordnance Survey in Britain), at the 1:25,000 scale. The IGM grid is at vertical and horizontal kilometre intervals, so we set the boundary of the study area by including those square kilometres in the boundary zone which had more than half their area within the 10  km radius and excluding those with more than half their area outside this distance. The result was the geometric shape shown in Figure 2.1, an area of 353 km2 with Tuscania at the centre. The area encompasses almost all of two IGM 1:25,000 maps, Tuscania (136 II NE) to the north and La Rocca (136 II SE) to the south, substantial portions of Canino (136 II NO) to the north-west, Commenda (137 III NO) to the north-east, San Giuliano (136 II SO) to the south-west and Castel d’Asso (137 III SO) to the southeast, as well as just intruding into Valentano (136 I SO) and Capodimonte (136 I SE) on the northern margin and Monte Romano (142 I NE) on the southern margin.

The topography of the selected study area consists of a gradually sloping plateau (Figs. 1.11, 2.2), about 300 m above sea level at the north-east towards Lake Bolsena and about 100  m above sea level at the south-west in the direction of Tarquinia and the coast. The plateau is dissected by the Arrone and Marta streams (Fig. 1.12) and their tributaries, many of the latter being dry torrent beds for most of the year. These valleys flow generally from north-east (from Lake Bolsena) to south-west across the survey area, making the natural lines of communication also in that direction, whether along the plateau tops or down the valleys. As described in Chapter 1, the geology of the plateau is predominantly volcanic (tuffs and ignimbrites) though there are also outcrops of sedimentary rocks (travertine and flysch) and marine sands and clays (Fig. 2.3). The wider stream valleys are floored with alluvial sediments, the age of which is discussed in Chapter 3. In terms of modern settlement, apart from Tuscania the only significant centres of population are the small villages of Arlena and Tessenano to the north-west. A series of modern roads radiates outwards from Tuscania around the compass in a clockwise direction from south-west (road SP3 to Tarquinia) to south-east (road SP11 along the line of the Via Clodia that joins the modern Via Cassia at Vetralla, the main route to Rome). There is no main communication route south, however. A minor road (a gravel strada bianca at the time of the fieldwork) runs in this direction from Tuscania for about 5 km and then becomes a farm track as far as the Marta. There is no vehicular crossing of the Marta downstream from the Tarquinia-Vetralla road until Tarquinia, so the valley remains a significant barrier to southwards movement from Tuscania. Before mechanization and agricultural subsidies favouring monoculture, the agricultural landscape of South Etruria was much more varied. Figure 1.13 shows simplified land use in this part of South Etruria according to the Carta dell’Utilizzazione del Suolo published by the Touring Club Italiano in the 1950s and 1960s, produced at 1:200,000 scale from vertical air photographs at that time. Instead of the huge tracts of cereal land, or olive groves or vineyards of today, there was a mosaic landscape with small areas of different kinds of crops. Many fields were used for ‘polyculture’, or mixed cropping: olives were

29

2 Methodologies

Orvieto Volsini Mountains

Sovana

Ti be r

Bolsena Lake Bolsena

Bisenzio

2

Fiora

1 3

4

5

TUSCANIA

Vulci

7

8

Castel d’Asso

Cimini Mountains

Ar

ro

ne

6

Musarna

Norchia

9 ta

ar

M

Blera

Tarquinia

1 2 3 4 5 6 7 8 9

136 I SO Valentano 136 I SE Capodimonte 136 II NO Canino 136 II NE Tuscania 137 III NO Commenda 136 II SO San Giuliano 136 II SE La Rocca 137 III SO Castel d’Asso 142 I NE Monte Romano

Mig

non

e

Lake Bracciano Tolfa Mountains

Land over 500 m

Cerveteri

20

0

Lake Vico

km

figure 2.1  The location of the Tuscania Archaeological Survey area on the 1:25,000 topographic maps of the Istituto Geografico Militare (IGM).

planted in widely spaced rows with rows of vines in between, and cereals were grown in between, a system of land use recorded by the Roman writers which is effective at conserving moisture (the roots grow to different depths, and the vines provide shade for the seed crops) but which is also impractical for modern mechanized farming. There were also extensive areas of pasture along the river valleys, especially in the lower tracts of the Marta

and Arrone, many of which had been cleared and drained for crops by the time of the fieldwork.

sampling strategy The next decision concerned whether we could investigate all 353 km2 in terms of the resources likely to be available and, if not, how much of the survey area we

30

chio Capec

300 m

Pisciacello

sampling strategy

Pantacciano

Marta

aF ont

200 m

arell Acqu

Maschiolo

Can

a

tina c

cia/ dell

Arrone m

200 m

0

20

Arlena

e

300 m

Canino

Le Tu far e

Tessennano

o

Musarna

pe

cc

hi

Vi co

m

Ca

di

0

10

ia

10

Fabbrichetta

Do

gan

e

Pi

an

100

m

TUSCANIA

Biedano

Marta

m 0 10

Mignattara

0m

Le

Norchia 0

5 km

figure 2.2  The 354 km2 study area defined for the Tuscania Archaeological Survey (the original 353 km2 defined within 10 km of Tuscania, plus an additional 1 km2 surveyed on the north): topography, watercourses and principal settlements.

could investigate, and how to select a representative ­portion of it. The project was planned to consist of five years of fieldwork, a period we regarded as long enough to allow

a reasonably detailed study without turning into the kind of open-ended commitment that funding bodies resist. The principal season of field-walking in the Mediterranean is September, for four reasons. First, the

31

2 Methodologies

TUSCANIA

Marine Tertiary clays and sands

0

5 km

Flysch Tuff Ignimbrite Travertine Recent alluvium

figure 2.3  Simplified geology of the Tuscania Archaeological Survey area, showing the dominance of volcanic sediments (tuffs and ignimbrites). (GIS map prepared by Nicoletta Vullo adapted from the Carta Geologica d’Italia, Foglio 136 Tuscania.)

32

sampling strategy

cereal land is ploughed by then after the summer’s harvest, making the period suitable for archaeological survey in terms of accessibility: there is no danger of damage to that year’s crop, as it has been harvested, and the seeds of the following year’s crops have not yet been sown. Second, as the land has been recently ploughed but (for the most part) not yet harrowed, the conditions are good for revealing archaeological artefacts on the soil surface. Third, the heat of the summer is beginning to dissipate, making working conditions easier and more efficient. Finally, September (or at least most of it) is within most European university summer vacations, so staff and students can be available for fieldwork. Within the five-year timescale, we therefore could plan for some 15–20 weeks of fieldwork, or 90–120 working days using a six-day week. Given the likely scale of the funding we would probably obtain, the necessity for close management and supervision of the field teams in terms of quality control and logistical factors such as vehicle availability and accommodation requirements, we planned for annual teams of a maximum of twenty people, though the initial season concerned with trialling procedures was to be smaller. Allowing for the accommodation needs of specialists remaining at the project’s base to work on the finds, the intention was to have two or preferably three teams in the field each day, our previous experience (e.g. Barker 1995a; Barker et al. 1986) having shown that teams of four–five members were generally preferable to smaller or larger teams in terms of accuracy, ground coverage, control of line-walking procedures and so on. For reasons discussed below, we planned to set the normal distance between walkers at 15  m. Given the time that teams needed to spend each day getting to and from a particular area, defining walking units, walking systematically and slowly across them, collecting surface material, and recording their procedures on standard record sheets, we reckoned that we would be lucky if two–three teams mounted each day could together cover a square kilometre of the open ploughed areas that dominated the countryside around Tuscania. Allowing for the certainty of delays from the weather, logistical problems, unexpected discoveries and

so on, it seemed likely that we would be able to investigate something in the order of 100 km2 of terrain at best. It was clear, therefore, that while in terms of the project’s research goals we aspired to collect high-quality settlement information from a study area measuring over 350 km2, we were unlikely to be able to cover even a third of it, and that we would need to collect information instead from a sample area. Ever since the New Archaeology of the early 1960s, the rationale for sampling in field survey had been much debated (e.g. Binford 1964, 1965; Flannery 1976; Plog 1976; Plog et al. 1978; Schiffer et al. 1978). Most archaeologists agreed that 100 per cent coverage remained the ideal, but if the research goals necessitate gathering information from a larger area than can be investigated with the resources likely to be available, then the strategy must be to attempt to collect information from a sample of land that is representative of the total study area. If the sample is representative, then in theory – as with a political poll – the results from the ‘sample population’ should be a good guide to the nature of the total or ‘target’ population. One common approach in Mediterranean field surveys prior to our fieldwork was to work within a series of transects spaced at regular intervals. This technique was used, for example, in the surveys of the Ager Tarraconensis in eastern Spain (Carreté et al. 1995), the Albegna valley to the west of Lake Bolsena (Attolini et al. 1991), the island of Melos in the Aegean (Renfrew and Wagstaff 1982) and in the Troodos Mountains of central Cyprus (Given et al. 1999, 2013). The Nepi Survey to the east of Tuscania, undertaken after our project, employed cardinal 5  kmlong transects centred on the town of Nepi (Mills and Rajala 2011; Rajala 2013). The selection of regularly spaced blocks of terrain within a grid, as we used in the later phase of the Biferno Valley Survey (Barker 1995a), worked to a similar principle, of collecting information at regular intervals across a landscape. A combination of the two approaches had been used in the Cecina valley, to the north of the Albegna valley in central Italy (Regoli 1992; Terrenato 1992). Both techniques have usually been employed when the archaeologists believe that the size and configuration of the sample units would be sufficient

33

2 Methodologies

to encompass a representative range of the natural features (landforms, topographies, vegetation and so on) of the total study area, on the assumption that the different features would be likely to produce different kinds of data. Another approach was to select areas on a statistically random basis. This technique had been recommended, for example, if the terrain was rather uniform, on the principle that a ‘probabilistic sample’ (say, 10 per cent) of the study area selected randomly should provide a representative guide to the diversity of material on the ground. A more sophisticated version of this was ‘stratified random sampling’, for example in cases where the study area consisted of two or three distinct landforms. In this system, separate grids would be placed over each of the constituent landforms (say, an alluvial floodplain, adjoining river terraces and adjacent hills), and comparable random samples then taken from each type. The final sampling technique was sometimes termed ‘judgement’ or ‘judgemental’. At the extreme, this may mean simply following one’s nose or visiting promising locations identified by local people, or looking in those parts of the survey area assumed from previous experience to be likely to be more productive than others. The classic example would be setting out to map Roman settlement in an area by looking for Roman settlements on either side of a Roman road, resulting in the self-­fulfilling prophecy that Romans lived mainly by their roads. However, this kind of selective survey has its place in sampling methodologies when used with the clear recognition of what it can show about settlement distributions and what it cannot. In many parts of the world, for example, prehistoric sites of any antiquity are likely to be found in only particular geomorphological ‘windows’ exposing surfaces of the period in question. Two examples of this in Italy are buried terramare settlements in the Po plain (Cremaschi 1997, 2010) and Neolithic settlements on sand dunes in Calabria (Ammerman 1985a, 1995). Palaeolithic surveys have often focused on searching for human occupation caves and rock shelters along promising valley scarps and outcrops. The ephemeral traces of Early Medieval settlements in Italy had sometimes been found by survey teams only when they investigated wooded hilltops where

documentary evidence indicated the likelihood of a settlement (Hodges 1982; Moreland 1986, 1987; Moreland and Pluciennik 1991; Moreland et al. 1993). In the case of the Tuscania Archaeological Survey, we ended up using a combination of the first three methods (Fig. 2.4), though in the final days of the last season we also visited four locations that documentary records suggested might have been occupied in the Early Medieval period (hence ‘Documentary Sites’), one of which, outside the survey area, yielded datable material (Table 8.2). We began by working within a series of four transects laid out on the IGM grid of kilometre squares. Tuscania is conveniently situated at the centre of a grid square, so we were able to lay out four transects using the IGM grid, each 1 km wide and 10 km long, running north, east, south and west of the town. Excluding the central square largely covered by the modern town, the Transect Sample (as we termed it) consisted of 40 km2, about 11 per cent of the total survey area. The use of transects commended itself at the beginning of the project in part because we were able to work outwards from the town, concentrating on the first 5 km north and south in the first season. This made for easy logistics, of course, but it also provided us with a first indication of the nature of the archaeology surviving in the 1980s in the immediate vicinity of the town, including in the area of the 1960s Forma Italiae survey (Quilici Gigli 1970). For the convenience of the recording system that we intended to use at the start of the survey (see below), we gave each 1:25,000 map a one- or two-letter prefix and numbered each square in it from top left to bottom right; each 1:25,000 map contains about 100 squares in its grid. Tuscania itself fell in square 84 of the Tuscania IGM map and hence was numbered as square T84, the adjacent square to the north became T74, and the one below it, number 14 in the La Rocca sequence, became R14 (Fig. 2.5). The Valentino map was not touched by the transects, the West Transect squares were mainly in the Canino map rather than the San Giuliano map, and the East Transect squares fell almost entirely in the Commenda map rather than the Castel d’Asso map, so the prefixes used for the Transect Sample were CP (Capodimonte), C (Canino), T

34

Arrone

300 m

Marta

Capecchio

sampling strategy

200 m

0

10

m

Tuscania (San Pietro)

10

0

m

10 0m

Transect Sample

0

10

Random Sample

km

Judgement Sample

Composite Sample

figure 2.4  The three landscape sampling strategies used by the Tuscania Archaeological Survey: ‘Transect’, ‘Random’ and ‘Judgement’.

35

2 Methodologies

(Tuscania), CC (Commenda) and R (La Rocca). The southernmost square of the South Transect fell in the Monte Romano map, but for the sake of simplicity we included it in the La Rocca sequence. Hence the Transect Sample consisted of the following square kilometres moving outwards from Tuscania: North Transect – T74, T64, T54, T44, T34, T24, T14, T4, CP94 and CP84; East Transect – T85, T86, T87, T88, T89, T90, CC10, CC11, CC12 and CC13; South Transect – R14, R24, R34, R44, R54, R64, R74, R84, R94 and R104; and West Transect – T83, T82, T81, C101, C100, C99, C98, C97, C96 and C95. The Transect Sample was completed in the first, second and third seasons of fieldwork. Although the Transects lie almost entirely on the volcanic geological formations that dominate the study area and for the most part proved extremely productive in terms of archaeological materials, they also had certain disadvantages in terms of their representativeness of the landscape. One of the features of the northern and southern transects is that they traverse the plateaus parallel to major stream valleys, so while giving us good coverage of the plateaus and plateau edges, they gave much less information about the nature of the archaeology in valley sides and bottoms. The western and eastern transects both crossed a series of stream valleys, but in the eastern transect, particularly in the sixth to tenth squares from the town, ground coverage was greatly restricted by woodland. Hence, we decided to complement the Transect Sample with a second sample of approximately the same size, of blocks of terrain (km2 again) selected on a statistically random basis. Given the overwhelming predominance of tufo geology, and the gradually shelving nature of the topography from north to south, it did not seem necessary to divide up the study area into major landforms in order to generate the kind of stratified random sample mentioned above. We therefore numbered each of the 353 squares in the study area consecutively from the top of the map to the bottom, so the top row consisted of squares 1–7 (left to right) and the bottom row of squares 347–353 (left to right). We then used random numbers tables to identify 40 squares between the numbers 1 and 353, to make up what we termed the Random Sample (squares RS1–RS40).

Three of the squares in the south-east quadrant of the survey area turned out to be problematical in terms of access: most of RS40 and all of RS37 and RS39 fell in what transpired to be an army firing range for heavy artillery! We were eventually able to gain access to RS40 on a non-­firing day, though because its use for crops and pasture had long since ceased, the area was extremely overgrown. It proved impossible to get permission to go further into the firing range to survey RS37 and RS39, so we used random numbers until we had generated two new RS37 and RS39 squares elsewhere in the south-east quadrant. Four of the Random Squares (RS2, RS10, RS21 and RS31) fell on the West Transect, but the rest fell outside the area covered by the four transects, so for comparative purposes (excluding the central square) we had two samples of equal size, of 40 km2 each, though in combination they added up to 76 km2. The terrain delineated by the Random Sample was investigated in the third, fourth and final seasons of fieldwork. In the final season, we added a third component to the sampling strategy, selecting parts of the landscape as a Judgement Sample to provide information complementary to that of the Transect and Random Samples. In particular, we wanted to see how the information from unbroken blocks of terrain – akin to the 100 per cent ‘total sample blocks’ used, for example, in the long-running Boeotia Survey (Bintliff and Snodgrass 1985; and since our fieldwork: Bintliff et al. 2007, 2017) – compared with that of the Transect and Random Samples. We could also see that certain parts of the survey area were poorly represented by the other two methods. Hence, we selected twenty ‘Judgement Squares’ (designated squares J1–J20) to create blocks of terrain either on their own or in combination with squares already investigated. These comprised just under 6 per cent of the total survey area. The main group of Judgement Squares was located to create a block at the corner created by the intersection of the southern and eastern transects: adding fourteen Judgement Squares here to the Transect and Random Squares created a block measuring 4 km by 7 km, or 28 km2, as large as the total study areas of some Mediterranean surveys. One particular reason for creating the block here

36

sampling strategy

J19 CP84

Capec chio

CP94 T4 RS6

RS16

T14 RS35

300 m Arrone

T24 RS36

RS3

RS15RS28

RS30 RS20 J20

T74

R14 RS33 J5

J6

R24

J2

J7

J8

R34

J3

J9

J10

R44

J4

J14 J13

RS25

pe

cc

hi

o

m

RS29

T83 T84 T85 T86 T87 T88 T89 T90 CC10CC11CC12CC13

Ca

m 100

RS14

T64

RS7

C95 C96 C97 C98 C99 C100 C101 T81 T82 0

J1

RS22 T54 200 m

RS23

RS9

10

RS1

T44

m

RS32

RS34

200 m

0

20

T34

Marta

RS38

RS19

300 m

RS5

J16 J17 J18

RS26RS13 RS12

RS37

0m

10

R54 RS39RS17 J12 R64 J15 RS11 J11

RS4

R74

RS18

R84 RS40

10

0

m

RS27

RS8

RS24

R94 R104

0

5 km

figure 2.5  The Tuscania Archaeological Survey, showing the identifiers of the 97 km2 selected for investigation.

was that this area was part of the 1970 Forma Italiae map, so it gave us better potential to evaluate the different kinds of information produced by the Forma Italiae methodology and our own (Fig. 2.6). Another reason was that the line of the ancient Via Clodia (Fig. 1.10) traversed this

part of the study area from south-east to north-west, so we hoped that the block would allow us to investigate the impact of the road on ancient settlement in its immediate vicinity in comparison with 2–3  km away (avoiding the circular reasoning mentioned earlier!). Another block of

37

2 Methodologies

8 km2 was created by adding a Judgement Square (J1) to the north of Tuscania where three Random Squares (RS6, RS35, RS36) lay adjacent to part of the North Transect (T4, T14, T24, T34). We also used three (J16, J17, J18) to create a horizontal transect to the south-west of Tuscania in an area where no Random Squares had fallen, placed across the line of the assumed ancient route to Tarquinia. Another square (J19) was selected in the closing days of the final season just outside the northern edge of the study area because Dennis (1848) mentioned a large and important site in the area (which we failed to locate). The effect of this was to make the total survey area 354 km2 rather than 353 km2. In total, therefore, we collected information from a Transect Sample of 40 km2, a Random Sample of 40 km2 and a Judgement Sample of 20 km2, plus the central square where Tuscania is located, making 101 km2, though because of the four Random Squares falling on the West Transect, our total sample consists of 97 km2, or 27 per cent of the total study area of 354 km2. In fact, the GIS analysis of the survey data by NV after the completion of the fieldwork demonstrated that we actually searched about 41.5 km2 or 12 per cent of the total area within 10 km of the town (Fig. 2.6). Even so, this area is one of the largest investigated systematically and intensively by a Mediterranean survey project. As a second exercise, the GIS was used by NV to investigate the extent to which the three landscape samples were representative of the survey area as a whole. The factors assessed were elevation, slope, aspect, geology and distance from Tuscania. The presence of these factors was calculated in percentages for the whole survey area, in the three samples separately and combined, and in the areas actually walked in the three samples separately and combined. A good level of representativeness would be indicated by similar percentages, with lower and higher percentages indicating respectively over- and under-­ representation. The plateau around Tuscania is rather homogeneous: the GIS analysis showed that more than 60% of the survey area has an elevation of between 100 m and 200 m above sea level, more than 80% has a very modest gradient (0–5 degrees) and more than 70% has a geology of volcanic origin, either pyroclastic tuff or

extrusive volcanics. Ideally a good portion of uncommon characteristics should be included in all samples as well as the dominant characteristics. This analysis showed a good level of representativeness of the ‘sampled’ (i.e. formally designated) areas and ‘surveyed’ (i.e. actually walked) areas with respect to the study area as a whole, with little variability between most values for sampled and surveyed areas. Positive and negative skews were most often associated with the Judgement Sample, particularly in the case of elevation and geology: in both cases the Judgement Sample tends to overrepresent the most common characteristics (elevation between 119 m and 215 m and tuff) at the expense of the least common ones. Almost no areas of travertine or flysch were included in the Judgement Sample, flysch in fact being found almost exclusively in the Random Sample. Nevertheless, despite these differences in geographical characteristics and their proportions in the three samples, the GIS analysis demonstrated that the combined sample is well representative of the survey area as a whole, with all the geographical characteristics of the Tuscania landscape being included in good proportion.

site and off-site/non-site archaeology Walking around the Tuscania countryside today, the visitor can see many examples of traditional farms: substantial farmhouses, barns and ancillary buildings constructed of tufo block walls and tiled roofs (Fig. 2.7a). Looking at them as archaeologists, we can imagine how, once abandoned, they will gradually collapse from the impact of time and weather: roofs will fall in, walls collapse, organic materials disintegrate, metal objects rust away, and while fragments of walls and floors might end up covered with soil blown in by wind or washed in by rain, what will be left on the surface for future archaeologists to find will be bits and pieces of the most indestructible kinds of rubbish such as broken potsherds and building debris. Nowadays, as in the past, most Mediterranean survey projects have a primary concern with mapping the distribution of archaeological sites that appear like this,

38

site and off-site/non-site archaeology

Capec chio

300 m

Arrone

Marta

300 m

Forma Italiae Survey Area

m

200 m

0 20

o hi cc

m

pe

0

10

Ca

100

m

200 m

Marta

10 0

m

0m

10 0

5 km

figure 2.6  The composite sample of 97 km2 (Transect, Random and Judgement Squares) selected for field-walking by the Tuscania Archaeological Survey showing (shading) the areas actually searched, amounting to 41.5 km2 or 12% of the total area of 354 km2. The area covered by the Forma Italiae survey is also shown. (GIS computation and mapping by Nicoletta Vullo.)

39

2 Methodologies

a

b

c

figure 2.7  Modern rural buildings in the Tuscania area at the time of the fieldwork, which will create very different kinds of archaeological ‘sites’ for future survey archaeologists to find: (a) farm; (b) field store; (c) shepherds’ milking ‘parlour’. (Photographs: Graeme Barker.)

40

site and off-site/non-site archaeology

as distinct concentrations of artefacts in the ploughsoil, that are assumed (and have often been proved by excavation) to be the residues of ancient settlements (Fig. 2.8). Methodologies have been developed for using differences in the size and artefact density of such concentrations, and in the kinds of artefacts represented, to distinguish between richer and poorer types of domestic habitation, as, for example, between the villa buildings of Roman rural estates and smaller farms, and also for distinguishing funerary structures from habitation structures (e.g. Alcock 1993; Carreté et al. 1995; Cherry et al. 1991; Launaro 2011; Lloyd 1995a, 1995b; Potter 1979; Whitelaw 2000; and see Chapters 6 and 7). Making distinctions between structures of varying functions is not an easy matter. One Roman Republican building meticulously excavated at San Martino in southern Tuscany by the Roman Peasant Project team was assumed from its surface remains to be a small farmhouse but was later interpreted as a temporary

stall for housing sheep and goats (Bowes 2020: 163–184). The Tuscania Archaeological Survey was typical of most Mediterranean surveys in its concern with locating and recording the distributions of significant artefact concentrations and making systematic collections of their surface artefacts, as a way of studying changing patterns of rural settlement through time. However, the same walk around the Tuscania countryside also reveals that it does not consist simply of farms on the one hand and emptiness on the other, the mixture of dots and white space that constitutes most archaeological distribution maps (Campana 2018; Witcher 2006). There is a wide range of small buildings, for example, scattered over the landscape, some used for storing equipment, others for animal feed or animals or as shelters for shepherds and herders (Fig. 2.7b). Some of these are made of stone and tile, others of organic materials such as reeds and thatch. The ‘archaeological signature’ created by one

figure 2.8  A dense scatter of pottery and tile fragments in the ploughsoil, marking the location of a partly ploughed out Roman farm (RS27:10). Looking north. (Photograph: Graeme Barker.)

41

2 Method olo gies

at scenic spots beside an official sign saying divieto di scarico – ‘no rubbish dumping’ (Fig. 2.9)! Many survey archaeologists have therefore endeavoured to embrace the potential complexity of the archaeological landscapes they are studying by investigating not just the main artefact concentrations assumed mostly to be ancient occupation sites or ploughed-out burials, but also the more diffuse or sporadic material across the landscape generally referred to as the ‘off-site’ or ‘non-site’ archaeology (e.g. Bintliff 1991, 1992, 2000; Bintliff and Snodgrass 1988; Bintliff et al. 2007; Bintliff et al. 2017; Cherry et al. 1991; Coccia and Mattingly 1992; Gallant 1986; Given et al. 1999, 2013; Haselgrove et al. 1985). If we liken the archaeological landscape to a three-dimensional contour map, we must imagine it consisting not just of major peaks (sites) in the midst of a featureless plain, but as a complex relief of major peaks, minor hills, ridges, undulations, depressions and so on, evidence of different kinds of past human

signature’ created by one shepherd we came across consisted of, at his main camp, a pen made of old bedsteads and pieces of fencing demarcating a circular area of dung, a ‘milking parlour’ made of a few sheets of corrugated tin (Fig. 2.7c) and an abandoned car containing some casualty lambs (Fig. 2.9), while scattered several kilometres away from the camp were his lunchtime picnic spots, marked by a few blackened stones and a couple of empty tin cans, where he rested with his flock in the heat of the day. There is a similar spectrum of archaeological signatures of religious faith, from the great church of San Pietro dominating the Tuscania skyline to the humblest wayside shrines, to say nothing of lost religious artefacts we might find on the roadside, such as a broken crucifix necklace. And the countryside is littered with abandoned material culture, of course, from village rubbish dumps to abandoned farm equipment and household rubbish, the latter all too frequently

figure 2.9  A long-abandoned Fiat Seicento near Tuscania, slowly being enveloped in vegetation. (Photograph: Graeme Barker.)

42

field-walking

activities. The mobile lifestyles of prehistoric huntergatherers and pastoralists could create such ‘non-site’ archaeology, and many agriculturalists in the past created similar spreads of material from manuring their fields if their manure heaps also acted as general-purpose rubbish heaps, with discarded archaeological artefacts accompanying the manure onto the fields. Thus, changing patterns and distributions of the ‘off-site’ archaeology surrounding the ‘site’ archaeology may well be an important indicator of changing patterns of land use (Bintliff and Snodgrass 1988; Whitelaw 2000; Wilkinson 1982, 1989). In Britain, where field-walking is usually small-scale and intensive within a few square kilometres (permission is needed from every landowner for every field, to start with), the scale of work is small enough and the finds generally few enough that each field can be gridded and all the finds on the surface collected within a grid of, say, 10, 15 or 20 m divisions in what has been termed the ‘traverse and stint’ method of field collection (Connolly 2008; Haselgrove et al. 1985). In the Mediterranean, projects that have used similarly intensive methods have invariably experienced a dramatic slow-down in coverage accompanied by the recovery of overwhelming quantities of material especially relating to the classical centuries, and especially tile fragments, requiring sometimes decades of study beyond the fieldwork (Bintliff et al. 2017; Jameson et al. 1994). Some projects of this kind have also attempted ‘siteless survey’, in which all data are collected systematically in grid units and sites are only defined post facto when the data have been analysed. Other projects have tried to avoid making hard-and-fast ‘site’ vs ‘offsite’ judgements in the field and instead used terms like ADABS (Abnormal Density Above Background Scatter: Carreté et al. 1995) or POSIs (Places of Special Interest: Given et al. 1999). One solution that we considered for the problem of how to map site and off-site archaeology reasonably quickly but also as objectively as possible was ­utilized in the Nemea Valley Survey in Greece. In this project provisional identifications of anomalies in the field data were made only as the data were processed and registered, the locations of the anomalies then being revisited for detailed artefact collection to decide whether or not

they should be formally classified as ‘sites’ (Wright et al. 1990). However objective the methodology, though, there invariably has to be a degree of subjectivity in the decisions to assign relative importance to different categories of field survey data: in the last resort, as John Cherry commented (1984: 119), ‘the recognition and delimitation of a site are an act of interpretation, not observation’. In the case of the Tuscania Archaeological Survey, we needed a field methodology sensitive enough to detect both sites and the major characteristics of ‘off-site’ patterning, but at the same time robust enough for us to tackle our central research questions about territoriality and town and country relationships within our target area of 100 km2. Hence, we decided on a compromise between ignoring off-site archaeology and gridding every field: we elected to identify sites (not ADABS!) in the field, separate them from the off-site archaeology and map the latter at the level of the individual field rather than at a more detailed scale of grid squares within fields.

field-walking In the past couple of decades remote sensing using cheap hand-held GPS (Global Positioning Systems) equipment and high-resolution satellite images has revolutionized archaeological survey, but the Tuscania Archaeological Survey was undertaken before either was easily available for civilian use, and when there were also considerable errors in accuracy built into what could be obtained. Hence the field-walking had to be based on normal mapreading, and there were many difficulties in locating survey units accurately on the map. Each team was allocated a portion of the particular square kilometre being searched. Frequently even locating this was not a straightforward task, because the best maps available at the time of the survey were the 1:25,000 IGM topographic maps and several had not been upgraded since the 1950s or 1960s, so buildings, roads and field boundaries marked on the map had often disappeared, and new ones appeared in unexpected places. At the 1:25,000 scale a square kilometre measures 4 cm by 4 cm on the map, clearly inadequate for an exercise in which the individual field would normally

43

2 Methodologies

be the main unit of study, so we had no alternative but to enlarge the maps as photocopies to about double size, though conscious of the errors inherent in this process. The team then had to define a unit for field-walking. Normally this was an individual field, bounded by a stone wall, hedge or ditch, but if a field measured more than a couple of hundred metres across, it was usually subdivided into smaller units making use, if possible, of dividing features such as an old wall line marked by clearance cairns or a drainage ditch. The teams were instructed to walk as much accessible land as possible in the square kilometre under study. Most units consisted of ploughed fields, but fields under crop were also investigated if the crop would not be damaged by trampling and if at least half the ground surface was visible between the trees or plants. Units of this kind included olive groves and vineyards, cereal stubble fields and fields with crops such as sunflowers, melons, tomatoes or lucerne, whereas intensively cultivated gardens had to be avoided because of potential damage, though in some instances vegetated units with much lower visibility were also able to be searched. Occasional pasture fields were usually possible to search because the grass was burnt dry by September, with substantial amounts of soil visible between the tufts. Scrub and open woodland were also searched whenever possible. The only land not searched consisted of dense woodland and fenced-off land, the latter happily uncommon except in the immediate vicinity of Tuscania. Large areas of woodland were also rare in many parts of the survey area, but some kilometre squares were heavily wooded, especially in the East Transect but also in some of the Random Squares and in the southern part of the main block of Judgement Squares. The data collected by the survey teams indicated that almost 55% of the land searched consisted of ploughed fields, the remaining 45% being covered by a wide variety of land-use types (listed in the notes to the Survey Gazetteer: Appendix II: Table A2.1), the most common ones being cereal stubble (8.73%), fallow land (6.22%), olive groves (5.48%), lucerne (3.42%) and cereals (3.31%) (Fig. 2.10). Having identified the unit of study, the team of walkers then traversed the ground normally at 15 m intervals

other

cereals

fallow

lucerne olives

stubble rough grazing

plough

figure 2.10  Percentages of land-use types on the ground searched by the survey teams. (Data computed by Nicoletta Vullo.)

(paced rather than measured with a tape), walking across the unit at right angles to a boundary line (Fig. 2.11). They were instructed to walk slowly but steadily, each member of the team picking up and bagging all artefacts (except obviously recent items such as plastic, modern glass, cartridge shells and so on) visible on the ground surface along an imagined 1 m strip at their feet. When the team arrived at the opposite field boundary, they returned to the original start line with a second set of transects 15 m apart parallel with and adjacent to the first set, continuing in this way until the entire field had been traversed. If the unit contained only ‘off-site’ or ‘sporadic’ material without any distinctive ‘site’ concentrations being noted during the line-walking, all the finds picked up by the individual team members would be put together and the unit would be given a single number, so the first unit in square T74 would be T74:1. Any concentrations were given new unit numbers and the material was collected and bagged separately. Thus, Unit RS39:1 was a large field containing a distinct artefact concentration within it, RS39:2, and the next field to be searched, adjoining Unit RS39:1, was classified as Unit RS39:3 (see Figure A2.8). It was impossible with this in-field methodology to make an absolute separation between site and off-site material. Occasionally sites were clearly visible to a team from the edge of the field before they started line-walking it (e.g. Fig. 2.8), but many sites would become apparent to the team only when it was in the process of line-walking

44

field-walking

figure 2.11  Field-walking in progress during the Tuscania Archaeological Survey. (Photograph: Tom Rasmussen.)

the field. Inevitably there were differences in the reaction times of different team members to note a change in surface distributions and call a warning before the rest of the team entered the area of the concentration and started to bag its finds with the material from the main unit. Detailed grid mapping by other projects of fields containing sites within them also makes it clear that instead of sharp boundaries the sites usually have a ‘halo’ of material of decreasing density moving outwards from the centre (Bintliff and Snodgrass 1985, 1988). We observed the same phenomenon at the sites where we undertook gridded collections. Another problem with the system we used is that it was likely to mask the complexity of the ‘off-site’ or ‘non-site/sporadic’ archaeology, bearing in mind the analogy mentioned above of the contour map with peaks and troughs of many different scales: with team members 15 m apart we were likely to miss some of the

very small concentrations of material (say measuring 3–5  m across), and also not distinguish more diffuse collections of material spread out over large areas that were still at variance with the ‘background noise’ of sporadic material across the rest of the unit. Nevertheless, while we were mindful of the information lost through the method we used, and in particular the dangers of the subjectivity involved in distinguishing between ‘site’ units and ‘off-site’ or ‘sporadic’ units in the field, we decided that it was a reasonable compromise between the need to measure changing densities of material rather than just ‘site-spotting’, and the need to recover information quickly about both site and off-site archaeology over a sufficiently large area to be likely to serve the project’s main research objectives. The detailed survey maps resulting from the five years of field-walking are shown in the Survey Gazetteer (Appendix II, Figs. A2.2–A2.10).

