Culverwell Mesolithic Habitation Site, Isle of Portland, Dorset: Excavation report and research studies 9781841710242, 9781407319193

Report on the excavations undertaken by the author at the Culverwell site in Dorset. Finds of virtually every period of

192 69 274MB

English Pages [281] Year 1999

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Front Cover
Title Page
Copyright
Dedication
Frontispiece
Table of Contents
List of Figures
List of Photographic Plates
List of Figures, Diagrams, Tables or Charts within the Text
Preface and Acknowledgements
Part 1. Excavation Report and Archaeological Assessments
Part 2. Environmental and Dating Evidence
Part 3. Interpretative Research Based on the Findings of the Excavations
Part 4. The Future of the Culverwell Mesolithic Site
Appendices
Bibliography
Figures
Plates
Recommend Papers

Culverwell Mesolithic Habitation Site, Isle of Portland, Dorset: Excavation report and research studies
 9781841710242, 9781407319193

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Culverwell Mesolithic Habitation Site Isle of Portland, Dorset Excavation report and research studies

Susann Palmer with contributions by

Kenneth D. Thomas, Myfanwie Stewart, John D. Gale, Helen Keeley and Beverley Collinson

BAR British Series 287 1999

Published in 2016 by BAR Publishing, Oxford BAR British Series 287 Culverwell Mesolithic Habitation Site, Isle of Portland, Dorset

© S Palmer and contributors and the Publisher 1999

The author's moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. ISBN

9781841710242

paperback

ISBN

9781407319193

e-format

DOI https://doi.org/10.30861/9781841710242 A catalogue record for this book is available from the British Library BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 1974 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by Archaeopress in conjunction with British Archaeological Reports (Oxford) Ltd / Hadrian Books Ltd, the Series principal publisher, in 1999. This present volume is published by BAR Publishing, 2016.

BAR PUBLISHING BAR titles are available from:

BAR Publishing 122 Banbury Rd, Oxford, OX2

EMAIL PHONE FAX

[email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

7BP, UK

TO

MY

FAMILY

Frontispiece: Aerial view of the Isle of Portland, seen from the Fleet with Chesil Bank on the right connecting the Isle to the mainland. The Verne Heights are in the centre of the photograph (northern end of the Isle) and the Culverwell Mesolithic site is at the southern end of the Isle (right of the photograph).

CONTENTS

List of Figures

iii

List of Plates

1V

List of Tables and Diagrams within the Text

Vi

Preface and Acknowledgements

Vil

PART 1 EXCAVATION REPORT 1.

AND ARCHAEOLOGICAL ASSESSMENTS

Introduction: Site Location History of site discovery and investigation Investigation Techniques

Stratigraphy of the site and its features Description of features and archaeological findings

18

Assessment of features

31

A Geophysical Survey of the Culverwell Site by

7.

Archaeological Material Excavated:

Artefacts:

49

John D. Gale

53

Descriptions and discussions

67

8. The Use of Non-knapped stones by Myfanwie Stewart

PART 2 ENVIRONMENTAL AND DATING EVIDENCE 9. Geology and Geomorphology

85

10. Dating

91

11. The Bioarchaeology of the Culverwell shell midden by K. D. Thomas and others

94

PART 3. INTERPRETATIVE REPORT

RESEARCH

BASED

ON

THE

FINDINGS

OF

THE

EXCAVATIONS

12. Economy: food, artefact & building material; skins; trade

115

13. Activity Zones on and near the Site

124

14. Continual or interrupted occupation: Planning

and intra-site structure

128

15. Population studies

134

16. Reconstruction of a Mesolithic-style hut or shelter (hypothetical)

136

17. Evidence for spiritual life of the site inhabitants

139

18. Comparisons with Site 1; Assessment of unexcavated sites on Portland

142

19. Culverwell in the wider Context of the later Mesolithic period

146

20. Relationship between the Mesolithic sites and other prehistoric sites on the Island

154

AND

PART 4. THE

FUTURE

OF THE CULVERWELL MESOLITHIC SITE

Management of a Heritage Project

157

APPENDICES 1. Ancient Monument Laboratory Report 2985 Series 68/79: Mollusc Analysis by Beverley Collinson

2. Artefact Classification

Sheet

Soil Samples by H. Keeley

161

164

3. Summary of the Report

165

BIBLIOGRAPHY

167

FIGURES

173

PLATES

215

il

LIST OF FIGURES (Please Note: All other figures, feature plans and section drawings executed on the site, other than those included with the rest of the archival material.)

below, are

Maps, Plans and Section drawings of Trenches and Features Fig 1 Southern England and the Isle of Portland — location maps. Fig 2 Culverwell site plan and location of trenches. Plan of the Culverwell Mesolithic site and various features. Fig 3 Fig 4 A and B. Section through Trenches 4 and 3. Fig 5 Profile of a transect through the floor and midden in the NNW of the site. Plan of Floor and Hearth 4 in Area A and B (part). Fig 6(1) Fig 6(2) Part of floor and features in Areas A and B (part). Part of Area A and B floor and features (part). Fig 6(3) Fig 6(4) Part of Area A and B floor and features X and Y (part). Part of Area A and B showing feature Z and postholes (part). Fig 6(5) Fig 6(6) Part of Area A and B showing slope into gully (part). Plan of Hearth 4 and adjacent paved area. Fig 7 Fig 8 Plan of paved area adjacent Hearth 4. Fig 9 Section through Hearth 4 and adjacent paved area. Plan of the floor in part of Trench 3. Fig 10 Plan of Hearth One and adjacent area. Fig 11 Fig 12 Section through the edge of the floor and Hearth one. Plan of Feature 3 in relationship to Feature 2 (Sondage 10). Fig 13 Fig 14 Section through Feature 2 and underlying floor and Feature 3 (Sondage 10 and 10A). Fig 15. NNW section through medieval lynchet indicating strata where the soil column and samples were taken. Fig 16 NNW section of Sondage 3 near Feature X (ritual feature). Fig 17 Section through the lynchet on a line through Features Y and X (ritual features). Fig 18 Plan of the ritual feature X and Y and the area round it with detail inset after excavation Trench 2: section through the pit. Fig 19 Plan of Trench 2 with details of the pit after excavation. Fig 20 Fig 21 A and B Profile of strata above and below the floor in Trench 6. Fig 22 Plan of the the limestone floor in the sondage through Trench 6. Trench 40: Sections and plan of workfloor area. Fig 23 Trench 40 Artefact density chart. Fig 24 Trench 41 section indicating original midden dump within the gully. Fig 25 Fig 26 Plan of an enigmatic stone feature in Trench 41. Figures of Archaeological Material: Fig 27 Examples of microliths, microburins and Krukowski microburins. Fig 28 Examples of picks. Fig 29 More examples of picks. Fig 30 Examples of burins. Fig 31 Examples of different types of scrapers. Fig 32. More examples of scrapers. Fig 33 Variously retouched flakes and blades.

Fig Fig Fig Fig Fig Fig

34 35 36 37 38 39

Variously retouched artefacts, some reminiscent of Late Palaeolithic types.

Various core tools: chopping tools and a transversely retouched axe. Cores retouched for use as scrapers. Axe and pick trimming flakes. Four examples of conjoining flakes and cores. Examples of cores and post-Mesolithic artefacts from the site.

Figures illustrating the contribution on Non-knapped Stones by M. Stewart:

Figures S1 — S87

Non-knapped stones.

iii

all prefixed with an S

LIST

OF

PHOTOGRAPHIC

PLATES

Frontispiece

Aerial view of the Isle of Portland from the Fleet

Plate 1 Plate 2 Plate 3 Plate 4 Plate 5 Plate 6 Plate 7 Plate 8 Plate 9 Plate 10 Plate 11 Plate 12 Plate 13 Plate 14 Plate 15 Plate 16 Plate 17 bottom Plate 18 Plate 19 Plate 20 Plate 21 Plate 22 Plate 23 Plate 24 Plate 25 Plate 26 Plate 27 Plate 28 Plate 29 Plate 30 Plate 31 Plate 32 Plate 33 Plate 34 Plate 35 Plate 36 Plate 37 Plate 38 Plate 39 Plate 40 Plate 41 Plate 42 Plate 43 Plate 44 Plate 45 Plate 46 Plate 47 Plate 48 Plate 49 Plate 50 Plate 51 Plate 52 Plate 53 Plate 54 Plate 55

Aerial view of the Bill area with the Culverwell Mesolithic site marked by an arrow Aerial view of the Culverwell Mesolithic site, looking West to East.

The Central part of the floor looking towards the Old Lower Lighthouse View of the Culver Well stream from the Mesolithic site View of the floor and other features looking towards the sea View of the floor and the wall of Feature 2 Part of the floor in Area A with two straight lines of limestone slabs Trench 4 with the edge of the floor and circular feature next to it End of the floor in the NW of the site The junction between the floor and Feature 2a Feature 2 looking east Section (sondage 10) through the Feature 2a showing a third feature below it Feature 2 and semi-circular arrangement of stones of Feature 3 Section through the gully of Trench 32 Two large unstruck pebbles in the bottom of the gully in Trench 32 A posthole in the gully near the edge of the floor (Trench 32) The top of a posthole in the midden in the gully adjacent the floor (Trench 32); pebbles in the gully

View of the floor and Hearth 1 Close-up of Hearth 1 Choppers, choppingtools and unstruck pebbles round Hearth 1 Artefacts associated with Hearth 1 Close-up of a pick and a pounder by Hearth 1 Close-up of a post-hole by the edge of Hearth 1 A charred object near the edge of Hearth 1 A cast of the charred object in situ The pit beginning to show as a discoloured circle in the midden The pit after excavation; sampling for archaeomagnetic dating The pit after partial excavation before removal of Stone 1 The stones at the bottom of the pit after excavation The small paved area at the edge of Hearth 4 A cluster of picks near the edge of Hearth 4 Part of the floor with the large stone X (ritual feature) showing Close-up of Stone X witb large beach pebble Close-up of Stone X and the beach pebble before full excavation The hole under Stone X The ritual feature after lifting Stone X and the artefacts found in the hole under it The floor after excavation of the lynchet showing a change of direction; the area round the ritual feature Another large triangular stone Stone S near Feature 2 Feature 41 with a large beach cobble (Trench 41) Trench 41: the edge of a compact shell midden visible at the base of deposits in the gully The compact midden material in Trench 41 Grooves resembling aard marks in Trench 41 Profile of the different strata in Trenches 4 and 31 Layers on top of the floor (Trench 3) Layers above the midden deposit in Trench 4 A cutting through the floor in Trench 6 A close-up view of the midden deposits directly under the floor Four examples of picks Three further examples of picks Three picks and other artefacts in situ on the workfloor of Trench 40 Picks in various stages of completion as found near Hearth 4 Examples of microliths from the Culverwell midden Examples of shell beads A countersunk pebble A countersunk pebble 1V

Plate Plate Plate Plate Plate Plate Plate Plate Plate Plate

56 57 58 59 60 61 62 63 64 65

A core and several conjoining flakes Small pieces of bone which appear to have been worked to form points Portland Site 1: part of a limestone paved area Site 1: close-up of a midden and paved area with two almost identical cores Site 1: Another paving of limestone slabs Site 1: enigmatic semi-circular stone structure associated with midden debris and human bones A sea-cave in the cliffs below the site and chert seams visible Seams of dark siliceous chert in the Jurassic limestone Culverwell Management Project: a protective cover over part of the site Mesolithic-type hut constructed during a Heritage project

Plates to accompany the contribution on the Use of Non-knapped Stone by M. Stewart, Plates S1 — S8

Examples of Non-knapped stones

prefixed by an S

LIST OF

FIGURES,

DIAGRAMS, TABLES OR CHARTS

WITHIN THE TEXT

Page no.

