Multi-period Occupation at Football Field, Worth Matravers, Dorset: Excavations 2006–2011 9781407316741, 9781407323305

Citation preview

BAR BRIT ISH SE RIE S 643

Lilian Ladle is an independent field archaeologist and a Visiting Fellow at the Centre for Archaeology and Anthropology, Bournemouth University. She directed a 13-year project on the multi-period site at Bestwall Quarry, Wareham and is currently directing the Druce Farm Roman Villa excavations near Dorchester. She was awarded the MBE in 2008 for services to archaeology. CONTRIBUTORS M. Allen, P.S. Bellamy, D. Challinor, M. Corney, T. Cowie, S. Cullinane, L. Evans, E.  Firth, J. Henderson, D.A. Hinton, R. Kenyon, A. Krus, R. Lancaster, M. Ladle, M.  Lyne, A. Morgan, B. O’Connor, C. Randall, J. Randall†, R.M.A. Trevarthen, P. Walton Rogers ILLUSTRATIONS J. Bartlet, M. Corney, G. Naylor

Multi-period Occupation at Football Field, Worth Matravers, Dorset

This book presents the detailed results of excavations at a small multi-period site in south-east Dorset. The site provided evidence for an Early Neolithic enclosure, a timberframed, Late Bronze Age roundhouse, Late Bronze Age/Early Iron Age middening activities, Iron Age and Roman settlement and a post-Roman inhumation cemetery. The midden and its associated features and finds are of particular significance, with evidence for deposition and accumulation over several centuries. Other important features include: a Late Iron Age stone-lined pit containing over 100 near-complete pots and a substantial quantity of animal bone, suggesting a large-scale communal gathering and feasting activity; a Roman barn, used as a shale workshop and as a burial place for infants; a 6th- to 8th-century AD cemetery of single, double and triple graves, illuminating the nature of change in burial practices in Dorset at this time; and significant pottery and animal bone assemblages, particularly from the Late Bronze Age/Early Iron Age and Late Iron Age.

LADLE

‘The presentation of the results of this excavation adds greatly to understanding of the archaeology of the site itself and of a wider area. … Elements of the report will be of value to specialists investigating more national or even international topics.’ Steve Wallis, Senior Archaeologist, Dorset County Council

BAR 643 2018

‘This is a very significant contribution to our understanding of an important site. … The LBE/EIA pottery assemblage and the LIA pottery assemblage are major groups which provide well dated and detailed information which is of considerable significance to British and European scholars.’ Professor Niall Sharples, Cardiff University

Multi-period Occupation at Football Field, Worth Matravers, Dorset Excavations 2006–2011

Lilian Ladle BAR BRITISH SERIES 643

2018

Multi-period Occupation at Football Field, Worth Matravers, Dorset Excavations 2006–2011

Lilian Ladle BAR BRITISH SERIES 643

2018

Published in by BAR Publishing, Oxford BAR British Series Multi-period Occupation at Football Field, Worth Matravers, Dorset © Lilian Ladle Late Iron Age pit 304. The Author’s moral rights under the UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reser ved. No par t of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any for m of digital for mat or transmitted in any for m digitally, without the written per mission of the Publisher.

ISBN 9781407316741 paperback ISBN 9781407323305 e-format DOI https://doi.org/10.30861/9781407316741 A catalogue record for this book is available from the British Library

BAR titles are available from: BAR Publishing Banbury Rd, Oxford, [email protected] + ( ) + ( ) www.barpublishing.com

,

Acknowledgements The fieldwork at Football Field, Worth Matravers, was carried out by members of the East Dorset Antiquarian Society (EDAS) between 2006 and 2011 on land earmarked for social housing, which was owned by society member Bob (Robert) Kenyon. Bob is thanked for securing funding from the Worth Community Property Trust which allowed the initial commissioning of archaeological specialists to work on various categories and aspects of the finds. A substantial grant from the Dorset Archaeological Committee funded a radiocarbon dating programme. The Worth Matravers community are also thanked for their generous donations and support. The results and conclusions presented in this volume were achieved by a huge amount of hard work, enthusiasm and the dedication of a number of individuals. From 2006 to 2010 fieldwork was directed by Phil Roberts and from 2011 to the completion of the project by Lilian Ladle, who was also responsible for overseeing the production of this report. A debt of gratitude is owed to Andrew Morgan (EDAS chairman) who skilfully managed the financial accounts and helped facilitate the pre-publication processes as well as being involved in the fieldwork and post-excavation work. Janet Bartlet’s superb illustrations of hundreds of finds enhance the text. Mark Corney and Gareth Naylor prepared the drawings for publication. Special thanks are extended to a small number of EDAS members who were crucial to the completion of the project, committing months of their time; this included Janet Bartlet, Sue Cullinane, Len Norris and Bryan Popple. Pam Norris and Vera McDonald washed all finds on site, often in very difficult circumstances. The following members both past and present helped during the excavation work: Kerry Barrass, Corinne Board, Gill Broadbent, John and Della Day, Sonia Ellingham, Alan Hawkins, Maureen and Malcolm Houghton, Dr Steve McDonald, Dr David and Dorothy Michell, Steve Morns, David Stewart, Gill Stolworthy and Peter Walker. †Dr Ken Wheatley undertook the metal detecting survey. In addition, Prof. Tim Darvill and Prof. Niall Sharples gave much appreciated advice.

Grateful thanks are extended to all the specialists named below who contributed towards the artefact studies and whose results and interpretations have enhanced the conclusions: Dr Mike Allen – preliminary charred seed work and land molluscs, Peter Bellamy and Mike Trevarthen (Terrain Archaeology) – worked flint and stone, Dr Dana Challinor – charcoal, Mark Corney – coins and Iron Age and Roman metalwork, Sue Cullinane – shale, Laura Evans – charred seeds, Emma Firth (AC Archaeology) – fuel ash slag, briquetage and fired clay, Prof. Julian Henderson (University of Nottingham) – prehistoric glass, Prof. David Hinton (Southampton University) – PostRoman buckle, Robert Kenyon – environmental change, Dr Anthony Krus – radiocarbon dating, Robert Lancaster – clay tobacco pipe, Dr Malcolm Lyne – Late Iron Age and Roman pottery, Andrew Morgan – landscape, geology and climate change, Dr Brendan O’Connor and Trevor Cowie – Late Bronze Age/Early Iron Age metalwork, Dr Clare Randall (Context One Archaeology) – human and animal bone, †Jane Randall – Late Bronze Age/Early Iron Age ceramic decoration, Penelope Walton Rogers (The AngloSaxon Laboratory) – Post-Roman fabric analysis, Dave Stewart – geophysical survey. Brendan O’Connor thanks Trevor Cowie, Susy Kirk, Kate Waddington and Adam Gwilt. Lilian Ladle prepared reports on the prehistoric pottery, fired clay objects and ironwork and with Dr Mike Ladle, a report on the marine mollusca. I am indebted to Mike Ladle, Andrew Morgan, Clare Randall and Matthew Tagney who painstakingly checked the text, figures and tables. Sigrid Padel and Gabrielle Delbarre are thanked for translating the summary into German and French. I am grateful for the helpful comments of the three anonymous referees. All text not attributed to individual contributors is by the author who accepts full responsibility for any errors or omissions. All photographs are from the Worth Project archive apart from Figure 4 which was donated by EDAS member Jo Crane. The archive was prepared for deposition by the author, Janet Bartlet, Phil D’Eath, Carol O’Hara and Andrew Morgan.

Contents List of Figures..................................................................................................................................................................... ix List of Tables...................................................................................................................................................................... xv Summary......................................................................................................................................................................... xviii Résumé.............................................................................................................................................................................. xix Zusammenfassung............................................................................................................................................................ xx 1. Environment and Project Background......................................................................................................................... 1 Location........................................................................................................................................................................... 1 Landscape, Geology, Environment and Climate Change  Andrew Morgan and Robert Kenyon.................................. 1 Project Background.......................................................................................................................................................... 4 Summary of Phasing........................................................................................................................................................ 8 Radiocarbon Dating......................................................................................................................................................... 9 2. The Neolithic, Early, Middle and Late Bronze Age Periods......................................................................................11 Neolithic Enclosure Ditch.............................................................................................................................................. 11 The Stone Alignment..................................................................................................................................................... 16 Other Early Bronze Age Features.................................................................................................................................. 17 Middle Bronze Age Activity.......................................................................................................................................... 18 Late Bronze Age Activity............................................................................................................................................... 18 Conclusions.................................................................................................................................................................... 21 3. The Late Bronze Age/Early Iron Age Transition....................................................................................................... 22 Introduction.................................................................................................................................................................... 22 The Cobbled Surface...................................................................................................................................................... 22 The Midden Deposit....................................................................................................................................................... 23 The Features................................................................................................................................................................... 23 Form and Type of Pits.................................................................................................................................................... 23 Cultural Deposits............................................................................................................................................................ 38 Postholes........................................................................................................................................................................ 40 Structural Elements Associated with a Possible Post-Built Roundhouse (House 2)..................................................... 41 Settings of Packed Limestone........................................................................................................................................ 42 Conclusions.................................................................................................................................................................... 42 4. The Middle and Late Iron Age..................................................................................................................................... 44 Middle Iron Age Activity............................................................................................................................................... 44 Late Iron Age Activity.................................................................................................................................................... 44 The Roundhouses........................................................................................................................................................... 48 Pits and Scoops.............................................................................................................................................................. 51 Deposits.......................................................................................................................................................................... 58 Burial 318....................................................................................................................................................................... 58 Conclusions.................................................................................................................................................................... 58 5. The Romano-British Period......................................................................................................................................... 60 Introduction.................................................................................................................................................................... 60 The Building.................................................................................................................................................................. 60 Features Associated with the Building........................................................................................................................... 63 Burials Associated with the Building............................................................................................................................. 69 Pits and Deposits............................................................................................................................................................ 71 Other Features................................................................................................................................................................ 73 v

Multi-period Occupation at Football Field, Worth Matravers, Dorset Conclusions.................................................................................................................................................................... 75 6. The Post-Roman Burials.............................................................................................................................................. 77 Introduction.................................................................................................................................................................... 77 Adult Burial from Site 1................................................................................................................................................. 77 The Post-Roman Cemetery............................................................................................................................................ 78 Later Features................................................................................................................................................................. 90 Conclusions.................................................................................................................................................................... 92 7. Flaked Stone, Worked Stone and Kimmeridge Shale................................................................................................ 95 Flaked Stone  Peter S. Bellamy and R.M.A. Trevarthen.............................................................................................. 95 Worked Stone  Peter S. Bellamy and R.M.A. Trevarthen........................................................................................... 104 The Kimmeridge Shale  Sue Cullinane and Lilian Ladle (with contributions from Emma Firth)............................ 110 8. The Early Prehistoric Pottery.....................................................................................................................................117 Introduction.................................................................................................................................................................. 117 Methods and Aims....................................................................................................................................................... 117 Fabric and Form Codes................................................................................................................................................ 117 Neolithic and Early Bronze Age Pottery...................................................................................................................... 119 Late Neolithic/Early Bronze Age Pottery.................................................................................................................... 121 Beaker Pottery from Other Features............................................................................................................................ 121 Middle Bronze Age Pottery.......................................................................................................................................... 123 Late Bronze Age Pottery.............................................................................................................................................. 125 Conclusions.................................................................................................................................................................. 129 9 The Late Bronze Age/Early Iron Age Pottery........................................................................................................... 131 Introduction.................................................................................................................................................................. 131 Radiocarbon Dating..................................................................................................................................................... 131 Fabric and Form........................................................................................................................................................... 131 The Pottery Assemblage............................................................................................................................................... 133 Assemblage from the Midden Layers.......................................................................................................................... 143 The Assemblage........................................................................................................................................................... 165 Cultural Associations................................................................................................................................................... 165 Vessel Types at Football Field...................................................................................................................................... 165 Vessels in Use at Football Field................................................................................................................................... 170 Discussion.................................................................................................................................................................... 174 Conclusion................................................................................................................................................................... 175 Comparative Analyses of the Main Decorative Surface Treatments Found on the All Cannings Cross Ware Assemblage from Football Field  Jane Randall†....................................................................................................... 176 10. The Middle Iron Age, Late Iron Age and Roman Pottery.................................................................................... 178 Middle Iron Age Pottery  Lilian Ladle....................................................................................................................... 178 Late Iron Age and Roman Pottery  Malcolm Lyne..................................................................................................... 181 11. Finds Of Metal, Slag, Glass, Fired Clay and Worked Bone.................................................................................. 192 Iron Age and Roman Coins  Mark Corney................................................................................................................ 192 Bronze Age Metalwork  Brendan O’Connor, with contributions by Trevor Cowie................................................... 192 Middle-Late Iron Age, Roman, Medieval and Post-Medieval Objects of Copper Alloy  Mark Corney and Lilian Ladle..................................................................................................................................... 195 The Post-Roman Buckle  David A. Hinton................................................................................................................ 198 A Note on the Mineral-Preserved Textile Adhering to the Buckle  Penelope Walton Rogers................................... 199 Objects of Lead  Lilian Ladle.................................................................................................................................... 199 Objects of Iron  Lilian Ladle...................................................................................................................................... 200 Fuel Ash Slag  Emma Firth........................................................................................................................................ 202 Prehistoric Glass.......................................................................................................................................................... 203 The Glass Object  Lilian Ladle.................................................................................................................................. 203 vi

Contents Archaeological and Scientific Aspects of the Football Field Glass Ring Julian Henderson ................................... 203 The Roman and Later Glass Mark Corney .............................................................................................................. 207 Daub and Fired Clay Emma Firth ............................................................................................................................ 207 Spindle Whorls and Modified Pottery Objects Lilian Ladle .................................................................................... 208 Briquetage Emma Firth ............................................................................................................................................ 210 A 17th Century Clay Tobacco Pipe Robert Lancaster ............................................................................................. 211 Worked Bone and Antler Clare Randall .................................................................................................................. 211 12. Radiocarbon Dating and Bayesian Modelling of the Post-Roman Cemetery Anthony Krus .......................... 216 Methodological Approach ........................................................................................................................................... 216 Stable Isotopes and Marine Correction ....................................................................................................................... 218 Post-Roman Cemetery Samples and the Model .......................................................................................................... 219 Conclusions ................................................................................................................................................................. 220 13. The Human Remains Clare Randall..................................................................................................................... 221 Introduction ................................................................................................................................................................. 221 Methodology ............................................................................................................................................................... 221 The Prehistoric Human Remains ................................................................................................................................ 221 Romano-British Infant Remains ................................................................................................................................. 222 Discussion ................................................................................................................................................................... 225 The Post-Roman Period .............................................................................................................................................. 226 The Post-Roman Cemetery ......................................................................................................................................... 226 Results ......................................................................................................................................................................... 228 Discussion ................................................................................................................................................................... 229 Conclusions ................................................................................................................................................................. 233 14. The Faunal Remains Clare Randall ..................................................................................................................... 234 Introduction ................................................................................................................................................................. 234 Methodology ............................................................................................................................................................... 234 Results ......................................................................................................................................................................... 234 Preservation and Taphonomy ...................................................................................................................................... 236 Late Neolithic/Early Bronze Age ................................................................................................................................ 236 Late Bronze Age.......................................................................................................................................................... 245 Late Bronze Age/Early Iron Age Features .................................................................................................................. 245 Late Bronze Age/Early Iron Age Midden ................................................................................................................... 249 Middle Iron Age Pit 1182 ............................................................................................................................................ 251 Late Iron Age............................................................................................................................................................... 253 Romano-British Period................................................................................................................................................ 258 The Mixed Midden...................................................................................................................................................... 260 Post-Roman Grave 425 ............................................................................................................................................... 260 Wild Species ................................................................................................................................................................ 261 Conclusions ................................................................................................................................................................. 262 15. The Environmental Remains .................................................................................................................................. 267 Introduction ................................................................................................................................................................. 267 Palaeo-environmental Assessment Mike Allen......................................................................................................... 267 The Charred Plant Remains Laura Evans (identification), Lilian Ladle and Clare Randall (discussion) .............. 268 Wood Charcoal Dana Challinor .............................................................................................................................. 276 Land Mollusca Mike Allen ....................................................................................................................................... 281 The Marine Molluscs Mike and Lilian Ladle........................................................................................................... 281 Fish Clare Randall ................................................................................................................................................... 283 16. Interpretation and Discussion ................................................................................................................................. 285 Introduction ................................................................................................................................................................. 285 Neolithic and Early Bronze Age ................................................................................................................................. 285 Middle and Late Bronze Age ...................................................................................................................................... 289 vii

Multi-period Occupation at Football Field, Worth Matravers, Dorset Late Bronze Age/Early Iron Age Transition................................................................................................................ 291 Middle Iron Age........................................................................................................................................................... 298 Late Iron Age............................................................................................................................................................... 302 Roman Period............................................................................................................................................................... 307 Post-Roman Period...................................................................................................................................................... 312 Late Saxon Worth Matravers........................................................................................................................................ 316 Conclusions.................................................................................................................................................................. 317 Archiving and Potential Future Work.......................................................................................................................... 317 Bibliography.................................................................................................................................................................... 318

viii

List of Figures Figure 1 Key to section and plan drawings........................................................................................................................... 1 Figure 2 Worth Matravers, Dorset, site location and geology.............................................................................................. 2 Figure 3 Location of Football Field and EDAS excavation areas in relation to Worth Matravers and Compact Farm....... 3 Figure 4 Aerial view of Worth Matravers............................................................................................................................. 3 Figure 5 Contour survey and archaeological interventions 1991-2011................................................................................ 4 Figure 6 Main topographical features in the Isle of Purbeck................................................................................................ 5 Figure 7 Site 2; Neolithic enclosure 1332 and its relationship with pits 1182 and 1131. Location of evaluation pits, Trenches 1 and 2 and Early Bronze Age postholes............................................................................................................. 12 Figure 8 Section of evaluation pit 1252.............................................................................................................................. 13 Figure 9 Section of evaluation pits 1308, 1270 and 1305................................................................................................... 13 Figure 10 Plan of Trenches 1 and 2 showing features visible at an excavated depth of 0.7m............................................ 13 Figure 11 Excavated section Trench 1................................................................................................................................ 14 Figure 12 Excavated section Trench 1, view from the north-west..................................................................................... 15 Figure 13 Excavated section Trench 2................................................................................................................................ 15 Figure 14 Sections of postholes 1616, 1618 and deposit 1628 in Trench 1 and scoop 1323 in evaluation pit 1320......... 16 Figure 15 Plan of stone alignments 1219/1477 and 1145/1486 and section C-D on the northern edge of the excavation area.................................................................................................................................................................... 17 Figure 16 Stone alignment, sections 1219 and 1145 during excavation, view to the north............................................... 17 Figure 17 Sections of Early Bronze Age postholes 1257, 1295, 1339 and scoop 1340..................................................... 18 Figure 18 Section of Late Bronze Age pit 1754 and associated posthole 1761 (Site 4)..................................................... 18 Figure 19 Late Bronze Age House 1, Site 2. View to the west........................................................................................... 19 Figure 20 Late Bronze Age House 1, plan.......................................................................................................................... 20 Figure 21 Late Bronze Age House 1, sections of structural postholes............................................................................... 20 Figure 22 Late Bronze Age House 1, sections of ancillary postholes................................................................................ 20 Figure 23 Late Bronze Age/Early Iron Age cobbled surface 1039 (Site 2). View to the west........................................... 22 Figure 24 Designated slot numbers across Site 2............................................................................................................... 24 Figure 25 Site 1 sections through midden layers (located on Fig. 28)............................................................................... 26 Figure 26 Site 2, sections 1 and 2 through midden layers on the northern edge of the site............................................... 27 Figure 27 Site 2, section 3 through the midden layer on the western side of site............................................................... 28 Figure 28 Site 1 plan Late Bronze Age/Early Iron Age features associated with the midden deposit............................... 29 Figure 29 Site 2 plan Late Bronze Age/Early Iron Age features associated with the midden deposit............................... 30 Figure 30 Late Bronze Age/Early Iron Age Type 1 pit 335, section................................................................................... 31 Figure 31 Late Bronze Age/Early Iron Age Type 1 pit 335, view from the north.............................................................. 31 Figure 32 Late Bronze Age/Early Iron Age Type 1 pit 326/390, section............................................................................ 32 Figure 33 Late Bronze Age/Early Iron Age Type 1 pit 326/390, view from the west........................................................ 32 Figure 34 Late Bronze Age/Early Iron Age Type 1 pit 1412, section................................................................................. 33 ix

Multi-period Occupation at Football Field, Worth Matravers, Dorset Figure 35 Late Bronze Age/Early Iron Age Type 1 pit 1412, view from the north ........................................................... 33 Figure 36 Late Bronze Age/Early Iron Age Type 1 pits 1513 and 1302, sections ............................................................. 34 Figure 37 Late Bronze Age/Early Iron Age Type 1 pit 1513. View from the east ............................................................. 34 Figure 38 Late Bronze Age/Early Iron Age Type 2 pits, 1126, 1503, 1574 and 393, sections .......................................... 35 Figure 39 Late Bronze Age/Early Iron Age Type 3 pits 1022, 1142, 1215, 1220, 1241, 1263, 1427, 1468 (Site 2) and 323 (Site 1), sections ................................................................................................................................................... 36 Figure 40 Quern 1036 with associated ceramic objects to the left. The quern covered Type 3 pit 1142. View from the south .................................................................................................................................................................... 37 Figure 41 Type 3 pit 1241, view from the north ................................................................................................................ 38 Figure 42 Deposit 1169 with pottery fragments and pebbles, view from the north-east ................................................... 39 Figure 43 Deposit 1175 with animal bone, view from the north ....................................................................................... 39 Figure 44 Limpet deposit 1500, viewed from the south-east............................................................................................. 40 Figure 45 Late Bronze Age/Early Iron Age postholes associated the midden deposit on Sites 1 and 2, sections ............. 41 Figure 46 Stone setting 1179 inserted into cobbles 1039, view from the west.................................................................. 42 Figure 47 Stone setting 1520 inserted into cobbles 1039, view from the north ................................................................ 43 Figure 48 Site 1 plan, Late Iron Age features and Post-Roman grave 425 ........................................................................ 45 Figure 49 Site 2 plan, Middle and Late Iron Age features ................................................................................................. 46 Figure 50 Middle Iron Age pit 1182 and Late Iron Age pit 1131, section ......................................................................... 47 Figure 51 Middle Iron Age pit 1182 and Late Iron Age pit 1131 (directly above), view from the north .......................... 47 Figure 52 Late Iron Age House 3, Site 1. View from the north ......................................................................................... 48 Figure 53 Oven/dryer 364 sections and plan ..................................................................................................................... 49 Figure 54 Oven/dryer 364 attached to the north-east side of House 3; burning evident as red soil. View from the north .... 49 Figure 55 Late Iron Age House 4, Site 2; foundation 1194 visible. View from the east ................................................... 50 Figure 56 Paving 1037 associated with Late Iron Age House 4, Site 2. View to the south ............................................... 51 Figure 57 Late Iron Age Type 1 pit 304, section ............................................................................................................... 52 Figure 58 Late Iron Age Type 1 pit 304 during excavation, view to the south.................................................................. 52 Figure 59 Late Iron Age Type 1 pit 304 after excavation. View to the north .................................................................... 53 Figure 60 Late Iron Age Type 1 pit 304 after excavation, view to the south ..................................................................... 53 Figure 61 Late Iron Age Type 1 pit 1740, section ............................................................................................................. 54 Figure 62 Late Iron Age Type 2 pits 327 and 324, sections............................................................................................... 55 Figure 63 Late Iron Age Type 3 pits 1209, 1408, 1483, 1532, 332 and 1549, sections..................................................... 55 Figure 64 Late Iron Age Type 3 pit 1131 with cow scapula in section, view to the south ................................................ 56 Figure 65 Late Iron Age Type 3 pit 1209, view to the east. ............................................................................................... 57 Figure 66 Late Iron Age Type 3 pit 1532 and stone deposit 1531, view from the north ................................................... 58 Figure 67 Fragmented pottery deposit 1019, view to the west .......................................................................................... 59 Figure 68 Grave 318, with skeleton 320, view to the south .............................................................................................. 59 Figure 69 Site 2, Roman building and associated features, plan ....................................................................................... 61 Figure 70 Site 2 after stripping and cleaning in 2010. View to the north-east................................................................... 62 Figure 71 Padstone 1008, was moved by mechanical excavator during the initial strip of the site. View to the south west......62 Figure 72 Features pre-dating the Roman barn, sections. .................................................................................................. 63 Figure 73 Length of wall 1018 in the foreground and elements of 1397 exposed beyond. View to the north-east .......... 64 x

List of Figures Figure 74 Door threshold 1419 and 1420 and large porch slab 1426. View to the south................................................... 65 Figure 75 Threshold elements of the Roman building, plan............................................................................................... 65 Figure 76 Threshold stones 1419 and 1420 with bolt holes and grooves. View to the south............................................. 66 Figure 77 Padstone 1415 with holes for wooden uprights, view to the west...................................................................... 66 Figure 78 Walls 1450 and beyond, 1398, with remnants of paving 1488. View to the north-east..................................... 67 Figure 79 Wall 1450 (centre) showing construction and wall 1398 in front. View to the south-west................................ 67 Figure 80 Wall 1450, section.............................................................................................................................................. 68 Figure 81 Triangular stone 1405 with pecked circular depression..................................................................................... 68 Figure 82 Burial cist 1123 with infant skeleton 1174 on the base. View to the south........................................................ 69 Figure 83 Infant remains within cist 1123, plan................................................................................................................. 69 Figure 84 Sump 1451 and gully 1454, view to the north-east............................................................................................ 70 Figure 85 Burial group 1025 comprising three infants, view to the west........................................................................... 70 Figure 86 Infant burial 1440 underneath threshold stones 1419 and 1420. View to the west............................................ 71 Figure 87 Roman Type 3 pits 1438, 1184, 1421 and accumulated deposit 1465, sections................................................. 72 Figure 88 Antler bone visible on the cleaned surface of deposit 1465, view to the north-east.......................................... 73 Figure 89 Pottery deposit 1505, view to the north.............................................................................................................. 74 Figure 90 The water pipe trench (Site 4) with recorded features, plan............................................................................... 75 Figure 91 Type 2 Roman pits 1749 and 1751, sections...................................................................................................... 75 Figure 92 Section through Trench 1748 with pit 1751, cobbles 1762 and deposits 1752 and 1759.................................. 76 Figure 93 Partial remains of adult female 311 in grave 425, view to the west................................................................... 77 Figure 94 Site 3 plan; the Post-Roman cemetery with grave types, pavement 1695 and Middle Bronze Age activity area 1700................................................................................................................................................................ 78 Figure 95 Type 1 Graves 1633, 1667 and 1682, plans........................................................................................................ 80 Figure 96 Type 1 Grave 1633 with skeletons 1632 and 1652, the head if 1632 rests on the stone anchor. View to the south 81 Figure 97 Type 1 Grave 1667 and skeleton 1669. View to the north.................................................................................. 81 Figure 98 Type 2 Graves 1664, 1670 and 1697, plans........................................................................................................ 82 Figure 99 Type 3 Graves 1660 and 1722, plans.................................................................................................................. 83 Figure 100 Type 3 Grave 1660 and skeleton 1662 during excavation, view to the north.................................................. 84 Figure 101 Type 3 Grave 1722 with double burial (skeletons 1724 and 1725), view to the south.................................... 84 Figure 102 Type 4 Graves 1640, 1643, 1646 and 1654, plans............................................................................................ 85 Figure 103 Type 4 Graves 1640 and 1654 before excavation. View to the west................................................................ 86 Figure 104 Type 4 grave 1640 with skeleton 1642 exposed. View to the north................................................................. 86 Figure 105 Type 4 Grave 1643, with the head of skeleton 1645 exposed. View to the north............................................ 87 Figure 106 Grave 1643 with skeleton 1645 exposed, view to the north............................................................................. 87 Figure 107 Type 5 grave 1649, plan................................................................................................................................... 88 Figure 108 Type 6 graves 1675 and 1678, plans................................................................................................................. 89 Figure 109 Type 6 grave 1678, with skeletons 1680 and 1681. View to the west.............................................................. 90 Figure 110 Type 7 grave 1685 with skeletons 1688, 1687 and 1693, plan......................................................................... 91 Figure 111 Type 7 grave 1685; initial cleaning showing horizontal slabs and skeleton 1687 partially exposed. View to the west.................................................................................................................................................................. 91 xi

Multi-period Occupation at Football Field, Worth Matravers, Dorset Figure 112 Type 7 grave 1685 with skeletons 1688, 1693 and 1687. View to the west..................................................... 92 Figure 113 Plan and sections of wall/pavement 1695/1775............................................................................................... 93 Figure 114 Wall/pavement 1695 with entrance (visible centre top) and its relationship to grave 1685. View to the west.... 94 Figure 115 Eastern extent of wall 1775 with layer 1777 exposed. View to the east.......................................................... 94 Figure 116 Flaked stone (1-9)............................................................................................................................................. 98 Figure 117 Flaked stone (10-18)....................................................................................................................................... 102 Figure 118 Saddle and rotary querns (1-3)....................................................................................................................... 106 Figure 119 Grinding stones and lid (4-8).......................................................................................................................... 107 Figure 120 Stone anchor and triangular stone with dished centre (9-10)......................................................................... 108 Figure 121 Kimmeridge shale objects by period.............................................................................................................. 110 Figure 122 Objects of Kimmeridge shale (1-15)...............................................................................................................111 Figure 123 Decorated Kimmeridge shale bracelet (16).....................................................................................................111 Figure 124 Shale working sites in the Isle of Purbeck...................................................................................................... 114 Figure 125 Early Neolithic pottery from Ditch 1332........................................................................................................ 120 Figure 126 Middle/Late Neolithic pottery from Ditch 1332............................................................................................ 121 Figure 127 Late Neolithic/Early Bronze Age pottery from Trenches 1 and 2, evaluation pits and posthole 1616.......... 122 Figure 128 Beaker pottery from other features................................................................................................................. 123 Figure 129 Middle Bronze Age pottery from clamp 1700 and unstratified contexts on Site 3........................................ 124 Figure 130 Late Bronze Age pottery fabrics by number................................................................................................... 125 Figure 131 Late Bronze Age pottery from House 1, posthole 1484................................................................................. 126 Figure 132 Late Bronze Age pottery from House 1, posthole 1596................................................................................. 127 Figure 133 Late Bronze Age pottery from postholes 1511, 1551 and 1557 associated with House 1............................. 128 Figure 134 Late Bronze Age pottery from pit 1750 and posthole 1761........................................................................... 129 Figure 135 Late Bronze Age pottery under cobble layer 1039......................................................................................... 130 Figure 136 Late Bronze Age/Early Iron Age pottery fabric groups.................................................................................. 132 Figure 137 Late Bronze Age/Early Iron Age pottery from Type 1 pit 335....................................................................... 134 Figure 138 Late Bronze Age/Early Iron Age pottery from Type 1 pit 326/390................................................................ 135 Figure 139 Late Bronze Age/Early Iron Age pottery from Type 1 pit 1412..................................................................... 135 Figure 140 Late Bronze Age/Early Iron Age pottery from Type 1 pits 1513 and 1302.................................................... 136 Figure 141 Late Bronze Age/Early Iron Age pottery from Type 2 pit 393....................................................................... 137 Figure 142 Late Bronze Age/Early Iron Age pottery from Type 2 pit 1126..................................................................... 138 Figure 143 Late Bronze Age/Early Iron Age pottery from Type 3 pits 1022, 1176/1196, 1215 and 1241....................... 140 Figure 144 Late Bronze Age/Early Iron Age pottery from Type 3 pits 1427 and 1807.................................................... 140 Figure 145 Late Bronze Age/Early Iron Age pottery from placed deposit 1169.............................................................. 141 Figure 146 Late Bronze Age/Early Iron Age pottery from placed deposits 1175, 1256 and limpet deposit 1500........... 142 Figure 147 Late Bronze Age/Early Iron Age pottery from postholes 1268, 1540 and 1547............................................ 143 Figure 148 Late Bronze Age/Early Iron Age pottery from midden layers 5 and 6, contexts 1560, 1108, 1226 and 1301... 145 Figure 149 Late Bronze Age/Early Iron Age pottery from midden layer 4, contexts 1170, 1193, 1294 and 1449.......... 146 Figure 150 Late Bronze Age/Early Iron Age pottery from midden layer 4, contexts 1460 and 1461.............................. 147 Figure 151 Late Bronze Age/Early Iron Age pottery from midden layer 4, contexts 1462, 1491, 1502 and 1521.......... 148 xii

List of Figures Figure 152 Late Bronze Age/Early Iron Age pottery from midden layer 3, contexts 1028, 1065, 1075, 1125 and 1225.... 149 Figure 153 Late Bronze Age/Early Iron Age pottery from midden layer 2, contexts 360 and 1017................................ 151 Figure 154 Late Bronze Age/Early Iron Age pottery from midden layer 2, contexts 1033, 1064, 1042 and 1067.......... 152 Figure 155 Late Bronze Age/Early Iron Age pottery from midden layer 2, contexts 1082 and 1089.............................. 153 Figure 156 Late Bronze Age/Early Iron Age pottery from midden layer 2, contexts 1211, 1280 and 1354.................... 154 Figure 157 Late Bronze Age/Early Iron Age pottery from midden layer 1, contexts 1016, 1027, 1032, 1035 and 1046.... 156 Figure 158 Late Bronze Age/Early Iron Age pottery from midden layer 1, contexts 1055 and 1061.............................. 157 Figure 159 Late Bronze Age/Early Iron Age pottery from midden layer 1, context 1100............................................... 157 Figure 160 Late Bronze Age/Early Iron Age pottery from midden layer 1, context 1104............................................... 158 Figure 161 Late Bronze Age/Early Iron Age pottery from midden layer 1, contexts 1108, 1112 and 1116..................... 159 Figure 162 Late Bronze Age/Early Iron Age pottery from midden layer 1, contexts 1120 and 1122.............................. 160 Figure 163 Late Bronze Age/Early Iron Age pottery from midden layer 1, context 1167............................................... 161 Figure 164 Late Bronze Age/Early Iron Age pottery of intrinsic interest from Site 1; contexts 302, 305, 307, 313 and 339............................................................................................................................................................... 162 Figure 165 Late Bronze Age/Early Iron Age pottery of intrinsic interest from Site 2; contexts 1013, 1096, 1098, 1133, 1422, 1553 and 1800............................................................................................................................................... 164 Figure 166 Distribution of Late Bronze Age/Early Iron Age pottery from Site 2, midden layer 1.................................. 166 Figure 167 Distribution of Late Bronze Age/Early Iron Age pottery from Site 2, midden layer 2.................................. 167 Figure 168 Distribution of Late Bronze Age/Early Iron Age pottery from Site 2, midden layer 3.................................. 168 Figure 169 Distribution of Late Bronze Age/Early Iron Age pottery from Site 2, midden layers 4, 5 and 6................... 169 Figure 170 Late Bronze Age/Early Iron Age jars and bowls – fabric groups................................................................... 172 Figure 171 Middle Iron Age pottery from pit 1182.......................................................................................................... 179 Figure 172 Middle Iron Age pottery from pits 1182, 326/390 and 393............................................................................ 180 Figure 173 Late Iron Age pottery from pit 304 (1-4)........................................................................................................ 183 Figure 174 Late Iron Age pottery from pit 304 (5-8)........................................................................................................ 184 Figure 175 Late Iron Age pottery from pit 304 (9-11)...................................................................................................... 185 Figure 176 Late Iron Age pottery from pit 304 (12-18).................................................................................................... 187 Figure 177 Late Iron Age pottery from pit 304 (19-25).................................................................................................... 188 Figure 178 Early/Middle Roman pottery (26-33) from occupation debris 315 on Site 1................................................. 190 Figure 179 Middle and Late Romano-British pottery (34-41).......................................................................................... 191 Figure 180 Late Bronze Age/Early Iron Age objects of copper alloy............................................................................... 193 Figure 181 X-radiograph of axe head (SF 117)................................................................................................................ 194 Figure 182 Iron Age and Roman objects of copper alloy (1-8)........................................................................................ 196 Figure 183 Roman objects of copper alloy (9-17)............................................................................................................ 197 Figure 184 Post-Roman copper alloy buckle (SF 471) from grave 1667......................................................................... 198 Figure 185 The mineral-preserved textile on the back of buckle SF 471......................................................................... 198 Figure 186 The weave structure of the textile shown in Figure 185................................................................................ 199 Figure 187 A part-mineralised fibre from the textile on buckle SF 471........................................................................... 199 Figure 188 Objects of lead................................................................................................................................................ 200 Figure 189 Objects of iron Late Bronze Age/Early Iron Age (1), Middle Iron Age (2), Late Iron Age (3-4) Romano-British (5-9)........................................................................................................................................................ 200 xiii

Multi-period Occupation at Football Field, Worth Matravers, Dorset Figure 190 Objects of iron Romano-British (10-16) and Post-Roman (17)..................................................................... 202 Figure 191 Fragment of Late Bronze Age/Early Iron Age glass ring SF 177 with drawn section above......................... 203 Figure 192 Compositional results for the Football Field glass ring (weight % calcium oxide and weight % alumina)...... 205 Figure 193 Compositional results for the Football Field glass ring (weight % potassium oxide and weight % of magnesia)...................................................................................................................................................................... 205 Figure 194 Plot of zirconium versus strontium concentrations of the Football Field glass with Levantine and Polish glasses.................................................................................................................................................................... 206 Figure 195 Objects of fired clay........................................................................................................................................ 208 Figure 196 17th century clay tobacco pipe with fine milling below the rim.................................................................... 211 Figure 197 Objects of worked bone (1-8)......................................................................................................................... 212 Figure 198 Objects of worked bone (9-11)....................................................................................................................... 213 Figure 199 Plot of the stable isotope results for the human bone samples....................................................................... 217 Figure 200 Results and structure of the primary Bayesian model for mortuary activity related to the Post-Roman cemetery............................................................................................................................................................................ 217 Figure 201 Results and structure of the alternative Bayesian model for mortuary activity related to the PostRoman cemetery................................................................................................................................................................ 218 Figure 202 Probability distributions for the span of mortuary activity related to the Post-Roman cemetery in the primary model and alternative model............................................................................................................................... 219 Figure 203 Element frequency by MNE, cattle; Late Bronze Age/Early Iron Age features............................................ 247 Figure 204 Element frequency by MNE, sheep/goat; Late Bronze Age/Early Iron Age features.................................... 248 Figure 205 Element frequency by MNE, cattle; Late Bronze Age/Early Iron Age midden............................................. 250 Figure 206 Sheep/goat toothwear by Payne Stage, Late Bronze Age/Early Iron Age midden......................................... 250 Figure 207 Element frequency by MNE, sheep/goat; Late Bronze Age/Early Iron Age midden..................................... 250 Figure 208 Element frequency by MNE, sheep/goat; Middle Iron Age pit 1182............................................................. 253 Figure 209 Element frequency by MNE, cattle; Late Iron Age........................................................................................ 256 Figure 210 Element frequency by MNE, sheep/goat; Late Iron Age pit 304................................................................... 256 Figure 211 Sheep/goat toothwear by Payne Stage; Late Iron Age................................................................................... 257 Figure 212 Element frequency by MNE, dog; Late Iron Age pit 304.............................................................................. 258 Figure 213 Element frequency by MNE, sheep/goat; Romano-British period................................................................. 260 Figure 214 Relative abundance of livestock species by period........................................................................................ 263 Figure 215 Site 2 chronological phases of dated excavated features................................................................................ 286 Figure 216 Neolithic to Middle Bronze Age landscape in Purbeck.................................................................................. 288 Figure 217 Late Bronze Age/Early Iron Age sites in Purbeck.......................................................................................... 291 Figure 218 Middle Iron Age sites in Purbeck................................................................................................................... 299 Figure 219 Late Iron Age sites in Purbeck........................................................................................................................ 303 Figure 220 Late Iron Age, Romano-British and Post-Roman settlement activity at Football Field................................. 305 Figure 221 Romano-British sites in Purbeck.................................................................................................................... 310 Figure 222 Post-Roman sites in Dorset and beyond together with other non-Purbeck sites mentioned in the discussion.................................................................................................................................................................... 313

xiv

List of Tables Table 1 Climatic phases of Southern England (based on Allen 2000; Allen and Scaife 2007)............................................ 6 Table 2 Radiocarbon determinations from Football Field.................................................................................................. 10 Table 3 Description of layers in Trenches 1 and 2.............................................................................................................. 14 Table 4 Midden slots with their excavated layers (in some areas Level 4 was excavated in two spits, a and b)............... 25 Table 5 Grave typology and associated skeletal material................................................................................................... 79 Table 6 Total flaked stone assemblage................................................................................................................................ 95 Table 7 Total flaked stone assemblage by phase and site................................................................................................... 96 Table 8 Late Bronze Age/Early Iron Age flaked stone assemblage.................................................................................. 100 Table 9 Prehistoric stone objects by phase....................................................................................................................... 105 Table 10 Quantification of stone types.............................................................................................................................. 105 Table 11 Number of shale objects by type and period...................................................................................................... 110 Table 12 Occurrence of early Prehistoric pottery............................................................................................................. 118 Table 13 Pottery fabrics through time............................................................................................................................... 119 Table 14 Occurrence of rim forms.................................................................................................................................... 119 Table 15 Neolithic pottery: occurrence of fabric types..................................................................................................... 120 Table 16 Late Neolithic/Early Bronze Age pottery: occurrence of fabric types............................................................... 122 Table 17 Middle Bronze Age pottery: occurrence of fabric types and minimum number of vessels............................... 124 Table 18 Late Bronze Age pottery: occurrence of fabric types and minimum number of vessels................................... 125 Table 19 Total numbers and weights of Late Bronze Age/Early Iron Age pottery sherds from all contexts.................... 131 Table 20 Numbers and weights of Late Bronze Age/Early Iron Age pottery sherds from the midden layers.................. 131 Table 21 Late Bronze Age/Early Iron Age pottery: occurrence of fabric types................................................................ 132 Table 22 Numbers of Late Bronze Age/Early Iron Age rim forms................................................................................... 132 Table 23 Late Bronze Age/Early Iron Age pottery from Type 1 pits 335, 326/392, 1302, 1412 and 1513...................... 133 Table 24 Late Bronze Age/Early Iron Age pottery from Type 2 pits 393, 1126, 1503 and 1574..................................... 137 Table 25 Late Bronze Age/Early Iron Age pottery from Type 3 pits and scoops............................................................. 139 Table 26 Late Bronze Age/Early Iron Age pottery from placed deposits......................................................................... 141 Table 27 Late Bronze Age/Early Iron Age pottery from postholes.................................................................................. 142 Table 28 Late Bronze Age/Early Iron Age pottery from midden layers 5 and 6.............................................................. 144 Table 29 Late Bronze Age/Early Iron Age pottery from midden layer 4.......................................................................... 145 Table 30 Late Bronze Age/Early Iron Age pottery from midden layer 3.......................................................................... 148 Table 31 Late Bronze Age/Early Iron Age pottery from midden layer 2.......................................................................... 150 Table 32 Late Bronze Age/Early Iron Age pottery from midden layer 1.......................................................................... 155 Table 33 Numbers and percentages of (illustrated) Late Bronze Age/Early Iron Age vessel types................................. 170 Table 34 Correlation of Late Bronze Age/Early Iron Age pottery typology..................................................................... 171 Table 35 Late Bronze Age/Early Iron Age pottery: comparison of vessel thickness........................................................ 172

xv

Multi-period Occupation at Football Field, Worth Matravers, Dorset Table 36 Late Bronze Age/Early Iron Age pottery: illustrated jar mouth diameters and number of vessels.................... 172 Table 37 Late Bronze Age/Early Iron Age pottery: illustrated bowl diameters and number of vessels........................... 172 Table 38 Middle Iron Age pottery fabrics, groups and number of sherds........................................................................ 178 Table 39 Estimated Vessel Equivalents of vessels from Late Iron Age pit 304................................................................ 182 Table 40 Estimated Vessel Equivalents for the Early Roman assemblage from occupation debris 315 overlaying paving 314 on Site 1 dating to c. AD 43-250.................................................................................................................... 189 Table 41 Estimated Vessel Equivalents for vessels from occupation deposits associated with the barn on Site 2 dating to c. AD 250-400.................................................................................................................................................... 190 Table 42 Late Bronze Age/Early Iron Age bronzes: XRF surface analysis (wt%)........................................................... 195 Table 43 Post-Roman human bone samples submitted for radiocarbon dating................................................................ 216 Table 44 Disarticulated adult human remains................................................................................................................... 222 Table 45 Perinatal remains from non-building contexts................................................................................................... 223 Table 46 Summary of perinatal and infant remains from the Romano-British building.................................................. 224 Table 47 Bone condition scores of skeletal material from the Site 3 cemetery and outlying skeleton 311 on Site 1...... 227 Table 48 Species abundance, NISP and MNI, all periods................................................................................................. 235 Table 49 Occurrence of bird NISP and MNI by phase..................................................................................................... 236 Table 50 Minimum Number of Elements, cattle, sheep/goat and pig for the Neolithic/Early Bronze Age, the Late Bronze Age and the Late Bronze Age/Early Iron Age features........................................................................................ 237 Table 51 Minimum Number of Elements, cattle, sheep/goat and pig for the Late Bronze Age/Early Iron Age midden, Middle Iron Age pit 1182 and Late Iron Age features........................................................................................ 238 Table 52 Minimum Number of Elements, cattle, sheep/goat and pig for the Late Iron Age pit 304 and Romano-British features................................................................................................................................................... 239 Table 53 Butchery............................................................................................................................................................. 240 Table 54 Animal bone; pathological information............................................................................................................. 241 Table 55 Tooth wear, all livestock species........................................................................................................................ 242 Table 56 Metrical information (mm) by period, species and element.............................................................................. 243 Table 57 Number of bone fragments by feature type....................................................................................................... 244 Table 58 Summary of gnawed, weathered and burnt bone fragments by period.............................................................. 244 Table 59 Species representation, NISP and MNI, by context; Late Bronze Age/Early Iron Age Type 1 pits 1513 and 326/390.............................................................................................................................................................. 245 Table 60 Species representation, NISP and MNI, by context; Late Bronze Age/Early Iron Age Type 1 pits 335 and 1412 and Type 2 pits 393, 1126 and 1503........................................................................................................... 246 Table 61 Species representation, NISP and MNI by context; Late Bronze Age/Early Iron Age Type 3 pits 354, 1142, 1176, 1215 and 1427....................................................................................................................................... 247 Table 62 Fusion in sheep/goat; Late Bronze Age/Early Iron Age features....................................................................... 248 Table 63: Species representation, NISP and MNI by context; Middle Iron Age pit 1182................................................ 252 Table 64 Species representation, NISP and MNI by context, Late Iron Age pits 304, 316, 324, 334, 342, 406, 1408 and upper fill of Late Bronze Age/Early Iron Age pit 1412..................................................................................... 254 Table 65 Species representation, NISP and MNI by context; Late Iron Age pits 1131, 1532, 1740, 1749 and 1209...... 254 Table 66 Species representation, NISP and MNI by context; Late Iron Age House, Site 1............................................. 255 Table 67 Species representation, NISP and MNI; Late Iron Age contexts 1184, 1106, 1160, 1521, 1095, 307 and 1019............................................................................................................................................................................ 255 Table 68 Species representation, NISP and MNI by context, Romano-British period..................................................... 259

xvi

List of Tables Table 69 Charred plant remains from selected features (all periods)............................................................................... 270 Table 70 Charcoal from Neolithic and Early Bronze Age features.................................................................................. 277 Table 71 Charcoal from Late Bronze Age/Early Iron Age pit contexts............................................................................ 278 Table 72 Charcoal from other Late Bronze Age/Early Iron Age contexts........................................................................ 279 Table 73 Charcoal from Late Iron Age and Late Iron Age/Romano-British contexts...................................................... 279 Table 74 Charcoal from Romano-British contexts........................................................................................................... 280 Table 75 Minimum numbers of hand-collected land snails from dated contexts............................................................. 281 Table 76 Land molluscs from flotation samples............................................................................................................... 282 Table 77 Minimum number of recorded marine molluscs................................................................................................ 282 Table 78 Distribution of fish bones across phases............................................................................................................ 284 Table 79 Late Bronze Age/Early Iron Age sites in Purbeck............................................................................................. 293 Table 80 Middle Iron Age sites in Purbeck....................................................................................................................... 300 Table 81 Late Iron Age sites in Purbeck........................................................................................................................... 304 Table 82 Comparison of the Worth Matravers ‘barns’ with other similar Purbeck buildings.......................................... 308 Table 83 Comparison of Romano-British sites in Purbeck............................................................................................... 309

xvii

Summary Prior to development of part of Football Field for a small social housing scheme, volunteers from the East Dorset Antiquarian Society excavated areas which were within the footprint of the proposed building project. Postexcavation work was funded by donations from the Worth Community Housing Trust, Viridor, collections from the local community and a grant from the Dorset Archaeology Committee. The results of the excavations have provided an opportunity to examine in detail the development of a small but highly complex landscape on the north-western edge of the Purbeck village of Worth Matravers. The earliest feature was part of a deep and wide enclosure dating to the Early Neolithic period (c. 3500 BC): the infilling was associated with groups of flintwork and pottery which were deposited over the following two thousand years. An enigmatic, undated double-row of small upright stones is likely to be early prehistoric in origin. There was ephemeral evidence for activity during the Middle Bronze Age in the form of a base for a pottery clamp together with a scatter of associated pottery. In the absence of supporting radiocarbon dates, the pottery suggests a late second millennium time frame for these activities. A single post-built house and pit was dated to the Late Bronze Age (c. 800 BC); these features would have been associated with agricultural activities. Agriculture continued to be a significant aspect for the next thousand years. During the Late Bronze Age/Early Iron Age (800-600 BC), a large proportion of the site was deliberately covered by a layer of small limestone pieces forming a pavement. A single post-built house was constructed at this time, and the site subsequently became the focus for depositional activities. Deliberate deposits of cultural material were laid within an accumulating mass of discarded rubbish, ultimately forming a midden and pits, both large and small were cut into its surface. Small-scale pottery making and the manufacture of shale objects were also carried out on

the site. Importation of continental copper alloy and glass items reflects access to cross-channel trading links and exotic goods. Middle Iron Age activity was represented by a single pit dated by metalwork to c. 400-200 BC. By the Late Iron Age (c. 150 BC – 50 AD), two stone-footed roundhouses and associated pits were established. Deliberate deposits of very large numbers of pottery vessels and animal bones in a deep, stone-lined pit and the continuing apparently ad hoc disposal of domestic rubbish suggest that specific discard policies were in place. A single crouched inhumation in typical local Durotrigian burial style was placed within the community’s area of activity. Coinage and imported pottery are indicators of small-scale trade and exchange. The pottery sequence of local and imported wares suggests that the site continued to be inhabited during the early Roman period. However, it was not until the 4th century AD that a further building sequence took place, with the construction of a stone-built, rectangular-shaped barn which was used as a shale workshop. It may have had an earlier agricultural function. A minimum of 20 infants were buried under the floor of this building with one child interred in a square limestone cist. Continental finewares were replaced by British equivalents during the mid-3rd century. Activity on the site continued into the early 5th century. The site appears to have been abandoned by the 6th century, when a small inhumation cemetery was laid out to the east of what had been the focus of prehistoric and later activity. Twenty six individuals were laid out in rows, many of the graves utilising Roman building material. Radiocarbon dating has placed the cemetery in the 6th-8th centuries AD. Comparison of the features and structures with those of other selected sites in Purbeck and beyond, has not only highlighted the site’s significance but also allowed the placement of its chronological phasing within a broad agricultural, economic and social framework.

xviii

Résumé L’activité du Second âge du Fer est représentée par une seule fosse, datée par la typologie du travail du métal à environ 400-200 ans avant J.-C. Vers la fin de l’âge du fer (environ 150 ans av. J.-C. à 50 ans après J.-C.) furent construites deux maisons rondes de pieds de pierre, et des fosses associées. Les dépôts intentionnels de très grand nombres de poteries et d’os d’animaux dans une fosse profonde, bordée de pierres, ainsi que la mise au rebut continue et apparremment ad hoc de débris domestiques, suggèrent que des règlements spécifiques liés à la gestion des déchets étaient mis en place. Une seule inhumation dans le style funéraire Durotrigien typiquement local a été placée dans la zone d’activité de la communauté. Une autre inhumation comparable, trouvé non loin de la maison ronde vers le nord, a été photographié et laissé sur place. La monnaie et la poterie importée sont des indicateurs de commerce et d’échange à petite échelle.

En amont des travaux de réaménagement d’une partie du champ Football Field en un petit lotissement de logements sociaux, des bénévoles de la Société d’Archéologie de l’est du Dorset (EDAS) ont fouillé la zone couverte par ce projet de construction. La post-fouille a été financée par le mécénat du Worth Community Housing Trust, de Viridor, par des dons de la communauté locale et par une subvention du Comité archéologique du Dorset. Les résultats des fouilles ont permis l’étude approfondie d’un paysage certes petit, mais très complexe, situé en bordure nord-ouest du village de Worth Matravers dans le Purbeck. La structure la plus ancienne faisait partie d’un large et profond enclos datant du Néolithique ancien (vers 3500 ans avant J.-C.). Son remplissage était associé avec des groupements de silex et de céramique déposés pendant environ les 2000 ans suivants.

La typologie des céramiques, qu’elles soient de fabrication locale ou importées, suggère que le site continua à être occupé au début de la période romaine. Toutefois, ce n’est qu’au IVe siècle après J.-C. qu’une nouvelle tranche de bâtiment avait lieu avec la construction d’une grange rectangulaire en pierre qui servit d’atelier de fabrication d’objets en schiste. Cet atelier a pu antérieurement être à usage agricole. Sous le plancher, au moins vingt enfants en bas âge furent enterrés, dont l’un dans une ciste carrée en calcaire, peut-être à la fin de l’utilisation de ce bâtiment. Au milieu du 3ème siècle la céramique fine d’importation continentale fut remplacée par des equivalents britan­niques. L’activité sur ce site se poursuivit jusqu’au début du 5ème siècle.

Une mystérieuse double rangée de petites pierres dressées et non datées est probablement d’une origine préhistorique ancienne. Il y avait des preuves éphémères d’activité pendant l’âge du Bronze moyen en forme d’une base d’un four de cuisson de céramiques et quelques tessons associés. En l’absence de datation au radiocarbone à l’appui, la céramique semble indiquer que ces activités datent de la fin du deuxième millénaire. Une seule habitation construite avec des poteaux et une fosse dataient du Bronze final (environ 1000 à 800 ans avant J.-C.); ces structures auraient été associées aux activités agricoles. Pendant le millénaire qui suivit, l’agriculture continua d’être une activité importante.

Le site semble avoir été abandonné à partir du 6ème siècle. Un petit cimetière fut alors aménagé à l’est de ce qui avait été l’aire principale des activités durant la préhistoire et pendant les périodes qui suivirent. Vingt-six individus furent inhumés dans des tombes disposées en rangées. Nombre d’entre-elles sont construites à partir de matériaux de construction romains. La datation au radiocarbone indique que le cimetière fut utilisé entre le sixième et le huitième siècle de notre ère, le style correspondant en effet à la tradition funéraire de l’ouest britannique.

Pendant l’âge de Bronze final et le Premier âge du Fer (800-600 ans av. J.-C.), une grande partie du site fut délibérément recouverte par une épaisse couche de petites pierres de calcaire qui formèrent ainsi une chaussée. Une seule habitation fut construite avec des poteaux à cette époque. Le site devint ensuite un centre de dépôt: des fosses, et grandes et petites, furent creusées, et on y posa des dépôts intentionnels d’objets culturels parmi une masse croissante de déchets, formant au final un fumier d’habitat. Une production à petite échelle de céramiques et la fabrication d’objets en schiste furent également entreprises sur le site. L’importation d’objets en alliage de cuivre et en verre d’origine continentale témoigne des liens commerciaux outre-Manche et de l’accès à des produits exotiques.

La comparaison entre les structures et l’architecture trouvées sur le site avec celles d’une sélection d’autres sites de la région du Purbeck et au-delà non seulement a mis en évidence l’importance particulière de ce site mais également a permis de le resituer chronologiquement dans un large cadre agricole, économique et social.

xix

Zusammenfassung Im Hinblick auf geplante Bauarbeiten einer Anlage für soziales Wohnen auf einem ehemaligen Fußballfeld unternahm eine Gruppe von Laien, Mitglieder von EDAS (East Dorset Antiquarian Society), Ausgrabungen des betreffenden Areals. Unkosten für die Auswertung der Funde wurden vom Worth Community Housing Trust, Viridor, Spenden von Gemeindemitgliedern und einem Zuschuß vom Dorset Archaeology Commitee getragen.