45

2 Methodologies

collecting

angles to one another crossing at the centre (Figs. 2.12, 2.13, and see Figs. 5.14, 5.17, 6.3, 6.4, 6.5). No prehistoric or post-classical sites were gridded, as the line-walking at the time of initial discovery effectively scoured their surfaces of all or most of their visible remains. The purpose of the gridding was twofold. The first was to aid the pottery specialists by ensuring that we had representative collections of all pottery: on Roman sites, for example, the black or red shiny surfaces of the finer table wares were usually easier to spot in ploughsoil than coarse wares with dull brown or ochre surfaces. The second was to get better information on the spatial characteristics of sites from artefact distributions and types: apart from the halo effect mentioned above, other information retrieved included, in some instances, indications of separate living and working areas on the evidence of different distributions of artefacts such as storage amphorae, coarse pottery, fine table ware, mosaic tesserae, fragments of painted wall plaster and fragments of statuary (Chapter 6). Our normal procedure was to retrieve all pottery and other finds seen within the 1 m transect in the 15 m and 5  m line-walking. This did not present any problems with units with prehistoric, Etruscan, Medieval and Post-Medieval material, but in common with most Mediterranean surveys we were faced with immense quantities of Roman material, especially brick and tile. There were a few cases of Roman sites where the finds even from 5 m line-walking were so plentiful and the unit so inaccessible to vehicles that pottery had to be sorted at the unit, with all fine wares and diagnostic coarseware sherds being collected and only a sample of coarseware body sherds being brought back for study. Such sites could also yield huge fragments of stone grinders, or complete roof tiles, making transporting the finds from the unit to the vehicle a major test of resolution and endurance for the field teams. Unit R14:13 immediately below Colle San Pietro, which as we describe in later chapters was probably a suburban extension of the ancient town, contained so much material in the ploughsoil that the finds from one 20 m grid square filled a large plastic dustbin (Fig. 2.12). In the case of the sites we gridded, either all material was collected

Most multi-period Mediterranean survey projects are faced with immense quantities of material on the ground, and decisions have to be taken about how to collect and what to collect. A few projects have endeavoured to ‘vacuum’ everything visible on the ground surface, but if the study area is anything more than a couple of square kilometres, this procedure can result in literally tons of material, mainly brick and tile of the classical centuries. At the other extreme, some earlier survey projects interested in classical settlement systems concentrated almost entirely on collecting fine wares, but such an approach biases the data heavily towards those sites occupied by the richer people of antiquity who could afford to have fine table wares, effectively excluding from consideration the people lower down the social scale. Also, fine wares are likely to be rare anyway, so surveys need to establish chronological sequences of coarse wares, even though the latter are usually much harder to date, if they hope to establish the period or periods of occupation of particular sites (see below). Most surveys in recent years have selected collection methods somewhere in between the two extremes. Examples include collecting all pottery, brick and tile; collecting all pottery and a sample of brick and/or tile; collecting all pottery but not brick and tile; collecting only diagnostic sherds such as rims, bases and decorated pieces; and collecting diagnostic pieces but also using hand-held clicker counters to record the numbers of sherds left lying on the ground as a guide to densities (Mattingly 2000). In our case, the normal system of collection during the line-walking at 15 m intervals was to pick up everything visible on the ground surface apart from patently modern material. Whenever ‘site units’ were recognized, their material was then sampled by line-walking at 5 m intervals, with the same principles of collection. However, we also returned to a representative series of Etruscan and Roman sites of different sizes to make more detailed collections of their surface remains, usually within a grid of 5 m x 5 m quadrants but sometimes using a single transect across the long axis of the site, or two transects at right

46

collecting

figure 2.12  Collecting surface material in a grid of 20 m x 20 m quadrants at the prolific unit R14:13, likely a suburban extension of Etruscan and Roman Tuscania (looking south). (Photograph: Tom Rasmussen.)

and brought back for study, or all material was collected, sorted, counted and weighed on site (Fig. 2.13) to give density distributions, with only a sample of diagnostic material being brought back. The main problem in terms of bulk material was tile. Tiled roofs have been used in the study area from the seventh century bc to the present day, and in terms of both numbers of objects and overall weight, the surface archaeological record was dominated by tile fragments. Most tile collected in the 15 m line-walking was retained, but some had to be discarded when the load was simply too much for the teams to carry. In the initial seasons in these instances, teams made a selection in the field of what seemed to be diagnostic pieces and discarded the rest, but unfortunately, we did not always record the proportion discarded. In the later seasons, we recorded the ratio of what was kept and what was discarded, as we began to establish that tile fabrics differed and that it was possible to establish a tile fabric series (see below),

establish its periodization at well-dated sites and use it to establish the periodization of sites with no fine ware and little or no coarse ware and off-site scatters (invariably dominated by tile). However, the fact that fabrics could usually only be established definitely by breaking the tile fragments to expose fresh unabraded surfaces meant that with our system of dividing them quickly by eye in the field the information on discarded tile in off-site units was of limited value, affecting the quantification exercises based on the sample we kept. The Rieti Survey attempted to overcome this problem by smashing all the tiles in the field at the end of each unit’s collection and recording the numbers of different fabric classes there and then (Coccia and Mattingly 1992). The procedure resulted in better quality tile data than the Tuscania Archaeological Survey’s but meant that the project was able to cover only 22 km2 in four seasons of fieldwork rather than the Tuscania Archaeological Survey’s 42 km2 in five field seasons.

47

2 Methodologies

a

b

figure 2.13  (a) Collecting surface material in a grid of 5 m x 5 m quadrants at T24:8; (b) weighing the tiles from the grid collection. (Photographs: Graeme Barker.)

48

recording

recording

modern ploughshare can reach and so effectively invisible to field-walking teams. Land-use patterns clearly affected the accessibility of land to survey teams. The state of the weather and the state of the agricultural cycle drastically affected the visibility of any archaeological artefacts lying on the ground surface. Thus, artefacts might be clearly visible in a freshly ploughed field, a rain shower might make some of them even more visible by washing the upper surfaces of the sherds (which then catch the sun’s rays), but a heavy storm might create muddy rivulets that smeared the soil surface and masked many of the artefacts. The use of caterpillar tractors and heavy steel ploughshares (Fig. 1.15) created enormous furrows a metre deep (Fig. 2.15), but the same field a few weeks later might be harrowed flat. The amount and intensity of the light cast onto the tops, bottoms and sides of the furrows varied considerably in sunny and overcast conditions, or as shadows changed in size and intensity through the day, so the weather conditions as a team walked over a field, and the direction taken in relation to the sun, might affect their ability to see surface artefacts (Fig. 2.16). Harrowing a poorly harvested field which had lots of stubble in it could create a good flat surface for field-walking, but one entirely covered by a thick carpet of dust and straw fragments. Most surfaces between rows of crops such as lucerne would not have been disturbed by farmers for many weeks before a survey team arrived, with wind and rain in the meantime smearing and wetting any surface artefacts. The potential biasing effects of differential ground visibility were one of the issues investigated by the GIS. For this exercise, NV took the most represented classes of Roman pottery (amphorae, African red slip, black-glaze, colour-coated, coarse, dolia, internal red slip, Italian terra sigillata and thin-walled) together with Etruscan bucchero and calculated their ratios in relation to land use according to presence/absence (True/False) (Fig. 2.17). This showed that 50–80 per cent of all pottery types were associated with ploughed land, but that most pottery types were also found in other types of land use including ones characterized by low visibility. While the percentage of plough associated with pottery is rather variable, and often low, the percentages of the other categories of land

Each team marked the location of the units it defined on their enlarged photocopy of the IGM 1:25,000 map using colour pencils to make them clearly distinguishable. Any colour could be used for the units, apart from green, which was used to mark vegetated areas with ground cover too dense to justify survey, or blue, which was used to mark any areas that were inaccessible because they were fenced. Having collected, bagged and labelled the material from a unit, the team then completed a standard pro forma record sheet on the unit. It is standard practice in plough-zone surveys to record a wide variety of information so that an assessment can be made of the scale and severity of different biases likely affecting the results. The main archaeological bias is the different visibility of different categories of material, both in the sense of literal visibility (shiny African red slip pottery being more visible than earth-coloured prehistoric handmade sherds, for example) but also in terms of the visibility of the different quantities of materials produced in different periods (Millett 1991, 2000). The variable skills and experiences of team members are another significant factor. Do all the team members have equal experience at recognizing different classes of artefacts such as prehistoric flints, handmade prehistoric and Early Medieval pottery, Roman coarse wares and so on? Are all the team members equally capable of sustaining accurate observations no matter how testing the weather conditions? How reliable are observations made in the last hour before the midday break, when the sun is normally extremely hot and the team already tired from several hours of field-walking, compared with observations made in the more pleasant working conditions of the morning or late afternoon? ‘Natural’ biases operate at a variety of scales. Erosion can remove archaeological sites and off-site materials from parts of the topography, such as hillslopes, and carry the artefacts down to lower elevations (Fig. 2.14). Alluviation and flooding can dump thick deposits of later sediment on top of earlier surfaces containing archaeological remains, making the archaeology lower than the

49

2 Methodologies

figure 2.14  The effects of erosion on surface archaeology: the dark soil in the centre of the picture contained a dense scatter of classical pottery (site T24:3), whereas the hilltop on the right, which has been ploughed down to bedrock, had none, but the ancient site and the dark soil were both originally on the hilltop – ploughing in recent decades has caused the soil to slip downhill, taking the archaeology with it. (Photograph: Graeme Barker.)

use tend to be rather high. The types of pottery included in the analysis were found in most of the non-ploughed units, and often in all of those units with the same kind of land use. A similar analysis was performed using distribution maps based on quantities of Roman sherds rather than presence/absence (Fig. 2.18). The amount of pottery per survey unit was reclassified into six classes, ranging from fewer than 5 fragments to more than 100. Taking two classes as examples, of 6–10 sherds and 11–20 sherds, the analysis showed that, while only a small percentage of land-use categories other than ploughsoil was associated with pottery, and the numbers of sherds of particular pottery types were often very small, all kinds of land use were associated with different kinds of pottery. Interestingly, survey units with more than five sherds were rather likely

to coincide with land-use categories characterized by low visibility, rather than ploughed fields. The implication is that while ploughed land undoubtedly favoured finds recovery, there does not seem to have been a significant bias against finds recovery in the more vegetated units that we searched. At the beginning of the Tuscania Archeological Survey, a Unit Record form was used that was adapted from one used in the Montarrenti Survey in Tuscany, which was being completed as the Tuscania Archaeological Survey began (Barker et al. 1986). However, when we began to computerize the database after the first and second seasons of fieldwork, we realized that parts of the form were too open-ended: good team leaders gave us all the information we needed, but there were other cases where information was insufficiently detailed. Hence a

50

recording

figure 2.15  Giant furrows and ridges produced by modern deep ploughing at T54:8, creating difficult conditions of light and shadow for field-walking. (Photograph: Graeme Barker.)

remodelled Unit Record was used in the third, fourth and fifth seasons that gave examples of types of slope, land use, topography, soil and light conditions, visibility and so on for the team leader to circle (or make another category if necessary), together with scales of average artefact sizes and densities. Putting the two sets of forms together into the single database was predictably problematical, with some categories of information from the first two seasons having to be entered simply as ‘not recorded’, but most information could be recovered using the recording form in combination with the photocopied map and Pottery Summary Sheets (see below). The Tuscania Archaeological Survey database established from the Unit Records contained 56 fields. The first fields included the unit number and type (Transect, Random or Judgement), its map reference, the date on which it was surveyed, the initials of the team members

and the estimated area of the unit in square metres. This last was problematical in the case of most field units (as opposed to ‘sites’), as it required a combination of paced measurements and estimates taken from the enlarged map photocopy, and considerable discrepancies emerged when the IGM maps were digitized for the GIS and the size of the larger (mostly non-site) units calculated accurately. The second set of information concerned the natural characteristics of the unit and of the conditions when it was surveyed: the underlying geology (though the main detail was taken off the IGM geological maps); soil type in terms of clay, silt and sand content; slope, aspect and height above sea level; present land use; topographic situation; the nature of the ground surface and its condition; the time of day, and the light conditions; and the distance between walkers and the collection method of the material. Of the ‘environmental’ data recorded by the survey

51

2 Methodologies

a

b

figure 2.16  The effects of light on the visibility of surface artefacts: RS27:10 in late afternoon sun: (a) looking north, and (b) looking west. (Photographs: Tom Rasmussen.)

teams, we have included only the entries for Elevation, Topography, Land Use and Visibility in the Survey Gazetteer, for reasons of space (Table A2.2). The hard copies of the survey forms are archived in the University of Cambridge’s McDonald Institute for Archaeological Research. The archaeological characteristics of the unit were recorded in the third group of fields: the average density of finds; the average size of tile fragments if present and the approximate proportion collected; the extent to which the finds were fresh or weathered; the range of materials found; and whether there were any standing remains or archaeological features in addition to surface artefacts. Finally, two questions had to be answered by the survey team: ‘is this a site?’ and ‘are the finds in situ?’, to which

they could answer ‘yes’, ‘no’ or ‘don’t know’. If they had decided that the material could be classed as a site, they gave its dimensions, orientation and what kind of site they thought it might be (domestic, industrial, burial, etc.), together with any comments, such as how it might relate to adjacent survey units. Archaeological features such as rock-cut tombs, ancient cuts for roadways, ruined buildings and so on were recorded within the same system of survey units, whether or not they were associated with ‘normal’ field units and whether or not they yielded artefacts. In total nearly 1,400 units were recorded by the survey teams: 1,218 with pottery, 56 with only tile, 52 recorded as standing structures and 68 with no surface archaeology. If the nature of plough-zone archaeology is complicated enough, how it may relate to any sub-surface

52

recording

100

AMP ARS

90

BUC

Percentage of pottery

80

BG

70

CC CW

60

D

50

IRS

40

ITS

30

TWW

20 10 0 clover/alfalfa

cereals

fallow

legumes

lucerne

olives

plough

rough grazing

stubble

Landuse

100

AMP

90

ARS BUC

80

BG

Pottery

70

CC CW

60

DL

50

IRS

40

ITS

30

TWW

20 10 0 clover/alfalfa

cereals

fallow

legumes

lucerne

olives

plough

rough grazing

stubble

Percentage of landuse figure 2.17  Ratios of (above) major classes of Etruscan and Roman pottery to land use and (below) land use to the same pottery classes, using True/False (presence/absence) maps of pottery distributions (see text for discussion). Pottery key: AMP – amphorae; ARS – African red slip; BUC – bucchero; BG – black-glaze; CC – colour-coated; CW – coarse ware; DL – dolia; IRS – internal red slip; ITS – Italian terra sigillata; TWW – thin-walled ware. (Analysis by Nicoletta Vullo.)

archaeology is even more problematical. Numerous Mediterranean surveys have investigated the relationship between surface and sub-surface archaeological remains not only by excavation but also by geophysical survey

and geochemical analysis. The findings of the Biferno Valley Survey were typical in revealing the complexity of the relationships (Barker 1995a). On certain soils there, geophysical techniques were effective at revealing buried

53

2 Methodologies

100

AMP ARS

90

BG

80

CW ITS

70

TWW

60 50 40 30 20 10 0 fallow

plough

rough grazing

stubble

fallow

plough

rough grazing

stubble

100

AMP

90

ARS BG

80

CW ITS

70

TWW

60 50 40 30 20 10 0 fallow

plough

rough grazing

stubble

fallow

plough

rough grazing

stubble

figure 2.18  Ratios of land use to major classes of Roman pottery and the same classes of pottery to land use based on groupings of (above) 6–10 sherds and (below) 11–20 sherds (see text for discussion). Pottery key: AMP – amphorae; ARS – African red slip; BG – black-glaze; CW – coarse ware; ITS – Italian terra sigillata; TWW – thin-walled ware. (Analysis by Nicoletta Vullo.)

features such as walls, hearths and pits, but on others less so. At several dense concentrations of surface remains regarded as sites (of different periods), test excavations revealed intact archaeological layers below the ploughing level, but at others it was found that ploughing had reached the subsoil so that all that remained were the artefacts in the ploughsoil. Test pits excavated at some

sites marked by sherd concentrations indicated that the sites had probably only ever been ephemeral settlements such as tented camps – that is, there never had been any substantial archaeology below the surface. In the Tuscania Archaeological Survey we investigated the nature of subsurface remains at a few sites using geophysical survey (e.g. Figs. 6.6 and 6.7) and augering, sometimes with

54

classifying the finds

poor results: deep ploughing has been going on so long in Etruria compared with the Biferno valley, for example, that the scale of attrition to the sub-surface archaeology of small archaeological sites is intense, though when we excavated at one small Etruscan unit, the ploughsoil remains turned out to overlie a quite substantial stonewalled building (see Chapter 5).

entire assemblage on pro forma Pottery Summary sheets, to ensure consistent treatment of the material. Students supervised by AM undertook this first stage of classification, dividing the pottery into four broad chronological categories: prehistoric, Etruscan, Roman and Medieval/ Post-Medieval. This was not an easy procedure, especially making distinctions between Etruscan and Roman coarse wares, the bulk of the material under study, but the starting point was the method of manufacture. Although there is a small quantity of handmade Etruscan pottery, and hand-made Roman pottery is a rarity, if a sherd appeared to be from a handmade vessel it was classified first as prehistoric. In comparison with Etruscan wheelmade coarse wares, Roman wheel-made coarse wares are on the whole more refined and better levigated and fired at rather higher temperatures. With these kinds of observations, the material was provisionally separated into its most likely period groups in readiness for specialist studies, when any misassigned sherds could be reclassified. The second stage of classification involved the detailed recording of all sherds by the specialists, with pottery record sheets appropriate to the material under study being completed for every unit. The smallest category was the c. 2,200 sherds of prehistoric pottery (see Chapter 4). At the other extreme was the daunting challenge posed by the almost 54,000 sherds of Roman pottery, the approach to which is described in detail in Chapter 6. A corner was chipped off the corner of each tile fragment and the freshly exposed surface was compared with those of a fabric series established by Chris Hunt early in the fieldwork using similar criteria to those used to establish the various pottery fabric series (Table 5.6). The number of fragments of each type was then recorded on a pro forma sheet for each unit (included in the Survey Gazetteer: Table A2.2), but not the weight, given the uncertainties mentioned earlier concerning the relationship of the sample brought back for study to the sample originally collected from the unit. In addition to flint, pottery and tile, the wide variety of other materials collected from the survey units was recorded in a separate catalogue. These included pieces of worked stone, fragments of glass, animal and human

classifying the finds All the material collected by the survey teams was washed, sorted and catalogued at a preliminary level at the project base during the field seasons, the final cataloguing being completed after the completion of the fieldwork. The preliminary cataloguing consisted of separating pottery, tile, brick and other materials, and subdividing the pottery into major period groups for later specialist study. The prehistoric pottery was studied by Francesco di Gennaro and GB at the end of the fieldwork, and the small number of prehistoric flints found by the survey was studied by Tim Reynolds at a later date. The Etruscan pottery was studied by Marco Rendeli during the fieldwork, with additional input from TR. Processing of the Roman fine and coarse wares from the Transect Sample was begun by Nick Whitehead and initial sorting of the Roman amphorae by Phil Perkins at the time of the first two field seasons, but the entire set of Roman material – by far the bulk of the pottery collected in the survey – was then studied by AM as the principal case study of her Oxford DPhil on ceramic methodologies for Mediterranean surveys (MacDonald 1999). The Medieval and Post-Medieval pottery was studied by Helen Patterson in the final field season and the following finds study season. In the first seasons of fieldwork, the pottery assemblages were inspected and given approximate date ranges. Feature sherds from all periods were recorded and sometimes removed for further study and illustration. A sample of Roman wares was placed in a form and fabric series, but body sherds were dealt with summarily. A more systematic system was established in 1990, with all sherd assemblages being divided into period groupings, then counted and weighed. This task involved recording the

55

2 Methodologies

bone, mosaic tesserae, wall plaster, coins and other metal objects (see the Comment column in Table A2.2). The finds records are archived with the field records in the McDonald Institute for Archaeological Research. On the basis of the finds classifications, the survey data can be classified into twenty major settlement phases (Table 2.1). The sequence begins with Period ‘0’, PreNeolithic, because as we describe in Chapter 4, we found a few lithic implements that may be Neolithic but may be earlier. The main sequence begins with Period 1, the Earlier Neolithic c. 5500 bc, and ends with Period 19, Modern, the nineteenth and twentieth centuries. These periods are based on artefact, especially ceramic, assemblage groups, and are not to be equated automatically with cultural change. The rationale for the definition of particular phases and their relationship, if any, to cultural and landscape change are discussed, as appropriate, in the later chapters. All of the pottery and other finds from the Tuscania Archaeological Survey, except the tile, were consigned after study to the Superintendency of Antiquities’ store in Tuscania. A small sample of every tile fragment entered in

the fabric count was also consigned, the bulk of the discarded tile fragments being deposited at the end of the fieldwork with the agreement of the authorities in a single location outside Tuscania, where there seemed the least likelihood of it confusing future researchers.

analysing the finds: interpretative issues In this section AM discusses general issues of interpretation which cross-cut the analysis of the finds but were most critical in the treatment of the Etruscan and Roman material that formed the bulk of the Tuscania Archaeological Survey finds.

Defining ‘Sites’ Although the field teams were asked to separate sites from offsite/sporadic spreads of material, these observations were not a straightforward guide when it came to analysing the material. There could be cases of ‘unreal sites’, for example when a concentration observed and classified as a site was more a function of natural processes such as hillslope erosion than human activity (e.g. Fig.  2.14). There could be ‘missed sites’ where a real site had been destroyed by ploughing and the material distributed widely across a field. There could be ‘part-sites’, where a discrete concentration with Etruscan and Roman sherds observed as a definite site in the field in fact consisted of, say, off-site or sporadic Etruscan material overlain by a Roman site. For the prehistoric material, the analytical problems were less of an issue: as we explain in Chapter 4, except for cases of isolated individual sherds almost all collections of prehistoric sherds were probably indicative of definite loci of prehistoric activity, as had been the experience of the Boeotia Survey, for example (Bintliff et al. 1999). For the Etruscan, Roman and later material, however, it was necessary at the post-survey stage to develop a set of criteria to measure the reliability with which units could be classified as sites. The system we used for the Etruscan and Roman sites was adapted from the system used in the Ager

Table 2.1  The periodization established for the Tuscania Archaeological Survey Period 0 Period 1 Period 2 Period 3 Period 4 Period 5 Period 6 Period 7 Period 8 Period 9 Period 10 Period 11 Period 12 Period 13 Period 14 Period 15 Period 16 Period 17 Period 18 Period 19

Pre-Neolithic, before Period 1 Earlier Neolithic 5500–4500 bc Later Neolithic 4500–3500 bc Chalcolithic or Copper Age 3500–2200 bc Earlier Bronze Age 2200–1400 bc Later Bronze Age 1400–950 bc Iron Age 950–700 bc Archaic Etruscan 700–500 bc Later Etruscan 500–300 bc Early Republic 300–170 bc Late Republic 170–30 bc Early Empire 30 bc–ad 120 Mid Empire ad 120–260 Late Empire ad 260–440 Late Antiquity ad 440–700 Early Medieval ad 700–1000 High Medieval ad 1000–1200 Late Medieval ad 1200–1500 Post-Medieval ad 1500–1800 Modern, ad 1800–2000

56

analysing the finds: interpretative issues

Tarraconensis Survey (Carreté et al. 1995). Their approach involved the calculation of the density of sherds per hectare belonging to each period group for each survey unit. Units where the density of sherds fell into the top tenth of values were classified as the ADABS referred to earlier and regarded as probable sites. Thus, collection units with the highest relative densities of sherds were taken to represent sites, rather than units simply with the greatest number of sherds. The disadvantage of the system is the difficulty of distinguishing sites for periods when pottery is scarce. If pottery from a particular period was found in less than 10% of assemblages, all units for that period would fall in the top percentile range, so a unit with only one datable sherd would be categorized as an ADAB. Also, the decision to use the top 10% was ­arbitrary – in theory the cutoff could have been 15% or 20% of density values. Another problem with the methodology was its implicit assumption that there was only one category of site in antiquity, whereas on the basis of other surveys, and excavations of sites discovered in surveys, we could assume that a range of different types of site existed around Tuscania in Etruscan and Roman times. We therefore started by dividing the Etruscan and Roman material into four groups by calculating density values for each period for all units, rather than the two groups, above and below 10 per cent, used in the Ager Tarraconensis Survey: the top tenth; below the top tenth but in the top quarter; in the top half but below the top quarter; and below the middle value. In addition to relative densities, the absolute number of sherds dating to a period was also used to contribute to site classification so that a unit with only a handful of sherds could be differentiated from a unit with a large quantity. The third factor was the observations made by the survey teams in the field. To establish the relative density of sherds for each unit, the absolute density was calculated first by dividing the number of sherds by the area of the unit. A ‘density correction factor’ method was devised by AG to take account of the distance between walkers and the observed visibility of the ground surface: if a team walked 15 m apart, and the visibility of the ground surface had been estimated

at 80 per cent, the density correction factor was 18.75 (15 over 80 x 100 = 18.75). The absolute density of sherds for each period was calculated by multiplying the number of sherds for the period in question by the density correction factor and dividing the total by the area of the unit in square metres (as measured by the survey team): thus, a unit of 2,000 m2 with 130 Roman sherds and a density correction factor of 18.75 has a density of Roman sherds of 1.2 per m2 (130 x 18.75 over 2000 = 1.2). The relative densities of sherds for each period were then established by arranging the absolute densities into the four categories mentioned earlier. The ranges chosen to divide sherd numbers for Etruscan and Roman pottery were: 100 or above, 50–99, 25–49, 10–24 and below 10. Although arbitrary, these divisions were chosen to allow units with a relatively large quantity of material to be distinguished from units with smaller numbers of sherds. For periodization within the Roman period, a notional figure for sherds in each unit was established by adding the ‘values’ of sherds in each Roman period: for example, as black-glaze body sherds are usually datable within two phases (Early and Late Republic), an assemblage with five black-glaze sherds was given a value of 2.5 for each period. The resulting figure need not represent the true number of sherds but gives a total ‘value’ per phase, often a fraction of a number rather than a whole. As fewer sherds could be dated to phases within the Roman period (rather than simply to the Roman period as a whole), the ranges established for pottery at this stage were set at a lower level: 10 or above, 5–9, 1–4 and (since ‘values’ were used) less than one. Each unit was assigned a grade for the Etruscan and Roman periods and sub-periods using a sliding set of scales based on relative densities of sherds (A–D), absolute numbers of sherds and whether a concentration was noted in the field. The classification system is shown in Tables 2.2 and 6.5. It divided units into a series of groups, with units most likely to represent sites at the top end of the scale and units least likely to represent sites at the bottom. Grade 1 and Grade 2 units had typical site characteristics (a relatively high density and large numbers of sherds, noted as a concentration in the field) and were interpreted

57

2 Methodologies

Settlement Densities

as ‘definite’ sites. Grade 3 and Grade 4 units generally had lower sherd densities and numbers and were regarded as ‘probable’ sites. Grade 5 units had few features usually attributable to sites and were regarded as ‘non-site’ units. The units classified as sites were divided into three size classes: ‘large’ for those bigger than 7,500 m2, ‘medium’ for those between 7,500 m2 and 1,500 m2 and ‘small’ for those under 1,500 m2. For the prehistoric and Medieval/Post-Medieval periods, for which the finds were almost always more exiguous than the Etruscan and Roman material and site status more difficult to assess quantitatively, we divided units qualitatively on the basis of their finds and the observations in the survey records into four categories: ‘definite’ sites, ‘probable’ sites, ‘possible’ sites and ‘sporadic material’. In the rest of the book, we mostly dispense with the use of quotes for the different categories of ‘site’ we defined (‘definite’, ‘probable’, ‘possible’) and for the designation of ‘non-site’ units as ‘sporadic’, but given the complexities of classifying ambiguous survey data described in this chapter, it should be remembered that all these terms should in reality be treated as if within quotation marks as here.

Calculating settlement densities was largely a matter of interest for the Etruscan and Roman phases given the richness of their data compared with relative paucity of materials for the earlier and later periods and the different challenges of site recognition that these represented. The conventional way for an archaeological survey to calculate the density of sites in a particular phase of settlement is to divide the number of sites by the survey area. On this calculation, the 365 Roman sites found in the 97 km2 sample represent a density of 3.8 sites per km2. The density of ‘non-site’ units can be calculated in the same way. A more accurate approach, however, is to use the area actually field-walked within each square kilometre, rather than the total area covered by the Transect, Random and Judgement Samples. We in fact searched about 41.5 km2, and using this figure the density of Roman sites changes from 3.8 per km2 to 8.8 per km2, clearly a significant difference. For comparison, site densities using both figures are given in later chapters, but the one related to the area actually walked is clearly the more useful guide.

Table 2.2  System used by Alison MacDonald for defining and classifying Etruscan and Roman sites in the Tuscania Archaeological Survey Concentration = Y Density category = Density category = Density category = Density category = Density category = Density category = Density category =

A A B B C C D

Number of sherds = Number of sherds = Number of sherds = Number of sherds = Number of sherds = Number of sherds = Number of sherds =

>20, 10–19, 5–9 1–4 >20, 10–19 5–9, 1–4 >20 10–19, 5–9, 1–4 >20, 10–19, 5–9

Grade 1 Grade 2 Grade 2 Grade 3 Grade 3 Grade 4 Grade 5

A A B B C C D

Number of sherds = Number of sherds = Number of sherds = Number of sherds = Number of sherds = Number of sherds = Number of sherds =

>20, 10–19, 5–9 1–4 >20, 10–19 5–9, 1–4 >20 10–19, 5–9, 1–4 >20, 10–19, 5–9

Grade 3 Grade 4 Grade 4 Grade 5 Grade 5 Grade 5 Grade 5

Concentration = N Density category = Density category = Density category = Density category = Density category = Density category = Density category =

Note: See text for discussion. Y = concentration observed in the field, N = concentration not observed in the field.

58

conclusion

Continuity or Discontinuity?

even fine pottery such as black-glaze and African red slip is not sufficiently datable to be sure of continuous occupation. In reality, the continuous occupation of a site represented as a surface collection of artefacts, and in what form (as continuity of occupation cannot be assumed to represent continuity in function), can only be demonstrated by excavation.

A well-established feature of a diachronic regional study using survey data is the assessment of degrees of continuity or discontinuity in settlement patterns from one phase to the next. Detailed information on the numbers of sites that continued or were abandoned from one period to the next has been presented in many survey reports, with high levels of site continuity usually interpreted as evidence of stability in settlement and landholding patterns, and low continuity and the foundation of large numbers of new sites as indicative of unsettled times, changes in land tenure systems and so on. Quantified data for continuity and change in site occupation were a valuable source of evidence for Alcock (1989, 1993) in her studies of the impact of the Roman conquest and annexation of land on Greek rural populations. Launaro (2011) and Witcher (2006, 2008) used similar data for assessing the demography of Roman Italy. For the prehistoric and Medieval/Post-Medieval periods, the uncertainties of dating much of the pottery were such that assessing continuity and discontinuity was very problematical, but for the Etruscan and Roman periods, with much larger data sets and clearer pottery definitions, the degree of continuity or change was investigated by dividing units into groups according to the presence or absence of pottery dating to succeeding or preceding periods. Units without later pottery were categorized as units abandoned before the subsequent period, and units without earlier material were classed as newly established. Units with material from successive periods were regarded as continuing into subsequent periods. However, these data need to be treated with caution. Continuity is particularly difficult to determine when diagnostic pottery is absent or cannot be recognized. For example, the main period of production of bucchero, the Etruscan fine ware, was the seventh to fifth centuries bc, and the production of black-glaze, the earliest Roman Republican fine pottery, did not begin until the early third century bc, so there is a gap of about a century when local fine wares are either lacking or remain unidentified. The Roman period is characterized by highly diagnostic wares, but

conclusion As is the case with all archaeological field projects, especially field surveys, the research objectives of the Tuscania Archaeological Survey resulted in a series of decisions having to be taken about how we would go about the fieldwork and the finds studies, decisions that have clearly filtered the data on which our interpretations are based. We have tried to be clear in this chapter about what we did, and why, and how we assess the strengths and weaknesses of the methodology. The Tuscania Archaeological Survey defined an area of 353 km2 as both large and small enough for the investigation of the research objectives discussed in Chapter 1, particularly the changing relations between the town and its surrounding countryside through the c. 3000 years of Tuscania’s habitation history. The project set out to study about 100 km2 of this study area in its five years of fieldwork, and in one sense almost achieved its objective in that the final area we defined consisted of 97 km2 or 27 per cent of the survey area. On the other hand, problems of inaccessibility (fenced land, thick woodland) and pressures of time meant that the area traversed by the survey teams was actually less than half that, some 12 per cent of the total survey area. Even so, a sample of this size is much larger than the size of sample thought necessary for most statistical inferences about a target population, while the GIS analysis indicates that each of the three sampling strategies (Transect, Random and Judgement) was reasonably representative of the landscape of the study area in terms of its natural characteristics, and in combination strongly so. We need to acknowledge some biasing effects resulting from differences in team skills, but the indications are that these are unlikely to have had

59

2 Methodologies

a critical impact on the quality of the data overall. In the final ­chapter we return to the question of how effective the three sampling strategies were, both individually and in combination, in terms of the information they provided about evolution of Tuscania’s cultural landscapes. In terms of how we dealt with the problem of differentiating ‘site’ and ‘off-site’ archaeology, and in the collection techniques used in the field, the Tuscania Archaeological Survey was like the Ager Tarraconensis Survey in Spain (Carreté et al. 1995) and the Rieti Survey in Italy (Coccia and Mattingly 1992) in standing midway between, on the one hand, site-based surveys such as the Albegna Valley (Attolini et al. 1991; Carandini and Cambi 2002), Biferno Valley (Barker 1995a, 1995b) and Liri Valley (Hayes and Martini 1994) surveys in Italy and the Melos (Renfrew and Wagstaff 1982), Methana (Mee and Forbes 1996), Laconia (Cavanagh et al. 1996) and Southern Argolid (Jameson et al. 1994) surveys in Greece and, on the other, those with a greater focus on mapping off-site material such as the Cecina Valley survey (Terrenato 1992) in Italy and the Boeotia (Bintliff and Snodgrass 1985; Bintliff et al. 2007, 2017), northern Keos (Cherry et al. 1991), Nemea (Wright et al. 1990) and Pylos (Davis et al. 1997) surveys in Greece.