7 11 29 50 51 52 61

Diagram of a lynchet. Thickness of Layer 11 in Trench 4. Diagram of marks on top of the midden in Trench 41. Location of geophysical grid and Transect A. Gale Figs. 2 and 3 Gradiometer and Resistivity survey. Gale Fig. 4 Resistivity Transect A. Microlith types.

63

Table 1. Total s and distribution of microliths.

64

Table 2. Microliths from different layers of Trench 4.

67 68 69 70 71 74

MS: Column 1 Trench 4. MS: Column 2 Trench 4; Column 3 Trench 41 MS: Column 4 Trench 41. MS: Column 1 comparison Purbeck Chert and molluscs; Column 2 comparison Purbeck chert and molluscs. MS: Column 3 comparison Purbeck chert and molluscs; Column 4 trench 41 comparisons. MS: Disk distribution.

75 87 95 96 98 99 100 101 102 103

MS: Summary smooth and abraded pebbles. Geological strata of Portland. KT: Table 1 Column samples from the midden; Table 2 Description of the layers sampled. KT: Table 3 List of marine invertebrate species. KT: Table 4 Marine molluscs from Sample Column 1. KT: Table 5 Marine molluscs from Sample column 2. KT: Table 6a Marine molluscs from Sample column 3. KT: Table 6b Marine molluscs from Columns 4 and 5 (Layer8 Trench 41). KT: Fig. 1. Abundance of the 3 main species through the midden. KT: Fig. 2. Changes in the age profiles of Monodonta lineata through the midden.

105

KT:

Table 7a Radiocarbon determinations on Monodonta shells, with calibrations.

107

KT:

Table 8 Bone fragments and teeth from the Culverwell excavations.

109

KT:

Table 9 Land and freshwater molluscs (plus a vole;

110 111 112 122 137 160 161

KT: Table 11 Land and freshwater molluscs from column 3. KT: Table 12 Land and freshwater molluscs in flotation samples; KT: Table 14 Preliminary identifications of wood charcoals. Portland chert artefact distribution in southern England. Construction of Mesolithic-style hut. Keeley: Soil profiles trench 3 NW face. Keeley: Soil profiles Section 1 Trench 4.

Vi

Table 10 Land and freshwater molluscs from column 2.

Table 13 post-midden deposits.

PREFACE

AND

ACKNOWLEDGEMENTS

The Culverwell research project was originated and directed by the author of this Report and was commenced when she formed the Portland Field Research Group during her studies at the University of London and the University of Southampton. She left the P.F.R.G. after a few years when it became mainly involved in local development issues. She later formed the Association for Portland Archaeology whose members since then have given much needed support to the work and hopefully will continue working for the future management of the site. The site was discovered by the author as a result of a field survey of the enthusiastic help of Messrs Richard Cooper of Portland and Alan Bromby, to Mr. R.W. Stone, who then owned the larger part of the site (Field 2154), initial years of work and to Mrs Bradley for allowing work for two years northerly part of the site (Field 2153).

Isle of Portland which was undertaken with then warden of Brownsea Island. Thanks is for giving permission for excavation during on the field which she then owned, forming

the due the the

The continued survival and research of the site was made possible by the intervention of Mr. Norris McWhirter without who the work would have been forced to stop many years ago as a result of several attempts to utilise the land for other purposes. To him much gratitude is owed for support, interest and encouragement. In the early years of my work on Portland I was greatly encouraged by the interest of the late W. F. (Bill) Rankine and the late A.D. (Donald) Lacaille, both known for their interest in early prehistory. Much interest was shown by the late Dr. John N. Carreck of Queen Mary College, London, who spent

most of his youth on Portland. The late Mr. Bob Cooper of

Portland was very staunch in his support and help in practical ways. All aspects of the work throughout has been done by volunteers from England and abroad, far too many to mention individually. Throughout the years the work was_ regularly supported by members of the Orpington and District Archaeological Society (ODAS) in diverse ways, in particular Messrs B. Bull, M. Meekums, T. Pavely and T. Gray. Many of the Director’s numerous students from the 1“ year of the University of London’s Extra-mural Diploma course, the I.L.E.A. and different branches of W.E.A. helped throughout the years and many of them worked on the site in order to gain their qualifications. The Director is well aware of the fact that without their constant and highly efficient work the site of Culverwell would never have been excavated. To them all many thanks. Other volunteers who have helped me constantly in one way or another are R. Kitchingman, C. Jessel, Mrs B. Stainton, Mrs. L. Camp, Mr and Mrs C. and M. Woods and the late Marchese Alessandro Nobili Vitelleschi. Several people, in addition to those already mentioned, the excavations had finished:

have regularly helped to process the archaeological material after

Jane Yeo, Dee Reilly, Brian MacKinnon, Mary Thompson and Doreen Riddle.

Many establishments and academics have given support and sustained interest in various aspects of the project: the Dorset Natural History and Archaeological Society and the Weymouth College of Higher Education for help with storage space and tools in the early years, English Heritage; the Prehistoric Society; International Quaternary Research Association; Portland Museum; the Weymouth Museum;

Portland Bird Observatory and Field Centre; British Museum:

Prehistoric and

Quaternary Departments and Laboratory; Weymouth and Portland Borough Council various departments; previous members of the Institute of Archaeology in different departments but in particular Dr. Geoffrey Dimbleby, Dr. John Cole, Dr. John Evans, Dr. Ian Cornwall and the late Dr. John Waechter. Much thanks to Dr. Martin Aitken then of the Donald Baden-Powell Quaternary Research Centre at Oxford University for the archaeomagnetic dating and other practical help; Dr. John Evans, University of Wales, Prof R.R. Newell of the Biologisch-Archaeologisch Institute, University of Groningen; Dr. Klaus Grote, University of Géttingen, Dr. Derek Mottershead of INQUA for producing the soil column, Drs D.Q. Bowen and C. Kidson of the Department of Geography, University College of Wales, Aberystwyth, for work on amino-acid dating. Thanks to the British Museum Laboratory for the first two carbon-dates supplied shortly after work on the site commenced. Thanks is due to the British Museum for allowing me facilities for in their Quaternary Department. I have been greatly encouraged by Department and I am delighted that they have agreed to accept exavations. I am sure they will be in good hands there and that benefit from their support. Vil

many years to research Mesolithic period assemblages the interest of Jill Cooke and the staff of the Quaternary the vast quantity of artefacts and archives from the future generations of archaeologists and students will

Since the late 1980’s Dr. Kenneth Thomas of the Institute of Archaeology, London, gave his constant support to the work which benefitted immensely from his greater knowledge of environmental issues; he instigated and supported several undergraduate and post-graduate studies on environmentally related subjects and examined many enigmatic finds from the site and commented on their possible nature and significance. To him also gratitude is owed for obtaining further carbon dates and thus benefitting the whole research. Above all, his contribution on the midden content and environment of the settlement site in this volume adds a most important extra dimension to the interpretation of the site and, indeed, to all future

research on Mesolithic settlement. Thanks also to all the people who contributed to the research and who are mentioned in Thomas’ Report. Thanks is due to the Natural Environment Research Council Radiocarbon Laboratory, East Kilbride, who through a Radiocarbon Dating Allocation to Dr. Thomas supported the work and provided much needed further dates. They are also supporting valuable work on stable isotope analysis through their Isotope Geosciences Labaratory.

Although Myfanwy Stewart only joined the work in the later years of the research, a great deal of thanks is due for her help in the field. Her contribution to this volume greatly enhances its value as a study of the Mesolithic way of life and reflects a special aspect of her personal interest in the site. I am also indebted to her for reading this manuscript in its draft form and extensively commenting on it. As is usual, the writer must, however, concede full responsibility for any omissions or remaining errors and for any controversial ideas, which may not necessarily also reflect the opinion of Myfanwy or of the other contributors to this Report. Many thanks to Dr. John Gale and the School of Conservation Sciences, University of Bournemouth for carrying out the geophysical survey. Thanks is due to H. Keeley and B. Collinson of English Heritage for their interest in samples from the site in the early days of the research and drawing attention to the potential of environmental studies of the site. Thanks to Dr. Christopher Buckland-Wright of University College for his interest in the bone fragments which came from the site during the early years of the research. At the invitation of the Director the site was visited in 1977 and the excavation records examined by a group of eminent archaeologists consisting of Drs. Paul Mellars, Roger Jacobi, John Wymer and the late Ray A.H. Farrar. Much gratitude to them all for their warm interest and subsequent support. Mr. Farrar was always so deeply interested in the archaeology of Dorset and in the writer's work in the county, that sadness is felt at the fact that he has not lived to see the culmination of

many years of work at Culverwell. A great debt of gratitude is also owing to The help of Denene Reilly in preparing survey work on site was done by Chris Thanks to Moira for her work on Trench

the following: Bambi Stainton the section and plan drawings and Moira Woods , who also 40 and permission to use some

for the superb drawings of artefacts in this Report. for publication is greatly appreciated. Much of the did some of the draft section and plan drawings. of her drawings.

Many thanks is due to Martin Blundell for many hours of help in getting this excavation report onto prepared for press. Thanks also to Keith Derwent for much help with the mysteries of computer work.

the computer and

Thanks are due for a measure of financial support: In 1973 a small grant was made available through the D.N.H.& A.S. from the Carnegie United Kingdom Trust for the purchase of a wooden site hut, now long ago destroyed by storms. Since 1989 some small contributions towards on-site expenses have been given by the Association for Portland Archaeology. Throughout the years all other financial requirements for the excavation work were met by the Director personally, unfortunately sometimes resulting in the need to economise as much as possible.

The conservation of the site and its management for the future is still an ongoing project, as discussed in a final chapter in this Report. A total grant of £15,000 has enabled a protective structure to be erected over part of the site and making it accessible to the public; the grant came from the Hanson Environment Fund under the Landfill Tax Scheme as administered by the Royal Society for Nature Conservation with the 10% contribution of £1,500 required under the scheme, coming from the EU Funded Portland Konver Project, administered by Weymouth and Portland Borough Council.

Vill

In implementing the conservation project for the site, much help has come from members of A.P.A. , particularly Rod Wild for doing all the monetary work and calculations; moral support and help in many ways came from the Planning Department, Weymouth and Portland Borough Council, the Portland Town Council, and English Heritage. The project is a registered and approved ENTRUST project (No. 325083.001). Thanks to them all. Thanks to local Councillors and other friends who helped in the struggle against a lot of odds to get the site preserved as a Heritage project. Finally , the greatest debt of gratitude is owed to my family, particularly my children Jesse, Marie, Samuel and Roy and my sister Salomina Fourie who for so long frequently and patiently had to put up with lack of attention, inconvenience and even financial and other deprivation so that the work could benefit. They gave moral support and practical help often and wherever possible and without them the project is unlikely ever to have got much beyond the starting post. Susann Palmer Director of Excavations and Research Association for Portland Archaeology Portland, 1999.