Hinweise auf Aktivität in der mittleren Eisenzeit wurden nur in einer Grube mit Metallwaren ans dem Zeitraum 400-200 v Chr gefunden. Gegen Ende der späten Eisenzeit (etwa 150 v.Chr.-50 A.D.) wurden zwei Rundhäuser mit Steinfundamenten gebaut und daneben befanden sich auch Gruben. Eine große Menge von absichtlich deponierten keramischen Gefäßen und Tierknochen in einer tiefen steingemauerten Grube, vermischt wie zuvor mit scheinbar ad hoc Beseitigung von Haushalt Müll, deutet auf gezielte Entsorgungs-Bräuche. Eine einzige Hockerbestattung im typischen Stil des durotrigischen Stammes befand sich inmitten des Bereiches der Siedlung. Eine andere ähnliche Bestattung, nicht weit vom nordlichen Rundhaus entfernt, wurde fotografiert und am Ort gelassen. Münzen und eingeführte Keramik bezeugen einigen Handels-und Tauschverkehr.

Die Ergebnisse der Ausgrabungen ermöglichten es, die Entwicklung einer begrenzten aber äußerst komplexen Landschaft am nordwestlichen Ende des Purbeck Dorfes Worth Matravers ausführlich zu untersuchen. Die älteste Fundstelle bildete einen Teil einer breiten und tiefen Einzäunung aus der Frühsteinzeit (etwa 3500 v. Chr.). Die Verfüllung enthielt Fundgruppen von Feuersteinwerkzeug und Keramik die sich während der folgenden 2000 Jahre gebildet hatten.

Stylistische Analyse von einheimischen und eingeführten keramischen Gefäßen deutet an, daß die Fundstelle während der frühen Römerzeit weiterhin bewohnt war. Aber, erst im 4. Jahrhundert n. Chr. gab es wieder Bautätigkeit, indem eine steinerne Scheune errichtet wurde, die als Werkstatt benutzt wurde. Es ist möglich, daß sie zuerst der Landwirtschaft diente. Mindestens zwanzig kleine Kinder waren unter dem Fußboden dieses Gebäudes beerdigt. Ein Kind lag in einem Sarkopharg aus Kalkstein: möglicherweise geschah dies nachdem das Gebäude jetzt keinen weiteren Zweck hatte. Feinwaren vom Kontinent wurden ab Mitte des dritten Jahrhunderts durch britische Waren gleicher Art verdrängt. Die Nutzung dieser Fundstelle setzte bis ins frühe 5. Jahrhundert fort.

Eine schwer zu deutende Doppelreihe von undatierten kleinen aufrecht stehenden Steinen schien auch aus der frühen Vorgeschichte zu stammen. Ein ephemes Anzeichen für Aktivität im mittleren Bronzezeitalter bestand aus der Basis eines Feldbrennofens mit Streugut von entsprechenden Scherben. Obwohl dafür keine 14C-Daten vorliegen, bezeugt die Keramik einen Zeitpunkt für diese Tätigkeiten gegen Ende des zweiten Jahrtausend vor Christi. Ein einzelnes Pfostenhaus und eine Grube stammen aus der späten Bronzezeit (etwa 1000-800 v. Chr.). Diese Funde sind offenbar mit Landwirtschaft verbunden. In den nächsten 1000 Jahren blieb Ackerbau ein wichtiges Element.

Vor Beginn des 6. Jahrhunderts scheint dieser Ort schon verlassen zu sein. Um diese Zeit wurde östlich vom Areal früherer Aktivitäten ein kleiner Friedhof angelegt. Sechs und zwanzig Personen wurden in Reihen bestattet. In vielen Gräbern hatte man römisches Baumaterial benutzt. 14C-Datierungen ergeben Bestattung vom 6. bis 8. Jahrhundert, welches dem Stil nach zur westbritischen Begräbnis Tradition paßt.

Während der späten Bronze- bzw. früheren Eisenzeit (800600 v, Chr.) wurde ein großer Teil des Areals absichtlich mit einer dicken Schicht von kleinen Kalksteinen bedeckt, die eine Art Pflaster bildeten. Ein einzelnes Pfostenhaus wurde zu dieser Zeit errichtet. Von da an war die Lage ein Bereich für Deposition. Kleine und größere Gruben wurden ausgehoben, um kulturelle Gegenstände absichtlich zu deponieren, umgeben von angesammeltem Müll, was zur Bildung von einem Kehricht-haufen führte. Einige Herstellung von Keramik und Schieferobjekten am Ort fand auch statt. Kontakt und Handel jenseits des Ärmelkanals ist durch exotische Gegenstände aus Bronze und Glas bezeugt.

Ein Vergleich der Funde und Bauten von dieser Ausgrabung mit gewissen anderen in Purbeck und Umgebung hat nicht nur ihre eigene Bedeutung bezeugt, sondern auch ihre Stellung inmitten eines breiteren agronomischen, wirtschaftlichen und sozialkulturellen Rahmens ermöglicht.

xx

1 Environment and Project Background This report details the results of five seasons of archaeological work at Football Field, Worth Matravers, which took place between 2006 and 2011, in advance of a community housing project. Fig. 1 is the key for all archaeological sections. Location The site is located in south-east Dorset on high land overlooking the English Channel to the south, the Purbeck ridge to the north and the modern town of Swanage to the east (Fig. 2). The site, previously Quarry Field, is now known as Football Field (centred on SY 975778) and is located on the northern edge of the village of Worth Matravers, immediately south of Compact Farm (Figs 3 and 4). The field is on a slightly sloping spur of land between 138m and 142m above Ordnance Datum (OD) (Fig. 5). On the western side the land falls sharply towards a dry coombe. The site has clear views to the English Channel and westwards along the coast as far as Portland. Landscape, Geology, Environment and Climate Change Andrew Morgan and Robert Kenyon

Figure 1 Key to section and plan drawings

Introduction

Chapman’s Pool. To the north, the land drops away along the edge of the plateau into the broad undulating Corfe Valley which stretches from Worbarrow Bay to Swanage Bay. This is bounded to the north by the Purbeck Hills, which form a chalk spine that extends across Purbeck from Arish Mell in the west to Handfast Point in the east. The Purbeck Hills reach a maximum height of 199m and form a natural barrier, accessed through openings at Corfe Gap and Ulwell Gap. Further north, the land falls away steeply through woodlands into the extensive low-lying landscape of the Purbeck Heathlands, which stretch from the Purbeck Hills to the fertile flood plain of the River Frome and the wetlands surrounding Poole Harbour.

Football Field, Worth Matravers, is located in the southern part of the area known as the Isle of Purbeck and lies on the edge of the Purbeck Plateau, where the ground falls away gently towards the valley of Hill Bottom which leads to the natural harbour of Chapman’s Pool. Purbeck is a peninsula dominated by the sea and has a distinct geographical identity that separates it from the rest of Dorset. It is bordered by the English Channel to the south and east, by Poole Harbour and the River Frome to the north. Its western limit is defined by the Luckford Lake stream, a small tributary of the Frome which continues along a small valley that cuts through the chalk ridge into the bay at Arish Mell (Legg 1989, 1-2).

Geology The geology of the area comprises a succession of strata from the Late Jurassic period to the current Quaternary period, constituting part of the Jurassic Coast which has been designated a World Heritage Site and provides a wide range of natural resources that have been exploited for millennia.

Topography Purbeck can be divided into five topographical areas (Fig. 6), each clearly associated with the underlying geology. The limestone Purbeck Plateau forms the southern boundary of Purbeck, ending in dramatic cliffs along the coast. To the south west is the Vale of Kimmeridge, a coastal area affected by landslips and enclosed by the escarpment of the Purbeck Plateau. Here the coast is formed by steep cliffs that are cut by the bays of Kimmeridge and

In the west, the earliest rocks are the extensive Kimmeridge Clay Formation, exposed along the unstable cliffs from Brandy Bay to St Aldhelm’s Head and comprise a series 1

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 2 Worth Matravers, Dorset, site location and geology

of thin-bedded, organic-rich, grey-black shales and clays interlaced with layers of non-organic mudstone and limestone. An important stratum in this formation is the Blackstone, a bituminous oil-shale up to a metre thick, and

is the source of Kimmeridge Shale which has been worked into objects since the Neolithic period. It can also be used as a fuel, but when burnt it gives off a pungent smell; it has been known to combust spontaneously. 2

Environment and Project Background Limestone beds are prominent along the cliff tops and continue along the edge of the coast to Durlston Head and provide some of Dorset’s most important building stone, which has been exploited for 2000 years. The thin-bedded limestone and shale rocks known as the Cinder Bed (Wood 2011, 72), and form the Purbeck Plateau, where Football Field is located, and yield materials that have been exploited for dry stone walls, roof tiles and paving slabs. During the Holocene epoch (9700 BC to present), the sea level has risen substantially, the coast has been eroded and the rivers have laid down channels of rich alluvial deposits which are agriculturally important when compared with the surrounding poor-quality heathlands. During this period the chalk ridge that connected the Purbeck Hills to the Isle of Wight was breached, and Poole Harbour formed from a submerged valley. There are a number of small natural harbours along the coast, including Worbarrow Bay, Kimmeridge Bay, Chapman’s Pool, Swanage Bay and Studland Bay. Environmental change Our understanding of the environmental history of the Isle of Purbeck is limited and no environmental studies have been undertaken on sites south of the Purbeck Ridge. Two projects have examined material from the clay and

Figure 3 Location of Football Field and EDAS excavation areas in relation to Worth Matravers and Compact Farm

Figure 4 Aerial view of Worth Matravers from the north. The excavation area and new housing development are within the red ellipse. The modern road is lower right and the Purbeck coast is visible at the top of the photo

3

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 5 Contour survey and archaeological interventions 1991-2011

gravels of the southern fringes of Poole Harbour: Wytch Farm, (Allen and Scaife 1991) and Bestwall Quarry, Wareham (Scaife 2009) where a range of highly valuable data for understanding the environmental development of the western harbour basin was recorded. The area is very different from South Purbeck in underlying geology, topography and hydrology. However, the results provide a useful backdrop against which to consider the situation to the south of the Purbeck Ridge, with respect to the type and degree of woodland cover, access to resources, the likelihood and date of episodes of clearance and the possible arable and animal economy of the site at Football Field.

for human settlement and population levels. Table 1 summarises the broad climatic phases which affected the environment and influenced human settlement in Purbeck in the Neolithic to Post-Roman periods. Project Background Planning background Football Field was bought by Signpost Housing Association Limited in 2002 and was the subject of two planning applications for housing development in 2003. Due to potential high archaeological costs, no development took place and the field was bought by Mr Robert Kenyon in 2007. In the same year, Performance Management Services International (PMSI) bought 7 ½ acres on the south side of the field; of this, one acre was earmarked for community housing. The remaining 6½ acres, in which there were known and excavated archaeological remains, were retained by Mr Kenyon. The extreme south west corner of the field was acquired by the Worth Community Property Trust with the intention of building five low-cost

Climate change Earth’s climate has changed continuously over time, driven mainly by natural events but has also been affected by human activities. Minor changes can have a significant impact by affecting the land available for cultivation, the length of the growing season and crop yield and the productivity of livestock. Changes determine the variety and abundance of natural resources and the opportunities 4

Environment and Project Background

Figure 6 Main topographical features in the Isle of Purbeck

A desk-based assessment for a proposed housing development was followed by an archaeological evaluation prior to any determination of a planning application (Oakley 2003). This work comprised eight 10m by 1.9m trenches, excavated by Wessex Archaeology in 2003 and revealed ‘substantial high-quality archaeological deposits’ in six of the eight trenches (Wessex Archaeology 2003a).

houses: this development was the subject of two further planning applications in 2008 and 2010. Planning consent was given in 2011 with relevant archaeological conditions under Planning and Policy Guidelines (PPG15). Building work began in July 2011 and was completed by April 2012. Site evaluation and previous work The archaeological potential of the site was first recognised by the Royal Commission on Historical Monuments (England) during their survey work of south-east Dorset in the 1960s and 1970s (RCHM 1970, 621).

In 2006, East Dorset Antiquarian Society (EDAS) carried out further evaluation of the proposed housing development footprint by magnetometer survey and a single evaluation trench, the results of which implied that there were no significant underlying archaeological remains (Roberts 2009, 211). This work was followed in 2007 by a magnetometer survey of much of the remainder of the field, targeting specifically two of the proposed garden plots. On the basis of geophysical anomalies, five 10m by 2m trenches were machine-stripped and in the garden areas, a further two 3m by 2m trenches were handdug (Fig. 5). One of the trenches, which had been sited close to the Southampton University excavation, revealed a substantial Late Iron Age storage pit with a nearby inhumation (Roberts 2009, 212).

Southampton University undertook fieldwork from 1990 to 1993 opening up an area approximately 16m by 25m in the south west sector of the field (Fig. 5). This produced settlement evidence in the form of structures and finds from at least the Early Iron Age up to the very latest Roman period. In addition, a 5m by 5m test pit and twenty four 1m2 square pits were dug at 20m intervals (Graham et al 2002, fig. 1.2). Rubble surfaces with potential archaeology were encountered in eight of these pits which were sited immediately to north, south and east of the main excavation. 5

Multi-period Occupation at Football Field, Worth Matravers, Dorset Table 1 Climatic phases of Southern England (based on Allen 2000; Allen and Scaife 2007) Approx. date calibrated BC

Archaeological period

Climatic zone

Pollen Zone

Climate and vegetation

1050 BC-Present

Early medieval Roman Period Late Iron Age Middle Iron Age Late Bronze Age

Sub-Atlantic

VIII

Deterioration Cold and wet, general deterioration. High rainfall. Decline of lime. Increase of ash, birch and beech.

3200-1050 BC

Middle Bronze Age Early Bronze Age Late Neolithic

Sub-Boreal

VIIb

Stable Warm and dry, low rainfall, wind-blown deposits. Woodland regeneration in Southern England.

4000-3200 BC

Middle Neolithic Early Neolithic

6300-4000 BC

Later Mesolithic

Atlantic

Vlla

Optimum Holocene Climatic Optimum, warm and wet. Increase of 2°C, poly-climax forest. Increase in elder. Some clearances.

8900-6300 BC

Mesolithic

Boreal

VI V

Ameliorating Continental climate. Warm and dry. Asynchronous expansions of mixed oak forest with hazel and succession from pine.

10,000-8900 BC

Early Mesolithic

Pre-Boreal

lV

Rapid amelioration Sharp increase in warmth at 8,000BC. Birch, juniper and pine woodland.

Pre 10,000 BC

Upper Palaeolithic

Arctic

III

Tundra

Declining warmth. First agriculture. Gradual woodland clearance. Loss of elm.

Archaeological organisation

Archaeological background

The most recent fieldwork (2008-2011) was also conducted by the East Dorset Antiquarian Society (EDAS), the site being evaluated and excavated by society members. The 2006-8 work was directed by Phil Roberts and subsequently by Lilian Ladle (Ladle 2012b). Mark Corney has advised on fieldwork and post-excavation processes since 2010. An £8000 grant from Comma, the Community Aggregates Fund, secured initial post-excavation specialist input for the pre-2010 finds and the Worth Community Housing Trust allocated £21,000 for specialist work associated with the housing development. The Dorset Archaeological Committee gave a grant of £7722 for radiocarbon dating of selected contexts.

In the late 1950s and early 1960s during fieldwalking exercises, J.P. Calkin and A.P. Brown discovered a potential Iron Age site south of Compact Farm (Brown, 1955, 77; Calkin 1960, 86). Five years later, R.A.H Farrar noted further material of Late Iron Age and Roman date in the same area (Farrar, 1965, 118). The site was located and recorded by the Royal Commission on Historical Monuments (England) in the 1960s (RCHME 1970, 621). Between 1990 and 1993, Southampton University undertook fieldwalking and excavation after a Roman grain dryer was located during ploughing. Excavation confirmed a period of occupation from at least the Early Iron Age to the latter part of the 4th century if not into the early part of the 5th century AD. The results were published in a University of Southampton monograph Purbeck Papers (Graham et al 2002).

The work of the project was submitted for the 2011 Dorset Archaeology Award and received a runner-up certificate. Since 2010, the results of the fieldwork have been widely disseminated through lectures to archaeology groups, amenity societies and conferences. In addition there have been articles in magazines and Interim reports in the Proceedings of the Dorset Natural History and Archaeological Society and in the CBA (Council for British Archaeology) newsletter. Three well-attended open days were organised during the excavations and a full-colour brochure was produced for the general public detailing key aspects of the site’s development. A video film was also created to promote the site to a wider audience via the internet.

Evaluation by Wessex Archaeology in 2003 confirmed the presence of stratified archaeological deposits within some of the proposed building plots on the western half of the site and within the easement of the access road. Archaeological material recovered from this work suggested Early Iron Age activity with little Late Iron Age or Roman material. In 2006, East Dorset Antiquarian Society (EDAS) were approached to conduct further evaluation of the site 6

Environment and Project Background (Trench 3) was extended when archaeological features were encountered. As this trench was in close proximity to the Southampton excavations, a decision was taken to examine the area in more detail in an attempt to clarify the relationship between the two sites. An area measuring approximately 14m by 18m was investigated. This work was undertaken as an EDAS training exercise.

and following a magnetometer survey of the proposed housing area, a single trench was opened; this proved to be archaeologically sterile (East Dorset Antiquarian Society 2007). The landowner requested further work to be undertaken within the proposed gardens of two of the dwellings as well as outside the development area. An additional magnetometer survey was conducted over most of the field and this produced anomalies indicating potential archaeological features. As a result, five 10m by 2m trenches were opened over the field and two trenches (3m by 2m) were opened in the proposed garden areas of the houses. Four of the test trenches produced archaeological sequences and one in particular (Trench 3), which was located directly east of the Southampton University excavation, yielded significant Late Iron Age features and finds including a deep storage pit and a human inhumation.

Between May and October 2010, in response to potential destruction of archaeological deposits by construction of the proposed site access road, an area approximately 7m by 10m was opened inside the field gateway. In order to access the archaeological levels, between 0.2m to 0.4m of overburden was removed by a small backacting mechanical excavator with a flat bucket, under archaeological supervision. The exposed subsoil was set out into forty eight individually numbered slots measuring approximately 1m by 2m (Fig. 24). These were handcleaned; a very difficult task due to the presence of very large amounts of limestone rubble. These slots were then taken down in 0.1m and 0.05m spits until a compacted surface of small limestone pieces was encountered. From February until May 2011, work took place on an extension of the site comprising a further 3m west to the edge of the current highway and 2m north (to investigate features encountered in 2010). As this was a much smaller area, only two subdivisions were made, the westernmost being termed the Access Road (AR) and the northernmost, the Stone Alignment (SA). Once again, the exposed levels were all cleared by hand and taken down in 0.1m to 0.2m spits.

The following year, Trench 3 was extended to determine any potential relationships between the storage pit and burial and those features recorded by Southampton University. An area measuring approximately 18m by 13.5m was subsequently excavated and revealed archaeological features dating from the Early Iron Age to the end of the Roman period (Roberts 2009, 211-213). Additional archaeological observation and recording took place in December 2010 when four geological test pits (all approximately 2m by 0.75m and 2m deep) were dug in the housing development area. No archaeological features were recorded, but evidence of quarry infilling with material from nearby Weston Farm was noted (RCHME 1970, 621).

Building work commenced in July 2011 with observation and recording of all groundworks. In addition, the complete footprint and part of the garden of one of the houses was excavated; the area under investigation measuring 16m by 16m. Further monitoring of groundworks took place in February and March 2012. The insertion of a water pipe in the field opposite the site was recorded in November 2011. The EDAS interventions have been divided into four subsites for this report (Fig. 5).

Prior to building work commencinging, EDAS were invited back in 2010 and 2011 to conduct work on the development access road, next to the field gateway. A rectangular area immediately inside the field entrance was investigated between May and October 2010. A further extension comprising an L-shaped area next to the main road was investigated between February and May 2011. When building work started in July 2011, unexpected archaeological deposits in the form of human skeletal remains were encountered at the extreme eastern end of the development area. In addition, an opportunity arose during 2011 to record archaeological sequences in a water trench in land immediately west of the site. In summary, the various interventions have exposed complex stratigraphic sequences dating from the Early Neolithic to the Post-Roman period.

• Site 1 – 2006-7 evaluation trenches and 2008 excavation area • Site 2 – 2010/2011 excavations on the access road • Site 3 – excavation associated with the house footprint • Site 4 – monitoring of a water pipe in the field to the west of the development A total of 235 discrete features dating from the Early Neolithic to the Post-Roman period were excavated. All features were hand excavated and the majority were 100% sampled; however, a number of features on Site 1 were 50% sampled and some not sampled at all. There were few linear features, most comprising shallow gullies which were 100% sampled. The large Neolithic ditch was initially sampled by evaluation pits in 2010 and excavated by sectioning in 2011. It is probable that some features were missed in the 2006 excavation due to the fact that the much of the rubble cover on the site was not cleared.

Archaeological methodology Football Field has been the subject of three major excavations (located on Fig. 5), discussed above; each was organised in a different way and with different objectives. The initial single EDAS test trench was dug by hand as were a further eight test trenches; one of these 7

Multi-period Occupation at Football Field, Worth Matravers, Dorset A unique series of context numbers was allocated and relationships between contexts were noted together with relevant stratigraphic information. Features were recorded by section at a scale of 1:10. Longitudinal sections of the excavated sides of Site 2 were produced at 1:20. Site plans were produced at scales of 1:25, 1:50 and 1:100. Features were recorded and planned using standardised recording forms derived from the AC Archaeology recording system. A comprehensive photographic record of digital images was maintained. Finds were washed (where appropriate), marked and catalogued by material type. An appraisal of material from each year involved the identification, cataloguing and spot dating of all finds. Unfortunately a comprehensive environmental strategy was only achieved for the 2010/11 work when in order to gain maximum information, a number of features were specifically targeted which were likely to produce material that would give information on the contemporary environment and also for potential radiocarbon dating. Both bulk and hand-retrieved samples were taken.

Neolithic/Early Bronze Age (c. 3500-1500 BC)

Themes and aims

Latest Bronze Age and Late Bronze Age/Early Iron Age (c. 950/800-400 BC)

A probable Early Neolithic enclosure ditch was located on the western side of the site on Site 2 and is dated by pottery and flintwork from its infilling which took place over several millennia. A radiocarbon date of 2023-1828 cal BC was obtained for an infill layer. The cultural material represents activity in a landscape which was increasingly being opened up for agriculture and husbandry. A miniature stone alignment was undated and may belong to the Early Bronze Age. Pottery and flint scatters imply nearby activity. Middle Bronze Age (c. 1550-950 BC) Settlement evidence is lacking but the period is represented by remnants of a probable pottery clamp and an associated pottery scatter on Site 3 together with a general spread of flintwork over all areas examined.

When the first EDAS project was initially undertaken, there was neither a project design nor arrangements for post-excavation analysis and publication. This author undertook responsibility for the project in 2009 and developed an excavation strategy and an outline proposal for post-excavation work and ultimate publication with major themes addressing the following criteria:

A single pit on Site 4 produced Late Bronze Age pottery in the plainware tradition.

Calibrate the sequence through absolute radiocarbon dating where possible.

A post-built roundhouse was located on the edge of Site 1 and pre-dated a deliberately laid surface of limestone rubble which covered much of the site. A further postbuilt roundhouse was probably constructed soon after the rubble was laid down. Over a period of time this surface was covered with a midden deposit of sticky loam which incorporated very large amounts of cultural material comprising pottery, animal and human bone, shale and metalwork. Pits and scoops were dug into the accumulating debris, many penetrating the rubble surface. A single pit produced evidence for in-situ pottery making. Mixed agriculture was practised nearby.

Settlement

Middle Iron Age (c. 400-150 BC)

Examine changes in the settlement status and function through time.

A single large pit was located on Site 2; two brooches, pottery, animal bone and charred grain hint at specific intentional depositional infilling events.

Chronology Place the sequence of structures, features and finds within the known framework of activities around Purbeck, Poole Harbour and Dorset as well as within Southern England.

Society

Late Iron Age (c. 150 BC- AD 50)

Examine cultural associations, social units, human remains and evidence for non-domestic activities.

Stone-footed roundhouses and associated pits were located on Sites 1 and 2. A single crouched female inhumation was recorded on Site 1. Two coins and a range of metalwork together with locally-produced and continental pottery were recorded on Sites 1 and 2.

Environment and economy Examine evidence for environmental change, agricultural production, food processing and storage, craft activities and trade and exchange.

Roman (c. AD 50-410) A rectangular stone-built barn was constructed on Site 2 around the early 300s AD and was predominantly associated with shale working. A minimum of 20 infants were interred within the confines of the building. The presence of coins, brooches, tweezers, glass and pottery

Summary of Phasing (Site 2 was the only area where there were intercutting features and where site stratigraphy could be plotted. This is illustrated on Fig. 215) 8

Environment and Project Background confirms rural status with access to some high-value objects.

charred grain and animal and human bone. Typically, the samples most suitable to provide the most reliable dates did not necessarily coincide with the stratigraphical units or finds which required dating. Consequently, a degree of pragmatism had to be exercised in the selection of materials for dating. Samples were selected on availability, suitability and condition. Single samples were selected from the infill of the Neolithic ditch, the Middle Iron Age pit, two Late Iron Age pits, the Roman barn and the isolated human burial.

Post-Roman (c. AD 410-650) A cemetery was set out on Site 3 where 26 east-west orientated burials were located. In addition to the single burials there were three double and one triple inhumation. The only grave good was a copper alloy buckle; a stone anchor was part of the grave furniture of a double grave. A number of graves utilised Roman stone roofing tiles. A single burial was located on Site 1.

Five samples associated with the Late Bronze Age/Early Iron Age midden deposit were selected with the objective of dating the duration of activity associated with this feature. These included a sample from a posthole from a post-built roundhouse which pre-dated and underlay the construction of the limestone rubble hard standing which made up the base of the midden deposit. Two samples related to the accumulation of the midden deposit itself and two samples were selected from pits which cut into the midden. One of the pits contained a fragment of continental glass and the other illustrated a complex infill sequence containing a range of ceramic material.

Radiocarbon Dating Introduction In total, 17 radiocarbon samples from Football Field were processed and dated at the Scottish Universities Environmental Research Centre at East Kilbride. The samples comprised two carbonised plant macrofossils, seven of animal bone and eight of human bone. The dating process is expensive: with limited financial resources, the project had to be realistic in choosing material for submission. A grant from the Dorset Archaeological Committee for the complete set of dates is gratefully acknowledged.

Six samples were selected from the Post-Roman cemetery in order to date the time frame of the burials and to attempt to define the spatial development of the site. Samples were chosen to reflect the linearity of the cemetery layout, the type of grave in which the individuals were buried and in two cases to date objects within the graves (a copper alloy buckle and a stone anchor). The human bones from the cemetery site were in poor condition, probably due to post-depositional soil conditions, which had to be taken into account in the selection of samples.

Objectives of the dating programme The scientific dating programme was designed to address the following objectives: • Date the infill of a potential Early Neolithic ditch • Date a post-built roundhouse underlying the Late Bronze Age/Early Iron Age cobble layer. • Date the accumulation of midden material on top of the cobble layer. • Date pits cut into the midden and cobble layer. • Refine the dating of a single Middle Iron Age pit together with a Late Iron Age pit which cut into it. • Define the dating of a large Late Iron Age pit containing significant artefactual deposits. • Date the stone-founded Roman barn. • Date an isolated adult human burial on Site 1. • Date Post-Roman cemetery activity on Site 3.

Results and calibration The results are tabulated in Table 2 and are conventional radiocarbon ages (Stuiver and Polach 1977) which are cited according to Trondheim convention standards (Stuiver and Kra 1986). The 14C ages are quoted in conventional years BP (before 1950 AD). The error which is expressed at the one sigma level of confidence includes components from the counting statistics on the samples, modern reference standard and blank and then random machine error. The calibrated age ranges are determined from the University of Oxford Radiocarbon Accelerator Unit calibration programme (OxCal4). The calibration plots were calculated using methods derived from Bronk Ramsey et al (2013) and Riemer et al (2013). All certificates and calibration curves are held in archive. Bayesian modelling was undertaken for the Post-Roman material and is discussed by Krus in Chapter 12.

Sampling strategy A range of material was available for radiocarbon dating but financial constraints meant that a limited number of samples could be submitted. These included charcoal, a

9

Table 2 Radiocarbon determinations from Football Field Laboratory Code

Sample

carbon

nitrogen C/N Ratio (molar)

Radio carbon Age (BP)

Calibrated date range (95%) confidence

3575±27

2023-1878 cal BC

Neolithic Monument on Site 2 SUERC-61231 (GU37944)

FFWM12 1630, Prunus spinosa from deep lowest fill of enclosure ditch 1332

-25.6‰

1839-1828 cal BC

Late Bronze Age/Early Iron Age midden activities SUERC-61168 (GU37940)

FFWM10 1017, sheep/goat tibia from accumulating midden deposit

-21.6 ‰ 4.8 ‰

3.2

2475±32

771-476 cal BC

SUERC-61169 (GU 37941)

FFWM10 1301, cattle phalanx from lowest level of midden deposit

-22.3‰

5.8 ‰

3.3

2506±32

790-538 cal BC

SUERC-61170 (GU37942)

FFWM1 1592, sheep/goat metatarsal from fill of posthole 1592 in roundhouse

-21.4 ‰ 5.2 ‰

3.2

2482±30

775-485 cal BC

SUERC-61163 (GU37938)

FFWM10 1141, sheep/goat tibia from fill of pit 1142 containing glass fragment

-20.7 ‰ 5.6 ‰

3.2

2432±32

651-683 cal BC

464-453 cal BC 445-431 cal BC

669-637 cal BC 623-616 cal BC 591-405 cal BC

SUERC-61167 (GU37939)

FFWM11 1525, sheep/goat radius from fill of pit 1513

-21.5 ‰ 5.2 ‰

3.2

2336±29

498-363 cal BC

2115±31

342-326 cal BC

Middle to Late Iron Age activity SUERC-61171 (GU37943)

FFWM 1151, Bromus secalinus, from fill of pit 1182

-22.2‰

SUERC-61162 (GU37937)

FFWM10 1134, cow scapula from fill of pit 1131 which was cut into pit 1182

-22.2 ‰ 6.5 ‰

3.3

2106±31

203-45 cal BC

SUERC-61161 (GU37936)

FFWM07 305, dog mandible from fill of pit with very large deposits of animal bone & pottery

-19.6‰

3.2

2093±31

197-43 cal BC

-19.2 ‰ 10.3 ‰

3.2

1747±31

226-386 cal AD

-19.8‰

8.9 ‰

3.2

1385±31

633-667 cal AD

205-48 cal BC

8.1‰

Roman occupation of stone-built barn SUERC-61173 (GU37945)

FFWM11 1440, infant burial under the threshold of the Roman barn

Isolated burial on Site 1 SUERC-61172 (GU37944)

FFWM08, human remains 311 from an isolated inhumation on Site 1

Post-Roman cemetery on Site 4 SUERC-61178 (GU 37947)

FFWM11 1632, human remains from grave 1633 which contained a stone anchor

-19.3 ‰ 8.5 ‰

3.2

1393±31

625-663 cal AD

SUERC-61179 (GU37948)

FFWM11 1642, human remains from grave 1640 where the individual’s legs were bound

-19.6 ‰ 7.8 ‰

3.2

1396±31

599-671 cal AD

SUERC-61180 (GU37949)

FFWM11 1680, human remains from grave 1678 which contained a double burial

-19.5‰

8.3 ‰

3.3

1413±31

583-655 cal AD

SUERC-61181 (GU37950)

FFWM11 1699, human remains from grave 1697 which contained a single burial

-19.7 ‰ 9.0 ‰

3.2

1547±29

424-575 cal AD

SUERC-61182 (GU37951)

FFWM 1687, human remains from a grave with three burials

-19.6 ‰ 9.2 ‰

3.2

1425±31

575-660 cal AD

SUERC-61177 (GU379460

FFWM11 1996, human remains from a grave which contained a buckle

-20.0 ‰ 8.8 ‰

3.3

1340±31

644-716 cal AD 743-766 cal AD

10

2 The Neolithic, Early, Middle and Late Bronze Age Periods Neolithic Enclosure Ditch

3 details the concordance of the fills of the trenches. Figure 10 illustrates the features visible within the ditch at a depth of 0.7m.

Part of a possible enclosure ditch was located in the entranceway to the field on Site 2. It was initially encountered during a test-pitting exercise in 2010 to determine whether any archaeological features survived beneath the packed limestone cobbles which covered the entire area (see Chapter 3). It also provided an opportunity to investigate the underlying geology. Eleven pits were excavated, varying in size from 1m by 1m to 2m by 1m (Fig. 7). The initial test pit 1252 measured 1m by 1m and was excavated to a depth of 0.7m (Fig. 8). Three fills were encountered, all comprising mid-brown loam which became progressively stickier towards the base of the pit. The upper fill 1253, contained occasional fragments of limestone; the middle fill 1282, was characterised by chunks of limestone, some of which had been burnt, rare charcoal flecks and raw red clay fragments. The base layer 1300 was similar, but the limestone pieces tended to be either horizontal or vertical in the soil matrix. Twenty one pieces of worked flint of Early Neolithic character were recovered.

Trench 1 (Figs 11 and 12) The south-western edge of the trench was defined by a deposit of tabular limestone (1612), mostly lying horizontally, but some were vertically or randomly placed and set into a mid-brown clay loam, the deepest layers of this were yellow clay. These stones were interpreted as bedrock which had been cut into in a stepped fashion, forming the southern side of the ditch. The northern side of the ditch was outside the excavation area. Seven layers of fill were recorded and were generally mid- to dark brown loam with a high percentage of clay. The uppermost layer 1594, lay directly underneath the cleared cobbles (1039); a single piece of pottery implies a Late Bronze Age land surface. The fills of 1602, 1606 and 1609, were similar varying from 0.1m to 0.40m in depth, and contained small pieces of limestone, rare charcoal, occasional crushed limestone and burnt clay fragments. Pottery was present but the sherds were very small. Under 1609, layers 1610 to the south and 1632 to the north merged into each other. Layer1610, at 0.4m to 0.5m in depth, comprised a darker loam with fragments of decomposing limestone which were denser where the layer abutted 1632. It was only apparent after investigating Trench 2 that 1632 was a deposit of small, consolidated, compacted limestone pieces embedded in a dark brown clay-loam. The deposit had a maximum depth of 0.6m and appeared to have developed over a very long period of time.

Further pits were cut with a view to understanding the nature of the soil build-up and how the flintwork had become incorporated into this. As work progressed, it was apparent that the fills were not part of a general soil horizon but were the infill of a very large feature. Evaluation pit 1270 was extended south and north (forming pits 1305 and 1308, Fig. 9): the profile of the excavated section suggested that a large ditch-type feature (designated 1332) crossed the western side of the site. The fills of these pits were similar to the initial evaluation pit, as was the nature of the flintwork; in addition 67 small sherds of pottery were retrieved including a rim sherd (Fig. 125.1) from an Early Neolithic plain bowl. Due to the size constraints of the site in 2010, it was not possible to investigate this feature further. However in 2011 when the site was extended westwards, there was an opportunity to examine a larger area. Figure 215 illustrates its stratigraphic position on Site 2.

Layer 1620 was a mid- to dark brown loam with rare charcoal flecks. Larger limestone pieces up to 0.2m long appear to have been either placed or dumped near the top and bottom of this layer. Finds included an Early Neolithic leaf-shaped arrowhead (Fig. 116.1) and Early Bronze Age and Neolithic pottery. The lowest fill 1629, was a yellowbrown, very sticky clay-loam with small limestone pieces and occasional tabular blocks. Towards the base the fill became sandier and levelled out at maximum depth of 1.3m. Three fragments of pottery of probable Neolithic date together with 14 flint flakes and a scraper (Fig. 116.2) were retrieved from the fill.

An ‘L’ shaped section was set out as two abutting trenches (Trench 1 and Trench 2, Fig. 7). The longer trench 1603, was aligned in a south-west, north-east direction and 1604, was aligned north-west to south-east at right angles to this. Trench 1 was 4.5m long and 2.5m wide. Trench 2 had a maximum length of 3.15m and was 1.5m wide. For the first 0.5m of depth, the trenches were excavated in 0.1m spits, and then numbers were allocated as the fills changed character. The maximum depth of the feature was 1.35m. Bedrock was encountered on the south-west side and on the base, but once again the limited size of the site precluded excavation of the full width of the feature. Table

Three features were cut into ditch layer 1620 at the extreme southern side and all were apparent at a depth of about 0.7m. Oval pit 1613 was partly defined by small limestone slabs on the south-eastern side of the ditch and measured 0.5m by 0.3m with a depth of 0.2m; the dark 11

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 7 Site 2; Neolithic enclosure 1332 and its relationship with pits 1182 and 1131. Location of evaluation pits, Trenches 1 and 2 and Early Bronze Age postholes

12

The Neolithic, Early, Middle and Late Bronze Age Periods brown clay-loam fill 1614 contained charcoal flecks and bright red decomposing clay together with a tiny sherd of pottery and three flint flakes. This small pit was 1.5m long with a width of 0.6m and a depth of 0.18m and was adjacent to deposit 1623. The fill consisted of blocks and slabs of limestone set into clay and possibly placed in the ditch edge; there was no dating material. Feature 1625 was a pit or posthole on the south-western side of the ditch and measured 0.6m long by 0.5m wide and was 0.45m deep. It was cut into the base of 1620 and also cut 1629. The fill (1626) of mid-brown clay loam with charcoal flecks contained some limestone blocks as well as decomposing limestone; two flint flakes were recorded. Trench 2 (Fig. 13) The uppermost layer 1560 lay directly underneath the cleared cobbles 1039; 52 sherds of Latest Bronze Age pottery together with flintwork and animal bone suggest nearby activity. Fills 1605, 1607 and 1608 were similar

Figure 8 Section of evaluation pit 1252

Figure 9 Section of evaluation pits 1308, 1270 and 1305

Figure 10 Plan of Trenches 1 and 2 showing features visible at an excavated depth of 0.7m

13

Multi-period Occupation at Football Field, Worth Matravers, Dorset Table 3 Description of layers in Trenches 1 and 2 Trench 1

Trench 2

Layer description

Colour

1594

1560

Blocky clay loam under cobbles Frequent small limestone pieces

Mid-brown

1602

1605

High % clay, limestone pieces, rare charcoal

Ginger/brown

1606

1607

(As Above)

1609

1608

High % clay, limestone pieces, rare charcoal

Ginger/brown,

Oblique arrowhead

1610

1611

Clay loam with crushed limestone Clay fragments and charcoal

Mid-dark brown

Y-shaped tool Fig 116.5

1620

1621

Mottled loam, decomposing limestone, limestone fragments,, some burnt, clay and charcoal

mottled

Leaf Arrowhead Fig. 116.1

1620

1622

clay loam, limestone, decomposing red clay

Dark to mid-brown

1629

1627

Sticky loam with limestone and charcoal

Mid-light brown

1630

Silty loam, occasional limestone pieces

Dark brown

1631

silt

Grey/black

to 1602 and 1606 described above, and the 15 sherds of pottery were all in the earliest All Cannings Cross tradition. Between 0.3m and 0.4m the clay-loam layer 1611 contained a high percentage of degraded limestone; charcoal flecks and burnt clay fragments were scattered throughout. The finds included eight pottery sherds of probable Early Bronze Age date and five flint flakes. A scrap of shale was also retrieved. Below this, layer 1621 varied in depth from 0.1m to 0.17m and comprised hard, compacted, light brown loam containing tabular limestone pieces, most aligned horizontally. There were moderate quantities of charcoal, burnt clay and burnt limestone. Three sherds of pottery and 16 flint flakes were of Late Neolithic/ Early Bronze Age date. Below this, the dark brown 0.1m deep layer 1622 contained small limestone pieces and charcoal flecks. Among the nine sherds of pottery were two sherds of Peterborough and Fengate-type wares (Fig.

Diagnostic Flintwork

Scraper Fig. 116.2

126.1 and 5) and a sherd of twisted-cord decorated Beaker (Fig. 127.4). Six flint flakes were recorded. The depths of layer 1627 varied from 0.32m at the western edge of the trench to 0.52m at the east. It comprised a mid-brown sticky loam containing a deposit of limestone pieces at the top of the layer that appear to have been placed horizontally. Among the 15 pieces of worked flint was a single chert item. One sherd of pottery was recorded. Abutting 1627 was deposit 1632 which had an abrupt almost vertical side on the north. This deposit of compacted small limestone pieces embedded in a very dark brown clay loam may have been a deliberate effort to consolidate the edge of the ditch. This deposit was also recorded in Trench 1. Two silty layers had formed on the base of Trench 2; they were absent in Trench 1. Layer 1630 was mottled brown in colour, 0.08m to 0.04m thick and contained occasional small limestone pieces. Two small flint flakes were retrieved. The primary

Figure 11 Excavated section Trench 1

14

The Neolithic, Early, Middle and Late Bronze Age Periods

Figure 12 Excavated section Trench 1, view from the north-west

fill 1631 was dark grey merging into black silt and varied from 0.08m to 0.03m thick. This layer had accumulated onto the natural limestone bedrock. A fragment of Prunus spinosa was dated to 2023-1878 cal BC or 1839-1828 cal BC (98% confidence; GU38093) and suggests that the ditch was filling by the Late Neolithic /Early Bronze Age period.

Three features were contained within the ditch fill in Trench 2; two postholes were apparent in layer 1608 and a placed deposit in layer 1627 (Fig. 14). Posthole 1616 was 0.35m in diameter and 0.4m deep with almost vertical sides and a flat base. Its fill 1615a, comprised a dense, black charcoal ash probably representing fuel waste. Stones around the sides

Figure 13 Excavated section Trench 2

15

Multi-period Occupation at Football Field, Worth Matravers, Dorset boundary of Football Field. Cultural material from the ditch filling must have derived from activity within its immediate vicinity. The Stone Alignment (Figs 15 and 16) Two rows of vertically placed, upright stones (1145 and 1219) on Site 2 were located during 2010 and 2011 and were initially visible as stone tips protruding through the hard standing 1039. The stones were aligned southeast to north-west and were approximately 1.2m apart at the southern end and 1.5m apart where the alignment continued under the northern baulk of the excavation. The feature was recorded as four separate elements (1145, 1219, 1486 and 1477) and investigated over two seasons; it appeared to continue outside the excavation area. Alignment 1145/1486 Figure 14 Sections of postholes 1616, 1618 and deposit 1628 in Trench 1 and scoop 1323 in evaluation pit 1320

The excavated extent of this eastern row of stones had a total length of 3.8m and comprised a series of limestone blocks designated 1145 and 1486. These measured 1.85m and 0.8m in length, with a small gap of 0.2m between the two. A further gap of 0.9m separated further uprights which were recorded in the northern baulk of the excavation (1617). A bedding trench was discernible here; it measured 0.3m wide and had a total depth of 0.58m (Fig. 15, section). The fill was a dark brown humic/clay loam which was cut by a later posthole 1466. The stones were placed two together, with little or no gaps between; were roughly rectangular with one pointed edge which was always placed directly in the ground. Generally, no bedding trench was evident; however there must have been such a feature. This must have been very narrow and cut to take two stones placed side by side. The northern extension 1486 was probably disturbed by later activity, resulting in the gap between the two sections.

of the feature displayed evidence of burning. The pit base 1615b was a mid-brown clay loam containing charcoal and fragments of burnt limestone. Four sherds of pottery included a stab-decorated wall sherd from an Early Bronze Age Beaker (Fig. 127.6) and a single flint flake. The fill had accumulated to at least a depth of 0.95m when the posthole was inserted. A second posthole 1618 was sited 1.5m away and also cut into layer 1608. This feature was 0.28m in diameter and 0.25m deep with near-vertical sides and a gently rounded base. The fill 1619, was mid-brown loam containing burnt clay fragments and one flint flake. At a lower level within layer 1622 an amorphous deposit of material 1628, measured approximately 0.7m in diameter and 0.15m deep and comprised dark clay loam with small limestone fragments and quantities of charcoal. Within this were fragmented and decomposing red and black burnt clay which could have originated from a pottery firing episode.

Alignment 1219/1477 The excavated western length of the alignment measured 4.55m in total, comprising a 2.85m length (1219), a gap of 0.6m and a further 1.1m (1447) measured to the site’s northern edge. In this instance, excavation revealed that there were never any uprights in the gap between the two sections. The longer length of the alignment 1219 consisted of two stones placed side by side and two deep; the lower stones were also placed in the ground with pointed edges lowermost. On average the stones measured 0.25m wide and 0.3m long, and were about 0.12m thick.

Small rounded scoop 1323, with limestone fragments placed around its edge, was located in the middle fill (1330) of evaluation pit 1320; it measured 0.28m in diameter and was 0.06m deep. The dark grey humic fill 1324, contained fragments of charcoal (Fig. 14). In conclusion, the ditch-type feature was probably constructed in the early-mid Neolithic period. It was a flat-bottomed and on the southern side, was roughly hewn from the underlying bedrock. It is possible that this section of the ditch terminated in the vicinity of Middle Iron Age pit 1182 (see Fig. 7), the cutting of which had destroyed evidence for its presence. The ditch was not apparent on the eastern part of Site 2 and due to the limits of the excavation area it was not possible to define the northern side of the feature. Over a period of about 2000 years, the ditch filled in and was re-cut; some of these episodes were discrete and short-term and others happened more slowly. It is likely that the feature extends around the western

Conclusion The alignment of this structure is undated and its function unknown, but the two double rows of deliberately set stones which predate the hard standing, suggest an early prehistoric date for its construction. The stones appear to delineate a pathway or walkway, and may be part of a larger monument which is possibly preserved to the north of the excavation area. 16

The Neolithic, Early, Middle and Late Bronze Age Periods

Figure 15 Plan of stone alignments 1219/1477 and 1145/1486 and section C-D on the northern edge of the excavation area

Other Early Bronze Age Features (Fig. 17) Four postholes 1257, 1295, and 1339 and one scoop 1340 were likely to be of Early Bronze Age, and were dated by the pottery from their fillings also by the fact that they were only visible once the cobbled surface had been removed (Fig. 7.) They were located south-east of the stone alignment and were all cut into a natural, decomposing limestone (the geological Cinder Bed layer). It is not possible to determine their function or use. Posthole 1257 This oval posthole was 0.3m long by 0.2m wide and 0.2m deep. The sides were vertical, the base was slightly rounded. The fill comprised a deposit of small, verticallyplaced limestone slabs (1254) which were embedded in a

Figure 16 Stone alignment, sections 1219 and 1145 during excavation, view to the north

17

Multi-period Occupation at Football Field, Worth Matravers, Dorset loam with a high percentage of clay. Charcoal chunks and burnt red clay were recorded together with three sherds of pottery, which included a wall sherd displaying a pinched cordon decorated with fingerprint impressions (Fig. 129.1). The feature extended underneath the undated (but at least Post-Roman) pavement (1695) to the south and the Post-Roman grave (1685) to the north-east. This feature may have been the remnants of a pottery clamp dating to the Middle Bronze Age. Very small sherds of similarly dated pottery (Fig. 129.2-6) were collected from the nearby stripped surfaces and some of the grave fills hint at Middle Bronze Age activity in the vicinity. Late Bronze Age Activity Pit and posthole Site 4

Figure 17 Sections of Early Bronze Age postholes 1257, 1295, 1339 and scoop 1340

A 45m length of narrow trench (1747) was dug for a water pipe in the field to the west of the site (for location see Figs 4 and 90). The trench was approximately 0.6m wide and 0.8m deep; a shorter 10m length (1748) was dug at right angles to this. Trench 1747 cut through a substantial pit 1750 and posthole 1761 (Fig. 18) which were dated by pottery to the Late Bronze Age. The full dimensions of the pit could not be determined, but it was at least 1.3m wide, 0.35m deep with steeply sloping sides and a flat base. The deposits were overlain by 0.25m of topsoil consisting of a mid-brown loam (1754). Three fills were recorded. The upper fill 1763, was a soft, crumbly, greyish-brown clay-loam containing decomposing limestone and poorlyfired pottery fragments. The middle fill 1764, had similar components but was charcoal-rich and mid- to dark brown in colour. The lowest fill 1765, was a pale grey clay-loam which also contained fragmented limestone. Pottery (34 sherds) was retrieved from the two upper fills (Fig. 134.1).

dark brown loam. Charcoal was noted near the base of the feature together with a small flint flake, a rim sherd and a sherd of toothed-comb decorated pottery. Posthole 1295 This sub-circular posthole was 0.22m long, 0.2m wide with a depth of 0.18m. The sides were vertical and the base slightly rounded. The fill was a mid- to light-brown loam and contained moderate amounts of limestone pieces with a larger block on the north side of the feature. This may have been a wedging stone for a timber upright. Three flint flakes were found near the base, together with a large piece of cattle metatarsal (see Randall, Chapter 14) which had been vertically placed in the posthole. Posthole 1339 This square posthole was lined with small thin limestone slabs and measured 0.15m across. It was 0.19m deep and had vertical sides and a flat base. The fill 1328 was a midbrown loam containing two pieces of animal bone and three of worked flint.

The pit was cut by substantial posthole 1761 which measured 0.48m wide and 0.32m deep having a vertical side to the south and a gently sloping side to the north. The fill 1766, comprised yellow clay with three limestone slabs, which had probably been packing stones. Two sherds of pottery, very similar to that from the pit, were recorded (Fig. 134.2).

Scoop 1340 This shallow, oval feature was 0.93m long by 0.52m wide and 0.08m deep with an undulating base. The fill 1341, of mid-brown loam had high clay content; there were frequent fragments of limestone and flecks of charcoal. The six undiagnostic wall sherds of pottery were in three fabric types which did not occur in the later periods. The feature was cut by a small vertically-sided stakehole 1342. Middle Bronze Age Activity Spread 1700, Site 3 (for location see Fig. 94) An amorphous spread (1700) was located on the southern edge of Site 3, its edges were not defined due to time constraints. The feature was at least 3m long by 2m wide and had a depth of 0.1m. It comprised dark brown sticky

Figure 18 Section of Late Bronze Age pit 1754 and associated posthole 1761 (Site 4)

18

The Neolithic, Early, Middle and Late Bronze Age Periods House 1, Site 2 (Figs 19 and 20, posthole sections Figs 21 and 22)

brown; otherwise the fills of the remainder comprised dark brown loam. Every posthole contained finds which included pottery, worked flint, worked stone, pebbles and animal bone. The two potential doorway postholes had particularly rich deposits of pottery; posthole 1484 yielded 65 sherds of mostly fresh, unabraded pottery weighing 2671g (Fig. 131). Posthole 1596 contained a lesser amount of pottery (16 sherds weighing 1006g).

The structural remains of a post-built roundhouse were located on the extreme western edge of Site 2 (Fig. 19). The house was on the 137m OD contour, on land which overlooked a dry valley to the west and the English Channel to the south, and was constructed in part, over the in-filled Neolithic ditch (1332). It was sealed by and pre-dated the hard standing (1039). A radiocarbon determination from a sheep/goat metacarpal from posthole 1592 gave a broad date of 775-485 cal BC (GU37942 at 95% probability). The structure therefore is likely to have been built at the very end of the Late Bronze Age, and is confirmed by the pottery assemblage.

Of the associated postholes (Fig. 22), four (1584, 1586, 1595 and 1564) were inside the building and two (1573 and 1580) were outside. The former may have been part of an internal structure; the latter, additional supports for the roof. These postholes were smaller than those forming the structural uprights, with widths varying from 0.19m to 0.3m and depths of 0.1m to 0.2m. One posthole was oval (1595) and the remainder were circular; all had moderately sloping sides and pointed or rounded bases. Three of the postholes (1584, 1564 and 1580) contained deposits of small limestone slabs which were wedged into the fill. The other postholes 1586, 1595 and 1573 were characterised by their dark, ashy/charcoal fills and the absence of large blocky limestone. Within the fill of posthole 1586 were fragments of burnt limestone. Debris from burning episodes may have become incorporated into the latter three features. In addition a shallow feature 1585 which measured 0.4m long by 0.3m wide and was 0.03m deep consisted of a deposit of charcoal in clay loam sitting on a clay base. It was interpreted as the base of a fire; but if so, was very close to the structural posthole 1574.

The surviving structural elements are illustrated on Fig. 20 and comprise a semi-circle of eight postholes with an external diameter of 6.7m, two internal postholes and a further six smaller postholes which were probably associated with the building. There was no external eaves-drip gully. The modern highway covers the western half of the structure. Postholes 1574, 1561, 1511, 1577, 1596, 1484, 1592 and 1599 are likely to have been part of an outer ring of posts supporting the roof, and it is possible that postholes 1596 and 1484 were part of an east-facing doorway. Postholes 1590 and 1551 may have been additional internal supports. Three postholes were circular (1484, 1590 and 1554); the remainder were oval in shape, with steeply sloping or almost vertical sides and flat or gently rounded bases. The postholes varied in width from 0.7m to 0.25m, and in depth from 0.5m to 0.22m, and were characterised by deposits of limestone slabs, most placed vertically, which may have been used as packing (Fig. 21). The fill of posthole 1577 was mid-

Two shallow, narrow gullies 1576/1058 and 1589 lay immediately east of the house and were sealed by 1039. As they were only 0.5m away from the posthole ring, it is unlikely that they were directly associated with the house:

Figure 19 Late Bronze Age House 1, Site 2. The posthole ring is marked by white posts; the modern road is centre with the English Channel coast beyond. View to the west

19

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 20 Late Bronze Age House 1, plan

Figure 21 Late Bronze Age House 1, sections of structural postholes

Figure 22 Late Bronze Age House 1, sections of ancillary postholes.

20

The Neolithic, Early, Middle and Late Bronze Age Periods Conclusions

there was likely to have been at least a one metre gap between the posts and the outer wall of the house. The gullies were 0.5m apart, measured between 0.17 and 0.18m wide and were only 0.05m deep, with rounded profiles Gully 1058/1576 was 2.66m long, and 1589 was at least 1.8m long. The fills were similar with black ashy loam and charcoal lumps. Part of section 1058 was lined with small limestone slabs. The function and date of use of these features is unclear.

Activity on the site began in the Early Neolithic period when a large enclosure ditch was constructed, and artefactual material derived from activities nearby was incorporated into the accumulating ditch fills over many centuries. The undated stone alignment is likely to have had a period of use which fell within the Early Bronze Age, and nearby postholes could have been associated with this. The ephemeral Middle Bronze Age possible pottery clamp was the only feature from this period. Later activity may have obliterated other features. A single Late Bronze Age pit and associated posthole could have been contemporary with the post-built roundhouse which was used and abandoned at the very end of the Late Bronze Age.

Archaeological evidence suggests that the house was dismantled and the structural components removed when the building went out of use; there was no evidence for burnt or rotting wood in the posthole fills. The large amounts of pottery from postholes 1484 and 1596 may have formed part of a deliberate closing deposit when the house was dismantled.

21

3 The Late Bronze Age/Early Iron Age Transition Introduction

as it accumulated. The soil matrix itself contained pottery, animal bone, shale fragments and occasional bronze items as well as charcoal and charred seeds. This soil was subsequently identified as a midden deposit; as such, it had cultural affinities with similar sites, in Wiltshire, at Potterne (Lawson 2000) , East Chisenbury (McOmish et al 2010) and All Cannings Cross (Cunnington 1923); in Berkshire at Runnymede (Needham and Spence 1996) and in the West Midlands at Whitchurch (Waddington and Sharples 2011).

In 2010, during the early part of the excavation, it became apparent that relatively large quantities of Late Bronze Age/Early Iron Age pottery were present, together with hand-worked shale, animal bone and a few metal artefacts (including a complete bronze axe). Although the depth of soil was only about 0.3m, the resemblance to other such deposits suggested that the accumulation was a result of middening activities. The Cobbled Surface

The compacted surface was recorded across most of the excavation area and comprised a tightly packed layer of small pieces of limestone measuring up to 0.1m by 0.1m which appeared to have been deliberately laid to a depth of approximately 0.1m on a mid- to dark brown clay-loam. In places the cobbles were placed over natural yellow clay. Cultural material in the form of pottery, worked flint and animal bone was randomly distributed on the surface, which was designated 1039 for most of its area on Site 1, but at the extreme western end was numbered 1558, 1559 and 1560.

During the Southampton University excavations of 199193, Early Iron Age deposits were encountered (Graham et al 2002, 6-9), including a compacted rubble layer which formed a ‘continuous surface flush with the clay subsoil along the western…. and eastern sides of the excavation’ (ibid. 9). This surface was also recorded in the EDAS 2008 excavations, but at the time was not recognised for what it was. In 2010, it became apparent that compacted limestone rubble (1039) extended across the whole of the excavated area (Fig. 23); this cobble layer was also recorded in 2011.