We recovered some 60,000 sherds of pottery as well as a variety of other artefacts in the fieldwork. Although the easiest areas to search were ploughed fields, terrain with other kinds of vegetation was also investigated by the survey teams and the post-fieldwork GIS analysis demonstrates that our policy of walking all accessible land, rather than just ploughed land, was justified in terms of the range and representativeness of the materials retrieved from the former as well as the latter. Building on the experience of other recent projects, we attempted to develop robust methodologies for analysing this large data set and establishing its patterning and variability for different periods of the past in terms of settlement densities, continuities, differences in categories of contemporary settlements, and the likely significance of off-site pottery distributions. The fieldwork was undertaken over 30 years ago when the archaeological techniques available for survey were very different from those of today (Campana 2018), but as the following chapters show, the project yielded one of the richest archaeological data sets for exploring the evolution of a Mediterranean landscape, a dataset that almost certainly could not be recovered today because of land enclosure and land-use changes.

60

3 The Natural Landscape and Its Evolution Tony Brown, Clare Ellis and Edward Rhodes

introduction

the structural components of the landscape

No attempt to evaluate the longue durée of human settlement and culture can ignore the environment as both a formative influence and a cultural artefact. This is particularly the case for multi-period studies in the Mediterranean, because of the inherent climatically driven fragility of the natural environment, a fragility recognized, for example, by the European Union in its programme of research on Mediterranean desertification (European Union 2018; Mairota et al. 1998). The fragility of the contemporary landscape has a history and is a product of that history. The environmental programme of the Tuscania Archaeological Survey area sought to provide data which would supplement regional studies in three principal areas: (1) natural changes of climate and landscape evolution; (2) patterns of landscape change within the survey area; and (3) the influence of human activities on the Tuscania landscape. To achieve these aims, the environmental work strayed beyond the area defined for the archaeological survey but was bounded by the interfluves of the catchment of the Marta river (Fig. 3.1). Whereas the archaeological survey area around Tuscania was defined to address the specific issues of territoriality described in Chapter 1, the environmental programme had to take account of other overlapping ‘functional areas’: the slopes and hydrological catchment for water, erosion and deposition; the atmospheric catchment for pollen and spores; and the broader physiographic region as a climatic area (Brown 2000). The methodology used here involved field reconnaissance and survey, first in the survey region and then downstream.

The archaeological survey area and the catchment of the Marta river are located in the volcanic foothills of the Monti Volsini, the northern sector of the central Italian volcanic complex that extends southwards to the Alban hills south of Rome. In geological terms this is a very young landscape (Blanc 1957). It is dissected by the Tiber, the Tiber’s right bank tributaries and a number of smallto medium-sized rivers, including the Marta, which drain to the Tyrrhenian Sea. The regional geology has its origins in the volcanotectonics of the northern Mediterranean interactions between the African and Eurasian plates, in particular the distensive tectonics of the Tyrrhenian coast and extensive faulting and subsidence (Locardi et al. 1976). The subsidence took place in the Lower Pliocene and led to the deposition of extensive marine deposits which underlie the volcanics. Volcanic activity then took place in this region from the Late Pliocene to the Late Quaternary, in several phases. The topography and existence of several lakes in this region follow from the explosive nature of the volcanics, which created large calderas such as Lakes Bolsena, Vico and Bracciano and the smaller crater lakes in their vicinity which are part of the crater complex, such as Mezzano west of Bolsena and Martignano east of Bracciano. Thinning away from the calderas are the pyroclastic deposits, largely lavas and tuffs (including leucilites, trachites and tephrites including pyroclastites and ignimbrites), which form the local bedrock over most of South Etruria (Fig. 3.2). The archaeological survey area itself is part of the southern slopes of the Bolsena palaeo-volcano, stopping just short of the crater rim, together with a smaller area

AB wishes to acknowledge the financial support of the British Academy, the British School at Rome and the Royal Society, the advice of the late Tim Potter and of Walter Dragoni, and the technical assistance of Anthony Gouldwell.

61

3 The Natural Landscape and Its Evolution

Mount Amiata

VI A

Study reaches

CA IA

ra Fio

SS

Existing lake Palaeolake Pollen data published Roman road Watershed Internal watershed (boundaries excluding lakes)

Lake Bolsena

Lake Mezzano

Survey Area

rta

Cimini Mountains

Lake Vico

Norchia

Tolfa Mountains

Narce

VIA F L A MINIA

a Lake Monterosi

Sabatini Mountains Lake Bracciano

VIA

Lake Martignano

Veio

RE

Palaeolake Baccano

LIA

tta

AU

he

lc Va

VIA

R N IA

Palaeolake Nepi Nepi

BE

Tre i

Falerii Novi

TI

Tarquinia

Civita Castellana

Ti be r

Tuscania

Ma

Vulci

ROME

0

Ostia Antica

20 km

figure 3.1  The Marta catchment, the Tuscania Archaeological Survey area (outlined in red) and the wider region, with sites outside the Marta mentioned in Chapter 3. (Illustration: T. Bacon.)

62

the structural components of the landscape

Lake Bolsena

Tuscania

Cimini Mountains

Lake Vico

Tolfa Mountains

Sabatini Mountains

Lake Bracciano

Ri ve

rT i

r be

TYRRHENIAN SEA

0

20 km

Volsini volcanics Vico volcanics Sabatini volcanics Alban volcanics

}

alkaline potassic volcanics (Pleistocene to recent)

Acidic volcanics (Plio-Pleistocene) Recent (Middle-Late Pleistocene and Holocene) terrestrial and coastal sediments Upper Miocene-Pliocene-Lower Pleistocene clays and sands Major travertine deposits Carboniferous-Lower Miocene sedimentary rocks Major fault Buried fault Group of faults

figure 3.2  The geology of South Etruria. (Illustration: T. Bacon.)

63

3 The Natural Landscape and Its Evolution

of the south-western slopes of the Vico palaeo-volcano. The volcanics in this region are as young as 25,000 bp (Fornaseri 1985), but the estimated date of the last eruptive phase of Vico is c. 90,000 bp and that for Bolsena is c. 60,000 bp (Locardi et al. 1976). The volcanic deposits in the survey area, as in most of the Marta catchment, are tephrites (tuffs, or tufo) and leucilites exposed in the valleys. At its southern edge, the survey area extends onto the Tertiary marine sediments which underlie the volcanics, sediments that include Pliocene conglomerates, clays, marls and sandstones (Fig. 2.3). These rocks are also exposed in the middle and lower valley of the Marta where it has cut down through the tuffs. To the west of the town of Tuscania, there are also outliers of Upper Cretaceous turbidites (sandstones). The nature of the Tertiary and Pleistocene rocks in the survey area has been important for human settlement in several ways. The soils that have developed on the volcanic formations are first-class arable soils (Fig. 1.11), because they are both extremely fertile and relatively easy to work. The tuffs and lavas are unusually easy to cut and dress (Fig. 1.14), but they are relatively resistant to erosion and weathering and so are very suitable for building stone. These young rocks are structurally competent, allowing the easy and relatively safe excavation of underground structures such as tombs and water tunnels, the famous cuniculi constructed by the Etruscans to divert water away from or into arable land (Judson and Kahane 1963; Fig. 5.8e; see also Chapter 5). They are permeable and so can form unconfined aquifers. The Tertiary clays and marls also provide a source of lime for cement and clay for pottery; the most common and convenient outcrops occur at their junction with the volcanics, such as in the lower slopes of the Marta valley close to the floodplain. Given the geological history of the region, it is clear that the geomorphological landscape is very young and has only been subjected to earth surface processes, predominantly fluvial, over the last c. 25,000 years or so. One result is that the landscape still resembles its original form: a series of gentle downslope ridges formed of pyroclastics and tuffs dissected by radial drainage from the crater rim of Lake Bolsena. This geological history makes the Marta

an unusual river, unusual even in this region, because it appears to be the only river which drains a caldera, whereas the other rivers in the catchment originate in the plain of the survey area or in the foothills to the east and are fed by surface run-off and springs at the base of the tuff (Fig. 3.1). The hydrology of most rivers in the region is directly controlled by the regional climate, which is a dry summer subtropical Mediterranean climate with a strong summer drought. As a result, all the other channels in the survey area except for the Marta, and the lower and middle reaches of rivers in the region, are ephemeral or dry up in most summers. This is the case with the Arrone river, a catchment to the west of the Marta with similar topography and geology, a small part of which is in the survey area. The Marta, however, even though its main stream is only about 50 km long, is a perennial river because of the relatively constant discharge from Lake Bolsena. Its channel is also unusually steep-sided (Fig. 1.12). Because lakes function as large sediment traps, the outflow from Lake Bolsena has, and had in the past, a negligible sediment load and thus a greater capacity to erode the Marta’s bed and incise its floodplain than the waters of the other rivers in the region. The unusual characteristics of the Marta river are further discussed later in the chapter. The geomorphology of the Tuscania region was initially characterized along the 10  km-long cardinal transects of the archaeological survey area, moving out from Tuscania (Fig. 2.4: Transect Sample). The North Transect runs parallel to the Fosso le Tufare running into the Fosso Capecchio (Fig. 2.2). This is a deeply incised small stream with sections of valley fill, the basal unit of which was cemented by carbonate. The East Transect crosses the Marta near the San Pietro acropolis and climbs up onto the tuff plateau, crossing a couple of small bedrock valleys, including the Leia, where there are occasional alluvial pockets. Most of the South Transect runs along the plateau top, gradually declining in altitude with the regional slope, and above the east-facing scarp slope of the Marta. The Marta valley is V-shaped, with little evidence of channel change or deposition until the most southerly square kilometre of the transect, where the Marta opens out onto a broader valley with a substantial alluvial terrace. From

64

the regional palaeoenvironmental record

Tuscania the West Transect crosses the interfluve and the valley of the Capecchio, where there is a ford. Below the ford is a terrace with limited exposures containing abundant tile, large pebbles/cobbles and the suggestion of an older finer unit. The main sedimentary unit is clearly of recent age. Back from the river here, there is also a modern brick/tile factory using clay from the Pliocene sediments behind the factory and underlying the tuff. The same clay is used by potteries in Tuscania.

At a similar distance to the north-west of Tuscania is Lake Lagaccione or Mezzano, a small lake lying on the edge of the Bolsena crater at 452 m above sea level. Preliminary studies by Hunt (1988) showed 7 m of sediment, the upper 4.1 m of which cover the Holocene. Although some clearance activity was recorded in the Mid Holocene, in Phase VIII, Hunt noted that human settlement from c. 1000 bc until recent centuries seems to have had little effect on the local vegetation. He pointed out that the same seemed to be true of both the Lake Vico diagram and the diagram from Lake Monterosi, a small lake 30  km south-east of Tuscania (Hutchinson 1970). The most likely explanation for this lies in the nature of the sites: they are all crater lakes surrounded by steep slopes which even today are not predominantly cleared. Some evidence of land-use change is evident in the pollen diagram from the basin of the former Lake Baccano 40 km to the south-east of Tuscania, which was drained in Roman times (Bonatti 1963). Bonatti divided the Holocene sequence here (which had one radiocarbon date only) into a Quercus (oak)-dominated phase c. 10,000–4500 bp, a Pinus (pine)-dominated phase c. 4500–3000 bp and another Quercus-dominated phase at the time when the lake was drained. However, since the aim of the work was climatic reconstruction, there is little detail on the later vegetational changes, and given the inadequate radiocarbon dating, neither the Baccano nor Vico sequences provide accurate palaeoenvironmental chronologies. One lake record in the region which has been ­radiocarbon-dated is Lake Monterosi, which lies in the Sabatini crater complex (Hutchinson 1970). Although very small (0.2 km2) and shallow (a maximum of 6 m in depth), it produced a 2.48  m-deep sequence with eight radiocarbon dates. A wide variety of analyses was performed on the core, including pollen, chrysomonadina (diatoms), porifera (sponges) and chironimidae (water fleas). The dating revealed a marked change in accumulation rate at a depth of about 1.5  m, estimated to be c. 2200 bp. There were also chemical changes at this depth (increases in carbon and nitrogen: Cowgill and Hutchinson 1964), as well as changes in the indicators of

the regional palaeoenvironmental record The principal source of palaeoenvironmental data for the region is the many crater lakes, but there have also been several studies of alluvial sequences which will be discussed in relation to the Tuscania record later in this chapter together with the long historical record from the Tiber. Apart from Lake Bolsena, the closest lake to the majority of the survey area is Lake Vico. Lying 19  km to the east of Tuscania at an altitude of 507  m above sea level, it has a lake area of 4.8 km2, a catchment area of 12 km2 and a maximum depth of 50 m. A pollen diagram of 7.5 m depth was published by Frank (1969), and a 15 m core by Magri and Sadori (1999) and Narcisi (2001) covering the last glacial/interglacial cycle. The Holocene only occupies the top 1.4 m and is not described in detail, but the most closed vegetation, at 100 cm depth, which is dominated by oak (Quercus), beech (Fagus), hornbeam (Carpinus), alder (Alnus) and hazel (Corylus), probably represents the Atlantic pollen zone c. 6000 bp, with indications of the highest available moisture around 5000 bp. Chestnut (Castinea sativa), which is regarded as having been introduced to the western Mediterranean from the eastern Mediterranean around 2000 bp (Beug 1977; Caramiello et al. 1994; Reille and Lowe 1993), is in the Lake Vico sequence from 80 cm depth. Indications of significant aridity about 4300 bp likely reflect the global climatic downturn identified as the ‘4.2 event’, with moister levels returning to previous levels by 2000 bp, when there is evidence of vegetation degradation.

65

3 The Natural Landscape and Its Evolution

lake chemistry (Hutchinson 1970). Bonatti (1970) suggested that these changes were caused by Roman settlement in the area and the building of the Via Cassia, which passes close to the lake shore. At this time there is also evidence of eutrophication and a rise in pollen types such as Urtica (nettle) characteristic of clearance and disturbed ground. In the Late Roman period, there is evidence of some regeneration, with a second significant episode of agricultural indicators (including cereal-type grains) in the Medieval period. However, Ward-Perkins (1970) suggested that there was survival well into the Roman period of large tracts of natural woodland, especially around Sutri, a thesis supported indirectly by the delays in the construction of the radial roads from Rome as a result of the need to fell woodland; this woodland had largely gone by the end of the Roman period (Ward-Perkins 1962). The greater sensitivity of Lake Monterosi probably reflects both its proximity to settlements and the far gentler slopes surrounding it than the other crater lakes in this area. More detail is available from the study of Lake Martignano, a small lake 38  km to the south-east of Tuscania, reported by Kelly and Huntley (1991). Although there are some problems with the radiocarbon dates, the upper 0.8 m covers the last 3000 years. The much greater taxonomic precision of this study allows more information to be gained on land-use change: for example, olive (Olea europaea) starts to rise a little after 3060 ± 220 bp and then increases more dramatically in the last 2000 years. Kelly and Huntley (1991) suggested that forest clearance reduced mesophyllous (soft and deciduous-leaved) forests and that changes in sclerophyllous (tough evergreen drought-resistant) vegetation reflect e­nvironmental –­ climatic – change. In their climatic reconstruction, using the pollen-climate response surfaces method, precipitation peaked at c. 1000 bc and then again at c. ad 1000, before falling to present levels. However, this sequence does not correlate well with either the lake level from Bolsena or the Tiber flood record (Camuffo and Enzi 1994; D’Onofrio 1980; Pavese et al. 1994; Fig. 3.3). Archaeological data from Lake Bolsena indicate dry phases around 1150–800 bc and 500 bc, a cold and wet phase during the Roman Republican

66

and Early/Mid Imperial periods (c. 200 bc–ad 200), a dry period around ad 1300–1500 and a pronounced wet period associated with the Little Ice Age c. ad 1500–1700 (Dragoni 1996, 1998). Lake Mezzano also experienced a period of lower lake levels c. 1000 bc, indicating less precipitation (Dragoni 1996, 1998). The proxy climate records from the Bolsena and Mezzano lake levels are preferred here, as they are closely related to run-off and the geomorphological changes which are evident in our own survey area. In general, the record in South Etruria is of warm climatic phases being associated with a decrease in water surplus (and therefore run-off), and cooler climatic phases being associated with a rise in water surplus. Since the two records (pollen and palaeohydrological) are based on different components of the hydrological cycle, absolute agreement should not be expected. It is worth comparing these records with others elsewhere in central Italy. South of the Tiber, a pollen diagram from Valle di Castiglione in the Alban hills 20 km to the east of Rome shows the development of Holocene forest and then, at 3480 ± 50 years bp, a fall in the concentration of arboreal pollen. The authors (Follieri et al. 1988; Magri 1989) ascribed this to environmental/climate change, but they also stated that human disturbance of the vegetation is evident in the zone above this level. In a detailed multidisciplinary study of the uppermost 10 m of this core, Alessio et al. (1986) argued that the appearance of beech (Fagus) around 5000 bp, and an increase in the sclerophyllous element, especially evergreen oak, around 3000 years bp, were both probably the result of increased human impact on the landscape. Other palaeoenvironmental studies at the junction between the Alban hills and the Pontine plain to the south have found evidence for vegetation change reflecting clearance activity c. 1000 bc, but with large-scale clearance taking place only in Roman times (Attema et al. 2000). In the Apennine mountains to the east, Lago di Fucino has low lake levels at ad 100–400 and ad 600–900, and a high level at ad 1100–1800 that likely relates to the Little Ice Age (D’Amato 1989; Dragoni 1996; Giraudi 1989). Further into the Apennines, in the Upper Sangro valley, significant human impact marked by deforestation began around 800 bc, according to

the regional palaeoenvironmental record

+ve

Mean temperature anomalies (°C)

a)

July Annual 5

January

Mean annual precipitation anomaly (mm)

-ve

b)

500

0 -200

0

High levels of Lake Bolsena

c)

3

6

Interpolated 14C age (10³ yr BP)

9

(200 BC — AD 200)

12

(AD 1500 — 1700)

Secular frequency distribution (FD)

0.06 0.05 0.04 0.03 0.02 0.01 0 -400

0

400

800

Years

1200

1600

2000

figure 3.3  Proxy palaeoclimatic records for the Tuscania region: (a) the Lake Martignano record, adapted from Kelly and Huntley (1991); (b) Lake Bolsena high lake levels, after Dragoni (1996); (c) the historical record from the Tiber, after Camuffo and Enzi (1996). (Illustration: T. Bacon.)

calibrated radiocarbon dates (Brown et al. 2013). This was the period of Samnite tribal confederacies, quasi-urban Iron Age societies contemporary with and in trading contact with the Etruscans (Lloyd 1995a). In conclusion, the lake pollen records of South Etruria, although not as sensitive to human impact as might be hoped, show significant disturbance of the natural

vegetation from around 3000 bp/1000 bc, followed by indications of a step change in deforestation and the appearance of the typical Mediterranean cultivars from around 2000 years ago or a little earlier, i.e. broadly coinciding with the beginning of the Roman Republican period. The lake level histories suggest a dry phase at the time of the initial impact c. 3000 bp/1000 bc, a cold wet

67

3 The Natural Landscape and Its Evolution

phase coinciding with the significant phase of human impact through the Roman Republican and Early/Mid Imperial periods (c. 200 bc–ad 200), and another during the Little Ice Age (c. ad 1500–1700), that correlate well with the record of Tiber floods and more widely in Italy (Giraudi 2012; Pepe et al. 2016; Walsh et al. 2019). The record for the Italian peninsula as a whole broadly parallels these trends (Stoddart et al. 2019).

little or no matrix. At site C2 on the downstream end of the same bend, 2 m of a sandy unit were exposed with laminated bedding, some containing small gravel clasts. The modern channel was filled with recent material including coarse sediment (mostly pumice) and pottery. Downstream there was no significant floodplain, only colluvial slopes extending to the channel. The banks were constructed of up to 3 m of brown clay-rich silt, without bedding or pottery. The easterly tributaries were almost entirely bedrock channels with only occasional pockets of alluvial deposits. Two sections were logged along the Pantacciano stream channel. Site P1a displayed a single alluvial unit: an unstructured brown (5 YR 4/3) sandy-silt with some clay, and blocky/prismatic ped development, overlay a basal pebble layer mixed with eroded and weathered tuff. The bedrock was unweathered tuff. The section was 30 m long, with an average depth of 2.5 m. Site P1b, 150 m upstream, revealed a similar but slightly deeper section (Table 3.1). Upstream of these sites, the Pantacciano is almost entirely incised into bedrock. Downstream, a terrace continues about a metre above the floodplain, but at lesser gradient, as the river became more incised. An exposure here revealed 5.5 m of alluvial sediment overlying tuff, indicating that erosion was exhuming an old valley. Although the valley of the Capecchio was generally confined by bedrock, two sites were located where the mini-transect formed by squares J16–J18 of the archaeological survey crossed the valley at Fontanelle delle Donne: site C1 (upstream) revealed about 2 m of fill, and site C2 (downstream) approximately the same. The summary of the two sections is shown in Table 3.2. From site M1 at Tuscania to Fosso Cadutella, there is a narrow strip of floodplain, before a bedrock-constrained section which ends with an alluvial reach at the junction

the alluvial stratigraphy of the marta valley Six sites were investigated in the middle and lower Marta valley within the area of the archaeological survey, and four further downstream (Fig. 3.4). The Marta has little or no floodplain until joined by the Fosso Acquarella south of Tuscania, and few associated sediments until joined by the Fosso Capecchio which forms the valley to the west of the town, and the Pantacciano and Leia streams draining the eastern part of the catchment including the western slopes of the Vico crater. From the town of Marta to the town of Tuscania, the valley is moderately steep, falling 140 m in 14 km, with rock benches, a gravel/boulder bed and no floodplain. The most upstream site on the Marta was M1, just upstream of the bridge below San Pietro. The incised channel revealed 3.5 m of alluvial fill, the top 2.5–3 m of which were dominated by a sequence of brown/dark brown (Munsell colour 5 YR 3/3) sandy silts fining upwards to clay-rich silts sitting unconformably on a gravel unit of unknown thickness. The silt units contained Roman pottery to the base, and the upper 0.3 m showed evidence of both soil formation and ploughing. The stream is incised into the floodplain, with a lower bench of recent age containing plastic refuse. There was a similar sequence at M2 a little downstream, though here the basal silts were grey brown (10 YR 5/3) due to partial gleying. Similar but less deep sequences were logged from the principal tributaries of the Marta. At site C1 on a bend by a ford on the Capecchio near the Fontanelle delle Donne in square J18, 1.5 m of crudely bedded sand and gravel with cobbles containing abundant modern tile overlay a silty sandy deposit 0.75 m thick containing lenses of gravels with

Table 3.1  The section at site P1b on the Fosso Pantacciano

68

Depth (m)

Sediment description

0–2.47 2.47–2.77

Brown sandy silt with some clay; blocky/ prismatic ped development Brown sandy silt with pebbles and weathered tuff

2.77–5.5

Tuff to stream level

the alluvial stratigraphy of the marta valley

Major areas of floodplain Gorge

Lake Bolsena

Lake Mezzano

Catchment boundary Topographic divide

Tuscania C1

Leia

Pa ntacc ia

Capecchio

Marta

no

Marta

M1 M2 P1

C2

ne

ro

Ar

Norchia M10

Lake Vico

Vetralla

Q1 M3 Corvena M5

ta

ar

M

M4

Tarquinia 0

10 km

figure 3.4  The catchment of the Marta river showing the Tuscania Archaeological Survey area and the geomorphological sites discussed in Chapter 3. (Illustration: T. Bacon.)

of the Marta and the Capecchio. This pocket of floodplain falls into the most southerly kilometre square of the South Transect. The valley bottom is a composite feature formed around the core of a large valley meander. The northerly edge is characterized by a break in slope associated with

the underlying marls, while the east contains the alluvial fill. There is a number of large blocks of tuff on the bench, which have probably been eroded from the slope to the north. An exposure here (site M10), just above the junction of the Marta with the Capecchio, revealed the contact

69

3 The Natural Landscape and Its Evolution

between the Miocene marl and the volcanics (Fig. 3.2). A spine of marl which sticks out into the southern floodplain here is responsible for the trajectory of the river. This structural feature also lines up with an exposure of the same marl on the western side of the floodplain. The stratigraphy of the exposure is summarized in Table 3.3. From the Capecchio junction downstream, a sinuous gorge section forms a natural barrier between the Tuscania and Tarquinia sections of the Marta valley. This gorge is formed by resistant Oligocene limestone. This reach must have controlled the upstream base level during the Holocene. The Corvena sluice at the end of this

gorge revealed extensive sections of the fill to a depth of 7  m (sites M3 and M4; Fig. 3.5). The upstream section, M3, has the most complex alluvial stratigraphy found in the Marta catchment (Fig. 3.6; Table 3.4). The upper sections show very variable and disturbed sequences, with Roman material within the top 2  m in some exposures, but Medieval and more recent material at the same depth in others. Underlying these silty sands there is a complex bedded unit which varies from sand to sandy gravel, which clearly indicates active migration by a meandering channel or channels. All the pottery from this unit is Roman. Two radiocarbon dates from the inset terrace

Table 3.2  Summary of two adjacent sections (sites C1 and C2) logged on the Fosso Capecchio Depth (m)

Unit

Sediment description

C1 0–0.90

1

0.90–1.5 (2 in places) 1.5–2.03

2 3

C2 0–0.85

1

0.85–2.10

2

Chaotic mixture of cobbles, potsherds, tile, pebbles and stones in a sandy matrix; elsewhere modern plastic rubbish evident. Seems to be a combination of post-WWII dumping with flood material. The cobbles are dumped and floods fill in the interstices Clast-supported cobbles and pot with finer fill – cobble dump Residual block of stratified sandy silt with occasional pebbles. Silt with fine sand, in some places clay; at 1.6 m a small stone layer. No pot sample taken. Two mottled colours: 10 YR 6/4, 7 YR 6/3. Blocky root holes. The main alluvial unit? Light brown sandy silt with clay; blocky; abundant worm holes, root channels; occasional stones (angular to sub-rounded, larger in top 20 cm). Pottery at 80 cm (sample C2, P1 – Roman?) Bedded grit with coarse sand, abundant small stones and pebbles. Horizontal laminations in the top 0.1 m. Below, small cross-bedded units. At 1.5 m, 4 cm-thick clay band, undulating, below more silty clay bands. All stones are matrix-supported; no imbrication; cross-bedded units of order of under 1 m in length. Stones volcanic, no limestone. Clay band at 2.05 m, water level at 2.07 m

Table 3.3  The alluvial stratigraphy of the exposure at site M10 Depth (m)

Unit

Sediment description

0–0.90

1

0.90–1.95

2

1.95–2.57

3

2.57–3.75

4

Sandy silt with a trace of clay, banded (silt/fine sand and clay drapes), dipping towards the river; root disturbed, worm holes. Plastic at 50 cm, very recent top of bank deposition Massive sandy silt, disturbed by roots but not fully bioturbated, faint evidence of silt/sand banding. Decorated potsherd found at 1.8 m (sample M10, P1) Horizontally bedded grit and small stones with sand and silt. Grit in elongated lenses dipping away from the channel. Clasts: limestone, volcanics. At base, 2 cm of silty clay under a 2 cm grit band. Abundant old root channels but no soil structure Crudely bedded, chaotic fabric: gravel, stones, pebbles, cobbles (max. 50 cm). Matrix supported in places. Orientation of discoid clasts generally horizontal, some big-small pebble ordering, many lithologies, limestone, volcanics, mud balls, clasts well rounded to subangular. Upper boundary very irregular, abrupt changes of over 50 cm in a few cm. Dip of this unit away from the river – old bar? Potsherd found at 1.7 m in gravel lens, possibly Medieval fine ware (sample M10, P2)

70

the alluvial stratigraphy of the marta valley

(see dating discussion, below) indicate that the incision into the floodplain pre-dates the fifteenth century ad. This section has many interesting features. The main body of the fill consisted of a basal sequence with crossbedding (unit 5) fining upward, then a thick ­horizontally bedded sand unit (unit 4), then an erosional ­bounding surface (a stratigraphic discontinuity) with finer silts above (unit 3). The finer silts contained an unusual wavy discontinuity identified as ploughing furrows, with several thin sand and stone lines, some of which were undulating, above. Inset into this stratigraphy was a bench with organic rich silty sands capped by a thin bed of gravel (units 7 and 6). This was interpreted as a cut-andfill cycle, which was confirmed by dating (see below). On a spur above the floodplain, a small quarry revealed a section cut into an extensive terrace forming the spur south from the gorge (Fig. 3.7). This was investigated in

order to establish a maximum age for the lower floodplain sequences. It has a complex stratigraphy of cross-bedded and tabular sands, gravels and cobbles, with both normal and inverse grading. A detailed stratigraphic description is given in Table 3.5 and summarized in Figure 3.8. In unit 5 a large boulder of Cretaceous limestone measuring some 10 cubic metres in size was embedded in the crossbedded sands, suggesting a catastrophic flow regime. The sand and gravel are capped by 1.4 m of a sandy silt of colluvial origin. A fragment of bone found at a depth of just over 4 m was used for radiocarbon dating, giving an estimated age of 15250 ± 60 bp (Beta-106624: see discussion of dating below). The bone was identified as a fragment of a thoracic vertebra of a large ungulate (A. Gouldwell: pers. comm.), and its C13/C14 ratio, indicative of uncontaminated collagen from a C3 plant-eater, confirmed its identification as an ungulate.

figure 3.5  The exposure at the Corvena sluice in the Marta valley, site M3 (see also Table 3.4). (Photograph: Tony Brown.)

71

3 The Natural Landscape and Its Evolution

0

M5

M4

M3

M10

M1 PW

1

R

19th M

2

R 3

cl ay m fin s ed e ilt s co ium and ar s se an sa d nd g st rit o pe ne bb s le s

cl ay fin e silt sa nd

6

420 ± 60 14th

PW 19th M 14th R

Post WW2 19th century AD Medieval 14th century AD Roman OSL date radiocarbon date

cl ay m fin s ed e ilt s co ium and ar s se an sa d nd g st rit o pe ne b s co ble b s bo ble ul s de rs

cl ay m fin s ed e ilt s co ium and ar s se an sa d nd g st rit o pe ne bb s le s

7

cl ay m fin s ed e ilt s i a co um nd ar s se an sa d nd g st rit on pe e bb s le s

835 ± 220

4

5

R R

figure 3.6  Generalized stratigraphies of alluvial sites in the Marta valley. The vertical scale is in metres. The radiocarbon and OSL dates are ‘before the present’. (Illustration: T. Bacon.)

Table 3.4  A summary of the exposures at site M3, downstream from the Corvena sluice Depth (m)

Unit Sediment description

0–0.42

1

0.42–0.79

2

0.79–2.38

3

Sandy silt with little clay. Blocky/prismatic jointing, worm holes, root penetration, snails, no banding, no stones. Estimated 30% sand, 60% silt and 10% clay. Colour: 7.5 YR 5/3. Boundary wavy to gradual, maximum 0.04 m. Decalcified Dark grey bands, commonly three but variable and subdividing; upper band is most developed – 0.08 m thick with worm casts just penetrating it. Coarse grit and very coarse sand with sandy silt (as Unit 1) in between. Bands contain small (c. 1–2 cm) pebbles, rounded tuff and pumice, limestone and some shell. Dark layers 7.5 YR 4/1. Banding is variable from coarse sand and grit to pebble and sand lenses. Single grain thickness laminations present in sand. Suggestion of fining upwards in some bands. Top boundary sharp, bottom variable from 0.5 cm to 10 cm, depending upon bioturbation. Decalcified Massive brown sandy silt with clay, same as Unit 1. Prismatic/blocky, same colour (7.5 YR 5/3), with shell remains and occasional pebbles (up to 6 cm diameter), limestone and volcanics. Carbonate c. 5%. Unit coarsening slightly at base (2.18–2.38), with some coarse sand and grit and very small pebbles including volcanics. At downstream end of exposure, ridges and furrows/swales present in upper 0.5 m of the unit, 4–5 in all visible. Negative features filled with coarse sand/grit as unit above. Features very regular: negative amplitudes upstream-downstream are 8, 7, 9 cm, positive amplitudes about 2 cm. Wavelength (ridge to ridge) is of the same order: 1.15, 1.49, 0.82, 1.45 cm. Excavation of one of the furrow bases showed soil

72

the alluvial stratigraphy of the marta valley

Table 3.4  (cont.) Depth (m)

Unit Sediment description

2.38–5.08

4

5.08–6.08

5

6.08–6.28

6

6.28–6.43 6.43–7.05

7

capping such as would result from rainfall impact on bare soil. These soils are prone to form a fine tilth, but also cap and crust. The interpretation is that the section cuts (probably slightly obliquely) a set of shallow ridges and furrows caused by ploughing, which were filled by flood-deposited sand and grit and then not reploughed until sometime later, when protected by flood silt or then too close to river. A series of weak grey sand bands in lower 0.55 m of this unit. Approximately four, each 1 cm thick, separated by sand/silt: (from bottom to top) 10, 5, 11 cm. Only pottery (Roman?) found at this level, at 2.88 m. At the level of these lower bands are large burrows/tree root holes filled with silt/clay from surface flooding and downwash Slightly cemented silty sand, fine to medium and well sorted. No stones but crude horizontal bedding, apedal. Little or no evidence of bioturbation. 10% + carbonate, and cementation increases towards base Cross-bedded and planar bedded sequence, fining upward from gravels at base to sand at top. Maximum cobble size c. 0.5 m. Pebbles subangular to rounded, including limestone and volcanics. Cementation of the whole unit. Pebbles and cobbles sub-imbricated (Inset terrace) Black brown grey organic silt with sand at top. Typical channel fill. Contains small plant remains. 7.5 YR 4/3 brown top, grey silt 10 YR 4/2 (Inset terrace) Pebble layer, coarse pebbles, clast-supported (Inset terrace) Grey organic silt bed, stiff, with large pieces of wood including root bases of birch (?), small twigs and seeds, together with occasional pebbles. Base is the gravel forming the channel bed Real base of entire unit unknown, but likely to be a minimum of 8 m (water level at 6.20 m)

figure 3.7  Site Q1 terrace exposure, logged as Table 3.5. Scale is in 20 cm divisions. (Photograph: Tony Brown.)