1X

PART 1.

EXCAVATION REPORT AND ARCHAEOLOGICAL ASSESSMENTS

1.1. INTRODUCTION,

SITE

LOCATION

AND HISTORY OF RESEARCH

The Mesolithic site of Culverwell is situated on the rocky prominence of the Isle of Portland in the county of Dorset, southern England at 50° 31’ 27" N. Lat., 02° 26’ W. Long. NGR SY 685694 on fields 2154 and 2153 (Fig. 1) .The site is at 31.05m (101.85 ft) above Ordnance Datum Newlyn, measured from the trigonometric point at the Old Lower Lighthouse at the Bill to the site datum post on the north-east corner of Trench 1 (Figs. 2 and 3). The reading was taken to the O.L.L. as the Benchmark, indicated on Ordnance Survey maps at the southern end of Field 2154, could

not be located at the time.

The site of Culverwell lies towards the bottom of the slope of a south-facing hill, about 274m from the present-day cliff tops, with wide open views to the sea (Plates 1 and 2). The hill on which the site is situated and the adjacent slopes are extensively covered with medieval field systems and lynchets, under which the Mesolithic site is buried. The Portland Race, where currents from the Channel meets the open sea, is less than a mile off the present Portland Bill coast and the area is notorious for its rough seas. Although exposed to the winds, the hill also gets the maximum sun in the summer months. The site datum post is approximately 270m east from the spring of Culver Well after which the site takes its name (Plate 4). For ease of reference it has been decided to refer to the Mesolithic site name as Culverwell, one word, and the name of

the spring as Culver Well, two words. Culver Well spring is approximately 26.38m (86.52ft) above Ordnance Datum, a measurement taken from the top of a rock one metre above the average present waterlevel of the spring, i.e. there is a slope of approximately 4.68m (15.35ft) over a distance of 172m (564ft) from the site down to the spring. The spring is still active today and a small stream runs from it to the cliffs where it falls into the sea. Thick gravel deposits and some deposits of tufa round the spring suggests that the spring flowed stronger in the past than today and had a much wider bed (Palmer 1970, 89). Riverine deposits are found in quite a large depression around the present-day spring and there are many bushes round the place where the water rises to the surface. Within 0.8km both east and west of the site there are other springs. To the east, by the village of Southwell, there used to be a waterfall until only a few years ago when the spring was blocked during quarrying and building operations. To the west of Culver Well, on the slope north of the Old Lower Lighthouse, there is a spring which runs underground at present but surfaces on the cliffs. This spring can clearly be heard running just under the surface and it was almost certainly a surface stream in the past. This stream is very near another Mesolithic site which was excavated by the present author (Mesolithic Site 1: Palmer 1969).

PREVIOUS RESEARCH IN THE AREA Portland has attracted attention

ever

since the last century.

Prestwich (1875) and Damon (1884) wrote largely accurate accounts of the geology of the Island and drew attention to the Pleistocene deposits near the Verne in the highest northern part of the Island; this was claimed to have produced many cartloads full of Pleistocene mammal bones and some late Palaeolithic tools (see also Palmer 1968: 117-9). In 1951 skeletal remains and artefacts were found during quarrying in a fissure in Bumper’s

Lane

Quarry,

finds was not clear and some of the been Mesolithic but it is possible remains of different periods (Palmer virtually every other period of the from Portland (Palmer 1998).

Easton but the date of the

material may, in fact, have that the fissure contained 1977: 148-9). Remains of past have been recorded

HISTORY OF DISCOVERY OF THE SITE During the early 1960’s the Weymouth and Portland Mesolithic artefacts were Fleet, the majority of them 1964). A trial excavation

author of this report first visited the area. Large quantities of typical found along the foreshore of the made from Portland chert (Palmer in one area of concentration at

Butterstreet Cove, west Fleet, was undertaken in order to assess

the likelihood of finding a stratified Mesolithic site on the foreshore of the Fleet. The results showed that it was unlikely that this evidence could be found due to tidal action and changing water levels (Palmer 1967). The

finding

Portland

chert

of such

large

quantities

did,

however,

direct

of artefacts attention

made

of

Isle

of

to the

Portland itself as it was reasonable to assume that the original natural source of the chert must be the cliff exposures on Portland. Extensive field walking on the Isle did indeed reveal that almost everywhere on the island stone age artefacts can be found;

the Mesolithic artefacts, mostly of Portland chert,

were

more concentrated near the cliffs in the southern part of the island . Over thirty areas of concentrations were isolated with numerous artefacts of a typical Mesolithic aspect but pointed picks were frequently the most numerous tool on these sites (Palmer 1977). A small number of concentrations with mostly artefacts of a Neolithic nature were found in the central areas of the Isle, away from the cliffs.

The Neolithic sites had

a higher percentage of artefacts made of material other than Portland chert (Palmer 1966). As

a result of coming

across

a previously unknown

small

collection of artefacts in the British Museum, deposited there at the end of the last century, it was also realised that a site of

Upper Palaeolithic type had previously existed in the Verne area in the north of the island (Palmer 1968). During 1967 and 1968 Palmer directed the area of very high tool density immediately Lower Lighthouse (Site 1 and Site 0), where revealed a shallow shell midden and other

excavation of an east of the Old several trenches features (Palmer

1967,1968,1969,1971 Report).

and

1977

and

see

Plates

58-61

this

While working on the site next to the O.L.L. in 1967 Palmer noticed that deep ploughing was taking place on the field registered as Field 2154 (now known as part of Culverwell site), approx .54km (about a third of a mile) further east towards the village of Southwell. During previous field walking on that field, very little evidence of Mesolithic activity had been noticed, but now, due to the deeper ploughing, it was very evident that a site of importance could be located here; a thick spread of shells and other occupation debris had been brought up by the plough. Initially the site belonged to a local farmer (R.D. Stone) who could only grant permission for excavation for a limited number of years as the County Council was planning to execute a compulsory purchase order on part of the site in order to make a pavement next to the Bill Road through the southern part of the site. In an effort to save the site, the present author had the site scheduled as an Ancient Monument and the site was purchased by a well-known public figure whose establishment was financially in a position to oppose the Council plans. Without this help, an important part of the site would have been destroyed. The road was moved a metre to the south and the pavement made next to it without encroaching onto the site. In the 1980’s it was again necessary to do battle for the site as the Wessex Water Board intended to put a main sewer pipe through the site, but this disaster was also averted.

In 1989 Field 2154 came into the ownership of the present author on the condition that it must never be used for anything else but for the benefit of archaeological research and education in general. A Plan for managing the site in the future is given at the end of this Report and is (at the time of writing) in the process of being implemented.

1.2. INVESTIGATION TECHNIQUES All work relating to the Culverwell site was carried out by volunteers from local areas and from all over England and the Continent. The work was at all times directed by a professional archaeologist, and many of the volunteers were qualified and highly experienced archaeologists. Professional and specialist collegues from various outside establishments regularly helped, also on a voluntary basis. It was only possible to work for one or two weeks per year, occasionally three, hence the many years (30) neccessarily spent on the work. Objectives of On-site work

A cursory examination of the terrain in 1967 soon after its discovery, had suggested that the site was potentially very important for this period and also well preserved, and a decision

was,

therefore,

taken

to

devote

as

much

time

in the area where the maximum quantity of shell debris was uncovered by the plough. This was numbered Trench 1. The trench was laid out at the base of the negative lynchet so that one side was exactly parallel to the apex of the lynchet between Fields 2153 and 2154 and the opposing side parallel to the road; the other two sides were at 90° to the lynchet (see Fig, 2). Within two days of beginning work, it was realised that a large and thick midden deposit was present and also a structure of horizontally laid down slabs of Portland limestone. The artefacts were distinctively Mesolithic. All work on the site was suspended,

was temporarily backfilled, pending the organisation of a largescale research project.

as

possible to researching the Culverwell site. At that time, very few Mesolithic sites in England had been extensively investigated and particulary habitation sites were rare. As explained in the introduction to this Report, the opportunity for research over a long period at first appeared as though it would be limited here too but when the ownership of the field changed, it became possible to set longer-term objectives for the research.

RESEARCH

At the beginning of work on the site, the following objectives were set:

2.

To confirm the dating of the habitation evidence, main features and finds and to see these in the context of any earlier and later features and/or fnds which may be present on the site. To gain a reasonably extensive insight into the lay-out of the site and to gain an understanding of the intra-site structure and the various features within it.

3.

If possible, to gain an understanding of the general life- style and social structure of _ the site inhabitants and the economic basis of that life-style.

4.

To understand the nature of the site and all its aspects in the context of its local environment.

1968 - 1997

Datum Post

Before work recommenced in August 1968 a square limestone column was erected at the base of the negative lynchet in Field 2154 at a point 45 cms (18 in.) east of the northeastern corner of Trench 1 opened the previous year. The magnetic points of the compass,

1.

the excavated area was marked and the trench

as at the time, was

post was to serve as the Datum Point for the laying out of all trenches on the site and all subsequent recordings and horizontal and vertcal measurements were taken from it.

TRENCH ORIENTATION AND LAY-OUT The medieval lynchets on the site are virtually parallel to the Portland Bill Road, which follows an approximately northnorth-west to south-south-west ( magnetic compass) line at this point.

or convenience sake, it was decided to lay all the

trenches out in a grid system in relationship to these two parallel lines (Fig. 2). They were, for ease of reference, in daily records and Plans regarded as the northerly and southerly reference lines, so that each trench has sides parallel to both the lynchet and the road to the Bill and with the other two sides of each trench therefore at an angle of 90° to the lynchet and the road. In terms of magnetic N/S reference points, the sections were, however,

5.

To acquire a large, representative collection of industrial material from all areas of the site which will (with further off-site research) allow adequare scope for studies placing the Culverwell habitation site in the wider Méesolithic context of England and Europe.

6. To gain an understanding of the methodology involved in the excavation and research of the complexities of a midden site for the benefit of future research. Exploratory excavation 1967 As there were initially no indications as to the lay-out and extent of the site, a trench 3.6m by 3.6m (12 ft sq) was opened

chiselled on this post. This

south-east these.

along

a north-north-west

and south-

alignment, with the other two at right angles to

Both Map North and Magnetic North are indicated on all plans and maps etc in this publication as also the lines along which section drawings were done. In 1968 three trenches were opened with their measurements taken from the Datum Post: Trench 1 covered the excavations of the previous year and the Datum Post was at its north-eastern corner. Trench 2 had its north-west corner 3.03m (10 ft) south-east from the Datum Post , and Trench 3’s north-eastern corner was 3.03m (10 ft) west from the D.P. (Fig.2). All following trenches were subsequently laid out in alignment to these trenches.