The cobble layer appears to extend over the south-west part of Football Field. The surface was recorded all over Site 2, and on parts of Site 1 and Site 4. However, the surface was absent on Site 3. The feature was at least 80m wide and

In antiquity, over a period of several centuries, soil built up over this surface; pits, postholes, scoops and gullies were cut into it, and deposits of material were placed in it

Figure 23 Late Bronze Age/Early Iron Age cobbled surface 1039 (Site 2). View to the west

22

The Late Bronze Age/Early Iron Age Transition (see O’Connor, Chapter 11), had slightly more ambiguous dates giving four readings, 751-683 cal BC; 669-637 cal BC; 623-616 cal BC and 591-405 cal BC (GU37938). The sample from the base of the midden (1301) is likely to be the most secure; the second sample (1217) derived from midden soil which may have been affected by later activities. The dates however are all within the postulated time frame for activity on the site.

at least as long, and appeared to continue into the gardens south of the site (see Fig. 5). Originally it must have been constructed to form a hard surface with a specific function. The underlying soil is clay-rich, and when wet, is extremely heavy and sticky (as the excavators found to their cost!). This surface sealed all earlier prehistoric activity. The Midden Deposit

The Features

Approximately 0.3m of soil accumulated on top of the cobbles and comprised a consistent dark brown loam with a high clay content, limestone fragments and patches of charcoal. Very large amounts of pottery, animal bone and shale were scattered throughout. In addition a small amount of metalwork which included a complete socketted axe and other items (see O’Connor, Chapter 11) were incorporated into the soil matrix. The realisation that a midden was present came early in the 2010 work on Site 2. A strategy was devised to maximise the archaeological potential by dividing the stripped area (which measured approximately 15.8m by 8.5m) into 47 slots (Fig. 24) which were then excavated in spits. These varied in depth for two reasons; first the initial stripping of the site by machine had not created a flat surface, and second, due to the varying amounts of limestone debris, it was not possible to excavate each slot in precise 0.1m spits. As a result, some layers were deeper or shallower than others. The slots and spits are tabulated (Table 4). Deliberately placed deposits of material consisting of varying amounts of pottery, bone, shale and limpets were located within the soil build-up. A series of pits, scoops, postholes and gullies were dug into the midden as it accumulated. In later centuries it was disturbed by Later Iron Age and Roman features, particularly at the western side of the site, and as a result it was particularly difficult to separate and date the animal bone.

A variety of features were cut into or deposited in the midden as it accumulated. Some of these penetrated the cobbled floor. They included pits, scoops and postholes, placed deposits of cultural material and arrangements of small packed vertical stones. A number of features cut through the pavement and included the large Type 1 pits and the smaller Type 2 pits. All the features containing packed limestone, and six of the posthole-type features also cut the pavement. Late in the post-excavation work, it was realised that a number of postholes may have related to a house. This is discussed below. Form and Type of Pits A total of 27 pits were of three types and were located on Sites 1 and 2. They may have been dug for a number of uses and varied in shape from circular to oval. All pits were dated by the pottery from the infilling sequence. • Type 1: large, circular pits with vertical sides, with diameters varying from 1.5m to 2.4m and depths varying from 0.9m to 1.45m, sometimes stone-lined and cut into the natural bedrock (335, 326/390, 1412, 1513, 1302). • Type 2: varying shapes (circular, oval and rectangular), moderately deep with steep sides and widths varying from 0. 5m to 1.22m and depths varying from 0.3m to 0.75m (393, 1126, 1136, 1503, 1574). • Type 3: shallow, mostly oval pits and scoops; the lengths varied from 0.4m to 0.9m and the widths from 0.35m to 0.72m. The surviving depths ranged from 0.1m to 0.35m (323, 354, 1022, 1142, 1176/1196, 1215, 1217, 1230, 1234, 1241, 1259, 1263, 1268, 1427, 1468, 1503, 1509, 1574 and 1807).

The midden deposit on Site 1 was encountered during 2008, but not identified until the post-excavation phase when the section drawings were inspected. Remnants of a the probable Late Iron Age House 3 overlay deposits (336 and 360) of clay loam soil (Fig 25) which contained All Cannings Cross-type pottery and animal bone in particular. Part of a right human femur was retrieved from the lowest part of the excavated section in layer 375. The cobble surface was not recognised at the time, but can clearly be seen in the section. In 2011, the human remains from grave 318 (see Fig. 48) were lifted and it was noted that the cobbled surface underlay the grave cut.

The Late Bronze Age /Early Iron Age features from Site 1 are located on Figure 28 and those from Site 2 on Figure 29, their stratigraphic relationship with features of other periods can be seen on Figure 215.

The south- and west-facing, longitudinal sections from Site 2 (Figs 26 and 27) illustrate the midden build up in that area; their positions are noted on Fig. 29. Two samples were submitted for radiocarbon dating; a cattle phalanx from above the cobble floor and a sheep/goat tibia from the midden build-up. The former (1301) was from the primary accumulation above the cobbles (layer 5) and returned a date of 775-485 cal BC (GU37941) with a 95% probability range. The second sample (1017) from the second layer of the midden, where a complete bronze axe had been found

Type 1 pits Of the five Type 1 pits, two were located on Site 1 (335 and 326/390) and three (1412, 1513 and 1302) on Site 2. Because of their location, it was only possible to halfsection the pits on Site 2; on Site 1, pit 335 was completely excavated, and due to time constraints, pit 326/390 was half-sectioned. Conventionally, a storage function is usually assigned to this type of pit. 23

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 24 Designated slot numbers across Site 2

in diameter and was 1.2m deep, its base formed by the natural limestone bedrock. The southern half of the pit was stepped and lined with limestone slabs for its complete depth. There was no indication that the northern half of the pit was ever lined. The stepped ledge was 0.4m below

Pit 335 (Figs 30 and 31) This pit was located on the eastern side of Site 1 and was initially identified by an arc of vertically-set stones visible on the stripped surface. The pit measured 2.3m 24

The Late Bronze Age/Early Iron Age Transition Table 4 Midden slots with their excavated layers (in some areas Level 4 was excavated in two spits, a and b) Clearance depth 1001 1001 1001 1001 1001 1001 1001 1001 1001 1001 1001 1001 1001 1001 1002 1002 1002 1002 1002 1002 1002 1002 1002 1002 1002 1002 1002 1002 1002 1002 1002 1002 1096 1096 1096 1096 1096 1096 1096 1096 1096 1096 1096 1096 1400 1400 1400 1400 1400 1400 1400 1400 1400 1401 1401 1401 1401 396 340 340 340 340

Slot No. (in m) 1052 1060 1078 1031 1023 1052 1045 1026 1014 1029 1034 1049 1015 1261 1189 1180 1166 1164 1019/80 1068 1087 1011a 1011b 1071 1062 1054 1266 1076 1020 1085 1149 1047 1099 1101 1103 1105 1107 1109 1111 1113 1115 1117 1119 1121 1402 1403

Layer 1 Fill 0.1-0.2 1053 1061 1079 1032 1024 1053 1046 1027 1007 1030 1035 1050 1016 1277 1190 1181 1167 1165 1081 1069 1088 1012 1043 1072 1063 1055 1267 1077 1021 1086 1150 1048 1100 1102 1104 1106 1108 1110 1112 1114 1116 1118 1120 1122 1411 1406 1410 1473

1448 1475 1404

1804 315 337 330 330 337

Layer 2 Fill 0.1

Layer 3 Fill 0.1

1148 1033 1042

1133 1125

1010 1130

1135

1017 1280

1133 1133 1133 1301

397 358 338 351 370

Layer 4b Fill 0.03

1157

1171

1157 1157

1171 1171 1171

1183 1179 1089 1070 1013 1051 1082 1064 1098

1272

1075

1084/87

1028 1065

1294

1128

1025

1170 1170 1193

1067

1201

1225

1224 1211 1432 1430 1437 1431

1233 1097

1168

1476

1491 1459 1460 1461 1462 1463 1501 1521 1502 1449

1490 1478 1407

Layer 4a Fill 0.05

1416 1425 1443 1431 1807 1354 360 336 361

1449

25

1193 1246 1228 1228 1228 1227 1227 1227 1228 1228 1228 1227 1227 1227

1456

Layer 5 over cobbles 0.01 1200 1200 1200 1228 1228 1200 1200 1200 1200 1200 1200 1200 1228 1301 1226 1226 1226 1226 1227 1227 1227/68 1226 1226 1226 1226 1227 1226 1200 1226 1226 1227 1244 1244 1244 1245 1245 1245 1244 1244 1244 1245 1245 1245

Layer 6 cobbles 0.8 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039 1559 1559 1559 1560 1560 1560 1558 1558 1558 1039 1039 1039 1039 1039 1039 1039 1039 1039 1039

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 25 Site 1 sections through midden layers (located on Fig. 28)

the top of the pit and was 0.4m wide. Four fills were identified.

Flecks of charcoal were noted. The undiagnostic pottery (30 sherds) was all Late Bronze Age/Early Iron Age in date; also present were 14 pieces of worked flint, quantities of animal bone including one worked, sharpened fragment and scraps of shale. The third fill (391), consisted of large quantities of limestone slabs, loosely and unevenly packed into a light brown clay loam which appeared to have been washed into the voids between the stones. It would seem that this layer had been deposited over a short period of time. The base fill (392) consisted of a dark brown humic, charcoal-flecked loam, the charcoal forming a discrete band on the base of the pit. The finds included 34 sherds of pottery, worked flint, animal bone and shale waste.

The partial lining (379) was made up of large, generally rectangular limestone blocks which had been placed vertically around the ledge and were set neatly and horizontally in a semi-circle. The upper fill (359) was 0.6m deep, and comprised mid-brown loam containing large quantities of limestone which varied from small pieces to very large blocks. Finds included Late Iron Age and Early Iron Age/Late Bronze Age pottery, two iron nails, shale waste and a fragment of bracelet and animal bone. The make-up of the second layer 380, was similar, but was stickier in texture and contained larger quantities of stone. 26

The Late Bronze Age/Early Iron Age Transition

Figure 26. Site 2; sections through midden layers on the northern edge of site 2, 2011 [1] and through midden layers on the northern edge of site 2, 2010 [2] (located on Figure 29)

the pits was 2.4m, pit 326 originally being approximately 1.6m in diameter and pit 390 being approximately 1.2m. Both pits were cut down to bedrock at a depth of 1.45m. This feature was half-sectioned.

Pit 326/390 (Figs 32 and 33) This complex, figure of eight-shaped pit was located only 1m from pit 335. It was formed from two pits: the earlier one 326, was stone lined and was cut on the west side by the unlined and slightly bellied pit 390. There was no visible difference between the infilling materials of the pits and therefore they must have been ultimately used and filled as a single entity. The maximum width across

A clay lining (388) approximately 0.1m thick had been applied to the sides of the double pit: this extended in a thinner layer over the base. Within pit 326 and embedded into the clay lining was a well-constructed section of 27

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 27 Site 2; sections through midden layers on the western edge of site 2011 [3] (located on Figure 29)

walling (352), which had been very carefully laid with horizontally placed stones which slightly overhung at the top of the pit. Large capping stones were laid around the top of the walled section. Fragments of fired clay and four sherds of pottery were retrieved from this structural deposit. A 0.1m deposit (355) across the top of the pit comprised a mid-brown charcoal-flecked clay loam containing small pieces of limestone. The finds consisted of 68 sherds of Late Bronze Age/Early Iron Age pottery, fragments of fired clay, worked flint, animal bone and shale fragments including a small piece of a roughout from a bracelet. A deposit of material (357) consisting of larger limestone blocks appeared to have been dumped near to the western side of the pit. Eight sherds of pottery and two animal jaws were recorded. This deposit cut into layer 356 which was a mid-brown, charcoal-flecked clay loam containing limestone pieces. A large amount of pottery was retrieved (154 sherds from at least 13 different vessels), as well as worked flint, fired clay and shale fragments. Animal bone was also present (58 pieces) and one fragment of human bone – part of a left femur. The middle layer (366) was a pale to mid-brown clay loam containing much decomposing limestone; other tabular fragments were mostly horizontally laid. The finds included pottery, worked flint, a shale bracelet roughout, animal bone and fragments of daub. Beneath this, layer 376 was similar, but much of the limestone had been burnt. The finds were also similar; of note was a fragment of quern and part of an adult human femur. The primary fill (layer 378) was a clay-rich mid-brown loam with thick lenses of charcoal. Also present were fragments of disintegrating fired clay and medium to large slabs of limestone, many of which

were burnt. A deposit of limpet shells was part of this layer. Finds included pottery and animal bone. On the base of the pit and directly above the clay lining was a very thin (0.01m) layer of fine, black, humic soil which was devoid of finds. This was similar to layer 389, which had accumulated in pit 390 beneath the clay lining. Pit 1412 (Figs 34 and 35) This pit cut through the cobble layer and was located on the eastern edge of site 2 (Fig. 29): it was only possible to half-section this feature. The sides were almost vertical and the base and lower sides were cut into the natural bedrock. The pit was probably circular with a maximum diameter of 1.6m and a depth of 1.08m; five discrete fills were recorded. The top of the pit (1413) was characterised by a jumble of large limestone blocks which masked the edges. The blocks were mostly set in dark brown loam which contained quantities of Late Iron Age pottery (157 sherds weighing 1650g) and 600g of animal bone. Fill 1423 (a mid- to dark brown loam) was recorded across the whole pit, but at the southern end and for half of its width it was only 0.08m deep. It then plunged steeply towards the northern side. On its northern and western edges this layer was defined by vertical limestone slabs: it is possible that this may have been a separate pit inserted at a later date and cut layers 1452 and 1507. The nature of the finds was markedly different from the layer above. The pottery was in the All Cannings Cross style and was found with typical midden finds of bone, fired clay, worked flint, pebbles, and 28

The Late Bronze Age/Early Iron Age Transition

Figure 28 Site 1 plan Late Bronze Age/Early Iron Age features associated with the midden deposit

(1508) was 70% clay, containing horizontal limestone pieces on top of the natural bedrock. These may have once formed part of a stone lining. Finds were sparse and included seven sherds of pottery, seven pieces of animal bone and a rough flint core.

shale. Most of the latter was waste, but a spindle whorl was noted. Layer 1452 comprised a sticky dark brown loam into which tabular limestone had been dumped, forming a horizontal layer dividing this layer from the underlying 1507. Two discrete lenses of clay were recorded. Similar finds were noted including a shale bracelet roughout and three fragments of human infant bone. Layer 1507 was recorded across the whole pit and comprised dark brown loam with large tabular limestones and small deposits of raw yellow clay. Fewer finds were retrieved: nine sherds of pottery and four pieces of animal bone. The base fill

Pit 1513 (Figs 36 (upper) and 37) The pit was located on the extreme western edge of the site: approximately half remains underneath the modern road. It measured 1.5m wide and was 0.9m deep, with 29

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 29 Site 2 plan Late Bronze Age/Early Iron Age features associated with the midden deposit

30

The Late Bronze Age/Early Iron Age Transition

Figure 30 Late Bronze Age/Early Iron Age Type 1 pit 335, section

slightly bulbous sides and a flat base. The pit cut through the cobble layer and had a complex infilling sequence with eight recorded fills. A lining (1530) consisting of approximately 0.08m to 0.17m of yellow clay had been applied to the sides and base of the pit. The uppermost layer 1514, extended across the pit: on top of this was a spread of limpet shells (1499), which is discussed under cultural deposits below. The dark brown loam layer sloped quite steeply from north to south and was characterised by very large limestone blocks which appeared to have been

thrown into the pit and was probably the final deposition in the infilling sequence. Layer 1517 was against 1514 and consisted of a very dark brown gritty loam which contained tiny fragments of decomposing, partially-fired red pottery and larger pieces of limestone, some of which had been burnt. A small quantity of Late Iron Age pottery was recorded in this fill. Shallow layer 1518 underlay 1514 and 1517, extended across almost the complete width of the pit and was slightly dished in the centre. It comprised dark red-brown, hard and compacted gritty

Figure 31 Late Bronze Age/Early Iron Age Type 1 pit 335, view from the north

31

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 32 Late Bronze Age/Early Iron Age Type 1 pit 326/390, section

loam with a high charcoal content and lumps of burnt and fractured limestone. This layer may have been burnt in situ and could have been the base of a firing clamp (spalled sherds of pottery were found throughout this pit). Layer 1519 also extended across the pit and consisted of a crumbly, charcoal-flecked, grey-brown loam with discrete lenses of charcoal and occasional pieces of burnt limestone. Finds included a ceramic spindle whorl (Fig. 195.8) and 64 sherds of pottery including rims from five different vessels. Layer 1522 on the northern side of the pit comprised a grey-brown, charcoal-flecked, compacted loam; eight pieces of pottery were recorded including spalled sherds. Layer 1524 was a dark grey-brown loam with sparse flecks of charcoal and graduated from soft at the top to compact on its base, the finds included 40 sherds of pottery, again including spalls, a hammerstone, a broken bone knife, and 26 pieces of animal bone. On the base, layer 1525 was similar in make-up to 1524, but was more compact. Among the 56 sherds of pottery was a substantial part of an open bowl (Fig. 140.4), together with sherds and spalls from at least another three vessels. Animal bone was present in this layer. A radiocarbon date from a sheep/goat radius (layer 1525) (GU37939) dates this pit to 498-363 cal BC, which suggests that it was cut towards the end of activity on the midden. Pit 1302 (Fig. 36, lower) It was only possible to record part of this pit, which was located in a small rectangular cutting on the extreme

Figure 33 Late Bronze Age/Early Iron Age Type 1 pit 326/390, view from the west

32

The Late Bronze Age/Early Iron Age Transition of pottery (120 sherds weighing 1045g). A possible floor (1301) comprising crushed limestone, some burnt and mixed with loam, butted up to the east edge of the pit and was laid above the cobble layer 1039. Type 2 pits Five Type 2 pits (393, 1126, 1136, 1503 and 1574) were fully excavated and may have been small storage pits. Pit sections are illustrated on Fig. 38. Pit 383 was located on Site 1, the remainder on Site 2. Pit 393 This rectangular pit, located on the north-west edge of Site 1was 1.22m long and 0.85m wide, with vertical sides and a base which sloped gently from south to north. The depth varied from 0.2m to 0.4m. The sides of the pit comprised limestone slabs, vertical at the top with smaller blocks below. The dark brown clay loam (394) contained limestone chunks and occasional heathstone pieces and sparse flecks of charcoal. Pottery (79 sherds) and animal bone (61 pieces) were recorded.

Figure 34 Late Bronze Age/Early Iron Age Type 1 pit 1412, section

north east corner of Site 2 (Slot 1261). Partial excavation suggested that it was a vertical-sided, deep storage pit and was probably lined with limestone. Only 0.23m of its mid-brown loam fill (1303) was removed. This contained limestone chunks, 76 fragments of animal bone and for the depth of the pit excavated, a relatively large amount

Pit 1126 This barrel-shaped pit measured 0.7m in diameter, was 0.75m deep and had three fills. Located on the north-

Figure 35 Late Bronze Age/Early Iron Age Type 1 pit 1412, view from the north

33

Multi-period Occupation at Football Field, Worth Matravers, Dorset east corner of Site 2, it cut through the cobbles and into the natural bedrock near its base. It had probably been stone-lined, with vertical stones set against a clay lining. It was apparent from the trench section that the pit had cut through the uppermost layer of the midden deposit and must have been one of the latest features of this period. The upper fill (1127) was a dark brown loam with frequent charcoal pieces and chunks of limestone. Very large quantities of pottery were retrieved (289 sherds weighing 6302g) including rims from 25 different vessels; in contrast, no bases were recorded. In addition there was a stone smoother, a flint knife and 74 fragments of animal bone. The middle fill (1152) was slightly lighter in colour and contained small, fragmented pieces of limestone, 15 sherds of pottery and four pieces of animal bone. The base fill (1154) of mid-brown loam was initially thought to be part of a separate feature but is more likely to have acquired its posthole shape through slumping on the western side of the pit. A large slab of limestone, probably part of the pit’s lining, lay against its eastern side. Pit 1503 The pit had poor definition in the midden layer and cut Type 1 pit 1513. It had moderately sloping sides and a rounded base and measured 0.6 in diameter and was 0.3m deep. The lower part of the pit was lined with yellow clay

Figure 36 Late Bronze Age/Early Iron Age Type 1 pits 1513 and 1302, sections

Figure 37 Late Bronze Age/Early Iron Age Type 1 pit 1513, note the burnt layer in the centre of the pit. View from the east

34

The Late Bronze Age/Early Iron Age Transition and was 0.35m deep. The sides were steep and the base sloped upwards to the east. Three fills were recorded; the upper (1137) comprised a mid- to dark brown loam with green/brown clay-like smears and frequent pieces of small limestone rubble. Two fills of clay (1138 and 1140) consisted of a yellow-brown layer on top of a mid-brown base fill. All fills yielded undiagnostic Late Bronze Age/ Early Iron Age pottery. Type 3 pits Sixteen Type 3 pits (and scoops) were recognised; two on Site 1 (323 and 354) and fourteen on Site 2 (1022, 1142, 1176/1196, 1215, 1217, 1220, 1234, 1241, 1263, 1268, 1427, 1468, 1509 and 1807). Seven of these have been selected for discussion and were clustered together on the eastern side of Site 2 (see Fig. 29) and all were dug through the midden deposit, some penetrating the cobble floor. The features are characterised by their depths, which are no more than 0.3m. Feature 1142 was initially identified as a hearth and1022 as a querning scoop; the function of the others is not known. Scoop 1807 was recorded in the final phase of work in March 2012 and was located about 0.15m below the topsoil just south of Site 2. Scoop 323 Sited in the centre of Site 1, this oval scoop with gently sloping sides and a rounded base was 0.9m long by 0.73m wide and was 0.19m deep. The western side may have initially been lined with small limestone slabs. The dark brown humic fill (363) contained tabular limestone pieces which gave the appearance of being placed in the feature. Two sherds of pottery were retrieved.

Figure 38 Late Bronze Age/Early Iron Age Type 2 pits, 1126, 1503, 1574 and 393, sections

Pit 1022

(1504); embedded in this were two fragments of shale and a large stone smoother which had been set upright. The pit fill was dark brown sticky loam (1501) and distributed evenly throughout was a deposit of bone (140 fragments), 14 lumps of shale and eight sherds of pottery weighing 74g. Vertical stones were recorded on the edge of the pit and may have been an upper part of the lining.

Sited on the northern side of Site 2, this (probably) circular feature was only one quarter sectioned due to its location on the excavation area boundary. Measuring 0.8m in diameter and with a depth of 0.35m, the pit had almost vertical sides and a flat base where it was cut into natural bedrock. The upper fill (1041) was coarse yellow-red clay with rare charcoal inclusions. Small fragments of pottery from five different vessels were identified together with two small pieces of animal bone and a large pebble. The lower fill (1044) consisted of a fine, dark brown clay containing lenses of dark grey silt. Two thin slabs of limestone separated the layers. Finds from the lower layer included five sherds of pottery from three different vessels and four pieces of animal bone. The pit was cut from the top of the midden layer and must have been one of the last features to be inserted into the midden build-up.

Pit 1574 Located in the extreme south-west corner of the site, the sub-circular pit was cut into natural clay and had very steep, almost vertical sides with a flat base; it measured 0.7m long by 0.6m wide and had a depth of 0.36m. The single fill (1597) was a dark brown loam into which large pieces of limestone had been placed; three of these were upright. Small quantities of finds were retrieved including animal bone, worked flint and one sherd of pottery.

Quern scoop 1036/1142 (Figs 39 and 40)

Pit 1136 (not illustrated)

This feature was initially identified by the presence of a large quernstone 1036 (see Bellamy Chapter 7, Fig. 118.1). When this was removed, it could be seen that it had been deliberately placed over a small scoop. The oval feature

Located on the northern edge of Site 2, this circular pit was ill-defined on the surface. It measured 0.5m in diameter 35

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 39 Late Bronze Age/Early Iron Age Type 3 pits 1022, 1142, 1215, 1220, 1241, 1263, 1427, 1468 (Site 2) and 323 (Site 1), sections

was defined by small limestone slabs, measured 0.4m long by 0.33m wide and was 0.1m deep; the base consisted of the cobble layer 1039. The dark brown loam fill (1141) was charcoal-rich and contained a variety of finds including 13 sherds of pottery from seven different vessels, four pieces of animal bone, two worked flints, a smooth shale pebble and approximately one third of an imported continental glass finger ring (See Henderson, Chapter 11, Fig. 191). A radiocarbon date (GU37938) places the pit in the 6th/7th centuries cal BC.

that they were earlier in the chronological sequence than the pottery from the pit above. Animal bone and worked flint were also present.

Scoop 1268 (Fig. 45)

The full dimensions of this scoop could not be defined as it lay on the south boundary of the excavation area. It appeared to be oval, measuring 0.7 into the site baulk and was 0.5m wide with a depth of 0.1m. The sides sloped gently to a flat base. The single fill (1216) was a dark to mid-brown sticky loam which contained frequent small pieces of limestone and rare charcoal. The finds included 13 sherds of pottery weighing 100g from eight different vessels including a handled jar (Fig. 143.4). In addition, worked flint, shale waste, fired clay, a single pebble and animal bone fragments were recorded

The following three scoops were cut into the cobbles; all had linings of small stones and were cut into natural yellow clay. They must have been some of the earliest features cutting the midden deposit as it was developing. Scoop 1215

This large scoop (which may have been a posthole) lay partially under the quern pit (1036/1142) described above and was visible as a dark stain when that pit had been excavated. This feature had steeply sloping sides and a rounded base with a diameter of 0.4m and a depth of 0.27m. The fill (1269) was a dark to mid-brown loam containing burnt and degraded limestone and was artefactrich. The pottery (15 sherds weighing 69g) included fragments from at least five vessels, their form suggesting 36

The Late Bronze Age/Early Iron Age Transition

Figure 40 Quern 1036 with associated ceramic objects to the left. The quern covered Type 3 pit 1142. View from the south

which had been placed on the base of the pit. Other finds included a partial shale bracelet roughout, worked flint and nine undiagnostic sherds of pottery weighing 38g.

Scoop 1241 (Fig. 41) This large shallow, oval scoop measured 1.2m long by 0.9m wide and was 0.15m deep. Its edges were defined by deliberately placed limestone cobbles, some of which were vertically placed. The single fill (1242) was a dark to mid-brown sticky loam containing rare charcoal flecks. A quantity of moderately sized limestone pieces appeared to have been dumped in the south-west corner. For its size, a relatively large amount of pottery was incorporated into the fill (41 sherds weighing 353g from eight different vessels. Worked flint, shale debris, fired clay, a single pebble and 34 animal bone fragments were recorded.

Scoop 1427 This almost circular feature measured 0.7m long and 0.6m wide and was 0.15m deep with gently sloping sides and a flat base. The dark brown loam fill (1429) was at least 70% charcoal. There was no evidence for in-situ burning: therefore the fill must have been introduced after the scoop was dug. Most of the pottery (86 sherds weighing 1180g) was located in the upper part of the fill, consisting of at least seven vessels including three jars and one large plain bowl as well as fragments of animal bone. It is possible that this pit may have been associated with a post-built house (see below).

Scoop 1263 This shallow almost circular scoop measured 0.6 long by 0.56m wide and was 0.1m deep with sloping sides and a flattish base. The single fill (1264) was a dark to midbrown sticky loam which contained frequent small pieces of limestone and sparse charcoal. Three fragments of animal bone were noted, including part of a cattle sacrum

Scoop 1468 This steep-sided circular feature had a diameter of 0.3m and was 0.1m deep with a slightly rounded base. The 37

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 41 Type 3 pit 1241, view from the north

sticky, mid-brown loam fill (1469) contained small pieces of limestone, charcoal and two sherds of pottery.

55 sherds of pottery weighing 826g and comprising at least ten different vessels, with a preponderance of furrowed bowls (Fig. 145). Eight pebbles lay on top of the deposit, and within the soil matrix were fragments from two shale bracelets, seven worked flints and seven animal bones.

Cultural Deposits An area of compaction was noted on Site 2 comprising a defined area (1431), (see Fig. 27, section 3) measuring approximately 2m by 1.75m. It consisted of a hard deposit of what appeared to be crushed and burnt limestone incorporating charcoal, pottery (12 sherds weighing 171g) and animal bone. This feature was 0.15m above the cobbles and extended out of the excavated area to the north.

Deposit 1175 (Fig. 43) Measuring 0.7m by 0.6m and contained within the upper layer (1102) of Slot 110, the associated finds appeared to be enclosed by pieces of limestone set in a sub-circular fashion. The finds included four sherds of pottery weighing 34g including a stab-decorated piece (Fig. 146.1), shale fragments and five pieces of animal bone.

Six deliberately placed deposits were recorded in the midden on Site 2. These were of two different types; the first group (1169, 1175, 1236 and 1256) comprised varying amounts of cultural material and were clustered together at the eastern end of the excavation area. The second group consisted of two discrete dumps of limpet shells (1499 and 1500) located at the western end of the excavation.

Deposit 1236 This small deposit measuring 0.4m by 0.3m in layer 1104 of Slot 1103 included six animal bones placed together; one of these was a horn core. Two undiagnostic sherds of pottery were also recorded.

Deposit 1169 (Fig. 42)

Deposit 1256

This was the largest of the four deposits and was contained within the upper layer (1114) of the midden in Slot 1113. It measured 0.8m long by 0.5m wide and had a depth of approximately 0.05m. The assemblage was dominated by

Located underneath the probable Late Iron Age House 4 (see Chapter 4) in layer 1254 of Slot 1103, the deposit measured 0.36m by 0.3m and was 0.03m deep. The 38

The Late Bronze Age/Early Iron Age Transition

Figure 42 Deposit 1169 with pottery fragments and pebbles, view from the north-east

Figure 43 Deposit 1175 with animal bone, view from the north

39

Multi-period Occupation at Football Field, Worth Matravers, Dorset Postholes

material comprised 19 sherds of pottery weighing 172g including part of the base and body sherds from a plain jar and sherds from three other pots (Fig. 146. 2-3). One piece of worked flint and three animal bones were also recorded.

Fifteen postholes, three on Site 1 (323, 345 and 347) and 12 on Site 2 (1139, 1220, 1222, 1223, 1247, 1287, 1481, 1495, 1536, 1540, 1547 and 1555) were dated by pottery which could positively be ascribed to this period. Further postholes (1472, 1479, 1481, 1542 and 1544) were undated but are likely to be contemporary. The postholes varied in size from 0.2m to 0.4m in width and from 0.08m to 0.3m in depth. Selected posthole sections are illustrated on Fig. 45 and their locations on Figs 28 and 29.

Deposit 1499 Situated on the extreme western margin of the site, this deposit had a width of 1m and was approximately 0.1m deep. Its true extent is unknown as the remainder of the feature lay outside the site. Consisting of hundreds of limpet shells, the deposit was located above the Type 1 pit 1513 which by then had been in-filled.

Postholes 345 and 347 These postholes, located in the northern third of Site 1, were only 0.1m apart and had very similar dimensions; posthole 354 was 0.28m in diameter and 0.2m deep and 347 was 0.31m in diameter and 0.2m deep. Their sides were near-vertical and the bases were slightly rounded. Both had dark brown, sticky loam fills (346 and 348). One small sherd of pottery was retrieved from posthole 346 and four sherds from posthole 347. Tabular limestone pieces, perhaps remnants of packing stones, were found in each posthole. To the west of posthole 345 was a cache of seven pebbles (331) which may have been deliberately placed.

Deposit 1500 (Fig. 44) This deposit measured 1.9m by 0.7m with a depth of 0.13m. It straddled two layers (1460 and 1461) and was located about 2.5m away from deposit 1499. Hundreds of closely packed limpets were recorded, only a sample of which were retained. Incorporated into the limpet mass were 29 sherds of pottery weighing 249g from at least six different vessels including jars and bowls (Fig. 146. 4-8) and 15 fragments of animal bone.

Figure 44 Limpet deposit 1500, viewed from the south-east

40

The Late Bronze Age/Early Iron Age Transition small to medium pieces of limestone and a pottery sherd which had been re-worked and partially pierced for a spindle whorl (see Fig 195.6). Four animal bone fragments were recorded. Posthole 1536 Cut into the midden layer 1460, this oval posthole measured 0.3m long by 0.2m wide and was 0.18m deep with steep, sloping sides and a flat base on which a small slab of limestone had been placed. The dark brown, sticky loam fill (1537) contained limestone pieces as well as a larger fragment which was upright in the centre of the feature. The sparse finds included one sherd of pottery and two flint flakes. Posthole 1547 Cut into the midden layer 1462, this oval posthole measured 0.33m by 0.25m and was 0.2m deep with vertical sides and a flattish base. The fill was mid-brown loam containing very small fragments of limestone. Finds included three small sherds of pottery weighing 4g, four pieces of animal bone and four worked flints. Posthole 1555 Figure 45 Late Bronze Age/Early Iron Age postholes associated the midden deposit on Sites 1 and 2, sections

Cut into midden layer 1521, this circular posthole had a diameter of 0.2m, was 0.23m deep with almost vertical sides and a narrow flat base. The fill was a dark brown loam with occasional pieces of charcoal. A single piece of limestone had been placed in the centre of the posthole and protruded about 0.05m above the cobbled surface and may have been post packing. One sherd of pottery was recorded.

The following postholes were located on Site 2. Posthole 1220 It is likely that this feature, measuring 0.35m in diameter but only 0.09m deep, was truncated by later activity. The surviving sides were vertical on the west and sloped moderately to the east, and the base was slightly rounded. The fill (1221) was a soft, dark brown, charcoal-flecked loam. Five sherds of pottery weighing 97 grams were retrieved together with eight pieces of animal bone and a large broken burnt stone pebble.

Structural Elements Associated with a Possible PostBuilt Roundhouse (House 2) (Fig. 29) Six features comprising five postholes and one small pit (1247, 1472, 1540, 1547, 1136 and 1468) were identified at a very late stage in the post-excavation work as potentially belonging to a post-built roundhouse on the central part of Site 2 and sited immediately east of the Late Bronze Age roundhouse (Chapter 2). Only the northern half of this putative building survived, probably due to intensive activity particularly in the Roman period when a stonebuilt barn occupied the site. The postholes, all of which cut through the cobble layer 1039, were steep sided with flat or very gently rounded bases and varied in shape from round to oval, in diameter from 0.35m to 0.24m and in depth from 0.3m to 0.2m. The fills were a consistent dark brown, sticky loam and large pieces of limestone had been wedged into the fills. Pottery from postholes 1540 and 1547 is illustrated on Fig. 147. 3-5). Postholes 1247 and 1547 were 5m apart and the total diameter of the house must have been in the region of 7m. There was no evidence for an eaves-drip gully. Within the posthole radius, two features may have belonged to the house. Pit 1136 was ill-defined on the surface and lay towards the centre of

Posthole 1287 Lying in the north-east extension of the site, this small posthole was 0.2m in diameter and 0.23m deep with vertical sides and a slightly rounded base. The fill (1288) of dark brown loam contained crumbly limestone fragments, two sherds of undiagnostic pottery and four pieces of animal bone. It was overlaid by horizontal limestone slabs (1289) of uncertain but later date. Posthole 1495 Cut into the midden layer 1491, this was probably another truncated posthole measuring 0.26m in diameter with a surviving depth of 0.17m. The feature was covered with yellow clay. The dark brown sticky fill (1496) contained 41

Multi-period Occupation at Football Field, Worth Matravers, Dorset the structure and measured 0.50m in diameter and was 0.35m deep. The sides were steep and the base sloped upwards to the east. Three fills were recorded; the upper (1137) comprised a mid to dark brown loam with green/ brown clay-like smears and at least 40% small limestone rubble. Two fills of clay (1138 and 1140) consisted of a yellow/brown layer on top of a mid-brown base fill. All fills yielded undiagnostic Late Bronze Age/Early Iron Age pottery (11 sherds in total weighing 117g).

0.27m high. They were firmly embedded in the cobble layer, but did not penetrate it.

Settings of Packed Limestone

Conclusions

Seven stone settings characterised by varying numbers of vertically placed small limestone slabs were recorded on Site 2 (1038, 1187, 1188, 1238, 1520, 1535 and 1538). They were either placed on, or penetrated the cobble layer; their function is unknown, but must have had purpose and meaning at the time they were set. Two are described in detail below.

It has to be stressed that the features discussed above are probably only a small proportion of those which are likely to be present within the midden deposit on the unexcavated area of this site. The amount of material from the midden itself and the number and variety of the excavated features demonstrate the complex build-up and the variety of activities which took place here over several centuries. On Site 2, there appears to be an area in the centre of the site where there are few features; this may have been the result of later Iron Age and Roman activity, principally the construction of buildings which would have destroyed earlier features. The integrity of the midden itself was compromised by the incorporation of later material.

Setting 1520 (Fig. 47) This setting measured 0.9m long and 0.08m wide and consisted of twelve thin but varying sized slabs of limestone placed closely together; the tallest stone was 0.27m high and the arrangement was set into the cobble layer.

Setting 1238 (Fig. 46) This arrangement measured 0.36m long and 0.29m wide and consisted of six thin vertical stones, placed close together and set at a slight angle. The tallest stone was

Figure 46 Stone setting 1179 inserted into cobbles 1039, view from the west

42

The Late Bronze Age/Early Iron Age Transition

Figure 47 Stone setting 1520 inserted into cobbles 1039, view from the north

43

4 The Middle and Late Iron Age The eastern side of the pit had been cut by a deposit of almost pure, yellow clay (1198), this measured 0.38m in width (north to south) and 0.35m in depth. Above the clay, the fill appeared identical to that of 1159. The clay encroached 0.2m into the pit and had cut through fills 1159 and 1151, with a small remnant of the grain-rich 1151 still in situ beneath it. The clay deposit was interpreted as being later than fills 1159 and 1151, and must have been placed in the pit late in the infill sequence. Apart from one large limestone pebble, there were no other finds. The fill sloped down from west to east.

A single Middle Iron Age feature and 20 Late Iron Age features were recorded on Sites 1 and 2 and in a services trench near to Site 2 (Figs 48 and 49). These comprised the structural remains of two stone-built roundhouses, an oven attached to one of the houses, areas of paving, pits, a single posthole and an inhumation. The features have been dated by radiocarbon, associated pottery and metalwork. Middle Iron Age Activity There was only one securely dated feature of this date. This was a large, deep Type 1 pit (1182) located on the central southern side of Site 2. Radiocarbon dating of a single Bromus secalinus (rye brome) seed returned dates falling between 342-326 cal BC and 205-48 cal BC (GU37943) at 95% probability indicating that the pit was in-filled during the 3rd- 4th centuries BC.

Fill 1163 comprised brown silty soil with traces of yellow and grey clay; flecks of charcoal and some burnt grain were also present. The finds included small fragments of bone, pottery, an iron object (possibly a pin), a bone pin, a stone pounder, a broken bone needle and a La Tène 1-type brooch (see Corney, Chapter 11, Fig. 182.1). Many badly degraded limpet shells were recorded in this layer as well as burnt limestone, one large limestone slab and several other smaller pieces. Again, this fill sloped to the down east.

Pit 1182 (Figs 50 and 51) This oval pit cut through the cobble layer 1039 and measured 1.98m by 1.62m with a total depth of 1.1m; it was fully excavated. The sides were cut into natural tabular limestone and yellow clay-rich soil and were nearvertical, except for a cut-out step to the west (see Fig. 51). The base was formed from bedrock and was flat, except where natural hollows occurred. Thin slabs of limestone had been placed over these to even it up.

Fill 1186 consisted of a layer of mid-brown loam with varying amounts of orange clay; towards the bottom, several large limestone slabs and blue limestone pieces occurred. The finds included pottery and a sheep/goat jaw bone. The fill sloped to the down east and varied in depth from 0.07m in the east to 0.4 in the west. The base fill (1191) lay directly on natural bedrock, had a maximum depth of 0.02m and comprised sterile dark grey silt. One of the limestone slabs lying on the bedrock was lifted, revealing a thin layer of fine greenish-grey silt. This artefact-free fill was only present on the north side of the pit and is therefore not on the section drawing.

Five discrete and complex fills together with a later deposit of clay were recorded. The upper fill (1159) had been disturbed by later activity associated with the cutting of pit 1131. This layer had a maximum depth of 0.25m and comprised dark brown, loamy soil and many limestone fragments; finds included pottery, bone, limpet shells, pebbles and an iron object, possibly a hinge (SF 276). The limestone pieces may represent a disturbed lining from an earlier phase of the later pit 1131. This fill could suggest a re-digging of the pit as it appeared to have a distinct edge to the west, cutting an earlier fill (1151),

Conclusions Pit 1182 was the only feature of Middle Iron Age date, although a few intrusive pottery fragments were retrieved from Late Bronze Age/Early Iron Age pits 320/390 and 393 on Site 1, suggesting that activity during this period may have been more widespread. The large size of the pit perhaps infers a storage function. At the end of its period of use it was backfilled, probably over a relatively short period of time, with the material initially tipped in from the western side. The presence of burnt grain and metalwork items suggests deliberate deposition of these materials.

Fill 1151 comprised dark brown/black loamy soil which incorporated very large amounts of burnt grain and charcoal. There were few finds; two pottery sherds, badly degraded limpet shells, a possible burnishing stone and a worked bone. The bottom of the fill sloped towards the east with a depth of 0.17m on the eastern side and 0.04m on the west. This fill may represent a re-digging of the pit as it had a near-vertical edge on the western side, cutting through an earlier fill (1163). On the base of the grain-rich fill, a thin compact layer (1155) consisting of burnt orange sand and areas of red/black burnt shale was present to a depth of 0.03m.

Late Iron Age Activity Large amounts of randomly scattered pottery, animal bone and shale were recorded in the clear-back and midden layer 44

The Middle and Late Iron Age

Figure 48 Site 1 plan, Late Iron Age features and Post-Roman grave 425

45

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 49 Site 2 plan, Middle and Late Iron Age features

46

The Middle and Late Iron Age

Figure 50 Middle Iron Age pit 1182 and Late Iron Age pit 1131, section

Figure 51 Middle Iron Age pit 1182 and Late Iron Age pit 1131 (directly above), view from the north

338 on Site 1 and in many of the upper slot fills on Site 2. Cut features were identified by pottery, metalwork and shale from their fillings. On Site 1, the features comprised part of a roundhouse (House 3) with an associated oven/ dryer, an area of paving, five pits, a single posthole and a single crouched burial. On Site 2, remnants of a further

house (House 4) were recorded together with eight pit type features, a single posthole and four discrete deposits of pottery. During the 2007 evaluation, a Late Iron Age pit was identified in Trench 4 and a further pit was identified in a service trench on the southern side of the field (for location of these features see Figs 5 and 220). 47

Multi-period Occupation at Football Field, Worth Matravers, Dorset The Roundhouses

defined, but the line of the internal facing slabs was more difficult to determine. No internal postholes or structures survived. Large amounts of unstratified Late Iron Age pottery (293 sherds weighing 3702g) were recovered from the putative floor area as were 268 fragments of animal bone. A shale spindle whorl (SF 47, Fig. 122.7) and small shale fragments were also recorded.

House 3 (Figs 48 and 52) This building was located on the north-west side of Site 1 and consisted of remnants of a semi-circle of flat limestone slabs with an internal diameter of 7.6m. The remainder of the structure lay outside the excavation area. The building was located directly east of the Late Iron Age roundhouses excavated by Southampton University in the 1990s and was approximately 5m away from their House 2 (Graham et al, 2002, 6).

The oven/dryer (Figs 53 and54) A stone-lined feature (364) appeared to be attached to the building on its north-east side. Built against the house wall (337) it measured 1.4m in length, was 0.4m wide with an excavated depth of 0.13m. Its edges were defined by small rectangular blocks of blue (possibly burnt) and white limestone (424) which were incorporated into a clay loam and were set at an angle into the ground. Evidence of burning, in the form of reddened soil, was apparent both outside and inside the feature (422). Three distinct

The roundhouse (with no foundations) had been constructed over Late Bronze Age/Early Iron Age midden deposits and survived as a single layer of internal and external limestone slabs with a core of rubble and dark clay loam (337). The wall was best preserved to the north where it was approximately 1m wide. The external face was well-

Figure 52 Late Iron Age House 3, Site 1. View from the north

48

The Middle and Late Iron Age

Figure 53 Oven/dryer 364 sections and plan

fills were recorded. The uppermost (399) was a thin layer (0.03m deep) of brown-red, very friable loam containing limestone crumbs and red clay fragments; larger lumps of clay were recorded near the house wall and on the edge of the feature. Underneath this was a second thin layer (421) of red clay 0.04m deep. The majority of the fill comprised a reddish clay loam (422) containing small limestone fragments and nine sherds of pottery weighing 110g. The blue limestone, burnt clay and reddened soil indicated that this structure had been used for activities involving heat. There was however no evidence for in-situ burning; the structure may have been used as a dryer or smoker where the heat was introduced in the form of hot stones or embers. The pavement Areas of paving consisting of horizontally laid, large limestone slabs (303 and 314) were located south of the building and were likely to have been associated with its use and that of pit 304 (described below). They were recorded as two elements but it is probable that they once formed a single entity. During the 2007 evaluation, a small area of four flat limestone slabs (303) was cleared around the western side of pit 304. This measured 1.75m long and 0.75m at its widest; the largest slab was 0.75m long by 0.75m wide. In 2008, a further area of limestone slabs (314) was uncovered to the west of 303 with a surviving length of 4m and a width of 2m. Similarly sized slabs had been used in its make-up.

Figure 54 Oven/dryer 364 attached to the north-east side of House 3; burning evident as red soil. View from the north

49

Multi-period Occupation at Football Field, Worth Matravers, Dorset During clearance over the slabs, 68 pieces of pottery (701g) and 57 animal bones were recorded. In addition, the layer directly over the stones yielded 55 sherds of dateable Late Iron Age pottery (705g) and 538 sherds (5671g) of Late Iron Age/Roman pottery which consisted of plain body sherds.

to approximately one quarter of the structure. The best preserved section (1194) was a 2.2m curving length of wall measuring 1m wide, with three courses surviving. The outer faces were roughly rectangular blocks of limestone with an internal core of limestone rubble set in a dark brown clay loam. There was no foundation trench, and the stones were placed directly onto the Late Bronze Age/Early Iron Age cobbled surface 1039. There was no evidence for mortar bonding. The internal diameter would have been about 7.5m.

The paving was probably part of a pathway or yard which appears to have gone out of use by the early Roman period when rubbish began to accumulate on its surface. Some of the flagstones may have been used to construct the Roman barns which were built immediately to the west.

The second section of wall (1195) measured 3m long with a maximum width of 1.25m. It was fragmentary, surviving as tumbled limestone blocks, mostly horizontal but some were vertically placed. This part of the wall had been badly disturbed, perhaps by ploughing. This section was also laid directly onto the cobbled surface. The construction and

House 4 (Figs 49 and 55) Located in the centre of Site 2, two discrete remnants (1194 and 1195) of the foundation courses of a probable roundhouse wall survived to a depth of 0.2m and amounted

Figure 55 Late Iron Age House 4, Site 2; foundation 1194 visible. View from the east

50

The Middle and Late Iron Age width of the walling is very similar to that of Southampton Roundhouse 1 (Graham et al, figs 1.5 and 1.6). No internal features were recorded. A total of 2038 sherds of Late Iron Age/Early Roman pottery weighing 17,732g was collected from the various slot fills and must be related to activities in and around the structure. An area of paving (1037) lay outside and to the south west of the house (Fig. 56), comprising six large limestone slabs with a surviving length of 1.9m and a width of 1.25m. The slabs were laid directly onto the cobbles and were probably part of a yard outside the building. This paving was partially underneath (and below the foundation level of) the later Roman barn (discussed in Chapter 5). The construction of this building over part of the house no doubt removed structural evidence, and is highly likely to have utilised a ready and convenient source of building stone. A further small area of paving (1289) was recorded in the rectangular cutting on the north east corner of the site and is also likely to relate to this phase of activity.

Age/Early Iron Age Type 1 pits described in Chapter 3 and probably had similar storage functions. Pit 304 (Figs 57-60) Sited on the southern side of Site 1, this pit was located during the EDAS evaluation in 2007 and subsequently completely excavated; unfortunately the only field records consisted of digital photographs and a partial section drawing. The upper part of the section was drawn, but the lower deposit was not. This was a great pity as the material placed in the pit after it went out of use would appear to have had enormous significance at the time of its deposition (see Lyne, Chapter 10 and Randall, Chapter 14).

Type 1 pits

The pit was initially identified by a semi-circle of small, flat limestone slabs in Evaluation Trench 3. The trench was extended to allow excavation of the pit which measured 1.76m in diameter and had a depth of 1.45m. The vertical sides were formed from small pieces of tightly packed, tabular limestone and were very carefully laid. Large vertical slabs were arranged on the lower part of the pit on the northern side (Fig. 59) but on the southern side, small stones continued to the base of the feature which was formed from the natural horizontally bedded limestone (Fig. 60). On the western edge of the pit, large limestone slabs (303) appeared to form a pathway.

Four Type 1 pits were recorded: three on Site 1 (304, 316 and 395), and one on the edge of Site 2 (1740). In construction, they are virtually the same as the Late Bronze

The fill was given a single context number (305) and was described on the section drawing as ‘plough soil’. This seems unlikely; it must have corresponded to fills found

Pits and Scoops Fourteen pit-type features which conformed to the typology set out in Chapter 3 were recorded on Sites 1 and 2 (Figs 48 and 49). One was dated by radiocarbon and the remainder by their artefact assemblages.

Figure 56 Paving 1037 associated with Late Iron Age House 4, Site 2. Note the Roman barn wall in the side of the excavation trench. View to the south

51

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 57 Late Iron Age Type 1 pit 304, section

Figure 58 Late Iron Age Type 1 pit 304 during excavation, view to the south

52

The Middle and Late Iron Age

Figure 59 Late Iron Age Type 1 pit 304 after excavation. Note the large upright slabs in this half of the pit and pavement 303 to the left. View to the north

Figure 60 Late Iron Age Type 1 pit 304 after excavation, view to the south

53

Multi-period Occupation at Football Field, Worth Matravers, Dorset in other similarly-dated features on the site. This was generally a dark brown clay loam.

represents an earlier feature which was later enlarged. The eastern side was lined with thin, horizontally placed limestone blocks (1743) to a depth of 0.75m. There was no evidence for lining on the lower wall of the pit; however any lining stones here could have been vertical and may have fallen into the fill. There were no lining stones on the western side of the pit. The natural limestone bedrock had been utilised for the pit base.

Very large numbers of finds were retrieved including 2316 sherds of pottery weighing 49,230g; much of this comprised complete or almost complete vessels (see Figs 173-7). A handle from a Catalan or Pascual amphora was also recorded. Large quantities of animal bone were present (744 pieces of weighing 8559g) as well as five fragments of adult human bone and one infant bone; Randall (Chapters 12 and 13) discusses the animal and human bone assemblages and their significance. Five shale objects were recorded including bracelet fragments, a lathe core and rough-outs. From the photographs and section drawing, it can be seen that substantial quantities of large limestone pieces were incorporated into the fill. These were neither recorded nor kept.

All three fills comprised dark brown sticky clay loam. The upper fill (1744) was characterised by small pieces of limestone and a lack of artefacts. The middle fill (1741) was characterised by the presence of some large chunks of limestone and a high density of charcoal, some of which was present in thin layers. A ceramic spindle whorl (SF 512, Fig. 195.9) was located on the top of this layer. Other finds included a quern fragment (SF 511), eight shale cores, 92 pieces of animal bone and 19 sherds of pottery weighing 328g. The lowest fill (1742) was also charcoalrich; a discrete layer covered the pit base. Limestone chunks were abundant. Finds included a stone pounder (SF 502), nine fragments of animal bone and five sherds of pottery weighing 17g.

Pit 1740 (Fig. 61) This pit (noted in the trench side) was located during work associated with the digging of a service trench in September 2011 on the extreme southern side of the site. It was approximately 5m east of Site 2 (see Fig. 220). The description is based on information gathered when cleaning the section. It is presumed that the trench cut through the pit approximately off-centre. Its internal diameter was 1.25m and the depth from the top of the stone lining was 1.1m. The pit was vertically sided on the east, but the lower part of the pit on the west appeared to be stepped. It is possible that the lower part of the pit

Two further Type 1 pits were identified on Site 1. Pit 316 was 1.5m in diameter, but no excavation was undertaken and no site records were made. A number was allocated for the fill (317) and a number of finds including small amounts of Late Iron Age pottery, animal bone and limpets were retained. Pit 395 was partially excavated. It was 1.7m in diameter, had vertical sides and was excavated to

Figure 61 Late Iron Age Type 1 pit 1740, section

54

The Middle and Late Iron Age 0.11m, and deeper features with depths varying from 0.2m to 0.3m. The shallower features were designated ‘scoops’. Six of these are illustrated on Figure 63.

a depth of 0.38m. The fill (333) was a dark brown clay loam containing lumps of clay and large amounts of apparently dumped limestone blocks. No section drawing was undertaken and no finds were recorded.

Pit 1131 (Figs 50, 51 and 64)

Type 2 pits (Fig 62)

This pit, located just to the south of House 4, was initially identified by an oval patch of charcoal-rich soil on the cleared surface. A single radiocarbon determination from a cattle scapula returned a date of 203-45 cal BC (GU37937) at 95% probability. The excavated feature measured 1.68m by 1.25m and was 0.3m deep with shallow sloping sides lined with a layer of small limestone slabs (1146). There were two layers of upright stones on the top of the pit. Between the two layers of stones, the soil contained very small chips of stone, shale, charcoal and infrequent burnt grain. Two fills within the stone lining were recorded. The upper fill (1132) was 0.15m deep and comprised orangeflecked, very dark brown loam, with a high percentage of charcoal and small limestone fragments. This fill also contained Late Iron Age pottery, bone fragments and burnt daub.

Two Type 2 pits were recorded on the northern half of Site 1; both were half-sectioned. Pit 324 Pit 324 was 0.7m in diameter and had a depth of 0.55m. The sides were very steep and the base was difficult to determine, but the section drawing indicated that it was rounded. The single fill (339) was a dark brown clay loam containing tabular limestone fragments. Finds were recorded throughout the fill and included large quantities of pottery (213 sherds weighing 1435g) and animal bone (100 pieces). In addition worked shale fragments were found. Pit 327

The lower fill (1134) was 0.15m deep and was composed of dark brown, soft crumbly loam with moderate amounts of blue limestone 0.05m to 0.08m in size. At the top of this fill a cattle scapula had been placed slightly off centre in the pit, with the distal end pointing north. Larger pieces of limestone were recorded at the base of the fill and probably represent part of the stone lining which had been disturbed. This fill contained Late Iron Age pottery and bone. The pit was on top of and cut into the earlier Middle Iron Age pit 1182.

Oval pit 327 measured 1.3m by 0.8m and was 0.32m deep with steeply sloping sides and a flat base. The fill (344) comprised dark brown very sticky clay loam with frequent blocks of limestone and finds distributed throughout. Twenty four sherds of pottery weighing 145g were recorded together with seven pieces of animal bone and small scraps of shale waste. Type 3 pits Eight Type 3 pits were located on Site 2 and were of two distinct types: very shallow with depths from 0.05m to

Figure 63 Late Iron Age Type 3 pits 1209, 1408, 1483, 1532, 332 and 1549, sections

Figure 62 Late Iron Age Type 2 pits 327 and 324, sections

55

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 64 Late Iron Age Type 3 pit 1131 with cow scapula in section, view to the south

a few small limestone fragments. Three sherds of pottery weighing 35g and two small animal bones were recorded.

Pit 1209 (Fig. 65) Located on the extreme eastern edge of the excavation area, only half of this pit was available to section. The feature was probably oval with a width of 0.52m, measuring 0.4m to the baulk; the total length must have been about 0.8m. The depth was 0.26m. The pit had very steep sides and a flat base and was cut into natural limestone on the north and natural yellow clay to the south and west. The fill (1210) was a very dark brown clay loam which contained charcoal fragments and degraded fired clay. Limestone slabs were set at an angle on the southern side of the pit. Four large jar sherds were lying on the top of the pit and in total 28 sherds weighing 332g were retrieved from the fill. Two animal jaw fragments had been placed on the base of the pit and a further 26 pieces of bone were randomly distributed within the fill. Part of a bone needle (SF 212) was recorded.

Scoop 1435 (not illustrated) This scoop, with shallow sloping sides and a flat base, was cut into the midden deposit. A large block of limestone, which was probably a later intrusion, masked the eastern edge of the pit which measured approximately 0.4m in diameter and was 0.1m deep. The mid-brown loam fill (1436) contained a single rim sherd. Scoop 1483 Sited to the west of House 4, this small sub-circular feature measured 0.57m by 0.5m with a depth of 0.07m. It was underneath a layer of possible opus signinum (1476) which may have been part of the later Roman barn’s floor. The top of the scoop comprised a layer burnt limestone slabs (1489), and the feature itself was edged by small limestone pieces. The centre fill (1485) was black, ashy, charcoal-rich loam with a few pieces of degraded limestone. The base fill (1487) consisted of a thin layer of raw yellow clay and degraded limestone. No finds were retrieved from the feature, but its position underneath part of a floor, and the fact that it was constructed above the cobbles, indicates that it probably pre-dates the Roman barn. It may have been the remnants of a hearth.

Pit 1408 This circular pit, located on the northern side of Site 2, measured 0.8m in diameter and was 0.2m deep. It was defined by limestone blocks on its outer edge and had steep, shallow sides and a flat base incorporating the cobble layer 1039. The fill (1409) varied from dark to midbrown loam and contained small lenses of charcoal and 56

The Middle and Late Iron Age

Figure 65 Late Iron Age Type 3 pit 1209, view to the east.