73

3 The Natural Landscape and Its Evolution

Marta

Q1 56m asl 0

60 Q1

1 50

2

40

3

4 30

5

T

Thoracic vertebra of large ungulate 15,250 ± 60 10,990 ± 1,570

20 6

? 7

cl ay f m ine sil ed s t iu an co m s d ar an se d sa nd g st rit on pe es bb co les b bo ble ul s de rs

10

0

figure 3.8  Site Q1 terrace stratigraphy and height relations. Vertical scales are in metres. The radiocarbon dates are bp (‘before the present’); see text for details. (Illustration: T. Bacon.)

74

DATING AND ALLUVIAL CHRONOLO GIES

Table 3.5  Section log of site Q1, a terrace section exposed in a quarry at Guado della Spina Depth (m) Unit 0–1.4

1

1.4–2.5

2

2.5–3.1

3

3.1–3.6 3.6–3.8 3.8–4.6

4 5 6

4.6–5.6

7

5.6

Sediment description Pedogenetically altered limestone up to sand with silt and pebbles: volcanic, 5 cm diameter Cross-bedded sand/grit with small pebbles, all rounded, mostly volcanic, with mica present Horizontally bedded sand/grit with some silt – laterally extensive across site Coarse well-sorted sand unit Mudstone bed with volcanic pebbles Cross-bedded coarse sand and grit with cobble lag at base; evidence of channels; bone fragment dated by radiocarbon (Table 3.6) Massive breccia-like sandstone with pebbles: sandstone, limestone pumice/tuff, ignimbrite Base of exposure

Note: Dip on massive units 2–3, direction 21, strike 110°.

Downstream of Corvena the valley widens and the floodplain reaches 1.5 km in width. The Marta has been channelized throughout this reach and its banks are graded and vegetated, obscuring the stratigraphy. However, the excavation of a large pit in the middle of the floodplain facilitated stratigraphic investigations at site M5 (Fig. 3.9). This revealed 3.6  m of a uniform dark-brown sandy silt lacking structure, textural variation, or inclusions. The trench, which was 2 m wide and extended for 30 m perpendicular to the river, showed no stratigraphic variation. Examination of the faces and spoil also produced no artefacts. The few observations possible from the channelized river banks also suggest at least 3.5 m of sandy silt above any underlying gravels or bedrock.

figure 3.9  The sediments of the lower Marta floodplain (site M5) exposed by the digging of an agricultural ditch. (Photograph: Tony Brown.)

inclusion into alluvium need to be evaluated. These are complex, including not only an upstream channel input and an input from scatters on the floodplain eroded into the channel, but also a direct input into the alluvium from manuring (Brown and Ellis 1996). One of the surprising results of the initial survey work in the Marta valley was the extremely low artefact content of some units: a detailed examination of over 1600 m2 of face in the silty unit at site M5, for example, did not produce a single artefact, yet contemporary river gravels in South Etruria, in the Treia valley, were as high as 10 per cent pottery and tile by weight (Brown and Ellis 1996). The pottery broadly suggests that the basal sands and gravels in the Marta valley are Roman in date, with no later pottery included. The

dating and alluvial chronologies (tb, with ce and er) In many ways dating has been the major problem in both environmental reconstruction in Mediterranean surveys and studies of landscape change and alluviation. The most commonly used non-chronometric method is pottery, particularly in areas where detailed typologies exist. However, in order for pottery to be utilized in even an approximate chronology, the controls on pottery

75

3 The Natural Landscape and Its Evolution

more variable and mixed coarse/fine units above the gravels contained Roman, Medieval and Renaissance pottery and the upper fine, overbank, unit contained Renaissance and Post-Medieval pottery. However, as can be seen from a comparison with the other dating techniques, the pottery dating was frequently in error, especially liable to overestimate the antiquity of the unit.

to date small quantities of carbon of known local origin such as identified macrofossils, a technique facilitated by the development of accelerator mass spectroscopy dating (AMS) where the carbon-14 (14C), or radiocarbon, isotope is measured directly rather than its radioactivity. It is, however, less common to find sites with well-preserved macrofossils in the Mediterranean than in northern Europe. A more subtle problem with radiocarbon dating in Mediterranean environments is a bias imposed by the conditions favourable to preservation. In the case of alluvial sediments, exposures are provided by river incision, and frequently organics are preserved at the base of the visible sequence, particularly on inset terraces. Lower organics are rarely obtainable and higher organics are destroyed by decomposition in the oxic or aerobic zone. This phenomenon may be the principal reason for the large number of Medieval dates recorded from such sections. Because of these problems, there is an urgent need for the application of other, preferably sediment-based, dating methods (Brown 1997, 2000). In this study, only two sites produced material suitable for radiocarbon dating, sites QM1 and M3, and one of the aims of the work was to trial other dating techniques. The radiocarbon dates from the Tuscania alluvial sites are shown in Table 3.6. As can be seen, the two samples from the inset terrace at M3 are conformable. The date from the terrace was from the collagen extracted from the ungulate vertebra using the AMS technique.

Radiocarbon Dating Radiocarbon dating is the most used chronometric technique, and alluvial or lacustrine sites are the most frequently used non-cultural contexts for radiocarbon dating of Mediterranean environmental change. There are three sources of potential error associated with the use of radio­ carbon dating to date alluvial environments: reworked or old organic material; pedogenic factors; and rootlet contamination. The extent to which these pose problems will obviously vary with the nature of the sample, and it is therefore necessary to consider the types of organics commonly found in floodplain and terrace sections that can be dated. Of most importance are wood, peat, organic mud and charcoal, especially in drier locations. Wood and charcoal produce the most reliable dates, but if not in situ will only give a minimum age of deposition, which in the case of charcoal could considerably pre-date the age of deposition of the sediment in which the charcoal is found. Observations of the exhumation of old tree stumps by natural riverbank erosion show that the reworking, or incorporation, of old wood can occur. This may be particularly common in Mediterranean environments, where decay can be limited by desiccation. Likewise, given the severe history of soil erosion that is typical in the Mediterranean, in-washed pedogenic carbon or charcoal fragments are common in both lacustrine sediments (such as Lake Monterosi) and alluvial sediments; clearly, these are likely to produce date reversals in conflict with their stratigraphic positions. Contamination by younger organic matter is also common in Mediterranean sites: the deep penetration of roots from exposures and banks makes radiocarbon unreliable in many cases. One of the ways of overcoming problems of contamination is

Palaeomagnetic Dating Palaeomagnetic dating has been used for many years for fired archaeological materials, but it can also be used on some natural sediments (Brown 1997, 2000; Tarling 1989). From the remanent magnetic signature of the material, an estimate is made of (a) its polarity (i.e. normal or reversed) and (b) the position of the magnetic North Pole. These are then compared with the known history of magnetic reversals and polar wandering, dated by other means. The most important characteristic of the sediment is that the natural remanent magnetism which is aligned to the ambient magnetic field must be locked into the grains at the time

76

DATING AND ALLUVIAL CHRONOLO GIES

Table 3.6  Radiocarbon dates from Marta valley alluvial sections Laboratory number

Site number

Depth (m)

Material

Age ± 1 sigma

Beta–85451

M3b/95

6.20 (inset)

detrital wood

420±60

UB–3692 Beta–106624

M3/TM3 MQ1

6.10 (inset) 3.90

wood bone collagen

370±43 15250±60

of formation or deposition. The most suitable terrestrial sediments other than lake deposits are aeolian or loess deposits, but a remanent magnetism may also be acquired by fine alluvial deposits (Batt and Nöel 1991; Clarke 1992; Ellis and Brown 1997). Although overbank silts could in principle acquire a detrital remanent magnetism, the lack of water column height and constant turbulence would seem to render this unlikely, and it is more probable that the signal is caused by post-depositional biophysical and biochemical processes. Studies have shown that palaeomagnetic dating seems to work in two situations: in relatively uniform sandy silts of overbank origin which have been rapidly accreted and relatively little bioturbated; and where the natural remanent magnetism is carried by detrital magnetic grains. These conditions are common in the Mediterranean region, with its history of severe soil erosion and corresponding high rates of alluvial deposition. Palaeomagnetic samples were collected from unit 5 (the grey silt) at site M3 (Fig. 3.5, Table 3.4). Seven cubic specimens were pushed into the field section of the Marta, and 89 were cut in the laboratory from a block sample taken at a depth of 5.75–6.29 m down from the floodplain surface. Analysis of magnetic anisotropy was carried out on all the specimens in order to ascertain the mode of particle alignment (hydrodynamic, gravitation, or geomagnetic forces) and the presence of post-depositional disturbance. The anisotropy of susceptibility results revealed that the grains were deposited from suspension under the influence of low velocity flow. Flow direction in the Marta unit sampled was fairly consistent, from 3.025  m down profile: E/W, SW/NE, SSW/NNE, WSW/ ENE, SW/NE, WSW/ENE, SSW/NNE and finally at 6.00–6.02  m E/W. It is probable, because the ellipses of anisotropy dip towards the south-west, that the general

Calibrated age ad 1450–1626

flow was from the north-east. As the degree of anisotropy of the specimens from the Marta was less than 3 per cent, the alignment of the coarse silt and fine sand grains was sufficiently weak so as not to affect the natural remanent magnetization, the basis for palaeomagnetic dating. The specimens were also analysed for their palaeomagnetic potential and were generally found to be good carriers of natural remanent magnetization. Some specimens from the Marta contained a weak secondary natural remanent magnetization, which was removed using alternating field (AF) demagnetization. Standard methods were used to obtain the characteristic remanent magnetization (Butler 1992; Irving 1964; Nöel and Batt 1990; Tarling 1989). The characteristic remanent magnetization derived from the Marta is good quality, with all but one of the suites of specimens having sigma 95 values less than 5 (Ellis 1995), which is the recommended cut-off value for palaeomagnetic data (Clarke et al. 1988). Three suites of specimens plot directly onto the Italian archaeomagnetic curve. However, this curve is limited in its duration, covering only ad 1510–1975 and 50 bc–ad 520 (Fig. 3.10a). The dates obtained for the Marta sediments lie between 500 bc and 50 bc. To check the validity of the dates obtained from the Italian reference curve, the raw data were corrected to Meridian using the method of Nöel and Batt (1990) and plotted onto the British archaeomagnetic curve (Fig. 3.10b). The range of dates is significantly smaller because of the higher resolution of the British archaeomagnetic curve. The declination and inclination coordinates for the Marta (samples 70–89) fit onto the British curve, yielding a possible date of 300–200 bc. This agrees with the broad pottery dating of the section, and other dated sections.

77

3 The Natural Landscape and Its Evolution

a)

45

Evans Rome Marta

Inclination

50

55

60

65

70 -30

-25

-20

-15

-10

-5

0

5

10

15

20

Declination

b)

50

Clarke et al. Clarke et al. Marta, M3 (5)

55

Inclination

60

65

70

75

80 -40

-30

-20

-10

0

10

20

30

40

Declination figure 3.10  Palaeomagnetic curves and the Marta samples: (a) the Italian curve; (b) the UK curve.

78

50

THE REGIONAL ALLUVIAL RECORD

Luminescence Dating

The related technique, optical stimulation luminescence (OSL), probably has a wider application to Mediterranean alluvial deposits, in part as it requires less sophisticated equipment (Bailiff 1992; Huntley et al. 1985; Rhodes and Aitken 1988). It was used in this study (Table 3.7). Visible light can be used to stimulate luminescence in quartz and infrared radiation with feldspars. As with TL, many factors affect the accuracy of ‘apparent’ luminescence ages, including mineralogy, light exposure before burial-darkening (Rhodes and Pownall 1994), overburden, erosion, and the nature of the radiation dose. This last factor is itself affected by sediment water content. Problems of radioactive disequilibrium were encountered in most of the Marta valley samples, meaning that the dose rate measured in the field (using a gamma-­spectrometer) was not in agreement with that measured by neutron activation analysis in the laboratory. Although not known for certain, the cause is probably related to the very high and variable dose rates from natural radioactivity in the area. However, good precision was achieved overall for fine-grained sediments (silts) and for K (potassic) feldspars. Some quartz samples were problematic due to feldspar inclusions causing scatter and zeroing correction problems.

All materials that contain, or are exposed to, radioactive substances such as uranium are continuously bombarded by radiation. This causes ionization and the trapping of electrons in the mineral. As radiation continues, trapped electrons build up, and if they can be measured and the radiation dose rate is known, then the time elapsed since the radiation started can be calculated. This measurement is possible because, if the mineral is heated, it emits extra light above that normally emitted proportional to the electrons trapped – thermoluminescence, or TL. If the heating is repeated, the extra light will not be given off, as electrons will not have had enough time to build up again. In addition to measuring the light emitted, the concentration of radioactive elements can be measured by mass spectrometry. Pottery is ideal for such dating because, if the temperature is high enough, firing reduces the TL signal to zero, as all trapped electrons are released when the pot is fired. However, sunlight can also reduce the TL signal of sediments to zero, so in this respect Mediterranean sediments are far more suitable for TL dating than northern European sediments. Thermoluminescence has been applied to river terrace deposits in semi-arid type climates outside the range of radiocarbon dating (Nanson et al. 1991). When TL dating was first applied to alluvial sediments, doubt was cast over whether they had been sufficiently zeroed to give the date of deposition (Berger 1984). Owing to the dynamics of the fluvial system and the generally short period of exposure of the sediment, the thorough penetration of light for all grains cannot be assumed (Bailiff 1992). However, Nanson and Young (1987) and Nanson et al. (1991) obtained TL dates from terraces of the Nepean river in New South Wales and from an alluvial sand below the floodplain of Coopers Creek in western Queensland which agreed well with other dating evidence. Their conclusion was that, if dates were from shallow sand flows deposited as sheets on floodplains, where residual TL and long-term sediment moisture contents could be accurately estimated, TL dating was reliable and could provide an excellent basis for alluvial chronologies.

the regional alluvial record: climate, people or both? South Etruria provided sites for Judson and for VitaFinzi in their pioneering studies of Mediterranean alluviation (Judson 1963; Vita-Finzi 1964; Vita-Finzi and Judson 1964). In his classic Mediterranean-wide Table 3.7  OSL dates from Marta valley alluvial sections Sample Site Estimated number number Depth (m) age ± 1 sigma

79

RO17

QT1

RO26 RO44

M11b M3

4.75

Comment

10,990 ± 1,570 Possibly a bit young when compared with the radiocarbon date 3.9 835 ± 220 Post-dates pottery 5.50 (inset) 75 ± 18 Rather young but in line with Post-Medieval radiocarbon date

3 The Natural Landscape and Its Evolution

synthesis, Vita-Finzi (1969) recognized two common alluvial units, one of which he demonstrated was of historical age and had a widespread occurrence throughout the Mediterranean basin, the so-called ‘Younger Fill’. This unit is often cut into, and partly derived from, an ‘Older Fill’, which is generally regarded as being of Pleistocene age. The Younger Fill is generally buff/grey/brown in colour and composed of silty fine sand with some rounded/ subrounded gravel lenses. It forms the floodplain of most Mediterranean valleys and is effectively a low terrace caused by Post-Medieval stream incision. In many cases, it was probably deposited relatively quickly by just a few flash floods, causing some erosion of existing valley sediments, of both channel and overbank origin. The major period of deposition was from the Late Roman to Early Medieval periods, between c. ad 300 and ad 1500. Vita-Finzi (1969, 1978; Vita-Finzi and Judson 1964) regarded the primary cause of Younger Fill deposition to be climatic change, a temporary southward shift in the depression belts of Europe. He cited as key evidence a decrease in the antiquity of the fill from north (45° N) to south (30° N). This interpretation of the data led to some debate with those who favoured a primarily anthropogenic origin (van Andel et al. 1986, 1990; Davidson 1980; Eisma 1964; Wagstaff 1981). The latter was argued especially on the basis of observations of a peak in deposition in the later centuries of the Roman empire, when it was known that agricultural decline took place and terraces commonly fell into disrepair. It was also pointed out that the diachrony of the Younger Fill could be the result of the complex response of alluviation to both climatic change and human impact (Lewin et al. 1991, 1995). The Younger Fill was, it was shown, in reality a complex unit or units and need not have the same primary cause in every location. Work in the Voidomatis basin in Epirus in northern Greece showed, for example, how much more complex the fill sequence was than the simple bi-partite division of Older and Younger Fills. This basin contains four terraces, distinguishable using lithological and mineralogical techniques, one of which can be directly related to glacial deposits in the headwaters (Bailey et al. 1990).

Four major episodes of aggradation were recognized, three of them Pleistocene (dated from first to last to before 150,000 years bp, c. 26,000–20,000 years bp, and c. 20,000–15,000 years bp), the fourth probably being related to overgrazing sometime around the eleventh century ad (Lewin et al. 1991). On the basis of this and other work in Greece (Pope and van Andel 1984), Lewin et al. (1991) questioned the validity and utility of the Older and Younger Fill model. In Italy, geomorphological work on the rivers draining the eastern slopes of the central Apennines in the Marche region set Holocene fluvial activity into a Pleistocene framework (Calderoni et al. 1991). In the upper Esino basin, thirteen fluvial phases were identified dating from the Last Interglacial c. 100,000 years ago to the present. In the Holocene, the difference in coastal and upper-middle valley reaches related not to sea-level or tectonic factors but to basin sediment output and the pattern of alluvial activity. The Musone river deposited a beach-barrier in the Bronze and Iron Ages, behind which sedimentation continued until Medieval times (Coltorti et al. 1991). Coastal recession due to fluvial deposition was also seen in the Misa and Cesano rivers (Coltorti 1991). Sections of these rivers were meandering from the Early Holocene to Roman times, but they then became braided as a result of an increase in sediment load (Coltorti 1991). The date of this transition, however, varies by as much as a thousand years between one basin and another in the region. In the Siena region of Tuscany, alluviation has been dated to later Medieval times and associated with land-use intensification (Hunt et al. 1992). Such variations are most easily explained by catchment factors over-riding or greatly modifying any climatic signal – in the case of the Marche valleys, an increase in sediment supply from spatially non-uniform deforestation in the Bronze and Iron Ages and again in the Renaissance. Further south, in the Biferno valley in Molise, Barker and Hunt (1995) and Hunt (1995a, 1995b) identified eight valley floors (floodplain surfaces) between later prehistory and the twentieth century, all of which correspond with periods of significant expansion in settlement and/

80

THE REGIONAL ALLUVIAL RECORD

or intensification in land use identified by the Biferno Valley Survey (Barker 1995a). Two significant periods of aggradation were the Samnite/Early Roman period (c. 300 bc–ad 200) and the Early Medieval period. The earlier aggradation is dated using pottery which, as discussed earlier, presents considerable problems of interpretation. The distinctive Medieval alluviation is radiocarbon dated to the eighth century ad (laboratory number: HAR-2557), although as mentioned above, such charcoal dates may also be problematical. Nevertheless, the repeated correlations between alluvial episodes and independent archaeological evidence for agricultural intensification in this valley are very striking. A regional approach to alluviation and land-use history in the southern Argolid, Greece, provided a complex but unusually complete picture for the Late Pleistocene and Holocene. Pope and van Andel (1984) identified eight alluvial units separated by periods of stability and soil formation. Three of these pre-date 6000 bc, and although they are assumed to be of climatic origin, they do not correspond to north European terrace chronology, as none dates to the Last Glacial Maximum, nor to the Late Glacial, nor to the Early Holocene, a period of postulated monsoonal intrusions into the Aegean (Kutzbach 1981). The earlier units are most probably associated with warmer, wetter periods, with little geomorphological activity during the colder steppe-climate periods corresponding to the northern hemisphere glaciations. The alluvial episodes post-dating 2000 bc were equated by van Andel et al. (1986, 1990) with cultural change, in particular periods of economic decline and a lack of terrace maintenance and soil conservation measures particularly in the Early Bronze Age and Hellenistic period. In the Southern Argolid and most of the Mediterranean, a distinction could be drawn between poorly sorted, ­matrix-supported, unstructured debris flow deposits and better-supported lenticular alluvial units. Van Andel et al. (1986) ascribed the former to periods of drier climate and less vegetation or extensive hillslope clearance and the latter to wetter conditions or the neglect of terraces producing gully erosion and concentrated run-off. The Argolid

study illustrated well how the complexity of both cultural and natural factors, and the interactions between them, could produce an environmental history lacking a one-toone match with either set of controls. Mediterranean rivers in their upper-middle reaches of young valleys naturally transport considerable quantities of bedload and are quite prone to changes in bed elevation relative to the height of the surface of the alluvial floodplain or terrace. Periods of rapid migration and erosion of older terrace material will produce an episodic record relating to local factors such as gradient, valley width, sediment supply and flood history. Owing to a combination of steep gradients and constrained width (both partly the result of the geological youth of the drainage network), compounded by the Mediterranean climate, valley history is characterized by aggradation and erosion cycles, producing cut-and-fill stratigraphy (Brunsden and Thornes 1979; Delano-Smith 1981). It is the last of these cycles that forms the Younger Fill(s), with evidence of earlier cycles either having been removed by erosion or being invisible beneath the floodplain water table. Studies of the stratigraphy of the lower reaches of valleys are more likely to preserve evidence of changes in the sediment budget of catchments in earlier periods, as indicated by the rapid regression of river mouths that so adversely affected Roman ports such as Ostia and Luni (Ward-Perkins et al. 1986). For example, stream valleys around Foggia in southern Italy have downstream evidence of downcutting in the Iron Age and aggradation in Roman times, whereas such evidence is lacking from the upper reaches (DelanoSmith 1981). Because the Mediterranean is a zone susceptible to erosion, due to high erosivities, erodibilities and tectonic activity, it needs only a small push for positive feedback mechanisms to lead to severe soil erosion. This has happened further back in the past before periods of major anthropogenic impact: the badlands of the Tabernas area in southern Spain, for example, pre-date the Bronze Age, the first period of dense settlement and intensive land use (Thornes and Gilman 1983). There is, however, little doubt that human manipulation and management of

81

3 The Natural Landscape and Its Evolution

lowest part of the crater rim at c. 350 m above sea level, the present lake level is 305 m above sea level and there is no geomorphological evidence of fluvial erosion by lake overspill. At the mouth of the river at the town of Marta is a sluice which regulates the contemporary flow of the Marta. The structure contains Roman tiles, is clearly of some antiquity, and probably has Roman origins. Given these observations, there is a strong probability that the upper part of the Marta channel is artificial and was cut at some time in classical antiquity. It would not be surprising if this is the case, given that studies of other crater lakes in Lazio, including Martignano, have shown that most were controlled by artificially constructed outlets, mostly tunnels, at various times between 500 bc and ad 100 (Castellani and Dragoni 1981, 1990, 1991). A tunnel would have been unnecessary for making an outlet on the southern side of Lake Bolsena because of the low level of the crater rim, and in any case was probably not feasible given the size of the lake and the technology then available (W. Dragoni, pers. comm.). Construction of the Marta sluice would have been most likely during a period of high lake level, perhaps around 700–400 bc, the high point of Etruscan power, when hundreds of cuniculi were constructed in the region (see Chapter 5), though it could also have been constructed in the next major phase of high lake level, between c. 200 bc and ad 200 (Fig. 3.3). On the basis of archaeological remains under and around Lake Bolsena, Dragoni (1996) postulated the following series of low lake levels: 1450–1250 bc, 1000–780 bc and 615–415 bc. During these periods there would have been no flow down the upper Marta, had its channel existed. This has important implications for ancient settlement around Tuscania. While dryland farming does not require irrigation, and farmsteads and villas probably relied on springs and cisterns for storing roof run-off (Thomas and Wilson 1994), the development of the town of Tuscania may well have increased local demand for water. The area around the town is poorly served by springs, and by constructing the Marta outlet on the southern rim of Lake Bolsena a perennial water supply would have been provided near to the town. Further work is required in the

the Mediterranean flora and the development of agriculture have had major effects on soil erosion by increasing soil erodibility, making even small variations in climate important in transporting this soil in pulses down valley, and in the subsequent trenching of these deposits (Barbero et al. 1990; Macklin and Woodward 2009; Pons and Quezel 1985; Thornes et al. 2009; Zanger 1992). The cultural factor is equally variable, due not only to population and technological changes but also to economic factors influencing the construction, maintenance or abandonment of terraces and the clearance of scrub and woodland (Brown et al. 2021; Wagstaff 1992). Work close to the Tuscania Archaeological Survey area, at Narce in the adjacent Treia basin (see Figure 4.1 for location), provides an intriguing contrast with the alluvial record described here and in the region (Brown and Ellis 1996). Trenches at Narce traversing the majority of the narrow floodplain contained abundant archaeological material, allowing Cherkauer (1976) to trace the thalweg (the line of maximum depth), both vertically and horizontally. The sequence indicated deposition sometime prior to 200 bc, erosion c. 200 bc–ad 200, deposition c. ad 200– 800, erosion c. ad 800–1000, and then deposition until ad 1800, after which incision occurred. The dating of the first recorded erosion phase (200 bc–ad 200) is synchronous with deposition downstream in the Treia catchment, deposition in the middle and lower basins of the Marta, and with the high lake level of Bolsena. The likelihood is that this major erosional signature reflects a combination of climatic and human agency: a shift to colder, wetter, conditions exacerbated by the impact on the landscape of the high rural populations of the Roman Republican and Early/Mid Imperial periods (Patterson et al. 2020).

the upper marta: natural or artificial? As suggested earlier, the geomorphology of the Marta river is anomalous. There is effectively no valley upstream of Tuscania, and walking the channel revealed that it was cut into bedrock. Although the channel crosses the

82

conclusion

Marta area and on the upper channel to confirm the general hypothesis that the upper channel of the Marta was created or enlarged in classical antiquity by the construction of an outlet from Lake Bolsena and, if so, whether the engineers who conceived and executed such a bold enterprise work were Etruscan or Roman.

may help to explain the substantial size and robustness of Roman structures in South Etruria such as bridges and viaducts (Potter 1979). The cause of this fluvial dynamism was probably fundamentally climatic, as postulated by Potter (1976b) from Cherkauer’s studies in the Treia basin (Cherkauer 1976). The climatic hypothesis is supported by the high level of Lake Bolsena c. 200 bc–ad 200 (Fig. 3.3). However, the continuation of sedimentation, predominantly of sands and fining-upwards units, was probably a function of increased sediment supply due to woodland clearance and increased arable cultivation in the later centuries of the Roman empire. The fluvial situation in the early post-Roman period mirrors the settlement archaeology in being poorly defined (Chapter 8), although several sites suggest a hiatus and then some incision, prior to increased sedimentation in the fifteenth, sixteenth and seventeenth centuries. There is substantial evidence that increased overbank deposition at this time was a response to an increase in flood frequency and magnitude, as illustrated by the Tiber flood record. The nature of the response was controlled by intensification in arable cultivation associated with the increase in population during the Renaissance and the ensuing growth of the city of Rome. The tendency for valleys to have filled with thick accumulations of overbank sandy silts is most marked in the small catchments to the north of Rome and to the south of Tuscania (Brown and Ellis 1996; Judson 1963). The south of the survey area has probably been more directly affected by geomorphic changes than the north. This work suggests that the floodplains within the area were hardly usable for arable farming during the Etruscan, Roman Republican and Early/Mid Imperial periods, only becoming of significant agricultural value in the post-Roman period as a result of the erosion of agricultural slopes. Like the regional vegetation record of South Etruria, the alluvial record of the Tuscania area suggests little significant human impact on the environment until the Roman Republican period. However, a combination of climatic conditions and land-use changes during the main centuries of the Roman Imperial period can be interpreted as

conclusion The earliest environmental information in the Tuscania area is from Late Glacial times, c. 15,000 years ago. Terrace fragments indicate erosion and rapid or even catastrophic transport and sediment deposition along what is now the Marta valley. The orientation and dip of the terrace beds and the topographic position suggest that the erosion was from the slopes of Monti Cimini around Lake Vico. The Late Pleistocene environment at this time consisted of a steppe landscape of low vegetation cover. In the Marta and the other valleys in the region, there is then a major hiatus in the depositional record, with considerable incision (approximately 40  m in the lower Marta) between the end of the Last Glacial and the Mid Holocene. This hiatus remains problematic, being interpreted alternatively as a major Late Pleistocene sudden erosional phase, or as a prolonged period when rivers were in a net erosional state. The overall rate of incision in the Marta would have been controlled by erosion of the resistant limestone reaches of the valley. The lack of sediments dating to the Early Holocene, even at sites extending to bedrock, and the pollen data indicating fully wooded conditions and hence supply-limited catchments would seem to support the hypothesis of progressive rather than abrupt erosion. The oldest Holocene sediments found in the Marta date to the Late Etruscan period c. 500–300 bc. During the Roman Republican and Early/Mid Imperial periods, the Marta and other rivers in the area were unstable braided and wandering gravel-bedded rivers, depositing coarse- to medium-calibre gravels and occasionally small boulders. Debris flows were also common in steeper valleys. These fluvial conditions, so unlike the modern rivers,

83

3 The Natural Landscape and Its Evolution

triggering the high landscape sensitivity which has characterized the region ever since, and which preconditioned the landscape to alluviation in the Late Medieval and PostMedieval periods. In the Pontine region south of Rome, too, geomorphological and palynological studies indicate that the first definite signs of anthropogenic impacts on the landscape were early in the first millennium bc, at

the time of the archaic Latial cities contemporary with the Etruscan city states north of the Tiber, but that major landscape change did not occur until Roman times, as in South Etruria (Attema et al. 2000). As in South Etruria, too, these major transformations to the countryside are thought to reflect the combined effects of climatic change and large-scale clearance.

84

4 PREHISTORIC LANDSCAPES Graeme Barker, Francesco di Gennaro and Tim Reynolds

introduction

sample sizes have to be recognized, the material collected makes a fundamental contribution to our knowledge of the prehistory of the study area, as well as informing on wider debates regarding prehistoric Etruria from the beginnings of the first farming communities (Period 1 in the periodization shown in Table 2.1) to the development of the ‘Villanovan’ Early Iron Age societies (our Period 6) that were the foundation of Etruscan state formation (Fig.  4.1). In this chapter dates are cited as bp (before the present) for the prehistoric periods before the Neolithic, and thereafter as bc (so a 5000 bp radiocarbon date becomes 3000 bc). All dates prior to the Iron Age are based on the calibrated radiocarbon chronology.

In common with other surface surveys of the volcanic regions of South Etruria, a feature of the Tuscania Archaeological Survey was the marked rarity of prehistoric lithic material found by the survey teams – some 200 worked artefacts in total – even though the teams generally included members experienced in recognizing prehistoric lithics in ploughsoil. (As a comparison, over 3,000 lithic artefacts were collected in a 35 km2 survey area on the Adriatic coast of Albania by the Mallakastra Regional Archaeological Project: Runnels et al. 2009.) Some 2,200 sherds of prehistoric pottery were collected, however, from 159 survey units. The total sample of material available for study is still small – over half of the collected units of prehistoric pottery (116) consisted of fewer than 10 sherds, there were 30 units with 11–50 sherds, 10 units with 51–100 sherds and only 3 units with more than 100 sherds. No significant concentrations of lithic artefacts were found anywhere, most such occurrences consisting of isolated artefacts. The low number of units with prehistoric material compares with the more than 550 units with Etruscan pottery and almost 1,400 units with Roman pottery. The disparities are all the more marked given that the prehistoric units span a period lasting at least 5000 years, whereas the c. 1400 years of Etruscan to Late Antique settlement yielded some 54,000 sherds. However, the discovery rate of units with prehistoric pottery was surprisingly consistent in the different sampling units, varying from 1.4 units per km2 in the Judgement Sample to 1.8 per km2 in the Transect Sample and averaging at 1.7 per km2 for the survey area as a whole (Table 4.1). (These densities are in terms of the total areas designated for field-­walking, whereas, as we explained in Chapter 2, we in fact only searched some 40 per cent of their combined area.) While the difficulties caused by the

the chipped stone collections (tr) All materials were examined in natural light, using a hand lens with x 5 magnification when appropriate. The following nine variables were recorded: 1. raw material; 2. blank type, including identification of primary, secondary and tertiary blanks and whether a piece was whole or fragmentary; 3. striking platform type: whether the platform was plain, cortical, crushed, lipped, prepared or dihedral; 4. edge damage: the presence of gloss, micro-flaking, lateral and transverse snaps, macro-flaking, crushing and half-moon snaps, also rolling, striations and abrasion; 5. retouch: its presence, location and form (for example, nibbling, abrupt, scaley and plano-convex); 6. maximum dimensions: length, width, thickness measured using calipers in millimetres (weight was not recorded, but as this has a direct relationship with volume this was not necessary); 7. presence of formal tool types such as scrapers, points, knives and burins; 8. blank form: flake, blade, bladelet and chip; 9. miscellaneous technological components

85

4 Prehistoric Landscapes

Neto di Bolasse

M NE NNI APE

Grotta all’Onda Romita di Asciano o

Arn

Consuma

Chiana

S TAIN OUN

Montarrenti

San Marco Tiber

Colline Metallifere

Lake Trasimene Petriolo Pienza

Populonia

Grotta dell’Orso

Mount Amiata Vetulonia

Om

br

on

Fiora

e

Grotta Lattaia

Paglia

Pitigliano

Orvieto (Volsinii)

Volsini Mountains

Grotta Bella

Al

AR

be

M

gn

a

Poggialti Vallelunga Castro, Ischia di Castro Visentium (Grotta di Carli) Grotta Grottadelle delleSettecannelle Settecanelle

A

M

EM

Poggio Olivastro

Mount Argentario

Lake Gran Carro Bolsena

Sorgenti della Nova

Tuscania Musarna

Vulci

Rinaldone Castel d’Asso (Valle Arcione)

Cimini Mountains

SABINA

Rogge di Canino Grotta del Vannaro, Corchiano Norchia Riparo Vetralla Grotta di Monte Venere Biedano Civita Castellana Tarquinia Cenciano Diruta a and rt Lake Tre Erici and Ma Faliscan La Ficoncella Luni sul Mignone Vico Nepi caves ne Narce Tolfa (Pian Cisterna, no ig Pian Conserva) M Bufalareccia Lake

TYRRHENIAN SEA

Tolfa Grotta Patrizi Mountains

Cerveteri (Caere)

Bracciano

La Marmotta

Tiber

Ar

ro

ne

Giglio Island

Rebibbia

AN GN M PA RO AM C

Veii La Polledrara di Cecanibbio

Rome

A

Tenuta di Torrenova Torre Spaccata and Casal Bruciato

50

0 km

Land over 200 m

figure 4.1  Etruria, showing the principal prehistoric sites and other locations outside the Tuscania Archaeological Survey area mentioned in Chapter 4.