Most trenches measured either 3.6m by 3.6m (12ft sq) or 3.05m by 3.05m (10ft sq). Several were subdivided into halves with each half then quartered in order to facilitate excavation and recording. Note:

When

work on the site commenced,

all measurements

were still taken in feet and inches and as Britain did not change to the Metric system until some years later when the excavation had already made considerable progress, it was decided to continue using Imperial measurements on the site and on all plans, section drawings, etc in order to avoid the possible confusion of two systems employed on different occasions. In this report measurements will be given in both systems where regarded appropriate (particularly for reasons of precision), but it must be kept in mind that it is not practical always to precisely translate fractions of measurement from one system to the other. Baulks between Trenches:

The majority of Baulks between trenches were originally 60cms (2 ft) wide but most became somewhat eroded by weather over the years and it was eventually found expedient to remove some of them in order to pursue particular features or to expose them more fully. This

form

meant

that

larger units

several trenches

were,

of excavation

which

in fact, combined

were

to

subsequently

referred to as Areas A,B,C and D; Area B included all trenches in Field 2153 and Areas A,C and D covered all excavated

trenches on Field 2154 which comprised the main area of the site and excavations (See Fig. 2). Trenches not linked up to these

four

larger

excavated

areas,

retained

their

individual

numbers. When it is necessary in this report to refer to a feature or find in relationship to a more precise locality, reference will continue to be made to a specific trench as being a smaller more precise unit of excavation. Sondages

In several places use was made of a smaller unit of excavation which was referred to as a Sondages’ where it was necessary to explore a particular feature or aspect of a feature without disturbing the adjacent features too much. For instance, in Area D small cuttings (Son. 10 and Son. 10A) were made through Feature 2a by removing only a few stones from that feature and leaving the rest intact for the time being. This was done in order to clarify the stratigraphy under the feature and to establish its relationship with other features in the area prior to excavating a larger trench. Many of these small excavation units were eventually extended to become larger trenches or to include a whole trench.

the shelly deposits and the presence of many stones, the depth of the Spits can only be given as an approximation of 5cms (2 in.). This technique was found particularly useful in excavating the thick midden and deposits in the pit in which changes in the layering were usually very subtle (see 1.4 Stratigraphy). On-site plotting of artefact finds In Trench 40 and Baulk 21 three dimensional recordings were made of every artefact found at the lower levels of the shell midden in order to determine whether it was likely that an intact habitation or working surface was present anywhere on the site under or within the large midden. For this purpose the trenches were sub-divided into smaller quadrats and all the artefacts at any one particular depth in the midden in these trenches were left in situ till they were all recorded and plotted on individual charts. In this way it was possible to ascertain that an early work level probably did exist near the base of the midden in the area of Trench 40. The artefacts found, were also individually examined to try and join them to others and to the cores found;

a number

of conjoins

were

obtained.

In all

other trenches three-dimensional recording was only done for finds and features regarded as being of unusual, or potentially unusual significance. Ideally these techniques of recording finds should have followed all over the site, as it may perhaps have enabled differentiation to be made between areas which were used for the dumping of refuse and those areas which, at some or

another,

were

also

used

as

work

zones.

However,

been some only stage with

regards to the haphazard manner in which a large shell-midden usually develops (see 1.4 this Report) and due to the fact that this work

was

most

time

consuming,

it was

decided

that it

would not be feasible to follow these techniques everywhere, especially not with limited time and resources. Trenches 40 and Baulk 21 must serve as examples of the type of research techniques which are possible on shell middens, even though the outcome may not always be rewarding. Recording of Information and Finds

Each evening the work in each individual trench was recorded for that day in a Site Book, with all the volunteers participating in the discussions about the day’s work so that everybody knew at all times exactly what was going on. _ The bags of finds from each trench and each layer were then given a preliminary sorting and significant finds noted and separately recorded in Small-finds Books. Whenever possible, the day’s finds were washed and marked every day on site., but due to inclement

sometimes staying.

weather

carried

conditions

and time factors, this task was

out at the home

where

volunteers

were

Excavation of ’Spits’ within Layers After completion of the season’s work on site, bags were re-

In every trench, as soon as the archaeological layers were reached, the midden and other deposits were excavated at intervals of ‘Spits’ so that any subtle changes in the layers would more readily be noticed and the recording of changes and finds could be more precise. Due to the coarse nature of

checked indoors during regular sessions throughout the following months (and years); knapping debris was separated from pieces which showed signs of deliberate retouch and both these elements were then weighed and recorded. Due to the vast quantity of material found, this immense task is still

continuing at the time of publishing this Report, and a further full-scale assessment of the artefacts (and their distribution) from the site will therefore be published separately at a later date. All the material will eventually be placed in the British Museum with the archives. Site plans and section drawings were completed each year and where necessary, older ones were updated or completed to include the most recent findings. A summary of the year’s findings were written each year and notes made of any particular problems or aspects which required further research during future work. Initially these summaries were published each year but it was decided that this could lead to confusion and misinterpretation as it frequently became necessary to modify preliminary conclusions.

Marking of Artefacts All finds (size permitting) were marked during the dig in accordance with the international convention: 1. Initials of Area name (PB); ii. Site name (CW); iii. Trench number (T) or Baulk number (Blk) or Sondage number (Son.); iv Layer number (L) - on larger artefacts further details, e.g. 'midden’' or ‘clay' etc was written on a few in each bag; v. Spit number (S); vi. date (e.g. 8.8.80). Finds from each trench and each layer or Spit within the layer were bagged separately for each day. Significant

finds

were

individually

marked,

numbered

and

recorded in a separate 'Small-finds Book' as soon as they were found and identified, giving details of where they were found, layers, spit, date etc and a short description of the object. They are all artefacts of some significance or specific category that will be discussed in the report and of which some are illustrated. Finds found together in or near any specific feature were numbered and bagged together, e.g. the pile of picks from Hearth 4.

Research Programmes

Throughout the years a number of students projects based on the site and the originals the Institute of Archaeology, London. carried out by students from universities and Morocco.

carried out research can be consulted at Projects were also in Italy, Lithuania

The environmental projects included work on the large shell midden on the site; the freshwater molluscs from the different

clay deposits; shellfish ecology and palaeo-coastlines in relation to the Culverwell site and a project comparing the present vegetation of the Bill area with likely past conditions. Archaeological studies included a project on the spatial analysis of artefacts on part of the site, in Trench 40 and another project related to the presence of large quantities of unstruck pebbles and burnt stones on the site (see M. Stewart, this Report). The latter project is incorporated as a contribution to this Report. Environmental research on the soil deposits on the site was initially carried out by Dr. Helen Keeley of the Ancient Monuments Laboratory and work on the molluscs from the soil samples by Beverley Collinson of the same establishment. This study is cited in the text of the present Report and appended in full as an Appendx. A small quantity of bone from the site was initially examined by C. Buckland- Wright of University College, London but a later study of them was made by Thomas ( this Report 2.10).

Since the late 1980's Dr. Kenneth Thomas of the Environmental Department at the Institute of Archaeology in London took samples of the soils and midden deposits of the site and these were analysed by him or under his supervision. He also examined many of the more enigmatic finds from the site and commented on their possible nature and significance.

Shell Count

Geophysical urvey

At the beginning of work on the midden the average weights and contents of 12 big marked buckets were calculated and

A geophysical survey of the site was carried out by John D. Gale and his team from the School of Conservation Sciences at Bournemouth University. This survey was carried out in 1997 when the excavation work on the site was nearing completion, primarily with the objective of determining the extent and layout of the gully which traverses the site. For various reasons it had not previously been possible to obtain the help of a survey

recorded, as follows:

average weight of midden refuse per bucket: 8.165kg (18 lbs); average weight of washed shells per bucket: 3.6kg (8lbs). The other 4.5kgs (10 lbs) consisted of earth, burnt stones and artefacts.

Subsequently, whenever excavation of the midden took place, a daily count (using marked buckets) was kept of the quantities of midden removed from the excavated trenches.

team;

the Director of the excavations had also been

advised

that it is likely that a geophysical survey of the site terrain may not be satisfactory. The main likely problem was the shallow soils on top of the solid Portland bedrock. As anticipated, the outcome was not fully satisfactory but was adequate to indicate that most areas of the habitation site had probably been explored in varying degrees.

It is realised that this is a very rough method of calculating the contents of the midden, but in combination

with calculations

based on site size and supported by other findings, this approximation can, with moderate validity, be used in discussing the scale of the site's occupation (see further discussions on the midden and population estimates Part 3 this Report).

The survey indicated an anomaly (possibly another gully or ditch) in Field 2154, approximately 12m _ south-east from the excavated Trench 41 (Fig. 2), i.e. possibly roughly parallel with the known gully on the excavated site (see further discussion, this Report 1.5 ‘Features’). An attempt was made to arrange

excavation

of

this

feature,

but

various

conditions

presented difficulties and it was decided to leave this for the future. Also, it was decided that the preservation work of the large site had to take priority for the time being (Management Plan, 4.20 this Report). Specialist Interest

A large number of professional collegues visited the site and gave help and advice in diverse ways. It has also been visited by groups from various national and international academic societies. An Interim Report in the Proceedings of the Prehistoric Society (Palmer 1976) drew attention to the many features of the site, as researched up to that time. At the invitation of the Director, the site was visited in 1977

and the excavation records examined by a group of eminent archaeologists consisting of Paul Mellars, Roger Jacobi, John Wymer and Ray A.H. Farrar (d. 1993). Section Drawings, Plan drawings, photographs of Features and other Archival Material

A very large number of plans, sections and features were drawn on the site. Some of them were redrawn and updated several times as research progressed and as new details were noticed or re-assessed. An equally large number of photographs and slides were also taken during every season of work. It is quite impossible, and also unnecessary, to include them all in this publication. A selection of the figures and plates showing the most significant, representative or unusual aspects of the site are included in this Report and all others will be placed with the rest of the archival material, such as Site Books

and Finds

Books,

in the British Museum

or National

Archaeological archives after all aspects of the research has been completed.

1.3. SITE

STRATIGRAPHY It is possible that a combination of these two situations, i.e. deliberate lynchet building and fortuitous lynchet development, may sometimes have been used in some areas of England and Europe and the techniques may even have been interchangeable in the same area. As a section only was excavated through a small part of a lynchet on Portland , it was not possible to come

Introductory Remarks

In order to understand the soil profile of the site, it is necessary to realise that the ancient habitation site was established on a well-drained south-facing (seawards) slope (now covered with a series of stepped terraces or medieval lynchets). Most of the site is also on a gentle westward slope in the direction of the

to any conclusions, in general, about the lynchet formation on

Culver Well stream, i.e. there is a two-directional slope in the

the island.

terrain involving the habitation site. By the period of the Mesolithic occupation (or possibly even earlier), there was at least one irregularly shaped gully running down the slope, approximately in a north-eastern to south-western direction. (It is possible that there may have been other gullies or depressions immedately outside the area explored under the present project — see chapters 1.4. and 1.6. ). Because of the exposed position

Immediately outside the down-slope of each lynchet i.e. negative lynchet, the deposits above the ancient remains are thinnest and sometimes incompletely developed as the plough would have spread the soil from here away to the next positive lynchet (step) downward.

on a hillside facing the sea, the site would

subjected agencies

to severe and,

even

denudation anciently,

and soil

always

soil

erosion

would

have

The disturbance of the ancient soils on the hill was therefore multiple: a. continual soil movements during the Mesolithic

have been

by

natural

formed

occupation,

but

slowly except in the hollows. The topsoil everywhere on the island is of a highly calcareous brownearth type. Medieval lynchets and

soil formation

soil pressure,

c.

accelerated

and

intensive

soil

disturbance

due

to

agricultural activities since medieval times.