Pit 1532 (Fig. 66)

Scoop 1549

Located on the extreme western edge of Site 2, the pit measured 0.65m in diameter, was 0.3m deep with steeply sloping sides and a rounded base and was cut into natural yellow clay. The fill was a fine, charcoal-rich, dark brown/ black loam. In places, the sides of the pit were bright red indicating in situ burning. Many large limestone blocks, some of which had been burnt, had been thrown into the pit. Finds included five sherds of pottery weighing 35g and six animal bones including part of a leg bone which was placed on the base of the pit.

This almost circular scoop, which was cut into clay, measured 0.45m by 0.4m and was 0.11m deep. The sides were steeply sloping and the base was flat. The dark brown loam fill (1550) contained minute fragments of decomposing red clay throughout, with a central deposit of charcoal. Four sherds of pottery weighing 14g and three pieces of fired clay were retrieved from the fill. Posthole 332 Located on the northern half of Site 1, this large, circular posthole measured 0.4m in diameter, had vertical sides and was 0.26m deep. Two fills were noted; a central deposit (349) of very dark, almost black loam contained small fragments of red, decaying pottery and four pieces of animal bone. The surrounding fill (350) was a mid-brown clay loam with possible packing material of limestone pieces, some of which were burnt. Also present were two sherds of pottery weighing 10g and four pieces of burnt clay. The fill 349 may represent a decayed wooden post.

Scoop 1528 (not illustrated) This feature was cut into clay and was defined on the surface by burnt limestone pieces and pottery fragments and measured approximately 0.25m in diameter, was 0.08m deep with shallow, steep sides and a rounded base. The mid-brown fill (1529) was a very stick clay loam with occasional fragments of charcoal. The 13 sherds of pottery (weighing 66g) were from a single vessel. 57

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 66 Late Iron Age Type 3 pit 1532 and stone deposit 1531, view from the north

Deposits

pieces of shale were embedded in the soil matrix which bound the stones together. The deposit can be seen to the south of pit 1532 on Fig. 66).

The deposits are categorised by the fact that the material appears to have been deliberately discarded or placed directly on a surface. Two such deposits were noted on Site 2 (for location see Fig. 49).

Burial 318 (Figs 48 and 68) A single grave was located on Site 1 about 1m to the northwest of pit 316. No site records were maintained for this feature; measurements were calculated from photographs. As there was no exhumation licence in place, the skeletal material was left in situ. The small oval grave cut (318) measured approximately 1m in length by 1m wide and had a maximum depth of 0.15m. Six sherds of pottery were retrieved from the grave fill (319). A probable female (320) was laid in the grave on its left side with the head to the east, arms bent in the chest region and legs folded at the knees. The disposition of the skeleton suggests that this inhumation is Late Iron Age in date, but this method of inhumation continued into the Roman period.

Deposit 1019 (Fig. 67) Located on the extreme southern edge of Site 2, this deposit comprised a spread of pottery which was uncovered during the first week of the excavation and given the slot number 1019; at the time this was not recognised as a discrete spread of material. In an area measuring approximately 1m by 1.2m, 77 sherds of pottery (weighing 406g) were embedded in sticky dark brown loam which also contained small pieces of limestone and 23 animal bone fragments. The material is likely to have been laid down in a single episode. Deposit 1531

Conclusions

Located immediately south of pit 1532, this feature comprised a deposit of horizontal and vertical stones which were set into the cobble layer (1039). When dismantled, three sherds of pottery (including two rims) and three

Only one securely dated feature is of the Middle Iron Age: it is possible that the focus of any activity in this period was located to the south, outside the excavation area. The 58

The Middle and Late Iron Age

Figure 67 Fragmented pottery deposit 1019, view to the west

Figure 68 Grave 318, with skeleton 320, view to the south

and must have been used initially for storage, probably of grain. Their final use was as rubbish dumps. The three burials (two found in the 1990s, see Chapter 16) imply that although the dead were buried with care, their graves were incorporated into the immediate surroundings of the living.

Late Iron Age settlement activity located by Southampton University is now seen to extend over a much larger area bringing the total number of houses to four; however they may not all have been in use at the same time. Very large deep, stone-lined pits were constructed near to the houses 59

5 The Romano-British Period Introduction

Three postholes (1495, 1470 and 1542, were aligned in an approximate south-west/north-east direction and were south-east of posthole 1515. None of these features cut the cobble layer.

Romano-British features were located on Sites 1, 2 and 4, and consisted of structural remnants of a stone building within which were buried skeletal remains of infants; pits and other deposits were also excavated. Much artefactual material not specifically related to features was recorded in the clearance layers of the excavations on Sites 1 and 2. During the 2007 evaluation, Roman material was recorded in Trench 4 (Fig. 5). The Roman occupation (based on the pottery and metalwork) spanned 400 years. Figure 69 illustrates activity on Site 2: only the western half of the site is depicted as there were no Roman features on the eastern half. A single feature was recorded on Site 1.

Posthole 1495 This circular posthole was 0.26m in diameter, had steeply sloping sides and a flattish base and was 0.17m deep. The fill (1496) of dark brown, very sticky clay loam contained small limestone fragments. Part of a partly perforated ceramic spindle whorl (SF 392) was recovered. Posthole 1470

The Building

The central posthole was 0.2m in diameter and 0.17m deep, funnel-shaped and defined by three thin vertical limestone slabs which must have formed part of the packing material. The upper sides sloped steeply and the lower sides were vertical. The base was flat. The dark brown fine loam fill (1471) contained 11 fragments of animal bone.

When Site 2 was stripped in 2010, it was apparent that archaeological remains of densely packed rubble debris (Fig. 70) survived just below the ground surface. Unfortunately, some of this material was removed before its archaeological significance was realised. It is likely that it was part of a collapsed wall of a stone building. The uncovering of a large padstone (1008) (Fig. 71) gave credence to this observation. The rubble spread (1003) diminished to the east of the site. Large rectangular slabs (1018) were apparent in the rubble and appear to have formed part of the building line. Excavation in 2011 uncovered a similar sequence with the rubble spread covering further structural elements. As the excavation progressed it was possible to determine the outline of a building. By scrutinising the plans, associated features and finds it was possible to reconstruct the development of this structure.

Posthole 1542 The eastern-most posthole was 0.3m in diameter and 0.12m deep, with steep sides and a flat base. The midbrown loam fill (1543) contained rare, small limestone fragments. Three upright limestone slabs were recorded and were probably part of the post packing. Phase 2: building construction A building, of which the partial ground plan was revealed (Fig. 69), was constructed without foundations directly onto the upper levels of the clay loam midden and was approximately 0.3m above the Late Bronze Age/Early Iron Age cobble surface 1039. It is likely that standing structural elements of the Late Iron Age roundhouse were plundered and utilised as building material for the Roman structure.

Phase 1: post-built structure (Fig. 69 plan, Fig. 72 sections) A single dated posthole (1515) pre-dated the building sequence. Further undated postholes 1470, 1495 and 1542 probably belong to this phase. The postholes may have related to a post-built structure.

The excavated limestone elements consisted of wall fragments (1018, 1178, 1397, 1398, 1399, and 1450), a doorway threshold (1419 and 1420), a padstone (1008) and a large stone of uncertain function (1405). A very large paving slab (1426) had been laid outside the doorway, and numbers of external paving slabs (1488) survived to the north. Scattered stone roofing tiles were recovered in the rubble layer. The building was better preserved on the western side, probably because it lay underneath the drystone boundary wall of the field. This latter feature must be of some antiquity as Roman rubble and artefacts lay directly underneath.

Posthole 1515 This was located under the very large paving slab (1426). The posthole was 0.24m in diameter, 0.2m deep with steeply sloping sides and a flat base and cut through the cobble layer (1039) into natural clay. The fill (1516) of mid-brown clay loam contained charcoal fragments. Finds of pottery (31 sherds weighing 165g) and animal bone were concentrated towards the base. Limestone slabs lay on top of this feature. 60

The Romano-British Period

Figure 69 Site 2, Roman building and associated features, plan

building 4 excavated by Southampton University (Graham et al 2002, 21) was rectangular with external dimensions of 6.2m by 13.5m. The Site 2 building may have been of a similar size. One doorway was recorded in the western wall

The long sides of the building were aligned south-west to north-east and its external width was approximately 5.6m; there was no evidence for corners within the excavation area, its length therefore must remain conjectural. However, 61

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 70 Site 2 after stripping and cleaning in 2010, showing the rubble debris (1003) and structural remains of the Roman building. View to the north-east

Figure 71 Padstone 1008, was moved by mechanical excavator during the initial strip of the site. View to the south west.

62

The Romano-British Period section (1399) survived and consisted of three rectangular blocks 0.5m to the south-west of the doorway. The doorway threshold on the north-west facing wall was very well preserved (Figs 74-76) and consisted of two substantial rectangular stones (1419 and 1420) fitted together and measured in total 1.48m long by 0.3m wide. The centre of each stone had been incised with square holes which may have taken door bolts; the function of the linear cuts is not clear. Outside the threshold was a huge sub-rectangular slab of limestone (1426), (Figs 74 and 75) measuring 2m long with a maximum width of 1.2m and a thickness of 0.08m. This slab had cracked in antiquity. Underneath was a layer of small levelling stones. The slab was laid close up against the threshold. Again, a parallel can be seen in a similar arrangement outside the Southampton building (ibid fig. 1.16). Adjoining the threshold was an irregularly shaped padstone 1415 (Fig. 77) with two square-cut holes with depths of 0.05m and 0.07m and another smaller hole. The larger square holes presumably held the base of a timber door frame. If the arrangement was the same as Southampton’s Building 4 (ibid, fig. 1.16), it seems that padstone 1415 had been displaced and would have originally have been placed at right angles to the threshold. There was no matching padstone next to threshold stone 1419. A further padstone (1008) (Fig.71) was moved by the excavator during site stripping. Its original placing is unfortunately lost; however, it is not likely to have moved very much and perhaps hints at a further doorway in the south-east facing wall. A large fresh sherd from a New Forest Parchment Ware bowl of Fulford type 67 with a date range of c. 270400 AD came from occupation context 1443 which lay against the threshold stones 1419 and 1420, giving a broad date for the use of this building.

Figure 72 Features pre-dating the Roman barn, sections.

and a further opening is hinted at by the presence of the padstone (1008), on the eastern side of the building albeit not in its original position. The lengths of walling 1398 and 1450, at right angles to the main building, suggest that an ancillary structure was attached on the western side and further structural elements may be preserved underneath the remaining boundary wall. Very large quantities of shale waste including lathe cores and bracelet fragments were found within the confines of the building and it is possible that it may have functioned as a manufacturing workshop (see Cullinane, Chapter 7). Description of the building elements Only fragments of the opposing long walls survived. They were laid directly on top of the midden deposit and comprised the first course of building stone; there was no evidence for a bedding trench. Three sections of remnant wall were recorded on the south-east facing wall. The best preserved 1018, consisted of a single course of six large, flat stones forming a wall 0.7m wide. This is the same width as the Southampton Building 4 (Graham et al, 2002, figs 1.15 and 1.16), although the construction technique there was slightly different and was generally only a single slab wide. Section 1397 was located 2m to the north-east (Fig. 73) and consisted of a 1.2m length of wall with a maximum width of 0.7m and comprising seven blocks. The longest (and thinnest) of these had been disturbed by the mechanical excavator during the site strip and was probably from a second course of walling. Section 1178 on the south edge of the site continued through the site boundary and may be preserved in the garden of the adjacent property (see Fig. 56). Here, two faces of wall 0.8m wide were recorded and consisted of well-shaped rectangular blocks. From the photograph it can be seen that some of these blocks appear to have tumbled into the building. On the north-west facing wall, only one small

Masonry structures, which may have been part of the building, survived by the north- west wall on the northern edge of the site and comprised two sections of stonework. The best preserved (1450) was constructed at right angles to the main building and consisted of a 1.5m length of limestone blockwork laid on loam with a loam infill. Three courses survived (Figs 78-80). On average the width of the wall was 0. 55m (narrower than the wall of the main building) and it was constructed of single and double blocks laid side by side. There was an approximate 1m gap between this length of wall and further fragment 1398 which was only partially exposed, the remainder being outside the excavation area. The two walls may have formed a corridor. A later pit 1465 (discussed below) may have disturbed the integrity of these remains. Features Associated with the Building Shaped limestone Within the building, and located only 0.2m from threshold 1420, was a large triangular shaped piece of limestone (1405) (Fig. 81) described in Chapter 6 (Fig. 120.10). Its 63

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 73 Length of wall 1018 in the foreground and elements of 1397 exposed beyond. View to the north-east

smallest face was uppermost. The stone had been re-used in the building of the wall and when found, a circular depression 0.2m in diameter and 0.05m deep, was apparent on its base. Its original function can only be conjectured, but the carved hollow must originally have been uppermost.

building with raw yellow clay; this was also present on the west side of the box and initially completely covered the top of it. The lid had slipped inwards and was covered by the clay. The box was rectangular measuring 0.5m by 0.3m and was 0.2m deep with a base formed of small limestone pieces and its construction cut through the underlying cobble layer 1039. Several of the upright stones had slipped inwards, as had the lid, which comprised two similar stones. Upon removal of the clay, which almost completely filled the box, and an underlying thin layer of

Limestone cist A small stone box (1123) formed from thin limestone slabs was bonded to the north-west facing wall (1018) of the 64

The Romano-British Period

Figure 74 Door threshold 1419 and 1420 and large porch slab 1426, section through the Early Neolithic ditch 1332 visible centre left. View to the south

Figure 75 Threshold elements of the Roman building, plan

65

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 76 Threshold stones 1419 and 1420 with bolt holes and grooves. View to the south

Figure 77 Padstone 1415 with holes for wooden uprights, view to the west

66

The Romano-British Period

Figure 78 Walls 1450 and beyond, 1398, with remnants of paving 1488 and lid SF 312 in the foreground. View to the north-east

loam, an infant skeleton (1174) was revealed lying against the western side of the box (Figs 82 and 83). The remains are discussed further in Chapter 12. A small nail (SF 195) was found in the loam layer. It is not clear whether the box was an integral part of the building, or a later insertion as a burial cist. As no floor levels survived, it is not possible to determine the box’s position in relation to the floor surface of the building. Further infant burials were placed and scattered within the building and are discussed below. Paving Scattered limestone slabs within the building may have been remnant flooring, but the best preserved area of paving (1488) lay outside and adjacent to the large slab 1426. Incorporated into this was a flat, circular stone (Fig. 119.8) which may have been used as a lid. Areas of crushed limestone Remnants of two possible floor areas (1431 and 1476) were recorded within the building (Fig. 69). The remains varied in thickness from 0.04m to 0.1m and comprised irregularly-shaped areas of pink/cream crushed limestone and burnt clay similar to, but softer than opus signinum cement.

Figure 79 Wall 1450 (centre) showing construction and wall 1398 in front. View to the south-west

67

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 80 Wall 1450, section

Gullies

infant burials discussed below were disturbed after burial and during the life of the building.

Two short gullies were recorded, one inside the building (1417) near threshold stone 1420 and the other (1454) outside, by the large slab 1426. Gully 1417 was 0.8m long with a width varying from 0.05m to 0.15m and was consistently 0.15m deep. It was filled with a soft dark brown loam which contained 21 sherds of pottery (weighing 318g), among which were sherds from a handled beaker and a seated-lid pot. A single fragment of infant bone demonstrates that the

Located to the north of slab 1426, and running in a southeast/north-west direction, a short length of gully (1454) and an associated sump (1451) were defined by small, vertical limestone slabs (Fig. 84). The gully may have been constructed to take water from the large porch slab and away from the building. The circular sump was in the centre of the gully and measured 0.30m in diameter and

Figure 81 Triangular stone 1405 with pecked circular depression

68

The Romano-British Period

Figure 82 Burial cist 1123 with infant skeleton 1174 on the base. View to the south

Burials Associated with the Building

had a depth of 0.1m. The sides were near vertical and the base was flat. Pottery from the fill comprised 55 sherds weighing 324g and included a fragment of samian. The gully fill (1455) was a very dark brown soft loam also containing pottery (31 sherds weighing 446g) as well as a lathe core (SF 460). The northern end of the gully was covered by two large paving slabs.

In addition to the cist burial, nine almost complete infant skeletons were located in the central section of the building. No grave cuts were visible. Scattered remains were also recorded in two pits. An estimated minimum of 20 infants was interred in the building. Deposit 1025 comprised three infants (Fig. 85), placed in the ground possibly at the same time, in flexed positions. The skull of one infant rested on a small slab of limestone and the inhumations lay about 0.1m above the cobbles. The bones had been slightly disturbed. A further three burials (1092, 1093 and 1262) had been placed close by, their heads in a west, north and west position respectively. The bodies had been interred at a slightly deeper level. A large amount of pottery (120 sherds weighing 789g) was found in the vicinity of burial 1262. Burial 1090 was placed near the north-west facing wall, with its head facing north; the remains were laid directly onto the cobble surface. On the other side of the building, disturbed burial 1440 had been placed by the side of and

Figure 83 Infant remains within cist 1123, plan

69

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 84 Sump 1451 and gully 1454, view to the north-east

Figure 85 Burial group 1025 comprising three infants, view to the west

70

The Romano-British Period in the centre of the threshold (Fig. 86). This may have been a foundation deposit, placed when the building was constructed.

to be a dumped and slumped deposit which accumulated after the building went out of use. Pits associated with the building (Fig. 87)

The remainder of the infant remains were scattered between pit 1421 and deposit 1443 (described below). Bones from a minimum of three infants were retrieved from the pit 1421 and collectively given a deposit number 1424. The digging of the pit had probably disturbed the burials. Deposit 1443 contained the remains of a further infant (1445). The infants appeared to have been interred during the lifetime of the building and were likely to have been offspring of the site’s inhabitants.

Type 3 pit 1184 This small pit measured 0.5m by 0.42m and was 0.11m deep with shallow sloping sides and a slightly rounded base. The sides and base of the pit were partially lined with yellow/orange clay; fill (1185) was a very dark grey, charcoal-rich loam. A number of objects were retrieved including slag, iron objects (SF 205 and SF 206), shale debris, seven pieces of animal bone and six sherds of pottery (weighing 92g) including a rim from a flanged bowl.

Pits and Deposits Seven pit type features can be dated to the Roman period (1184, 1421, 1435, 1438, 1465, 1749 and 1751). Five were associated with the building on Site 2; two were located in the water trench on Site 4. There were no Type 1 pits. Two pits fell into the Type 2 category (1749 and 1751), both on Site 4; three (1184, 1438 and 1421) were Type 3 and all of these were categorised by dense charcoal fills. Although described as a ‘pit’ in the record, feature 1465 appeared

Type 3 pit 1438 Located on the northern side of the excavation, pit 1438 measured 0.6m long and 0.4m wide with a depth of 0.2m. The pit had steeply sloping sides and a rounded base. Clay was noted on its lower sides and base (which was covered by a thin layer of fine grey silt). The fill (1439) was an

Figure 86 Infant burial 1440 underneath threshold stones 1419 and 1420. View to the west

71

Multi-period Occupation at Football Field, Worth Matravers, Dorset floor deposits 1431 and 1476. The pit therefore was later than and cut through the floor. Deposit 1465 Initially recorded as a pit, this feature appeared to be composed of material which had accumulated to a depth of 0.5m between the sides (1450 and 1398) of the remnant ancillary building. Four layers were recorded; the uppermost (1441) was a 0.2m dark brown loam defined by a spread of red deer antler bone on the surface (Fig. 89, SF 238). In addition there were large slabs and lumps of limestone, which had the appearance of either being dumped or fallen. Fragments of shale and 67 sherds of pottery weighing 551g were recorded. The second layer (1446) was dark brown/ black loam to a depth of 0.1m containing fewer, smaller pieces of limestone. The finds were similar to the upper fill but fewer. Layer 1447 was 0.12m of dark brown loam with discrete deposits of limestone, particularly on the south-west side. Large numbers of finds were recovered including an iron stud (SF 318) and nail (SF 317), a ceramic counter (SF 316) and burnisher (SF 488), a shale lathe core (SF 345), 51 fragments of animal bone and 53 sherds of pottery weighing 404g. Two layers were recorded on the base of the pit. To the east, 1457 comprised dark brown/ black loam from which several pieces of tabular limestone were retrieved including one roof tile (SF 346). Other finds included pottery (11 sherds weighing 152g), fragments of fired clay, and shale pieces one of which was part of a bracelet (SF 347). Merging into this fill on the west was a dark brown loam (1458); limestone was almost absent from this layer. Among the finds were a single hobnail (SF 360), 81 pieces of animal bone, shale including part of a bracelet (SF 349) and 25 sherds of pottery weighing 230g. Animal bone was spread across the two basal contexts 1457 and 1458; much of this consisted of cat remains which had been disturbed in antiquity. This deposit represented the largest group of cat bone from the site. The number of loose teeth from all contexts implied material which had moved around a great deal, having either lain on the ground surface for some time or had been redeposited.

Figure 87 Roman Type 3 pits 1438, 1184, 1421 and accumulated deposit 1465, sections

almost black loam, resulting from a heavy concentration of charcoal. One sherd of pottery, a cattle tooth and a large piece of worked flint were retrieved. Type 3 pit 1421 This irregularly-shaped feature was located just inside the south-east facing wall of the building. It measured 1.5m long and was 0.95m at its widest with a depth of 0.28m and was edged with small flat stones. The sides sloped gently and the base utilised the cobble layer 1039. The fill of dark brown, soft, crumbly loam (1422) contained a high percentage of degraded limestone; at least half of this was burnt red. The lower sides of the pit were red, indicating burning in situ. Small fragments of charcoal were recorded all through the fill, but there was a particular concentration in the centre and just above the base of the pit. Finds included large amounts of pottery (323 sherds weighing 3764g), four iron nails, fired clay, shale fragments, 26 pieces of animal bone (mainly sheep/goat and cattle), and the scattered remains of a minimum of three infants which were dispersed within the fill and on the base of the pit. It is likely that the digging of this pit had disturbed the burials. Just underneath the top of the pit were discrete deposits of crushed/burnt clay/limestone; this was identical to the

Spreads A number of spreads of material were recorded and probably represent separate episodes of disposal both deliberate and accidental. Five discrete deposits of pottery were located on Site 2; four of these were within the confines of the building and are illustrated on Fig. 69. Deposit 1004 A discrete collection of 15 scattered sherds from a cooking jar represent a pot which was broken in situ. Deposit 1443 Located against the threshold stone 1420 was an oval area of dark brown loam measuring 0.35m by 0.25m with depths varying from 0.05m to 0.12m which contained large 72

The Romano-British Period

Figure 88 Antler bone visible on the cleaned surface of deposit 1465, view to the north-east

Other Features

quantities of pottery (249 sherds weighing 2573g), limpet and oyster shells. There were three iron items including a fragment of doorstrap (SF 315, Fig. 190.10) part of a handle (SF 31) and a nail (SF 352).

These comprised pits and deposits located on Sites 1, 2 and 4 and were located in areas which were not directly associated with the building. The pit typology is detailed in Chapter 3.

Deposit 1505 (Fig. 89)

Type 2 pits

Located on the extreme western edge of the site, this substantial pottery deposit consisted of 257 sherds weighing 6585g which were spread over an area measuring 0.7m by 0.7m. Most of the vessel pieces were joined and had belonged to a very large storage jar. Other finds included animal bone, a piece of lead displaying cut marks (SF 391, Fig. 188.1) and a small shale offcut.

Larger pits were absent on Sites 1 and 2, but were recorded in the trenches on Site 4 (Fig. 90) and must have been related to activities taking place in the Roman buildings (Site 2 and Southampton University excavation). Approximately 0.3m of dark humic topsoil (1754) covered both pits. Large limestone slabs were noted in the trench side and were also present in the spoil. These may have been associated with activities centred on the pits.

Deposit 1506 This comprised a spread of 60 sherds of thin black pottery (weighing 362g), fragments from a probable cattle skull and a sea bream vertebra. These were spread over an area measuring about 0.4m in diameter. The deposit contained much crushed limestone (1476) which was part of the spread identified as flooring. The deposit lay on the building line of the structure. The sherds were from a Bestwall form 11/2 tankard and an indeterminate cavettorimmed jar and dated to c. 300-370 AD.

Pit 1749 (Fig. 91, upper) This was the westernmost Roman feature. The pit appeared to be circular with steeply sloping sides and a rounded base. The diameter was about 1.15m and the depth was 0.45m. The fill of sticky dark brown loam (1756) contained randomly-placed large blocks and thin slabs of limestone which appear to have been thrown in from the south side of 73

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 89 Pottery deposit 1505, view to the north

the pit. Finds included shale, nine animal bone fragments and three sherds of pottery weighing 22g.

samian sherds and one New Forest sherd. Large chunks of cattle and horse bone were also present as well as 29 iron hobnails.

Pit 1751 (Fig. 91, lower)

Deposit 1759 (Fig. 92)

This probably circular pit was 1.4m in diameter with almost vertical sides; its base utilised the cobble layer 1039 which extended to the western edge of this pit and no further. The fill (1755) was a dark brown loam containing flecks of disintegrating limestone and randomly spread blocks of limestone. A discrete patch of burnt soil contained 20 pottery fragments weighing 233g. A small amount of shale waste was recovered. The pit cut an undated posthole (1758).

Located on Site 4 at the extreme eastern end of Trench 1748, a small deposit of pottery (11 sherds weighing 115g) was distributed within dark brown loam at a depth of 0.4m from the present ground surface. Two further deposits were recorded on Sites 1 and 2. Deposit 315

Deposit 1752 (Fig. 92)

Located on the south-west corner of Site 1, this deposit was above the Late Iron Age paving 314 (see Fig. 48 for location). The context note is brief and mentions dark, sticky, black soil with charcoal fragments and limestone debris throughout. The site plan indicates an area about 4m by 4m and the depth of soil taken off was about 0.1m.

Located at the eastern end of trench 1748 and within layer 1754, the deposit was a dark brown/black, soft, crumbly layer of loam about 2.5m long, between 0.2m and 0.3m deep, and was 0.25m below the present ground surface. The 19 sherds of pottery (weighing 227g) included two 74

The Romano-British Period

Figure 91 Type 2 Roman pits 1749 and 1751, sections

burial. Otherwise, the deposit has the hall-marks of midden material which accumulated over several hundred years. Deposit 1172 This was located on the extreme eastern boundary of Site 2 and consisted of the lower part and base of a probable jar (32 sherds weighing 462g) which had been placed into the midden fill (Slot 1109). Later ploughing may have truncated the vessel. It is likely to have been associated with activities centred on the barn.

Figure 90 The water pipe trench (Site 4) with recorded features, plan

Conclusions On Site 2, four postholes may relate to a structure which was replaced by a large stone building at the beginning of the 4th century and which may have functioned as a barn. Shale objects and waste imply that it may have been a workshop at some time in its history. Infant inhumations associated with the structure illuminate burial practices at this time.

Finds were numerous and included a small amount of earlier pottery. The Roman material consisted of very large quantities of pottery (259 sherds weighing 3470g) comprising local Black Burnished wares together with samian products which have a date range of AD 43-250 (Lyne, Chapter 10). The only examples of Roman glass were from this area and derived from at least three vessels. Animal bone was spread throughout (64 fragments mostly representing sheep/goat). Part of an adult human limb bone and a bone from a neonate were also noted. Perhaps the most enigmatic finds were those of metal; three pairs of tweezers (Fig. 183.11-13; SFs 9, 15 and 19) and a complete pin (Fig. 182.4; SF 14). It is unusual to find multiple tweezers. There is therefore a possibility that the human bone and metal finds might represent a disturbed

Activity however was not confined to Site 2. Although no discrete features were identified on Site 1, large numbers of Roman artefacts attest occupation very close by and are likely to relate to Roman features excavated in the 1980s. Examination of features in the water trench (Site 4), on the other side of the present road into the village, confirm that activity extended well-beyond the major excavated areas.

75

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 92 Section through Trench 1748 with pit 1751, cobbles 1762 and deposits 1752 and 1759

76

6 The Post-Roman Burials Introduction

the present ground surface and had been badly disturbed by ploughing. A regular grave cut was not apparent, but slabs of limestone were noted around the skeleton, some flat and some upright. The inhumation was in a Type 2 grave (for typology, see below). The stones must have marked a shallow grave which was approximately 1.35m long and 0.65m wide. The skeleton itself had been placed on the cobble layer 1039 which extended into Site 1. The body was orientated south-north with the head (which was not present) at the south end of the grave. Photos show that the upper arms were placed by the side of the body and the left lower arm was extended across the chest (Fig. 93). No grave goods were recorded. A recalibrated radiocarbon date of 630-780 cal AD (SUERC-61172) at 95% confidence places the inhumation probably in the 7th century AD towards the end of the use of the site for burials.

Excavated remains from 27 individuals were recorded during the archaeological work. A single grave was located on Site 1 and 21 graves on Site 3. Initially the individual from Site 1 was thought to be Roman in date but subsequent radiocarbon determination confirmed that all the burials fall into the Post-Roman period (see Krus, Chapter 12). The skeletal material is considered by Randall in Chapter 13. Adult Burial from Site 1 An extended, supine, probable female inhumation (skeleton 311, grave 425) was recorded on Site 1 by brief site notes and photographs during 2007 as no exhumation licence had been obtained. As this burial was threatened by one of the new housing boundaries, a decision was taken to lift it in 2011 when an exhumation licence was in place.

This isolated grave is located about 30m west of the cemetery discussed below. There may be other burials of this date on the site, but currently, Grave 425 has to be considered as an outlier to the main cemetery. It is notable that the grave orientation is south-north unlike the cemetery where the graves are all west-east.

The grave was located just south-east of late Iron Age pit 304 and is illustrated with the Late Iron Age features on Fig. 48. The remains were encountered about 0.3m below

Figure 93 Partial remains of adult female 311 in grave 425, view to the west.

77

Multi-period Occupation at Football Field, Worth Matravers, Dorset The Post-Roman Cemetery (Fig. 94)

excavation area. The eastern extent is unknown. The grave fills were all similar, consisting of a mid-brown crumbly clay/loam containing small limestone fragments. There was no evidence for wooden coffins, either from coffin nails or as stains of decaying wood in the soil. The bodies had probably been placed in the graves in simple shrouds. A single grave good from grave 1667 (a small copper alloy buckle; Figs 184-87) was initially used to date the cemetery to the period AD 500-700. Bone samples from six individuals were submitted for radiocarbon dating and Bayesian analysis and the recalibrated results suggest that the cemetery was in use from cal AD 420-650 to cal AD 650-810 (see Krus, Chapter 12).

A Post-Roman cemetery comprising 21 graves containing 26 individuals was located about 90m from the field entrance (see Fig. 5) and was designated Site 3. It was during ground clearance in July 2011 for the building of the development’s five houses that the first human remains were encountered 0.3m below the present ground surface. A project design was rapidly developed and building work ceased until all human remains were located, recorded and lifted. The site was approximately rectangular, measuring 13m by 17m and was cleared by mechanical excavator and then hand-cleaned. As the burials were located at varying depths and time was limited, the mechanical excavator was used to carefully remove further depths of soil. Seven grave types were encountered and varied from simple earth-cut to complex stone-lined ones. Apart from a single child and three adolescents, the inhumations were adults. There were three double burials, one triple burial, and the remainder were single burials: these were in short north-south rows and the bodies had all been carefully placed with heads at the west end of the grave. The full extent of the cemetery has not been defined; however the west, north and south limits appear to be confined to the

Earlier features An activity area (1700) interpreted as remnants of a Middle Bronze Age pottery clamp (see Chapter 2) was located to the west of the cemetery and small fragments of pottery of this date were spread throughout the soils in this area. Artefactual evidence (prehistoric pottery and worked flint and a few sherds of Roman pottery) from clearance of the cemetery indicates activity in the second millennium BC, with little happening afterwards until the development of the graveyard.

Figure 94 Site 3 plan; the Post-Roman cemetery with grave types, pavement 1695 and Middle Bronze Age activity area 1700

78

The Post-Roman Burials Grave orientation

aligned east-west and burial 1662 inclined very slightly to the north-west. There was a 2.5m gap between this grave and 1737. Row 4 appeared to be carefully laid out with six interments (1654, 1640, 1660, 1643, 1697 and 1715), all on an east-west axis. The three graves (1722, 1633 and 1733) of Row 5 were similarly aligned. The partly excavated grave 1778 was the only one in Row 6; however further graves may lie in the unexcavated area to the east.

The number of graves (21) is too small for statistical analysis however; the following observations can be made. Fifteen graves were laid out in a relatively strict east-west alignment. At the western edge of the cemetery, graves 1667 and 1685 were set out slightly to the northwest and graves 1649 and 1737 were aligned slightly to the south-west. It may be significant that these four graves were on the western extremity of the burial ground. There appeared to be six north-south rows of graves (see Fig. 94 and Table 5), some more carefully laid out than others. There was generally an approximate 1m gap between the rows and the graves varied from being 0.20m to 3m apart. Row 1 contained one grave (1685): this was the triple burial. There were four graves in Row 2 (1667, 1649, 1664 and 1719); Grave 1667 which was not orientated on a true east-west axis contained a grave good. There was a 3m gap between that grave and 1649. There were six burials in Row 3: four (1678, 1646, 1670 and 1675) were

Grave typology Seven discrete grave types occurred within the cemetery (see Fig. 94). The grave typology is set out in Table 5 together with the sex and ages of the inhumations (including the outlier). The skeletal material, which was very poorly preserved, is considered in detail by Randall in Chapter 13. Type 1: Plain earth-cut (four examples 1633, 1667, 1682 and 1715)

Table 5 Grave typology and associated skeletal material. Skeletons in bold type indicate bone material submitted for radiocarbon dating Grave Type Row No

length

width depth Skele- Age ton

1633

1

5

2

0.6

0.26

1632

Adult ?

1667

1

2

1.6

0.53

0.28

1669

25-35 Female

1682

1

3

1.7

0.5

0.1

1684

17-25 ?

1715

1

4

1.5

0.45

0.1

1717

Adult ?

1664

2

2

1.75

0.56

0.1

1666

35-45 ?Female

1670

2

3

1.9

0.55

0.2

1672

17-25 Female

1697

2

4

1.7

0.4

0.25

1699

25-35 ?Female

1660

3

4

1.95

0.7

0.29

1662

25-35 ?Male

1722

3

5

1.8

0.8

0.35

1724

35-45 ?Male

1640

4

4

1.55

0.6

0.15

1642

17-25 ?Female

1643

4

4

1.7

0.4

0.39

1645

35-45 ?Female

1646

4

3

1.7

0.4

0.2

1648

Adult ?

1654

4

4

1.28

0.4

0.2

1656

c.5 years

1778

4

6

0.5

0.1

1779

16-17 Female

1649

5

2

1.7

0.5

0.15

1651

25-35 ?

1675

6

3

2.02

0.5

0.1

1677

35-45 Female

1678

6

3

1.9

0.7

0.2

1680

40-45 Male

1737

6

3

1.9

0.66

0.28

1713

Adult Male

1685

7

1

1.8

1

0.1

1687

25-35 Female

1719

?

2

1721

25-35 ?Female

1733

?

5

425

2

outlier 1.35

1.5

Sex

Sex

1652

17-25 ?

1725

17-25 ?Female

1681

15-16 ?

1688

45-49 Male

Skele- Age ton

Sex

1693

?

?

0.5

0.08

1735

?

?

0.65

0.3

311

45-49 Female

79

Skele- Age ton

15-16

Multi-period Occupation at Football Field, Worth Matravers, Dorset Type 2: Rubble lining or kerb alongside (four examples 1664, 1670, 1697 and outlier 425)

Type 4: Cist-type with split limestone slab walls and occasional lids (five examples 1640,

tiles. There was no evidence for wooden coffins either by the presence of coffin nails or from the staining of decomposing wood. The grave lengths, where measurable, varied from 1.5m to 2.02m, the widths from 0.4m to 0.9m and the depths from 0.08m to 0.39m. All grave cuts appeared to be very shallow and there were no instances of intercutting. All interments were supine with generally very poor bone preservation.

1643, 1646, 1654, and 1778)

Type 1 graves

Type 5: Stone lumps at either side of head and irregular arrangement of slabs/rubble at foot end of grave (one example 1649)

Four Type 1 graves were identified and were typified by the difficulty of locating a grave cut; if visible at all this was only generally apparent as the skeleton was removed. The defining characteristic was that the grave was simply cut into the earth, with no addition of any stone.

Type 3: Stone lumps/slabs at head and foot position only (two examples 1660 and 1722)

Type 6: One vertical stone at either head or foot (three examples 1675, 1678 and 1737)

Grave 1633 (Figs 95 upper and 96)

Type 7: Horizontal slabs on top of grave (one example 1685)

This grave in Row 5 was the first to be located and contained two individuals, skeletons 1632 and 1652. No grave cut was apparent, but by comparing this with other double graves on the site, it was estimated to have been approximately 2m in length and 0.6m wide. The depth was 0.26m. The burials were initially identified by the presence

It was not possible to type Graves 1719 and 1733; the former was disturbed by the digging of house foundations and the latter was composed of fragmentary skeletal elements. All stone used in the grave construction was Purbeck limestone, including some re-used Roman roof

Figure 95 Type 1 Graves 1633, 1667 and 1682, plans

80

The Post-Roman Burials of the skull of skeleton 1632 which had been partially uncovered during the topsoil strip. The bodies had been placed in the grave together. The feet were present, and initially thought to be animal bone, were removed, and not therefore shown on the drawing. It was not possible to sex either individual. The taller skeleton 1632 was an adult and the smaller skeleton 1652 was in the age range 17-25. Skeleton 1632 had probably been placed in the grave first, with the legs close together, and the arms extended and placed over the pelvis. The skull rested on a shaped limestone 1635 (SF 517; Fig. 120.9) and had probably been displaced by ploughing. The stone was subsequently identified as an anchor (see Bellamy, Chapter 7) and may hint at sea-faring activities or links during the individual’s lifetime. The re-calibrated radiocarbon date range is cal AD 640-780 (SUERC-61178). The smaller skeleton 1652 had been placed very close to 1632 and the head, which was very poorly preserved, was level with the latter’s shoulders. The right arm was slightly bent at the elbow and the hand rested in the pelvic area. The left arm was flexed at the elbow and the lower arm rested in the area of the pelvis of skeleton 1632. Grave 1667 (Figs 95 centre and 97) This was the northernmost burial in Row 2. The grave cut was only visible once the skeletal elements (1669) had been exposed; this measured 1.7m long, 0.5m wide with a depth of 0.28m. The body of a probable female aged between 25 and 35 years had been placed in the grave with the head on its side and facing south. The arms were extended by the sides of the body. The feet were missing but as this skeleton was very badly preserved, it is likely that these small bones had decomposed. The grave was unusual in that it yielded the only item of personal ornament. This was a small copper alloy buckle (SF 471, Fig. 184-7) which rested on the pelvis. This is discussed in detail by Hinton and Walton Rogers in Chapter 11).

Figure 96 Type 1 Grave 1633 with skeletons 1632 and 1652, the head if 1632 rests on the stone anchor. View to the south

Figure 97 Type 1 Grave 1667 and skeleton 1669 partially exposed with the copper alloy buckle visible on the pelvis. View to the north

81

Multi-period Occupation at Football Field, Worth Matravers, Dorset Grave 1682 (Fig. 95 lower)

arm bones. There was evidence for plough damage.

Located in Row 3, the burial was on an axis which favoured a north-west/south-east orientation; the head end of the grave was only 0.1m away from Grave 1675. The grave cut was difficult to determine but is likely to have measured 1.7m long by 0.5m wide with a depth of 0.15m. The skeleton (1684) was very poorly preserved. The head faced south, the left arm was flexed and placed over the chest and the right arm was slightly bent at the elbow with the hand resting in the pelvic area. The legs were extended and the feet were absent. It was not possible to sex the individual but the age at death ranged from 17-25 years.

Type 2 graves The three examples of Type 2 graves were characterised by partial linings of limestone rubble which defined the grave cut. A further example of this grave type which was located on Site 1 is discussed above. Grave 1664 (Fig. 98 upper) This grave in Row 2 was located by the presence of protruding limestone slabs. On the north side these comprised six upright stones by the upper part of the skeleton (1666) and a single stone by its lower left leg. A single stone had been placed on the south side by the thorax area. The grave measured 1.7m long by 0.56m wide and was 0.1 deep and was only 0.2m away from the triple grave 1685 which lay to its west. The head of the skeleton faced north. Both arms were flexed at the elbow, with the lower arms on the chest, and the hands rested near the neck. The legs were extended. The inhumation was that of a probable female aged 35-45 years old.

Grave 1715 (not illustrated) This inhumation was located in Row 4 after a second machine strip and was the southern-most grave in the row. No grave cut was visible and only a partial skeleton (1717) remained. The grave was at least 1.5m long and 0.5m wide; the surviving depth was 0.1m. The skeletal remains from an adult consisted of very fragmented, extended upper and lower leg bones and scraps of disintegrating

Figure 98 Type 2 Graves 1664, 1670 and 1697, plans

82

The Post-Roman Burials was bent at right angles at the elbow and rested on the abdomen. The legs were extended. The burial was that of a probable female aged between 25 and 35 years old. The re-calibrated radiocarbon date range is cal AD 420650 (SUERC-61181: this appears to be one of the first inhumations in the cemetery.

Grave 1670 (Fig. 98 centre) This grave in Row 3 measured 1.9m long by 0.55m wide with a depth of 0.2m and was initially identified by three projecting stones. Small upright limestone slabs defined the north side, a single slab was placed at the foot end and five stones were placed by the legs on the south side. A single stone was placed at the side of the head on the south side. The upper body of the skeleton (1672) had been positioned very close to the limestone uprights. The head faced north, the right arm lay along the body and the hand was possibly under the pelvis. The lower left arm was placed across the abdomen at a slight angle. The legs were extended. The burial was that of a female aged between 17 and 25 years old.

Type 3 graves Two examples of Type 3 graves were recorded: they were characterised by large slabs or lumps of stone at the head and foot ends. Grave 1660 (Figs 99 upper and 100) Located in Row 4, the grave cut was sub-rectangular being narrower at the head and foot ends and measured 1.95m long by 0.7m wide with a depth of 0.25m. Two large vertical slabs were present at the west end of the grave and similarly at the east end, where two further stones appeared to have been displaced and lay over the feet of the skeleton (1662). The head was in line with the body, the arms were slightly flexed at the elbows and the hands were placed together in the pelvic area. The legs were extended. The burial was that of a probable male aged between 25 and 35 years old.

Grave 1697 (Fig. 98 lower) This grave in Row 4 measured 1.57m long, 0.47m wide and was 0.25m deep. It was defined by two upright limestone slabs on either side of the middle of the grave and two further stones on either side near the east end. Two small stones had been placed at the head end on the south side and there was a further small stone on this side but by the knee. The head of the skeleton (1699) faced north – the skull was badly crushed. The left arm was extended down the left side of the body and the right arm

Figure 99 Type 3 Graves 1660 and 1722, plans

83

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 101 Type 3 Grave 1722 with double burial (skeletons 1724 and 1725), view to the south

Figure 100 Type 3 Grave 1660 and skeleton 1662 during excavation, view to the north

are likely to have been placed in the grave at the same time.

Grave 1722 (Figs 99 lower and 101)

Type 4 graves

This grave measured 1.8m long by 0.8m wide and 0.35m deep and contained two skeletons. It was the northernmost burial in Row 5. One large vertical stone defined the west end of the grave and a similar sized stone was located at the east end; two small stones were placed on the north side of the grave. The larger skeleton (1724) was that of a probable male aged between 35 and 45 years, and the smaller skeleton (1725) was that of a probable female aged between 17 and 25 years. The head of skeleton 1724 rested on the western stone and faced north. The right arm was flexed over the abdomen and the left arm was slightly bent toward the north and rested on the abdomen of skeleton 1725. The legs were extended and well apart.

This was the most common grave type with five examples, all having limestone slab walls and occasional horizontally placed lids or coverings. Grave 1640 (Figs 102, 103 and 104) This grave in Row 4 was located by five horizontally laid limestone slabs which covered skeleton 1642 (apart from a gap in the upper leg area). At the head end of the grave were three vertical slabs. Six verticals had been placed on the south side; these were set at a slight angle and had probably been displaced by ploughing. Two of these (SF 465, SF 481) were re-used Roman roof tiles. This grave was unique in that the body had been placed on a thin layer of yellow clay (1692). The grave measured 1.55m long and 0.6m wide and the maximum depth was 0.15m. The skeletal elements were poorly preserved; however it was possible to determine that the inhumed was a probable

Skeleton 1725 lay to the north and had been laid on its right side. The head rested on a small stone. The right arm was bent up towards the chest and the left arm was flexed downwards towards the pelvis. The legs were close together and bent towards skeleton 1724. The two bodies 84

The Post-Roman Burials

Figure 102 Type 4 Graves 1640, 1643, 1646 and 1654, plans

0.39m. Initial clearance exposed a cist burial with vertical limestone slab sides and two very large horizontal slabs covering the lower part of skeleton 1645 (Fig. 105). The head was in line with the body, the arms were flexed at the elbows and the hands were placed together in the pelvic area. The legs were extended and relatively wide apart (Fig. 106). The body was that of a probable female aged 35 -45 years old.

female between 17 and 25 years of age. The fragmented skull rested hard against the western-most limestone upright and the head appeared to have been positioned in line with the body. The arms were extended with the hands resting probably together, in the upper leg area. The legs were extended and were very close together, particularly the lower parts. It is possible that they had been tightly bound. The re-calibrated radiocarbon date range is cal AD 630-770 (SUERC-61179).

Grave 1646 (Fig. 102 lower centre)

Grave 1643 (Figs 102 upper middle, 105 and 106)

This grave, located in Row 3, measured approximately 1.7m long and 0.4m wide with a recorded depth of 0.2m. Four vertical stones lined each side of the lower end of the grave, three small stones had been placed in the area

Located in Row 4, this was the best-constructed and bestpreserved grave in the cemetery and also the deepest. It measured 1.7m long by 0.4m wide and had a depth of 85

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 103 Type 4 Graves 1640 and 1654 before excavation showing the limestone slabs covering the inhumations. The head slab of grave 1722 can be seen in the foreground. View to the west

Figure 104 Type 4 grave 1640 with skeleton 1642 exposed, note the lower legs are placed very close together. View to the north

86

The Post-Roman Burials

Figure 105 Type 4 Grave 1643, with the head of skeleton 1645 exposed. View to the north

Figure 106 Grave 1643 with skeleton 1645 exposed, view to the north

survived. The head faced north and the arms and legs were extended.

of the skull and three large horizontal slabs covered the lower part of the skeleton (1648) which was very poorly preserved. It was only possible to say that this was an adult. The upper part of the body was absent, fragments of arms indicated that they had been flexed and possibly folded in the abdominal area. The legs were extended.

Grave 1778 (not illustrated) This grave was located on the extreme eastern edge of the excavation area in Row 6 (see Fig. 94): as a result, only the western end of the burial was excavated. The placement of the stones strongly suggests that the grave was Type 4. Two horizontal slabs covered the skull, which was very fragmented (skeleton 1779); vertical stones surrounded it. The individual was a female aged 16 to 17 years old. The grave was only 0.3m below the present ground surface and apart from the skull, which was removed for analysis; the remainder of the grave and skeleton was left in situ.

Grave 1654 (Fig. 102 lower) a child aged about five years old. Not surprisingly the grave was the smallest in length at 1.28m. It had a maximum width of 0.4m and was 0.2m deep. Eleven horizontal limestone slabs (the largest at the western end) covered the cist; three of these were reused Roman roof tiles (SF 485, SF 467 and SF 468). The stones on the south side of the cist had fallen inwards. Only fragmentary remains of the child (skeleton 1656) 87

Multi-period Occupation at Football Field, Worth Matravers, Dorset clear grave cut. A single vertical limestone slab marked the foot of the grave which measured 1.9m long, 0.7m wide and 0.2m deep and was cut to take two individuals. A shallow oval scoop (1673) measuring 0.42m long by 0.34m wide was recorded north-west of this grave and can be seen on Fig. 109. It had steep sides and an undulating base and could possibly have held a grave marker. The larger skeleton 1680 was male and 40 to 45 years old. The head faced south, the left arm was placed along the body and the right arm was flexed at the elbow with the hand placed in the pelvic area. The legs were extended and the feet were relatively well-preserved. The re-calibrated radiocarbon date range is cal AD 610-770 (SUERC-61180).

Type 5 graves Only one grave of this type was recorded; it was characterised by lumps of stone at the head end and lumps or horizontal slabs of stone at the foot end. Grave 1649 (Fig. 107) This grave in Row 2 measured 1.7m long by 0.5m wide and was 0.15m deep. Two lumps of stone had been placed at either side of the head of skeleton 1651. Four horizontal slabs were arranged at the east end of the grave. The skeletal elements were poorly preserved, but it could be seen that the head was aligned with the body, the left arm was gently inflexed towards the pelvis and the lower part of the right arm was bent over the lower abdominal area. The lower legs (and absent feet) rested on and above a horizontal slab. The sex of the individual was undetermined but the age range was 25 to 35 years.

The sex of the smaller skeleton 1681 was indeterminate but its age was between 15 and 16 years. The body had been placed on its right side with the head also facing south. The right arm was flexed up to the shoulder. The left arm was flexed towards the abdomen and the hand may have clasped that of skeleton 1680. The legs were bent at the knees.

Type 6 graves

Grave 1737 (not illustrated)

The three type 6 graves were identified by a single vertical limestone stone slab placed at either the head or foot end of the burial.

This grave in Row 3 was located during the digging of house foundations. It was at a lower depth and was not visible even after two strips by the mechanical excavator. A vertical slab defined the foot of the grave which was 1.9m long and 0.66m wide with a depth of 0.28m. The head was aligned with the body; the trunk unfortunately was machined through (the spoil was searched and some abdominal and arm bones were recovered). The legs were extended, close together and portions of the feet survived. The skeleton (1713) was male but it was not possible to determine the individual’s age.

Grave 1675 (Fig. 108 upper) The grave in Row 2 was the longest in the cemetery measuring 2.02m; it was 0.5m wide and 0.1m deep. A single vertical stone marked the foot of the grave and a further vertical stone was placed on the extreme north-west side. The head of skeleton 1677 faced south, and the arms had been placed by the side of the body. The legs were extended and fragments of the foot bones were present. The individual was female and was between 35 and 45 years of age.

Type 7 graves

Grave 1678 (Figs 108 lower and 109)

The grave type was defined by slabs laid on the surface of the burial and no apparent grave cut; there was only one example.

Located in Row 3, this was the northernmost grave in the cemetery and was the only one which was identified by a

Figure 107 Type 5 grave 1649, plan

88

The Post-Roman Burials

Figure 108 Type 6 graves 1675 and 1678, plans

between 25 and 35 years old, had been placed with the upper body on its side and the legs extended. The head faced south and the arms were flexed at the elbow with the hands placed in the pelvic area. Unusually, the feet bones were well-preserved in all three skeletons. A re-calibrated radiocarbon date range for this skeleton gave a date of cal AD 600-720 (SUERC-61182).

Grave 1685 (Figs 110-112 and 114) This unique grave, the only one in Row 1, contained three individuals (skeletons 1688, 1693 and 1687) and was the westernmost interment lying only 0.35m away from the linear pavement 1695. There was no apparent grave cut but the top of the burial was defined by an arrangement of irregular-sized, horizontally placed limestone slabs which only partially covered the burials. The grave measured 1.8m long, 1m wide and was 0.1m deep. The skeletal remains were probably the best-preserved in the cemetery.

Un-typed graves Two graves could not be assigned to type, the first because it was destroyed by the digging of house footings and the second because of its fragmentary nature.

The three bodies had been placed in the grave at the same time. Skeleton 1688, the southernmost burial, had probably been deposited first. This was a male aged between 45 and 49 years whose head and upper body had been covered by four limestone slabs. The head faced north, the right arm was flexed at the elbow with the hand resting in the pelvic area and the left arm was similarly arranged but with the hand behind the pelvis. The legs were extended, slightly pointing in a northern direction. The middle skeleton (1693) was an adolescent aged between 15 and 17 years old (the sex was undeterminable), and had been placed on its right side with the head facing south. The arms were extended along the body and the legs were slightly bent at the knees and inclined towards skeleton 1688. Four large horizontal stones covered this body. One of the stones partially covered the right leg of the northernmost skeleton (1687). Two smaller stones completely covered the feet of 1693 and 1687. This skeleton, a female aged

Grave 1719 This grave in Row 2 was located underneath pavement 1695 and was recorded during the mechanical excavation of house foundations. The skeleton may have been disturbed probably by the pavement construction. There were no indications of a grave cut. Skeletal material (1721) was recovered from saved spoil from the area and the remains were identified as probably female, aged between 25 and 35 years old. Grave 1733 The southernmost grave in Row 5 was only visible as a slightly darker soil mark measuring approximately 1.5m long by 0.5m wide. Within this was a scatter of 89

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 109 Type 6 grave 1678, with skeletons 1680 and 1681. View to the west

fragmentary, disintegrating bone. It was not possible to determine the sex or age of this individual.

numbered 1695 at the west and 1775 at the east and which was cleared at a late stage of the work. In section 1695 at least two layers of limestone rubble survived; this had a defined north-east facing edge. There was a possible entrance at the extreme western end formed from large rectangular limestone blocks 1702 and 1703 (Fig. 113, A-B). The larger block 1702 (which measured 0.4m by 0.21m by 0.15m) had slipped westwards and there was a gap of 0.7m between this and the smaller block 1703 (which measured 0.3m by 0.18m by 0.13m). The possible wall may have returned south-west near to this gap. There was no evidence for the rubble continuing beyond this.

Later Features (Fig. 113). A tumble of limestone blocks and rubble measuring 17.5m long and up to 3m wide aligned in a north-west/south-east direction was recorded on the south-western side of the cemetery (see Fig. 94). Its original length is not known as the feature continues in a south-east direction outside the excavation area. Due to time constraints and pressure to remove the human remains, it was not possible to satisfactorily clean and excavate this feature which was 90

The Post-Roman Burials

Figure 110 Type 7 grave 1685 with skeletons 1688, 1687 and 1693, plan

Figure 111 Type 7 grave 1685; initial cleaning showing horizontal slabs and skeleton 1687 partially exposed. View to the west

91

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 112 Type 7 grave 1685 with skeletons 1688, 1693 and 1687. View to the west

The extension 1775 was investigated and a section through (Fig. 113, C-D) hinted that the feature may have been a wall with six courses of stonework surviving. The limestone structure was visible on the base of the top layer of midbrown loam (1776). The second layer 1777 exposed a well-constructed, double line of stones, the largest of these measuring 0.4m by 0.3m by 0.15m; the infill consisted of small rubble. The width of the wall here was approximately 0.8m. The third layer (1728) also contained horizontally laid limestone slabs, but generally of a smaller size. One large piece of unworked shale was found between the limestone slabs on the base of the feature. The lowest layer (1736) was a lighter brown loam with high clay content and with rare charcoal and no limestone.

and Graves 1685, 1664 and 1737 were laid out only 0.2m from its north-east facing edge. During observation of the digging of a water trench to the houses, a U-shaped, undated ditch (1775) was noted. It was 1.4m wide with a depth of 0.6m ran in a north-south direction and lay approximately 2m from the edge of the wall feature. Conclusions The burial of 26 individuals dating from c. cal AD 420810 in a well-laid out cemetery points to an organised community with established burial traditions. The focus of this community however has yet to be located. The use of Roman roofing tiles (and possibly building stone) in the graves implies that the Roman buildings had gone out of use and were either falling down or had been deliberately demolished.

Much of the stone appeared to be tumble which had mostly fallen to the south and probably accumulated as the wall deteriorated. Low numbers of finds were incorporated in the soil matrix (1653) into which the stones were placed. These varied from 14 sherds of Middle Bronze Age pottery (probably related to feature 1700 discussed in Chapter 2), to seven tiny abraded sherds of Roman pottery, two iron nails and small fragments of shale. Skeletal material from Grave 1719 was recorded underneath the spread

The wall feature on the southern side of the cemetery is enigmatic. Its form and function remain conjectural; the absence of firm dating evidence does not help. It is likely that the stonework comprised foundation layers for a building or a boundary wall. On balance it would seem that the feature post-dates the cemetery. 92

The Post-Roman Burials

Figure 113 Plan and sections of wall/pavement 1695/1775

93

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 114 Wall/pavement 1695 with entrance (visible centre top) and its relationship to grave 1685. View to the west

Figure 115 Eastern extent of wall 1775 with layer 1777 exposed. View to the east

94

7 Flaked Stone, Worked Stone and Kimmeridge Shale Flaked Stone Peter S. Bellamy and R.M.A. Trevarthen

for more detailed quantitative analysis. However, it was clear that there was a high degree of residuality in the assemblage: because there were insufficient quantities of flint from stratigraphically secure contexts, analysis of the flaked stone assemblage was restricted to qualitative assessment of the assemblage from each of the main chronological phases.