86

THE CHIPPED STONE COLLECTIONS

Table 4.1  Density of locations where prehistoric pottery was found in the Tuscania Archaeological Survey Sample unit Random Transect Judgement

Size (km2)

Locations with prehistoric pottery

Locations per km2

36 40 20

59 73 27

1.6 1.8 1.4

discussed below in terms of raw materials, edge conditions, typology and technology. Comments on the lithic material of particular periods of prehistory in relation to the other data on prehistoric settlement are made as appropriate later in the chapter.

Raw Materials

Note: The densities are in terms of the total areas designated for field-walking, not the c. 40% of these actually searched.

There were seven raw materials discerned: basalt, chert, flint, obsidian, quartz, quartzite and slate. There were only single pieces of basalt, slate, quartz and quartzite, six pieces of chert and twelve pieces of obsidian, the rest of the collection being of flint of varying quality, small pebbles predominating. The small size of the sample means that it is not possible to identify patterns of preferential selection of raw materials, except perhaps in the case of the obsidian (discussed under ‘Technology’ below). This is also the only obviously exotic material, as deposits of this volcanic glass occur at restricted locations in the Mediterranean, none in peninsular Italy. Obsidian was widely traded throughout Italy in the Later Neolithic (4500–3500 bc) and the Chalcolithic, or Copper Age (c. 3500–2200 bc). Extensive characterization programmes indicate that the material found in the peninsula generally derives from the islands of Palmarola (off the coast of Lazio), Lipari (off the coast of Sicily) and Sardinia (Hallam et al. 1976; Tykot 1996). Work at the quarry sites suggests that they were not closely controlled by local people; probably both local islanders and mainland communities with seagoing craft were able to access the obsidian sources directly, trading finished pieces to communities further afield for whom it would have been a highly exotic material (Robb and Farr 2005; Tykot et al. 1999; and see below, ‘Technology’).

including utilization, presence/absence of Janus flakes, Siret flakes, double-bulbs, plunged pieces, hinges, burning, core edge-trimming flakes, crested blades, microburins, platform/core tablets and impact fractures. A cursory attempt was made at refitting pieces within a collection when sample size allowed. All tools (i.e. artefacts deliberately altered to suit a function) were drawn at 1:1. The lithic collections recovered by the survey are small and difficult to characterize. Material collected from 145 survey units was examined, but only 113 such units had lithics that were definitely humanly worked. The 206 worked pieces from the many square kilometres of terrain field-walked reflect extremely low densities. (In arable areas of Britain and elsewhere, surveys commonly record a background of 1–3 pieces per 100 m2.) Only three units produced more than ten pieces of worked flint, the remainder of the sample generally consisting of one or two pieces. The 32 tools represent about 15 per cent of the total struck stone pieces in the collection. As a percentage of any single assemblage, 15 per cent would be within expected limits, but spread across 113 sample units this figure severely limits any information that can be gleaned about how the landscape was exploited. The same is true in terms of technology: cores comprised 4 per cent of the collection, again an acceptable frequency for a single sample of this size, but spread across a landscape they can reveal little behavioural patterning. No preferential blank production, core maintenance or specialized reduction could be identified. No clear associations indicative of activity sets beyond basic flaking could be discerned. No collection alone could be considered to be a ‘site’ in a behavioural sense. The combined sample is therefore

Edge Conditions The edge conditions of the different collections were varied, with all states of wear from heavily rolled and abraded to fresh being present. Indeed, edge conditions varied within collections from a single unit and cannot be used to generalize about depositional environments. Most pieces showed some traces of wear and there were a few that could have

87

4 Prehistoric Landscapes

been utilized, their wear reflecting cutting activities. One had cereal gloss. Signs of classic plough damage, such as the complex edge nick, were rare, with most damage being attributable to crushing through soil processes and trampling. There was no sign of impact damage (the kind present on projectile heads, for example) on any of the pieces.

RS38:2 (Fig. 4.2:1) and, at the other end of the chronological spectrum, gunflints from, e.g., RS18:7 (Fig.  4.2:3) and J18:2 (Fig. 4.3:13). Several tools have a Middle Palaeolithic appearance, such as the sidescrapers from RS27:17 (Fig.  4.3:5) and T86:19 (Fig. 4.3:10). Backed blades and bladelets (e.g. RS12:25: Fig. 4.2:5) and microliths (e.g. R14:13: Fig. 4.3:11) could all fit within assemblages spanning the Late Pleistocene to the Mid Holocene, so they are shown as ‘Epipalaeolithic-Neolithic’ in Table 4.2. Two sickles (J3:39 and RS28:3: Fig. 4.4:1) indicate cereal harvesting, a stone tool-based activity most likely to date from the Neolithic to the Early Bronze Age. Despite these typologically distinct

Typology A total of 32 chipped stone tools was recovered (Table 4.2; Figures 4.2–4.5). The only truly characteristic tools in the collection are the barbed and tanged arrowhead from

Table 4.2  Chipped stone tools recovered by the Tuscania Archaeological Survey (compiled by Tim Reynolds) Tool type

Survey unit

Barbed-and-tanged arrowhead Tanged bladelet Microlithic element Retouched bladelet Borer Sidescraper Sidescraper Flake knife Sickle element Sickle element Retouched flake fragment Bilaterally retouched bladelet Bilaterally retouched blade Gun flint

RS38:2 J8:7 R14:13 J8:7 R34:22 T86:19 RS27:17 T24:6 J3:39 RS28:3 J3:39 J3:38 R34:9 RS18:7

Scraper on quartz blade Scraper Scraper

J18:2 J16:10 J17:10

Scraper

J17:11

Truncated blade Truncated flake Scraper Scraper Sidescraper Scraper Backed blade segment Backed bladelet Transverse arrowhead Endscraper Burin Burin Knife/scraper

J17:11 J3:36 J3:3 R14:1 T64:4 RS16:11 J16:7 RS12:25 RS18:14 RS39:27 R34:9 R34:9 C99:4

Possible age Late Neolithic – Early Bronze Age Epipalaeolithic – Neolithic Epipalaeolithic – Neolithic Epipalaeolithic – Neolithic Late Neolithic Middle Palaeolithic? Middle Palaeolithic? Middle Palaeolithic? Epipalaeolithic – Early Bronze Age Epipalaeolithic – Early Bronze Age Epipalaeolithic – Neolithic Epipalaeolithic – Neolithic late eighteenth/early nineteenth century

Middle Palaeolithic?

88

Epipalaeolithic – Neolithic Epipalaeolithic – Neolithic Epipalaeolithic – Neolithic

THE CHIPPED STONE COLLECTIONS

1

3

2

5

6 4

8

9 7

3

0 cm

figure 4.2  Stone tools from the Tuscania Archaeological Survey: 1. barbed and tanged arrowhead (RS38:2); 2. bilaterally retouched blade (R34:9); 3. gunflint (RS18:7); 4. borer (R34:22); 5. backed bladelet (RS12:25); 6. scraper (RS16:11); 7. knife/ scraper (C99:4); 8. endscraper (RS39:27); 9. conical scraper (R14:13). (Illustrations: Tim Reynolds.)

89

4 Prehistoric Landscapes

1 2

4

3

5

7

6

9

8

11

12

10

3

0

13

cm

figure 4.3  Stone tools from the Tuscania Archaeological Survey: 1. scraper (J17:10); 2. knife (T24:6); 3. scraper (T64:4); 4. scraper (J18:2); 5. scraper (RS27:17); 6. scraper (J17:11); 7. scraper (J16:10); 8. truncated flake (J3:36); 9. scraper (R14:1); 10. scraper (T86:19); 11. microlithic element (R14:13); 12. transverse arrowhead (RS18:14); 13. gunflint (J18:2). (Illustrations: Tim Reynolds.)

90

THE CHIPPED STONE COLLECTIONS

4 1 3

2

5

8

6 7

9

12 11 3

0

10

cm

figure 4.4  Stone tools and utilized pieces from the Tuscania Archaeological Survey: 1. sickle (RS28:3); 2. utilized bladelet (RS22:7); 3. backed blade (J16:7); 4. scraper (J3:3); 5. burin (R34:9); 6. bilaterally retouched blade (T82:2); 7. retouched bladelet (J8:7); 8. truncated blade (J17:11); 9. utilized bladelet (J17:12); 10. bilaterally retouched blade (J3:38); 11. tanged bladelet (J8:7); 12. burin (R34:9). (Illustrations: Tim Reynolds.)

tool forms being present, however, there is insufficient patterning in their associations and frequency to allow for accurate use in establishing chronology. About half the tool count is based on blade or bladelet blanks. Given the predominance of flakes in the blank

collection, this suggests preferential selection of blades and bladelets for tool production, although sample sizes and visibility of blades to fieldworkers may also have affected such patterning. The forms of retouch used for shaping tools are limited to simple direct flaking and

91

4 Prehistoric Landscapes

backing, with pressure flaking present on the barbed and tanged arrowhead. Stepped and scaley retouch are absent, which might suggest that refreshing tool edges was not a significant factor in the use of materials. A number of functions is indicated including scraping, cutting, boring, harvesting and hunting, but given that spatial patterns in this material are as likely to derive from discard as from activity, it is not possible to discern patterning in the distribution of activities across the landscape.

category and the fact that it was being worked down to small stubs suggests that it was a highly valued exotic item in this part of Italy, though it did not receive special attention in terms of knapping, being subjected to direct hard hammer percussion just like the other materials. Where obsidian was a rare and exotic material in Mediterranean Neolithic and Chalcolithic communities, it appears to have been used for cutting soft tissues, though very probably its meaning and cultural significance shifted with distance from point of origin (Robb 2007). Potential uses cited include butchery, human scarification and cutting and shaving hair (Broodbank 2013: 231). The frequency of crushed and plain platforms suggests the use of the hard hammer technique, while the presence of two Siret flakes could support this as well. Shatter is poorly represented (only six pieces), but this could be a result of field sampling. The same may be said for burnt pieces. The latter are frequent, as prehistoric knapping often took place around hearths; there were 26 burnt pieces, and more might be expected. There seems to have been little intensive input to lithic exploitation: there is no sample which could represent a knapping area alone. Most of the collection would seem to represent chance finds from a landscape characterized by a low intensity of lithic use in prehistory. There are certainly no signs of quarrying, specialized tool manufacture, or repeated use of materials. There is no Levallois technique present (the technique commonly used in Middle Palaeolithic technologies in Italy as elsewhere: Mussi 2001) and prepared platforms are rare (one out of 88). There is a single core tablet, while platform edge rejuvenation flakes are absent; all raw materials appear to have been treated the same way. The only pattern visible for reduction is to be found in the frequency of bladelet cores (five out of eight pieces), but their products are not as well represented in the collection, a phenomenon perhaps reflecting, if not sampling error, the fragility of small bladelets. The presence of bladelet cores might be another indication that occupation sites were generally rare in the Tuscania area in the earlier (i.e. Palaeolithic) periods when stone was the primary material for tools and weapons, because people venturing into the volcanic

Technology Technology can be monitored by studying cores, flaked lumps and characteristics of blanks. The total number of cores recovered was eight, 4 per cent of the total collection (Fig. 4.5). There is a clear over-representation of bladelet cores when the frequencies of recovered blanks are considered: there were 97 flake blanks used, 29 blades, 12 flake-blades and only 34 bladelets. Not only are flakes and flake-blade cores under-represented, but blade cores are absent. This could be a product of sampling but is perhaps more likely to indicate that a large amount of material was being imported into the survey area as blanks, with bladelets the only exception to this. This hypothesis is supported by the different frequencies of primary, secondary and tertiary blanks: there were only two primary pieces compared with 60 secondary and 144 tertiary ones. The roughing out and early stages of reduction would be expected to create a significant number of primary pieces, but these appear under-represented in the collections. Equally, there are only two cortical platforms among the blanks, which are dominated by plain and crushed platforms (59 and 24 examples respectively). Further field testing would be needed to resolve this, but the lack of good chipping stone in the volcanic regions of South Etruria, and the need to import it, are likely to be one of the main reasons for the rarity of lithic material found in many surveys in this region compared with other parts of Italy (di Gennaro and Stoddart 1982). Flaked lumps were also recovered from RS18:15, RS39:27, R34:9, J18:2 (two), RS18:2 and RS18:15, the latter two of obsidian. The frequency of obsidian in the flaked lump

92

THE CHIPPED STONE COLLECTIONS

1 2

3

5

4

7

6

3

0

8

cm

figure 4.5  Cores from the Tuscania Archaeological Survey: 1. obsidian bladelet core (RS18:15); 2. flake core (R34:9); 3. bladelet core (T54:15); 4. bladelet core (R34:9); 5. flake core (J6:8); 6. bladelet core (J8:7); 7. bladelet core (R34:9); 8. bladelet core (R34:9). (Illustrations: Tim Reynolds.)

93

4 Prehistoric Landscapes

zone needed to come with pre-prepared cores for retooling while carrying out activities such as hunting and gathering rather than being able to rely on picking up suitable local materials.

Fabric 4 (2.7%) Manufacture Colour of core Colour of margins Colour of surface Feel Fracture Inclusions (frequency; size) Surface treatment

the pottery (fdg, gb) The prehistoric pottery was examined by FdG and GB and nine fabrics were defined in terms of colour, inclusions and surface treatment (Table 4.3). All the inclusions can be found in the local clays and the similarities between many of the fabrics should be stressed. In some cases

Fabric 5 (2.4%) Manufacture Colour of core

Table 4.3  The fabrics of the Tuscania Archaeological Survey prehistoric pottery and their frequencies

Colour of margins Colour of surface

Fabric 1 (48.6%) Manufacture Colour of core Colour of margins Colour of surface Feel Fracture Inclusions (frequency; size) Surface treatment

Feel Fracture Inclusions (frequency; size) Surface treatment

handmade varies from red to orange to brown varies from red to orange to brown varies from red to orange to brown rough jagged quartz (high; 1mm); black grits (frequent; >1mm) smoothed, occasionally burnished Fabric 6 (1.5%)

Fabric 2 (12.8%) Manufacture Colour of core Colour of margins Colour of surface Feel Fracture Inclusions (frequency; size) Surface treatment

handmade black to brown to orange black to brown to orange black to brown to orange rough jagged grit (frequent; >2mm); mica (rare; >1mm); quartz (rare; >1mm) smoothed

handmade dark brown to black brown to orange dark brown to black rough to semi-smooth jagged quartz (common; 1mm) often well burnished Fabric 8 (0.8%)

Manufacture Colour of core

94

handmade dark grey to black

THE POT TERY

Table 4.3  (cont.)

assigned to the first three fabrics. Furthermore, Fabrics 1 and 2 are very similar in their composition and treatment, being essentially the same except for the larger size of white grits in Fabric 2. In combination they make up almost two-thirds of the total sample. Fabric 3, while similar in composition to Fabrics 1 and 2 in its mix of inclusions, is markedly better fired, and sherds of this fabric are commonly burnished. Fabric 7 seems a finer version of Fabric 3, with a dark core and frequent burnishing but without grit inclusions. While most of the prehistoric pottery collections consisted wholly or principally of undiagnostic body sherds, the frequency of diagnostic sherds such as rims, bases, decorated sherds and handle fragments (each given a C prefix and a unique number in the cataloguing) was sufficient to enable us to assign the majority of the larger collections to a particular period or periods with a reasonable degree of confidence (Table 4.5). However, the comparison of the fabric frequencies in these units did not reveal a simple trend through time in terms of either the changing proportions of the more common fabrics or the presence or absence of the minor fabrics (Table 4.4).

Fabric 8 (0.8%) Colour of margins Colour of surface Feel Fracture Inclusions (frequency; size) Surface treatment

dark grey to black grey rough jagged mica (infrequent; >1mm); grit (infrequent; >1mm); quartz (very common; >2mm) smoothed Fabric 9 (0.5%)

Manufacture Colour of core Colour of margins Colour of surface Feel Fracture Inclusions (frequency; size) Surface treatment

handmade grey/brown grey/brown red/brown rough jagged grit (rare; >1mm); quartz (rare; >1mm); mica (frequent; ~2mm) smoothed

the fabrics defined are probably our artificial separations of different mixes of quartz, mica and grit inclusions of pottery that were fundamentally the same to the ancient potters in terms of source material and manufacturing process. Ninety per cent of the pottery, in fact, can be

Table 4.4  Frequencies (%) of the major (F1–3) and minor (F4–9) pottery fabrics in the principal prehistoric units Site

N

P1

P2

RS28:3 J3:3 T88:11 R54:1 R34:9 RS18:3 C99:8 CP94:7 RS33:8 CP94:5 R14:60/70 J18:2 J18:1 RS33:12 R14:13

117 72 127 66 68 86 50 147 42 66 69 68 45 58 91

A A C

A B A B

C

B

C

P3

B B A C B

P4

P5

P6

B C C B A B B B B

C A B A B B B A

C

C A A

F1

F2

F3

F4–9

59.8 51.4 54.3 53.0 29.4 77.9 44.0 47.6 9.5 75.8 46.3 45.6 35.5 44.8 42.9

20.5 30.6 1.6 24.3 41.1 3.5 43.5 3.0 4.3 20.6 26.7 17.2 3.2

11.1 18.0 32.3 9.1 25.0 10.5 48.0 8.8 90.5 16.7 44.9 22.0 17.8 15.5 47.7

8.6 11.8 13.6 4.5 8.1 8.0 4.5 4.5 11.8 20.0 22.5 6.2

Note: N = number of sherds. Periods (P): 1. Earlier Neolithic; 2. Later Neolithic; 3. Copper Age; 4. Earlier Bronze Age; 5. Later Bronze Age; 6. Iron Age; A – definite; B – probable; C – possible. Connecting lines indicate that material could belong to different periods.

95

4 Prehistoric Landscapes

Table 4.5  Period and site classification of Tuscania Archaeological Survey units with prehistoric material

Lithics Transect Squares North T84:3 T64:2 T63:2 T63:3 T44:11 T34:13 T34:12 T24:6 T14:22 T4:2 T4:6 T4:11 CP94:3 CP94:4 CP94:5 CP94:6 CP94:7 CP94:8 CP94:10 CP94:22 East T85:29 T86:19 T88:1 T88:8 T88:9 T88:11 T88:18 T89:1 T89:3 T89:4 CC13:102 South R14:13 R14:15 R14:31 R14:60/70 R24:18 R24:22 R34:8 R34:9 R34:10 R34:11 R34:22

P

P

P

P

P

P

P

Period 1 Earlier Potsherds Period 0 N Pre-Neolithic Neolithic

1 1U 3U 1 2U 1U 2 13 1 3 7 66 1 66 28 148 1 5 11 2 13 17 34 127 9 10 24 24 1 107 20 10 69 1 13 1U 68 2U 24 11

Period 2 Later Neolithic

Period 3 Chalcolithic

Period 5 Period 4 Later Earlier Bronze Age Bronze Age

Period 6 Iron Age

●?

○

○

●? MP?

●?

●?

●?

●?

●?

●?

●

○ ●? ●?

●? ●? ○

●?

●?

●?

●?

○ ●?

○ ●? ●? ●?

○ ●? ●?

○ ○

○ ●? ○ ○

○ ●?

○ ●? ●?

●? ●? ●? ●

○

○

MP?

● ●?

○

Epi/Neo

●

Epi/Neo

○

●?

96

●?

○

○

○

●?

●?

○ ● ●? ●?

THE POT TERY

Table 4.5  (cont.)

Lithics R34:25 R44:21 R54:1 R54:3 R54:6 R54:11 R54:13 R54:20 R64:1 R64:5 R64:6 R74:1 R74:2 R74:3 R74:5 R74:7 R94:3 R94:5 R94:10 R104:3 West T82:3 C101:2 C100:4 C100:5 C99:4 C99:8 C98:16 C97:13 C97:17 C95:108 C95:110 Random Squares RS1:1 RS1:8 RS1:10 RS3:9 RS8:1 RS12:20 RS12:25 RS12:26 RS15:3 RS18:1 RS18:2 RS18:3 RS18:5

P

P

P P P P

Period 1 Earlier Potsherds Period 0 N Pre-Neolithic Neolithic

14 16 13 3 3 1 1 5U 15U 25 86 6U

Period 3 Chalcolithic

○

1 9 66 12 6 9 1U 2U 17U 6U 8U 3U 4U 5 18U 7U 2U 2U 1U 2U 2 2U 3 6 1 50 1 1 24 1 1

Period 2 Later Neolithic

○ ○ ●? ○ ○

○

○ ○

Period 5 Period 4 Later Earlier Bronze Age Bronze Age

●? ●? ○

Period 6 Iron Age

○

○ ●? ○

○

○

○

○

○

●? ○ ○ ○ ○ ○

○ ○ ○ ○ ○

○ ○ ●? ○ ○ ○

○ ○

●? ○

○ ●? ○ ●? ○

Epi/Neo

●?

●?

97

● ●?

○

○ ○ ○

○ ○

○ ○

○ ○

○

○

○ ○

○

4 Prehistoric Landscapes

Table 4.5  (cont.)

Lithics RS18:11 RS18:14 RS18:16 RS19:15 RS19:16 RS22:5 RS22:16 RS24:55 RS24:58 RS26:1 RS26:2 RS26:11 RS26:17 RS26:19 RS26:20 RS27:5 RS27:17 RS28:2 RS28:3 RS28:5 RS28:7 RS29:5 RS29:6 RS29:8 RS30:4 RS32:1 RS32:4 RS32:9 RS32:10 RS33:1 RS33:5 RS33:7 RS33:8 RS33:10 RS33:11 RS33:12 RS33:15 RS33:16 RS34:5 RS35:21 RS35:26 RS36:29 RS36:30 RS37:1 RS38:2 RS38:5

P

P P

P

Period 1 Earlier Potsherds Period 0 N Pre-Neolithic Neolithic 1 3 1 1 9U 15 10 1U 2U 3 6U 5 5 4 2U 6 23 8U 9U 20U 1U 4U 2U 2U 1U 4U 6U 1 2 5 42 13 14 58 3U 5 2U 1U 1 1 1U 12 13

Period 2 Later Neolithic

Period 3 Chalcolithic

Period 5 Period 4 Later Earlier Bronze Age Bronze Age

Period 6 Iron Age ○

Epi/Neo

○

●? ●?

●?

○ ○

●?

○

○

○ ●?

○

○ ○

MP? Sickle

○

●? ●

●

○ ○ ○ ●? ●? ○

●?

○ ○ ○ ● ●? ●? ●?

○ ●?

○ ○

●

○

●?

●? ?

?

98

●?

○

THE POT TERY

Table 4.5  (cont.)

Lithics RS39:3 RS39:22 RS39:24 RS39:27 Judgement Squares J1:11 J2:13 J2:15 J3:1 J3:2 J3:3 J3:22 J3:32 J3:34 J3:36 J3:37 J3:38 J3:39 J4:1 J4:23 J7:8 J8:7 J8:10 J11:4 J12:2 J15:6 J15:7 J16:7 J17:11 J18:1 J18:2 J18:5 J18:6 J20:20 J20:23

P P

P

P P P P

P P

P P P P P P

Period 1 Earlier Potsherds Period 0 N Pre-Neolithic Neolithic 1U 1U 3U 2

6 36 6U 1 2U 72 3U 6U 3U 5U 3U 7 2 1U 1U 13 1U 1U 2 3U 2U

45 68 4 4 2 3U

○

Period 5 Period 4 Later Earlier Bronze Age Bronze Age

Period 6 Iron Age

○

●?

○ ○

○

○

○

●?

●?

●?

●?

○

●?

○

○

●?

Epi/Neo

Epi/Neo MP?

Period 3 Chalcolithic

○

●

Epi/Neo Sickle

Period 2 Later Neolithic

○ ●? ●? ○ ○

●? ●? ●? ○ ○

○ ○ ○ ○

Note: Lithics P – presence (MP – Middle Palaeolithic; Epi/Neo – Epipalaeolithic/Neolithic); Potsherds N – numbers of sherds (U = p ­ robably prehistoric but undiagnostic); Period 0 – Pre-Neolithic (lithics); Period 1 – Earlier Neolithic 5500–4500 bc; Period 2 – Later Neolithic 4500–3500 bc; Period 3 – Chalcolithic or Copper Age 3500–2200 bc; Period 4 – Earlier Bronze Age 2200–1400 bc; Period 5 – Later Bronze Age 1400–950 bc; Period 6 – Iron Age 950–700 bc; ● – definite site; ●? – probable site; ○ – possible site. Lithic periodization after Table 4.2.

99

4 Prehistoric Landscapes

Although Fabrics 1 and 2 are preponderant in all phases, the one obvious trend in Table 4.4 is the increasing frequency through time of the finer Fabric 3, that is very similar to Etruscan impasto. It makes up almost 50 per cent of the collection from, for example, the late prehistoric and Etruscan settlement zone on the southern margins of Colle San Pietro (R14:13 and R14:60/70). However, Fabric 3 is also common in some of the units belonging to the Neolithic and Chalcolithic, and very rare at some Bronze Age units. Factors behind such discrepancies might include the production of different kinds of wares for domestic, ritual and/or funerary purposes, and/or for different social groups, and the importation of pottery from outside the region. The outcome is that it is impossible to use fabric frequencies as a simple guide to chronological period for the many units with small collections of undiagnostic body sherds, even though we may suspect that most of those with Fabric 3 sherds are more likely to belong to later rather than earlier phases of prehistoric pottery use. In the summary list of prehistoric sites (Table 4.5) the many units with such pottery are shown simply as ‘prehistoric’. Even the larger collections of prehistoric pottery are very small compared with, for example, many prehistoric assemblages found by the Biferno Valley Survey, where the surface collections from the principal prehistoric settlements frequently numbered hundreds of sherds (Barker 1995a, 1995b). Much of the disparity probably reflects differential destruction caused by the contrasting agricultural histories of the two regions. The volcanic soils of South Etruria, which are both fertile and relatively easy to work, have been a prime agricultural resource for thousands of years, as the settlement systems reconstructed by our survey demonstrate. In contrast, while initial (Neolithic) agricultural settlement in the Biferno valley was even earlier, most arable land use there in prehistory and in the historical periods concentrated on the limestone soils especially, with large areas of poorly drained heavy soil being drained and taken into systematic cultivation only some fifty years ago. Deep ploughing using caterpillar tractors was being introduced into the greater part of the Biferno valley only during the 1970s,

and the large sherds with fresh breaks found at many prehistoric sites recorded by the Biferno Valley Survey indicated that buried settlement layers were being actively destroyed and their material culture being brought to the surface around the time when the survey teams were searching the ploughsoil. In contrast, mechanized agriculture was well established in South Etruria by the 1930s, so the prehistoric sites we found in the Tuscania Archaeological Survey had been exposed to half a century of such ploughing. Most handmade prehistoric pottery is much more friable than the wheel-made pottery produced in later periods, and more vulnerable to breakage and weathering. In the Montarrenti Survey in northern Tuscany, where a small valley was surveyed every year for five years (Barker et al. 1986), there was one example of a prehistoric site found in the first year as large sherds with fresh breaks that became progressively smaller and more abraded in the second, third and fourth seasons of fieldwork. In the fifth year, though, there were large fresh sherds again on the surface as the plough cut deeper into hitherto undisturbed layers. Presumably many of the prehistoric assemblages we found in the Tuscania Archaeological Survey had endured repeated cycles of such exposure and destruction. In the Biferno Valley Survey, the wealth of prehistoric material retrieved from the ploughsoil (pottery, lithics and groundstone tools, fragments of daub from wattleand-daub structures and so on) and other characteristics of the find locations, such as patterning in the area and density of finds, were such that it was possible to separate the assemblages with reasonable confidence into ‘sites’ and ‘sporadic material’. The former were mostly taken to reflect significant locations of either domestic activities – settlements of some kind – or funerary activities, and the latter to be evidence for ‘off-site’ activities such as hunting and pastoralism. A typical small (c. 150  m2) Bronze Age settlement produced about 200 sherds from the c. 30 cm-thick ploughsoil when the site was first discovered, and limited excavation then found below the ploughsoil some 6,000 sherds in shallow but intact stratified deposits including pits cut into the subsoil full of discarded household debris such as pottery, flint, fragments of animal

100

pre-neolithic activity (‘period 0’)

bone, charcoal and ash. It was calculated that if there were similar densities throughout the unexcavated portion of the site, the total assemblage buried from view would number some 180,000 sherds (Barker 1995a: 140). Very few of the prehistoric collections made in the Tuscania Archaeological Survey measure up to similar criteria for ‘settlement sites’; fragments of daub, for example, were found at only one site (J3:3). Nevertheless, given the severe processes of attrition that have clearly affected the survival of prehistoric material in the Tuscania ploughsoil, it seems reasonable to conclude that most of the locations where prehistoric pottery was found, even in small numbers, should be taken as prima facie evidence for domestic activities (or burials, in some cases) rather than as isolated occurrences such as the pot accidentally dropped by the weary traveller, or smashed in a ritual, or for some reason secretly buried away from a settlement, etc. In contrast, occurrences of isolated lithics without associated potsherds have been taken to represent ‘offsite’ activities such as hunting, pastoralism, or clearing or working land for cultivation. Exactly the same conclusion was drawn by the Boeotia Survey in Greece, in a comparable lowland region with a long history of intensive arable farming (Bintliff et al. 1999): occurrences of even a few prehistoric sherds on the surface were assumed to be the vestiges of buried domestic sites characterized by pits dug into the subsoil, the sherds being brought to the surface from the pits by the action of the plough, whereas sporadic occurrences of lithics on their own were classified as evidence of off-site activities. While the Tuscania prehistoric data are certainly limited, their similarities with the Boeotia Survey data encourage us to explore their implications for changing settlement patterns and processes in the remaining sections of this chapter.

lower sections of the Tiber valley that form its boundary, giving us a framework of occupation history in which to contextualize the prehistoric sites we found. The locations of the most important of these are shown in Figure 4.1. Early humans probably reached the Italian peninsula as early as elsewhere in southern Europe, around 1.5 million years ago (Arzarello 2018; Dennell and Roebroeks 1996). Two evocative Lower Palaeolithic sites near Tuscania on the Tyrrhenian littoral are La Polledrara di Cecanibbio (Santucci et al. 2016) and La Ficoncella (Aureli et al. 2016), where hominins had butchered an extinct form of elephant (Palaeoloxodon antiquus) almost 500,000 years ago. Whether the hominins who made the stone tools found with the carcasses had managed to drive the animals into soft ground, or were scavenging animals killed by other predators, is unclear (Lemorini 2018). The first widespread evidence for hominin settlement begins with the Middle Palaeolithic, the period of the Neanderthals. The origins of the Neanderthals go back to at least 400,000 years ago, but in Italy there is a significant expansion in the evidence for Middle Palaeolithic occupation especially during the warm humid phases of the Last Interglacial or Marine Isotope Stage 5 that is globally dated to 130,000–71,000 years ago (Aureli and Ronchitelli 2018; Mussi 2001; Peretto 1992). Several Middle Palaeolithic occupation sites have been excavated in the Tiber valley around Rome such as an elephant butchery site at Rebibbia, Casal de’ Pazzi (Anzidei et al. 1984), along the Tyrrhenian coast of Lazio and Toscana (Tuscany), and also east of the Tuscania Archaeological Survey area, for example in caves to the east of Lake Vico near Civita Castellana. Most of these date to MIS 5 and/or to MIS 4, the latter a period of oscillating climates with episodes of marked cold and aridity globally dated to 71,000–57,000 years ago (Aureli and Ronchitelli 2018). Studies of faunal samples and lithic collections from caves south of Rome indicated that Neanderthal groups practised a mixture of scavenging and hunting (Stiner 1991, 1994), though there is increasing evidence for the sophistication of Neanderthal hunting systems at least in the later stages of their history before their extinction around 45,000/40,000 years ago (Brown et al. 2011). They probably ranged across

pre-neolithic activity (‘period 0’) A few lithics of typical Palaeolithic or Mesolithic type were found by the survey teams, as mentioned earlier, but no discrete assemblages of lithic material. However, a number of Palaeolithic and Mesolithic sites have been found elsewhere in South Etruria and the middle and

101

4 Prehistoric Landscapes

the landscape in small groups (around 25–50 people?), that stayed more or less within the same territory each year (Kuhn 1991; Spinapolice 2018). The occasional presence of such mobile Neanderthal groups in the Tuscania Archaeological Survey area is indicated by the flake knife from T24:6 (Fig. 4.3:2) and scrapers such as those from RS27:17, J17:11 and T86:19 (illustrated respectively as Fig. 4.3:5, 6 and 10). The first Upper Palaeolithic (blade-based) industries in Italy date to around 45,000–40,000 years ago (Mussi 2001). These, like similarly dated lithic assemblages elsewhere in Europe, have been regarded as indicating the more or less contemporary dispersal into the region of anatomically modern humans – Homo sapiens – as part of their dispersals out of Africa. This consensus has been complicated by the recognition of some Homo sapiens traits in skull remains found in the Apidima Cave in Greece dated to around 200,000–150,000 years ago (Harvati et al. 2019), part of the increasing evidence for successive dispersals by archaic Homo sapiens into Eurasia. A variety of Upper Palaeolithic blade-based industries were used in Italy from their first appearance during MIS 3 ­(57,000–29,000  bp) through the main phase of MIS 2 (29,000–15,000 bp). This period was characterized by a progressively cold and arid climate that culminated in the Last Glacial Maximum (LGM) around 20,000 years ago, when global sea levels fell to around 130 m below their present levels because of the expansion of ice caps, and in central Italy permanent snow lines descended from the Apennine peaks to about 1,000 m above sea level. Populations seem to have expanded considerably as the climate ameliorated after 18,000 bp: there are many ‘Epigravettian’ (Late Palaeolithic and Epipalaeolithic) sites throughout Italy, as populations expanded from LGM refugia (Guerreschi 1992; Mussi 2001). Pollen analysis shows that the landscape of South Etruria, as throughout lowland Italy at this time, was dominated by dry open grassland (Alessio et al. 1986; Follieri et al. 1988; Frank 1969; Magri and Sadori 1999; Narcisi 2001). People survived in these harsh conditions especially by hunting migratory herd animals such as steppe horse (Equus hydruntinus) and red deer. Small parties of hunters probably foraged away from the main

camps in what are termed logistical systems of hunting, ambushing the herds at intercept locations, such as narrow passes and gorges, during their seasonal migrations between winter lowland and summer upland grazing areas, and killing them with spears tipped with small ‘backed’ lithics (i.e. with the non-cutting edge battered or blunted) (Bietti and Stiner 1992; Donahue 1988). Similar logistical foraging systems have been proposed for Greece (Runnels 2009). Rock shelters with Epigravettian material dating to the Terminal Pleistocene (the period of deglaciation between the LGM and the transition to the Holocene c.  11,700 years ago) have been excavated on the southeastern edge of the Tuscania Archaeological Survey area at Riparo Biedano near Norchia (Pennacchioni and Tozzi 1985) and to the east at Valle Arcione near Castel d’Asso (Brocato et al. 1998; Giacopini and Mantero 1998), and an open site was investigated to the east of the survey area at Cenciano Diruta, Vignanello (Pennacchioni and Tozzi 1984). All three are in typical intercept locations in gorges of the streams that flow down from the Monti Cimini to the coastal plain. The flint assemblages consisted predominantly of backed blades and bladelets. We found a few such tools as isolated pieces in the survey (e.g. Fig. 4.2:5, Fig. 4.3:11 and 4.3:12; Fig. 4.4:3 and 4.4:7). Many Epigravettian sites have been found across South Etruria between the sea and the Middle/Lower Tiber valley, both small rock shelters and open sites (Bietti et al. 1983). As mentioned earlier, it is impossible to be sure of the precise date of our material as the technology continued to be used by Mesolithic hunter-gatherers in the Early Holocene, as well as to a limited extent in the Neolithic, but the likelihood is that at least some of the backed pieces we found in the survey represent tools used by Late Glacial hunters searching for migratory game like the herbivore found in the Q1 geomorphological exposure (Fig. 3.8). Skeates (2017) suggests that many of the small caves and rock shelters used by Epigravettian hunter-gatherers were probably special places in their social landscapes, for activities such as rites of passage, as well as (or because of?) being good camping places for hunting.