It is appropriate here to give an explanation about the formation of lynchets as it is necessary for an understanding of _ the stratigraphy of the site as it was encountered during excavation. Lynchets are largely the natural outcome of ploughing with a small aard or plough on the slope of a hill, diagonally to the slope, during medieval times (diagram below). The plough went back and forward, diagonally to the of slope and, as the furrows

e.g. water,

human activity and animals; b. post-depositional movement, i.e. a degree of soil disturbance in post-mesolithic prehistoric and Roman times by natural agencies, humans and animals walking over the land or taking part in other activities on the Island;

direction

due to natural agencies,

It is also important to note that much of the pre-lynchet topsoils would have been removed from their original positions in the area of the negative lynchets and redeposited and reformed lower down on the next positive lynchet as top-soils; new topsoils would then have started forming where the previous soils had been removed and could in, part at least, |§ incorporate some material from the much older layers. Soil formation on top of the solid limestone beds would have been very slow on slopes and the recuperation of the soil profile would therefore have also been very slow and probably only partial.

on the field are never

varied, soil automatically piled up on the edge of the field. Any big stones encountered during the ploughing would also have been piled up on the edge of the field, parallel to it. Some archaeologists argue that lynchets were at least sometimes deliberately formed to prevent the soil from eroding «as

une?

et.

or Pr Pen a, SN

-a™N

—*:

TON ~

eNOS ~~

-a

~

altaee

-~Se

NEGATIVE

downhill and to keep the plough and oxen ploughing. Soil spreading out from the plough, denudation on the higher part of the slope and lower down and so form a ledge above the next down the hill.

POSITIVE >

7~

he

-_—~ =—

wa

level during would cause then bank up parallel field

on

the

fields

downhill,

the result

is a series

stepped terraces following the original contours of the hill.

of

LYNCHET

This ‘built-up’ or accumulated level top edge of the system of the field is known as the ‘positive’ lynchet. The denuded area, sloping downwards down the hill, towards the next level stretch of field, is known as the ‘negative’ lynchet. When this process is repeated

LYNCHET

Survival of Mesolithic Layers and Features The important aspect, for the stratigraphy of

the Culverwell

Mesolithic

soils

site,

is the

fact that

the

various

would

be

thickest under the two positive lynchets, forming the northern boundaries of Fields 2153 and 2154 and here too it can be expected to find the most complete soil profile and the best preserved ancient remains.

In the eastern half of the site (excluding those trenches which are under the lynchet) the soils above the archaeological layers are very thin indeed so that the Mesolithic midden is sometimes

almost immediately under’ the _ present ploughsoil. Archaeological layers in the northern part of the site are likewise immediately under the recent soils and also immediately on top of the natural loam. This is due to the original contouring of the hill-slope and, in particular,

slow soil

A. Aproximately half-way up the hill the profile in the hole was as follows: 1. Topsoil - loamy brownearth, average 30 to 15cms (between 1 ft to 6 in.) dark greyish-brown.

formation in those areas away from the gully.

2. Pale yellowish-brown clay loam - average 7.5 to 15cms (3-6 in.).

Fortunately, the deeper layers containing the archaeological features in most areas at Culverwell have not been severely disturbed as shown by the fact that an ancient buried soil (Layer 4, see Figs. discussion below),

4, 15, 16, 17; Plates 3, 43, 44, 45) and is still present almost everywhere

immediately above the Mesolithic levels, even though very thinly so in some areas. Also, although the present topsoil on the site and the yellow loam immediately below the brownearth does contain some shells as well as small quantities of Mesolithic material, the finds from Layers 1 to 2 are mostly post-Mesolithic and from Layer 3 downwards, these later finds decrease significantly and Mesolithic material begin to predominate.

The presence of a large pre-Mesolithic natural gully on part of the site (see 1.4.°Features’ in site description), has ensured excellent protection of the archaeological remains in those areas too. There was obviously more scope for the intact survival of the deposits in the gully and the shell midden is here thickest. In Trench

40

3. Stiff clay loam - mottled brown-yellow, average 60cms (1ft thick).

(see Site description

and

Figs.

23,

24;

4.

Bedrock - pale yellow rubbly, friable sandy-clay loam with a darker band about 2.5cms (lin ) thick near the top.

B. At the top of the hill, above the topmost lynchet: 1. Topsoil

- brownloam, only about

30cms ( 1ft) thick.

2. Brown clay loam with pieces of slatty limestone, about 15cms (6in.) thick. 3.

Stiff yellow-brown clay loam, about 30cms (1ft) thick.

4. Pale yellow rubbly clay for about 23cms (9in.) and then less rubbly stiff yellow-brown clay.

Plates

50,56) an intact working surface has been researched, producing conjoined flakes. It has also been possible to refit flakes from the lower layers of Trench 4 and several other trenches, including a core with two conjoining flakes, illustrating that post- Mesolithic changes have not severely affected the deeper layers of the Mesolithic site (Fig. 38).

Note: The 2nd layer in section B was probably modified by agricultural activities at the top of the hill, i.e. above a lynchet.

In the southernmost part of the site, south of Feature 2 (see Site

SOIL PROFILE OF THE ANCIENT HABITATION SITE

description and Fig. 4), the deposits above the archaeological remains are also thick and it appears that there was another gully or a minor hollow here. A faint linear mark in the field, towards the fence and parallel to the line of the fence and the other lynchets, suggests that it is possible that there was originally another positive lynchet here, followed by the negative lynchet under what is now the road which marks the southern boundary of Field 2154. This would mean that what is now Field 2154 would originally have been two fields. If so, this other lynchet must have been destroyed fairly early in postmedieval times, possibly to provide space for a cart track to the Bill, as it is not indicated

on the earliest Portland

C.

On field 2154,

west of the excavated archaeological site:

Same as for A.

General concepts regarding the Site Layers Decoding the stratigraphy of a site containing mainly shellbearing deposits is very complex, as discussed at length by various workers in this field (e.g.Stein 1992). As pointed out by that author, there are different interpretations possible for the concept of Layers and this can cause confusion in discussing a site such as Culverwell. Stein's system of nomenclature is relevant to the situation on the Culverwell site only in a limited sense, but it can be usefully summarised as follows:

Tithe

maps. Natural stratigraphy of the area outside the site Several small sondages were excavated in field 2154 in 1968, to

the west of the occupation area and well away from the known limits of the site, in order to determine the natural stratigraphy of the terrain without the archaeological deposits. In 1975 a series of holes were dug, averaging a depth of about one and a half metre each, at regular intervals up the slope of the hill in order to erect sturdy fencing for cattle. This afforded an opportunity of studying the natural stratification of the hill-side:

a. Layers which have implications of chronology and which permits division into different periods are referred to as Chronostratigraphic units. b. Layers which are divisible from others on the basis of their characteristic contents of artefactual material (Ethnostratigraphic units).

c. Layers which are recognisable mainly biological fossil content, e.g. human/ (Biostratigraphic units).

according to the animal remains

d. Layers which are recognisable on the basis of their observable changes regarding the colour, texture, compactness or nature of the matrix of the midden material ( brown/yellow, plastic/gritty clay/sand/loam etc - Lithostratigraphic units).

disturbance but contents still the same as above.

Layer 1 is thick in Keeley Section 1 because it is here that the medieval ploughsoil banked up to form a positive lynchet, as explained above.

The Culverwell Stratigraphic Units The Culverwell stratigraphic profiles contain elements of these types of units which

were used, to some

all

extent in each

case, in the recognition of strata or ‘layers’. The Mesolithic occupation layers were recognised largely by their high content of molluscs and the artefactual or archaeological material but within these layers referring to the habitation period, were subdivisons discernible mainly on the basis of colour, texture or compactness or matrix. The post-occupation — strata were recognisable to a lesser extent by the content or, rather, the limited presence of Mesolithic living debris but also by the nature of the soil deposits. The total record of these different strata enabled some chronological conclusions to be reached. A study of a specimen profile of the soils above the midden of the Culverwell site was carried out in 1979 Helen Keeley of the Ancient Monuments Laboratory (then) Department of the Environment. Her report, in appended to this Report. Her samples were all taken north face of Area A, in Trenches 3,1 and 4,

shellby Dr of the full, is in the

in the area of the

lynchet between field 2153 and 2154 (see Figs. 2 and 3), i.e. in an area where the soil profile has been protected from disturbance.

A soil column, taken at the point of Keeley's Section 1, was made by Dr. Paul Farrel of Portsmouth Polytechnic, Geology Section, by arrangement with Dr. Derek Mottershead, in 1979

to enable permanent demonstration of the site's stratigraphy. It is possible to bracket the layers which are noticeable on the site together into four distinct Zones, with some chronostratigraphic implications for each. It must, however, be stressed that the different layers within the zones are not everywhere present and it is not implied that there was a homogenous development of all layers all over the site:

A.

RECENT LAYERS AND POST-LYNCHETS

Layers 1 and 2 - preponderance of medieval artefacts and post-prehistoric in origin.

to present-day

Layer One includes Keeley Samples I, II and III, Section I: 0 to 8 cms very dark-grey brown clay loam, present surface soil, contains much debris dating from late medieval periods up to the present e.g. coal (from harvesting machines), pottery, glass, metal and china of all periods and also some disturbed ancient material. 8 to 18 cms slightly lighter coloured clay loam, represents the plough soil over the centuries; contents same as above. 18 to 32 cms is brown clay loam, moderate medium blocky, too deep for the plough and slightly less signs of very recent

In some areas of the site, e.g. Trenches

1, 2,

5, 6 below this lynchet, the layer is thinner and averages about 18 to 30 cms at maximum. The distinction between Keeley's levels (samples) I, If and III is also not clear in areas outside the lynchet. As the artefactual content of all these samples are the same and are predominantly recent, they have been regarded as belonging to one recent Layer for site record purposes. The layer has everywhere been subjected to disturbances by farming activities, ploughing, burrowing and grazing animals and natural elements Layer 2 - Keeley Sample IV 32 to 43 cms a yellow-brown clay loam, a quite distinctive thin layer over much of the site although hardly distinguishable in the trenches immediately below and outside the down-slope of the lynchet where it has probably been largely removed by ploughing or has become incorporated into the shallow plough soil above it. This layer still contains medieval material and also Mesolithic artefacts from lower disturbed levels. This layer represents the present B-horizon from which material was washed and re-deposited lower down.

B. DEPOSITS OF PRE-LYNCHET

PERIODS

Layers 3 — 5 are everywhere more recent than the Mesolithic occupation of the site (i.e. date from later prehistoric and early historic times) but also pre-date the formation of the lynchets in medieval times.