Introduction A total of 2041 pieces of worked flint and chert were recovered from the excavations. The whole assemblage was examined and catalogued and the flaked stone from each site is presented in Table 6 and by phase in Table 7. Close examination of the assemblage revealed that there was very little flint recovered from securely sealed contexts: most appeared to be contaminated with residual or intrusive material. About 82.5% of pieces came from stratified contexts, with most of it coming from Late Bronze Age/Early Iron Age contexts on Site 2 (Table 7). None of these contexts contained a large enough sample of pieces to allow for any meaningful quantitative analysis.

Raw material The raw material used was primarily flint with a small number of pieces of Portland chert, other limestone chert and Greensand chert. Flint formed 97.94% of the total assemblage (1999 pieces). The flint ranged in colour from pale grey opaque to glossy mid-brown translucent and mid-brown opaque. In general, it appears to have been of relatively poor quality. Where it could be determined, about 36% of the pieces appear to have derived from beach pebbles or cobbles; 60% had a thin smooth weathered cortex, perhaps indicating gravel flint or field nodules. About 2% of the assemblage had a thin un-weathered cortex and the occasional piece had iron-stained cortex. The precise source of the raw material is not known. The pebble flint may have been collected from nearby beaches, but there appear to be very few flint pebbles along the present-day rocky coastline. The poor quality of the raw material would suggest that it had not been brought from a great distance. The nearest possible source of chalk flint is the chalk ridge about 5 km to the north.

Methodology All material was counted by context and classified into a series of broad general categories commonly employed in flint analysis to produce an overall catalogue. The definitions of the main lithic categories follow Andrefsky (1998), Ballin (2000) and Inizan et al. (1999). In addition, a number of technological and other attributes that may have chronological or diagnostic significance (see Harding 1991, table 16, for example) were noted but not quantified. Conditional factors such as burning, patination or damage were also noted.

Chert does not appear to have been utilised as a significant raw material on the site, as it only formed 2.06% (42 pieces) of the total assemblage. Limestone chert nodules occur naturally in the limestone bedrock, but only 0.64%

Following this initial cataloguing, the material was grouped by context type and phase to assess the potential Table 6 Total flaked stone assemblage Site

flakes

blades

cores

tools

Evaluation Trenches

16

1

2

7

Site 1

205

8

11

18

17

Site 2

828

57

65

62

Site 3

150

4

1

8

Site 4

14

Total

1213

70

79

95

chips

Total

tool type

26

1 B + T arrowhead; 1 oblique arrowhead; 3 scrapers; 1 piercer; 1 knife

49

308

4 scrapers; 2 piercers; 2 notched flakes; 2 truncated blades; 2 denticulate retouched flakes; 4 retouched flakes; 1 core tool; 1 hammerstone

159

289

1460

2 leaf arrowheads; 2 oblique arrowheads; 23 scrapers; 1 knife; 2 piercers; 1 spurred flake; 1 notched flake; 1 flake tool; 1 lathe bit; 3 denticulate retouched flakes; 19 retouched flakes; 6 hammerstones

27

40

230

3 scrapers; 1 scraper/knife; 1 piercer, 2 retouched flakes

3

17

381

2041

203

Indeterminate

95

Multi-period Occupation at Football Field, Worth Matravers, Dorset Table 7 Total flaked stone assemblage by phase and site Phase

Site

Raw Mat.

flakes

blades

cores

tools

chips

Indeterminate

Total

tool type

Neolithic/ EBA

2

Flint

200

35

5

12

67

26

345

1 leaf arrowhead; 2 oblique arrowheads; 8 scrapers; 1 Y-shaped tool

Neolithic/ EBA

2

Chert

2

1

3

MBA

3

Flint

7

LBA

1

Flint

9

LBA

2

Flint

129

LBA

2

Chert

3

LBA

4

Flint

6

LBA/EIA

1

Flint

84

2

4

6

9

20

125

2 scrapers; 1 notched flake; 3 retouched flakes

LBA/EIA

2

Flint

389

7

39

28

51

175

689

1 leaf arrowhead; 6 scrapers; 2 piercers; 1 knife; 5 denticulate retouched flakes; 8 retouched flakes; 5 hammerstones

LBA/EIA

2

Chert

6

1

1

1

3

12

1 retouched flake

LBA/EIA

3

Flint

1

1

MIA

2

Flint

3

2

6

LIA/RB

1

Flint

38

16

65

LIA/RB

1

Chert

LIA/RB

2

Flint

LIA/RB

2

Chert

PostRoman

3

Flint

50

PostRoman

3

Chert

3

Unphased

1

Flint

2

1

Unphased

2

Flint

6

2

1

Unstratified

1

Flint

88

6

4

Unstratified

1

Chert

6

1

Unstratified

2

Flint

17

Unstratified

2

Chert

Unstratified

3

Flint

87

Unstratified

3

Chert

Unstratified

4

Unstratified

Eval

7 1 7

10

11

25

1

11

1 piercer

40

222

8 scrapers; 1 lathe bit; 2 retouched flakes

2

5 6

1 3

3

5

1

1 51

2

5

7

9

1 3

3

10

1 piercer; 2 truncated flakes

25

99

1

2

17

83

2 scrapers; 1 piercer; 1 flake tool; 1 lathe bit; 1 retouched flake; 1 hammerstone

2 scrapers; 1 piercer

3 3

2

8

3

2

1

12

5

17

128

1

8

7

31

2

18

129

3

4

7

Flint

8

3

11

Flint

16

1

2

7

Total

1213

70

79

95

1

2 denticulate retouched flakes; 1 retouched flake

1

1 1

2 scrapers; 2 denticulate retouched flakes; 1 retouched flake; 1 core tool; 1 hammerstone

5

17

26

203

96

381

2041

1 oblique arrowhead?; 1 scraper; 1 scraper/knife; 1 retouched flake; 1 core tool

1 barbed & tanged arrowhead; 1 oblique arrowhead; 3 scrapers; 1 piercer; 1 knife

Flaked Stone, Worked Stone and Kimmeridge Shale The 12 tools consist of three arrowheads, eight scrapers and a retouched flake (Table 7). The arrowheads include one leaf arrowhead of Green’s Type 2B (Green 1980) (Fig. 116.1) and one near complete and one broken oblique arrowhead. The leaf and the oblique arrowheads were recovered from Ditch 1332 (contexts 1620 and 1609 respectively). Six of the eight scrapers are broken (one burnt) which makes it difficult to determine the preferred blank size and shape: three appear to be on primary flakes and the remainder are on a range of different size and thickness flakes, including three squat flakes and one long flake. Generally the blanks appear to be relatively small with none over c. 40mm in length and most less than 30 mm. The scrapers are all end scrapers (Fig. 116. 2-3), except for one thumbnail scraper from context 1348 (Fig. 116. 4). The scraping edges were often formed by minimal fine regular acute or abrupt retouch. The character of these scrapers suggests a Neolithic date, with some, such as the examples from contexts 1629 and 1275 perhaps belonging to the Early Neolithic, and the thumbnail scraper from context 1348 perhaps associated with Late Neolithic (Beaker) activity. The other implement from Ditch 1332 is a Y-shaped tool (Fig. 116. 5) which is also Late Neolithic in date.

of the assemblage was made up of this material, which is of poor quality. A greater number of pieces of better quality Portland chert were recovered, but this still only formed 1.32% of the assemblage. Two flakes of Greensand Chert were noted. Assemblage description The assemblage was divided into the main chronological phases by site (Table 7) and the material from each phase is described separately below. The bulk of the flint came from Late Bronze Age and Late Bronze Age/Early Iron Age transition contexts, with a smaller number of pieces from Neolithic/Early Bronze Age contexts. There was a relatively small amount of flint from Middle Iron Age, Late Iron Age/Roman, and post-Roman contexts. Flint from Neolithic and Early Bronze Age contexts A total of 364 pieces of flint and chert were recovered from Ditch 1332 and other features and deposits below the soil layer sealed by the Late Bronze Age/Early Iron Age cobbled surface on Site 2 (Table 7). This assemblage consists of 57.97% flakes and broken flakes, 10.14% blades and broken blades, 1.45% cores, 3.48% tools, 19.42% chips and 7.54% miscellaneous debitage. Generally the flint is in a slightly edge-damaged condition with a smaller proportion of pieces in a sharp condition; 49% of pieces are patinated and 6% are burnt. The patinated pieces vary from items with very slight patinated specks to a small number of pieces with fairly heavy blue or white patination.

Flint from Middle Bronze Age contexts Seven pieces of flint were recovered from the possible Middle Bronze Age pottery clamp on Site 3. One piece is burnt and the remainder are patinated. This material consists of five small broad flakes and two broken flakes. Two flakes were from beach cobbles. None of this material is diagnostic and is likely to be residual in this context.

The flakes (of which 47% are broken) include primary, secondary and tertiary flakes, perhaps with a slight majority of tertiary flakes present. In general, the flakes are intermediate to broad in shape with fairly irregular plan form. The primary flakes tend to be larger, thicker and squat and were removed by a hard hammer. A proportion of the other flakes appear to be soft hammer removals, but the majority are indeterminate. Within this assemblage are a small proportion of thinner long flakes, some with nearparallel sides, but most of intermediate plan form. A small number have signs of platform trimming, but overall there is little evidence for deliberate platform preparation. One flake was struck from a ground flint axe.

Flint from Late Bronze Age contexts Flaked stone from Late Bronze Age contexts was recovered from Sites 1, 2 and 4, though the vast majority is from Site 2 (Table 7). Most of the flint was collected from the soil layer immediately beneath the limestone surface 1039, with small quantities derived from the postholes of the Late Bronze Age house on Site 2 and other features. An assemblage of 228 pieces of flint and chert was present in Late Bronze Age contexts. This comprises 60.25% flakes and broken flakes, 2.87% blades and broken blades, 4.10% cores, 4.92% tools, 10.24% chips and 17.62% miscellaneous debitage. Much of the flint is slightly edgedamaged but a significant minority of the pieces are in a fresh condition and there are occasional rolled pieces. Nearly 30% of the assemblage is patinated and only 0.8% is burnt. The patinated pieces vary from items with very slight patinated specks to a small number of pieces with fairly heavy blue or white patination. All the patinated and burnt pieces came from Site 2.

The blades are small thin and broad with slightly irregular, but generally parallel sides. Both secondary and tertiary blades are present. There is some evidence for platform preparation. The majority (68%) of the blades are broken. The cores are flake cores. One multi-platform core from context 1606 is on a fragment of a ground flint axe. A fairly crude single-platform core is on a thick flake of beach cobble flint, which together with a broken fragment from a multi-platform flake core would fit comfortably in a Late Bronze Age industry and may be intrusive. The other core is a small beach pebble with a number of small flake removals from a single platform.

The flakes, of which 36% are broken, are mainly secondary and tertiary flakes, with a small quantity of primary flakes. This assemblage includes both small thin intermediate to broad flakes (many of which are patinated) and larger 97

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 116 Flaked stone (1-9)

98

Flaked Stone, Worked Stone and Kimmeridge Shale thicker irregular flakes, which are generally not patinated. The smaller thin patinated flakes are all from the soil layer beneath the cobble surface 1039 on Site 2 are very similar to the material from the Neolithic/Early Bronze Age features described above and therefore are probably residual in the Late Bronze Age contexts. The remainder of the flakes are generally broad with an irregular plan form, thick plain or cortical butts and pronounced bulbs of percussion, characteristic of Middle-Late Bronze Age flint industries.

Flint from Late Bronze Age/Early Iron Age Transition contexts

Blades do not form a significant component of this assemblage and are generally small thin and broad. Almost all are broken and nearly all are patinated. They are very similar to those recovered from the Neolithic and Early Bronze Age contexts and are almost certainly residual in a Late Bronze Age context.

This Late Bronze Age/Early Iron Age assemblage comprises 57.92% flakes and broken flakes, 1.21% blades and broken blades, 5.32% cores, 4.23% tools, 7.26% chips and 24.06% miscellaneous debitage. The condition of the flint is generally slightly edge-damaged, but with a significant minority of pieces in a sharp condition and a small number of rolled pieces. Just over 6.4% of the assemblage is patinated and 3.75% is burnt. The majority of contexts contain very small quantities of flint and where larger numbers are present there is generally a clear residual component to the assemblage. However, within the midden there are a small number of contexts that appear to have evidence for contemporary flint working. Unfortunately there are no secure sealed contexts and it is not possible to confidently isolate this contemporary industry from the background scatter of residual flint. This material appears to be distributed throughout the midden and is most likely to have been derived from elsewhere. The following is a qualitative description of this industry.

A total of 827 pieces of flint and chert was recovered from Late Bronze Age/Early Iron Age transition contexts, almost all from Sites 1 and 2, with only one piece of possibly accidental miscellaneous debitage from Site 3 (Table 8). A total of 639 pieces of worked flint and chert came from midden contexts, with 188 pieces from pits, gullies and other features.

The cores are flake cores and most appear to have been worked out. Four of the cores are on cobble flint, of which two appear to have utilised cobbles which had originally been broken by a very heavy hard hammer strike and then the broken pieces used for flake production. Eight of the ten cores are multi-directional flake cores, with the core rotated to use an earlier flake bed as the striking platform. There is no evidence for platform preparation and the majority of removals appear to have been squat flakes. The other two cores were single-platform flake cores. Two of the multi-directional cores were small, under c. 25mm across. The 12 tools recovered from Late Bronze Age contexts consist of eight scrapers, one piercer, one lathe bit and two retouched flakes (Table 7). Four of the scrapers are patinated with fine regular acute or abrupt retouch, very similar in character to the scrapers from Neolithic and Early Bronze Age contexts and are likely to be residual. One other scraper has coarse abrupt unpatinated retouch on the distal end of a long patinated flake blank. The remaining three scrapers were on thicker squat blanks and included one with coarse inverse abrupt retouch on the proximal end of a thick squat flake and two with irregular retouch on the distal end of thick flakes. The piercer has a short point formed medially on an irregular piece of debitage (Fig.116.6).

Midden contexts The raw material is relatively poor quality cobble and pebble flint; a significant proportion has a thin smooth pale buff cortex. The poor quality of the raw material has resulted in a relatively large quantity of miscellaneous debitage, which appears to be mainly fractured broken cobble and nodule pieces, probably accidental products of the knapping process. The nodules and cobbles appear to have been initially broken apart using a very heavy hard hammer blow, but there are no prismatic cores or other evidence of the use of an anvil during this initial reduction process. A number of flakes have the remains of large prominent demi-cones of percussion surviving on their dorsal surface, derived from this initial breaking of the nodule/cobble. The majority of the flakes are secondary and tertiary flakes, with only a very small number of primary flakes. The flakes are predominantly squat and characterised by thick plain and cortical butts with obtuse percussion angles and very prominent bulbs and cones of percussion, often with bulbar scars and radial fissures present, indicating removal with a heavy hard hammer. They are generally an irregular plan shape with very irregular angular dorsal surfaces and usually with step or hinge terminations.

A possible lathe bit was recovered from context 1568 and consists of a transverse section of a thick flake with very abrupt retouch on both snapped edges and with small secondary removals at the angle between the original flake edge and the snapped face, probably caused by use in a lathe. It is very similar to Romano-British lathe bits from Rope Lake Hole (Cox and Woodward 1987) and Ower (Woodward 1987b) and other sites on Purbeck and must be intrusive in this context. The retouched flakes included one flake with minimal irregular retouch on the distal end of a squat flake, which could be damage rather than intentional. The other retouched flake had lightly patinated coarse abrupt retouch on a heavily patinated broken flake.

Blades are almost completely absent from this assemblage, but those present are almost all patinated small thin blades 99

Multi-period Occupation at Football Field, Worth Matravers, Dorset Table 8 Late Bronze Age/Early Iron Age flaked stone assemblage Context

Site

Raw Mat.

flakes

Broken flakes

blades

Broken blades

cores

Midden

1

Flint

20

13

1

1

3

Midden

2

Flint

216

123

3

3

33

25

Midden

2

Chert

2

3

1

1

1

Pit 326/390

1

Flint

19

10

Pit 335

1

Flint

14

8

Pit 1142

2

Flint

1

Pit 1234

2

Flint

Pit 1428

2

Flint

2

Pit 1510

2

Flint

1

Pit 1126

2

Flint

Pit 1136

2

Flint

1

Pit 1184

2

Flint

1

Pit 1196

2

Flint

1

Pit 1241

2

Flint

1

Pit 1302

2

Flint

3

Pit 1412

2

Flint

5

Pit 1427

2

Flint

Pit 1513

2

Flint

Scoop 1216

2

Flint

1

Scoop 1231

2

Flint

1

Scoop 1263

2

Flint

2

1

3

Scoop 1269

2

Flint

2

1

3

Gully 1202

2

Flint

1

1

Postholes

2

Flint

1

Postholes

2

Chert

1

Other contexts

2

Flint

7

Other contexts

3

Flint Total

1

tools

chips

Indeterminate

Total

3

12

53

31

151

585

1 leaf arrowhead; 5 scrapers; 2 piercers; 4 denticulate retouched flakes; 8 retouched flakes; 5 hammerstones

3

11

1 retouched flake

5

3

6

43

2 scrapers; 3 retouched flakes

1

3

2

29

1 retouched flake

1

12

10

1

2

3 3

1

1 1

3

2

1 knife

2

1

1 1

1

2

1

5

2

4

3

6

3

16 1

1 1

300

tool type

3

4 2

1

1 denticulate retouched flake

1

4

2

9

2

1 10

179

2

4

6

44

100

1

35

60

8

28

1

1

199

827

1 scraper

Flaked Stone, Worked Stone and Kimmeridge Shale the occasional piece of worked flint or chert and these are considered individually below.

and bladelets, similar to those recovered from the Neolithic and Early Bronze Age contexts. They are almost certainly residual in a Late Bronze Age/Early Iron Age context.

Pit 326/390

The cores appear to be a mixture of beach cobbles (45%) and weathered field nodules (12%) and are generally fairly crude flake cores. Many have several shattered surfaces where the nodule/cobble has split along planes of weakness in the flint, perhaps when the nodule/cobble was initially smashed open. A number of cores retain traces of the cone of percussion from this initial blow. Multi-platform cores comprise 64% of the Late Bronze Age/Early Iron Age core assemblage. These have crude flakes removed from three or more directions by the simple expedient of rotating the core and using previous flake beds or other fractured surfaces as a platform. A number have incipient cones indicating miss-hits. The remainder of the cores have flakes removed in one or two directions (9% each); two are on very thick irregular flakes (6%) and the others are broken core fragments (12%).

Forty-three pieces of flint were recovered from this pit, primarily flakes and broken flakes (Table 8). The majority of pieces are in a fresh condition and are similar in character to the Late Bronze Age/Early Iron Age flint industry in the midden, but with a small residual component also. Five tools were identified; two scrapers and three retouched flakes on thick flake blanks. One scraper was burnt and most of the scraping edge had spalled off, but it appeared to have originally been an end scraper with steep regular retouch on a broad thick flake. The other scraper has coarse abrupt retouch forming a denticulated edge, a straight distal scraping edge, and truncation and retouch on the proximal end removing the bulb (Fig. 117.14). The retouched flakes are on irregular flake blanks with areas of abrupt retouch along a number of edges, some of which may be accidental edge damage rather than deliberate modification of the pieces.

Of the 26 tools recovered from Late Bronze Age/Iron Age midden deposits, 22 are on blanks that suggest they form part of the Late Bronze Age/Early Iron Age industry. These comprise four scrapers, two piercers, eleven retouched flakes and five hammerstones. The scrapers are small (33mm to 42mm across) and are formed by coarse abrupt retouch on broken thick blanks or thermal pieces (Fig. 116.7-8). One piercer had a long thick point formed on the distal end of a flake. A burin-like flake was removed from the ventral surface of the point, presumably use damage (Fig. 116. 9). The other piercer has a relatively short point formed on the distal end of a primary flake (Fig. 117.10). The retouched flakes were made on a series of differingsized thick irregular blanks with coarse abrupt retouch, in four instances forming a denticulated edge (Fig. 117.1113). Three of the blanks had the bulb wholly or partially removed by retouch.

Pit 335 The flint from this pit is more mixed and in a more edgedamaged condition than that in Pit 326/390. It contains more residual pieces, but still has an identifiable Late Bronze Age/Early Iron Age component. A single small thick broad flake has fine abrupt retouch along part of the distal end. Pit 1142 Most of the flint from this pit consisted of small chips from sieved samples with no diagnostic pieces. Pit 1412

The five flint hammerstones recovered from midden deposits ranged in size from 60mm to 113mm across and included one ovoid example with pecking over almost all of its surface, two core hammers and two irregular beach nodules with more limited pecking.

The flint from this pit was small in size and comprised mainly residual pieces with very little material that could be considered contemporary with the use of the pit.

The residual tools recovered from the midden deposits are a patinated broken leaf arrowhead, a patinated end scraper of Late Neolithic/Early Bronze Age character, a small thin flake with fine inverse retouch on the proximal end and one edge, and a long Portland Chert flake with oblique abrupt retouch removing the proximal end.

The other Late Bronze Age/Early Iron Age features contained little diagnostic flintwork. Three tools were recovered. These comprised a residual patinated flint knife on a long thin flake from Pit 1126; a well-made end scraper of Late Neolithic/Early Bronze Age character from pit 1022, and a broken flake with coarse denticulated retouch from Scoop 1216. Only the latter appears to be Late Bronze Age/Early Iron Age in date.

Other features

Pits and other contexts The flint from Late Bronze Age/Early Iron Age non-midden contexts is similar to that from the midden and contains a mixture of potentially contemporaneous flintwork and clearly residual material. Most of the features contain very small quantities of material and this probably represents accidental inclusion. Only four pits contained more than

Flint from Middle Iron Age Contexts Only six pieces of flint were recovered from the single Middle Iron Age feature (pit 1182) on Site 2. Two pieces were heavily patinated, including one blade fragment, which may be residual. One broken flake was burnt. 101

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 117 Flaked stone (10-18)

102

Flaked Stone, Worked Stone and Kimmeridge Shale Flint from Late Iron Age and Roman contexts

Discussion

A total of 167 pieces of flint and chert were recovered Late Iron Age and Roman contexts on Sites 1and 2 (Table 7), consisting of 53.29% flakes and broken flakes, 1.20% blades and broken blades, 5.39% cores, 5.99% tools, 8.98% chips and 25.15% miscellaneous debitage. Overall the flint is in a slightly more rolled condition than the earlier material; 7.8% of the assemblage is patinated and 6.6% is burnt. Very few contexts contained more than a small number of pieces and generally the assemblage appears mixed. There is no definite evidence for contemporary knapping associated with Late Iron Age or Romano-British contexts, although there are a small number of very crude flakes made from very poor raw material, which may represent a late phase of knapping. The number of pieces however is not large enough to enable any meaningful analysis. There is no evidence of an actual flint industry for the production of shale-working tools, though a small number of possible tools are present in the assemblage. One very crude lathe bit was recovered from context 1402 (Fig. 117.15). Two flakes with very steep retouch on the truncated distal end were recovered from Site 1 (Contexts 305 and 383), and may also be associated with shale working (Fig. 120.1617). One square flake tool superficially looks like a small broken gunflint, but may also have been used for shale working (Fig. 117.18).

The overall character of the flint recovered from the excavations is of a mixed assemblage with a small earlier prehistoric component and a much larger later prehistoric (Late Bronze Age/Early Iron Age) element. There is a significant amount of residuality in many of the excavated features. The assemblage appears to be broadly similar to the flint recovered from the 1990-93 Southampton University excavations (Marshall 2002), but unfortunately this earlier collection has not been analysed in detail, though the brief description in the publication strongly suggests the flint from the earlier and the current excavations have similar characteristics. Flint was recovered from all excavated areas, but Site 2 produced significantly more flint than the other areas. This may be a function of the more intensive investigation of Site 2, compared to the other areas. Nevertheless, the bulk of the flint appears to be associated with the Late Bronze Age/ Early Iron Age midden on both Sites 1 and 2. The midden contained a flint industry that shares many characteristics with Late Bronze Age or later industries found elsewhere (Ballin 2002; Fasham and Ross 1978; Humphrey and Young 1999; Young and Humphrey 1999) where there is use of local poor quality raw materials worked using a simple hard hammer core/flake technology. The cores are irregular with expedient flake removal, rather than carefully controlled core reduction. The flakes are squat, irregular with thick, wide striking platforms, obtuse striking angles and a large number of step terminations. There appears to be a relatively restricted range of formal tool types associated with this industry, consisting mainly of scrapers and piercers. There is a range of miscellaneous retouched flakes, usually with coarse, sometimes denticulated, retouch. Undoubtedly a range of unretouched pieces were also utilised but there were no macroscopic signs of use/wear. Overall, this industry shares many common characteristics with the Middle and Late Bronze Age flint found at Bestwall Quarry (Bellamy and Montague 2009).

Flint from Post-Roman contexts Eighty-six pieces of flint and chert were recovered from Post-Roman contexts on Site 3. The assemblage comprises 61.63% flakes and broken flakes, 3.49% blades and broken blades, 3.49% tools, 11.63% chips and 19.77% miscellaneous debitage. The flint is mainly in a fresh condition with only very slight edge damage. Most of the flint is white patinated (88%) and 7% is burnt. The flint was derived from grave fills with only a sparse number of pieces recovered from each grave, except for grave 1685 which contained 12 pieces of flint and chert. It is likely that the flint is residual in a Post-Roman context, though the good condition of the flint suggests it was derived from some activity nearby and had not been very disturbed before being incorporated in the Post-Roman graves. The flint includes a significant proportion of broad thin flakes and a number of blades and blade-like flakes with parallel flake scars, similar in character to the flint recovered from the Neolithic/Early Bronze Age contexts on Site 2. Three tools were recovered including one scraper with fine abrupt retouch at the distal end of a cobble flint flake, one small fragment of another scraper and a borer/awl on the proximal end of a flake with the long point formed by both direct and inverse retouch.

The flint appears to have been distributed throughout the midden deposit, with material found both in the lowest and the highest levels. No meaningful spatial distribution of the flint across the midden could be deduced from the excavation data. The differences in the excavation of Sites 1 and 2 mean that it is not possible to compare the midden assemblages from the two sites in detail: in particular it is not possible to make any meaningful comparisons of the density of flint distribution. There is no clear evidence recorded for possible in situ knapping scatters either, in the midden or elsewhere on site. The impression gained is that the flint was incorporated within the midden as part of the pattern of refuse disposal. No evidence was recovered to indicate any special deposition practice involving the flint. The assemblage from the midden includes both waste material created as part of the knapping process and used and discarded tools. It has not been possible to determine what proportion of the industry had been removed for

A brief examination of the unstratified flint from Site 3 indicated it also contained a probable Neolithic component with some blades and blade-like flakes present and a small single platform flake core. The tools included an end scraper and a flake knife with fine semi-abrupt retouch, a small fragment of a flake with similar retouch and a possible broken oblique arrowhead. 103

Multi-period Occupation at Football Field, Worth Matravers, Dorset use elsewhere and had not become incorporated into the midden.

4. Thumbnail scraper, context 1348, Ditch 1332, Neolithic/Early Bronze Age

The earlier flint assemblage appears to be primarily a flake industry with a small blade component. Much of this earlier flint is patinated, unlike the later flint industry. There is no indication of any in situ flint knapping in any of the excavated areas and the vast majority of the material is residual. Indeed, it is unclear whether any of this earlier flint is contemporary with the contexts in which it was found. This earlier industry contains elements that date from the Early Neolithic to the Early Bronze Age and may represent a very mixed assemblage representing use of this area over a very long period of time. Ditch 1332 contained the greatest proportion of early flintwork, but this included both Early Neolithic tools (leaf arrowhead, scrapers) and Late Neolithic/Early Bronze Age tools (oblique arrowhead, thumbnail scraper). The relatively small quantity of flint from this feature, together with the small proportion of diagnostic material, means it is difficult to determine the date of this feature from the flint alone.

5. Y-shaped tool, context 1611, Ditch 1332, Neolithic/ Early Bronze Age 6. Piercer, context 375, Late Bronze Age 7. Scraper, context 1133, midden, Late Bronze Age/Early Iron Age 8. Scraper, context 1055, midden, Late Bronze Age/Early Iron Age 9. Piercer, context 1491, midden, Late Bronze Age/Early Iron Age Fig. 117 10. Piercer, context 1233, midden, Late Bronze Age/Early Iron Age 11. Retouched flake, 1082, midden, Late Bronze Age/ Early Iron Age

The early material also has a widespread distribution across almost the whole of the excavated area. It is noticeable that a very significant proportion of the flint from Site 3 appeared to belong to an earlier prehistoric industry, though it was all recovered from later features or was unstratified.

12. Retouched flake, 1168, midden, Late Bronze Age/ Early Iron Age 13. Retouched flake, 1233, midden, Late Bronze Age/ Early Iron Age

There is relatively little flint recovered from Iron Age, Roman and Post-Roman contexts and most of this appears to be residual Neolithic to Late Bronze Age material. There is no clear evidence for Middle and Late Iron Age and Roman flintworking, other than the occasional example of possible shale-working tools. This pattern largely reflects that found in the much larger assemblage recovered at Bestwall, where most of the flint from Late Iron Age and Roman features appears to be residual Late Bronze Age and earlier flint, with occasional shale-working tools present (Bellamy and Montague 2012). The evidence from both Bestwall and Worth suggests that the production of flint artefacts in Purbeck ceased in the Late Bronze Age or Early Iron Age, with flint only being used later in the Late Iron Age and Roman period for producing specialised shale-working tools.

14. Scraper, 366, Pit 326/390, Late Bronze Age/Early Iron Age 15. Lathe bit, context 1402, Romano-British 16. Truncated flake, context 305, Romano-British 17. Truncated flake, context 383, Romano-British 18. Flake tool, context 1097, Romano-British Worked Stone Peter S. Bellamy and R.M.A. Trevarthen Introduction

Illustrated flint

Forty-one pieces of worked stone were recorded. Many of these objects were broken and found in residual contexts, so their precise dating often remains in doubt. The artefact types recovered and the date of the contexts from which they were recovered are recorded in Table 9.

Fig. 116 1. Leaf Arrowhead SF 440, context 1620, Ditch 1332, Neolithic/Early Bronze Age 2. Scraper, context 1629, Ditch 1332, Neolithic/Early Bronze Age

Raw material Petrological identification was undertaken by visual examination using a hand lens. The raw materials used for the stone objects derive from a number of sources. By far the greatest majority were local limestone, heathstone, ferruginous sandstone, sarsen, chert and quartzite, which

3. Scraper, context 1275, Ditch 1332, Neolithic/Early Bronze Age

104

Flaked Stone, Worked Stone and Kimmeridge Shale can be found locally. Together these make up 97.56% of the worked stone assemblage. A small number of pieces from more distant sources are present in the assemblage. These comprise one piece each of micaceous sandstone and granite from Devon/Cornwall. The granite fragment was unworked. The stone types are quantified in Table 10.

from Bestwall Quarry (Bellamy 2009, fig. 187, 1-3). The other fragments are too small to determine their original size and shape, but appear to have flat or longitudinally concave grinding surfaces, similar to some of the Middle Bronze Age saddle querns from Bestwall Quarry (Bellamy 2009).

Querns

Rotary quern

Six quern fragments were recovered from the site: five saddle querns and one rotary quern fragment.

One fragment of rotary quern was recovered, from Late Iron Age pit 1740.

Saddle querns

It is an upper stone pecked to shape with a domed top (but with some areas of the original rather irregular unworked surface of the stone still present) and a regular concave grinding surface and tapering oval handle socket (Fig. 118.3). It has been broken and most of the edge is damaged. It is of Curwen’s Wessex type (Curwen 1937) and is similar to the Type 1a quern from South Cadbury (Barrett et al. 2000, fig 103.2), which is roughly dated to the Middle-Late Iron Age. It corresponds to quern type R3 from Danebury, which is also roughly dated to the MiddleLate Iron Age (Brown 1984, 407) but is unlike the Iron Age querns from Hengistbury Head (Cunliffe 1987). No

Four of the five saddle querns were made from heathstone and one was of micaceous sandstone. None were complete, but one was almost complete, though broken into many pieces (Fig. 118.1). This was approximately rectangular in shape with a rough unworked underside and a dished grinding surface formed by pecking. One saddle quern fragment (Fig. 118.2) from a clearance level, has a longitudinally slightly concave grinding surface and a convex underside carefully shaped by pecking, which is similar to a group of Middle Bronze Age saddle querns Table 9 Prehistoric stone objects by phase Phase

Saddle quern

LBA/EIA

1

LIA

1

RB

2

Rotary quern

Grinding stone

Rubber

Hammer stone

Lid

Stone Anchor

1

3

1

1

1

1

3 1

Unphased 1

Total

5

Misc. object

3

Post-RB

unstratified

Stone Building Material

1

2

1

3

1

2

9

3

6

10

1

1

1

1

14

Table 10 Quantification of stone types Stone type

No of pieces

%

Probable source

Chert

5

12.19%

Local

Heathstone

5

12.19%

Local

Purbeck limestones

18

43.90%

Local

Quartzite

6

14.64%

Local

Sarsen

3

7.32%

Local

Tertiary ferruginous sandstone

3

7.32%

Local

Micaceous sandstone

1

2.44%

Devon/Cornwall

Non-local unworked stone Chalk

4

Purbeck

Granite

1

Cornwall

105

1

Multi-period Occupation at Football Field, Worth Matravers, Dorset polish over much of the surface. Heathstone. Diameter approx. 180mm; thickness approx. 83mm. Object SF 511; context 1740; phase Late Iron Age. Grinding stones Nine pebbles with evidence for crushing or grinding on one or more facets were recovered from the excavations. Five are quartzite, three of sarsen and one of Tertiary ferruginous sandstone. The pebbles are mainly elongated ovoid in shape, with one flattened and ovoid. They are between 67–170 mm in length and between 65–1036g in weight. Eight have flattened facets on one or both ends through use; the remaining example has a flattened facet around the edge of a flattened ovoid pebble (Fig. 119. 2). These objects may have been used for a variety of grinding tasks and are relatively common on Middle and Late Bronze Age and Iron Age sites and cannot be closely dated. The stratified objects were all found in Late Iron Age or Roman contexts. Illustrated grinding stones (Fig. 119)

querns of this type were found during the earlier Compact Farm excavations (Graham et al. 2002) or at Bestwall (Bellamy 2012).

4. Smooth cobble with heavy pecking and crushing on both ends forming facets. Sarsen. Length 120mm; width 78mm; thickness 66mm. Object SF 386; context 1504; phase Late Iron Age. 5. Smooth pebble with flattened facets formed by heavy pecking or crushing. Sarsen. Length 67 mm; width 38mm, thickness 30mm. Object SF 272; context 1352; unphased. 6. Smooth pebble with a flattened pecked and crushed facet at one end and a flat roughened facet at other end. Sarsen. Length 74mm; width 39mm, thickness 38mm. Object SF 459; context 1606; unstratified. 7. Smooth flat pebble with crushing along edges forming a flat facet. Tertiary ferruginous sandstone. Length >57mm; width 80mm, thickness 23mm. Object SF 231; context 1002; unstratified.

Illustrated querns (Fig. 118)

Rubbers

1. Saddle quern, almost complete but broken into many pieces. Roughly sub-rectangular in plan shape with a concave grinding surface formed by pecking, but with a number of irregularities present in the surface. Irregular unworked under surface. Micaceous sandstone. Length 445 mm; width 322mm; thickness 63mm. Object SF157; context 1036; phase Late Bronze Age/Early Iron Age. 2. Broken saddle quern, very well made and finished. Possibly ovoid or sub-rectangular in plan shape with a flat slightly polished grinding surface. Well-shaped convex under surface, pecked all over. Heathstone. Length >114mm; width 155mm; thickness 4mm. Object SF 157; context 1002; phase Romano-British. 3. Approximately one-quarter of a rotary upper stone. Pecked over much of surface to shape, with steep sides and domed top. Convex hopper. Tapering oval handle socket. Pecked concave grinding surface with use

Three pebbles have very slight evidence of use, mainly one side or edge that appears to be smoothed through use, and have been classified as rubbers. These are all on flattened pebbles or fragments, two of Tertiary ferruginous sandstone and one of siliceous limestone. One is broken. They range in size from 47mm to 97mm in length and from 40g to 371g in weight.

Figure 118 Saddle and rotary querns (1-3)

The precise function of these tools is not clear, but they are likely to have had a range of uses including smoothing leather and sharpening. These objects are not closely datable: the only dated example is from a Late Iron Age context. Hammerstones Six hammerstones were recovered from the excavations. These are all roughly spherical or ovoid pebbles of chert 106

Flaked Stone, Worked Stone and Kimmeridge Shale

Figure 119 Grinding stones and lid (4-8)

107

Multi-period Occupation at Football Field, Worth Matravers, Dorset or quartzite with evidence of heavy pitting and crushing through percussive use over a considerable portion of the surface of the pebble. They measure between 65mm and 77mm across and from 308g to 383g in weight. These objects cannot be closely dated but three of the four stratified objects were found in Late Bronze Age contexts. Stone lid An almost circular flat stone disc of Purbeck Limestone was recovered from layer 1402. It had been carefully shaped by bifacially flaking the edges to form the disc. Its original purpose is unclear, but it may have been used as a lid or cover for a storage container or small pit. Fig. 119 8. Lid? Flat stone disc made from a tabular piece of limestone with carefully flaked edges. Purbeck Limestone. Length 290mm; width 285mm, thickness 29mm. Object SF 312; context 1402; phase Late Iron Age/Romano-British. Stone anchor A single-hole stone anchor was recovered from Site 3 and reused as a pillow stone in post-Roman grave 1633 (see Fig. 96). It was formed from a sub-rectangular piece of Purbeck limestone which has some suggestion of rough flaking on one face to shape it. This is not completely certain, as the piece is very worn and eroded. It is slightly ovoid with flattish top and bottom surfaces and is slightly rounded at the top and bottom of the object. A single perforation, 42mm by 20mm across, was bored through from one face, 78mm from one end. The hole is oval and exhibits traces of wear and part of the stone adjacent to the hole has broken away. This form of anchor is the earliest type known, but is not well dated in this country. This example is significant because it is from a securely dated context (Gordon Le Pard, pers. comm.).

Figure 120 Stone anchor and triangular stone with dished centre (9-10)

purpose or function of this object is unclear, but it may have been a mortar or perhaps an offering stone. Fig. 120 10. Triangular-shaped stone with a pecked sub-circular dished recess offset towards one of the edges. The top is slightly larger than the base. Eroded on two edges. Purbeck limestone. Width of top 500mm x 544mm x 480mm, thickness 160mm; circular depression diameter c. 160mm, depth 60mm. Object SF 569, context 1405; phase Late Roman.

Fig. 120 9. Single-hole stone anchor. Sub-rectangular stone with single perforation at one end. Wear traces and eroded, worn condition suggest it was heavily used. Purbeck limestone. Length 505mm; width 265mm, thickness 83mm; weight 16.1k. Object SF 517; context 1635; phase Post-Roman.

Stone building material

Triangular stone

A quantity of Purbeck limestone was recovered from the site, most probably derived from building material. It included pieces of stone roof tile and other worked pieces, as well as featureless flat fragments and blocks.

A single roughly shaped triangular piece of worked Purbeck stone (see Fig. 81). was recovered from the building debris associated with the Roman barn and may have been incorporated into the western wall The base slopes up very slightly to the top and the pecked dishing is off-centre. This depression has gently sloping sides and a rounded base and displays no signs of rubbing. When found, the dished surface was facing down. The original

Stone roof tile Four fragments of Purbeck limestone roof tile were recovered from the site. None was complete and all 108

Flaked Stone, Worked Stone and Kimmeridge Shale However, pebbles of this type are not readily available on the nearest beaches to the site and it is difficult to believe that the smallest pebbles were deliberately collected and transported.

retained traces of a single nail hole. The edges of the tiles were snapped rather than chipped and the nail holes were chipped and punched through from one face. The tiles were not complete enough to determine the original shape. One broken tile was reused as the lining of a Post-Roman grave (1654), the others were recovered from RomanoBritish or unphased contexts.

Tesserae Twelve small pieces of Purbeck limestone were identified during the excavation as possible tesserae. These are all small cubes or cuboidal fragments of limestone ranging in size from 13mm to 41mm across. The majority are worn with rounded edges on all sides and some have one or two faces that are much less eroded than the others. Most appear not to have been deliberately shaped and are likely to be accidentally broken pieces of tabular limestone. It seems unlikely they represent elements of a tessellated pavement.

Tabular limestone Seven large fragments of tabular limestone have no distinguishing marks, but may represent broken roof tile. These pieces were all (re)used as grave linings in PostRoman graves 1640, 1646 and 1654. Pivot stones? Two broken slabs of Purbeck limestone with the remains of a single perforation on the broken edge were recovered from the site. They appear to be irregularly shaped, measuring over 435mm across and 50mm thick. The perforations were originally about 60mm to 70mm in diameter and were formed by pecking from one face only; the other face in both examples appears to have broken off around the hole. The original function of these two pieces is unclear, but they may have been used as door pivot stones. Both were found reused as grave linings in Post-Roman grave 1654.

Foreign stone A single small piece of granite with no signs of working was recovered from the fill of evaluation slot 1316. The granite has a Cornish origin. Four small pieces of chalk, one burnt and the others very eroded, were recovered from slots 1031, 1068 and 1320. The nearest source of Chalk is the Corfe ridge about 4km to the north.

Block

Discussion

A single rectangular block of Purbeck limestone was retained from the excavations. There was no sign of deliberate dressing of this block, which was probably selected for retention as it had a natural fissure running across it. This block was found in wall 1450.

Worked stone was recovered from Late Bronze Age/Early Iron Age, Late Iron Age, Roman and Post-Roman contexts. This includes both objects and building materials. The range of objects present is typical of sites of Middle-Late Bronze Age and Iron Age date and includes evidence for food preparation (querns, and perhaps some of the rubbers and grinding stones) and craft activities (grinding stones, rubbers and hammerstones). Many of the items are broken and most appear to have been discarded, so reflect depositional activity rather than use.

Unidentified object One broken fragment of Purbeck limestone 56mm thick has an area pecked flat on one worn face. The original size, shape and purpose of this object is uncertain. It was recovered from the fill of Post-Roman grave 1722.

The vast majority of the stone objects were made from local materials with little evidence for longer-range contacts. Two items (one unworked) were of rock, which have a Devon/Cornwall origin.

Unworked stone

The retained stone building materials represent only a small proportion of the building stone uncovered on the site. The majority of these were reused pieces from the Post-Roman graves. The stone lining of the graves included at least one reused stone roof tile fragment and two other reused broken door pivot (?) stones. The most interesting stone object found in the Post-Roman graves was a stone anchor reused as a pillow stone in grave 1633. This gives a terminus ante quem of the Post-Roman period for this stone anchor, one of the first dated examples of this type. It is not clear whether the anchor was another reused Roman item or was in use during the Post-Roman period prior to its deposition in the grave.

Pebbles and cobbles A total of 791 unworked ovoid pebbles of chert, limestone or sandstone were retained from the excavation. These measured from about 13mm to 107mm across, with the majority measuring about 40mm to 50mm. These pebbles were found in 170 different contexts spread across the site, with the largest number from Late Bronze Age/Early Iron Age or Late Iron Age/Romano-British contexts. It is possible that a number were used as sling stones, but the small size of many would suggest this was not the case for all. It has been assumed that the pebbles were imported from local beaches (Graham et al. 2002, 42). 109

Multi-period Occupation at Football Field, Worth Matravers, Dorset The Kimmeridge Shale Sue Cullinane and Lilian Ladle (with contributions from Emma Firth) Introduction A total of 157 Kimmeridge shale objects (weighing 18,657g) were presented for analysis, although this mostly comprised undiagnostic pieces and waste. These data are presented in Table 11. A selection of significant objects has been chosen for illustration (Figs 122-3). Methodology All shale was counted and weighed by context and all objects were given Special Find (SF) numbers. No further analysis of the undiagnostic fragments was undertaken and this material was discarded. Source of material The material is most likely to have been sourced from the oil-bearing shale beds around Kimmeridge Bay (which is approximately 5.5km west of Worth Matravers). The shale outcrops are found in two sections of the coast on either side of the bay; the larger, most accessible, and therefore the most exploited exposure is the outcrop to the east beneath Clavell’s Hard (Calkin 1955, 47).

Figure 121 Kimmeridge shale objects by period

Table 11 Number of shale objects by type and period Object type (by number)

Neo/ EBA

MBA

LBA

Armlet

LBA/ EIA

MIA

LIA

Unstrat LIA

LIA/RB

RB

Totals

20

1

11

17

17

6

72

Armlet core (handmade)

3

3

Bead

1

1

Burnisher

1

1

Disc

1

1 3

3

2

33

39

Pebble

1

1

2

Platter

1

Furniture Lathe core

4

Ring

1

1

Roughout

6

3

1

Spindlewhorl

2

2

2

Vessel

1

13

3 1

7 1

1

5

Totals

1

36

2

22

Waste by weight

32g

6743g

442g

345g

80g

3

1

Unidentified object

27g

1

110

1

3

10

24

25

47

157

1609g

1263g

4147g

14,688g

Flaked Stone, Worked Stone and Kimmeridge Shale

Figure 122 Objects of Kimmeridge shale (1-15)

Figure 123 Decorated Kimmeridge shale bracelet (16)

111

Multi-period Occupation at Football Field, Worth Matravers, Dorset Results

10 Finished handmade armlet 13 Disc with lathe mounting and part-turned ring on edge of disc 14A Core with lathe-mounting and central square hole 14B Core with lathe-mounting and central square recess 14C Core with lathe-mounting and paired circular bored holes 15 Part-finished lathe-turned armlet 16 Finished lathe-turned armlet 18 Unfinished lathe-turned products 19 Finished hand-made vessel 21 Other hand-made products 22 Other lathe-turned products 23 Other products

Undiagnostic fragments These were counted and weighed, recording features such as scoring, cut edges and cut marks. In total, 53 pieces weighing 613g from Late Bronze Age/Early Iron Age contexts, 36 pieces weighing 2730g from Late Iron Age contexts and 60 pieces weighing 1865g from Late Iron Age/Romano-British contexts were recorded. There was no unworked or undiagnostic shale from Romano-British contexts, strongly suggesting that such material was taken off site or possibly used as fuel. The data for this group is available within the site archive. The fragments varied in size and condition, and included many laminated pieces together with quantities of burnt material. There were seven rounded shale pebbles which could have been selected from the shore at Kimmeridge for specific purposes, some showing signs of wear. Such pebbles were present in several phases of the site and may have been collected for fuel, or perhaps for use as rubbing stones which may explain some of the worn surfaces. In some instances, natural lamination and wear made it difficult to determine whether edges had been worked. Shale beach pebbles have also been noted at other sites, for instance at Bestwall Quarry (Cox 2009, 277), where it was suggested that they may have been brought to the site with seaweed which had been harvested for fertiliser.

This typology identifies the shale at its various stages of working for both the hand-made and the lathe-turned armlet industries, and covers other worked items. Table 11 lists the objects by phase and suggests that the industry on the site had two peaks; one in the Early Bronze Age/Early Iron Age and the second during the later Roman period. Shale objects The Late Bronze/Early Iron Age hand-worked industry is dominated by armlet production and is characterised by the presence of chiselled armlet fragments which probably snapped during the manufacturing process, together with a small number of blank discs. The subsequent latheturned industries of the Late Iron Age/Romano-British periods are recognised by large numbers of waste cores and a smaller number of broken armlet fragments. There are no complete examples of finished armlets from either industry. In addition to armlet fragments, there are small numbers of spindle whorls, a single perforated bead, tableware and possible furniture fragments.

Diagnostic worked shale The diagnostic shale falls into two groups; the first is flat pieces displaying evidence of working, such as cut edges or scoring to surfaces. The second consists of objects such as armlets, unidentified worked objects and waste from shale armlet production. This group can be sub-divided into two distinct industries; the prehistoric handmade armlet industry predominantly of the Late Bronze Age/ Early Iron Age, and the Late Iron Age/Romano-British lathe-turned industries.

Condition The majority of the shale objects are in a stable condition although many have laminated surfaces, often with only a half-profile present. It is not known whether breakage occurred post-deposition or during the production of the object. A small number of objects show evidence of burning.

Typology The shale objects have been categorised according to the Wytch Farm typology (Cox and Mills 1991, 173); some of those categories are not present at Worth and are omitted from the list.

Methods of shale working The prehistoric hand-working industry

1 Unworked shale 2 Undiagnostic worked shale 3 Blank disc with chiselled faces 3i Blank disc with chiselled face and edges 3ii Blank disc with sawn edges 5 Blank with punched hole 6 Blank with part-chiselled central hole 7 Central chiselled disc or fragments from category 6 above 8 Roughout (handmade) 9 Part finished handmade armlet/ring

Calkin (1949, 37-38) described the sequence of handmade armlet manufacture, which he recognised from the large amounts of shale and flint waste present during his excavations at Kimmeridge. Hand-made shale armlets at all stages of manufacture are present in the Football Field assemblage; these were manufactured using flint tools that leave distinctive chisel marks and conchoidal fractures on the surface of the shale. These marks were subsequently polished out of the finished armlet. Some armlet fragments in the assemblage retain conchoidal chisel marks of uniform 112

Flaked Stone, Worked Stone and Kimmeridge Shale shape on their surfaces. These pieces were probably about to be trimmed prior to polishing but cracked before this stage could be completed. SF 81 (not illustrated) is such an object; it appears to have some polishing on external edges, but the internal surface is still rough and shows chisel marks.

partially shaped ring c. 26mm in diameter again from the midden deposit (Fig. 122.4), which could have been used as a pendant. A small fragment of rim from a handmade vessel (Fig. 122.6) was recovered from pit 335, from its uppermost fill, and may have been intrusive. Middle Iron Age

The Late Iron Age and Romano-British lathe-turning industry

Two objects were recovered from fill 1163 of pit 1182; both were parts of circular armlets which were in the final stages of finishing before they snapped and were discarded. In keeping with traditional descriptions of these objects (Calkin 1955, 46) the term ‘armlet’ is used here to describe rings of varying sizes. However, one of these objects was very small (Fig. 122.5) with an internal diameter of 34mm and was presumably intended for some other purpose.

During the 1st century BC, shale workers began to employ the lathe. For a time it seems that hand-working and latheturning of shale was coterminous. This is apparent at Green Island where both hand-worked and lathe-turned shale products were present in association with imported Late Iron Age pottery (Cunliffe 1987, 46-47). The earliest type of lathe-fixing seems to have been type 14A, with 14B coming in a little later; both were superseded by the type 14C fixing, which is typically found in 2nd/3rd century AD contexts, as seen at Rope Lake Hole (Cox and Woodward, 1987, 164).

Late Iron Age The handmade armlet industry continued into the Late Iron Age with little or no discernible change; however it is well-documented that lathe-turning of shale was introduced around 50 BC (Cox and Mills, 1991, 170) and the material from Football Field appears to be no exception. The assemblage is dominated by fragments of hand worked armlets at various stages of manufacture, although there are some notable exceptions. There are two possible spindle whorls, both of which show wear to the central perforation (Fig. 122.7 and 8).

Calkin (1955, 58-64) described the stages of lathe-turning a shale armlet in great detail, discussing the various fixings and the tools used. He noted that iron tools were unsuitable and discussed the highly-specialised flint tools, which continued in use through to the 4th century AD (ibid, 60). Shale objects by period The objects were considered by period and phased accordingly; not all periods represented on site yielded shale. This material was absent from Middle Bronze Age, Late Bronze Age and Post-Roman contexts. Table 11 and Figure 121 illustrate the numbers of shale objects by type and through time.

There are few examples of lathe-turned products from Iron Age contexts, the earliest being laminated cores with a central square hole (Fig. 122, 9-10), Wytch Farm Type 14A). Another core (Fig. 122.11) has a square recess with a single central pivot on the reverse (Wytch Farm Type 14B). A fragment of a near-finished lathe-turned armlet has an internal burr still present (Fig. 122.12). A further core (SF 75 is not illustrated but is worthy of note being part of a type 14A lathe core with a small fragment of partturned ring on the edge of the disc (Wytch Farm category 13); it is the only example of a lathe core with part of the ring still attached.

Neolithic/Early Bronze Age The earliest evidence of shale exploitation during the prehistoric period is small laminated pieces weighing 12g from the middle fill (1620) of the Neolithic/Early Bronze Age ditch (Figs 10 and 11). The fragments unfortunately display no diagnostic features other than cut edges.

Iron Age occupation debris Of the 24 objects in this phase, 23 came from clearance levels on site 1; all show characteristics of the prehistoric handmade tradition and were associated with occupation debris broadly dating to the Iron Age. Although technically unstratified, these objects have been included to give an accurate representation of the scale of shale working during the prehistoric period.

Late Bronze Age/Early Iron Age The earliest observable handmade armlet production at Football Field commences in the Late Bronze Age and continues into the Early Iron Age. A total of 36 objects were recovered, of which 20 were handmade shale armlet fragments and six were roughouts prepared for shaping into armlets. Of particular note is a fragment of finished armlet from the midden deposit (Fig. 122.1) which is decorated with four parallel lines, and resembles a very similar armlet recorded at Rope Lake Hole from an 8/6th century BC context (Cox and Woodward 1987, 155; 170). Also of note were two possible spindle whorls from the midden deposit, which show signs of use and/or wear around the central perforation (Fig. 122.2-3), and a small

The majority of objects in this phase are fragments of partly finished hand-made armlets in the final stages of manufacture which were broken during this process. A single hand-made circular spindle whorl, with one side flattened and a countersunk perforation (Fig. 122.7, SF 47) was recovered from layer 338 within roundhouse 340. Also of note is a small handmade bead (Fig. 122.13, SF 113

Multi-period Occupation at Football Field, Worth Matravers, Dorset 167) which was recovered from the midden deposit. The bead has an exterior diameter of 11mm and a tiny central perforation. Although smooth, it is sub-circular and shows no sign of final polishing or decoration. Both of these objects have been identified as Late Iron Age on the basis of their typological similarity to other similarly-dated objects.

Romano-British The bulk of the Romano-British industry is represented by waste lathe cores and a number of discarded broken armlet fragments. The lathe cores are of a fairly uniform size, with an average 45mm external diameter. Of the 33 cores recorded in this phase, 31 were identified as Wytch Farm category 14C type, one of which (SF 115, Fig. 122.14) had a hole bored through the centre, perhaps for re-use as a spindle whorl.

Late Iron Age/Romano-British transition Although this phase is dominated by hand-made technology, a significant number of lathe-turned products are present. A small fragment of decorated platter is notable (Fig. 122.15, SF 95) and can be compared to a similarly-decorated fragment found during the Southampton excavations (Mills 2002, fig. 1.24/14) and also from Greyhound Yard (Woodward et al 1993, 145); these typically date to the 1st century AD. Of the finished hand-made armlet fragments, of particular interest is SF 58 (Fig. 123.16), decorated with two bands of diagonal lines, which was found in an unstratified clearance context.

The majority of the type C cores, which Calkin’s typology dated to the 2nd/3rd centuries AD, were found in contexts relating to the building (Fig. 69) which was interpreted as a shale workshop. A similar contemporaneous industrial building is described by Woodward at Rope Lake Hole (1987, 146). Discussion Shale has been recovered widely in Purbeck. Figure 125 illustrates the sites where this material occurs with dates

Figure 124 Shale working sites in the Isle of Purbeck: 1 Acton, 2 Bucknowle, 3 Chapmans Pool, 4 Church Knowle, 5 Cleavel Point, 6 Football Field, 7 Corfe Castle Sandyhills, 8 Corfe Castle Steeple, 9 East Creech, 10 Eldon’s Seat, 11 Encombe, 12 Encombe Gwyle, 13 Furzey Island, 14 Green Island 15 Gallows Gore. 16 Kimmeridge, 17 Clavell’s Hard, Metherhills, 49 Smedmore, 20 Kingston Barn, 21 Kingston Hill, 22 Kingston plantation, 23Norden, 24 Povington, 25 Swalland Farm, 26 Rope Lake Hole, 27 Smedmore Hill, 28 Swanage Whitecliff Farm, 29 Sheepsleights, 30 Tyneham, 31 Wilkeswood, 32 Worth Matravers St Aldhelm’s Head, 33 Worth Matravers west of Gallows Gore, 34 Wytch Farm

114

Flaked Stone, Worked Stone and Kimmeridge Shale than four lathe cores, two discs and one armlet fragment. The lathe cores and discs all have a square central hole which is a defining feature setting them apart from the later lathe-turned armlet material. The presence of these Wytch Farm category 14A and B cores would suggest that a small element of shale working existed during the Late Iron Age, but on a much smaller scale compared with that of the earlier and later phases.

ranging from the Late Bronze Age to the Roman period. Most are clustered near to the coast, fairly near to the Kimmeridge area however other sites are located close to the River Corfe and on the margins of Poole Harbour. The range of shale objects recovered from the excavations at Football Field not surprisingly bear similarities to material found during earlier excavations at Compact Farm (Graham et al 2002) and at other production sites in the Isle of Purbeck.