102

TRANSITIONS TO FARMING, c. 5500–3500 bc

Following the transition to the Holocene, over the next few millennia permanent snowlines retreated to their present positions, sea levels rose to present levels and forests expanded over most of the landscape; oak forest dominates the Etrurian pollen diagrams at this time (see Chapter 3). Mesolithic societies in central Italy killed game such as red deer, roe deer, cattle and pig, and in upland areas species such as chamois and ibex, as well as smaller animals and birds. From the ninth millennium bp onwards, there is widespread evidence for a greater emphasis on fishing and shellfish collection (Mannino and Richards 2018). There is hardly any evidence for plant gathering, but that is almost certainly due to the poor survival of organic materials. One or two of the microlithic tools we collected could be Mesolithic, but the general absence of evidence for Mesolithic settlement in inland South Etruria is striking, suggesting that the dense forests that developed on these volcanic plateaulands in the Early Holocene were much less attractive for settlement than the coasts, the major river valleys such as the Tiber, and inland lakes such as what is now the Fucine basin, all of them ‘mosaic environments’ favoured by Mesolithic hunter-fisher-gatherers (Martini and Tozzi 1996).

and economic relationships with the existing populations (Barker 2005; Broodbank 2013; Forenbaher and Miracle 2005; Martins et al. 2015; Robb and Miracle 2007). Before the major impacts of radiocarbon dating in the late 1960s and early 1970s, Italian Neolithic pottery was used to define three phases of Neolithic settlement. The Early Neolithic was defined by impressed ware, pottery decorated with impressed and incised patterns of a kind that was commonly found at the beginning of the Neolithic ceramic sequence in the Mediterranean basin and taken as a signature of the arrival of agricultural colonists: ‘impressed ware people’. There are few impressed ware sites in Etruria, notable assemblages coming from an open settlement at Pienza near Siena (Calvi Rezia 1969, 1972) and, closer to Tuscania, from the Grotta delle Settecannelle near Viterbo (Ucelli Gnesutta and Bertagnini 1993). The Middle Neolithic was defined by a series of painted wares in the south and east of Italy and by plainer fabrics in the west and north. The western and northern pottery is very similar and is usually termed Sasso-Fiorano to reflect this: Sasso refers to the Grotta Patrizi at Sasso Furbara, a cave with burials located south of the Tuscania Archaeological Survey area, and Fiorano to an open site on the southern margins of the Po plain. These broad regional differences were seen to continue into the Late Neolithic, with various painted wares and finely burnished wares predominating in the south and east and dark burnished wares in the north and west, the latter both generally named after the northern settlement of Lagozza. The application of radiocarbon dating and increasingly better stratigraphic control in excavation demonstrated a considerable degree of regional and chronological variation in these ceramic sequences (Skeates 1994; Skeates and Whitehouse 1994). Studies of animal bones and plant remains associated with them have shown considerable variability in the timing, character and scale of the onset of farming (Pluciennik 1997). The first pottery was being used as early as 6500 bc in parts of southern and eastern Italy, associated, for example, at the famous ditched villages of the Apulian Tavoliere with definite evidence for food production (Skeates 2000) interpreted as evidence

transitions to farming, c. 5500–3500 bc (c. 7500–5500 bp) The Neolithic is commonly defined in Europe as the period when pottery and polished stone tools first appear in the archaeological record, and when farming began. In Italy as elsewhere in Europe, though, there have been major debates about the timing, nature and reasons for people transitioning from foraging (hunting, fishing and gathering) to farming: whether agriculture was introduced by Neolithic people migrating from the Levant, and/or adopted by the existing population of Mesolithic foragers. New excavations combined with bone isotope studies of diet and migration and insights into demography and genetic relationships from ancient DNA suggest a process of ‘leap-frogging’ of small migrant groups across the Mediterranean, who developed complex social

103

4 Prehistoric Landscapes

of farming communities newly arrived from the Greek mainland. The impressed ware assemblage in the Grotta delle Settecannelle has also provided a date in the mid seventh millennium bc, but in most of Etruria the first Neolithic pottery does not appear in the archaeological record until about a thousand years later and the evidence for farming being associated with its use is much more mixed (Barker 1999; Pluciennik 1997). In many parts of central Italy, too, the first pottery assemblages included from the outset both ‘Early Neolithic’ impressed wares and ‘Middle Neolithic’ wares, for example at Portonove-Fosso Fontanaccia on the Adriatic coast dated to ­5900–5200 bc (Conati Barbaro 2019) and the remarkably preserved lakeshore settlement at La Marmotta on the edge of Lake Bracciano dated to 5700–5200 bc (Fugazzola Delpino 2002a, 2002b). The Neolithic sequence for Etruria is still probably the least well understood and dated in Italy. One well-­ excavated Neolithic settlement in the region is San Marco outside Gubbio on the eastern fringe of Etruria in Umbria (Malone and Stoddart 1992), where impressed and incised sherds (some 13 per cent of the assemblage) were associated with material with affinities to (though not identical to) the Sasso repertoire, dated to c. 5500– 5250 bc. Less certainly, both Pienza-type impressed ware and Sasso material were associated in the Grotta dell’Orso in Tuscany (Grifoni and Cremonesi 1967), dated to c. 5200–4800 bc (Alessio et al. 1973), while Sasso material was dated to the late sixth millennium bc at the Grotta Bella in Umbria (Delpino and Fugazzola Delpino 1987) and the early fifth millennium bc at San Rossore in Tuscany (Grifoni Cremonesi 1987). The present indications are that ‘Early Neolithic’ impressed wares and ‘Middle Neolithic’ Sasso material were both being used in this part of Italy in the second half of the sixth and first half of the fifth millennia bc (di Gennaro 1998). There are a few rather unreliable dates for Lagozza-type pottery in Etruria in the later fifth and earlier fourth millennia bc (Skeates 1994: 70). There is a fifth millennium bc date for incised and painted pottery found in the Grotta di Monte Venere near Lake Vico, but the reliability of the association is not clear. Hence, the following discussion considers

104

the Neolithic material from the Tuscania Archaeological Survey simply in terms of an Earlier Neolithic phase c. 5500–4500 bc (our Period 1) and a Later Neolithic phase c. 4500–3500 bc (Period 2).

Earlier Neolithic Settlement, c. 5500–4500 bc Nothing was known of Neolithic presence in the region of the Tuscania Archaeological Survey prior to the project (Fugazzola Delpino 1987; Negroni Catacchio 1987), but we recorded 22 units with Neolithic pottery: 7 from the Random Sample, 12 from the Transect Sample and 3 from the Judgement Sample. The most notable feature of the distribution is the general preponderance of sites to the south of Tuscania (Fig. 4.6). For the Earlier Neolithic, impressed ware sherds at R24:18 (Fig. 4.7a) and J3:3 (Fig. 4.7b, the latter associated with a typical Earlier Neolithic scraper: Fig. 4.4:4) are within the Pienza tradition and jars with dimples or small lugs near the rim at RS28:3 and J4:1 are typical Sasso forms. Possible (much damaged) impressed ware sherds were also found at RS26:19 and R34:25, and a sherd from the later settlement of R54:1 (C92 in the typology: Fig. 4.7c) may also belong to this phase. The shapes represented are mostly typical baggy storage jars and containers, together with (in the case of the Sasso material) finer rounded bowls and gently carinated cups. The Tuscania Earlier Neolithic material is typical of the regional ceramic traditions. Apart from Pienza and Sasso, good analogies can be seen between the Tuscania material and that found by surveys and excavations immediately to the north in the Fiora valley (Negroni Catacchio 1987, 1988a), especially from excavations at Poggio Olivastro near Canino (Bulgarelli et al. 1998); from earlier excavations at the Grotta dell’Orso near Sarteano (Grifoni and Cremonesi 1967) and the Grotta Lattaia in Umbria (Grifoni Cremonesi 1969); to the south, from survey finds in the Tolfa hills at localities such as Bufalareccia (di Gennaro 1998: 74) and at La Marmotta, and in the environs of Rome such as Tenuta Torrenova and Quadrato di Torre Spaccata (Anzidei 1987; Bietti Sestieri 1984).

TRANSITIONS TO FARMING, c. 5500–3500 bc

Capec chio

300 m

Arrone

m

RS1:8 200 m

0

20

RS1:10

Marta

300 m

RS28:2

RS28:3

200 m RS29:8 T88:11

C100:4

T89:4

hi

o

m

cc

RS33:16

pe

0

10

Ca

100

m

T88:18

R24:18

J18:1

R34:22

J3:3

R34:25 R44:21

0m

RS39:27

R54:6

10

R54:1 R54:3

J4:1

10

Marta

0

m

RS18:3

Earlier Neolithic

Later Neolithic Definite

0

5 km

Probable Possible

figure 4.6  The Tuscania Archeological Survey: distribution of Neolithic units. Shading denotes the areas walked.

Most of our Earlier Neolithic sites are in rather similar locations, generally being situated below the plateau edges overlooking watercourses. Of the principal impressed ware sites, for example, J3:3 overlooks a fosso, the dry valley of a seasonal stream or torrent that flows

into the Marta, R24:18 is on a natural platform overlooking the Marta (Fig. 4.8) and RS26:19 is above the floodplain of the Arrone at a confluence with one of its tributary valleys. J4:1, one of the main Sasso sites, is on the eastern (left) bank of the Marta, again on the scarp below

105

4 Prehistoric Landscapes

a

b

0

c

5 cm

figure 4.7  Some Earlier Neolithic pottery from the Tuscania Archaeological Survey: (a) R24:18; (b) J3:3; (c) R54.1. (Photographs: Graeme Barker.)

the plateau rim. The other (RS28:3) is on a spur above the confluence of two streams that form the Fosso Arroncino di Pian di Vico, the principal tributary of the Arrone. The Sasso site J4:1, a small rock shelter which may well have been a burial chamber like the eponymous site (though no human bones were visible to our survey team), was also in a mid-slope position overlooking the confluence of the Marta with one of its dry-valley tributaries. Whenever the distribution of surface material provided indications of an occupation area, it was invariably extremely small, c. 500 m2 or less. At J3:3, for example, the material was found within an area measuring c. 30  m x 17  m, with most of the finds (including daub from wattle and daub walling as well as artefacts) within a zone measuring less than 10 m x 10 m. Similar concentrations were observed at RS26:19. The rock shelter J4:1 was also very small (10  m wide by 5  m deep), the finds coming from disturbed soil within the shelter and from the ground immediately in front of it. The nature of the transition from hunting to farming in Etruria and the extent to which Earlier Neolithic people here were committed farmers are unclear (Barker 1999).

The discovery at Petriolo near Siena of Early Neolithic pottery in association with a microlithic lithic industry of classic Mesolithic type dated to c. 5000 bc (Donahue et al. 1993) suggests that one process was the adoption by the indigenous population of hunter-gatherers of elements of the ‘Neolithic package’ from their neighbours before they finally switched to farming. Without animal bones or plant remains from our Earlier Neolithic survey sites at Tuscania, or dietary information from isotopic studies of bones, for example, we can make no assumptions about whether their inhabitants were hunters, or farmers, or both, but the balance of the evidence indicates that most people in this period were combining small-scale mixed farming with hunting and gathering. San Marco, for example, consisted of pits and occupation debris suggestive of a cluster of simple shelters, in part semi-­subterranean (Malone and Stoddart 1992). Organization at the household level was inferred from the thin section analysis of a sample of sherds (Skeates 1992). The botanical remains indicate the cultivation of a variety of cereals (emmer, einkorn, club wheat, bread wheat and six-row barley), pea and vetch,

106

TRANSITIONS TO FARMING, c. 5500–3500 bc

figure 4.8  The typical location of an Earlier Neolithic site below a plateau edge overlooking a watercourse: R24:18, above the Marta, looking north towards Tuscania in the distance. The Etruscan farm we excavated, R24:19 (Guidocinto), was to the north on the same terrace. (Photograph: Graeme Barker.)

and the gathering of a variety of fruits and nuts including blackberry, raspberry, elder, grape, fig, wild plum and hazelnut. Flax was also collected, presumably for its oils and fibres, and there were numerous seeds of edible plants such as Galium, Papaver, Polygonum and Rumex that may have been either crop weeds gathered along with the cereals during the harvest or collected separately from fallow ground. The faunal sample was dominated by the bones of domestic sheep and goats, cattle and pigs, with game represented by red deer, roe deer and hare. The killing ages of the domestic stock suggested a relatively unspecialized husbandry system in which meat was the main resource rather than the ‘secondary products’ of the live animal such as milk and wool (Malone and Stoddart 1992). Pollen and molluscan analysis show that the San Marco encampment was on damp, probably seasonally

waterlogged, soils in a well-wooded landscape. The location is typical of many Early Neolithic farming sites in the Mediterranean. It is presumed that, inhabiting a wooded landscape poorly suited to extensive stock-keeping, early farmers relied heavily on the crops they grew in garden patches around their settlements (using hand implements such as hoes and mattocks and digging sticks to cultivate the ground), while maintaining a diversity of food sources to ensure against crop failure and food shortage. The preference of the Earlier Neolithic sites at Tuscania for locations by watercourses and alluvial soils accords with this model of intensive small-scale cultivation combined with herding, hunting and gathering (Bogaard and Styring 2017). Excavations of Earlier Neolithic occupation sites in the region invariably produce evidence for rudimentary pit

107

4 Prehistoric Landscapes

shelters rather than substantial houses, as at San Marco in Umbria and in the Biferno valley in Molise (Barker 1995a), the nearest example to Tuscania being the pit complex excavated at Poggio Olivastro (Bulgarelli et al. 1998). The communities who inhabited these sites are thought to have been characterized by segmentary tribal structures without marked ranking (Robb 2007; Whitehouse 1995). The survey evidence from Tuscania is consistent with this model in indicating a pattern of very small residential units consisting of one or two households, with no evidence for larger agglomerations such as hamlets or villages, or of richer or poorer sets of material culture at different habitation sites. The wealth of evidence missing from such sites is emphasized by the extraordinary survival of organic materials at La Marmotta that included not just abundant plant remains and animal bones but a 10  m-long dugout canoe made from a single oak tree dated to 5450 bc (Fugazzola Delpino and Mineo 1995). We have no evidence for how the J4:1 rock shelter was used. Many caves and rock shelters were used for both habitation and for single primary burials like those known in some Neolithic villages (Robb 1994). Others appear to have been primarily or wholly reserved for ritual activities involving the deposition of ‘valuables’ such as fine pottery and/or the performance of elaborate funerary rites. The Grotta Bella near Terni had human remains and habitation evidence (Guerreschi et al. 1992), whereas the Grotta delle Settecannelle at Ischia di Castro on the edge of the Tuscania Archaeological Survey area appears to have been primarily a funerary cave: there was a juvenile burial within a circle of stones and disarticulated fragments of several other individuals, accompanied by a range of fine pottery including Cardial impressed ware and lithic artefacts such as an ochre-stained grinding stone, associated with a 14C (radiocarbon) date of 7044–6375  cal. bc (Ucelli Gnesutta and Bertagnini 1993). The Grotta Patrizi south of the survey region was similarly used for burial: a trepanned individual was placed in a rock-enclosed niche accompanied by pots, a grinding stone, animal bones and lithics, and there were also the skulls and disarticulated bones of two adults and four juveniles (Radmilli 1951– 1952). It has been argued that many caves represented a

bridge between the lived-in and supernatural worlds, the rituals performed at them – concerned with themes of death, mortality, fertility and the supernatural – serving to maintain the social norms that bounded these dispersed, small-scale and loosely structured societies (Skeates 1991, 1997a; Whitehouse 1990, 1992).

Later Neolithic Settlement, c. 4500–3500 bc Although Fabrics 1 and 2 usually dominate the Later Neolithic assemblages from the survey, there are also occurrences of typically Lagozza finer fabrics, better fired and finished by burnishing to give a darker shinier surface than the other pottery. Several assemblages also have the open straight-sided bowls (often in the finer fabrics) typical of the Lagozza tradition throughout west-central and northern Italy. Also typical of the period are elaborate handles such as ‘trumpet-lugs’, handles on the rim of vessels curving at each end like the bell of a trumpet. One unit with Later Neolithic pottery of this kind was R34:22 (Fig. 4.9), where the lithic material included a small groundstone implement (SF496) probably used for burnishing pottery, and a flint borer (Fig. 4.2:4). Similar pottery was intermixed with the Earlier Neolithic material in RS28:3, a unit that also included the sickle illustrated as Figure 4.4:1, and in the surrounding unit RS28:2. Later Neolithic material was found with Chalcolithic material at several units, including T88:11, RS18:3 and R54:1. The collections from RS33:16 and T88:11 (Fig. 4.10) included fragments of typical Lagozza handles. A similar assemblage that included a classic Lagozza flûtes-de-Pan lug was found during investigations of the Etruscan settlement at Musarna on the eastern edge of our survey area (Recchia and Boccuccia 1998). As with the Earlier Neolithic pottery, the Later Neolithic material fits well within the regional traditions of Etruria, good analogies being found with sites both nearby in the Fiora valley (Negroni Catacchio 1987, 1988a), further north in Tuscany (Consuma: Castelletti et al. 1992; Grotta all’Onda: Amadei and Grifoni Cremonesi 1987; Grotta dell’Orso: Grifoni and Cremonesi 1967; Neto di Bolasse: Sarti 1985), to the east near the Tiber valley (Grotta del

108

TRANSITIONS TO FARMING, c. 5500–3500 bc

0

5 cm

figure 4.9  Later Neolithic material from R34:22: groundstone burnisher (SF496), two struck flints and four sherds. (Photograph: Graeme Barker.)

Vannaro: Cazzella and Moscoloni 1976), and to the south around Tolfa (Pian Cisterna and Pian Conserva: di Gennaro 1998) and Rome (Torre Spaccata: Anzidei 1987; Bietti Sestieri 1984). The Later Neolithic in the Italian peninsula was characterized by interlinked changes in settlement, subsistence, communication systems and ritual life, together marking significant transformations in social organization (Robb 2007). Most settlements are larger than Earlier Neolithic sites and excavations on them have revealed evidence for rectangular houses and grain silos, very different from the pit complexes of most Earlier Neolithic domestic sites. Agriculture was fully established now as the primary subsistence base, with stock managed both for secondary products as well as for meat (Barker 1995a). The production of items such as fine-quality pottery, flint, greenstone and obsidian indicates a new level of specialization in craft production. The discovery of such a­ rtefacts at a distance from their areas of production demonstrates well-developed exchange systems (Ammerman 1985a; Malone 1985; Skeates 1993a). The emphasis of cave ritual also changed increasingly from the community to the individual (Robb 1994; Skeates 1997a; Whitehouse 1992). Taken together, the evidence suggests that Later

Neolithic societies were characterized by increasing social inequalities. The exchange of exotic prestige items was a mechanism by which inter-community alliances were formed, enabling individuals and communities to accumulate ‘debts, wealth, prestige and status in relation to neighbouring groups and individuals’ (Skeates 1997a: 54). Presumably the exotic goods were restricted as valuables not just by their inherent scarcity but also by social and cultural sanctions (Skeates 1995). The major settlements and ceremonial caves were both the focus of such exchanges, within a variety of overlapping social, economic and ritual contexts. The evidence of the Tuscania Archaeological Survey supports this model of increasing complexity. First, except in one case (R44:21), the surface finds areas of the Later Neolithic sites are several times larger than those of the Earlier Neolithic sites, each measuring at least 2,000–3,000  m2. Of course we could not always be certain that the material was in situ – in some cases there were clear indications of sherds eroding downslope – but there were more cases of large sites where the topography indicated that the material could not have moved significantly. Furthermore, at several sites the area of finds was also marked by patches of dark soil presumably pulled

109

4 Prehistoric Landscapes

0

5 cm

figure 4.10  Later Neolithic material from T88:11: seven sherds and (bottom right) a greenstone axe (SF524). (Photograph: Graeme Barker.)

up from buried occupation deposits. It seems likely, therefore, that the Later Neolithic phase of settlement at Tuscania was characterized by significantly more substantial habitation forms than before. Second, the locations of the sites are more varied than those of the Earlier Neolithic. Some are like those of the Earlier Neolithic, midway downslope below the plateau overlooking a stream valley, such as RS28:2/RS28:3 and T89:4. Others, such as T88:11, are right down on the alluvium of the valley floors. More are at higher elevations, either on the plateau edge (R34:22, R44:21, R54:3) or hilltops (RS18:3: Fig. 4.11). The diversity could indicate the development of primary settlements based on mixed farming supported by satellite seasonal camps (for example for shepherding), or of neighbouring communities

pursuing rather different subsistence goals in terms of their mixes of crop cultivation and animal husbandry. Whatever the underlying reasons, the trend would suggest a degree of subsistence intensification or articulation in how the resources of the study area were now being exploited compared with in the Earlier Neolithic. Third, there are indications of exotic materials reaching the Tuscania settlements. At R44:21, for example, there was a sherd of a fine reddish fabric that is probably Diana ware, a fine ware that was widely exchanged in the Later Neolithic throughout the peninsula. The trumpetlug handles probably belong to high-status pottery from beyond Etruria (Malone 1985), as does the vessel with the flûtes-de-Pan lug at Musarna. We also found a dozen or so obsidian bladelets and flakes at a variety of locations,

110

CHALCOLITHIC, c. 3500–2200 bc

figure 4.11  The hilltop location of Later Neolithic site R18:3. The location was also the site of a major classical villa. Looking north. (Photograph: Graeme Barker.)

including four in square J17 (two in unit 11 and two in unit 12). Whilst they could be Later Neolithic, Chalcolithic or even Early Bronze Age in date, some of them could well belong to this phase: there was a series of obsidian flakes at Musarna, for example, though characterization was unable to locate their exact source (Francaviglia 1998). The greenstone hammer-axe (SF524) from T88:11 (Fig. 4.10) is another example of a material that was widely traded in the period. The contextual biographies of many axes suggest that their importance changed with distance from source, first from functional to social, as markers of social identity and power, and ultimately to ceremonial and symbolic when buried with a particular deceased individual (Evett 1973; Leighton and Dixon 1992; Skeates 1995). Fine barbed-and-tanged arrowheads were also manufactured at this time, perhaps primarily as status markers rather than as weapons. The one found in the

survey from RS38:2 (Fig. 4.2:1) could be such an artefact, though these arrowheads continued to be made until the Early Bronze Age.

chalcolithic, c. 3500–2200 bc The earliest copper objects in Italy (as commonly elsewhere in Europe) are found in Early Neolithic contexts, beginning in northern Italy in the fifth millennium bc. Copper tools and ornaments are found throughout Italy by the first half of the fourth millennium bc as metal objects, and ultimately knowledge of metalworking, spread throughout the central Mediterranean in the context of Later Neolithic networks of communication and exchange (Dolfini 2010, 2013a; Skeates 1993b). Though there was no dramatic switch from a stone-using to metalusing technology in Italy, the thousand years or so after c.

111

4 Prehistoric Landscapes

3500 bc (Period 3 in our sequence) is normally termed the Chalcolithic or Copper Age because it was characterized by the first systematic use of copper in some regions, including Etruria. However, the rarity of copper was still such that many Italian prehistorians prefer to term the period Eneolithic to emphasize the importance of its continuity with Later Neolithic culture. Most of the metal has been found in cemeteries. Three principal cemetery groups have been defined: Remedello in the Po valley, Rinaldone in Etruria and Gaudo in the Naples region. In the north the dead were buried in tombe a fossa, simple trench graves, whereas the Rinaldone and Gaudo cemeteries consisted especially of rock-cut chambers, often referred to as tombe a forno (oven-shaped tombs), that were reached by a short entrance tunnel sealed by a stone slab. Caves were also used. The dead were buried usually in a crouched position on their left side, accompanied by fine pottery and artefacts such as superb pressure-flaked daggers and arrowheads and stone axes, as well as thin triangular copper daggers, axes, and smaller implements such as awls. In central Italy the vast majority of the metalwork has been found in Rinaldone cemeteries either immediately around the Colline Metallifere (‘OreBearing Mountains’) in west-central Tuscany or a little further south in the Fiora valley, still within 100  km of the ores. Beyond here, for example on the eastern side of the Apennines, people were mostly buried with just pottery and stone implements (Skeates 1997b). Metallurgical analysis confirms what the distributions suggest, that the Rinaldone smiths were using local ores (Barker and Slater 1971). The sophistication of the mining is evident from Libiola in Liguria, where copper was being mined systematically by the second half of the fourth millennium bc (Maggi and del Lucchese 1988; Maggi and Vignolo 1987). The miners cut shafts into the hillside and extracted the ore using wooden picks, mallets, shovels, wedges and stone hammers. Similar copper mining was practised in Iberia, southern France, the Balkans and the Aegean. The restricted circulation of metal and the distinct regional styles of Copper Age pottery have been taken as evidence for a decline in long-distance exchange networks in the Italian Copper Age (Skeates 1997b), though there is

widespread evidence for maritime contact right across the Mediterranean at this time. In fact, the Italian cemeteries of high-status individuals were part of a wider phenomenon that Broodbank (2013: 305) describes as ‘increasingly “flashy” communities right across Mediterranean Europe, ancestral to the Bronze Age societies of the following millennium’ marked by ever-increasing social inequality. On the evidence of tomb arrangements, grave goods and skeletal variability in the Poggialti Vallelunga cemetery near Pitigliano in the Fiora valley (Negroni Catacchio 1988b), it is likely that Rinaldone society was characterized by hierarchically structured clans (Negroni Catacchio 2006). Spindle whorls and strainer sherds (the latter probably from ceramic sieves used in milk and cheese production) on Rinaldone settlements, as well as changes in herd age structures (Wilkens 1991), suggest that secondary products were increasingly important in animal husbandry, with animals on the hoof increasingly valued as a source of wealth and status in addition to the resources they provided for subsistence. The appearance of new settlements such as Castellaro di Uscio in the Ligurian mountains (Maggi 1990) coincides with pollen evidence there for the beginnings of systematic deforestation and the establishment of pasture and cultivation (Brown et al. 2013; Cruise 1991). Microwear studies of the lithic tools at Castellaro indicate their use for skinning and hide-working as well as butchery, suggesting a degree of specialization in leather-working by this community parallel with the mining activities of the people at Libiola. In combination, the evidence for the clustered burials of high-status (especially adult male) individuals in central Etruria, for settlement expansion and subsistence intensification, and for increased specialization in the production of both subsistence and non-subsistence goods, suggests the emergence of distinct social hierarchies in this part of Italy in which rival lineages and clans competed to control access to resources at the expense of their neighbours. The Tuscania region was on the fringe of the largest and richest Rinaldone cemetery cluster, in the Fiora valley. No Rinaldone graves had been found in the survey area, but isolated examples have been found in and around Tarquinia (Barich et al. 1968; di Gennaro and Pennacchioni

112

CHALCOLITHIC, c. 3500–2200 bc

1988) and south-east at Norchia and Vetralla. Settlement evidence was also sparse. A review of Rinaldone archaeology at the time of the project (Negroni Catacchio 1988b) noted a few caves in northern and central Tuscany with possible habitation evidence but was able to list only three open sites: two in the Fiora valley, at Ischia di Castro and Vulci; and one south of Tuscania in the Mignone valley at Tre Erici, where as part of their investigations of the protohistoric and classical acropolis of Luni, Swedish researchers had excavated some Neolithic/Copper Age domestic deposits (Östenberg 1967). Another possible domestic site had been reported near Tarquinia, at La Tolfa and Pian de’ Santi (di Gennaro and Pennacchioni 1988). The picture has been transformed by the Tuscania Archaeological Survey (Fig. 4.12). We located two definite, thirteen probable and fifteen possible Copper Age sites. It is possible that some of these might be ploughedout burial sites, but the absence of elaborate flintwork and the relative paucity at all the sites of the finer fabrics that dominate the Rinaldone burial repertoire indicate that our material probably derives from domestic contexts. It is possible that funerary material is mixed in with domestic material in unit R34:9, one of the definite sites (Fig. 4.13; Table 4.6), and its adjacent unit R34:11, because finer fabrics made up respectively about a quarter and a third of the pottery collected. The location of this site two km south of Tuscania is also somewhat different from that of the other Copper Age units (see below). Unit R34:9 also produced the largest collection of worked flint (Table 4.6), including a bilaterally retouched blade (Fig. 4.2:2) and two burins (Fig. 4.4:5,12), and four cores for manufacturing bladelets (e.g. Fig. 4.5:4,7). One of the cores is an opposed platform type, while the others are more informal. Six of the pieces are burnt (a significant part of the total number of burnt pieces from the survey). The material suggests that activities at this site included the production of lithic blanks and general toolkit maintenance, though the number of blanks and bladelets recovered is low compared with the number of cores. Although the two definite sites are new locations rather than existing Neolithic sites, more than half of the probable sites are new (Table 4.7), suggesting that Copper Age

settlement in this part of Etruria was characterized by significant expansion as well as continuity. The expansion also resulted in sites now being located in all four quadrants of the survey region, rather than within the restricted zones of Neolithic settlement. Furthermore, some sites were bigger than hitherto: the distinct artefact clusters that could be discerned indicated seven measuring 3,000–10,000 m2 as well as seven measuring 3,000 m2 or less like the Later Neolithic units. There is also an impression of distinct settlement clusters (Fig. 4.12): although RS18:2 was noted as a definite site, the collections from the adjacent units RS18:3 and RS18:16 probably belong to the same habitation zone, and similar associations are likely for R34:9 and R34:11, and for R54:1, R54:3 and R54:6. The pottery from the three RS18 units included sherds from a vessel with an impressed rim (C277), and with plain or impressed cordons just below the rim (C271, C281). These are similar to Copper Age material from caves further north in Etruria such as the Grotta dell’Orso, the Romita di Asciano near Pisa (Peroni 1962– 1963) and the Grotta all’ Onda near Viareggio north of the Arno (Amadei and Grifoni Cremonesi 1987), and from open settlements near Rome such as Casal Bruciato and Torre Spaccata (Bietti Sestieri 1984: 28–33, 51–59). There was also an obsidian core from RS18:2 and another from RS18:15 (Fig. 4.5:1). R34:9, the other definite unit that may have a funerary as well as domestic component, included pinched below-rim cordons (C6, C7) and a cordon cut with wedges (C9: Fig. 4.13, bottom right), the latter again found in the caves of northern Etruria. The R54 material included impressed cordons (C94) and a sherd with zigzag incisions (C92), a motif found in the Chalcolithic collections from the Rome survey. The sherds from CP94:3 included one with a cordon decorated with diagonal slashes (C135) and another with impressed cordons laid out at right angles to one another horizontally and vertically (C140), a motif particularly characteristic of the Grotta dell’Orso assemblage. Apart from the obsidian cores, the only indicator of possible non-local elements in the material recovered from the survey was a sherd from J12:2 with impressed dots (C458), possibly punteggio

113

4 Prehistoric Landscapes

Capec chio

300 m

CP94:3 T4:6

T14:22

Arrone

200 m

0

20

RS1:8 RS1:10

Marta

300 m

m

RS22:16

RS3:9

200 m

RS33:16

o

C100:5

hi

C98:16

CC13:102

T88:9

cc

m

T88:11

pe

0 10

T89:4 T82:3

J8:7

Ca

100

m

C95:108

C99:8

C97:13

R34:9 R34:11 RS12:20

RS39:27

RS8:1

Marta

10 0

m

RS18:16

RS18:3

J12:2

0m

RS18:2

10

R54:1 R54:3 R54:6

Definite Probable Possible 0

5 km

figure 4.12  The Tuscania Archaeological Survey: distribution of Copper Age units. Shading denotes the areas walked.

114

CHALCOLITHIC, c. 3500–2200 bc

0

5 cm

figure 4.13  Copper Age pottery from unit R34:9. (Photographs: Graeme Barker.)