Layer 3 - Keeley Sample V 43 to 57 cms is a brown to dark brown firm clay loam containing many flecks of diminutive shell fragments, some ancient artefacts but also a small quantity of recent material, derived from the soils above it. Keeley refers to this layer as the top of the pre-lynchet soil. It is absent, or only very thinly represented, in all the trenches where it was not protected from erosion, as explained above. In some areas of the site it appears to have become incorporated into the next layer down as a result of disturbances, anciently and more recently. Also, in some areas, e.g. in the western part of Area C, this layer sometimes incorporates a very thin lens, rather sparsely distributed, of very tiny pebbles at or near its base, giving almost an effect of "pebble dash";

this effect could be due to

worm action from the next layer down. Some of these stones could also have been anciently derived by natural agencies, such as hillwash and rainwater, from the protruding top of the Mesolithic occupation midden slightly higher up the slope. Layer 4 - Keeley Sample VI At 57 to 67 cms is a very distinctive layer of dark yellowishbrown, strongly prismatic firm clay/clay-loam with minute fragments of shell but very few stones. This layer is an ancient soil which became modified and subsequently buried, by a

process of diagenesis, and so sealed the archaeological remains below it. A soil of this type is sometimes referred to as a palaeosol, although some pedologists nowdays prefer not to use this terminology; it is also described as a Yelict soil’ and its main interest is that it takes a very long period of exposure to the elements to form and is therefore definitely early prehistoric in origin (Cornwall 1958: 52-55).

C. MESOLITHIC OCCUPATION PHASE: (as all recognised in Trench 4):

These layers all represent the shell midden and did not form part of Keeley’s detailed soil analysis, although noted by her. (But see Collinson in Keeley’s report, appendix) The mollusc content of the midden is further discussed in Thomas (this Report).

Deposits of this nature can frequently be relatively stone free, although the stoneless condition could also be partly due to the activity anciently of worms: they burrow by swallowing the earth in front of them and then go back up to evacuate their castings. These eventually sink down into the soil with the larger particles in the soil, leaving a zone relatively free of stones.

The sub-division into all these layers are very complex and not present nor very distinctive in all trenches. It was specifically first made in Trench 4, where the deposits of midden are very thick (Plate 43). This was necessary in order to: a. facilitate the detection of changes in the soil deposits and any ’gaps’ or changes in the deposits of artefactual or organic material, if any; b. to give some indications, in general terms of the development of the midden at least in one small part of the site; c. to facilitate the recording of any changes in as meaningful a way as possible and d. to enable some general comparisons with other parts of the site.

This layer is most clearly visible in the northern face of Area A under the lynchet and in the trenches which fall within the gully on the western side of the site where the depth was greater and disturbance

less

(Fig.

4,

15

and

16; Plate 44).

This

layer

is

present all over the site, although sometimes very thin and sometimes it shows minor signs of disturbance or modification by later prehistoric or historic activities.

The sub-division was based on very localised observance of changes in physical characteristics as regards colour, compactness of shells and debris and the nature of the matrix

This deposit could represent clayey loam which was originally washed down the slope onto the site not too long after the Mesolithic occupation, or perhaps even during it and so terminating the occupation. The Atlantic climate must have provided ample opportunities for the site to become covered by hillwash. It’s main significance is that it can be stated without hesitation that the Mesolithic occupation layers are well stratified below this layer and the next.

containing the midden material, as will be detailed below. It must be stressed that, although the midden zone as a whole has

a measure of chronostratigraphic significance (i.e. identifying the Mesolithic phase from earlier and later development phases), the different ‘layers’ which are discernible within the midden in some areas of the site, does not necessarily have any chronological significance, apart from all being broadly referable, in origin, to the Mesolithic occupation phase. The other dimension to this situation is that of localised postdepositional changes and non-anthropogenic phenomena affected to a midden which can change the shape, size and content of the midden, further blurring the concept of chronological significance.

Layer 5 - Keeley sample VII Below 67 cms from the surface of Keeley’s column is a yellowish-brown layer of clayey-loam, similar to that above but stonier and containing many snail and shell fragments. In the northern part of the site ( e.g. in Area A) it covers and is amongst the top of the large stones of the floor which here lies

It must be considered that during the Mesolithic occupation, small mounds

of rubbish was, almost certainly, left on the site

in a fairly random manner even though the main dumping may generally, or periodically, have taken place in a specific area. These small mounds would have overlapped in some places, some could have been left for a few weeks or longer and then again returned to, in which interim period there could have been a storm or children could have played on them, or some other

on the shell-midden, as will be discussed. In some areas (e.g. in Trench 11 and Trench 27, Area D see Figs.2, 14) it is above a

thin deposit of midden which there overlies the floor.

LAYERS 6- 11

In the

areas outside the floor, it is also immediately above the shell-

midden. In the south of the site it also covers the rough stones of a possible windbreak (see 1.4. and 1.5. description and discussion of Features) and some of this deposit there can be found amongst the stones of that feature on top of shells. It also varies greatly in thickness in different parts of the site and is barely visible in some trenches.

disturbance

could

have

taken

place,

altering

the

nature

or

appearance of the small heaps of rubbish. The effects of apparent ‘layering’ within the midden are, therefore, to some extent probably due to insignificant situations during occupation and _ post-depositional factors after occupation. Only when a change is consistent and everywhere visible, may it perhaps have significance, e.g. a change in diet or a break or change in occupation.

This Layer could possibly belong to a phase immediately after the Mesolithic site occupation; it was originally probably the basal part of Layer 4 and everywhere contains derived Mesolithic debris.

At Culverwell the layer numbering ‘Layers 6-11‘ were used in practice during work, as a matter of expediency in recording, after they were first identified in Trench 4. For instance, in most of the areas of shallow midden deposits (particularly in the eastern areas of the site) a thin layer of loose shells in a greyish

Keeley is of the opinion that both Layers 4 and 5 represent hillwash which covered the site as a result of a wet episode or period shortly after the occupation period or maybe even at the end of it.

10

matrix was encountered and was referred to as “Layer 6’, i.e. as most closely resembling Layer 6 of Trench 4 (see Fig. 4 and further discussion below). The apparent differences between Layers 6-11 could be part of the site formation processes of a midden site, e.g. compression due to the weight of the overburden at greater depths and also reflecting different climatic and environmental factors.

midden did not develop simultaneously or in the same manner. As it happens, there is a reasonable agreement between the different types of dates for Culverwell, taken from either organic and non-organic samples fromTrenches 3,4 and 41.( see 2.10 Dating

and

2.11

Thomas);

this

could

be indicative

or

suggestive of the fact that the Culverell midden has remained reasonably unaffected by various factors since its deposition. In real terms, it is likely that during the initial stage of settlement on the site, one or two dumping areas only were in use and during the subsequent time span of the occupation, the intervening and surrounding spaces between refuse heaps became filled with other heaps or detritus as a result of further dumping and the inevitable lateral spreading due to various activities on the site. It may perhaps be possible to argue that the main dumping area within the gully on the western side of the site (including Trench 4) was filled in a reasonably systematic manner over a period of time, but it would remain debateable what happened at the same time on the edges of this feature. It is, for instance, quite impossible even to suggest which areas of the midden were used as latrines. Vertically and

It is not, at all, suggested that these layers are everywhere present in the midden all over the site, nor is it suggested that these layers are everywhere continuous or that they developed in the same manner or at the same time all over the site. In fact, it can be categorically stated that the midden must have developed in a fairly haphazard manner, as can be demonstrated by, for instance, comparing different sections through the midden with each other (see all Section drawings). It is therefore not feasible to insist that an attempt should be made to precisely correlate these layers all over the site, except in general terms, i.e. ‘grey coloured midden’, ’scattered midden’, thin, loose midden deposit’, compact dark midden’, etc. The thicknesses of layers also vary all over the site from one point to the next, as demonstrated by the diagram below. Also, the implication is that there is caution required in the dating of different areas of very large middens, particularly in

horizontally

all

over

the

site,

the

layers

of

a

midden

is,

therefore, a mixture of numerous different events and possible situations.

rT IN iil litf7_ FOSS eat pt aie | ATE AANA

Z NSS — FN Np Pept e n n a ° a aS. The example in the graph above illustrates the irregular thickness of Layer 11 in diferent quadrates of Trench 4. The margins between layers within the midden at Culverwell are not distinctive enough to support arguments for seasonal breaks in the occupation of the site, although seasonal variations in the lifestyle and economy of the inhabitants is certain to have occurred whether or not they may have stayed on this site for most or all of the year (see further discussion in the report 3.14 below on this aspect).

regards to any relative short-term dating. The apparent layering within the midden at any one point, or at a different number of

points, cannot always without other corroborating data be used as necessarily an absolute indicator of the chronological indicator of the development of the midden (or site) as a whole, as itis possible that all parts of the whole of the large

11

Some of the visible differentiation of layer formation in Trench 4 could also have been caused by the degree of stoniness or compactness of the debris thrown into the gully or the predominance of a particular shell type or even on the degree of fragmentation of the shells. It should also be considered as feasible that at least some of the visible differences may be effects created by varying day to day or week by week activities; movement of shells by human or natural agencies after they were dumped or by weather conditions during the deposition period of the midden and (very important) by postoccupation factors - so many factors can create slight visual variations in arubbish dump without being of vital significance.

indicates that Layer 6 is the uppermost part of the midden deposit in this particular trench. It should, however, be kept in mind that colluviation almost certainly did affect the spread of the latest deposits of midden material to some extent in at least some parts of the site and some of Layer 6 may, therefore, at least sometimes represent earlier midden material which have been redistributed. The midden deposits which were dumped towards the end of the occupation period would have been unprotected and exposed to the elements for some time, depending on how soon afterwards the yellow-brown clay-loam (Layer 5) was deposited on top of the site.

In the section face of Trench 4 immediately adjacent the floor it can be seen that Layer 6 also represents the deposit of midden which here accumulated amongst the stones of the floor and on top of the floor; i.e. Layer 6 was in part at least deposited after this part of the floor was laid down on top of the peak of the lower midden deposits in the lower areas which had previously accumulated up to that stage in the site’s history (Fig. 4 and 16).

There is also the possibility that at least some of the differences may be illusory to some extent, created by differing degrees of dampness, during and after the site’s occupation, particularly within the lower-lying areas of the site. Stein (pg. 9-10 and various comments in following chapters) has pointed out the possible importance of post-depositional dampness at many of the sites excavated by her and her colleagues, and to this she attributed the fact that these middens could be divided into two basic layers: an upper light coloured midden deposit and a lower dark coloured deposit. The same broad division was, indeed, noticeable at Culverwell, except on the periphery of the site where the midden is very shallow and a sub-division of the midden would hardly be valid.

A midden deposit which compares in texture and colour to that of Layer 6 in Trench 4 is found all over the site on top of the floor

(e.g.

in

Trench

or even all, of these

layers are less in the nature of continuous layers over a large area and more in the nature of large or small ‘pocket’s or lenses which are not necessarily present everywhere on the site. Keeping all the above problems in mind, we are here giving the ‘layers’ of the midden as_ studied in the profiles of Trench 4 (Fig. 4). A detailed description of the midden contents is given in Thomas (this Report).