The Romano-British lathe-turned armlet industry is represented by a reasonable quantity of lathe cores and a number of broken armlet fragments. The small number of armlet fragments (6) in comparison to lathe cores (33) in this phase strongly suggests that the finished products were being produced for a wider market.

During the Late Bronze Age to Early Iron Age there appear to be significantly more roughouts and partfinished handmade armlets (Wytch Farm categories 8 and 9) compared to the number of blank discs (Wytch Farm categories 3i, 3ii, 4i, 5 and 6): such a small quantity of blank discs could possibly indicate that they were being prepared at source and then imported to the site where they were finished.

All but one of the armlets securely dated to the RomanoBritish period are undecorated with simple ovoid section. One armlet fragment and three lathe cores in this group were of fossiliferous shale. Lawson (1975, 243) discusses the difficulties in working this type of shale, suggesting that hitting a fossil in the shale would probably have snapped both the object under construction and possibly the flint chisel. Despite this, the shale could perhaps have been selected for its unusual appearance.

The handmade armlets occur in a variety of features, the largest concentration of which (seven objects) was recovered from the fill of storage pit 335. The average internal diameter of the handmade armlets is 43mm, suggesting that an individual may have been given an armlet as a child and this would have been worn into adulthood. A small number of shale armlets have been found on the wrists of adults in burials of this date, for example at Tollard Royal, where a large shale bracelet of 56mm internal diameter was found on the wrist of a male crouched burial (Wainwright, 1968, 137).

The absence of large pieces of armlet makes accurate measurements difficult; however internal diameters of measurable finished lathe-turned armlets average 46mm, loosely reflecting the average diameter of the cores. The smaller armlets were, as with the prehistoric examples, likely to have been given to children. A small number of larger armlet fragments are present, with diameters of up to 60mm. Armlets with a large diameter have been found on the upper arms of skeletons in burial contexts, for example at Maiden Castle (Wheeler, 1943, 355).

The very small number of finished armlet fragments suggests that during the Late Bronze Age/Early Iron Age there was small-scale manufacturing of hand-made armlets at the site, and that once made, these were traded elsewhere. This seems likely given the small but significant number of flint shale working tools identified at the site.

There was very little burnt shale from the Romano-British phase. It may have been used for fuel at nearby pottery production sites (Hearne (1992a; 95; Cox 2012, 197). There may have been a profit in selling or trading the waste material.

Only one Middle Iron Age feature (1182) was excavated with a resulting paucity of shale objects. This consisted of two armlet fragments and a small quantity of undiagnostic material. It is possible that production declined at the site during this period: the absence of contemporary features could suggest that the focus of occupation had perhaps moved elsewhere. The presence of burnt shale recovered from this pit suggests that it was perhaps being used as fuel. Certainly, during historical periods, shale was frequently collected for this purpose (Hutchins 1768, 111). It is possible that this material could have been consciously selected and intentionally deposited. The bulk of the burnt shale at Worth was recovered from prehistoric contexts.

While the majority of the shale objects from the later industry are lathe cores and armlet fragments, there are also a small number of other items of interest. Among these are three possible tray fragments and a fragment of possible furniture foot or leg. The pieces however are very small and only very cautious identifications can be made. It is impossible to determine whether these objects were manufactured at the site, or were acquired as traded goods. The nearest site known to have been producing such items during this period was Bucknowle (Cox, 2009, 81 and 84). The relatively limited number of finished objects throughout all periods suggests that Worth was a smallscale, workshop-level production site, which was probably engaged in the trade of finished objects.

The majority of objects produced in the Late Iron Age were handmade armlets and differ little from the material found in the Late Bronze Age/Early Iron Age. There is little definitive Late Iron Age lathe-turned material, other

115

Multi-period Occupation at Football Field, Worth Matravers, Dorset Catalogue

8. Context 1413, SF 303. Spindle whorl with wear around the central perforation. Category 21; fill of pit 1412. Late Iron Age. 9. Context 305, SF 45. Lathe core with square hole. Category 14A; fill of pit 304. Late Iron Age. 10. Context 1072, SF 148. Lathe core with square hole. Category 14A; fill of Slot 1071. Late Iron Age. 11. Context 1455, SF 460. Lathe core with square hole. Category 14B; Late Iron Age. 12. Context 305, SF 94. Fragment of near-finished armlet with internal burr. Category 15; fill of pit 304. Late Iron Age. 13. Context 1097, SF 167. Hand-made bead, unpolished, near-spherical. Category 21: fill of Slot 1098. Unstratified Late Iron Age. 14. Context 1005, SF 115. Spindle whorl made from a re-used lathe core. Category 14C and 22; clearance. Romano-British. 15. Context 314, SF 95. Fragment of decorated platter. Category 21; surface of limestone paving. Late Iron Age/Romano-British.

Fig. 122 1. Context 1108, SF 214. Fragment of finished armlet, decorated with four parallel lines; internal diameter 35mm. Category 10; fill of Slot 1107. Late Bronze Age/ Early Iron Age. 2. Context 1017, SF 122. Spindle whorl with wear around the central perforation. Category 21; fill of Slot 1015. Late Bronze Age/Early Iron Age. 3. Context 1116, SF 207. Spindle whorl with wear around the central perforation. Category 21; fill of Slot 1105. Late Bronze Age/Early Iron Age. 4. Context 1165, SF 188. Small, partially shaped ring, Category 9; fill of Slot 1164. Late Bronze Age/Early Iron Age. 5. Context 1163, SF 197. Small ring with internal diameter of 34mm, Category 10; fill of pit 1182. Middle Iron Age. 6. Context 359, SF 105. Rim fragment from a hand-made vessel. Category 19; Late Bronze Age/Early Iron Age pit 335. Possibly residual. 7. Context 338, SF 47. Spindle whorl with wear around the central perforation. Category 21; layer in House 3. Late Iron Age.

Fig. 123 16. Context 302, SF 58. Fragment of hand-made armlet decorated with two bands of diagonal lines. Category 10; clearance. Unstratified.

116

8 The Early Prehistoric Pottery

Introduction

Fabric and Form Codes

The large assemblage of pottery from this relatively small site dates from the Early Neolithic up to the end of the Roman period and comprises significant additions to the known groups of pottery from the area and beyond. The assemblage has been divided into three broad phases. The earliest prehistoric pottery encompasses material from the Early Neolithic up to the Late Bronze Age. There is a substantial assemblage of Late Bronze Age/Early Iron Age pottery which is dealt with in Chapter 9 and the Later Iron Age and Roman pottery is reported on in Chapter 10. One of the major aims of the ceramic analysis was to provide information for relative dating of the many contexts and features.

The following codes were used for analysis of all the prehistoric pottery. The abbreviations in the fabric codes denote the various categories, densities and size of inclusions: Inclusion type: F (flint); fe (ironstone pellets); G (grog); H (heathstone); M (mica); S (sand); Sh (shell); SS (shale); R (rock = limestone); U (unidentified rock) Density: 1 sparse; 2 medium; 3 dense Inclusion size: S (small); M (medium); L (large) Fabric

Methods and Aims (These methods and aims also apply to pottery discussed in Chapters 9 and 10). From 2008, the material was assessed on an annual basis, recording number and weight of sherds by context. A fully illustrated archive was compiled during the assessment stage by recording quantification details (number and weight of sherds), fabric types and featured sherds. The fabric type series was derived from that used by Lilian Ladle and Ann Woodward at Bestwall Quarry (Ladle and Woodward 2009, 201-2) with codings for rim type, base angle form and decorative type and motif. A full archive record was prepared using the Prehistoric Ceramics Research Group Guidelines for Analysis and Publication (PCRG 1995). Fabric identifications were determined by using a low-magnification eye glass (x8) and the following categories of information were recorded on pottery record sheets: Sub Site number, context details, sherd type, fabric code, abbreviated fabric description, number of sherds, weight, form code, rim/base diameter, surviving percentage of diameter, decoration code, abrasion, thickness, colour and probable date. The information was entered into an Excel spreadsheet. In addition, each featured sherd was illustrated in pencil on a separate form and drawn to scale in ink by Janet Bartlet. The early prehistoric pottery can be divided into four major groups: Neolithic, Early Bronze Age, Middle Bronze Age and Late Bronze Age. Detailed period groups are described and illustrated with relevant discussions on the cultural associations for each period, comprising the nature of the depositional practices, vessel form and function, production and long-distance links.

117

Code

Description

Abbreviation used in spreadsheet

F1

grog

G

F2

grog plus flint

GF

F3

sparse medium/large flint

1MF; 1LF

F4

sparse small flint

1SF

F5

medium flint

2MF; 3MF

F7

fine sand

FS

F8

coarse sand

CS

F9

sand plus sparse flint

S1SF; S1MF; S1LF

F10

large flint

23LF

F11

limestone

R

F12

grog plus medium flint

G2SF; G2MF; G2LF

F13

heathstone plus limestone

HR

F17

grog and limestone

GR

F18

sand plus unidentified rock

SUR

F19

sand, limestone and flint

SRF

F20

grog, limestone and sand

GRS

F21

flint, grog and sand

FGS

F22

sand plus small ironstone

Sfe1

F23

sand plus large ironstone

Sfe2

F24

sand plus flint

SF

Multi-period Occupation at Football Field, Worth Matravers, Dorset F25

grog plus sand

GS

F26

sand, limestone and mica

SRM

F27

sand plus limestone

SR

F28

sand, limestone and ironstone

SRH

F29

shell

Sh

F30

sand plus shale

SS

F31

shell and limestone

ShLS

F32

limestone plus ironstone

RH

F34

grog, sand and ironstone

GSfe

F35

grog, limestone and ironstone

GRfe

F37

sand, grog and unidentified rock

SGU

F38

limestone, ironstone, sand and grog

RHSG

R9 external expansion R10 internal expansion R11 bead rim R12 internal bevel R13 external bevel R14 lid-seating groove R15 internal groove R16 internally and externally expanded R17 flattened bead with groove under R18 bead with internal bevel R19 applied bead, internally apparent R20 collared with broad band under rim R21 incurved R22 flat or rounded top with external groove R23 internally squeezed, externally rolled R24 slight external expansion R25 flat with external bevel Base form codes

The categories F6, F14 and F16 noted at Bestwall Quarry were absent at Football Field; F15 was the same as F25 and is omitted from the list. Additionally F33 and F36 were not used.

BI right angled (90° – 100°) B2 100° – 120° B3 more than 120° Form and decoration codes

All of the pottery was hand-made, probably in small batches. As a result, the potting mixtures and resulting colours would vary from batch to batch. For the process of analysis it was logical therefore to merge some of the fabric groups: Group 1: grog only

F1

Group 2: grog and flint

F2, F12

Group 3: grog and mixture of inclusions

F17, F20, F21, F25, F34, F35

Group 4: flint only

F3, F4, F5, F9, F10

Group 5: sand only

F7, F8

Group 6: sand and mixture of inclusions

F19, F22, F23, F24, F26, F30

Group 7: limestone only

F11

Group 8: limestone and mixture of inclusions

F13, F27, F28, F32, F38, F31

Group 9: shell

F29

Group 10: unidentified rock

F18, F37

R rim B base BA base angle W wall DW decorated wall H handle AC applied cordon PC pinched cordon FN fingernail impression FP fingerprint impression IC incised lines TC toothed comb impressions S stabs The numbers and weights of early prehistoric pottery are detailed on Table 12; however because of the small, Table 12 Occurrence of early Prehistoric pottery Period

Number of sherds

Weight (g.)

% number

% weight

Av. weight

Neolithic

36

132

5

2

4g

Rim form codes

Neo/ EBA

120

297

16

4

2g

R1 simple rounded R2 flattened R3 hooked R4 beaded with upstanding rim R5 proto-bead with groove R6 upright and rolled R7 everted R8 short upstanding

EBA

41

154

5

2

4g

EBA/ MBA

45

237

6

3

5g

MBA

171

656

23

9

4g

LBA

225

5889

45

80

17.5g

Totals

748

7365

118

The Early Prehistoric Pottery Description of illustrated sherds

fragmentary and abraded condition of many of the sherds (and the mixed contexts), it was only possible to allocate a broad period to some of them. It has been possible to tabulate the use of fabric types through time (Table 13). It can be seen that there were marked changes in the potting mixes over a 2000 year period. It is not possible to say whether this was due to innovation, necessity or culture. For example, by the Late Bronze Age, grog-only tempered fabrics formed a small part of the assemblage, and flint alone as filler ceased to be used. By this time in general, the mixes incorporated a larger variety of fillers. The only mix to be used consistently was F25 (grog and sand) and indeed this became progressively more frequent. In many cases, due to the extremely small sherd size (and lack of featured sherds), it was difficult and sometimes impossible to assign a period with certainty to the material and in these instances, fabric type was often used as a potential indicator. In total, 748 sherds weighing 7365g were recorded. The pottery rims were also tabulated (Table 14); during the whole period, nine types were in use: the simple, rounded rim was by far the most common, and was the only type to be used throughout.

Fig. 125: Early Neolithic 1. Evaluation slot 1270, lower fill 1275 (Fig. 9). Two conjoining sherds from a possible carinated bowl with an externally thickened and rolled rim (R6). F2. 2. Trench 1, layer 1620 (Fig. 11). Two conjoining sherds from a simple open bowl with an externally thickened and rolled rim (R6). F4. 3. Evaluation slot 1320, lowest fill 1333 (section not illustrated). Externally expanded rim from an open bowl (R9). F2. 4. Trench 2, layer 1621 (Fig. 13). Simple tapered rim from an open bowl (R1). F25 5. Evaluation slot 1305, lowest fill 1307 (Fig. 8). Simple tapered rim from an open bowl (R1). The fabric is uncommon: voids were apparent where the unidentified rock had fallen out. F18. 6. Evaluation slot 1305, lowest fill 1307. Tip of rim damaged but likely to be simple, from a probable open bowl; diagonal fingernail impressions on outer rim (R1). F1. 7. Evaluation slot 1313, lowest fill 1337 (section not illustrated). Wall sherd from slightly carinated bowl with horizontal row of diagonal fingernail impressions. F2.

Neolithic and Early Bronze Age Pottery Pottery from the Neolithic enclosure ditch and evaluation slots Without exception, all of the earliest material was derived from layers within enclosure ditch 1332 and from evaluation trenches excavated to determine stratigraphy underneath the consolidated cobble layer 1039. All layers within the evaluation trenches were found to be part of the ditch fill.

Fig. 130: Middle/Late Neolithic 1. Trench 2, layer 1622 (Fig. 13). Angular rim sherd from a Mortlake style, bipartite Peterborough ware bowl; irregular diagonal slashes decorate the rim, which has an external bevel (R13). FI.

Table 13 Pottery fabrics through time (with numbers of confidently identifiable sherds) Period

F1

F2

F3

F4

F5

F7

F8

F9

F11

F17

F18

F19

F20

F25

F27

F29

F34

Group

1

2

4

4

4

5

5

4

7

3

10

6

3

3

8

9

3

Neo

45

21

9

10

1

1

1

1

3

2

2

20

2

EBA

20

8

2

7

MBA

106

5

13

26

3

LBA

1

8

41

79

6

12

1

Totals

173

14

61

135

19

21

4

36

1 11

13

14

5

6

1

12

7

4

17

1

25

18

12

Table 14 Occurrence of rim forms (Neolithic to Late Bronze Age) Period

R1

R2

Neolithic

3

Early Bronze Age

5

4

Middle Bronze Age

2

1

Late Bronze Age

13

5

Totals

23

10

R3

R5

R6

R9

2

1

R10

1

1

R12

1

Totals

1

7

1

12

1

4 27

2

1

2

2

2

2

3

3

2

4

119

R13

2

50

Multi-period Occupation at Football Field, Worth Matravers, Dorset and simple (R1), one of which had exterior diagonal slashes. Two were upright and externally rolled (R6), and two had an external expansion; one of these also had a defined bevel (R9 and R13). Decoration was restricted to fingernail impressions. The later Neolithic pottery belongs to the Peterborough Ware tradition of decorated round and flat-bottomed bowls; the Football Field assemblage displays fingernail, fingerprint and stab decoration motifs. Cultural associations Without exception, the pottery was all derived from the infill of ditch 1332 and was generally located towards the base of the feature. The small sherd size and moderate level of abrasion point towards material which had already been broken up before it entered the ditch fill. The earliest pottery fits well within the defined vessel forms for southern Britain (Cleal 2004, 165) which encompass open, closed, slack-sided and S-shaped profile roundbottomed bowls. The nearest comparable Early Neolithic pottery is from Bestwall Quarry (Woodward 2009a, 202). This came from pits beneath a Neolithic monument and was also defined by a series of open bowls. Other assemblages of plain bowls occur at Maiden Castle (Cleal 1991a, 171-185), Flagstones, Dorchester (Cleal 1997, 88) and Rowden (Cleal 1991b, 98-9). The single instance of a slightly carinated vessel bearing diagonal fingernail impressions just below the carination may be a later vessel in the sequence (Gibson, 2002, 74). Cleal (1995, fig. 16.2), in her seminal study of Wessex pottery fabrics, suggests that the earliest Neolithic fabrics are grog-free. At Football Field and at Bestwall Quarry (Woodward 2009,

Figure 125 Early Neolithic pottery from Ditch 1332

2. Evaluation slot 1320, lowest fill 1333 (section not illustrated). Peterborough ware wall sherd with cuneiform stab decoration possibly from underneath a carination. F1. 3. Trench 2, layer 1621 (Fig. 13). Peterborough ware wall sherd with diagonal fingernail decoration from rounded part of a probable bowl. F1. 4. Evaluation slot 1305, middle fill 1307 (Fig. 9). Mortlake style, Peterborough ware wall sherd with deep fingerprint impression. F3. 5. Trench 2, layer 1622 (Fig. 13). Base sherd from possible Fengate-type urn. F1.

Table 15 Neolithic pottery: occurrence of fabric types

Fabric and form The Neolithic pottery assemblage derived from within the fillings of the enclosure ditch 1332 but was too small for meaningful analyses. It can be seen from Table 15 that a minimum of 50 vessels were identified. The sherds generally were very small and moderately abraded with an average weight of 6g; eight sherds were less than 4g. Twelve fabric types were recognised and fall into eight groups. The majority contained inclusions of grog (33 of the 50 vessels), either grog alone or with the addition of other material. In seven instances, the fabric contained only flint; nine contained sand, two of these included an unidentified rock and one fabric contained limestone. The earliest assemblage was characterised by simple open bowls (Fig. 125. 2, 3, 4, 5 and 6) with single instances of a closed form (Fig. 125.1) and a carinated form (Fig. 125.7). Rim forms were not extensive: three were slightly pointed 120

Fabric type

Group

Fabric description

Minimum number of vessels

F1

1

Grog

4

F2

2

Grog plus flint

13

F3

4

Sparse medium/large flint

3

F4

4

Sparse small flint

4

F7

5

Fine sand

1

F8

5

Coarse sand

4

F11

7

Limestone

1

F17

3

Grog and limestone

1

F18

10

Sand plus unidentified rock

2

F20

3

Grog. Limestone and sand

1

F25

3

Grog plus sand

14

F27

8

Sand plus limestone

2

The Early Prehistoric Pottery Fig. 127 1. Trench 1, layer 1620 (Fig. 11). Beaker wall sherd decorated with a pair of stab-and- drag triangular shaped impressions. F1. 2. Trench 1, layer 1620 (Fig. 11). Beaker shoulder sherd decorated with vertical fingernail impression on a slightly raised cordon. F1. 3. Trench 1, layer 1620 (Fig. 11). Very abraded Beaker wall sherd decorated with three clear lines of twisted cord and two further lines visible on the sherd break. F1. 4. Trench 2, layer 1622 (Fig. 13). Beaker wall sherd decorated with lines of twisted cord. F25. 5. Trench 1, layer 1620 (Fig. 11). Carinated Beaker wall sherd with remnant fingernail decoration. F1. 6. Posthole 1616, fill 1615 within Trench 2 (Fig. 13). Beaker rounded wall sherd decorated with horizontal rows of shallow lentoid stabs. F25. 7. Posthole 1616, fill 1615 within Trench 2 (Fig. 13). Beaker abraded rim sherd (R1) with faint incised line under rim and deeply impressed diagonal line beneath, perhaps part of a lattice design. F2. 8. Evaluation slot 1270, middle fill 1275 (Fig. 9). Slightly hooked Beaker rim sherd (R3). F5. 9. Evaluation slot 1270, middle fill 1275 (Fig. 9). Slightly flaring simple Beaker rim (R1). F1. 10. Evaluation slot 1305, middle fill 1307 (Fig.9). Flattened incurved rim (R2) from a possible food vessel. F1. 11. Evaluation slot 1305, lowest fill 1310 (Fig. 9). Simple Beaker rim (R1) with a slight external thickening. F1. 12. Evaluation slot 1305, lowest fill 1310 (Fig. 9). Rim with an internal bevel (R12) from a possible food vessel. F25. 13. Evaluation slot 1313, lowest fill 1337. Beaker wall sherd with slight incised horizontal lines, a poorly executed chevron and fingernail decoration. F2. 14. Evaluation slot 1313, middle fill 1322. Beaker flat rim sherd (R2) with five horizontal rows of twisted cord decoration. F1. 15. Evaluation slot 1320, lowest fill 1333. Beaker flat rim sherd (R2). F25.

Figure 126 Middle/Late Neolithic pottery from Ditch 1332

202) however, grog-tempered fabrics are dominant: it is possible that the potting mixes in south-east Dorset were developed independently. Although the numbers of identifiable Middle and Late Neolithic pottery sherds were small, they are characterised by decorative schemes. The single rim sherd from a Mortlake style bowl together with the stab, fingernail and fingerprint impressions on the other sherds fit well into the late sequence. Similar sherds and parts of larger vessels were recorded by Cleal on Cranborne Chase (Cleal 1991c, fig. 7.10; P110, P118, P155). The base sherd from the Fengate-style urn illustrates the transition between the round bottomed bowl sequence and new-style flat bases which became the norm in succeeding periods. The decorated single Mortlake sherd with a deep fingerprint impression can be compared to a potentially similar vessel from Bucknowle, near Corfe Castle (Woodward 2009b, fig. 61.8). Unfortunately because of the small size of the sherds it was not possible to estimate vessel size or capacity.

Beaker Pottery from Other Features Other Beaker sherds were recorded from a posthole (1257) and clear-back under the cobble layer (1039) on Site 2. Further residual sherds were recorded from Post-Roman grave fills and general clearance on Site 3. Fig. 128 1. Posthole 1257, lower fill 1258 (Fig. 17). Beaker slightly everted simple rim sherd (R1). F7. 2. Posthole 1257, lower fill 1258 (Fig. 17). Beaker wall sherd with lightly impressed toothed comb decoration. F20. 3. Clear-back 1559 under cobbles 1039. Beaker wall sherd with two close rows of toothed-comb decoration. F7. 4. Grave 1654, fill 1655. Beaker simple rim sherd (R1) with toothed-comb decoration under rim. F1.

Late Neolithic/Early Bronze Age Pottery Material from this period was from the Trenches excavated across Ditch 1332, evaluation slots and from a posthole within the ditch fill. 121

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 127 Late Neolithic/Early Bronze Age pottery from Trenches 1 and 2, evaluation pits and posthole 1616

5. Unstratified layer 1663 from clearance, Site 3. Beaker wall sherd, with slight thickening and row of light diagonal fingernail impressions; the inside of the sherds bears a horizontal fingernail impression. F1. 6. Unstratified layer 1663 from clearance, Site 3. Rounded Beaker wall sherd with lentoid stab decoration. F1. 7. Grave 1682, fill 1683. Slightly flattened and everted Beaker rim (R2) with internal geometric incised line. F1. 8. Unstratified layer 1718 from lower level clearance Site 3. Rounded Beaker wall sherd with lightly incised horizontal line and small vertical stabs. F1. 9. Grave 1737, fill 1734. Beaker rim sherd with external fold and bevel (R13). F1.

Table 16 Late Neolithic/Early Bronze Age pottery: occurrence of fabric types Fabric type

Group

Fabric description

Minimum number of vessels

F1

1

Grog

24

F2

2

Grog plus flint

7

F4

4

Sparse small flint

2

F5

4

Medium flint

1

F7

5

Fine sand

4

F17

3

Grog and limestone

1

Fabric and form

F20

3

2

The Later Neolithic and Early Bronze Age pottery was derived from a variety of contexts on Sites 1 and 2; Table 16

Grog. Limestone and sand

F25

3

Grog plus sand

9

122

The Early Prehistoric Pottery impressions as well as stabs and incised lines. Similarly, it is not possible to classify and date the assemblage; however it is likely that the Football Field material had a currency of use from approximately 2500 -1700 BC. Nature of deposits The pottery was derived from a number of contexts on Sites 2 and 3. On Site 2, the majority of sherds were found in evaluation slots 1270, 1305, 1313 and 1320 (Fig. 7) and layers 1620/1622 of Trenches 1 and 2 (Figs 11 and 13), all comprising material accumulated in the initial infillings of ditch 1332. The Beaker pottery was often associated with earlier sherds, a probable indication (together with the very small sherd size) that the sherds originated from nearby locations and were incorporated into the accumulating ditch fill. Two sherds (Fig. 128.5 and 6) came from a posthole (1615) within the ditch filling (Fig. 13) which also contained a further three small plain sherds. Two sherds were derived from posthole 1257 (Fig 128.1-2 and Fig. 17). The Site 3 material associated with the Post-Roman cemetery (Fig. 128.4-9) was unstratified within the soil matrix, but is likely to have originated from unidentified but nearby activities. Cultural associations Because of the small sherd size and lack of vessel profiles, it was not possible to accurately assign and characterise the Football Field material. The nearest large assemblage comes from Bestwall Quarry, which lies about 8km north of the site; some of the decorative motifs found at Bestwall can be seen in the Football Field assemblage, including toothed comb and fingernail impressions (Woodward 2009, figs 135, 136, 139 and 140). Twisted cord was one of the most commonly used Beaker decorative techniques (Gibson and Woods 1997, 133): this was absent at Bestwall, but was recorded on three sherds from Football Field. The sharp carination on Fig. 127.5 can be compared to a vessel from Mount Pleasant, Dorchester (Longworth 1979, fig. 51 P215).

Figure 128 Beaker pottery from other features

demonstrates that a minimum of 50 vessels were present and were identified as Beakers. Only one profile could be reconstructed (Fig. 128.7); this sherd came from a beaker with a slightly out-turned rim and an approximate diameter of 125mm. The 11 surviving rim forms consisted of four simple and slightly rounded (R1) and four where the rim had been flattened (R2). There were single examples of a hooked rim (R3) and internal (R12) and external (R13) bevels. The sherds generally were very small and moderately abraded with an average weight of 3.5g.

Middle Bronze Age Pottery

Eight fabric types were in use: the overwhelming majority (43), contained inclusions of grog (F1, F2, F17, F20 and F25). Seven of the 50 vessels did not contain grog, the fabric types being F4, F5 and F7. This can be compared to the Beaker assemblage from Bestwall Quarry where 58 out of the 64 vessels contained grog inclusions (Woodward, 2009, 204). The fabrics are similar to those in use in the earlier Neolithic period but the mixes are less varied. All of the sherds came from thin-walled vessels with thicknesses varying from 4mm to 8mm.

The small assemblage of Middle Bronze Age pottery all came from contexts on Site 3 and comprised 164 sherds weighing 604g. The material is likely to have derived from activity nearby. Fabric and form Six fabric types were identified: F1, F2, F17, F20, F25 and F27, the latter being the only type not to include grog in the mix. Fabric type F1 dominated the assemblage (73%): the remainder were much lower percentages. The average sherd size was very small (3.6g) and most pieces displayed heavy levels of abrasion commensurate with movement within the ploughsoil. At least 49 different vessels were

Due to the small sherd size it has not been possible to construct vessel profiles, but the range of decorative motifs include twisted cord, toothed comb and fingernail 123

Multi-period Occupation at Football Field, Worth Matravers, Dorset represented (Table 17), including bucket and globular urns.

Fig. 129 1. Clamp 1700, fill 1701. Shoulder sherd from a biconical urn with raised cordon with a row of diagonal, crisp fingernail impressions, probably executed by a left hand. F1. 2. Unstratified clearback 1663 over the cemetery area. Flattened rim sherd (R2) with slight fingerprint impression below rim. F1. 3. Unstratified clearback 1663 over the cemetery area. Shoulder sherd with a raised fingernail-impressed cordon. F25. 4. Unstratified clearback 1663 over the cemetery area. Rounded shoulder sherd with fingernail impression (probable row). F1. 5. Fill 1689 under pavement 1695, probable extension of clamp 1700. Rim sherd with internal bevel (R12). F1. 6. Clearback 1659 around grave 1649. Wall sherd with narrow incised grooves. F1.

There were 12 diagnostic sherds in total. Four rim sherds were identified, comprising two simple, rounded types (R1), a single flat rim (R2) and one slightly everted rim (R12). Decorative motifs were recorded on six sherds and consisted of fingerprint and fingernail impressions, some on raised cordons; there were also examples of incised lines. Two small base sherds were identified. Most of the pottery was unstratified and was found during handclearance of the site and in the Post-Roman grave fills. Only one feature (1700), a possible pottery making clamp, was identified (position located on Fig. 94); similar pottery was recovered from this and from underneath the adjacent pavement 1695 and the fill of grave 1685. Table 17 Middle Bronze Age pottery: occurrence of fabric types and minimum number of vessels Fabric type

Group

Fabric description

Minimum number of vessels

F1

1

Grog

20

F2

2

Grog plus flint

4

F17

3

Grog plus limestone

5

F20

3

Grog plus limestone and sand

5

F25

3

Grog plus sand

13

F27

8

Sand and limestone

2

Cultural associations The very large, well-dated Middle Bronze Age assemblage from Bestwall Quarry belongs to the Deverel-Rimbury pottery tradition and was reported in detail by Woodward (2009, 213-271). There is unfortunately a lack of similarly dated pottery from elsewhere in Purbeck (ibid, 241). Ann Woodward and Stuart Needham have recently reassessed the Eldon’s Seat assemblage in the light of recent research and have placed some of the Period 1 material (originally defined as Ultimate Bronze Age) into the Middle Bronze Age Deverel-Rimbury sequence. Nine

Figure 129 Middle Bronze Age pottery from clamp 1700 and unstratified contexts on Site 3

124

The Early Prehistoric Pottery were subtle variants of those already identified. Because of the larger numbers of sherds it has been possible to detail the relative proportions of the major fabric groups during this period (Fig. 130). The vast majority of sherds belong to fabric groups 3 and 5 which consist of potting mixes that include grog and sand. There appears to be a marked increase in the variety of potting mixtures used.

samples were submitted for radiocarbon dating and six of these gave dates between the 16th and 11th centuries BC. (Woodward and Needham forthcoming) A single sherd of grog-tempered, fingernail-decorated pottery was recorded during evaluation work at Belle Vue, Swanage (Ladle 2010). Most of the small number of diagnostic sherds from Football Field probably came from coarse-ware bucket urns decorated with fingernailimpressed cordons. Two sherds (Fig. 129.3-4) may have derived from globular urns, indicated by the rounded profile in the former and incised decoration on the latter.

The number of rim forms also increased; the two simplest types continued to be the commonest: the rounded R1 (11 examples), followed by the flattened R2 (4 examples). There are two examples of R5, R9 and R12 and single examples of R10 and R6. Similar changes were also recorded in the Late Bronze Age pottery assemblage from Bestwall Quarry (Woodward 2009, 224-5). The Football Field vessels follow the same code forms as the Bestwall examples (ibid., 245 and 247).

Late Bronze Age Pottery The assemblage of Late Bronze Age pottery comprised 748 sherds weighing 7365g and derived from contexts associated with post-built roundhouse 1 on Site 2, a pit and posthole on Site 4 and unstratified material collected from clearance under the cobble layer 1039. There were an estimated 73 vessels.

Form 1: shouldered bowl Form 2: open bowl Form 3: biconical bowl Form 4: jar with angular high shoulder Form 5: shouldered jar with sinuous profile Form 6: jar with high, rounded shoulder Form 7: high-shouldered jar with internally bevelled rim Form 8: ovoid jar Form 9: cup

Fabric and form It was possible, because of the size of this assemblage, to carry out meaningful analysis. Thirteen fabric types were identified and show a marked change in preferred potting mixes from the previous period (Table 18). Grog alone and grog with flint (F1 and F2) was much rarer, accounting for only 0.6% of the assemblage. Although grog continued to be used, it was always mixed with other materials. New fabrics were introduced (F22, F29, F32 and F34), and apart from F29 which was tempered with shell, the other fabrics

Site 2; post-built roundhouse 1 Pottery was recorded in a series of postholes belonging to a post-built roundhouse on Site 2 (Fig. 20). This was sealed by the cobble layer 1039 dating to the Late Bronze Age/Early Iron Age, and pottery sherds associated with dating of the construction of this house are likely to have become incorporated into the Late Bronze Age surfaces.

Table 18 Late Bronze Age pottery: occurrence of fabric types and minimum number of vessels Fabric type

Group

Fabric description

Minimum number of vessels

F1

1

Grog

1

F2

2

Grog plus flint

1

F7

5

Fine sand

11

F8

5

Coarse sand

4

F11

7

Limestone

5

F17

3

Grog and limestone

5

F20

3

Grog. Limestone and sand

7

F22

6

Sand plus small ironstone

2

F25

3

Grog plus sand

20

F27

8

Sand plus limestone

9

F29

9

shell

6

F32

8

Limestone plus ironstone

1

F34

3

Grog, sand and ironstone

1

The largest number of sherds (87 weighing 2933g), many of them very large chunks, derived from the probable doorway posthole 1484 (fill 1581) and its associated fills 1567 and 1568 (Fig. 21). Substantial remains of three large jars were found in the fills. Two are illustrated below; the third, from which there were no featured sherds, displayed traces of red coating.

Figure 130 Late Bronze Age pottery fabrics by number

125

Multi-period Occupation at Football Field, Worth Matravers, Dorset 3. Posthole 1596, fill 1597. Simple, upright rim (R1) from an open bowl. Form 2, F25.

Fig. 131 1. Posthole 1484, fill 1567. Abraded everted rim and body sherds (R1) from a coarse ware shouldered jar. Form 5, F20. 2. Posthole 1484, fill 1567. Very large high-shouldered jar with external expanded rim (R10), horizontal row of fingerprints on the shoulder and traces of wiping on the body. Form 7, F25. 3. Posthole 1484, associated fill 1567. Small, unusual rim sherd, eternally expanded (R9), F25. 4. Posthole 1484, fill 1581. Slightly everted rim sherd (R1) from a shouldered jar. Form 5, F20.

The remaining structural postholes contained smaller numbers of pottery, and the sherd sizes were also small. Fig. 133 1. Posthole 1511, fill 1512. Small, flat rim sherd (R2), the back displaying apparently deliberate, crossed lines. F7. 2. Posthole 1511, fill 1512. Rim sherd from shouldered bowl, simple upright rim (R5) with a proto-bead and traces of a slight external groove. Slight trace of red coat. Form 1, F25. 3. Posthole 1511, fill 1512. Rim sherd from shouldered bowl (R1). External red coat with traces of wiping and two crossing lines of black (painted?) decoration. Form 1, F7. 4. Posthole 1551, fill 1554. Rim sherd with external expansion (R9), possibly from an open bowl. Form 2, F25. 5. Posthole 1551, fill 1554. Simple upright rim sherd (R2), probably from an open bowl. Form 2, F25. 6. Posthole1577, fill 1578. Simple upright rim sherd (R1) with traces of red coat, probably from an open bowl. Form 2, F25.

The other doorway posthole (1596, Fig. 21) yielded 16 sherds weighing 1006g from four vessels, the larger sherds deriving from coarse ware jars and the finer sherds from probable bowls. Fig. 132 1. Posthole 1596, fill 1597. Large base sherd from probable high shouldered jar with external, vertical wiping. F 25. 2. Posthole 1596, fill 1597. Wall sherd decorated with horizontal row of fingerprints, probably from the above vessel. F25.

Figure 131 Late Bronze Age pottery from House 1, posthole 1484

126

The Early Prehistoric Pottery

Figure 132 Late Bronze Age pottery from House 1, posthole 1596

2. Layer 375, underneath cobble layer on Site 1. Wall sherd probably from the same vessel, decorated with fine incised lines. F11. 3. Layer 375, underneath cobble layer on Site 1. Rim sherd (R12) with well-defined internal bevel. Probably from a high-shouldered jar. Form 7, F27. 4. Layer 368, underneath cobble layer on Site 1. Rim sherd (R6), upright and rolled with slight external expansion. Probably from a biconical bowl. Form 3, F32. 5. Layer 368, underneath cobble layer on Site 1. Slightly everted jar rim (R1). Probably Form 4. F27. 6. Layer 375, underneath cobble layer on Site 1. Rounded and internally bevelled rim (R12), probably from a jar, Form 4. F27. 7. Layer 368, underneath cobble layer on Site 1. Base sherd (B3), probably from a thick-walled jar. F7. 8. Layer 362, underneath cobble layer on Site 1. Base sherd (B3) with very little wall, internal groove where base and wall meet. F7. 9. Under cobble layer 1039, Site 2. Rounded wall sherd with four incised lines. F29.

On Site 4, posthole 1761 and pit 1750 were recorded in the water pipe trench (location Fig. 5 and section Fig. 18). The posthole appeared to cut the pit, and the pottery from the features (2 and 34 sherds respectively) was similar. At least four vessels were represented. Fig. 134 1. Pit 1750, fill 1764. Near-rim and shoulder plus large base sherd from a wide-mouthed, round-shouldered jar. F11 (shell has leached out of the limestone in places). Form 6. There were sherds from at least two other vessels in this pit. 2. Posthole 1761, fill 1766. Upright rim sherd (R2) with traces of a slight cordon below the rim, most likely from a straight-sided jar. F11. A number of Late Bronze Age sherds (162 weighing 1108g) were recovered after the cobbles (1039) were removed on Site 2, and from layers identified underneath the cobbles on Site 1. Fig. 135

Cultural associations

1. Layer 375, underneath cobble layer on Site 1. Base sherd (B1) from a thin-walled, upright vessel. There is a trace of residue on the inside. F11.

When John Barrett assessed assemblages of Late Bronze Age pottery from central southern England in the late 127

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 133 Late Bronze Age pottery from postholes 1511, 1551 and 1557 associated with House 1

1970s (Barrett 1980), his seminal study determined that crucial ceramic changes had occurred in the late second millennium BC. The globular and bucket urns of the Middle Bronze Age developed into new forms which were dominated by a characteristic range of vessels comprising mostly undecorated jars, bowls and to a lesser degree cups, often thin-walled, generically known as Post-Deverel Rimbury plainware dating to the 10th/9th centuries BC.

and have placed it in a Post-Deverel Rimbury plainware tradition; two radiocarbon dates gave the parameters of this PDR sequence between the 11th and early 9th centuries BC. The large assemblage from Bestwall Quarry (Woodward 2009, 244-253) approximately 8.5km north of the Worth site, was dated by radiocarbon from 1020-930 cal BC to 825-750 cal BC. The inland site at Coburg Road near Dorchester yielded a small assemblage of Late Bronze Age pottery (Cleal 1992, 36-38). No radiocarbon samples were submitted and dating was by comparison with sites in the Thames Valley which date to the early 11th /10th centuries BC.The settlement and pottery production site at Tinney’s Lane, Sherborne in north Dorset is approximately 50 km north west of Worth Matravers and had a short period of use between the 12th and 11th centuries BC. The manufactured pottery forms (Tyler and Woodward, 2012, 224-231) comprised jars, bowls and a few cups which are stylistically similar to vessels from Football Field.

The assemblage from Football Field, although small, conforms to these changes and can be compared to material from other Dorset sites including Eldon’s Seat (Cunliffe and Phillipson 1968), Bestwall Quarry (Woodward 2009), Coburg Road, Dorchester (Cleal 1992) and Tinney’s Lane, Sherborne (Best et al 2012) and to similarly dated sites further afield. The site at Eldon’s Seat lies approximately 3.2 kilometres east of Worth Matravers and was excavated between 1963 and 1966. The pottery was divided into two groups: Period 1 assigned to the Ultimate Late Bronze Age with bucket urns, carinated jars and bowls, and Period 2 vessels which were assigned to the Early Iron Age (Cunliffe 1968, 211). Ann Woodward and Stuart Needham (Woodward and Needham forthcoming) have reassigned some of the Eldon’s Seat assemblage in the light of recent research

Outside the immediate area, the very large assemblage from Runnymede Bridge, Berkshire (Needham 1996) yielded similar jars bowls and cups, but Needham (ibid, 106) stressed that vessel attribution was often impossible due to incomplete profiles and the study concentrated on style, fabric and finish. Nevertheless, vessel comparisons 128

The Early Prehistoric Pottery

Figure 134 Late Bronze Age pottery from pit 1750 and posthole 1761

radiocarbon determinations from posthole 1593 returned a date of 790-485 cal BC at 95% probability (SUERC61170), placing the demolition of Roundhouse 1 at the very end of the conventional Late Bronze Age. At Bestwall Quarry, occupation had ceased by the 790s BC (Bayliss et al 2009, 149); activity at Worth would seem to have carried on a little longer.

can be made with the Football Field assemblage. The site, which comprised a large area of dense, cultural deposits, was dated by radiocarbon from 1020- 790 cal BC (Needham 1996, table 63, 230). Conclusions The majority of the vessels are likely to have been locallymade using readily accessible material for temper. The pottery forms correspond well to those recorded from other sites and the pots were used for storage, food preparation and serving.

The deposition of large amounts of pottery in posthole 1484 (48 sherds weighing 1921g) consisting of large chunks from at least four vessels may have constituted a closing deposit for the roundhouse. The construction of the limestone cobbled area over the demolished structure completely covered any building remnants and effectively created a physical barrier between what had gone before – and what was to follow.

The pottery assemblage, although not large, provides much-needed information about the traditions associated with its manufacture and use in this part of Purbeck. The

129

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 135 Late Bronze Age pottery under cobble layer 1039

130

9 The Late Bronze Age/Early Iron Age Pottery Lilian Ladle Introduction

from a posthole from the Late Bronze Age house, from a sample taken from just above the cobble layer 1039, from the midden matrix and from a pit containing a fragment of glass (see Henderson, Chapter 11). Despite the problems associated with dating material from this period, the returned dates confirmed activity between 790-363 cal BC. This fits well with the approximate end of the Late Bronze Age falling around 800 BC and the Middle Iron Age starting around 350 BC.

Excavation over three seasons produced 6721 sherds of pottery from this period, weighing 65,458g which largely derived from a midden deposit and features associated with this, on Sites 1, 2 and 4. Pottery of this period was completely absent from Site 3. Table 19 details pottery from clearance levels, midden layers and features. Table 20 shows numbers and weights of pottery from layers of the midden which had accumulated over the cobble deposit 1039. The assemblage is dominated by a range of pottery comprising bowls and jars in the All Cannings Cross tradition which is conventionally dated between the 8th and 6th centuries BC. This type of site is exceptionally rare in Dorset, and therefore the pottery has been considered in detail.

Fabric and Form In total, 27 different fabric forms were identified, a large increase when compared to the 11 fabrics identified in the previous period. Table 21 shows the occurrence of the fabric groupings which are described in Chapter 8. It can be seen that fabric groups 3 and 5 remained dominant. Within Group 3, two new varieties of mix (F21 and F35) were developed, as they were within other groups. For example in Group 4, fabric F10 which incorporated flint of various sizes was introduced. The Group 6 fabrics, which included sand as a base with the addition of other material diversified and fabrics F19, F23, F24, F26 and F30 were added to the repertoire. Two new fabrics F13 and F28 were added to Group 8 and a new group was added to the range (Group 10). Two fabric types were abandoned: the grog-only fabric (F1) which had been the preferred mix in earlier periods and the flint-tempered F3. Figure 136 compares the fabric groups of this period.

Radiocarbon Dating Samples of animal bone from four features containing Late Bronze Age/Early Iron Age pottery were submitted for radiocarbon dating (see Chapter 1). The bones came Table 19 Total numbers and weights of Late Bronze Age/ Early Iron Age pottery sherds from all contexts Feature or context

Number

Weight

clearance

271

3143

Midden layers

4381

37015

Pits

1769

20798

Scoops and postholes

59

384

Deposits and gullies

149

1617

Other minor features

92

1464

Totals

6721

65458

Fabric analysis is not an exact science, unless petrology is used (this was not an option for the Football Field assemblage). When the pots were being made, the potting mixes would vary from batch to batch. No doubt the potters were well aware of the correct proportions needed to fire pots successfully, and indeed the subtle additions of the inclusions probably had a bearing on the type of vessel being made and its ultimate purpose. During analysis there was a problem in that sometimes sherds from a single vessel appeared to be of different fabrics: caution had to be exercised in assigning fabric type.

Table 20 Numbers and weights of Late Bronze Age/Early Iron Age pottery sherds from the midden layers depth (m)

No

Wt (g)

Average sherd wt

layer 1

0.1-0.2

1801

14850

8.2

layer 2

0.1

994

7852

7.8

layer 3

0.05-0.1

870

8695

9.9

layer 4

0.08

224

1658

7.4

layer 5

0.3

425

3422

8

layer 6

0.01

67

538

8

4381

37015

Totals

Rims With regard to rim types (Table 22), from the 719 rims available for study, 23 different forms were identified with the simple rim form R1 dominating the assemblage and the flat R2 rim being the second most frequent. The numbers of rims from forms R5, R6, R7, R9 and R11 varied between 14 and 23 and the remainder were in single figures, with eight types only having one example (R3, R4, R12, R14, 131

Multi-period Occupation at Football Field, Worth Matravers, Dorset Table 21 Late Bronze Age/Early Iron Age pottery: occurrence of fabric types Group

No sherds

Fabric description

F2

2

16

Grog plus flint

F4

4

4

Sparse small flint

F5

4

7

Medium flint

F7

5

1755

Fine sand

F8

5

371

Coarse sand

F10

4

3

Large flint

F11

7

364

Limestone

F13

8

15

Heathstone plus limestone

F17

3

214

Grog and limestone

F18

10

9

Sand plus unidentified rock

F19

6

1

Sand, limestone and flint

F20

3

472

Grog, limestone and sand

F21

3

3

Flint, grog and sand

F22

6

310

Sand plus small ironstone

Rim form

Nos.

Description

F23

6

255

Sand plus large ironstone

R1

362

Simple, rounded

F24

6

10

Sand plus flint

R2

105

Flattened

F25

3

1787

Grog plus sand

R3

1

Hooked

F26

6

38

Sand, limestone and mica

R4

1

Beaded with upstanding rim

F27

8

666

Sand plus limestone

R5

17

Proto-bead with groove

F28

8

114

Sand, Limestone plus ironstone

R6

18

Upright and rolled

F29

9

85

Shell

R7

17

Everted

F30

6

11

Shell plus limestone

R8

2

Short upstanding

F32

8

46

Limestone plus ironstone

R9

23

External expansion

F34

3

102

grog, sand and ironstone

R10

8

Internal expansion

F35

3

56

Grog, limestone and ironstone

R11

14

Bead

F37

10

2

Sand, grog and unidentified rock

R12

1

Internal bevel

R13

4

External bevel

F38

8

5

limestone, ironstone, sand and grog

R14

1

Lid-seating groove

R16

7

Internally and externally expanded

R17

1

Flattened bead with groove under

R18

2

Bead with internal bevel

R17, R20, R23 and R25). The vessel forms included bowls, jars and cups, both in fine and coarse wares.

R20

1

Collared with broad band under rim

R21

3

Incurved

Bases

R22

3

Flat or rounded top with external groove

R23

1

Internally squeezed and internally rolled

R24

5

Slight external expansion

R25

1

Flat with external bevel

Total sherds

Number of sherds

Fabric Type

Figure 136 Late Bronze Age/Early Iron Age pottery fabric groups Table 22 Numbers of Late Bronze Age/Early Iron Age rim forms

6721

From the entire pottery assemblage, there were only 131 bases, forming a relatively small component of the assemblage. In most pottery reports, this rim to base survival rate appears similar; there are always more rims 132

The Late Bronze Age/Early Iron Age Pottery than bases. Some broken bases were shaped and fashioned into spindle whorls (see Chapter 11) and some may have been utilised as lids. It is possible that the base, which is often the thickest part of a pot, was the logical piece to save for re-shaping; or perhaps this part of the pot may have had specific meaning and the usual discard procedures did not apply. Another possibility is that bases were deliberately kept, ground down for grog and combined with clay to make new pottery vessels.

Hengistbury Head (Cunliffe 1987), Eldon’s Seat (Cunliffe and Phillipson 1968), All Cannings Cross (Cunnington 1923) and Longbridge Deverill Cow Down (Hawkes and Hawkes 2012). In addition, pottery from various Purbeck sites recorded by Calkin (1949) is also cited. The sites are partially referenced in the text, but full references are located in the bibliography. The feature types included pits, scoops, deposits and postholes which were cut or placed into the accumulating midden.

The Pottery Assemblage

Type 1 pits

The relatively large quantity of pottery is considered as two assemblages: first, that from features cut or placed into the midden deposit, and second, unstratified material from the midden layers. Where possible, published parallels with similar vessels from other major Wessex assemblages were noted; these sites are Potterne (Lawson 2000),

Four Type 1 pits were identified, and were characterised by their large size and complex infillings. A further pit (1302) could only be partially excavated due to its location. The pits were likely to have had at least initially, a storage function. Table 23 summarises the pottery assemblages from these pits.

Table 23 Late Bronze Age/Early Iron Age pottery from Type 1 pits 335, 326/392, 1302, 1412 and 1513 Feature

context

type

No sherds

Fabrics

Fig.nos

Diagnostics

335

359

Top fill

6

F7,F15,F22, F23,F27

137.1-2

R1, FP

335

380

2nd fill

30

F7,F22,F23,F24,F27,F29

335

392

Base fill

34

F7,F8,F22,F23,F27

137.4-7

R1(4), R24, IC lines, dots, handle, red coat (2)

326/392

352

Part of wall

4

F7,F11,F20,F22

326/392

355

Top fill

88

F7,F8,F11,F17,F20,F21,F23, F25,F27,F29

138.1-2

R20,IC lines, red coat (3)

326/392

356

2nd fill

154

F7,F8, F11,F23,F25,F27

138.3-5

R1(6),R5,R6,R9(2),R12,R21,FP, red coat (3)

326/290

357

Deposit in pit

8

F7,F8,F25,F27

326/290

376

4th layer

27

F7,F8,F11,F22,F23,F25

138.6-6

R1(2),R6, red coat(2)

326/290

378

Base fill

23

F7,F27

138.8

B1

1302

1303

fill

120

F7,F8,F11,F17,F20,F25,F26 ,F27,F30

140.5

R1(3),R13,dots, IC, slashes, red coat(8)

1412

1423

Top fill

143

F7,F8,F11,F17,F20,F22,F23, F25,F26,F27

139.1-9

R1911),R2,R5,R7(2),R16,R22, dots, IC lines, red coat (4), 4x overfired

1412

1452

2nd fill

11

F7,F11,F17,F20,F25,F27

139.10

Furrows, red coat (3)

1412

1507

3rd fill

9

F7,F20,F25,F27

139.11

IC line

1412

1408

Base fill

8

F20,F25

139.12

R5

1513

1514

Top fill

20

F7,F17,F20,F25

R1, red coat (2)

1513

1518

2nd fill

4

F7,F8,F25,F27

R1

1513

1519

3rd

64

F7,F20,F25

1513

1522

4th fill

8

F7,F20,F25,F27

R1(2), red coat (5), spalls

1513

1524

5th fill

40

F7,F8,F20,F25,F27,F29

R1(3), red coat (5), spalls

1513

1525

Base fill

56

F8,F20,F25

fill

140.1-2

140.3-4

133

R1(3),R292), red coat (15), IC lines, B2-decorated

R1(4), red coat (2), spalls

Multi-period Occupation at Football Field, Worth Matravers, Dorset 2. Top fill 355. Wall sherd, probably from a jar with two horizontal lines of incised decoration. F15. 3. Second fill 356. Rolled rim (R6) from small jar. F27. 4. Second fill 356. Rim sherds (R1) with internal bevel and trace of red coat on the inside, from small jar. F7. 5. Second fill 356. Rim (R12) and upper body sherds from jar with upstanding simple, rounded and slightly flattened rim (R1), diameter 120mm. F27. This can be compared with a Type JB2.0 jar from Hengistbury (fig. 134, 1258) with a broad Early Iron Age date. 6. Fourth fill 376. Rolled rim sherd (R6) from thick-walled bowl. Top of rim is slightly irregular and faintly indented. F7. 7. Fourth fill 376. Carinated sherd from a red-coated bowl, with near-vertical side above the shoulder. F7. 8. Fifth fill 378. Base sherds (B1) from straight-sided jar with a base diameter of 150mm, lightly-incised groove just above base. F7.

Pit 335 (Figs 30 and 31) In total, 70 sherds of pottery weighing 514g were recovered from three of the four fills of this stepped pit. Ten fabric types were identified and the average sherd size was 7.34g. Fig. 137 1. Top fill 359. Simple rim (R1) with evidence of folding on the inside, from a bowl. F15. 2. Top fill 359. Probable shoulder sherd from a jar with fingerprint impression, vessel is wiped above the fingerprint and the fabric is ‘rough and pimply’ below. F15. 3. Base fill 392. Simple rim (R1) from a red-coated bowl. F7. 4. Base fill 392. Simple rim (R1) from a thin-walled bowl. F27. 5. Base fill 392. Slightly externally expanded rim (R24) from a jar. F22. 6. Base fill 392. Simple, slightly pointed rim (R1) from a bowl. Top of rim slightly abraded, probably through wear. F7. 7. Base fill 392. Round-bodied jar with a vertically perforated lug. Sherds decorated with angled, grooved lines. F27. This is similar to a vessel belonging to the Type 20 series handled jars from Potterne (fig. 52.50) dating to the 9th-8th centuries BC.

Pit 1412 (Figs 34 and 35) This pit could only be half sectioned, as it extended into the eastern section of the 2011 excavation on Site 2 and was probably circular. Five separate fills were recorded and pottery (170 sherds weighing 1591g) was collected from all of the fills. In total six fabric types were identified and the average sherd weight was 9.35g; it was noted that the average sherd size was larger nearer the pit base.

Pit 326/390 (Figs 32 and 33)

Fig. 139

Ten separate fills were recorded from this unusual figureof-eight shaped double pit and pottery (304 sherds weighing 3232g) was collected from six of these. In total 11 fabric types were identified and the average sherd weight was 10.63g.

1. Top fill 1423. Jar rim sherd in coarse fabric, internally and externally expanded (R16). Burnt residue on the internal surface. F25. 2. Top fill 1423. Jar rim sherd (R1), with evidence of exterior folding. Interior surface smooth but exterior surface abraded. F25. 3. Top fill 1423. Jar rim sherd (R22) with rounded top and exterior incised groove. The outer surface displays evidence of over-firing. F25.

Fig. 138 1. Top fill 355. Jar rim sherd (R20), with pronounced collar just underneath the rim. F7.

Figure 137 Late Bronze Age/Early Iron Age pottery from Type 1 pit 335

134

The Late Bronze Age/Early Iron Age Pottery 4. Top fill 1423. Bowl, everted rim sherd (R7), diameter 200mm. F7. 5. Top fill 1423. Bowl, simple rim sherd (R1) in very fine fabric with traces of red coat. F7. 6. Top fill 1423. Base fragment (B3) in fine fabric from probable bowl.F25. 7. Top fill 1423. Decorated wall sherd with two ranks of impressed dots with square bases forming a v-shaped

pattern. Traces of white, possibly chalk, infill on several of the dots. The sherd has broken along incised triangular lines which contained the pattern. F17. Probably from a jar. 8. Top fill 1423. Decorated wall sherd with sharply incised vertical lines. F25. 9. Top fill 1423. Decorated wall sherd with impressed wedge shapes. F19.

Figure 138 Late Bronze Age/Early Iron Age pottery from Type 1 pit 326/390

Figure 139 Late Bronze Age/Early Iron Age pottery from Type 1 pit 1412

135

Multi-period Occupation at Football Field, Worth Matravers, Dorset 10. Second fill 1452. Near-rim sherd from a red-coated, furrowed bowl. F7 (Potterne Type 3.1, fig. 48.18). 11. Third fill 1507. Decorated wall sherd from jar with single incised curved line. F27. A Type 20 Jar from Potterne (fig 51.43) dates to the 9th-8th centuries BC and displays similar decoration. 12. Base fill 1508. Rim sherd with proto-bead and external groove (R5), possibly from a bowl or small cup. F25.

Compares with a Potterne Type 11 bowl. 2. Third fill 1519. Wall sherd from probable jar decorated with random, short, incised strokes. F25. 3. Sixth and lowest fill 1525. Rim sherds (R1) from bipartite, short-necked, red-coated bowl; diameter 180mm. F20. Compares with Potterne Type 3.1, (fig. 48.24). F20. 4. Rim sherds (R1) from a tripartite bowl, 330mm in diameter, which had been placed on the base of the sixth and lowest fill (1525). F25. About half of this pot was located; the remainder may survive in the unexcavated section. Conforms to Potterne Type 3.3 bowl, (fig. 48.30) with a 7th to early 6th century BC date.