Table 4.6  Lithics (all of flint) from Copper Age unit R34:9. (Compiled by Tim Reynolds.) Blank type

Support form

flake flake flake flake flake fragment fragment blade-flake blade flake blade flake blade fragment flake flake blade

secondary primary secondary tertiary primary secondary secondary tertiary tertiary tertiary secondary tertiary secondary tertiary tertiary tertiary tertiary

Platform type

Edge damage

plain cortical plain plain plain 3 platforms, 2 directions crushed plain

Termination form feathered

hinged

micro-flaking

hinged

burnt burin along the platform flaked lump bladelet core

burin on Siret flake, burnt backed utilized blade

plain crushed

plain crushed

Comment

micro-flaking hinged

115

1 platform burnt

4 Prehistoric Landscapes

Table 4.6  (cont.) Blank type blade blade burnt natural flake flake blade blade flake burnt natural flake flake bladelet flake flake bladelet bladelet bladelet flake bladelet bladelet flake flake bladelet flake flake flake shatter fragment fragment

Support form tertiary tertiary tertiary tertiary tertiary tertiary secondary tertiary tertiary tertiary secondary secondary tertiary tertiary secondary tertiary secondary tertiary tertiary tertiary tertiary tertiary tertiary tertiary secondary tertiary tertiary

Platform type

Edge damage

plain

Termination form snapped snapped

crushed

Comment rectangular blade retouched notch burnt burnt

transverse snap

transverse snap

crushed crushed

transverse snap transverse snap transverse snap transverse snap transverse snap transverse snap transverse snap

crushed plain

transverse snap

plain crushed crushed opposed 2 platforms and direction

hinged snapped snapped snapped snapped snapped feathered snapped feathered feathered feathered

micro-flaking micro-flaking

retouched bladelet

burnt

opposed platform bladelet core bladelet core

Table 4.7  Existing and new units with prehistoric pottery, phase by phase

Definite Probable Possible Total

Neolithic

Chalcolithic

Earlier Bronze Age

Later Bronze Age

Iron Age

existing: new 4 8 10 22

existing: new 0:2 5:8 4:11 9:21

existing: new 0:1 10:15 16:20 26:36

existing: new 5:0 13:2 29:4 47:6

existing: new 2:0 3:1 10:6 15:7

Note: The figures show all units, including linked units that may belong to one phase or the succeeding phase or both, but the general trends in expansion or contraction are probably valid.

decoration, a style of decoration particularly common in the Copper Age ceramic repertoire of the eastern side of the Apennines (Skeates 1997b). Most of the probable Copper Age units are on narrow strips of alluvial soil by watercourses: for example, RS1:10, RS8:1, T14:22, CP94:3, T88:11, C99:8, J8:7 and J12:2. Other

sites are either at the plateau edge (RS22:16, R54:1/3/6) or in dry valleys on the plateau (T4:6). The two definite sites are in quite different locations, however. R34:9 and R34:11, the units with an unusual proportion of finer wares, are on a platform overlooking the Marta below the plateau cliff at a point where the survey team noted possible

116

CHALCOLITHIC, c. 3500–2200 bc

traces of chambers cut into the tufo scarp, suggesting that these units may contain the vestiges of long-wrecked burials at the plateau edge as well as domestic occupation. In contrast, the RS18 group of units (Fig. 4.14) indicates an extensive habitation zone: RS18:2 is immediately below the spur end occupied by RS18:3, and RS18:16 is on the slopes below. Together, these three units indicate a distinct Copper Age habitation area on the northern end of the La Madonella spur. The Copper Age elsewhere in central Italy has provided clear evidence for some settlements being well defended with ditches (Skeates 1997b). There are no indications of this yet in Etruria, though di Gennaro and Pennacchioni (1988) point out that a site such as Tre Erici was in a potentially defensible location at the foot of a natural tufo acropolis. The same could be said of a few of the Tuscania Archaeological Survey units: apart from the spur-end

location of the RS18 units, RS8:1 is at the confluence of two stream valleys overlooked by the steep southern scarp of the Bicocca Verde promontory and T14:22 is at the foot of a spur formed by the Fosso le Tufare and a tributary valley. Far more of the Tuscania Copper Age units, however, are in locations with no defensive qualities. Was warfare a significant component of Copper Age life? Stone ‘statue menhirs’ from Liguria, some of which might have been grave markers, show warlike individuals equipped with the kind of daggers and axes found in the cemeteries (Anati 1981). The copper daggers themselves, however, are of almost pure copper, lacking the hard cutting edge that is achieved with a tin-bronze mix, and being so thin they would probably have buckled if used in c­ ombat – and most of them have survived undamaged (Dolfini 2011). The pressure-flaked daggers and arrowheads were similarly fragile implements and have invariably been

figure 4.14  View over part of the RS18 Copper Age habitation zone. The survey team in the middle distance is recording unit RS18:16. The vehicle on the horizon (to the right) is by RS18:3. Looking west. (Photograph: Graeme Barker.)

117

4 Prehistoric Landscapes

found in mint condition. Hence, it is generally argued that the primary importance of such artefacts must have been for display rather than actual fighting, marking the high status of particular adult males. On the other hand, the discovery that the Copper Age ‘Ice Man’ was equipped with a well-hafted copper axe, presumably to equip him better for his journey into the Alps (unsuccessfully as it turned out, as he had an arrow embedded in his back), suggests that the importance of copper was not entirely social and symbolic (Höpfel et al. 1992). While ritualized combat may have been a feature of competition among Rinaldone elite groups, it is difficult to find evidence in the character and locations of the Tuscania sites for raiding or warfare at the community scale.

boundary with the end of Copper Age material culture is unclear (and there are more similarities than differences between Late Copper Age and Early Bronze Age pottery), and there are debates over the chronology of the succeeding Bronze Age divisions as well. It is especially difficult to place survey material within the sequence given the homogeneity of Bronze Age fabrics and the fact that ‘type fossils’ regarded as indicative of particular phases are frequently missing from small ploughsoil assemblages. In the classification of the Tuscania Bronze Age material, therefore, we have simply divided assemblages with Bronze Age pottery where possible into two broad categories: ‘Earlier’ (Period 4) and ‘Later’ (Period 5), combining respectively the Early and Middle, and Late and Final, Bronze Age phases of the Italian chronological framework. Classic Apennine decorated pottery was generally extremely rare in the Tuscania Archaeological Survey material, though sherds with typical curvilinear and geometric cut-out decoration were found at units such as CP94:7 (C147, C164, C166; Fig. 4.15, top left), RS33:8 (C315, C316, C318), T88:8 (C247) and R14:60 (C47, C48). Slab or ‘axe’ handles are also common in the Apennine Bronze Age, especially with a central perforation as in examples from RS19:16 (C282) and J2:13 (C404), or solid as in the case of the fragments from RS22:16 (C286), RS26:17 (C292) and C99:8 (C207). Typical material of the Later Bronze Age includes a Subapennine knobbed handle (C23: Fig. 4.16, top left) and geometric incised motifs (C48: Fig. 4.16, centre) from R14:13, and (particularly characteristic of Final Bronze Age funerary urns) sherds decorated with combinations of incisions and dimples or pits from RS33:8 (C317, C318). All the Bronze Age units are invariably dominated by Fabric 1 sherds of large storage vessels, sometimes decorated with plain or pinched cordons, though one characteristic of the Later Bronze Age assemblages is the increasing frequency of Fabric 3 sherds from well-burnished carinated cups. Two units with Bronze Age pottery also produced very small lithic assemblages: J3:38 and J18:2 (Table 4.8). The material from J3:38, all of flint, included a single tool, a bilaterally retouched bladelet (Fig. 4.4:10), and

bronze age, c. 2200–950 bc The peninsular Bronze Age is conventionally divided by Italian prehistorians into four phases: Early, Middle, Recent and Final, dated respectively to 2200–1700 bc, 1700–1400 bc, 1400–1200 bc and 1200–900 bc, though the start of the Early Iron Age has now been pushed backwards to at least 950 bc (Barker and Stoddart 1994; Bietti-Sestieri 2010; Guidi 1998). These phases have been defined primarily from changes in pottery styles, in particular body decoration and handle types. The central feature of the sequence is what is known as ‘Apennine’ pottery, dating mainly to the Middle Bronze Age, characterized by drinking vessels in finely burnished fabrics decorated with elaborate handles rising from the rim and with geometric or curvilinear motifs marked out with incised lines and then filled either with more incised lines or dots (punteggio), the lines and dots filled with white paste after firing. The beginnings of handle elaboration and Apennine-style decoration (hence sometimes referred to as ‘Protoapennine’ pottery) have been taken as marking the Early Bronze Age. The later Bronze Age phases are defined by simpler handle forms, frequently animal-shaped, and by a decline in Apennine decoration – hence the term ‘Subapennine’ is sometimes used. However, because of the lack of well-dated long stratigraphies spanning the third and second millennia bc, the

118

BRONZE AGE, c. 2200–950 bc

0

5 cm

figure 4.15  Earlier Bronze Age pottery from unit CP94:7. (Photograph: Graeme Barker.)

0

5 cm

figure 4.16  Later Bronze Age pottery from unit R14:13. (Photograph: Graeme Barker.)

119

4 Prehistoric Landscapes

Table 4.8  Lithic assemblages from Bronze Age units J3:38 and J18:2 (all of flint except where other material is shown). (Compiled by Tim Reynolds.) Blank type J3:38 blade bladelet flake flake flake flake flake flake bladelet bladelet

Support form

Platform type

secondary secondary tertiary secondary tertiary tertiary tertiary tertiary secondary tertiary

crushed

bladelet flake J18:2 blade-flake

tertiary tertiary

crushed crushed

secondary

plain

fragment fragment flake blade bladelet bladelet bladelet flake flake

secondary secondary secondary tertiary tertiary secondary secondary tertiary tertiary

crushed crushed crushed plain plain

plain

crushed crushed crushed

Edge damage

Termination form

Comment

transverse snap

feathered transverse snap transverse snap transverse snap transverse snap transverse snap

crushed micro-flaking transverse snap crushed

there was a single burnt piece, but no cores of flaked lumps; 50 per cent of the pieces have transverse snaps. The adjacent unit J3:39 yielded a sickle blade. J18:2 had no cores or utilized pieces of flint, but there were two flaked lumps, a single burnt flint bladelet, a quartzite scraper on a flake-blade (Fig. 4.3:4), a plunged bladelet of obsidian and a hinged flake of slate. A gunflint of recent (eighteenth-/nineteenth-century) date was also found in this unit. Close parallels with our Bronze Age pottery can be found in the pottery of the four principal excavated domestic sites in adjacent parts of Etruria, all of which have both Earlier and Later Bronze Age occupations: Sorgenti della Nova (Cardosa and Passoni 2006; Negroni Catacchio 1981, 1986, 1989; Negroni Catacchio and Cardosa 2006) and Pitigliano (Aranguren et al. 1985) in the Fiora valley,

hinged snapped

hinged

plunged feathered feathered

burnt Janus flake burnt

bilaterally retouched blade

scraper; material: quartzite flaked lump flaked lump material: slate gunflint material: obsidian burnt

respectively 25 and 30 km to the north-west of Tuscania, Luni sul Mignone 20 km to the south (Östenberg 1967) and Narce 50 km to the south-east (Potter 1976a).

Earlier Bronze Age Settlement, c. 2200–1400 bc The earlier phases of the Bronze Age in Etruria, as elsewhere in the peninsula, were marked by a substantial expansion of settlement (Barker and Stoddart 1994; Guidi 1992). The number of known sites increases significantly in the principal Chalcolithic settlement zones such as the lower and middle sections of the Albegna, Fiora and Tiber valleys (Bietti Sestieri et al. 1991–1992; Negroni Catacchio and Miari 1991–1992), and in inland valleys and basins hitherto settled sparsely (Balista et al. 1991–1992; Malone and Stoddart 1994). A few sites

120

BRONZE AGE, c. 2200–950 bc

are found at the highest elevations in the Apennines (Barker 1995a: 139; Barker and Grant 1991). Another indicator of rising populations is pollen evidence for increases in the amount of cleared land on the lowlands in the primary settlement areas (Alessio et al. 1986; Bietti Sestieri et al. 1991–1992; Hunt and Eisner 1991; Kelly and Huntley 1991), and for new clearances in the mountains (Brown et al. 2013; Cruise 1991), though these still are very small-scale relative to the clearance activities registered in the pollen diagrams of Etruria after 1000 bc (see Chapter 3). The Tuscania Archaeological Survey located 1 definite, 25 probable and 36 possible units with Earlier Bronze Age material, compared with the (respectively) 2, 13 and 15 units with Chalcolithic material (Fig. 4.17). These units are located throughout the survey area and at every distance from Tuscania. Although Apennine Bronze Age pottery was placed with burials in Etruria, or lodged in ritual deposits, both of these activities were largely restricted to caves (Whitehouse 1995). One such cave near Tuscania was the Grotta di Carli at Ischia di Castro (Casi and Mieli 1998). In this light it is reasonable to assume that most if not all of our Earlier Bronze Age units primarily represent locations of domestic activity. If this is indeed the case, they represent a doubling of the settlement evidence in terms of both the total number of units (30 Copper Age, 62 Earlier Bronze Age) and the more reliable definite and probable categories (15 Copper Age, 26 Earlier Bronze Age). There were many new foundations (Table 4.7). This table includes all units, including those that may belong to one phase or the next, or both, but if we exclude these, the number of definite or probable units with Earlier Bronze Age pottery and definitely lacking Copper Age pottery is still six compared with only one existing foundation with pottery of both periods. It seems clear that the first half of the second millennium bc witnessed a substantial expansion in the utilization of the survey area, in line with the general trend of settlement increase in the region (Palmisano et al. 2017). The density indicated by the Tuscania Archaeological Survey is, in fact, unequalled in South Etruria, though this presumably

reflects the intensity of our survey rather than any inherent preference for settlement in this part of the region. Relatively high densities were also found in the Fiora valley (Negroni Catacchio and Miari 1991–1992), which has been subject to less intensive but more protracted fieldwork over many years. Apart from the now disputed evidence that substantial long-houses partly cut into the rock of the Luni sul Mignone natural acropolis are Bronze Age (Östenberg 1967), excavations of sites such as Narce (Potter 1976a) and Torre Spaccata (Anzidei et al. 1985) indicate that most habitation sites in Etruria in the first half of the second millennium bc consisted of conical wattle-and-daub huts. Until recent decades, huts of this kind were a frequent sight on the lowlands of central Italy, used seasonally by transhumant shepherds from the mountains (Barker 1981: fig. 50; Barker and Rasmussen 1998: 35; Close-Brooks and Gibson 1966), and similar huts were built by fishermen by lakes such as Bracciano and Bolsena (Fugazzola Delpino 1983–1984: fig. 5). In the Biferno valley, the Bronze Age sites found by the archaeological survey were of two main size categories: most of the ploughsoil pottery scatters were smaller than 50 m x 50 m in extent, while a few sites were 50–100  m in diameter (Barker 1995a). Excavations at two of the latter indicated that they had consisted of about half a dozen huts, suggesting that the smaller ploughsoil scatters were likely to be the vestiges of one or two huts only. Most of the Tuscania Archaeological Survey sites of this phase are also very small, c. 30 m x 30 m or less, but there are also groups of adjacent units which suggest that there may also have been larger settlement clusters like the larger Biferno valley sites (Fig. 4.17). Small ‘single household’ units include RS19:16 on the slopes above the Marta river, RS38:5 on a plateau-edge location above the Fosso della Cadutella near Tessennano, and J2:13 and J3:38 in somewhat similar locations above the Marta near Tuscania. There are indications of larger ‘hut clusters’ at six locations, one of the most striking being RS1:10 on the northeastern edge of the survey area (Fig. 4.18), where there was a series of distinct areas of dark organic soil, though unfortunately they and the finds with them were

121

RS38:5

RS19:15

CP94:22 CP94:5 CP94:6 CP94:3

CP94:7 RS19:16

T14:22 RS1:1 RS1:8

T34:12

m

RS28:2

RS22:10

200 m

J20:20 RS33:1

C95:108

C97:13

T82:3

C98:16 C99:8 C95:110

o

hi

cc

RS26:17

J18:2 J18:1 J18:5

RS26:20

J18:6

RS12:26

CC13:102

T89:4 T89:3 T88:9

J2:13

J8:7

J2:15

R34:9

T88:8 T89:1

T85:29

RS33:5/7 RS33:8 RS33:11

R14:31 R14:60/70 RS33:10 RS33:12 R24:22

m

T88:1

T85:29

pe

100

1

T84:3 R14:13

Ca

m

C97:17

00

RS1:10

200 m

0

Marta

Arrone

300 m

20

300 m

Capec chio

4 Prehistoric Landscapes

J3:1 J3:38

R34:25 R34:11

RS12:20

0m

R54:3 RS4:3

10

RS4:1 R54:1

R54:11 RS4:11

J12:2

RS8:1 R74:3

Marta

10 0

RS18:3

m

RS18:16

RS18:2

Earlier Bronze Age

Later Bronze Age

Definite 0

5 km

Probable Possible

figure 4.17  The Tuscania Archaeological Survey: distribution of Bronze Age units. Shading denotes the areas walked.

122

BRONZE AGE, c. 2200–950 bc

figure 4.18  RS1:10, the location of a likely significant Bronze Age habitation area consisting of several huts marked at the time of survey (though very badly ploughed out) as areas of dark organic soil on the slopes and valley floor where members of the survey team are visible. Looking south-west. (Photograph: Graeme Barker.)

so badly damaged by ploughing that the location could only be designated as a possible settlement zone despite there being both Earlier and Later Bronze Age material here. The others, categorized as definite sites, consisted of units RS26:17 and 20 on the plateau and adjacent slopes at the confluence of the Arrone with one of its tributaries; units 8 and 10 of RS33 on the edge of the Marta alluvium below Tuscania; units 3, 5, 6, 7 and 22 in CP94, a group of discrete pot scatters above the Fosso le Tufare at the Fonte di Paolo spring; units 8, 9 and 11 by the Fosso Puntacciarlo in T88; and units 2, 5 and 6 on either side of the Fosso Capecchio in J18. The evidence of site distribution and morphology appears to support the picture emerging elsewhere at this time of dispersed populations in which the individual household was once more the primary unit of production

and consumption, though with groups of households coalescing from time to time. The arrangement of hut groups and burials at Torre Spaccata suggests a series of family households, without marked ranking, in contrast with the structuring of Chalcolithic societies (Anzidei et al. 1985). The best insights into lifeways at these small domestic sites have been revealed at Croce del Papa near Pompeii in the shadow of Vesuvius: preserved under the pumice, ash and pyroclastic flows of an eruption in the eighteenth century bc was a series of small thatched huts within stockades, animal pens, a threshing floor, grain silos, wells and the footprints of cattle preserved in cart tracks leading off to surrounding fields (Albore Livadie 2002, 2008). A remarkable Early Bronze Age landscape of ploughed fields has been similarly preserved by the Pomici d’Avellino eruption at Gricignano d’Aversa

123

4 Prehistoric Landscapes

(Saccoccio 2020; Saccoccio et al. 2013). The botanical and faunal data from a number of sites indicate similar systems of mixed farming, with people growing a range of cereals and legumes and keeping a variety of stock, but especially relying on sheep and goat husbandry: meat, milk and wool were all important (De Grossi Mazzorin 1985, 1995; Wilkens 1991, 1991–1992). On the evidence of thin-section and X-ray analyses, most Apennine Bronze Age pottery was manufactured at the household level (Barker 1995a: 154; Loney 1995), but close similarities between the decorative styles of different parts of the peninsula demonstrate that there must have been effective systems of communication, with the Apennine mountains themselves far less a barrier than they had been hitherto. Interestingly, isotope studies of bones from the cemetery of Sant’Abbondio near Pompeii indicated that men were moving and marrying, whereas most women stayed local (Tafuri 2005). One of the features of prehistoric societies characterized by competitive behaviour among powerful individuals, but not yet by the self-perpetuating elites of more developed systems of inequality, is a marked tendency to instability and oscillations in power structures. It seems likely that the earlier centuries of the second millennium bc in Etruria, though a period of expansion in demographic terms, were marked by some degree of disintegration of the kind of established social hierarchies and systems of resource control that we can discern in Etruria in the Copper Age (Barker and Stoddart 1994; Guidi and Stoddart 1998). Early and Middle Bronze Age metalwork, while comparatively rare, is rather evenly distributed across the peninsula, in contrast with the concentration of Rinaldone metalwork in Etruria. It is difficult to imagine that households were able to operate entirely self-sufficiently, though: keeping a team of plough cattle, for example, must have been beyond the means of most individual households in Etruria given the difficulties of maintaining such animals through the summer drought before the days of modern fodder crops. Presumably mechanisms of collaboration were developed to share risk, mechanisms of the kind that are thought to have

characterized life in the Greek Aegean region at this time (Halstead 1987, 1992).

Later Bronze Age Settlement, c. 1400–950 bc In the second half of the second millennium bc, there is consistent evidence for significant developments in social complexity in Etruria, more marked than anywhere else in central Italy. They denote the beginnings of the process of transformation in social complexity that was to crystallize a few centuries later in the Etruscan city states (Bietti Sestieri 2010; di Gennaro 1999; Guidi 1998; Minniti 2012; Pacciarelli 2000; and see Chapter 5). There was an increasing trend towards the abandonment of open sites in favour of naturally defended tufo promontories and hilltops (Barbaro 2010; di Gennaro 1982, 1986, 1988; Guidi 1998; Miari 1987). The best excavated site of this kind is the 15-ha tufo outcrop of Sorgenti della Nova in the Fiora valley, where larger dwellings on the summit surrounded by smaller huts on lower terraces have been taken as evidence that Later Bronze Age societies were characterized by marked social stratification (Cardosa and Passoni 2006; Negroni Catacchio 1981, 1986, 1989; Negroni Catacchio and Cardosa 2006; Negroni Catacchio and Domanico 1986). Space was carefully divided between areas of public life, sleeping quarters and storage. Gender differences can be discerned at the site, but entwined with age, status and kin relationships rather than performed in isolation (Dolfini 2013b). The period was also marked, as in many parts of Europe, by the introduction of cremation for burial, the treatment of the dead providing further evidence for social hierarchies based on formalized inequality. People were burned on funeral pyres and their ashes placed in the ground in large funerary urns. The ashes of some individuals were buried with nothing, others with perhaps a few pots and a couple of metal objects such as a razor or a pin, and a few with distinctly richer collections of pottery and metalwork. New settlement clusters around the Colline Metallifere and the Tolfa mountains suggest an intensification in metal extraction and processing (Giardino 1984). Certainly, metalwork was much more plentiful

124

BRONZE AGE, c. 2200–950 bc

than hitherto, with effective tools and weapons being produced in high-quality tin-bronze as well as items for dress and display such as cloak pins. Presumably c­ ompetition for controlling access to the ores and the production and circulation of metal was an increasingly important feature of elite power in this critical phase of social intensification. However, the major centres of wealth remained the traditional settlement foci in South Etruria, such as the Fiora valley, rather than the immediate localities of the ores. By the end of the second millennium bc, Vulci, Tarquinia and Cerveteri were all significant ‘central places’ with settlement at each extending well over 100 ha (Guidi 1992, 1998). They were occupied by communities among whom were dominant social groups whose warrior attributes were heavily emphasized in their burials (Barbaro 2010; Blake 2014; di Gennaro 1999). The Later Bronze Age phases of settlement in the Tuscania area are represented by 5 definite, 15 probable and 33 possible units, slightly fewer than in the Earlier Bronze Age (Fig. 4.17). Most were first established in the preceding Bronze Age phases, suggesting that around Tuscania continuity was the principal characteristic of Bronze Age settlement through the second millennium bc. However, there was one particularly important development: the beginning of the first significant occupation of Tuscania’s Colle San Pietro acropolis. Later Bronze Age pottery has been found in unstratified deposits near the Early Medieval church (Fugazzola Delpino and Delpino 1979; di Gennaro 1986, 1999a). We recovered further material from the units on the southern edges of the acropolis mound (units 13, 15, 31 and 60 of R14), and beside the alluvium of the Marta valley immediately below (units 8, 10, 11 and 12 of RS33). The latter included sherds with the line and dimple decoration characteristic of the urns used in cremation burials. Although the same pottery occurs within the assemblages of settlements such as Sorgenti della Nova, it is likely that the Tuscania settlement had its cemetery nearby in the Marta valley. By the late second millennium bc, the Colle San Pietro acropolis was probably the focus of a typical hilltop community of the kind revealed by the excavations at Sorgenti della Nova or the intensive survey of the ‘Il Pizzo’ promontory of Nepi (di Gennaro et al. 2002; Rajala 2007, 2013).

Few hilltop sites in southern Etruria have been investigated on any scale like Sorgenti della Nova, and none has produced convincing evidence for defence works. Nevertheless, their frequency indicates that the final phase of the Bronze Age witnessed a dramatic process of settlement nucleation, with people being gathered together in local centres for security, under the control of local elites (di Gennaro 1982, 1986, 1988, 1998, 1999b; di Gennaro and Passoni 1998). On the basis of the size of the hilltops, which average 4–5 ha, di Gennaro has estimated their populations as between 100 and 1,000  ­people,  and  the density of occupation evidence suggested that most of the hilltop communities were nearer the upper rather than the lower figure. However, we should probably not exaggerate the extent of nucleation: much of the argument rests on negative evidence, on the lack of systematic survey around the hilltop sites. In this context it is striking that there were numerous occupation sites in the countryside surrounding Tuscania, the focus of probably the most intensive survey of any around the hilltop centres of Etruria. In fact, the evidence for the local settlement cluster around Tuscania fits well with the other evidence mentioned above for internal differentiation at Sorgenti della Nova, and for marked disparities in the cemetery grave goods. It is tempting to conclude that by the later second millennium bc the Tuscania acropolis was home to hierarchically structured clans and also acted as the focus for surrounding communities bound to the hilltop elite by other forms of clientship. We cannot be sure of the size of territory for which Tuscania was such a focus, but neighbouring hilltop settlements are generally 12–15 km away, just beyond the edge of our survey area, such as Rogge di Canino to the west, Tarquinia to the south-west, Norchia to the south and Musarna to the east, suggesting that Tuscania’s territory was probably little more than the 4–5 km that are visible (in some directions at least) from Colle San Pietro. The Later Bronze Age sites in the more distant parts of the study area, such as those in CP94 to the north, were probably beyond the remit of the Tuscania dominant group. It is unclear whether local elite groups were in some kind of client relationship at this time to regionally dominant

125

4 Prehistoric Landscapes

groups of the kind presumed for Tarquinia, Vulci and Cerveteri, as was subsequently the case in Early Iron Age and Etruscan times, but the balance of probability is that they were. The emergence in South Etruria of these nested local and regional settlement hierarchies, with their implications of markedly stratified societies, seems to have been accompanied by developments in agricultural support systems. New crops indicate intensification in systems of crop husbandry (Follieri 1981), for example, as does faunal evidence for the increasing importance of animal secondary products, with sheep the most frequent stock at most sites (De Grossi Mazzorin 1995, 1998; De Grossi Mazzorin and di Gennaro 1992). The occurrence of bones of older cattle and of donkeys at Sorgenti della Nova suggests that one feature of elite power may have been control of the means of agricultural production (plough cattle) and transport (pack animals), much as has been argued for the MinoanMycenaean palace elites of the Aegean at about the same time (Broodbank 2013). What seem to be ritual deposits of pig bones at Sorgenti della Nova (De Grossi Mazzorin 1998) could be a further indicator of how local elites may have used consumption rituals as a means of legitimating their authority and underpinning clientship relations. The closing centuries of the second millennium bc have been characterized as the beginning of the emergence of significant ‘connectivities’ or long-distance trading relations linking Mediterranean coastal communities all the way from Sardinia to Cyprus (Broodbank 2013: 449–505; Dawson and Nikolapoulou 2020). Some Villanovan metalwork in graves has been interpreted as evidence for marriage exchanges between Italian and Sardinian families, and these more local social networks are likely to have been much more common than economically purposeful long-distance trade the length and breadth of the Mediterranean (Knapp et al. 2021).

Etruria were for long regarded as evidence for new people spreading southwards down the peninsula, ultimately from central Europe, bringing the rite of cremation with them (Hencken 1968). However, most scholars now accept that the archaeological record for this period in Etruria demonstrates that we are primarily witnessing dramatic internal ­transformations in social complexity, a culmination of the processes of differentiation and stratification already apparent in the region during the later second millennium that presaged Etruscan urbanism (di Gennaro 1982; Bietti-Sestieri 2010; Perego and Scopacasa 2016; Spivey and Stoddart 1990). Powerful elites capitalized on Etruria’s metallurgical and agricultural wealth and developed extensive trans-Apennine and Tyrrhenian maritime trading networks (Foxhall 1998). There is evidence of significant intensification in the scale and nature of arable farming and pastoralism as well as viticulture, accompanied by an upscaling in the palaeoecological evidence for human impact on vegetation, including around Tuscania (Chapter 3). Iron Age settlement, though not its initial phase (950– 800 bc), is attested on Tuscania’s Colle San Pietro acropolis (di Gennaro 2018), but the survey of the countryside produced only two definite, four probable and sixteen possible units with Period 6 or Iron Age pottery (Fig. 4.19). Both the total number of units and the number of definite/ probable units (six) represent only a third of their Bronze Age equivalents. Only one unit, CP94:4, could be taken as indicating a ‘new foundation’, but its location within the cluster of units with Later Bronze Age material in CP94 makes this unlikely. One of the ­difficulties in recognizing Iron Age units at the time of the survey was that the ceramic typologies were based substantially on cemetery assemblages. It may be significant that the two ‘definite’ Iron Age units, RS33:12 and R14:13, are both extremely near Tuscania, and it is possible, indeed likely, that their assemblages contain funerary material, such as a white slipped sherd (C185) in RS33:12. In fact, one distinctive feature of both assemblages is the variety of fabrics: the coarse domestic wares are always predominant, but there are isolated sherds of the finer fabrics, again perhaps indicative of mixed domestic and funerary material. Apart from these units, the study area is almost devoid of material that

Early iron age, 950–700 bc The ‘Villanovan’ (Early) Iron Age of Etruria is so called after a cremation cemetery excavated in the nineteenth century outside Bologna. Similar cemeteries found in

126

EARLY IRON AGE, 950–700 bc

RS19:5

CP94:4

Capac chio

300 m

RS35:26

RS1:1

J1:11

J34:12

200 m

0

20

Marta

Arrone

300 m

m

RS28:2 RS28:3

200 m

J20:20

C97:17

R14:31

hi cc

RS26:20 RS26:19

RS33:12

o

m

pa

0 10

R14:13

Ca

100

m

R14:15

J2:13

J18:2 J18:5

RS26:2

J18:6

RS37:1

J3:1

0m

10

Definite

Marta

10 0

m

RS18:11

RS24:55

Probable Possible 0

5 km

figure 4.19  The Tuscania Archaeological Survey: distribution of Iron Age units. Shading denotes the areas walked.

127

4 Prehistoric Landscapes

can be assigned to the period with any degree of confidence, the exceptions being a twisted cordon sherd (C290) at RS26:20, a typical Villanovan bossed sherd (C116) at CP94:4 and a series of possible Iron Age fabrics and forms from the J18 units. While the uncertainties about the nature of domestic assemblages have to be acknowledged, it is likely that the Tuscania acropolis and its immediate environs had become, probably from the intervention of the proto-urban centres (di Gennaro 2018), the primary focus of settlement in this part of Etruria by the eighth century bc following a process of abandonment (voluntary or otherwise) of small domestic sites in the countryside. The concentration of settlement on Tuscania’s hilltop was part of the larger process of settlement transformation in South Etruria during the Early Iron Age (Guidi 1998; Palmisano et al. 2017). Most of the tenth-century bc hilltop sites measure 1–5  ha with a few measuring 5–15 ha, but in the ninth century bc a series of huge settlements developed that entirely dominated this system, each measuring between 100 and 200  ha (di  Gennaro 1988; Guidi 1998). Five of these centres were to become the five major Etruscan cities of the region: Caere (modern Cerveteri), Tarquinia, Veii, Volsinii (modern Orvieto) and Vulci; to which a sixth can now be added, Visentium (Bisenzio) on the edge of Lake Bolsena (Ceci and Cifarelli 1995). These Villanovan settlements were substantial population centres on an entirely new scale, even though their huge hilltops were probably not occupied in their entirety. There were distinct habitation zones characterized by larger and smaller structures, each habitation zone within its enclave of arable land (Bartoloni 1989a). Excavations at Cerveteri and Tarquinia have demonstrated the existence of possible cult sites among the huts, a further suggestion of family or clan organization. In the large cemeteries that surround the major centres, the people buried seem to have been differentiated in terms of allegiances to particular kin groups or clans that were also stratified according to wealth and status (Fugazzola Delpino 1984; Hladikova 2013; Teegen 1995). The evidence points to an Early Iron Age social structure of territorial tribes reminiscent of the curiae of Roman society.

The diverse origins of the metalwork buried in the richer Early Iron Age graves, especially in the female graves, suggest a system of ‘ritualized friendship’, in which leading families cemented their alliances through marriage partnerships and prestige gift-giving, though violence is also attested (Bartoloni 1989b; Cuozzo 2014; Giardino et al. 1991). Another mechanism was feasting, almost certainly including wine drinking: Early Iron Age viticulture is attested, for example, at the settlement of Gran Carro on the eastern shore of Lake Bolsena (Costantini and Costantini Biasini 1987). Although vines were indigenous to Italy and viticulture probably developed earlier in the Bronze Age, it is notable that the first clear evidence for viticulture coincides with the beginnings of contact between the Villanovan elites and the Greek colonists of southern Italy in the eighth century. Wine drinking was a critical component of the Greek symposion, the ritualized banquet designed to display the power and influence of the host. The earliest Greek imports in Villanovan cemeteries are invariably objects associated with wining and dining, and were also the first objects also to be imitated in the craft workshops that served the Villanovan elites (Ridgway 1992). The adoption of the Greek heroic ideology is another manifestation of how Villanovan elites in Etruria adopted aspects of Greek culture to emphasize their political control and their distance from the rest of the population (Guidi 1998: 147). By the eighth century bc, therefore, Tuscania was part of a network of nucleated settlements embedded in client relationships with major proto-urban centres that were to emerge a century later as Etruscan cities – in Tuscania’s case, almost certainly with Tarquinia (di Gennaro 2018; di Gennaro and Rendeli 2019). ‘Local leaders would have been linked to their clients in a set of mutual obligations, such as receiving gifts, tribute, labour and so on, in return for providing protection, communal services, and foodstuffs. In turn, they would have been beholden to more powerful leaders in parallel systems of clientship’ (Barker and Rasmussen 1998: 80). The regional power bases of the paramount chiefs of the southern centres are thought to have been such that they were able to compel Greeks seeking to trade for metal

128

EARLY IRON AGE, 950–700 bc

with Populonia and Vetulonia, the major Early Iron Age centres of metalworking by the Colline Metallifere ores, to seek safe conduct from them to sail along their coasts or travel by land or river through their territories (Cristofani 1983; Torelli 1982). At the local level, the scatter of possible Early Iron Age material throughout the Tuscania Archaeological Survey area but lack of evidence for distinct farmsteads suggest

that this particular Early Iron Age settlement likely functioned as an ‘agro-village’ in which people walked out to their fields each day, returning in the evening to their homes on the southern flanks of the Colle San Pietro acropolis. The hilltop had become the physical expression of the networks of protection and obligation that underpinned Early Iron Age societies, and the Etruscan societies they became.