Layer 6 consists of exactly the same types of molluscs, artefact debris and other material as found in all layers and in all areas throughout the whole midden on the site but it is not nearly as compact as the lower layers. It is also much lighter in colour than the dark grey or black of the lower midden deposits in this trench. It has a matrix of brown or yellowish-brown clay Chart

reading

10YR4/3),

not

well

21;

Plate

46).

It is

also

is in

fact, characteristic

of virtually

all the

However, in giving the above description of this type of midden deposit and in outlining its distribution, it cannot be accepted unequivocably that it is the ‘same continuous Layer 6’ which is found all over the site, although in some areas it may well be so (as pointed out above in connection with stratigraphical problems relating to middens). It may have formed earlier or later in some areas of the site and some of its lateral distribution may, at least in part, be due to movements during and after the Mesolithic occupation of the site. Therefore, although midden deposits of this description was referred to as "Layer 6’

Layer 6:

(Munsell

Fig.

occupation debris on the eastern periphery of the site. In Area B in the northernmost part of the site a similar loose, light coloured midden deposit is thinly present on top of the floor, varying in thickness from about 4 to 6 cms but it becomes thicker towards the southern part of Areas A and C and in Area D it is about 7 to 9 cms thick on top of the floor (Fig. 4; Plate 43).

Frequently changes occurred in the middle of a trench which were not reflected in the nearest adjacent sections or trench It is therefore certain that some,

see

extensively at the top of midden deposits in the eastern part of the site, outside the floor; it becomes progressively thinner towards the outer limits of the occupation area (Fig. 5). This type of midden,

faces.

6,

wherever

structured.

it was

encountered

on

the

site,

it would

correct to interpret this as meaning

‘deposits

Layer

general

be

more

comparable

to

Visually, it has an overall light grey appearance, partly due to the high content of fragments of Portland and Purbeck chert, burnt and unburnt grey limestone fragments and charcoal but the colour could possibly also be due in part, at least, to changes due to exposure after the site was abandoned. The shells of different species of molluscs are fairly haphazardly distributed throughout the width and depth of this layer, although small clusters of a particular mollusc are quite common in any one place; often it is also noticeable that in some areas the shells of limpets (Patella sp) appears to have been stacked one inside each other.

pattern

Stratigraphical interpretation of the midden deposits in Trench 4

they occur. Its distribution can be taken to define, approximately, the limits of the settlement site (i.e. keeping in

6

of Trench

4’.

Even

so,

in

reasonable to assume that this type of material belongs to a late phase of history, even though there is nowhere development or chronology between deposit and the more compact midden Trench 4 (and elsewhere on the site).

terms

it seems

pale, loose, top midden the site’s occupational evidence for any break in the formation of this deposits lower down in

This layer contains vast quantities of all types of Mesolithic artefacts and its distribution coincides with the distribution

12

of these

Mesolithic

artefacts

and

features,

whereever

mind problems relating to colluviation).

thickness of extremely hard and compact shells with many burnt stones and the usual artefacts (see graph above for Trench 4). Although it has an overall dark grey to black appearance, the matrix is still the dark brown clay (Munsell colour 10YR4/3), although there is a lesser clay content in this layer than higher up in the midden. This layer represents the earliest development of the midden in the gully in the area of Trench 4 as the Mesolithic occupants threw their rubbish down in this part of the site; it continues into the eastern part of the adjacent Trench 31.

Layer 7 (Munsell colour 10YR3/3) is very thin and sometimes grade into Layers 6 or 8; it is present also in pockets amongst the stones of the floor in Trench 4, and it is therefore possibly in part pre-floor but also post-floor. It is also dark brown or yellowish-brown but is slightly more firm and better structured clay than Layer 6; one could argue that it is feasible that this sort of deposit on top of a developing midden, could have been the outcome of an episode of increased rainfall or, even just one single severe storm.

It may, in fact, merely be the older more

D. PRE- MIDDEN

compact base of Layer 6. It contains the same molluscs, burnt and unburnt stones and artefacts as Layer 6. Layer 8 (Munsell colour 10YR4/3)

Layers 12 and 13 are in origin natural (i.e. geological) deposits under the occupation debris of the midden but both contain a quantity of intrusive occupation debris, i.e. in origin these layers are natural but they have become modified by the activities of the people who used the site, possibly, firstly, in

resembles Layer 6 and 7,

except for the fact that it is much more compact (Figs. 4, 14, 16,

17). Its contents (i.e. artefacts, burnt/unburnt stones, etc) is exactly the same as in the above layers but it has visibly more pockets of winkles (Littorina littorea) present in the midden amongst the other shells. This does suggest that this layer of rubbish was deposited during a phase or period when winkles often formed the main molluscan diet for the day, either by

Late Palaeolithic and then Mesolithic

still present in large quantities. In sections through Trenches 3, 4 and 6 it is visibly a compact deposit underneath the stones of the floor (e.g. Plates 46,47). This layer is, therefore, of some chronological importance as it was found that in some areas of the site at least the big stones of the floor lies partly immediately on top of this layer, i.e. the Mesolithic site inhabitants here sealed part of this layer off by placing the floor on top of it. After the floor was laid down, the midden appears to have continued developing and Layers 6/7 were deposited. In the Sondage through the floor in Trench 6 it was also found that the limestone slabs lay directly on top of a midden deposit comparable to Layer 8 of Trench 4.

Layer 12 is only present in the lower lying parts of the site, as in the gully, where it overlies Layer 13. Layer 13 alone is present in every part of the site under the occupation debris and also outside it (see above in this chapter for the natural stratigraphy of the area surrounding the site).

Layer 9 ( Munsell colour 10YR4/3) is still the same colour as the above layers but it resembles Layer 7 more in that it has a firmer clay matrix but with a lot of minutely fragmented shells. The topmost few centimetres of this layer contains a higher proportion of fragmented limpet shells which has been referred to as Layer 9a in Trench 4 even though it is suspected that this fragmentation is the result of post-depositional activities, natural and human. Otherwise the contents of this midden deposit is no different from those described above.

Layer 12 is a dark brown to yellow brown clay (Munsell colour 10YR4/4 or 4/5) — which is present under the midden deposits in pockets or thick layers only in those areas where the site sloped down fairly rapidly in the direction of the Culver Well stream or where there appears to have been a gully prior to the beginning of the site occupation, as observed eg. in Trenches 3/1 (Hearth 1), Trenches 19/25 (Hearth 4), Trench 4,

Layer 10 when present, appears to be more in the nature of lenses, also more clayey than the above layers and is slightly more yellow brown (Munsell colour 10YR4/4). It is possible that these lenses are the remnants of a thin, uneven clay layer which was originally more widespread but became broken up by subsequent activities or circumstances on the site. It could, therefore, represent a minor climatic episode but still does not appear to be of major significance, not enough to constitute a change in occupation. The contents of this midden ‘layer’ are identical to that in the above layers: the same molluscs,

31,

is fairly distinctive and is a deposit

32

and

therefore occupation

41

(Figs.

4,40,43).

pre-dates the midden

In origin

development

of the site and, in some places,

this

basal

clay

and Mesolithic represent the old

post-Devensian land surface as also. the earliest occupation surface. Weather conditions and human activity modified it during the Mesolithic occupation on the site. The presence of large quantities of artefacts and unstruck smooth beach pebles in the upper levels of this layer, particularly in the bottom of the gully, e.g in Trench 41, is

artefacts, burnt and unburnt stones, etc.

11

times (Figs. 42, 25; Plate

40). It is, of course not possible to say for how long the site (and the original natural terrain) was in use before the midden had developed enough to occupy all of its present area, nor is it clear whether the initial occupation was confined to only a part of the present site or, equally possible, whether it was spread out over the site. This would mean that not all the artefacts found directly on top of these two layers are necessarily in a strict sense contemporary, but in general terms, some of the debris associated with these layers include the earliest known from the site. Conjoining cores and flakes found in Layer 12 of Trench 41 (Fig. 38) demonstrate the fact that at least in some areas the earliest artefacts on the site had not beem substantially moved by various post-depositional factors (see Trench 41 discussion). Several conjoining flakes and cores found in Trench 4 and other layers and areas of the sites also demonstrate that the lowest parts of the midden have remained reasonably undisturbed since deposition (Plate 56).

choice or circumstances, even though all the other molluscs are

Layer

SURFACES:-

of variable

13

STRATIGRAPHY OF DEPOSITS ABOVE AND BELOW THE FLOOR

compatible with the fact that they were directly and deliberately dumped there, either as a rubbish clearance strategy or for some other unknown reason, and sank down in the clayey deposit (Plate 15; see also section 3.16.). The layer also everywhere contains shell remains which generally become progressively less downward and are not nearly as concentrated as in the compact shell layers above.

The Stone slabs of the Floor, where present, do not strictly represent a layer but must here be considered in relationship to the above discussed strata. For the purpose of this discussion on the soil profile, it is necessary to point out that the floor directly lies on top of the compact midden which was identified as Layer 8 in Trench 4, without any other intervening soil deposits between it and the shell layers. This layer under the eastern edge of the floor in Trench 3 was dated to 7150 + 135 BP (see 2.9. ‘Dates). This layer can also be seen in Trench 4, immediately adjacent the floor, and in a sondage which was cut through the floor down to the natural in Trench 6 (see Fig. 2 site plans). In Trench 4 it can also be seen that the

This clay deposit would almost certainly have been subjected to modification

by

human

activity,

weather

conditions

and

periodical flooding by rain water coming down the hill and being diverted into the gully before and during the Mesolithic occupation; the water table here would have been slightly higher than in other areas of the site for much of the rainy seasons.

There

is,

however,

no

evidence

for

perrmanent

ponding of water and this is unlikely to have occurred here because of the general slope of the original terrain.

floor is covered by midden material of Layer 6/7, i.e. the slabs

of limestone forming the floor seals off the lower deposits of midden in part of the site.

The finding of concentrations of numerous round and oblong unstruck beach pebbles on top of this clay and within the top 15cms of it (see description of Site 1.4. and 1.5.) and Stewart this

Report),

does,

however,

suggests

that these

items

were

Where the floor is under the medieval lynchets, the thicker deposits of layers 1—5 survive clearly formed above the stone slabs. The relict soil level of Layer 4 can, in particular, be seen to lie undisturbed over the floor (Plate 44) In those areas, outside the lynchets, where the post-Mesolithic soil profile has been denuded or modified, these layers are very shallow and not all of them are always present, as discussed above for the rest of the site. The fact that the profile is still complete (particularly with the distinctive relict soil of Layer 4) at least in some areas, indicates that originally it could have been present everywhere on the floor but became disturbed or modified in medieval times.