Pit 1513 (Figs 36 and 37) This clay-lined circular pit could only be half sectioned as it was located on the extreme western edge of Site 2. Seven fills were recorded and pottery (192 sherds weighing 1819g) was recovered from six of them. In total seven fabric types were identified and the average sherd weight was 9.47g.

Pit 1302 (Fig. 36) This pit was only partially excavated due to its position in a small extension of the site. However the amount of pottery retrieved from its presumed upper fill (1303) was substantial: 120 sherds weighing 1045g (average sherd weight 9g). Nine fabric types were present.

Fig. 140 1. Third fill 1519. Jar base sherd (B2), diameter 110mm. Incised lines inscribed on the exterior of the base. F25.

Figure 140 Late Bronze Age/Early Iron Age pottery from Type 1 pits 1513 and 1302

136

The Late Bronze Age/Early Iron Age Pottery Fig. 140

Pit 1126 (Fig. 38)

5. Body sherd from a jar decorated with random impressed circular dots contained within wide geometric furrows. F25.

Complex, barrel-shaped pit cut into the upper midden make-up with three fills. A total of 304 sherds of pottery weighing 6391g (average sherd weight 21.02g) were recovered from the two upper fills, the uppermost (1127) containing the majority of the pottery. Eleven fabric types were recorded. For its size, this pit yielded the largest amount of pottery. There were many large, over-fired, coarse, undiagnostic wall sherds in Fabric F20, probably from large storage jars. However there were few con-joins and no rims or bases, so vessel forms or sizes could not be assigned. Spalls of pottery suggest that wasters were present. Parts of at least 15 different vessels had been placed in the pit and the diagnostic forms are illustrated below.

Type 2 pits Four Type 2 pits were fully excavated; these too may have been used for storage. The pottery assemblages are summarised on Table 24. Pit 393 (Fig. 38) A rectangular pit located on Site 1 with a single fill contained 81 sherds of pottery weighing 953g with an average sherd size of 11.76g. A small amount of pottery from this pit dated to the Middle Iron Age (see Chapter 10).

Fig. 142 1. Upper fill 1127. Internally and externally expanded rim sherd (R16) from a jar. Diameter 220mm. Outer surface wiped. F7. Similar to Hengistbury Type JB1.0 (fig. 133.657). 2. Upper fill 1127. Everted and externally bevelled rim sherd (R13) from a coarse pot, probably a jar. F25. 3. Upper fill 1127. Heavily abraded rim sherd (R1) from thin-walled, long-necked, red-coated bowl. Diameter 190mm. F7. Compares with Potterne Type 2, fig. 47.12. 4. Upper fill 1127. Near-rim sherd (R1) and carination from a biconical bowl, similar to Potterne Type 1 above. F22.

Fig. 141 1. Fill 394. Internally and externally expanded rim sherd (R16) from an open, shallow dish, diameter 220mm; well-incised groove below rim. F22. This is an example of a Potterne Type 11 bowl, (fig. 49.39) with dates ranging through the 9th-8th centuries BC; it was also noted at Eldon’s Seat (Type 6, fig. 15.97). 2. Fill 394. Small rim sherd (R9) with external expansion and slight, uneven groove. F7. 3. Fill 394. Rim sherd (R13) with very slight external bevel and traces of red coat. F7.

Table 24 Late Bronze Age/Early Iron Age pottery from Type 2 pits 393, 1126, 1503 and 1574 Feature

context

type

No. sherds

Fabrics

Fig. Nos

Diagnostics

393

394

Single fill

81

F7,F8,F11,F20,F22,F25,F27

141.1-3

R1, R9, R16, red coat (4)

1126

1127

Top fill

289

F7,F8,F11,F20, F22,F25, F26,F27

142.1-13

R1(16),R2(5),R6,R9,R16, slashes, notches, red coat (13)

1126

1152

Base fill

15

F7,F8,F22, F25,F28,F34,F35

R1,R2(2)

1503

1501

Top fill

8

F2,F17,F20,F25,F27

R1,R10,red coat (1)

1574

1579

Single fill

1

F7

Figure 141 Late Bronze Age/Early Iron Age pottery from Type 2 pit 393

137

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 142 Late Bronze Age/Early Iron Age pottery from Type 2 pit 1126

5. Upper fill 1127. Rim sherd (R1) from a biconical bowl, with firing cracks on the rim. Diameter 180mm, F7. Compare with Potterne Type 1, fig. 47.3. 6. Upper fill 1127. Upright rim sherd (R1) with small notches on the rim edge from a goblet-shaped bowl; traces of red coating. F25. Potterne Type 3.1, fig. 48.24. 7. Upper fill 1127. Rim sherd (R1) from over-fired, thinwalled bowl. Diameter 130mm. F7. 8. Upper fill 1127. Externally expanded rim sherd (R9) which has been crudely folded over. Tiny notches have been incised into the edge of the rim and a groove has been lightly incised under the rim. Probably from a jar. F8. 9. Upper fill 1127. Rim sherd from large bowl (R1). One fingernail impression on the rim top (accidental?). Traces of red coat. F25. Potterne Type 2, fig. 47.12. 10. Upper fill 1127. Two simple rim sherds (R1) from a barrel-shaped jar with an incurving rim. F22. Similar vessels occur at Longbridge Deverill (Type P50, fig. 3.28, pit II.55, 1) and Eldon’s Seat (fig.11.16) and date to the Earliest Iron Age. 11. Upper fill 1127. Slightly everted, rolled rim (R6), overfired. F25.

12. Upper fill 1127. Rim sherd (R1) from a long-necked bowl. F25. 13. Upper fill 1127. Rim sherd (R1) from high-shouldered bowl. F25. Type 3 pits Seventeen features defined by their shallow depths of no more than 0.35m were excavated. The pottery assemblages are summarised in Table 25 and a selection of diagnostic sherds have been chosen for illustration. Pit 1022 (Fig. 39) This pit was only one quarter excavated; in total, 15 sherds of pottery weighing 92g in seven different fabrics were recovered from two fills and had an average sherd weight of 6.13g. Fig. 143 1. Lower fill 1044. Slightly everted and rolled rim sherd (R6) from a shouldered jar with a rim diameter of 140mm. F25. Potterne Type 20, fig. 51.45. 138

The Late Bronze Age/Early Iron Age Pottery Table 25 Late Bronze Age/Early Iron Age pottery from Type 3 pits and scoops Feature

context

Type

No. sherds

Fabrics

323

363

fill

2

F7,F28

354

354

Cut/fill

1

F7

1022

1041

Top fill

10

F7,F20,F22,F23,F27

1022

1044

Base fill

5

F11,F25

1142

1141

fill

13

F8,F11,F22,F24,F25,F27

1176/96

1177

=1197

18

F7,F8,F20,F22,F25,F27

1176/96

1197

=1177

1

F7

1215

1216

fill

13

F8,F11,F20,F25,F27,F35

1217

1218

fill

5

F11,F20,F25

1230

1232

=1240

7

F7,F8,F11,F25

1230

1240

=1232

7

F7,F25

1234

1235

fill

27

F7,F8,F11,F20,F25,F27,F29

1241

1242

fill

41

F7,F11,F17,F20,FF23,F25, F27, F25

1259

1260

fill

2

F7

1263

1264

fill

9

F7,F20,F25,F27

1268

1269

fill

15

F7,F8,F11,F20,F25,F29

1427

1429

fill

86

F1,F5,F11,F22,F25,F27

1468

1469

fill

2

F7,F8

1509

1510

fill

5

F8,F25

1807

1805

Top fill

27

1807

1806

Base fill

8

Fig. nos

Diagnostics

Fig.143.1-2

R6,R7 residue R1(2),R2, red coat (1)

Fig.143.3

R1,R2, furrows, red coat (1)

Fig.143.4

R1(30,handle, red coat (2)

R1

Overfired (1) Fig.143.5-7

R1(5),R2 ,IC, FN, furrows

R1(4),R2,FN, red coat (20 Fig.144.1-3

R2,B1, IC

F7,F8,F17,F20,F25,F27,F29, F34

Fig.144.4-6

Furrows, reds coat (2)

F7,F11,F25,F30

Fig.144.7

IC

2. Lower fill 1044. Everted rim sherd (R7) from a shouldered jar with a rim diameter of 140mm. F25. Typologically similar to the above.

Fig. 143 4. Fill 1216. Part of a handle from a round-bodied jar. F8. Potterne Type 30, fig. 52.54.

Pit 1176/1196 (not illustrated)

Scoop 1241 (Figs 39 and 41)

A shallow oval pit was initially recorded as two features; 40 sherds of pottery were recovered weighing 231g, with an average weight of 5.75g and six fabric types were recorded.

An oval scoop with edges defined by vertical slabs of limestone. The single fill yielded 41 sherds of pottery weighing 353g; the average weight was 8.6g.

Fig. 143

Fig. 143

3. Fill 1177. Furrowed wall sherd from a carinated bowl. F25. Potterne Type 3.1, fig. 48.18.

5. Fill 1242. Near-rim sherd from a bipartite jar with diagonal fingernail decoration on the carination. F20. Potterne Type 33, fig. 54.64. 6. Fill 1242. Body sherd decorated with incised linear pattern, probably from round-bodied jar. F35. 7. Fill 1242. Finely made body sherd from near carination of a possible cup, decorated with two well-executed incised horizontal lines. F17. Potterne Type 81, fig. 60.101.

Scoop 1215 (Fig. 39) Cut into the eastern baulk of Site 2, this presumably oval feature yielded 13 sherds of pottery weighing 100g from the single fill. The average sherd weight was 7.69g. 139

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 143 Late Bronze Age/Early Iron Age pottery from Type 3 pits 1022, 1176/1196, 1215 and 1241

3. Fill 1429. Conjoining sherds from the belly of a roundbodied bowl decorated with random incised lines. Belly diameter 215mm. F25. Profile compares with a similar Type 7 bowl from Potterne (ibid, 160, fig. 49.37 which had an Early Iron Age date.

Scoop 1427 (Fig. 39) Sub-circular scoop with a charcoal-rich single fill from which 86 sherds of pottery weighing 1180g (with an average sherd weight of 13.72g) were recorded. Fifty of these (weighing 728g) were from a red-coated bowl which displayed evidence of a beige-coloured external residue.

The following two features were noted during final groundwork in March 2012 and were located 21.5m from the site entrance. Approximately 0.2m of material comprising topsoil and the uppermost surface of the midden had been removed to level the road surface.

Fig. 144 1. Fill 1429. Probable bowl rim sherd (R2) with very slight external expansion and remnants of diagonal slashes immediately under the rim. F27. Similar to a Type 1 bowl from Potterne (fig. 47.2) broadly dated to the 9th-6th centuries BC. 2. Fill 1429. Base sherd (B1) from jar. F11.

Scoop 1807 (not illustrated) Oval scoop measuring cut into the top layer of the midden. A total of 27 sherds of pottery were recorded weighing 93g and having an average sherd weight of 3.4g.

Figure 144 Late Bronze Age/Early Iron Age pottery from Type 3 pits 1427 and 1807

140

The Late Bronze Age/Early Iron Age Pottery Fig. 144

Fig. 145

4. Fill 1805. Furrowed body sherd from a red-coated bowl (furrows 3mm wide). F25. 5. Fill 1805. Near-rim sherd from furrowed bowl (closely spaced furrows 5mm wide). F25. 6. Fill 1805. Furrowed body sherd probably from jar (furrows 7mm wide). F20. 7. Deposit within 1807. Body sherd from a jar decorated with short, vertical incised lines. F7. Similar vertical incisions were recorded at Potterne on a Type 51 jar (ibid, 171, fig. 58.89).

Located on Site 2, four ‘deposits’ containing a selection of cultural material including pottery were placed into the midden build-up. In addition two large deposits of limpet shells were also recorded (pottery was retrieved from one of these). The pottery assemblages are summarised in Table 26.

1. Simple, upright rim (R1), from bipartite carinated, furrowed bowl. F7. Potterne Type 3.1, fig. 48.22. 2. Simple, upright rim (R1), slightly everted, from bipartite carinated, furrowed bowl. F7. Potterne Type 3.1, fig. 48.26. 3. Body sherd from bipartite, very sharply carinated, furrowed bowl. F29. Longbridge Deverill Type PO4, fig. 3.2/23. 4. Body sherd from bipartite, sharply carinated undecorated bowl. F29. Potterne Type 3.1, fig. 48. 26. 5. Flat rim sherd (R2), possibly from a cup. F29. Similar to Potterne hemispherical Type 80 cup, fig. 60.98. 6. Wall sherd decorated with possibly chalk-filled impressed dots. F25. 7. Base sherd (B2), from probable jar. F27. 8. Thick-walled base sherd (B2) with a diameter of 230mm from probable jar. F20.

Deposit 1169 (Fig. 42)

Deposit 1175 (Fig. 43)

Contained within Slot 1113 (fill 1114), the deposit contained 55 sherds of pottery weighing 826g with an average weight of 15g. Other finds included pebbles, a ceramic pebble, and parts of shale bracelets.

Contained within Slot 1101 (fill 1102); four sherds of pottery were recorded weighing 35g (average sherd size 8.45g). Other finds included shale fragments and animal bone.

Placed deposits

Table 26 Late Bronze Age/Early Iron Age pottery from placed deposits Deposit

No. sherds

Fabrics

Fig nos

Diagnostics

1169

55

F7,F8,F11,F17,F20,F25,F27,F28,F29

145.1-8

R1(8),R2(2),B2(2), furrows, dots

1175

4

F7,F11, F17,F25

146.1

stabs

1236

2

F11

1256

9

F7,F20,F25,F27

146.2-3

R1,B2

1500

29

F7,F11,F20,F25,F27,F34

146.4-8

R1(3),R10, red coat

Figure 145 Late Bronze Age/Early Iron Age pottery from placed deposit 1169

141

Multi-period Occupation at Football Field, Worth Matravers, Dorset Fig. 146

Limpet deposit 1500 (Fig. 44)

1. Body sherd decorated with short vertically incised lines similar to jar sherd Fig. 144.7. F17.

A total of 29 sherds of pottery weighing 249g with an average sherd size of 8.58g were retrieved from the deposit of limpet shells.

Deposit 1256

Fig. 146

Located in Slot 1103 (fill 1254), 19 sherds of pottery weighing 172g (average sherd weight 9g) were found with animal bone and worked flint.

4. Rim sherd (R10) internally expanded with a flat top from dish. F20. 5. Simple upright rim (R1) with traces of red coat, probably from bowl. F25. 6. Simple upright rim (R1) from a bowl, with very faint scratched lines. F7. 7. Simple upright rim (R1) from a long-necked bowl. F7. Potterne Type 2, figs 47.11 and 12. 8. Body sherd from a jar with exaggerated thickening on the carination and a trace of a horizontally incised line. F20.

Fig. 146 2. Simple rim sherd (R1) from a long-necked carinated bowl with a trace of red coat. F7. Potterne Type 2 bowl fig. 47.15. This bowl type has a long currency of use from the late 8th to the early 6th centuries BC. 3. Base sherd (B2) diameter 80mm from probable jar, F7.

Figure 146 Late Bronze Age/Early Iron Age pottery from placed deposits 1175, 1256 and limpet deposit 1500 Table 27 Late Bronze Age/Early Iron Age pottery from postholes Feature

Context

type

No sherds

Fabrics

Fig. nos

345

346

fill

1

F27

347

348

fill

4

F28

367

367

cut/fill

1

F23

1220

1221

fill

5

F7,F23,F25

1247

1248

fill

3

F25

1268

1269

fill

15

F7,F8,F11,F20,F25,F29

1287

1288

fill

2

F22,F25

1536

1537

fill

1

F25

1540

1541

fill

14

1547

1548

fill

1555

1556

fill

Diagnostics

red coat

R1

Fig.147.1-2

R1(4),R2, FN(2), red coat

F7,F17,F25

Fig. 147.3-4

R1(3), furrows

3

F7, F11

Fig. 147.5

R1, FN

1

F27

spall

142

The Late Bronze Age/Early Iron Age Pottery Postholes Pottery was retrieved from 12 of 17 postholes; Table 27 summarises the assemblages. Generally the numbers of sherds were small varying, between one and five, but 15 and 14 sherds respectively were present in two postholes (1268 and 1547). Posthole 1268 (Fig. 45) Lying below and therefore earlier than the Type 3 pit 1142, the posthole had a single fill. In total 15 sherds of pottery weighing 69g (average sherd weight 4.6g) were recorded. Fig. 147 1. Fill 1269. Simple rim sherd (R1) from a jar with welldefined finger-nail impressions on the top of the rim. F11. The decoration is unusual in that the impressions perforate the rim itself. 2. Fill 1269. Flat rim sherd (R2) from a jar decorated with diagonal fingernail impressions on the outside edge of the rim and having a slight external groove. F25. A similar vessel with a neat rim-edge fingernail pattern is recorded as a Type 51 jar from Potterne (fig. 57.85).

Figure 147 Late Bronze Age/Early Iron Age pottery from postholes 1268, 1540 and 1547

Assemblage from the Midden Layers During the post-excavation process, the slot fills of the midden were assigned layer numbers (Table 4), Layer 1 being the highest and Layer 5 the lowest. Layer 6 comprised material incorporated into the cobbles themselves. Table 20 records the numbers and weights of pottery sherds, together with the average sherd weight from each layer: it can be seen that the pottery quantities diminish towards the base of the deposit. It has to be assumed that the earliest material was deposited as the cobbles were laid down and that the latest material in the sequence was located in the upper midden; however much of the material was displaced and disturbed by later activities. The midden levels are considered from the earliest to the latest.

The following two postholes were associated with a probable house which was identified during the postexcavation work (Fig. 29). Posthole 1540 (not illustrated) Located in midden layer 1463, the oval posthole yielded 14 sherds of pottery weighing 225g with an average sherd size of 16g. Sherds from four vessels were present. Fig. 147 3. Fill 1541. Upright, simple rim sherd (R1) from a longnecked, red-coated, carinated bowl, diameter 150mm. F7. Similar to the Type 2 bowl from Potterne (fig. 47.15). 4. Fill 1541. Simple, slightly everted rim (R1) and body sherds from red-coated, biconical bowl, diameter 220mm. F25. Similar to a vessel from Eldon’s Seat (Type 1, fig. 16.129).

Midden layers 5 and 6 With an approximate depth of c. 0.12m, the deposit consisted of material lying directly above, and incorporated into the cobble matrix (1039). It comprised 492 pottery sherds weighing 3960g with an average weight of 6.2g and is detailed on Table 28.

Posthole 1547 (Fig. 45)

Everted rim sherd (R7) from a jar.

This oval posthole was cut into midden layer 1462. Three small sherds of pottery were recorded weighing 4g (average sherd weight 1.3g).

Context 1560: incorporated into the cobble matrix

Fig. 147

1. Everted rim sherd (R7) from a jar. F25. 2. Upright rim sherd with unusual external bevel (R13) decorated with a faint fingernail impression. F25.

Fig. 148

5. Fill 1548. Simple rim sherd (R1) from a jar displaying diagonal fingernail impressions underneath the rim. F7. 143

Multi-period Occupation at Football Field, Worth Matravers, Dorset Table 28 Late Bronze Age/Early Iron Age pottery from midden layers 5 and 6 Layer

No. sherds

Fabrics

Fig. nos

Diagnostics

1171

18

F7,F8,F11,F17,F25,F27,F29

1168

187

F4,F7,F8,F11,F20,F22,F25,F27,F28,F29,F35

1193

8

F11,F17,F20,F25,F27

B3(2), red coat(1)

1200

15

F3,F7,F17,F22,F25,F27,F32

Red coat (2), overfired(1)

1226

13

F7,F8,F20,F25,F27,F29

1227

34

F7,F8,F11,F17,F20

1228

24

F7,F11,F17,F25,F27

1233

67

F3,F7,F8,F11,F17,F20,F22,F23,F25

1244

3

F7,F11,F27

1245

30

F7,F8,F11,F17,F20,F23,F25,F27,F29,F34

1246

3

F17,F25,F34

1272

24

F7,F8,F20,F22,F25,F27,F29

1301

7

F1,F7,F20,F25

1456

16

F1,F7,F11,F17,F20,F25,F27

R1, IC, red coat( 3)

1558

9

F7,F20,F25,F27

Red coat (3)

1559

18

F7,F11,F17,F20,F25,F27

R1 red coat (2)

1560

50

F7,F8,F11,F17,F20,F22,F25,F27

Red coat (3) 148.3-13

148.14

R1(7),R2(5),R9,R16 ,B1,IC,FP, FN, furrows, slashes

R1,R2, red coat (1)

R1,R2,B1,B3 R1,R2, red coat (1) R1,R2, red coat (3) 148.15

148.1-2

Wedges, red coat (1)

R1, R7, FN, red coat (3), spalls (3)

11. Slightly everted simple rim sherd (R1) from a cup, diameter 60mm. F7. Similar to Potterne Type 86 cup with high, rounded shoulders (fig. 60.116). 12. Furrowed wall sherd from a bowl. F25. 13. Wall sherd from jar decorated with incised lines and sharp slashes. F20. A similar vessel comes from Eldon’s Seat (catalogued as an earlier period bucket urn, fig. 14.81).

Context 1168: immediately above cobbles around pit 1182 Fig. 148 3. Simple, pointed upright rim (R1) with slight internal bevel and fingerprint decoration, probably from a jar. F17. Similar fingerprinting is seen on Type 51 jars from Potterne (figs 57.81 and 58.86) and Type P26 jar from Longbridge Deverill (fig. 3.16/22). 4. Large rim/body sherd (R9) from a wide, open dish with wiped exterior. Rim has pronounced external expansion and an internal groove. Diameter 210mm. F25. Potterne Type 11 Bowl, fig. 49.39, 9th to 8th centuries BC. A similar vessel comes from Eldon’s Seat (Type 6, fig. 17.158). 5. Unusual flat rim sherd (R2), with a slight groove and faint incised diagonal lines on the body, from a probable jar. F7. 6. Internally and externally expanded rim sherd (R16) with groove on top of the rim, from a probable jar. F25. Such grooves are rare and may have been formed to take a lid. A complete Type P34 barrel jar from Longbridge Deverill (fig. 3.21/7) has such a groove. 7. Flat upright rim sherd (R2) from jar with wiped exterior and fine diagonal fingernail impressions. F7. 8. Flat rim sherd (R2) from probable bowl. F25. 9. Simple rim sherd (R1) from bowl. F25. 10. Simple rim sherd (R1) from a bowl, with faint horizontal incised lines beneath the rim. F25.

Context 1226: clearance over the cobbles in the southwest quadrant of Site 2 Fig. 148 14. Flat rim sherd (R2) from a red-coated bowl. F7. Context 1301: overlying cobbles in Slot 1261 Fig. 148 15. Body sherd from belly area of probable jar decorated with unusually long (14mm) vertical impressed wedges. F7. Midden layer 4 This layer comprised approximately 0.05m of deposit over Layer 5 and contained 224 sherds of pottery weighing 1658g with an average sherd weight of 10.3g. The assemblage is detailed on Table 29. 144

The Late Bronze Age/Early Iron Age Pottery

Figure 148 Late Bronze Age/Early Iron Age pottery from midden layers 5 and 6, contexts 1560, 1108, 1226 and 1301

Context 1170 under Slot 1149

4. Carinated wall sherd from a bipartite bowl. F25

Fig. 149

Context 1193 under Slots 1047 and 1085

1. Flat rim sherd (R2), probably from a dish. F25. 2. Chunky, rounded rim sherd (R1). F27 3. Fine, flat rim sherd (R2) probably from a bowl. F27.

Fig. 149 5. Base sherd (B3) from probable jar. F25.

Table 29 Late Bronze Age/Early Iron Age pottery from midden layer 4 Layer

No. sherds

Fabrics

Fig. nos

Diagnostics

1157

9

F7,F8,F11,F20,F25,F28

1170

13

F7,F8,F17,F20,F25,F27,F28

149.1-4

R1,R2(2), red coat (1)

1193

8

F11,F17,F20,F25,F27

149.5

B3(2), red coat (1)

1294

32

F7,F8,F11,F17,F25,F27

149.6-9

R1(3),R7, red coat (9)

1449

91

F7,F8,F11,R1,

149.10-15

R1,R6,R9,R10,B1,FN,IC, red coat (5)

1459

49

F7,F8,F18,F20,F25,F27,F34

1460

34

F7,F8,F20,F22,F25,F27,F28

150.1-6

R1(5),R2,R16,R24(3),B3,IC, stabs, red coat (9)

1461

50

F1,F7,F8,F11,F20,F25,F27,F34

149.7-12

R1(5),R24(2), red coat (10), slash, residue

1462

64

F7,F8,F11,F20,F22,F25,F27

150.1-7

R1(6),R2,R9(2),R12,B1, B2(2), IC, handle, red coat (12)

1463

4

F25,F29

1491

98

F7,F8,F11,F25,F27,F34

1501

8

F2,F17,F20,F25,F27

1502

12

F7,F25,F26

151.10-12

R1(3), red coat (8)

1521

20

F7,F20,F25,F27

151.13-15

R1(2),R4,R7, red coat (10), slash, spall

R1(4), red coat (21)

Red coat (3) 151.6-9

R1,R5,R16, FN, red coat (3), residue R1,R6, red coat (1)

145

Multi-period Occupation at Football Field, Worth Matravers, Dorset 9. Simple rim sherd (R1) with an internal fold on the rim, from a small jar or cup (rim diameter 110mm). F17. Similar to Potterne Type 86 cup, fig. 60.116.

Context 1294 beneath paving slabs 1037 Fig. 149 6. Fine, upright, simple rim sherd (R1) from red-coated bowl, diameter 180mm. F7. Probably a Potterne Type 3. 7. Flat rim sherd (R2) with rounded edges and lentoid impression under rim. F11. 8. Slightly everted rim (R7) from a bipartite red-coated bowl, F25. Potterne Type 3.1, fig. 48.26.

Context 1449 under crushed limestone deposit and layer 1416 Fig. 149 10. Flattened rim sherds (R2) from a small shouldered cup. Longish diagonal lines are lightly incised underneath the rim and the carination is decorated with a row of

Figure 149 Late Bronze Age/Early Iron Age pottery from midden layer 4, contexts 1170, 1193, 1294 and 1449

146

The Late Bronze Age/Early Iron Age Pottery 10. Red-coated body sherd from a bowl, with a very sharp, angular carination and decorated with a vertical slash. F11. A similarly shaped vessel from All Cannings Cross (plate 28.5) has an imprecise zigzag just above the widest part of the carination. 11. Flat rim sherd (R24) with slight external expansion and evidence of folding. F7. 12. Flat rim sherd (R24) with slight external expansion, wiped internally and externally. F7.

fingernail impressions. Rim diameter 100mm. F7. Can be compared to Potterne Type 83 cup (fig. 60.109). 11. Externally expanded flat rim (R9) with slight dishing. F25. 12. Upright simple rim (R1). F25. 13. Internally expanded and folded rim (R10) and body sherd from a small jar or cup. Body decorated with horizontal slashed lines. F7. 14. Sharply carinated wall sherd from a large jar. F20. 15. Wall sherd from near the base (B1) of large jar, the junction between the base and wall is rounded; base diameter 220mm. F8.

Context 1462 above pit 1551 Fig. 151

Context 1460 above 1462 and 1463

1. Flat rim sherd (R2) from a biconical bowl decorated with short horizontal, slashed lines. F25. Can be compared to a Type 1 bowl from Potterne (fig. 47.9) with a very long currency of use from the 10th/9th -6th centuries BC. 2. Simple upright rim (R1). F7. 3. Rim sherd with thick external expansion (R9). F25. 4. Lower part of a lug from a jar. F7. At Potterne, the round-bodied, Type 20 jars often had evidence for vertically pierced lugs (fig. 52.49-51) and dated to 8th6th centuries BC. 5. Base sherd (B2) from a thick-walled jar. F20.

Fig. 150 1. Jar rim with flat top and external groove (R22). F7. Similar to a Type JB 2.0 jar from Hengistbury Head (fig. 123.1507). 2. Jar rim sherd (R1) with a slight external rolling. F7. 3. Jar rim sherd (R24), flat with very slight external expansion and groove. F7. 4. Probable bowl rim sherd both internally and externally expanded (R16). F7. 5. Body sherd in fine fabric decorated with crude stabs. F7. Similar decoration is recorded on A Type 51 jar from Potterne (fig. 58.88). 6. Base sherd (B3) with possible knife mark, probably from a bowl. F7.

Context 1491 below 1432 Fig. 151 6. Slightly internally and externally expanded rim sherd (R16) with internal residue; probably from a thinwalled jar. F27. 7. Simple upright rim sherd (R1) from a red-coated bowl. F7. 8. Proto-bead rim (R5), with faint external groove. F7. 9. Small body sherd decorated with row of horizontal, vertical fingernail impressions. F11.

Context 1461 below wall 1450 and pavement 1453 Fig. 150 7. Simple, very rounded rim sherd (R1) from large redcoated jar. F8. 8. Simple rim sherd (R1). F20. 9. Simple rim sherd (R1) with a very slight external thickening. F7.

Figure 150 Late Bronze Age/Early Iron Age pottery from midden layer 4, contexts 1460 and 1461

147

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 151 Late Bronze Age/Early Iron Age pottery from midden layer 4, contexts 1462, 1491, 1502 and 1521

Context 1502 below layer 1478

Context 1521 below layer 1478

Fig. 151

Fig. 151

10. Very slightly thickened simple rim (R1) from a jar, with internal traces of residue. F25. A similar jar sherd is recorded at Longbridge Deverill (Type P37, fig. 3.30, Pit II.2/1). 11. Simple upright rim (R1) with an unusual internal indentation from a red-coated bowl. F26. The bowl has a rounded carination; the best parallel is from Longbridge Deverill (Type P13, fig. 3.15, WH II.3/16. 12. Simple upright rim (R1) from a red-coated bowl, with severe abrasion on the immediate internal rim surface. F26.

13. Short, rounded upstanding rim (R4) from a jar. F25 14. Short, everted rim (R7) with slight external groove and a diagonal slash. F7. 15. Simple upright rim (R1). F25. Midden layer 3 The layer consisted of approximately 0.1m of deposit overlying Layer 4 and containing 870 sherds of pottery weighing 8968g with an average sherd weight of 10.8g (Table 30). Context 1028 third fill of Slot 1011, same as 1065 Fig. 152 1. Simple upright rim (R1) from a red-coated bowl. F7.

Table 30 Late Bronze Age/Early Iron Age pottery from midden layer 3 Layer

No. sherds

Fabrics

Fig. nos

Diagnostics

1028

14

F7,F25,F29

152.1

R1(3),red coat (4)

1065

10

F7, F22,F25

152.2-3

R9, FP, slash, red coat

1075

13

F7, F20,F22,F23,F25,F27,F28

152.4-6

R1,R16,R23

1125

35

F7, F17,F20,F25,F27

152.7-8

R1(2),R2, filled dots

1128

12

F7, F11,F25

1225

13

F7,F8,F17,F22,F25,F27

R1 152.9

148

B1, red coat, residue, over-fired

The Late Bronze Age/Early Iron Age Pottery Context 1065 same as 1028 in Slot 1011

Context 1225 below 1201 in Slot 1096

Fig. 152

Fig. 152

2. Two conjoining body sherds decorated with a horizontal row of fingerprints from the carination of a probable jar. F7. 3. Chunky, externally expanded rim (R9), the pointed edge of which is decorated with nicks; from a large red-coated jar. F7. A complete vessel catalogued as a situla jar was recorded at Longbridge Deverill (Type P22, fig. 3.17, Pit 1.1/1); this also displayed a defined T-shaped rim.

9. Base sherd (B1) from a jar with traces of exterior residue. F17. Midden layer 2 This comprised 26 contexts which produced assemblages of pottery. In total 994 sherds weighing 7852g were recorded; the average sherd weight was 8.72g. The assemblages are detailed on Table 31. A selection of pottery from 11 contexts has been chosen for illustration.

Context 1075 below 1070 in Slot 1068

Context 360 on Site 1

Fig. 152

At the time of excavation, this layer was not recognised as midden deposit. It comprised a sticky dark brown loam similar to the deposits on Site 2 and was recorded as a single fill, the depths varying from 0.28m to 0.45m. A total of 59 pottery sherds weighing 610g (average sherd weight 10.3g) were recovered.

4. Simple upright rim (R1) from a bowl with a rounded carination, similar to Fig 154.11. 5. Thick-walled internally and externally expanded rim (R16), probably from a bowl. F22. 6. Rim sherd showing very slight external bevel, external expansion and internally squeezed (R23). F23.

Fig. 153

Context 1125 below 1042 in Slot 1023

1. Wall sherd from the rounded belly of a jar decorated with a horizontal row of fingernail impressions above fine incised lines.

Fig. 152 7. Simple upright rim sherd (R1) from a bowl. F7. 8. Flat rim sherd (R2) with horizontal row of impressed dots filled with white paste. F25.

Context 1017 second layer of Slot 1015 This artefact-rich layer had depths varying from 0.15m

Figure 152 Late Bronze Age/Early Iron Age pottery from midden layer 3, contexts 1028, 1065, 1075, 1125 and 1225

149

Multi-period Occupation at Football Field, Worth Matravers, Dorset Table 31 Late Bronze Age/Early Iron Age pottery from midden layer 2 Context

No. sherds

Fabrics

Fig. nos

Diagnostics

336

14

F7,F8,F11,F23,F27

360

59

F7,F8,F20,F22,F23,F27,F28

361

1

F7

Red coat (1)

1013

2

F7,F25

IC

1017

262

F7,F8,F11,F20,F22,F25,F27,F29,F30,F37

1025

4

F7,F11,F26,F29

1033

80

F7,f8,F17,F20,F22,F23,F25,F27,F28,F34, F35

154.1-3

R1(40,R2,furrows,wiped, red coat (1)

1042

23

F7,F8,F17,F20,F22,F25,F26,F27,F28,F34, F35

154.5-6

R22,B2, red coat(1), furrows, wiped

1051

4

F22,F25

1064

10

F7,F8,F17,F22

154.4

R1,boss

1067

38

F7,F22,F23,F25,F26,F27,F32

154.7-10

R2,R6,PC,FN, furrows

1082

43

F4,F7,F11,F20,F22,F23,F25,F26,F27,F28, F34

155.1-9

R1(4),R2,R6(2),B3, red coat(12), residue

1089

31

F7,F8,F20,F22,F23,F25,F26,F27,F34

155.10-13

R1(3),R9,R10, residue, red coat (4)

1098

39

F7,F11,F17,F20,F22,F23,F25,F28,F32,F34

R1(2),R7,B(3),residue, red coat (1)

1148

28

F7,F8,F11,F17,F20,F24,F25,F26,F27,F28, F29

R2, B, red coat(5)

1183

22

F7,F8,F17,F20,F22,F23,F25,F29

R1(20,R2

1201

42

F7,F8,F11,F17,F20,F25,F27

R1(20

1211

61

F7,F8,F11,F20,F22,F23,F25,F27,F28,F34

1224

14

F7,F11,F17,f20,F25,F27,f29

1280

70

1354

Red coat(4) 153.1

153.2-20

FN,IC, red coat (10), residue

R1(20),R2(4),R9 (2), R13, R24, B(4), IC, FP, stabs, furrows, Red coat (47) R1

Red coat(2)

156.1-2

R1(4),R2, IC, residue(2), furrows

F2,F7,F8,F20,F25,F27

156.3-8

R1(5),R2(2),B1, dots, IC

44

F7,F8,F11,F17,F18,F20,F25,F27,f29

156.9-16

R2(2),FN,FP,IC, red coat (3)

1416

4

F1,F17,F25

R2,B3

1425

18

F7,F11,F25,F27

R1(3),B, red coat (2)

1431

13

F7,F8,F11,F27

FN

1432

14

F7,F11,F25,F27

R1(3),B3, red coat (7)

1490

2

F7,F25

R1, red coat (2)

to 0.24m. It was from this layer that the socketed axe (O’Connor, Chapter 11, Fig.180.1) was retrieved.

4. Upright simple rim (R1). F11. 5. Bowl rim (R1), slightly everted. F7. 6. Red-coated bowl rim with slight external expansion (R9). F7. 7. Simple rim (R1) from thin-walled bowl. F25. 8. Rim with slight external expansion (R24). F27. 9. Upright rim (R1) with single furrow, from a bowl. F25. 10. Slightly everted simple rim (R1) from a furrowed bowl. F25. Potterne Type 3.1, fig. 48.18. 11. Simple upright rim (R1) from a furrowed bowl. F25. Similar to above. 12. Rounded and everted rim sherd (R7). F25.

Fig. 153 2. Incurved, simple rim sherd (R1) from a round-bodied jar decorated with a horizontal row of short vertical slashes. F25. Compares with a similar pot from Longbridge Deverill Type P55, fig. Pit II.33B 3.25/5 and All Cannings Cross, plate 29.9. 3. Chunky, simple rim (R1) with remnants of chalk-filled dots. F29. 150

The Late Bronze Age/Early Iron Age Pottery

Figure 153 Late Bronze Age/Early Iron Age pottery from midden layer 2, contexts 360 and 1017

13. Externally expanded and folded rim sherd (R9) with flattened top, decorated with irregular short slashes. F11. 14. Very thin, upright simple rim sherd (R1) from a cup, with an approximate diameter of 90mm. F25. This can be compared to a simple, hemispherical Type 80 cup from Potterne (fig. 60.98) dating to the 7th century BC. 15. Large body sherd from a jar decorated with lightly incised diagonal lines forming a loose chevron pattern. F25. 16. Body sherd from under the neck of a jar decorated with a horizontal furrow under the neck and joining furrowed chevrons. F25. An almost identical decorative scheme is found on a Type 32 jar from Potterne (fig. 54.63). 17. Near-carination body sherd from a thick-walled jar with remnants of finger-nail impressions on the carination and a poorly executed diagonally incised line above. F27. 18. Small body sherd decorated with a horizontal row of fingerprints with clear nail impressions. F25. From the carination area of a probable jar.

19. Sharply carinated body sherd from finely-made furrowed bowl. F25. 20. Body sherd from a sharply carinated furrowed bowl. F25. Context 1033 layer of Slot 1031 under layer 1032 A total of 80 pottery sherds were recorded weighing 648g (average sherd weight 8.1g). In addition shale bracelet fragments, a worked bone comb (Fig. 198.9) and antler tine fragments were found. Fig. 154 1. Rim (R2) and body sherds from a small, over-fired, thin-walled cup (rim diameter 90mm). F7. Similar to Potterne cup Type 81 (fig. 60.102). 2. Simple, upright rim sherds (R1) from a furrowed bowl. F22. 3. Chunky upright simple rim (R1). F25

151

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 154 Late Bronze Age/Early Iron Age pottery from midden layer 2, contexts 1033, 1064, 1042 and 1067

8. Upright and rolled rim sherd (R6) from a jar. F23. 9. Wall sherd from a jar with a raised cordon. F25. Similar to Potterne Type 58 jar (fig. 59.93) and a small sherd from Gallow’s Gore (plate IIA). 10. Fine jar body sherd with a horizontal row of welldefined, short diagonal slashes; internally wiped. F27.

Context 1064 second layer of Slot 1062, under 1064. Fig. 154 4. Oval boss from a thin-walled vessel. F22. Bosses are unusual on Late Bronze Age/Early Iron Age pottery; three bosses were applied to a globular pot from Gallow’s Gore (fig. 6.6), and a sherd with a boss from Eldon’s Seat (from an earlier period bucket urn, fig. 12.31) can be compared to this example. A semicomplete vessel from All Cannings Cross has four extant bosses on the upper part of the pot (fig. 3.10/55 and plate 42, upper).

Context 1082 second layer of Slot 1071, under 1072 Fig. 155 1. Rolled rim (R6) with external groove and slight internal bevel and trace of residue underneath the rim; probable jar. F22. 2. Simple upright rim (R1) from a probable long-necked bowl, Potterne Type 2. F25. 3. Simple rim (R1), slightly everted and red-coated. F34. 4. Simple, very pointed rim (R1), crudely-made and probably from a cup. F23. 5. Finely-made red-coated body sherd decorated with short diagonal slashes contained within horizontal grooves. F27. This decorative motif is found on both bowls and jars. 6. (and 7 below) Jar sherds from the same vessel. Very abraded rolled rim (R6). F27. 7. Base fragment (B1). F27. 8. Waisted base sherd (B3). F26. 9. Jar base sherd (B2), diameter c. 120mm. F25.

Context 1042: second layer of Slot 1023, under layer 1024 Fig. 154 5. Flat-topped rim sherd (R22) with well-defined exterior groove. F32. 6. Base sherd (B2) from a thick-walled jar (base diameter 140mm) with vertical wipe marks. F20. Context 1067: second layer of Slot 1047, under 1048 Fig. 154 7. Flat rim sherd (R2) with very slight external expansion and slight nicks on the rim edge. F25. 152

The Late Bronze Age/Early Iron Age Pottery

Figure 155 Late Bronze Age/Early Iron Age pottery from midden layer 2, contexts 1082 and 1089

Context 1089 second layer of Slot 1080

Context 1280 fill under pavement 1289, north of pit 1126 and under Slot 1261

Below fill 1081 which contained large limestone pieces and Late Iron Age pottery.

Fig. 156 3. Flat rim sherd (R2) from a wide, incurving open dish, similar to an example from Eldon’s Seat, (fig. 17.160). F25. 4. Flat rim sherd (R2) with a wiped interior and trace of red coat. F25. 5. Bowl rim sherd (R1). F7. 6. Base sherd (B1) from a jar with a poorly executed incised line near the base angle. F25. 7. Wall sherd with a horizontal incised fine line. F25. 8. Wall sherd decorated with impressed chalk-filled dots. F8.

Fig. 155 10. Sherd from a large red-coated dish (diameter 230 mm) with exterior residue and internally expanded rim (R10). F23. This compares with Type 6, Eldon’s Seat (fig. 15.98). 11. Simple (R1) chunky rim. F22. 12. Externally expanded rim (R9) with a slightly dished top. F22. 13. Simple rim sherd (R1) from a long-necked, red-coated bowl with crude incised line under the rim. F27. Compares to Potterne Type 1 bowl, (fig. 47.3) with a long currency of use from the 10th-6th centuries BC.

Context 1354 fill underneath Late Iron Age burial 318 on Site 1

Context 1211 second layer of Slot 1121 below fill 1122

Fig. 156

Fig. 156

9. Simple rim (R1) with very slight internal bevel, wiped externally. F29. 10. Abraded rim sherd (R1) with fingerprints on the side of the rim and random lightly incised lines below. F8. 11. Wall sherd with a trace of red coat and decorated with a lightly incised line. F17. 12. Wall sherd decorated with a lightly incised line. F17.

1. Flat rim fragment (R2) from a bowl decorated with incised lines under the rim. F27. 2. Body sherd decorated with geometric incised lines. F17.

153

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 156 Late Bronze Age/Early Iron Age pottery from midden layer 2, contexts 1211, 1280 and 1354

13. Sherd from the belly of a jar and decorated with fingernail impressions. F25. 14. Sherd from the belly of a jar decorated with sharp fingernail impressions. F27. 15. Wall sherd from a large jar decorated with poorly incised curved lines. F27. 16. Sharply carinated wall sherd from a jar with a horizontal row of well-executed fingerprints on the carination. F17. Probably similar to a Potterne Type 51 jar.

Context 1016 top fill of slot 1015 comprising sticky loam Fig. 157 1. Upright simple, pointed rim sherd from a bowl (R1). F25. Context 1027 single fill of Slot 1026 with multi-period finds Fig. 157

Midden layer 1

2. Chunky simple rim (R1), very slightly out-turned. F27. 3. Simple and slightly everted rim (R1) from a small furrowed cup. F7. Similar to Potterne cup Type 81 (fig. 60.104) dating to the 8th-7th centuries BC. 4. Flat rim sherd with very slight internal and external expansion (R16) from a probable jar. Internally the rim is folded and externally the edge is decorated with close-together, light finger-nail impressions. The outside surface has a brown slip. F30. 5. Rolled rim sherds (R6), internal surface wiped. F7.

This deposit varied in depth from 0.1m to 0.2m and generally was the latest material in the midden. However later disturbance both in the Late Iron Age and in the Roman period resulted in many Level 1 contexts containing mixed assemblages of material. This layer consisted of 40 contexts which produced assemblages of pottery. In total 1801 sherds weighing 14,850g were recorded; the average sherd weight was 8.5g. The assemblages are detailed on Table 32. A selection of pottery from five contexts has been chosen for illustration.

154

The Late Bronze Age/Early Iron Age Pottery Table 32 Late Bronze Age/Early Iron Age pottery from midden layer 1 Context

Slot

No. sherds

Fabrics

Fig. nos

Diagnostics

338

330

2

F2,F27

Red coat(2)

358

337

9

F7,F11,F20,F23,F25,F27

R1,furrows, red coat (6)

397

315

15

F7,F8,F11,F22,F27

IC

1006

1007

7

F17,F20,F25,F27,F35

1016

1015

15

F7,F8,F20,F25

Red coat (1)

1024

1023

33

F7,F8,F23,F27,F28,F29,F34

1027

1026

45

F7,F8,F11,F17,F20,F22,F23,F27,F29

1030

1029

6

F7,F11,F20,F22,F34

1032

1031

29

F7,F8,F20,F23,F25

157.1

R1(3), red coat (1)

157.2-5

R1(6),R6(2),R16, FN, furrows, slip

157.6

FP

B1,B, furrows R7

1035

1034

18

F7,F22,F25,F27,F29,F35

157.7-8

R6,R12, IC

1046

1045

34

F7,F17,F20,F22,F23,F25,F27,F32,F35

157.9-13

R5,R9,B1,B3,FC,IC, red coat(2)

1048

1047

34

F7,F8,F20, F22,F23,F25,F27,F35

1050

1049

10

F7,F17,F20,F22,F25,F27,F34

1055

1054

99

F7,F17,F20,F22,F25,F27,F34,F35

158.1-6

R1(7),R9,IC, dots, furrows

158.7

R1,R12, red coat (1)

1061

1060

31

F7,F8,F11,F20,F22,F23,F27,F34,F38

1069

1068

1

F34

R1,R9

1077

1076

9

F1,F7,F25,F28,F32,F34

R1(2), red coat (2)

1079

1078

6

F8,F17,F22,F25,F34

R1

1086

1085

5

F17,F25,F27,F32

B1

1100

1099

191

F2,F7,F8,F11,F17,F22,F23,F25,F27,F28, F33,F34,F35

159.1-13

R1(19),R2(3),R7(2), wedge, furrows, Red coat (4)

160.1-9

R1(9),R2(5),R6,(2), R7, dots, IC, grooves, Red coat (10)

161.1-5

R1(5),R2(2),R5,R7,R9,B2,IC, red coat (6), residue

1102

1101

33

F7,F11,F17,F22,F23,F25,FF28,F34,F35

1104

1103

195

F7,F11,F20,F22,F23,F24,F25,F27,F28,F32, F34,F35,F38

R1(2),R2(3),IC, wiped, red coat(2)

1106

1105

2

F22,F34

1108

1107

145

F3,F7,F11,F17,F20,f22,F23,F25,F27,F28, F29, F32,F35,F38

1110

1109

61

F1,F7,F8,F11,F20,F22,F23,FF25,F27,F28,F 32,F34

1112

1111

153

F3,F7,F8,FF11,F20,F22,F23,F25,F27,F32,F 34,F35

1114

1113

38

F7,F8,F17,F22,F25,F27,F28,F29,F32,F35

1116

1115

43

F7,F11,F20,F25,F27,F32

1118

1117

108

F7,F8,F11,F20,F22,F23,F25,F27,f29,F30

1120

1119

121

F7,F8,F11,F17,F20,F22,F23,F25,F27,F28, F34

162.1-5

R1(5),r2,R6,R12, wiped, haematite (2)

1122

1121

79

F7,F11,F17,F20,F22,F23,F25,F27,F28,F35, F38

162.6-8

R1(2),R2,PC, dots, slashes, red coat (2), residue

R1(5),R2,R9,B3, red coat (1) 161.6-7

R1(60,R2(2),IC, furrows, red coat (9), R1(4),R2, red coat (2)

161.8-11

R1(5),R7,R14,B2,B3, dots, furrows, red coat (2) R1(3),R2, IC, wiped, red coat (1)

1150

1149

7

F7,F11,F22,F25

R1, haematite (4)

1165

1164

58

F7,F8,F11,F17,F20,F22,F25,F27,F28

R1(2),R2(2),R9, red coat (3), spalls

1167

1166

40

F7, F17,F20,F22,F25,F27,F28

163.1-3

R1(2),R2(2),R5,B2,B(2),furrows, red coat (1)

1181

1180

12

F7,F11,F17,F32

R1,B1,furrows, red coat (6)

1267

1266

11

F8,F17,F25,F27

R1, slashes, red coat (5)

1277

1261

206

F7,F8,F11,F20,F22,F25,F26,F27,F30

1407

1402

20

F7,F8,F17,

1411

1402

2

F20,F27

1473

1419

3

F25,F27

163.4-7

R1(4),R7,R9, FN, dots, furrows

163.8-10

Red coat (2)

R1(7), slip, red coat (1) Red coat (1)

155

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 157 Late Bronze Age/Early Iron Age pottery from midden layer 1, contexts 1016, 1027, 1032, 1035 and 1046

13. Base sherd (B1) possibly from same vessel (Fig. 157.10) illustrated above. F7. A similar, curiousshaped pot was located at Potterne (fig. 60.118). In that instance it was catalogued as a miniature vessel (Type 87). The Football; Field example is larger but the profile is comparable.

Context 1032 upper fill of Slot 1031 Fig. 157 6. Body sherd decorated with horizontal row of deep fingerprints. F8. Such decoration is normally restricted to jars.

Context 1055 upper fill of Slot 1054

Context 1035 upper fill of Slot 1034

Fig. 158

Fig. 157

1. Flat rim with slight external expansion (R9) and a slight internal bevel, probably from a bowl. F25. 2. Simple upright rim sherd (R1) from long-necked Potterne Type 2 bowl. F27. 3. Simple upright rim sherd (R1) from a long-necked Potterne Type 2 bowl with an approximate diameter of 180mm. F20. 4. Simple, pointed rim sherd (R1) from a probable cup similar to Potterne Type 80 (fig. 60.98). F35. 5. Simple rim sherd (R1) from a biconical furrowed bowl, Potterne Type 3.1 (fig. 48.18). F7. 6. Wall sherd decorated with chalk-filled impressed dots, contained within parallel incised lines. F25.

7. Unusual flat rim sherd with a sharp external bevel (R12). F27. 8. Wall sherd decorated with short rows of incised horizontal lines. F35. Context 1046 single fill of Slot 1045 Fig. 157 9. Externally expanded rim sherd (R9) with diagonal slash decorations on the body which impinge onto the underside of the rim. F17. 10. Small, fine rim sherd with proto-bead (R5), very slight internal thickening and a defined external groove possibly from the vessel (Fig. 157.13) illustrated below. F7. 11. Sherd from a jar; shoulder decorated with a horizontal row of fingerprints. F23. 12. Base sherd (B3) with an approximate diameter of 120mm, probably from a jar. F7.

Context 1061 upper fill of Slot 1060 Fig. 158 7. Jar rim sherd with a small internal bevel (R12). F37. Similar to Potterne Type 33 slightly carinated bipartite

156

The Late Bronze Age/Early Iron Age Pottery

Figure 158 Late Bronze Age/Early Iron Age pottery from midden layer 1, contexts 1055 and 1061

Figure 159 Late Bronze Age/Early Iron Age pottery from midden layer 1, context 1100

157

Multi-period Occupation at Football Field, Worth Matravers, Dorset jar (fig. 54.64) with a long date range of 10th-6th centuries BC.

8. Flat rim (R2) with a suggestion of external thickening. F23. 9. Slightly pointed simple rim (R1) from a carinated bowl. F23. 10. Body sherd decorated with a horizontal row of impressed wedges. F27. 11. Body sherd from the belly of a jar, decorated with a diagonal fingernail impression. F27. 12. Body sherd from the belly of a jar with a single lentoid impression on the carination.F23. The sherd has an internal residue. 13. Upright base sherd (B1) from a probable jar, decorated with short incised lines near the base. F23.

Context 1100 upper fill of Slot 1099 Clearance by mechanical excavator left only 0.1m of midden material in this layer. The fill was finds-rich and included 191 sherds of pottery weighing 1721g with an average sherd weight of 9g. Fig. 159 1. Near-rim (probably R1) and body sherds from a large vessel, with a rim diameter of c. 320mm and evidence of a sharp carination, decorated with curved-line furrows. Concreted deposit on the inside of the pot. F28. The decoration suggests that the vessel was probably a jar. 2. Fine, slightly flattened rim sherd with proto-bead and external groove, probably from a bowl (R5). F7. 3. Flat upright rim (R2) very slightly externally expanded and wiped. F27. 4. Simple everted rim (R7) decorated with two wide furrows. F25. 5. Simple rim (R1) very slightly everted with one slight impressed circle. F17. 6. Small everted rim sherd (R7), F33. 7. Simple rim sherd (R1) from a bowl with a rounded carination, similar to Potterne Type 7 (fig. 49.37). F20.

Context 1104 single fill of Slot 1103 This contained large slabs of limestone and mixed-date finds. In total 195 sherds of pottery weighing 1121g (average sherd weight 5.79g) were recorded. Fig. 160 1. Everted rim sherd (R7) with traces of red-coat, from a large jar with a rim diameter of 240mm. F38. Not enough profile was present to determine whether this was either a Potterne Type 31 or 32 carinated jar; its deep everted rim suggests a Type 32 vessel (fig. 54.60) with a broad date range of 10th-6th centuries BC.

Figure 160 Late Bronze Age/Early Iron Age pottery from midden layer 1, context 1104

158

The Late Bronze Age/Early Iron Age Pottery 2. Flat rim sherd (R2) with rounded edges from an internally wiped large vessel. F28. 3. Flat rim sherd (R2) with rounded edges and very slight internal bevel. The interior surface is scored with fine incised lines and the exterior is red-coated. F7. 4. Slightly everted simple rounded rim (R1) from a bowl. F11. 5. Slightly everted simple rounded rim (R1) from a bowl. Decorative band below the rim consists of short diagonal lines contained within parallel incised lines. Similar decorations are found on both bowls and jars at Potterne. F25. 6. Simple rim sherd (R1) with a very slight external expansion from a finely-made bowl decorated with diagonal fingernail impressions contained within parallel incised lines. F22. 7. Upright rolled rim (R6), the outside of which is decorated with closely spaced diagonal slashes. F27. 8. Body sherd decorated with impressed dots above an incised line. F25. 9. Coarse, abraded base sherd (B2) with a diameter of 140mm and slightly ‘waisted’ above the base. A similar base profile is recorded on a Type 40 jar from Potterne (fig. 55.72). F32.

Early Iron Age. In total, 145 sherds of pottery weighing 1185g were recorded with an average sherd size of 8.2g. Fig. 161 1. Simple, slightly flattened upright rim (R1) from a redcoated bowl. F22. 2. Everted rim sherd (R7) decorated with long incised vertical lines contained within two parallel incised lines. F22. 3. Simple, upright rim sherd (R1) probably from a jar, with heavy residue on both internal and external faces. F22. 4. Carinated body sherds from a large vessel with red coating above the carination and absent below. F25. Despite the lack of rim, this can probably be compared to the Potterne Type 3.4 bipartite bowl series which are characterised by flaring rims (fig. 49.31). 5. Base sherd (B3) with internal residue from a jar. F25. Context 1112 single fill of Slot 1111 Contained two bone needles (Chapter 11, Fig. 197.3-4) and 153 sherds of pottery weighing 786g (average sherd size 4.7g).

Context 1108 single fill of Slot 1107 Mixed date for finds but predominantly Late Bronze Age/

Figure 161 Late Bronze Age/Early Iron Age pottery from midden layer 1, contexts 1108, 1112 and 1116

159

Multi-period Occupation at Football Field, Worth Matravers, Dorset Fig. 161

Fig. 162

6. Thick flat rim (R2) heavily abraded under the rim surface. F23. 7. Body sherd decorated with wide furrows forming triangles. At Potterne, both bowls and jars display this motif. F28.

1. Rim sherd (R1) from a jar displaying horizontal wiping under the rim. F25. 2. Flat, upright rim sherd (R2) with horizontal wiping. F7. 3. Rolled rim sherd (R6). F7. 4. Rim sherd with internal bevel (R12); interior smooth and outside slightly abraded. F25. 5. Thick-walled gently rounded rim (R1) with internal wiping and patches of grey concretion on the outside of the vessel. F7.

Context 1116 single fill of Slot 1115 Contained large limestone rubble and shale spindle whorl. In total 43 sherds of pottery weighing 462g (average sherd weight10.7g) were recorded.