129

5 ETRUSCAN URBANIZ ATION, c. 700–300 bc Tom Rasmussen, Marco Rendeli and Graeme Barker

The status of pre-Roman Tuscania was no doubt fluid over the centuries of its existence. In fact the term ‘town’ can really be used to describe it only from the sixth century bc onwards, as it developed from its earlier preurban or ‘proto-urban’ phases of the Early Iron Age (c. 900–750 bc) and the Etruscan Orientalizing period (750– 580 bc, so termed because of the appearance in Etruria of east Mediterranean artefacts at this time). To accommodate such developments that many nucleated centres of population must have undergone through these periods, the term ‘higher-order settlement’ has been proposed to be applied to any settlement more than 2 ha in area ‘that exercised or might have exercised control over a territorial unit’ (Sewell and Witcher 2015: 4.2). ‘Control’, however, clearly can be fluid too, as well as being more – or less – consensual. There is control in the form of landowning and tenure, and there is political control; although they overlap, they are not the same. Through survey it may be possible to map the extent of the agricultural land that Etruscan Tuscania had control over, but after the stateformation processes were complete the ultimate power, and the political control, must have lain with Tarquinia – perhaps in the form of land taxes of one kind or another, of requirements to raise troops for the controlling city, or supplying goods or labour in times of need, and so forth. Over twenty years ago it was remarked that: ‘The archaeological history of the Etruscans, as currently written, is the history of individual city states interacting with one another and the city of Rome. The cities also interact with the wider Mediterranean world, particularly the Greeks. The one thing that the cities do not seem to interact with is the landscape of Etruria which they occupy’ (Perkins 1999: 2). Perkins set out to thoroughly address this issue for the Albegna valley, and already the previous year had seen the publication of a wider and

It was a pleasure to us to look over so much of old Etruria, and a wonder to think how populous it must have been … (Elizabeth Caroline Hamilton Gray 1843: 338)

introduction The organization of Etruria into separate states, it is generally agreed, was a process of ‘state formation’ that had been set in train by the seventh century bc and was completed in the sixth. Statehood implies a territory with a controlling centre, which need not be but usually is urbanized. State formation and urbanization are linked, therefore, but are not interchangeable (Stoddart 1999). The largest and most powerful of the urbanized centres were the cities with their densely nucleated populations. A population figure of around 5,000 is often mentioned in the archaeological literature as being the minimum requirement for city status (Osborne 2005: 1), but in Etruria the cities that controlled individual states were often more populous. Tuscania, a small town, was clearly not one of these but must have been under the political control of such a city, and all the evidence goes to show that the city in this case was Tarquinia (Fig. 5.1). Literary evidence, admittedly of Roman Imperial date (see Chapter 1), suggests that the territory of Tarquinia extended as far as Lake Bolsena (Spivey and Stoddart 1990: fig. 7c). Archaeologists plotting Etruscan territories with methods such as Thiessen polygons, both ‘weighted’ to take account of natural features and boundaries (Spivey and Stoddart 1990: fig. 7b) or further modified by means of the XTENT formula (Redhouse and Stoddart 2011: fig. 8.2), all conclude that Tuscania was in Tarquinian territory, which encompassed most or all of the river valley of the Marta and its catchment area.

130

introduction

M NE NNI APE

o

Arn

Chianti Hills

Tiber

Chiana

S TAIN OUN

Murlo/Poggio Civitate

Colline Metallifere

Lake Trasimene

Col di Marzo

e

Mount Amiata

br

on

Sovana

Om

Podere Tartuchino

Paglia

Orvieto (Volsinii)

gn

a

Volsini Mountains

e on

RE

Montebello

Viterbo Norchia xxorchia

Tarquinia Gravisca (Graviscae)

TYRRHENIAN SEA

e on

n

ig

M

ar

ta

Giglio Island

Vulci

San Giuliano San Giovenale Pian della Conserva

Tolfa Poggio Mountains San Pietro

La Castellina

Pyrgi

50

0 km

SABINA

Cimini Mountains Lake Vico Nepi (Nepet)

Cerveteri (Caere) R

O

Lake Bracciano

Veii R ACN OM ACM A APMA N PGAN GA N A

M

Capena

Tiber

A

Arr

MM

M

Giglio wreck

Mount Argentario

Montefiascone Acquarossa Musarna Tuscania Castel d’Asso

Fiora

be

Marsiliana d’Albegna

MA

Lake Bolsena

Bisenzio (Visentium)

Al

Doganella

Rome

Land over 200 m

figure 5.1  Etruria, showing the principal sites and other locations outside the Tuscania Archaeological Survey area mentioned in Chapter 5 (ancient names in brackets).

131

5 ETRUSCAN URBANIZ ATION

more generalized study (Barker and Rasmussen 1998). This attempted to set the landscape firmly in the foreground of archaeological approaches to the Etruscans, making use of the preliminary results of the Tuscania Archaeological Survey and the early data from the British School at Rome’s South Etruria Survey (Potter 1979). Use of an incomplete set of Tuscania Archaeological Survey data had also been made by MR in an earlier synthesizing work on the Etruscan settlement patterns of southern Etruria (Rendeli 1993a). Yet, as we describe in the next section, despite the ensuing decades of Etruscan archaeology, the Tuscania Archaeological Survey data remain a vital resource to fill what is still a significant gap in our knowledge of Etruscan land use. The field-walking was more intensive and systematic even than at Veii, and the investigation of an agricultural zone around a small urban centre in the Etruscan heartland offered the chance to ask questions about the relationship between the centre and outlying regions, about the intensification and waning of agricultural activity over time, and about relations with areas beyond the survey area, in Tuscania’s case with its more powerful neighbour Tarquinia.

iron were placed within many of the tombs of the major centres such as Vulci and Caere (modern Cerveteri). The tombs, too, contained increasing amounts of imported Greek pottery, while native Etruscan pottery, always a major component of funerary assemblages, underwent profound developments. The material culture of the smaller settlement sites, which concern the present survey, reveals transformations that are no less dramatic. In the first place, tile is now present in the archaeological record for the first time, as the population – whether through native invention or Greek influence – began to build their homes and other buildings with tiled roofs. An immediate upshot is that building remains become much more substantial than hitherto: one square metre of roof tile averages 60 kg in weight and necessitates stronger materials for walls which may be of stone or at least have stone footings, compared with the wattle and daub or pisé walls of earlier houses (Rathje 2004: 64). The use of tiles also lessened fire risk and enabled buildings to be sited more densely (Damgaard Andersen 1997: 374–375), not only in the burgeoning urban centres but also in villages and hamlets, and it coincided with a great increase in numbers of rural sites in many parts of the countryside (see below). It is, of course, possible to argue that field-walkers are much more likely to spot sites that are littered with tile fragments than sites with pottery alone and that the increase in numbers of sites may therefore be exaggerated, but all the evidence goes to show – from tombs and cemeteries as well – that a remarkable expansion in land use occurred in the seventh century. New crops were also introduced, such as the vine and olive, or if introduced earlier were certainly first cultivated on a large scale at this time (Barker 1988: 781; Ciacci et al. 2012). The sunken pithos (storage jar) and associated grape pips at the Podere Tartuchino farm in the Albegna valley (Perkins and Attolini 1992: 87, 122) strongly suggest wine production there; while the Auditorium site in the north of Rome with its olive-press of the early fifth century (Terrenato 2001b: 8–9) and the cargo from the Giglio wreck recovered off the Maremma coast with its organic remains including olive stones (Bound 1985: 67,

etruscan urbanization and its landscape impacts It is a matter of convention that for the period ending c. 750/700 bc archaeologists speak in terms of the Early Iron Age in central Italy, and Etruscan Orientalizing for the following period (Barker and Rasmussen 1998: 5). The problem of nomenclature, inevitably compounded when ‘Villanovan’ is also used synonymously with ‘Early Iron Age’, is well known. It is as misleading to think in terms of a new population of Etruscans appearing now for the first time as it is to believe in an influx of ‘Villanovans’ at the end of the Bronze Age. Nevertheless, in terms of material culture the changes visible from the early seventh century bc onwards were very significant, especially where burial practice was concerned. In South Etruria inhumation replaced cremation (though never exclusively), tumuli and chamber tombs replaced the old trench (fossa) graves and items of exotic materials of ivory, gold, silver and

132

etruscan urbanization and its landscape impacts

fig. 6; Peña 2011: 185) point equally strongly to the manufacture of olive oil. To accommodate this new agricultural activity, one must envisage scrub and forest clearance on a massive and unprecedented scale, as has been shown further south around San Giovenale (Hemphill 1988) and is indicated by increased erosion rates in the Marta valley (Chapter 3). Although there must have been earlier land clearance contemporary with Late Bronze and Early Iron Age finds, it would have been on a small scale compared with what was needed for the population increase that occurred throughout southern and central Etruria from the eighth to the sixth centuries bc (Cifani 2002a; and see Chapter 9). Intensive fieldwork at Tarquinia has shown that there was already a highly organized settlement on the habitation plateau, the Pian di Civita, by the eighth century and a dense population with a concomitant building programme in the seventh (Bonghi Jovino 2005). Data from a non-systematic survey at Vulci in the 1970s by the GAR (Gruppo Archeologico Romano) suggest that there was a ‘proto-urban’ phase in the eighth century bc prior to full urbanization in the seventh and later (Pacciarelli 1989–1990). For Caere, Zifferero (2005: 264), utilizing data from Enei (2001), has argued that by the mid sixth century there was a fully urbanized centre with control over a wide surrounding area; the town wall dates to the early fifth century at the latest (Bellelli 2016: 50–53). At the site of Veii, it has been argued both that the proto-urban settlement of the eighth century consisted of independent villages (Rendeli and Sansoni 2004: 15–16) and that the Early Iron Age population was already fairly integrated (Biagi 2019: 51; Schiappelli 2004: 159–161). In the second half of the seventh century, there was an urban grid layout on the hill of Piazza d’Armi (Cascino et al. 2012: 341), and from the early sixth century Veii was clearly a wellstructured urban community with firm control over a large territory populated with medium- and small-scale settlements (di Sarcina 2012a: 349). Air photography and large-scale geophysical survey have revealed an extensive and complex road system connecting the different zones of the Etruscan city, and the city with its hinterland (Campana 2018: 110).

We can be fairly confident about the settlement landscape of Veii, as this area must be one of the most intensively surveyed of the Mediterranean, since field-walking here goes back to the core of the South Etruria Survey of the 1950s and 1960s (Potter 1979; Potter and Stoddart 2001: 10–16). It was then integrated into the wider Tiber Valley Project begun in 1997 (Patterson and Millett 1998), not all areas of which received the same intensity of landscape survey (Patterson and Coarelli 2008; for the geographical area of the project: Patterson and Witcher 2002: 87, fig. 1; for the ethnic divisions of this territory into Etruscans, Umbrians, Faliscans, etc.: Cifani 2002a: 221, fig. 1). The same is true of the territories of the Etruscan coastal cities, where exhaustive field-by-field survey has been the exception rather than the rule. The important treatment of the territory around Tarquinia by Perego (2005) has identified many sites, not through survey but for the most part by the study of archive material, published notices and reports. The terrain included in this study extends northwards for 10 km as far as Montebello (Perego’s site 75) and even further to the south. In this large area very few of the sites listed are ‘new’. Moreover, tombs and necropolises form the bulk of them as opposed to habitation sites; in the Orientalizing period the ratio is more than 4:1 (46 necropolises, 11 habitation sites), in the Archaic period more than 2:1 (45 necropolises, 19 habitation sites). Nor should this be surprising, as tombs are a far more conspicuous feature in the landscape than small habitation sites. The raising of defensive stone walls at these sites can also give an indication of when urbanization was under way or had been completed. A conference devoted to this theme was held in 2005 (La Città Murata 2008) and drew attention to re-datings, mainly to earlier dates, of a number of major wall-circuits. Tarquinia’s was in place by the mid sixth century (La Città Murata 2008: 161; cf. Leighton 2004: 125), as was the wall in opera quadrata at Veii (La Città Murata 2008: 138), and Caere’s walls probably not long after (Bellelli 2016: 50–53). Vulci’s wall is still of uncertain date. Smaller towns within the city states seem to have been walled somewhat later. For San Giuliano, Norchia and Musarna, the date is probably fourth century

133

5 ETRUSCAN URBANIZ ATION

(La Città Murata 2008: 298, 300, 301). Musarna was founded in only the fourth century, but the other two are much older settlements. It was probably around this same time that the Etruscan walls of Tuscania were constructed (Quilici Gigli 1970: 158, figs. 228–229). As for the chronology of the process of urbanization for South Etruria as a whole, the parameters resulting from archaeological evidence, according to Carafa (2004), lie somewhere between the second half of the eighth and second half of the seventh centuries, but the crucial phase is probably towards the end of this period. The process of agglomeration towards urban centres was taking distinct shape around 725 bc, about the time that is usually considered as the dividing line between the Early Iron Age and Orientalizing periods in the region. It is the seventh century that sees the development of ‘palaces’ and sophisticated residencies across Etruria, such as at Acquarossa and Murlo (Riva 2006: 113). As to how the system ­operated – i.e. the nature of the political control exerted by the major cities on the smaller settlements within their territory, and of the relationship between towns and farms/rural sites – little is at present known. What field survey can do is at least identify evidence of various scales of human activity and provide both a counterbalance to the archaeology of the city elites as well as some understanding of the countryside which elites exploited and on which they ultimately depended (Stoddart 2007: 240).

that was assumed to give a ‘firm’ date, the earlier century a ‘possible’ date (as in Figs. 5.8a–d). Hence, we may talk of a ‘probable’ site ‘possibly’ dated to the sixth century, and so on. Fine wares, which are the most easily dated, are not common on rural sites. Of these, the imported Greek painted wares (Corinthian, Attic) were more associated with cemeteries and sanctuaries than with habitation sites, while the black or grey-black bucchero fine ware was more common but often found in small fragments that did not lend themselves to close dating or were of forms that are long-lived. Coarse ware formed the bulk of the Etruscan pottery from the Tuscania Archaeological Survey. Since the first integrated publication of the South Etruria Survey (Potter 1979), precision in the dating of both Etruscan fine wares and coarse wares has made considerable strides, and it is now often possible for finds from sites consisting entirely of coarse wares to be dated to within a few generations instead of to longer periods. In Etruscan archaeology it has often been a convention to divide the earlier periods into Early, Middle and Late Orientalizing (respectively 750–680 bc, 680–630 bc and 630–580 bc), followed by Archaic (580–480 bc). That is how the pottery sequences are arranged in the reworking of the South Etruria data (Di Giuseppe and Patterson 2009) and for the area of the Tiber Valley Project more generally (Di Giuseppe 2020: 84). It serves well enough with painted fine wares, but these are nearly absent in our material. Bucchero, the principal Etruscan fine ware of the Orientalizing and Archaic periods (Gran Aymerich 2017; Rasmussen 1979), is much more common but not often closely datable for the reasons stated above. We have therefore reverted in our Survey Gazetteer (Table A2.2) to using chronological divisions by century, as Potter (1979) did, and they were also employed in the Albegna Valley Survey (Cambi 2002; Perkins 1999). It is worth noting that Cambi pairs the earliest two centuries (700–500 bc), as we do in our final discussion in Chapter 9 (Table 9.2), whereas Perkins kept them separate, as we do for the more fine-grained discussion in this chapter. Of the coarse wares (Tables 5.1–5.3), only Fabrics 1, 5 and 6 have real numerical significance (Table 5.1). The commonest are Fabrics 1 and 5, which have overlaps in

establishing chronologies Pottery Technically, Etruscan pottery is rather different from that produced in the Early Iron Age: fine wares were made on a fast wheel, the more abundant coarse wares usually continued to be made on a slow wheel or sometimes by hand. We made a distinction between sherds that may be ‘firmly’ dated and those that may be ‘possibly’ dated to a particular century – as opposed to the terminology applied to sites which were graded either as definite or probable (see Chapter 2). Where pottery from a site could be given a date range of more than one century, it is the later century

134

establishing chronologies

Table 5.1  Tuscania Archaeological Survey: Etruscan coarseware fabrics: descriptions and percentages Fabric 1 (78.7%) ‘Rosso bruno’. Varies from grey to orange-brown. Scattered rounded cavities. Inclusions: many quartz, scattered white calcareous/grit (> 2mm), some black (feldspar?). Fracture: coarse. Fabric 2 (1.9%) Variant of Fabric 1, creamier in tone. Fabric 3 (0.2%) Variant of Fabric 1, greyish with larger inclusions. Fabric 4 (1.2%) ‘Chiara sabbiosa’. Also known as ‘coarse creamware’ (Perkins and Walker 1990: 31). Dull orange core (Munsell 5 YR 6/4), lighter surface (7.5 YR 7/4). Crumbly texture. Rounded cavities. Inclusions: many large augite (> 2mm), some quartz, some feldspar, few grains mica. Fracture: rough. Fabric 5 (12.3%) Orange. Munsell 5 YR 5/4 – 5 YR 4/4 (i.e. darker in centre). Angular cavities. Inclusions: few specks mica, scattered feldspar, considerable quartz, few large calcareous (> 2mm), few iron. Fracture: fairly rough. Fabric 6 (5.3%) Orange. Munsell 5 YR 5/3 – 4/3. Hard and compact. Some rounded cavities. Inclusions: many quartz, some augite, scattered specks mica, some iron. Fracture: fairly smooth. Fabric 7 ‘Fine creamware’. Cream, Munsell 5 YR 8/2. Compact, with smooth surface. Many small elongated cavities. Inclusions: few augite. Fracture: smooth. Fabric 8 (0.02%) Orange-brown. Homogeneous fabric, no cavities. Inclusions: many mica, few iron, some rounded black (augite), very few white calcareous. Fracture: powdery. Fabric 9 (0.1%) Brown. Homogeneous fabric, no cavities. Inclusions: black rounded (augite), many quartz, few mica grains. Fracture: rough. Munsell 7.5 YR 5/6. Fabric 10 (0.01%) Grey-brown. Munsell 7.5 YR 4/2 (fabric), 2.7 YR 4/6 (surface). Some rounded cavities. Inclusions: very few quartz, very few black (feldspar?), few calcareous, very few rounded orange (iron), very few specks mica. Fracture: rough. Fabric 11 (0.2%) Fabric varies from orange to grey. Frequent rounded cavities. Inclusions: frequent quite large (< 2mm) black (feldspar?), few quartz. Fracture: very rough. Fabric 12 (0.02%) Red-brown. Munsell 5 YR 4/6. Scattered round cavities. Inclusions: frequent large (> 2mm) quartz, few large (> 2mm) white calcareous, some black augite. Fracture: rough. Note: The percentages are of the total number of diagnostic sherds (N = 4,830) from all sites except the excavated site R24:19. Fabric 7 is a fine ware.

classification. Large storage vessels – the dolia and pithoi – are usually manufactured in Fabric 1, and sometimes the uncommon and grittier Fabric 11. The latter may belong to the Archaic period only; otherwise, all the fabrics have a chronological span that covers the whole Etruscan period. Fabric 1 is by far the most common, accounting for nearly 79 per cent of all coarseware finds (Table 5.1) and at 65 per cent the jar, or olla, is the most popular shape as well as being the best indicator – within wide parameters – of

physical composition, with the latter being usually darker in the core. Fabric 6 is rather similar in colour to Fabric 5 but is altogether harder and smoother. Fabric 4 is the coarse creamware described by Murray Threipland and Torelli (1970: 78), also known as ceramica chiara sabbiosa (Perkins and Walker 1990: 31), itself a less refined version of fine creamware (here, Fabric 7). Local transport amphorae are of Fabric 1 or Fabric 5; most of those that are identifiable are of Types 3a and 3b in Py’s (1984)

135

5 ETRUSCAN URBANIZ ATION

Table 5.2  Tuscania Archaeological Survey: Etruscan coarseware shapes and their percentages of diagnostic sherds (N = 1,637) Shape

date (Table 5.2). The coarseware shapes are illustrated in Appendix 1 (Figures A1.1–A1.8), and their principal occurrences are listed in the table accompanying these figures (Table A1.1). The Etruscan coarse wares are fairly standardized through central and southern Etruria, but only up to a point. They also show considerable variability, which is not surprising given that there are likely to have been numerous production centres. No kilns were noted in the Tuscania Archaeological Survey, but kiln wasters were recorded at T34:17 (1 sherd) and CP94:7 (3 sherds). Perkins (1999: 138) noted coarseware wasters also in the Albegna valley. Comparisons can be made with material from other localities, including Tarquinia, Caere and as far south as

%

jar bowl storage (unspecified) lid amphora basin dolium pithos brazier jug

64.7 8.6 7.3 6.7 5.2 3.5 2.8 0.6 0.1 0.1

Table 5.3  Tuscania Archaeological Survey: Etruscan coarse ware: numbers of sherds per fabric per shape/ type. F – fabric. Shape/Type

F1

F2

Jar Type 1 Jar Type 2 Jar Type 3 Jar Type 4 Jar Type 5 Jar Type 6 Total

3 28 34 37 58 32 192

1 2 1 4 1 3 12

Bowl Type 1 Bowl Type 2 Bowl Type 3 Bowl Type 4 Total

13 14 5 6 38

Base Type 1 Base Type 2 Base Type 3 Base Type 4 Base Type 5 Total

3 19 2

Basin Type 1 Basin Type 2 Basin Type 3 Basin Type 4 Basin Type 5 Total Total

24 4 1 7 5 17 271

1 2 3

F3

F4

F5

F6

2 6 2 10

11 7 10 10 2 5 45

5 3 8 12 2 30

1

2

1 1 3

1 3

1

1

1 1

1 1

F7

F8

F9

F10

F11/12

15 42 48 61 80 44 290

1 1

17 15 6 11 49 3 19 3 1 1 27

1 1

2

3

1

1

1 1 18

1

3 15

Total

1 49

33

136

1

1 1

5 5 7 1 5 23 389

establishing chronologies

Table 5.4  Tuscania Archaeological Survey: Etruscan bucchero fabrics: descriptions and percentages

Rome, but are best applied sparingly for the reasons above. Tol (2012: 35) also makes the point that most coarse ware is not closely datable and that by giving each sherd its fullest date range of production – as we have to – we run the risk of artificially prolonging the time span of each site. Turning to fine wares, fine impasto ranges through the eighth to sixth centuries and is most prevalent in the seventh. There are two kinds: red impasto and brown impasto, and in both cases the fabric is essentially that of coarse ware but more levigated and refined, and often with a well-polished surface. It was not common in the Tuscania countryside, and the material from the few sites where it was relatively abundant (e.g. T85:7, J2:7) is clearly funerary. Fine creamware (Fabric 7), sometimes called also plain white ware, is here listed within the coarse wares as its separation from Fabric 4 (coarse creamware) is a matter of degree only. It, too, is rare on rural sites: 26 fragments were recovered from the survey, together with 13 from the major site R14:13 on the southern edge of Tuscania; it was rather more frequent at the Etruscan farm site we excavated, Guidocinto1 (R24:19). Bucchero fabrics are usually classified according to the colour of both surface and core. So at Tarquinia, the large quantities of bucchero excavated from the acropolis area are separated into 37 fabrics from black to light grey (Locatelli 1999: 200). At Tuscania, given the very small size of most fragments recovered, the material did not lend itself to such precision of classification, and the fabrics distinguished are fewer in number (Table 5.4). Again, as at Tarquinia, the darkest/blackest – and thinnest – fabrics usually signify an early date. The finest is bucchero B8, which is rare; more common is the rather thicker bucchero B3. Bucchero B1 is altogether greyer, while bucchero B6 is grey in the core and brownish on the surface. The majority of the sherds were not diagnostic as to shape, but what could be ascertained was that drinking vessels were well represented, amounting to nearly a third (32.7 per cent of the material) (Table 5.5). The larger fragments came from R14:13 and a selection of these is illustrated (Figs. 5.2 and 5.3).

Fabric B1 (61.4%) Dark/very dark grey. Fine clay with few and small cavities and inclusions. Inclusions: white (calcareous), red (ferrous), also few mica grains. Fracture: smooth. Munsell 7.5 YR 5/ sandwiched between 7.5 YR 4/. Fabric B2 (9.6%) Grey. Many small round cavities. Some white (calcareous) and larger (< 1mm) orange (ferrous) inclusions. Fracture: fairly smooth. Munsell 7.5 YR 5/. Fabric B3 (23.2%) Dark grey/black. Compact. Well burnished smooth surface. Very few (< 1mm) white (calcareous) inclusions, very few mica specks. Fracture: smooth. Munsell 7.5 YR 3/. Fabric B4 (0.2%) Light grey. Small attenuated cavities. Few white (calcareous) inclusions (< 1mm). Fracture: smooth. Munsell 7.5 YR 6/ sandwiched between 7.5 YR 5/. Fabric B5 (1.4%) Dark grey. Few large attenuated cavities. Widely spaced small white inclusions (calcareous/grit). Very few mica grains. Fracture: smooth. Munsell 7.5 YR 4/. Fabric B6 (3.0%) Grey-brown with lighter slipped surface (7.5 YR 6/2). Sparse small white transparent inclusions (quartz), few larger (< 1mm) rounded black inclusions (feldspar?). Fracture: smooth. Munsell 7.5 YR 5/2. Fabric B7 (0.2%) Kiln waster. Grey-black core. Many white (calcareous) inclusions, some white transparent inclusions (mica? quartz?). Fracture: rough. Fabric B8 (1.1%) Black core and surface. Very fine hard fabric without internal cavities. Many small white (calcareous) inclusions, and a few mica grains. Fracture: smooth. Munsell 7.5 YR 2/. Note: The percentages are of the total number of diagnostic sherds (N = 664) from all units except the excavated site R24:19, together with 46 unfabricked sherds from the survey.

As with coarse ware, kiln wasters were found at R14:13 (bucchero B2) and R14:15, suggesting at least some local production. The primary source, however, was probably Tarquinia, which was both a production centre and a probable centre of redistribution for bucchero from Caere, in the same way that south of the Tiber valley Satricum

The locality is referred to as Guadocinto on some maps, but as the BSR and Soprintendenza excavation permit documentation refers to the site as Guidocinto, and that spelling has been used in earlier publications, it has been retained here.

1

137

4

1

2

3

5

6

7

8

9

10 11

13 12

14

16 15 0

5

17

cm

figure 5.2  Tuscania Archaeological Survey: bucchero from R14:13: 1. amphora foot (square B4, Fabric B1); 2,3. oinochoe handles (2: square B3, Fabric B1; 3: square C1, Fabric B1); 4. oinochoe foot (square B4, Fabric B1); 5. kantharos (square C3, Fabric B3); 6. chalice/kantharos/kyathos (square D3, Fabric B1); 7. kantharos/kyathos (square C1, Fabric B1); 8. kantharos/kyathos handle (square B4, Fabric B3); 9. kyathos (square B3, Fabric B1; Rasmussen 1979: kyathos Type 1e; cf. Pianu 2000: pl. 10.89,90); 10. cup handle (square B3, Fabric B3); 11. cup (square B3, Fabric B1); 12. cup/kotyle handle (square D1, Fabric B1); 13. flaring foot (square C1, Fabric B5); 14. low base (square C1, Fabric B1); 15–17. bowls: ring feet (15. square D3, Fabric B1; 16. square C3, Fabric B2; 17. square B4, Fabric B1). (Illustrations: Sally Cann.)

138

18

19

20

21

22

23

24 26 0

5 cm

25 figure 5.3  Tuscania Archaeological Survey: bucchero from R14:13: 18. lid/plate (square D3, Fabric B1; cf. Gran-Aymerich 2017: form 2221a1); 19–26. bowls: rims (19. square B4, Fabric B1; 20. square D3, Fabric B1; 21. square B3, Fabric B2; 22. square C1, Fabric B5 – Rasmussen 1979: bowl Type 3, cf. Pianu 2000: pl. 3.18; 23. square C3, Fabric B2, cf. Gran-Aymerich 2017: form 2984a1/b1; 24. square C3, Fabric B1; 25. square D1, Fabric B1 – Rasmussen 1979: bowl Type 4, cf. Pianu 2000: pl. 3.24; 26. square D1, Fabric B1 (or goblet? cf. Gran-Aymerich 2017: pl. 46). (Illustrations: Sally Cann.)

139

Table 5.5  Tuscania Archaeological Survey: Etruscan bucchero shapes and their percentages of diagnostic sherds (N = 175); percentage of drinking vessels: 32.7% Shapes

%

bowl kantharos/kyathos kantharos/kyathos/chalice oinochoe jug miniature bowl cup/kotyle cup kantharos kotyle plate lid kyathos miniature kyathos olpe chalice

50.8 14.0 7.0 4.3 3.8 3.8 3.2 2.7 2.1 2.1 2.1 1.6 1.1 0.5 0.5 0.5

0

5 cm

figure 5.4  Fragment of Etrusco-Corinthian ware from T85:7. (Photograph: Tom Rasmussen.)

Fig.  2.12) and only a sample removed for fabric analysis using a tile fabric series developed by Chris Hunt (Table  5.6). The associations with pottery indicated that many tile fabrics were extremely long-lived, such as Fabric A, which was used from the Roman to the Post-Medieval centuries, while others can be assigned only to long periods covering many centuries, such as Fabric C (Etruscan into Roman), Fabric D (Medieval), Fabric H (Roman Republican and later) and Fabric K (Roman). Fabrics B, G, M and O, however, were consistently associated with Etruscan pottery, and Fabrics F, J and L were commonly associated with Etruscan pottery. These tiles were notably heavy, coarse-grained, with a reddish exterior and usually darker grey or brown in the interior. Decorative architectural terracottas also formed part of the Etruscan roofing system. Few such terracottas were recovered, but the most striking was one from site R54:2 featuring a complex meander of a kind that can be paralleled throughout South Etruria and Latium in the later sixth century (Winter 2009: 353, scheme 5.D.1; Fig. 5.6). Most of these revetment plaques have a figured frieze

is likely to have been a centre for the redistribution of bucchero for coastal Latium (as noted by Tol 2012: 369). Several of the Tuscania types have good parallels with the bucchero found at Gravisca, on the coast near Tarquinia, where the chronological parameters are 600/580–480 bc. Etrusco-Corinthian amounts to thirteen fragments, mostly very small, of which the majority have conventional linear (banded) decoration. The most striking pieces are from T85:7 (Fig. 5.4) and belong to the Gruppo ad Archetti Intrecciati (interlaced semicircles) first isolated by Colonna (1961). Imported fine wares are rare. There are 34 Attic fragments; most consist of very small black-glazed fragments/scraps, though there is a little black-figure (Figs. 5.15, 5.22) and red-figure (Fig. 5.23).

Tile No doubt not all Etruscan buildings had tiled roofs, but tiles often provide the bulk of the finds from those that did. In some cases tiles were so numerous that they were weighed en masse on site as at R14:13 (Fig. 5.5; see also

140

settlement densities

A4

B4 3.0

C4 37.4

A3

B3 5.0

A2

C3

B2 21.3

A1

41.5

41.9

5.2

B0

D3

C2

B1

55.7

57.4

51.6

population was one that lived in towns, and between the towns the countryside consisted of unpopulated spaces. It was the same for George Dennis (1848), and although he realized that much Etruscan wealth was provided by crop cultivation and cattle rearing – for he knew his Vergil well (Georgics 2: 513–533) – he had little idea that the countryside was itself populated. The countryside must indeed have been rather sparsely inhabited in the Bronze Age and throughout much of the Iron Age, but a major result of the South Etruria Survey was the discovery that as the first millennium bc progressed so the landscape became increasingly filled with rural habitations (Potter 1979). In his discussion of this process Potter distinguished between ‘open’ and ‘defended’ sites, which may also be described as ‘dispersed’ and ‘nucleated’ respectively (Potter 1979: 14). It is the open/dispersed sites, which were both small-scale and undefended, that gradually spread over the landscape, a form of settlement that was sustainable only in politically stable times. The South Etruria Survey was conducted over a very wide area north of Rome (Fig. 1.3) and more recently was extended to the further side of the Tiber valley (Patterson et al. 2020). Not all the area was surveyed to the same intensity: some parts were systematically field-walked, in others the finds and the data were recovered in a more random manner. The Tuscania Archaeological Survey was envisaged from the outset as a systematically intensive survey of a relatively restricted area, while at the same time large enough to encompass a varied range of terrain and possible settlement types – from a small urban centre to various types of agricultural and rural installations. As discussed in Chapter 2, the quality and quantity of the distributions of surface material that we encountered with Etruscan and/ or Roman pottery gave us confidence to classify sites as either definite or probable, and distinguish these from non-site units, rather than use the three categories of site (definite, probable, possible) that we employed for the sparser surface archaeology of the prehistoric and postRoman periods. We recorded Etruscan pottery at 602 units that included 166 definite and 143 probable sites. To these

D4

47.9 D2

65.7 C1

59.7 D1

25.0

33.6

14.4

3.9

3.1

4.3

C0

D0 60

0 m

figure 5.5  Tuscania Archaeological Survey: gridded site R14:13: tile weights (kg). The grid squares measured 20 m x 20 m.

below the meander and relief tongues above; but in our example the meander comes immediately below the convex crowning moulding. Many Etruscan buildings were provided with architectural terracottas to protect their upper parts, but the decorative treatment here is evidence that there is likely to have been a temple/shrine or highstatus residence at or near the site.

settlement densities The quotation at the beginning of this chapter by Hamilton Gray, who wrote the first popular account in English of the Etruscans, was from when she was contemplating the lands south of Tuscania. From that vantage point she mentions that she could see in the distance (in different directions) Tuscania, Tarquinia, Viterbo and Montefiascone. For her, the Etruscan

141

5 ETRUSCAN URBANIZ ATION

Table 5.6  Tuscania Archaeological Survey tile fabric series. (Descriptions: Chris Hunt.) sub-millimetre to c. 8 mm. Voids ±5%, mostly rather angular, 1–2 mm across. Inclusions ±50%, so looks very rubbly, subrounded to angular includes rare large (5–8 mm) grey speckled white volcanic, possibly a welded tuff or very fine pumice. Grey/brown mottled meta-sedimentary, possibly a clay-ironstone or siltstone. Common small (