Context 1122 single fill of Slot 1121 Mixed assemblage containing 79 sherds of pottery weighing 724g (average sherd weight 9.2g).

Fig. 161 8. Simple upright rim (R1) from a furrowed bowl with traces of red-coating. F7. 9. Everted simple rim (R7) from a bowl with a shallow furrow under the rim and a diagonal furrow above the carination; traces of red-coat. F25. 10. Simple upright rim (R1) from a bowl; interior and exterior patches of brown slip. F7. 11. Very unusual sherd possibly from near the rim of a carinated vessel, possibly a bowl, displaying a lip under the rim. Despite the small sherd size, it would appear that the lip did not extend around the complete pot. F27. A similar rim was recorded at All Cannings Cross (plate 36.1a) where it was unique and was thought to have helped support a lid.

Fig. 162 6. Slightly everted flat rim (R2) and body sherds from a crudely-made jar. The sub-circular depressions may be spall sites. There are traces of external burnt residue and internally there are small patches of grey concretion. Both the interior and exterior have been heavily wiped. F20. Can be compared to Potterne Type 20 round-bodied jars (fig. 51.42) dating to the 8th-6th centuries BC. 7. Body sherds with a raised cordon decorated with short diagonal slashes. F11. Similar to the shouldered tripartite Potterne jar Type 58 (fig. 59.93). It was thought that this vessel style was unique to that site: the Football Field pot differs in that the decoration is on the belly of the pot rather than its neck. 8. Body sherd, probably from a jar, decorated with a sharply incised triangle and a single row of impressed dots on two sides. There are remnants of white chalk in both the incised lines and the dots. F38.

Context 1120 upper fill of Slot 1119 Containing a deposit of tabular limestone pieces and a mixed assemblage of finds including 121 sherds of pottery weighing 1154g (average sherd weight 9.5g).

Figure 162 Late Bronze Age/Early Iron Age pottery from midden layer 1, contexts 1120 and 1122

160

The Late Bronze Age/Early Iron Age Pottery

Mixed assemblage containing 40 sherds of pottery weighing 562g (average sherd weight 14g).

7. Red body sherds decorated with an incised vertical line which encloses at least five vertical rows of impressed dots, some of which are filled with a brown-coloured paste. F8.

Fig. 163

Context 1407 layer under 1402

1. Base sherd (B2) from jar; exterior pimply, interior smooth. F17. 2. Body sherd decorated with a smooth dimple contained within a wide, circular furrow. F29. 3. Part of a rod-type handle from a jar. F25. A jar from All Cannings Cross (plate 38.4) had two handles of similar form.

Comprising a spread of limestone rubble. In total 20 sherds of pottery weighing 235g were recorded (average weight 11.75g).

Context 1167 upper fill of Slot 1166

Fig. 163 8. Simple rim fragment (R1) from a carinated bowl which displays areas of grey slip on the outside. Compares with Potterne Type 2 bowl fig. (47.14). F27. 9. Simple upright rim sherd (R1) from a red-coated, widemouthed dish with an approximate rim diameter of 210mm. Splashes of grey concretion near the rim. The profile is similar to an illustrated example of a deepbodied dish from Eldon’s Seat (Type 6, fig. 15.96). F7. 10. Simple upright rim sherds (R1) from a carinated bowl, diameter 160mm. F27. Similar to Potterne longnecked bowl Type 2 (fig. 47.15) with a date range of late 8th-6th centuries BC.

Context 1277 upper fill of Slot 1261 Containing 206 sherds of pottery weighing 1837g, and an average sherd weight of 8.9g. Fig. 163 4. Exaggerated, externally expanded rim (R9). F8. 5. Carinated bowl sherd with simple everted rim (R7) similar to Eldon’s Seat Type 4 (fig. 17.144). F25. 6. Small carinated sherd decorated with a horizontal row of fingernail impressions, from a small jar. F25.

Figure 163 Late Bronze Age/Early Iron Age pottery from midden layer 1, context 1167

161

Multi-period Occupation at Football Field, Worth Matravers, Dorset Contexts with pottery of intrinsic interest

Fig. 164

Pottery from the following contexts, which are all associated with the midden deposit, has been selected because of the intrinsic interest of form or decoration.

1. Round-bodied jar with a slightly everted flat rim (R2) and a rim diameter of 170mm. F23. The profile compares with Potterne jar type 20 (fig. 51.42) and another from Longbridge Deverill (Type P 23, fig. 3.11/17 with dates of the 8th-6th centuries BC). 2. Rim sherd (R1) from a carinated bowl with a diameter of 120mm. F11. Similar to Hengistbury Type BA, which includes vessels with sharp shoulder angles (fig. 56/1500) with a broad Early Iron Age date, and Potterne Type 3.4 (fig. 49.32). 3. Cup-type bowl with high, rounded shoulder and a rolled rim (R6), with vertical wipe marks; diameter

Context 302 clearance This was the general overburden layer on Site 1 and yielded very large numbers of finds of Roman, Late Iron Age and Late Bronze Age/Early Iron Age dates. A total of 113 sherds weighing 1084g (average sherd size 9.6g) were recorded.

Figure 164 Late Bronze Age/Early Iron Age pottery of intrinsic interest from Site 1; contexts 302, 305, 307, 313 and 339

162

The Late Bronze Age/Early Iron Age Pottery

4. 5. 6. 7. 8. 9.

110mm. F25. Compares with Potterne cup type 86 bowl (fig. 60.116) dating to the 8th-7th centuries BC. Slightly flattened, incurving sloping rim (R2) from a jar. Decorated with faint lines which were incised after firing. F23. Body sherd decorated with deep, incised dots contained within a vertical incised line and the remnants of the apex of an incised triangle. F7. Body sherd decorated with at least four rows of deep, incised dots. F7. Body sherd decorated with deeply incised lines, probably triangular pattern. F7. Body sherd decorated with a curved line; vertically wiped externally and horizontally wiped internally. F11. Body sherd decorated with a deeply incised horizontal line. F27.

Context 339 residual pottery in fill of Late Iron Age pit 324 In total 11 sherds weighing 55g with an average sherd weight of 11g were recorded. Fig. 164 15. Unusual rim: flat form with an external bevel (R25), the latter decorated with short diagonal slashes. F25. 16. Small body sherd decorated with sinuous dendritic irregular incised lines. F28. 17. Body sherd decorated with a shallow furrow and impressed circular dots, some of which still have traces of white paste. F3. Context 1013 upper fill of Slot 1011 Most finds were Late Iron Age; two sherds of Late Bronze Age/Early Iron Age pottery were recorded weighing 54g (average sherd weight 27g).

Context 305 fill of Late Iron Age storage pit Residual finds comprising 47 sherds weighing 782g (average sherd weight 16.6g).

Fig. 165

Fig. 164

1. Body sherd from under the rim of a bowl with a horizontal furrow and three incised diagonal lines which would have formed part of a triangle. F7. 2. Body sherd from a round-bodied jar with unusual decoration of a very lightly worked furrow and very short, fine horizontal lines. A further, deeper line was inscribed after firing. F25.

10. Jar with long neck and a flat, very slightly externally expanded, upright rim (R2) with a diameter of 150mm. F29. 11. Short upstanding rim (R8) from a jar with two lightly incised horizontal lines below the rim. F8. 12. Fragment from a jar base (B3) approximately 160mm in diameter. F11.

Context 1096 clearance level above Slots 1099 to 1121

Context 307 clearance above Late Iron Age pavement 307

All the pottery was Late Bronze Age/Early Iron Age comprising 50 sherds weighing 629g (average sherd weight 13g).

This context contained only one Late Bronze Age/Early Iron Age sherd and it weighed 7g.

Fig. 165

Fig. 164

3. Sherd from the carination of a jar or bowl decorated with short, deep diagonal slashes. F23. 4. Body sherd decorated with small stabs, with remains of white chalk in-fill. Remnants of a furrow on one side. F34.

13. Simple upright rim sherd (R1) with short diagonal lines on the outside of the rim and vestiges of these on the body. F11. Context 313 fill under the Late Iron Age pavement 303

Context 1098 second fill of Slot 1084

A total of 96 sherds weighing 1000g were recovered (average sherd weight 10.4g).

Contained 39 sherds of pottery weighing 262g (average sherd weight 7g).

Fig. 164

Fig. 165

14. Large sherd from a shouldered jar with light fingerprints on the outside of the flat rim (R2). Before firing, a badly executed line was incised on top of the rim; internal wiping. F2. Compares with Potterne shouldered jar Type 51 (fig. 51.81) with a very long currency of use from possibly the early 10th century into the early 6th century BC.

5. Upright rim with very slight internal bevel (R12) and decorated with very faint, short, incised zigzag lines. F7.

163

Multi-period Occupation at Football Field, Worth Matravers, Dorset Context 1133 midden layer below Slots 1017, 1033, 1035 and 1050

Context 1422 Roman pit 1421 Two sherds of residual pottery weighing 33g (average weight 11g).

Immediately above cobbles. Copper alloy items were recorded; these included a ring and parts of a rolled bead (see O’Connor Chapter 11). The pottery assemblage comprised 199 sherds weighing 1121g (average sherd weight 6g).

Fig. 165 10. Fine bowl sherd decorated with diagonal incised lines between which is a parallel row of impressed dots placed above the carination. F20.

Fig. 165 6. Jar sherd decorated with a vertical incised line and a row of impressed dots, all of which were filled with chalk. F20. 7. Body sherd from the carination of a jar decorated with a horizontal row of fingerprints. F11. 8. Body sherd from a fine bowl decorated with narrow furrows which formed a triangular pattern. F7. 9. Carinated sherd displaying a horizontal row of short diagonal slashes. F11.

Context 1553 Fill between the Early Bronze Age stone alignments 1447 and 1486 Three sherds of pottery were recorded weighing 39g (average sherd weight 13g). Fig. 165 11. Flat rim sherd (R2) with a very slight rim indentation. The body is decorated with randomly applied boneimpressed circles and a single ring and dot motif. F25.

Figure 165 Late Bronze Age/Early Iron Age pottery of intrinsic interest from Site 2; contexts 1013, 1096, 1098, 1133, 1422, 1553 and 1800

164

The Late Bronze Age/Early Iron Age Pottery case, due to the relatively shallow depth of the midden and its subsequent later disturbance.

Context 1800 Clearance under access road Finds included 28 sherds of pottery weighing 128g (average sherd weight 5g).

Cultural Associations The period 850-600 BC is associated with the adoption and development of ironworking, and in the pottery record, the appearance of carefully made and decorated finewares in a completely new range of forms. Utilitarian, possibly conservative coarsewares associated with storing foodstuffs harked back to styles of the previous millennium. There was a marked increase in the varieties of potting fabrics, rim forms and decorative motifs which may indicate new cultural influences and values, combining imitation and innovation.

Fig. 165 12. Body sherds from the carination of a red-coated bowl decorated with diagonal furrows, and tiny abraded nicks on the carination. F8. The Assemblage The Football Field material in the main comprised jars and bowls, with the addition of a very few cups. Within the bowl category were a number of shallow, wide-mouthed dishes. Most sherds were very small (the average sherd size across the assemblage is 9.46g) and it was not possible to assign form to many of the sherds. However it is likely that the coarser sherds would have derived from jars and the finer ones from bowls.

Between 1911 and 1922, Maud and Howard Cunnington excavated a site at All Cannings Cross Farm, Wilts (Cunnington 1923) where a very large assemblage of pottery was recorded. The resulting analysis provided the basis for future typology, and the site itself gave its name to the specific pottery types which have continued to be found from an ever-increasing number of sites in central southern Britain (Cunliffe 2005, figs 5.1, 5.3 and 5.4). Recently, Paul Tubb looked at the Wiltshire sites in great detail, and concluded that further sites in Purbeck (see below) are also part of the cultural and landscape changes which were taking place at this time, particularly in relation to ‘dark earth or midden’ sites (Tubb 2011). Similar sites are recorded in the Thames Valley at Runnymede Bridge (Needham and Spence 1996) and Wallingford (Cromarty et al. 2006) and further afield at Whitchurch in Warwickshire (Waddington 2011). Local sites are discussed below. It is highly likely that new sites are yet to be discovered.

Because of the shallowness of the midden deposit (maximum depth 0.3m) on Sites 1 and 2 and the homogeneity of the soil make up, it was unfortunately not possible to determine stratigraphical depositional differences. An added problem was that the midden had been disturbed by later activity. Because the excavation strategy encompassed taking the soil deposit off in approximate layers on Site 2, and despite difficulties associated with intrusive structural debris, it was still possible to plot the distribution of pottery in each layer and this is illustrated on Figures 166-169 (also see Table 4). The depth of the top of the midden (Layer 1) varied from 0.1m to 0.2m and it was from this layer that the largest numbers of pottery sherds were recorded (Fig. 166). The sherds were spread across the site but with a greater concentration towards the east. This area was least affected by later activity. Figure 167 shows that the largest concentration associated with the 0.5-0.1m deep Layer 2 is on the northern part of the site, where particularly large numbers of sherds occurred. Pottery from Layer 3 (Fig. 168) which was 0.1m deep has a general spread in the centre of the site, with the highest densities again occurring on the northern side.

The Late Bronze Age/Early Iron Age pottery from Football Field falls into the broad All Cannings Cross tradition, spanning the 9th to the 7th centuries BC; and is typified by: • round-shouldered jars with out-curved rims; • bipartite bowls with beaded rims and sharp shoulder angles; • tripartite jars; • wide-mouthed dishes; • cups and miniature vessels; • barrel-shaped jars; • large decorative repertoire mainly consisting of geometric incised lines, often in-filled with impressed and stabbed motifs and fingernail and fingerprint impressions. In addition, many pots have a red finish.

Pottery weights from layers 4, 5 and 6 (with an approximate aggregated depth of 0.11m) were combined and are illustrated on Figure 169 where it can be seen that the sherds were generally sparse, but the highest concentrations were located on the southern part of the site. This material was incorporated into the cobble layer itself (where excavated) and the layer directly above this.

Vessel Types at Football Field Although the number of sherds from Football Field was relatively high, there were no complete pots: indeed the average sherd size was only 9.5g and it was not always easy to identify vessel types. Rims, bases, body shape and decoration were used as identifiers, and all typical vessel categories appear to have been in use on the site. A total

It should follow that the earliest ceramics would have been associated with the construction of the cobble layer and the initial accumulation over this (Layers 6, 5 and 4) and that the topmost layer (Layer 1) would have incorporated the latest pottery in the sequence. This was not always the 165

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 166 Distribution of Late Bronze Age/Early Iron Age pottery from Site 2, midden layer 1

166

The Late Bronze Age/Early Iron Age Pottery

Figure 167 Distribution of Late Bronze Age/Early Iron Age pottery from Site 2, midden layer 2

167

Multi-period Occupation at Football Field, Worth Matravers, Dorset

Figure 168 Distribution of Late Bronze Age/Early Iron Age pottery from Site 2, midden layer 3

168

The Late Bronze Age/Early Iron Age Pottery

Figure 169 Distribution of Late Bronze Age/Early Iron Age pottery from Site 2, midden layers 4, 5 and 6

169

Multi-period Occupation at Football Field, Worth Matravers, Dorset Bronze Age/Early Iron Age pottery, it appeared best to compare the Football Field assemblage with what was already known, rather than create a new type series. Table 34 compares the Football Field types with those from other major excavated sites in the region. The vessels were recorded as bowls, jars or cups, using profiles, rims, thickness, finishes and decoration as indicators. Large dishes were also recorded; at Eldon’s Seat they were termed ‘pans’ however at Potterne they were considered an element of the bowl category. The published series varied from simple (Hengistbury with 7 vessel types and Eldon’s Seat with 11) to relatively complex (Potterne with 32 types and Longbridge Deverill with 50 variants). The report on the small assemblage from the adjacent Compact Farm excavation (Seagar Smith 2002) was also considered.

Table 33 Numbers and percentages of (illustrated) Late Bronze Age/Early Iron Age vessel types Vessel type

number

% assemblage

Jars

101

32.8

Bowls

92

29.9

Cups

12

3.9

Dishes

6

1.9

Miniatures

1

0.3

Unassigned

96

31.2

Totals

308

100

of 308 sherds were chosen for illustration; these highlight the variety of forms present on the site, but because of their small size, 96 of these could not be assigned a vessel type.

Vessels in Use at Football Field Compared with the total number of sherds present, there were relatively few vessels where rim or base diameters could be taken and there were very few instances where base to mouth measurements could be recorded.

Jars were the commonest form, followed by bowls, with a very small percentage of dishes, cups and miniatures (Table 33). It has to be stressed that this is a very small proportion of the total number of vessels which must have been discarded at Football Field, but illustrates the relationship between the vessel types and probably reflects the relative proportions of types in use. Table 33 may also give an indication of the proportions of vessel usage, for example for serving or storage: and through that, an insight into potential on-site activities.

The Potterne typology series is probably the most useful for vessel typology and comparison, principally because approximate dates for currency of use are assigned to the vessel types, and is therefore used here for dating wherever possible. Table 34 correlates the major sites used for vessel comparison, and indicates how the Football Field types fit within the published typologies. A comparison was made between the thicknesses of illustrated jars, bowls, dishes and cups where identification to type was positive (Table 35) and it can be seen that thickness may be a rough indicator of type. Looking at the complete data base (held in archive), the thinnest sherd measured 2mm and the thickest measured 20mm. It would seem that jars, which are often the largest pots, were the sturdiest vessels, commensurate with a need for durability and stability. Bowls and cups were much thinner, an indication that they were easier to carry and did not have to contain large or heavy quantities of (possible) foodstuffs or liquids. Dishes, probably used for serving food, varied in thickness and in size, which probably related to the material they held.

Jars were probably used for storage of dried materials, perhaps grain, flour or dried fruits. They may also have stored other non-food commodities. It is assumed that bowls were probably used for serving food; the smaller ones may have been used to eat from. The wide, open dishes also may have been for serving food. The assumption is that cups were used for drinking, but the small surviving numbers hint that they may not have been the everyday preferred option. It is possible that small bowls were used for drinking. Finally, the use of the tiny miniature vessels is open to speculation. For this report, the pottery typology was compared in the main, to the major published assemblages from Potterne (Wilts) and Hengistbury Head (Hants) and from the recently published site at Longbridge Deverill, Cow Down (Wilts). Although All Cannings Cross is the ‘type-site’, and the material was described in detail, a type series does not exist for this assemblage. Pottery types from nearby Eldon’s Seat were also used as comparisons, however that assemblage has recently been re-evaluated (Needham and Woodward forthcoming) and some types have been re-dated. Pottery from local Purbeck sites which Bernard Calkin published from the 1940s to the 1960s was also considered; however only very basic vessel or sherd type information was available on those assemblages.

Jars Ten specific jar types were recorded (Table 34) with a date range of 10th/9th to the early 6th centuries BC, indicating a long currency of use for a number of these vessels. Some of the forms evolved from Middle and Late Bronze Age types, particularly the barrel shapes and fingerprint/ tip decorated forms. Barrel-shaped jars were absent from Potterne but present at All Cannings Cross and Longbridge Deverill. Sharp carinations appear to be an innovation. It was possible to construct mouth diameters for 19 vessels from Football Field and these varied from 100mm to 320mm (Table 36). Although the sample is small, it indicates a range of vessels varying from small to very

As there is no standard typology for the transitional Late 170

The Late Bronze Age/Early Iron Age Pottery Table 34 Correlation of Late Bronze Age/Early Iron Age pottery typology P (Potterne), LD (Longbridge Deverill), H (Hengistbury), CF (Compact Farm), ES (Eldon’s Seat) Type

Description

Jar

Barrel

Jar

Globular, flaring rim

20

Jar

Carinated with lugs

30

Jar

Carinated, everted rim

31

Jar

Carinated, tripartite

32

Jar

Slight carination, bipartite

33

Jar

Situlate, shouldered, upright neck

51

Jar

Shouldered, tripartite, cordons

58

Jar

High-shouldered, fingerprint dec.

Gp18

Jar

Rounded body

Gp12

Bowl

Biconical, sometimes beaded

1

Gp7

BA

10/9-early 6C

142.4; 144.1; 151.1; 155.13

Bowl

Carinated, long neck, furrowed

2

Gp4

BB

Late 8-early 6C

142.3; 142.5; 142.9; 145.2; 146.2&7; 147.3; 152.2; 158.2&3; 163.8&10

Bowl

Carinated, bipartite, upright rim

3.1

Gp4

8-7C

139.10; 140.3; 142.6; 143.3; 145.1&3: 149.8; 153.10&11; 158.5; 164.4

Bowl

Carinated, large, tripartite

3.3

Gp4

7-early 6C

140.4: 163.5

Bowl

Carinated, bipartite, furrows

3.4

Gp4

10/9-early 7C

161.4; 164.2

Bowl

Round shouldered

7

EIA

144.3

Bowl

Carinated, accented curve

Dish

Wide, open, shallow

11

Gp33

R1

Cup

simple hemispherical

80

Gp7

R25

Cup

Carinated, biconical, may be furrowed

81

Cup

Sharply carinated

Cup

Rounded shoulder, beaded rim

Miniature

P

LD

H

Gp 25

R2

Gp 31

JB2.0

CF

ES

Date BC

PI1 R4

Football Field Fig. no. 142.10; 148.13; 153.2

8-6C

137.7; 138.5; 139.11; 143.1&2; 150.1: 151.4; 162.6; 164.1

EIA

143.4

PI2

10/9-early 6C

142.1

10/9-early 6C

153.16; 160.1

Gp 31

10/9-early 6C

143.5; 151.10; 158.7

Gp26

10/9-early 6C

144.7; 147.2; 148.3; 152.3; 150.5; 156.16; 164.14

Gp18

JB1.0

8-7C JB1.0

154.9; 162.7 EIA 155.10

PII1

BA

PII3

Gp9

151.11 9-8C

140.1; 141.1; 148.4; 155.10; 156.3; 163.9

7C

145.5; 153.14; 158.4

Gp7

8-7C

143.7; 154.1; 167.3

83

Gp7

9-7C

149.10

86

Gp7

8-7C

148.11; 149.9; 164.3

87

PII6

157.13

studs or rivets found on bronze vessels. The second vessel was very small, perhaps a miniature pot (Fig. 157.11-12). Three plain examples were recovered from Potterne. The Football Field example is decorated on the shoulder with rough fingerprints and resembles a tiny situlate jar. The function of this vessel type is not known. They could have been children’s pots or maybe contained small quantities

large; their main use would have been to store dry goods such as cereals and pulses. Two unusual jar types were recorded; one of them was a bossed vessel (Fig. 154.4). This was recorded locally at Gallows Gore and Eldon’s Seat and also at All Cannings Cross, where Cunnington postulated that bosses imitated 171

Multi-period Occupation at Football Field, Worth Matravers, Dorset Table 35 Late Bronze Age/Early Iron Age pottery: comparison of vessel thickness (jars, bowls, dishes and cups) Thickness (mm)

3

4

5

6

7

8

9

10

11

12

13

14

15

Total Nos vessels

Jars (numbers)

0

2

14

9

11

18

17

21

5

4

4

3

3

111

Bowls (numbers)

4

21

29

13

12

7

5

3

0

0

0

1

0

95

1

1

1

1

1

Dishes (numbers) Cups (numbers)

2

6

2

7 11

3

Table 36 Late Bronze Age/Early Iron Age pottery: illustrated jar mouth diameters and number of vessels Diameter (mm)

100

110

120

140

160

170

180

200

220

240

250

320

No. vessels

1

1

2

3

1

3

2

1

2

1

1

1

Figure 170 Late Bronze Age/Early Iron Age jars and bowls – fabric groups

of foodstuffs or perhaps exotic material.

and may have had such a function. The residue is likely to be remains of food burnt during cooking processes.

Figure 170 shows the fabric groups used for making jars: it can be seen that Groups 3 (grog tempered with mixed inclusions), 5 (sand only) and 8 (limestone plus inclusions) predominate, being 43%, 22% and 15% of the total. Groups 3 and 5 are both tempered with sand and Group 8 includes large percentages of limestone. The temper mix was no doubt determined by availability, proposed use of vessel and in all likelihood, long-held custom. Both internal and external burnt residue is recorded on seven of the illustrated jars; this type of vessel is often referred to as a cooking pot

Bowls Seven bowl types were recorded at Football Field (Table 34) with dates ranging from the 10th/9th to the early 6th century BC, showing similar dates of use to those of jars. They appear to develop from forms used during the Late Bronze Age. At Bestwall Quarry, three types (shouldered, open and biconical) were recorded (Woodward 2009, 245247) and radiocarbon dates place the pottery sequence

Table 37 Late Bronze Age/Early Iron Age pottery: illustrated bowl diameters and number of vessels Diameter (mm)

100

110

120

130

140

150

160

180

170

180

190

200

215

220

240

330

No of vessels

1

1

3

2

4

1

4

6

1

6

1

3

2

2

1

1

172

The Late Bronze Age/Early Iron Age Pottery walled types were in use. The dishes are not decorated, although three were red-coated. One vessel had patches of external residue indicating that it may have been used for cooking foodstuffs. The vessels may have had multiple functions being used for the cooking, preparation and serving of foodstuffs.

there from the 960s or 950s cal BC to the 790s cal BC (Bayliss et al, 2009, 146-149). There was no transitional pottery from Bestwall Quarry despite the very large amounts of Middle and Late Bronze Age material and Late Iron Age and Roman wares. All three types of bowl were recorded from the small Late Bronze Age assemblage from Football Field, and several are recorded as having traces of external red coat. The styles continued in use and developed into new forms. Furrows are a new decorative device for bowls and are commonly horizontal on the upper part of the vessel, but are also used in geometric designs such as chevrons. The currency of use for this design harks back to bronze and even gold examples such as the Ringlemere cup which dates to around 1950-1750 BC (Needham et al, 2006, 26). However its use on carinated bowls appears first in the Early All Cannings Cross group which Cunliffe dates to the 9th to 8th centuries BC (Cunliffe 2005, fig. A2.1214) and where it is often a feature of short-necked bowls. Later examples of this type often display longer necks. In general, bowls had fewer decorative motifs than jars. Other motifs included parallel horizontal grooves containing fine diagonal slashes and other incised patterns. Occasionally rims were decorated with notches or fingernail impressions.

Cups At Potterne, this vessel type was loosely described as being ‘generally smaller than bowls or jars’ and was divided into six variants which comprised a very small proportion of the pottery assemblage. Four of these variants are present at Football Field (Table 34), are always thin walled, and vary between 3mm and 5mm in thickness, with diameters ranging from 60mm to 110mm. Rims were present on ten of the eleven cups; forms R1, R2, R6 and R10 were recorded with 6, 2, 1 and 1 examples respectively. Six fabric types were noted (F7, F17, F22. F23, F25 and F29) and apart from F29 which is shell-tempered, all are sand-based; this mix may have been particularly suitable for very small, thin-walled vessels. To a degree, the forms emulate larger jars and bowls, and their use and function was probably associated with small quantities of foodstuffs and drinks. Their limited number suggests that they were never a common element of the pottery repertoire. Decorative motifs appear to be limited and consist of rim grooves, horizontal furrows, fingerprints and short slashes. At Potterne their currency of use was from the 9th to the 7th centuries BC.

Table 37 details the 32 examples where rim diameters could be measured. These vary from100 mm to 330mm, suggesting a variety of uses for this pot type: indeed the two smallest examples at 100mm and 110mm could well have been used as cups. It is thought that bowls may have been used for serving, eating from, or for the display of foodstuffs. The vessel thicknesses generally varied from 3mm to 10mm with only one thicker illustrated example (Table 35). The proportions of fabrics in use were very similar to those used in jar manufacture (Fig. 170) but two fabric groups (3 and 5) were dominant at 43% and 35% respectively. Both of these groups include sand as a major component of the mix. Many of the bowls were fired red, and red coating was a common feature.

Decorative motifs Large repertoires of decorative motifs are recorded on pottery from all sites where transitional pottery is known. These were recorded in detail at Potterne where designs were assigned to six broad categories (Gingell and Morris 2000, 153). The aspect of decoration both as inscribed motifs and finish is considered by Jane Randall below. Decoration may have had a practical use, making the vessel easier to handle or hold, or it may have been a significance marker, characterising specific uses or functions. The following decorative schemes were noted:

Dishes

1. finger-tip decoration by fingers or fingernails; 2. applied strips or cordons; 3. incised parallel lines (vertical and horizontal) forming sharp incisions, grooving and furrowing; 4. diagonal lines comprising chevrons, triangles, diamonds, excluding horizontal and vertical patterns; 5. continuous sinuous patterns around the widest part of the vessel; 6. impressed or stabbed zones, often associated with, and contained within linear motifs.

Dishes are wide, open and characterised by their shallowness in proportion to their width. Although only seven vessels were positively identified, the number is likely to have been much higher. At Potterne, they existed as a sub-class of the bowl category with a date range from the 9th to the 8th centuries BC. On other sites, these vessels are sometimes referred to as pans. There were three measurable diameters from the Football Field assemblage (210mm, 220mm and 230mm) and it is likely that the other dishes were of similar sizes. No bases survived. Five rim forms were recorded: one vessel had a simple rounded rim (R1), two had flattened rims and four displayed expanded rims (R9, R10 and R16). Five fabrics (F7, F20, F22, F23 and F25) were recorded and all are sand-based. The vessel thicknesses are recorded on Table 35 where it can be seen that both thin-walled and thick-

At Football Field there is an additional category: 7. applied bosses. These were absent at Potterne and Hengistbury Head, but recorded at All Cannings

173

Multi-period Occupation at Football Field, Worth Matravers, Dorset was perhaps used to store water. Residue analysis was not undertaken.

Cross, Longbridge Deverill and on the Purbeck sites at Eldon’s Seat and Gallow’s Gore.

Discussion

Fingernail and fingerprint impressions have a long currency of use stretching back to the early Neolithic period and continuing into the Early Iron Age. They are found on the carinations of high-shouldered vessels such as jars and are not usually associated with bowl forms. Fingernails and fingerprints also decorated the rims of jars. This style of decoration fades out during the Middle Iron Age. The other decorative motifs continue and evolve into the early Late Iron Age decorative repertoire.

On-site production Evidence for prehistoric pottery production is often ephemeral and difficult to pinpoint. Bonfires and fire pits are likely to have been small and may have only been used once, their archaeological footprint leaving little trace. In addition, the number of vessels in each firing would have been small. It is this author’s belief that prehistoric potters were skilled and therefore few firings would result in spoilt vessels. Indications for on-site firing at Football Field include the re-use of a pit as a possible clamp, and spalled and over-fired pottery sherds.

Surface treatment Various surface treatments of vessels, applied before firing when they were probably in the leather-hard stage, were noted on both on the exterior and interior surfaces and were recorded accordingly: 1. 2. 3. 4.

The clay-lined pit 1513 (Fig. 29 location and Fig. 36 section) displayed evidence for possible pottery production. A distinct layer (1518) with a depth of 0.15m comprised burnt red loam which contained fractured and burnt limestone and may have been either the base of a pottery clamp, or debris from firing. Pottery was found both above and below this layer; significantly most of the pottery sherds were located below. Amongst these were 43 rounded spalls blown from a vessel wall during the early stages of firing when the temperature rose too quickly (Gibson and Woods 1997, 156), strongly suggesting pottery firing in the vicinity, with debris being placed in layers in the pit. In addition spalls, although not common, were found scattered in the slot fills. A number of sherds of over-fired pottery were also noted: these were characterised by their light weight and grey colouration and may have been wasters.

red coat; white chalk infill; other colour coating; wiped.

Red coat The exterior surfaces, generally of bowls (and very occasionally jars) were coated with an iron-oxide slip or perhaps a dry application of this mineral, which occurs locally in the heathstones of the Bagshot sands and gravels. After firing, the outside surface of the vessel was glossy and red. White chalk infill Impressed and incised designs, usually on black/grey fired pots, were filled with a chalk paste accentuating and highlighting the design.

Similar local assemblages Excavated evidence at Worth Matravers indicates that Late Bronze Age/Early Iron Age presence was based on regular, but probably not continuous activity resulting in large deposits of accumulated and discarded artefactual material (the midden) into which pits and other smaller features were dug. Pottery was recovered from cut features as well as the midden. During Southampton University’s work on the Compact Farm site (see Fig. 4) in the early 1990s, similar deposits dating to the ‘Early Iron Age’ were encountered. These produced much smaller pottery assemblages, corresponding to those found during the 2010-11 excavations with similar vessel types of tripartite bowls and situlate and ovoid jars (Seager Smith, 2002, 55). A proportion of jars and bowls were noted as having red coating. There were few decorated sherds: these included fingernail impressions, incised lines and pastefilled, impressed dots.

Colour coats Clay slips were applied to the exterior of the pot giving the fired vessel a cream, brown or grey finish. Only four examples of such finishes were recorded at Football Field. Internal and external wiping This was recorded on a number of vessels and was likely to have been part of the manufacturing process; however the lack of apparent smoothing may have been deliberate and associated with the proposed use of the vessel, perhaps to give a better grip. Residues In total 19 sherds bore evidence of burnt material adhering to either the inside or outside of the vessel, suggesting that the pots had been placed over a heat source and burnt food (or drink) survived as carbonised residues. A white residue, possibly limescale, was recorded on a single vessel which

Paul Tubb’s overview of Late Bronze Age/Early Iron Age transitional sites in the Vale of Pewsey in Wiltshire (Tubb 2011) suggested a number of comparable sites in Purbeck (ibid, 182-192), which had yielded pottery assemblages 174

The Late Bronze Age/Early Iron Age Pottery imported. It is also possible that the original influence for the characteristically decorated All Cannings Cross style pottery came to Britain through these routes (Cunliffe 2005, 590); this may have included high-shouldered bronze buckets (or situlae), the style of which was imitated in several jar styles and could have been emulated in local Purbeck pottery assemblages. The question remains whether the Purbeck pottery styles were influenced from the Wiltshire heartlands, or was it the other way around?

concurrent with this period and similar to the Wiltshire assemblages. Many of the Purbeck sites in the vicinity of Worth Matravers were recorded during the 1930s and 40s by local school teacher Bernard Calkin, when they were destroyed through limestone quarrying (Calkin 1949). The Purbeck All Cannings Cross style pottery assemblages all comfortably fit within the Late Bronze Age/Early Iron Age transitional period. The sites (from east to west) are Herston, Acton, Gallow’s Gore, Sheepsleights, Westhill, Encombe, Kimmeridge, Hobarrow, Worbarrow and Flower’s Barrow (see Chapter 16, Fig. 217). All are located in a ten mile strip, within one to two miles apart; the majority have views over the English Channel. Also noted are Blashenwell which is slightly further inland, but is still located on the limestone uplands; and Fitzworth on the southern shore of Poole Harbour. Since that time, only one other site has been excavated in detail. Sir Barry Cunliffe undertook work in 1963, 1964 and 1966 at Eldon’s Seat, when comparable assemblages were recorded (Cunliffe 1968).

Timescale of use Maud Cunnington realised that the assemblage from All Cannings Cross which was excavated between 1911 and 1922 was of great importance, indeed when the site was published in 1923 it was regarded as a locus classicus (belonging to the highest class) by E. Goddard, Hon. Secretary of the Wiltshire Archaeological Society who wrote the Foreword to the published volume (Cunnington 1923). The artefacts from the site were of a single period with no earlier or later material and were dated to the ‘period of overlap between the late Hallstatt and the earlier La Tene 1’ which now equates to the Late Bronze Age/ Early Iron Age transition. When this material was written up there was little pottery available for comparison, but it had been noted that similar material had been excavated from Hengistbury Head and from a number of sites in Wiltshire. However these sites also produced finds from other periods and close dating of the pottery relied on associated metalwork.

Further along the coast approximately 25km to the east, the extensive multi-period site at Hengistbury Head was excavated between 1979 and 1984 (Cunliffe 1987). The pottery assemblages are again similar in form and decorative details to those from the sites noted above. All of these sites have produced ceramic material loosely described as All Cannings Cross Type and broadly dated (at the time of publication) from the 8th to 4th centuries BC. The forms and decorative motifs show both a continuation of the traditional styles of previous centuries and a completely new repertoire of vessel types and decoration. It is clear that the plainware and decorated traditions of the Late Bronze Age as seen at Bestwall Quarry (Woodward 2009, 244) were subsumed into innovative vessel forms. However, activity at Bestwall ceases in the 790s BC (Bayliss et al 2009, 149): prehistoric ceramic traditions which had been focussed on this Poole Harbour site for the previous 2000 years came to an apparently abrupt end. There was no evidence for any All Cannings Cross type pottery on this extensive (55ha) site.

Although much smaller in volume, the Football Field assemblage can be compared to that from Potterne where the accumulation of various layers of the midden deposit was estimated to have begun in the Late Bronze Age and where activities had ceased by about 600 BC (Lawson, 2000, 271-274). As the Worth deposit was relatively shallow by comparison, the pottery here was confined to a single layer which was (at its maximum) only 0.3m deep and which had been disturbed in places by extensive Late Iron Age and Roman occupation activity. Conclusion

Areas of influence

The Late Bronze Age/Early Iron Age pottery at Football Field is closely associated with the construction of the cobble layer which extends over much of the site, the build-up of soil comprising the midden deposit which overlay this feature and the subsequent insertion of pits, scoops and placed deposits into the midden as it was developing.

The assemblage from Football Field is important in that it has yielded a considerable amount of ceramic material. Although the sherd size is generally small, the variety of identifiable vessel types is relatively large. The question of cultural links with the well-known Wiltshire midden sites (All Cannings Cross, East Chisenbury and Potterne) and the recently published settlement site at Longbridge Deverill is crucial; as are relationships with potentially similar sites in Purbeck (Tubb 2011, 182-192). All of the Purbeck sites are located within a couple of miles of the present coastline and continental maritime trade must have been of importance. It is accepted that the south coast of Dorset and Hampshire was a contact zone which was linked with Armorica and Normandy and it was through coastal settlements that Armorican metalwork was

Stylistically, the pottery conforms to the All Cannings Cross typology, with dates ranging from the 9th to the 6th centuries BC and with a wide range of vessel types and forms and can be usefully compared to material from a wide range of Dorset and Wiltshire sites. It is likely that form was related to function and use. Reasons for pottery deposition within the midden and 175

Multi-period Occupation at Football Field, Worth Matravers, Dorset his Wiltshire sites (both in the Vale of Pewsey), where slip-coating containing haematite particles was used (Middleton 1987, 252-3).

features are beyond recovery. The small size of the sherds and low number of conjoins is probably consistent with the dumping of already broken material, which may have moved around substantially. The extensive variety of recorded types implies access to the repertoire of newly evolving styles and designs, the potters possibly emulating material from other areas but at the same time adapting and developing designs and motifs, some of which had enjoyed a long currency of use. Specific designs may have had significant meaning, but again the reasoning behind this is irrecoverable.

For the 2014-15 work, museums and other organisations in Dorset and Wiltshire were approached for permission to examine reserve collections, and test suitable sherds (i.e. those, which on visual examination were thought to have a red finish, and were also fineware if possible) from Late Bronze Age/Early Iron Age sites in the geographic area between South Dorset and the Vale of Pewsey. Despite the proposed test being destructive, there were no refusals to participate, and 21 sherds from 18 sites, (7 in Dorset, 11 in Wiltshire), including two sherds from Football Field, were tested. This involved thick sections being sawn off and the freshly cut surface being polished then photographed using a high-resolution digital microscope (VHX-5000). The surface boundary between the red layer and the fabric body was then examined to see if it was either smooth (indicating a slip coat) or uneven (indicating a pressed-on powder).

On this site, middening by disposal of rubbish in the immediate vicinity was recorded in the earlier excavations, and appeared to continue into at least the Late Iron Age (Graham et al, 2002, 10). The recent work confirms this. As a result, there were many mixed contexts with both earlier and later material present. This was particularly apparent in areas where later features penetrated the midden deposit. The tradition of depositing rubbish was long-standing and only ended when activity on the site ceased at the end of the Roman period.

The results for the two sherds from Football Field, together with the single sherds from nearby sites at Herston and Gallows Gore, showed conclusively that their red finish was applied by the burnishing of iron oxide powder into leather-hard clay, thus substantiating Middleton’s findings for Purbeck sites. This methodology was not found elsewhere, although two results from sites near Salisbury were inconclusive. More surprising perhaps were the results from the rest of Wiltshire, where only samples from three sites bordering the Vale of Pewsey had a slip coating, whereas those from the middle and south of the county, including samples from Longbridge Deverill Cow Down, had no applied coating whatever; the typical redbrown appearance had been achieved just by burnishing the vessel surface before firing.

Comparative Analyses of the Main Decorative Surface Treatments Found on the All Cannings Cross Ware Assemblage from Football Field Jane Randall† All Cannings Cross (ACC) ware, with its distinctivelyshaped jars and bowls, always includes a minority of decorated vessels: some with a red finish derived from an iron ore, traditionally referred to as ‘haematite coating’: others having impressed or incised patterns, occasionally with a white inlay. Ladle (this volume, Chapter 16) raises the question as to whether the ACC ware pottery styles found at Football Field, and at other Purbeck sites, were influenced from the Wiltshire heartlands (Vale of Pewsey) where ACC ware was first recorded and where the largest concentration of sites in England has subsequently been discovered. Or was it the other way around, with European ideas coming first to Purbeck’s shores and spreading inland? It was the possibility that studies of ACC decoration might also contribute to this important topic of debate that initiated this research in 2014 for an undergraduate dissertation (Randall 2014). Comparative analyses were carried out (2014-5) on two aspects of the decoration (red finish and pattern types) found on the Football Field assemblage.

The detail of these results is presented with a degree of caution due to single samples from most sites and the assumption of ‘household production’ of pottery as defined by Peacock (1982, 8) or at least, mainly the use of localised materials (Morris, 1994, 27). The difference in methodology in the application of red finish between all tested examples from both the South Dorset and Vale of Pewsey sites is consistent, and may stem from the geology, rather than technology. The iron ore (haematite) which produces the red colouration was available in the Vale of Pewsey; but the most abundant source of iron ore in the Isle of Purbeck (which was also available at Hengistbury Head), was limonite (also known as goethite), which converts to haematite by losing water molecules when heated. It may be that this conversion took place during the pottery firing; hence the appearance of the red colouration would not be so effective if the ore were dispersed in a slip (slips of other colours were found on a tiny number of Worth sherds), as opposed to burnishing in the powder. Experimental trials would be useful in checking the results of the two methodologies using limonite ore; but meanwhile, no conclusions can be drawn concerning the

Analysis of the application of red-finish This analysis followed work in the 1980s by Middleton, who tested 35 ACC ware sherds from a total of 12 sites in Kent, Hampshire, Wiltshire and Dorset, to ascertain the application methodology of the red surface finish. His results, using a scanning electron microscope, showed that the application method of burnishing iron oxide powder into leather-hard clay, was used at four south Dorset sites (three from Purbeck and the other being Hengistbury Head), differed without exception from that found in 176

The Late Bronze Age/Early Iron Age Pottery one very simple addition not found at Potterne; this was a single incised line just below the rim of the vessel as shown in Fig. 139.3.

different technologies and/or the lack of transfer between South Dorset and Pewsey Vale. Analysis of decorative patterns

By far the most frequently-occurring decorative motif at Football Field was the corrugation of the furrowed bowls as illustrated in Fig. 143.3; 105 sherds had this type of decoration compared with 29 (28%) which had a red finish. This is surprising, because the haematite-coated furrowed bowl has traditionally been seen as typical of Wessex ACC ware (Cunliffe 2005, 92) and because the Football Field assemblage has a relatively high proportion of red-finished sherds overall (more than 11%), compared to from 1%-3% at Potterne. Hence there seem to have been cultural or aesthetic choices made as to whether furrowed bowls were red-finished or not. It would be interesting to see whether the proportion of furrowed bowls that were red-finished changed through the period of the Late Bronze Age/Early Iron Age transition, as the memory faded of the Late Bronze Age furrowed bronze vessels upon which they are generally assumed to have been based, such as the carinated and furrowed cup from the Welby Hoard (Powell 1950, 35; pl.V1).

Each decorated sherd was examined, and those with any incised or impressed pattern were classified according to the system of six design groups and 39 decorative motifs devised by Gingell and Morris (2000, 153-7; 331-337) for the ACC ware typesite at Potterne, Wiltshire. This was a very large, deep midden, a small part of which was excavated in layers in 1982-5 (Lawson, 2000). The incidence of similar patterns at both sites was then compared, using percentages of the whole Football Field assemblage, and of the relevant layers from Potterne (midden layers there started earlier and ended later than those at Football Field). In the course of this work, sherds with a red finish were also counted to compare incidence at the two sites. The results showed some considerable differences. Whereas just 4% of the Worth sherds were decorated, at Potterne in the later part of the Late Bronze Age/Early Iron Age transition, the proportion of decorated sherds rose to 18%. However at Football Field, finger-tipping made with the ends of the fingers indenting around the necks or shoulders of vessels occurred infrequently, whereas at Potterne, except in the uppermost layers, it was always more common than parallel-line patterns. Fingertipping was virtually the only decoration of the Late Bronze Age plain ware which preceded ACC ware. Less common in Wessex than in other central southern areas of England (Barratt 1980, 309-11), it tends to be found in ACC assemblages on the shoulders or rims of larger, coarseware vessels (Cunliffe 2005, 93). The dominance of this decoration at Potterne could be seen as either an indicator of food preparation on a larger scale, or a longer continuation of a traditional pattern in this particular region. Cunnington also noted its dominance at All Cannings Cross, remarking that ‘sherds of vessels with a finger-tip decoration were more numerous there than all the other pottery put together’ (Cunnington 1923, 33).

Conclusions Overall, these analyses have demonstrated that two assemblages of ACC ware, both from middens, but of very different sizes and locations, differ very substantially in their proportion and variety of decoration. That the assemblage from Potterne, a huge site probably drawing participants from a very wide catchment, is actually not typical in many ways of that from a much smaller site with a more localised catchment, is perhaps not unexpected. However it has been shown that further analysis of the decoration on other smaller assemblages, may be another way of adding to the detail of cultural groupings in Late Bronze Age/Early Iron Age transition societies. A query has been raised against the generally-held perception that a red finish was commonplace on furrowed bowls of ACC ware, and it has been demonstrated that quantitative studies of red-finish, which only use optical observation, would be unlikely to produce accurate results.

At Football Field, only 10 examples of the 39 decorative motifs recorded at Potterne were found, together with

177

10 The Middle Iron Age, Late Iron Age and Roman Pottery The pottery assemblages discussed below comprise material belonging to the Middle Iron Age, the Late Iron Age and the Roman period. Due to the mixed nature of many of the assemblages, attribution was often by the character of the featured sherds, particularly in relation to the small Middle Iron Age assemblage.

Eight rim forms from 32 vessels were identified, the most common being the simple R1 (10 examples). Rim R9 with a pronounced external expansion and R18 with an external bead and internal bevel were represented by four examples each. There were three examples of the proto-bead with groove under (R5), the everted R7 and the simple bead rim (R11). Flat rim R2 was represented by two sherds as was the internally bevelled R12. There was a single example of the rolled rim Type R25. The small number of rim sherds precludes meaningful analysis. It seems that complex rim decoration, which had been a common feature of earlier pottery, had gone out of use (or fashion). However grooving directly underneath the rim, which was occasionally noted in the preceding centuries, became more common, with 18 of the 32 rim sherds displaying this characteristic. Generally the pots were plain, and decoration was uncommon; when it occurred it took the form of incised curves or large dimples. The surviving fragments from Football Field suggest that cooking jars were the dominant vessels in use.

Middle Iron Age Pottery Lilian Ladle Introduction In Dorset, there is an inherent difficulty in assigning pottery to the Middle Iron Age. The various researchers are not clear or in agreement regarding the chronological boundaries of the period. Indeed the present chronological divisions are purely a device for archaeological purposes enabling approximate dating initially through artefactual seriations, and now through absolute radiocarbon dating. A charred wood (Bromus secalinus) sample from pit 1182 on Site 2 was submitted for dating and returned two dates: 342-326 cal BC or 205-48 cal BC (SUERC61171) confirming that the infilling of this feature and its associated finds falls within the conventional Middle Iron Age.

The pottery series from local sites have been used for identification and comparison and include Hengistbury Head (Brown 1987), Rope Lake Hole (Davies 1987) and Stoborough (Lyne 2002).

Only one feature (1182) is of secure Middle Iron Age date and 184 sherds of Middle Iron Age pottery weighing 2299g were retrieved from its fills. The latest infill sequence of this pit was compromised by a later pit (1131) which was cut into its uppermost fill (Fig. 50). Two features on Site 1 (pits 326/390 and 393) although dated on their ceramic assemblages to the Late Bronze Age/Early Iron Age, also produced a small amount of Middle Iron Age material.

Table 38 Middle Iron Age pottery fabrics, groups and number of sherds

Fabric and form A total of 13 fabric types (F7, F8, F11, F17, F20, F22, F25, F26, F27, F28, F29, F30 and F34) were identified. Although the sample size is too small for meaningful analysis, the fabrics fell into six of nine the groups discussed at the beginning of Chapter 9 and are detailed on Table 38. Group 3 predominated (F17, F20, F25 and F34) and although grog-based, the mixes contained sand and/or limestone inclusions. The sand-based Group 5 (F7 and F8) was well-represented. Indeed, apart from Group 9 (F29) which included shell, the other groups (6, 7, and 8) contained sand and limestone mixes. This mix repertoire was not as diverse as in the earlier period. The reasons for this may have been associated with availability of resources and possible constraints on the potters themselves. 178

Fabric Type

Group

No. sherds

Fabric description

F7

5

26

Fine sand

F8

5

17

Coarse sand

F11

7

13

Limestone

F17

3

4

Grog and limestone

F20

3

8

Grog, limestone and sand

F22

6

4

Sand plus small ironstone

F25

3

45

Grog plus sand

F26

6

1

Sand, limestone & mica

F27

8

22

Sand plus limestone

F28

8

3

Sand, limestone plus ironstone

F29

9

1

shell

F30

6

3

Sand plus shale

F34

3

1

Grog, sand and ironstone

The Middle Iron Age, Late Iron Age and Roman Pottery Pottery from pit 1182 (Fig. 50)

3. Top fill 1159. Rim sherd (R12) with internal bevel and an external groove where two attempts had been made to score an incised line. The interior of the sherd is burnished. F22. Similar to a Type JC 2.0 Hengistbury high-shouldered, rounded profile jar (fig. 136.1299). 4. Top fill 1159. Flat rim (R2) with well-defined and skilfully executed groove. F25. 5. Top fill 1159. Small, very thin, simple rim sherd (R1) with an external groove. F27. 6. Second fill 1151. Two rim sherds (R17), with a flattened top and groove under. Remnants of grey concretion on the inside. F27. Hengistbury Type JC 3.1 high-shouldered jar (fig. 139.1729). 7. Second fill 1151. Slightly everted R1 from a largenecked jar, similar to Fig 171.1. F2.

There were five discrete fills in this sub-circular pit. A total of 179 sherds of pottery weighing 2198g were recorded from eight of the eleven recorded contexts. All thirteen fabric types were represented and the average sherd weight was 12.4g. Fig. 171 1. Top fill 1159. Everted rim sherd with slight external expan­sion (R9) from a jar. F27. A similar rim from Rope Lake Hole (fig. 81.49) belonged to a largenecked storage jar. 2. Top fill 1159. Rim sherd (R9) with a defined groove under the external expansion. F22.

Figure 171 Middle Iron Age pottery from pit 1182

179

Multi-period Occupation at Football Field, Worth Matravers, Dorset 8. Second fill 1161 (equivalent to 1151). Simple rim (R1), the top surface is much abraded and there is internal residue adhering. F22. Possibly from a saucepan pot similar to Hengistbury Type PB1 (fig. 180.1264). 9. Second fill 1161. Simple upright rim (R1), F7. 10. Third fill 1156. Flat topped bead rim (R11) with a slightly wavering incised groove below. F25. Probably from a small jar. 11. Third fill 1156. Flat topped bead rim (R11) with a welldefined groove below; the interior of the rim is very smooth. Diameter 90mm. F26. Similar to Hengistbury Type JC 2.0 (fig.136.1726). 12. Third fill 1156. Bead rim with an internal bevel (R18) and a slight internal groove with a well-defined, but poorly executed external groove, external residue. F20. From a jar. 13. Third fill 1156. Bead rim with internal bevel (R18) and well-defined external groove. Part of a curvilinear design under the rim. F20. From a small bowl or jar in the Maiden Castle/Marnhull style (Cunliffe 2005, 107). 14. Third fill 1156. Chunky simple rim (R1). F20. 15. Third fill 1156. Base sherd (B2) from a thick-walled jar. Grey concretion on outer side of base. F25. 16. Third fill 1163 (equivalent to 1156). Wall sherd decorated with a shallow, curvilinear furrow and short random incised lines. F25. 17. Third fill 1163. Bead rim with an internal bevel (R18) in a very pimply fabric. F8. 18. Third fill 1163. Small externally rolled rim sherd (R25) from an upright vessel, perhaps a saucepan pot. A similar vessel from Hengistbury is fig. 180.1264. F27. 19. Third fill 1163. Proto-bead rim (R5) with a very shallow groove. F7.

20. Third fill 1163. Wall sherd with two very fine parallel incised lines, internally wiped. F27. Fig. 172 1. Base fill 1199. Small rim sherd (R9) with an external expansion. Exterior has a trace of red-coat. F26. 2. Base fill 1199. Wall sherd decorated with a narrow curved furrow. F25 The following illustrated sherds are intrusive in two Late Bronze Age/Early Iron Age pits on Site 1 Fig. 172 3. Second fill 356, Pit 326/390. Large everted rim (R7) sherd from a Hengistbury Type JC 3.0 jar decorated with a circular dimple (fig. 135.665). This type is also recorded at Stoborough (Lyne, fig 6.1). F25. 4. Single fill 394 of pit 393. Proto-bead rim R5 with two grooves, the lower better defined than the upper. Two faint diagonal finger-marks on a burnished surface. F7. 5. Single fill 394 of pit 393. Everted rim (R7) with faint uneven dashes under the rim. F7. 6. Single fill 394 of pit 393. Proto-bead rim R5 with no groove, similar to Hengistbury Type JC 2.0 jar (fig. 136.1731). F27. 7. Single fill 394 of pit 393. Slightly everted R2 rim, with faint lines etched on the flat portion and a poorly executed, partial groove underneath. Another Type JC 2.0 jar, rim diameter 170mm. F7. 8. Single fill 394 of pit 393. Base fragment probably from a large jar displaying small internal dimples. F7. Locally, similar decoration is found on Late Iron Age

Figure 172 Middle Iron Age pottery from pits 1182, 326/390 and 393

180

The Middle Iron Age, Late Iron Age and Roman Pottery to determine the natures, forms, frequencies and sizes of the filler and naturally-present inclusions. A numbered fabric series was then devised.

vessels from Rope Lake Hole (Woodward 1987, fig. 81, 74-5) and Stoborough (Lyne, fig. 6.2). Conclusions

The fabrics

Although the Middle Iron Age assemblage is small, it can be compared to similar assemblages recorded on the coast at Rope Lake Hole, by the River Frome at Bestwall and Redcliffe, and on the shores of Christchurch Harbour at Hengistbury Head. The range of ceramics spans an approximate 300 year period and an overlap of styles and techniques can be seen at both ends of the time frame. Red-coating, which characterised many of the Late Bronze Age/Early Iron Age ceramics is present in the earliest material, which interestingly was located on the base layer of pit 1182. Much of the material is in the Maiden Castle/Marnhull style which was current from approximately 400 to 200 BC and is characterised by large ovoid jars and smaller round-bodied bowls, often with bead rims. Decoration consists of scrolled arcs, wavy lines and dimples (Cunliffe 2005, 107). A couple of rim sherds derived from saucepan pots. These straight sided vessels were recorded at Hengistbury but not previously at any Purbeck site. Their distribution is from Sussex in the east to Gloucestershire in the west and the Cotswolds to the north; they tend to be conspicuous by their absence in Dorset. The latest Hengistbury vessels are dated to c. 100 BC and are associated with Dressel 1 amphorae (Cunliffe, ibid, 106).

National Roman Fabric Reference Collection codings are bracketed at the ends of fabric descriptions where applicable (Tomber and Dore 1998) F4. Handmade with sparse small flint filler F7. Handmade Durotrigian/BB1 fabric with fine sand filler (DOR BB1 variant) F8. Handmade Durotrigian/BB1 fabric with coarse sand filler (DOR BB1 variant) F9. Handmade with fine sand filler, as well as sparse, medium and large flint inclusions F15. Handmade with grog and sand filler F22. Handmade with sand and small ironstone inclusions F23. Handmade with sand and large ironstone inclusions F25. Handmade with grog and ferrous inclusions F26. Handmade with fine sand and sparse small limestone inclusions F27. Handmade with fine limestone filler F28. Handmade with fine limestone, ironstone and sparse shell inclusions F30. Handmade with sand and shale filler F39. South Gaulish samian (LGF SA) F40. Martres-de-Veyre samian (LMV SA) F41. Central Gaulish samian (LEZ SA) F42. Very-fine whiteware with profuse