Shells incorporated into this Layer have been dated to 7,285 + 60 BP (see 2.9. ‘Dating’). Layer 13 (Munsell colour 10YR5/6) is also a natural deposit, present over the entire site, but in the gully and hollows it is covered by the clay of Layer 12. It is a pale yellow gritty, sometimes very sandy, loam with many naturally occuring angular pieces of limestone and chert and is comparable to the Pleistocene solifluction deposit which is found as the ‘Head’ above the Raised Beach at the Bill (Arkell, 1953: 331-336) In origin, this layer is, therefore, a Late Pleistocene deposit. the big gully,

it is the basal

deposit

on which

Midden material, therefore, occurs

variously between, above and directly under the stones of this floor wherever this feature occurs on the site. This indicates that the erection of the floor (and any structures which may originally have been associated directly with it) came sometime after the first deposition of living rubbish but before the occupation of the site was terminated. The characteristics of the artefacts associated with the floor and the shell deposits above and below the floor are everywhere the same.

thrown into an area where the deposits were damper than in the rest of the site. This is borne out by the presence of small freshwater shells (see Thomas). Test augering carried out in unexcavated areas of the field, just west of Trench 31, suggests that this clay could possibly be up to a metre thick in some places as it was not always able to reach its bottom with the augur.

Outside

IMMEDIATELY

In the deposits above the floor under the lynchet a number of large limestone slabs were found in such a haphazard way that they were clearly not in any original context. As these were not found in areas outside the floor, it can be suggested that they may have been derived from some or other features previously directly associated with the floor and which had been disturbed and so become incorporated into post-mesolithic layers.

the

midden and occupation evidence occur, i.e. in some areas, it is

the original (natural) surface on which the first Mesolithic occupants of the site actually lived; a little intrusive Mesolithic debris is present everyuwhere only in the top few centimetres of this deposit. As such it is also the main pre-occupation buried land surface of the site.

In Trenches 1 and 7, east of the through the floor in Trench 6 (see seen that most of the floor is in a leads to the deeper gully. In these vary in thickness from about 30cms

A small trial pit dug well outside the western limits of the site, has shown that further down this loam becomes progressively more rubbly with sizeable pieces of limestone slabs and forms part of the geological layers of Portland, as discussed in this Report (see 2.8. ‘Geology’).

floor and in the Sondage Fig. 2 Site Plan) it can be shallow depression which areas the midden deposits to 17 or 18cms.

Immediately adjacent the easternmost edge of the floor a series of hearths are contained within small hollows within the clay of Layer 12, indicating that here the slope down towards the gully had already commenced (the point where the slope starts is actually visible in the profile of Trench 1 and 3 about 45cms

Molluscs found within the top of this layer within the gully (Trench 41) have been dated to 7,525 + 60 BP and is well in agreement with the radiocarbon dates obtained from Hearth 1.

14

outside the eastern edge of the floor. (Fig. 12). The layers below the midden under the floor (where exposed by excavation) is either that of Layer 12 or Layer 13 preMesolithic natural, depending on where the floor also straddles the gully. This indicates that the floor was laid down only in part over midden contained within the gully; in the north and south of the site the midden under the floor is generally shallower and outside the area of the gully.

of the site, even though the main deposits are comparable to those described above ( Fig. 14 and 1.4.Site description). In these areas the midden deposits can only in general terms be compared to those of the described section in Trench 4 (i.e. light coloured midden,

much higher clay content mixed in with the living debris. Also, in a sondage which was cut through the man-made stone structure of Feature 2 and the southern part of the main Floor (Sondage 10A and see site description), it was observed that the midden deposit here is 'split' into three thin layers, separated by thin brown shelley clay lenses (Fig. 14). All of these deposits are still associated with the typical Mesolithic artefacts encountered all over the rest of the site.

In the northernmost part of the site (above the lynchet in Field 2153 — see Fig 2. Site Plans aand Fig. 5) the slabs of the floor were laid down directly onto the rubbly natural loam of Layer 13 with only a very thin sprinkling of shell midden below it. This could suggest that the floor here was laid down at an early stage of the site occupation.

It is unlikely that this localised situation indicates a break in occupation; it is more likely that it indicates a difference in the original topograhy of the terrain here. The evidence is compatible with the suggestion that here there was an

LAYERS ON THE EASTERN AND SOUTHERN PERIPHERY OF THE OCCUPATION ZONE Areas east and southeast of the floor: All the deposits on the easterly side of the site are very thin, i.e. in those areas where the original ancient terrain was more level and where no gullies or hollows appear to have been present. Here a thin layer of greyish uncompacted midden material, similar to Layer 6 of Trench 4 and varying from about 16cms to only 5 or 6cms in thickness, is present almost immediately under the recent soils of Layers 1 —2. This suggests that the main site activities took place nearer to the gully and floor throughout the occupation even though some activities would have taken place further away from that feature, near a pit approximately 7m (24ft) east of the floor (see Site description). Immediately beyond the pit, the shell deposits rapidly thin out. It should also be considered that here some of the occupation debris may have been eroded away or displaced as post-mesolithic soil formation would have been much slower on level terrain than in lower-lying areas.

intermittent

secondary

nature,

i.e.

which

have

been

moved

of mud

either in another,

shallow,

PROFILES IN TRENCHES 31, 32 AND 41 These trenches are south-west of the Floor

( Figs.2 and 25 and

section 1.4. descriptions; Plates14 and 43)

and were excavated

with the intention of of the gully.

further researching the nature and extent

Trench 31 was sited, in part adjacent Trench 41 and wholly adjacent Trench 4, which revealed an edge of the floor on top of the midden and also an area of the gully (and of which the stratigraphy has been discussed in detail above); Trench 32 is immediately south of Trench 4 and immediately adjacent

In these shallow areas any sub-division of the shell midden is also not pronounced. The explanation for this could be that either less dumping of occupation had anciently taken place in these areas or else that some rubbish may periodically have been removed. Being on level ground, probably in an area of daily site activity (see site descripton) the development of the midden to any depth may here not have taken place in the same manner as in the gully and lower lying areas. It is also difficult in these areas to define which of these thin midden deposits are a

accumulation

gully or else mud was caused to “bank up’ against or around some feature(s), possibly associated with Feature 2 (as described). This situation is also compatible with the fact that the occupation area is here on southward-sloping ground, just outside the well-drained area of the floor. The archaeological evidence suggests that there may have been a series of small gullies here (see Site description). Water and mud coming down the hill would have had the tendency to byepass the higher parts of the floor on top of the peak of the midden deposits and would have accumulated in lower areas. Unfortunately it was not possible to research this area more extensiveky because of the presence of the fence and road.

As the first one or two seasons of excavation took place in areas where the topsoil was very thin and the post-Mesolithic soil profile was incomplete, the shell deposit was here referred to as ‘Layer 3’ until further work indicated that it could generally be correlated to Layer 6 in Trench 4.

of

darker midden, etc.) but here there is a

Trench 41.

The four trenches, therefore, gives a cross-section

of part of the whole gully. Trench 31: This trench provides a_ section through approximately the middle of the gully. A very dark brown, almost sterile, clay deposit was revealed below the uppermost level of uncompacted midden, which is elsewhere equated with

Layer 6 and in Trench 4 was shown to be mostly above the floor level. Below this dark clay is another thick deposit of compact midden material of generally similar appearance to

from

elsewhere by water and other agencies and re- deposited and which are still (relatively speaking) in the original position where they were left by the site occupants.

that of Layers

two deposits

Southerly areas of the site near the road: The profile of the deposits in the southern part of the site (1.e. associated with Feature 2 and south of it also differ from those all over the rest

8-11

(of Trench 4), i.e. the clay is between the

of midden.

The very dark coloured

clay was

referrred to as ‘Context 31', being characteristic of this Trench.

The original

15

surfaces of Layers 12 and 13 at this point

slopes

13 at a point where the floor of the gully rises up fairly steeply to the highest edge of the gully ( immediately just outside the area of present excavation).

very rapidly down (noticeable already in the adjacent Trench 4) into lower ground, presumably to the point of maximum depth of the gully. The ‘Context 31’ clay, therefore, appears to suggest an accumulation of mud in this lower terrain, just outside the living area of the floor. It may suggest a wet episode which occurred during the site occupation but during which mud only accumulated in the deeper parts of the gully outside the living areas.

The profile presented by the stratigraphy of this trench can therefore be explained by accepting that the contours of Layers 6-8 in the easterly section of this trench defines the outline of part of an actual original mound of refuse within the gully. It is apparently also the furthest east (as far as known) that such a mound had developed away from the area of the ‘Floor’. As suggested by the details of the stratigraphy, the mound, Layer 8, slumped in time and material moved to lower points in the gully, with other material possibly brought from elsewhere higher up the slope of the site, developing over this area of ‘inverted stratigraphy’. The lowest levels of Layer 8 and all of Layer 9 does not appear to have been part of this ‘slumping’ process and is present and compact within the gully depth and therefore represents the original base of the small ‘mound’ of rubbish.

Trench 32: The layering in this trench was also found to be problematical due to the fact that it was excavated over part of the easterly edge of the gully, which here runs in a general south-south-west direction; it is also just at a point adjacent where the floor was laid down on top of midden within the gully (see site plans and descriptions). Here the gully edge is very stoney and forms an irregular ridge and lacks the gradual slope of the western

edge,

as found in Trench

41, described

below. The midden material of Layer 9 is present in pockets amongst the stones of the ridge which is haphazardly wellbedded into the natural deposits of Layer 13. Layer 9 is, therefore, here the last of the occupation layers on top of Layers 12 and 13. In areas west of this stoney ridge the layers are approximately comparable to that in the eastern end of Trench 41, i.e. into the gully.

The AMS Radiocarbon dates from Monodonta shell samples taken in Trench 41 appears to reflect the situation indicated by the above stratigraphy (see also 2. 9.‘Dating’ and 2.10. Thomas, this Report). The oldest dates of 7,525 + 60 BP for Layer 13 and 7,285 + 60 BP for Layer 12 corresponds to the time when the first deposit of shell debris from cooking

Trench 41: The stratigraphy here was found to differ in some respects from the average nature of the strata described above for Trench 4 even though these two trenches are partly adjacent each other.

activities were left

maximum depth of the gully, but the includes the south-westerly edge of the upwards and indicates the sequence of occurred there. The geophysical survey gully does not continue in a south or beyond this trench (see 1.6. Gale, this

The inversion of the value of dates for Layers 7 to 8b

8 is,

S55BP) can be regarded as reflecting the situation when a mound of shell midden near the edge of the gully had formed, ceased to be in use and began to slump down with other deposits developing over it. The implications of the possible time lapses, if such in real terms, suggested by the dates are discussed in other sections of this Report.

DISTURBANCE OF THE MIDDEN: A CONSIDERATION OF DEPOSITIONAL AND POSTDEPOSITIONAL INFLUENCES ON STRATIGRAPHY On

throughout the trench, compact but the

‘Features’)

all

archaeological

sites

there

horizontal and vertical movement

is

usually

a

degree

of

of artefacts, both during the

period of occupation and also during shorter or longer time lapses afterwards. It can be expected that this must particularly have been the case on a shell-midden site like Culverwell.

top is less so in the easternmost third of the trench where it is much more diffused. At the base of Layer 8 and on the top of Layer 9 a

(Layer 7

6,800 + 60 BP; Layer 8a 6,730 + 55 BP and Layer 8b 6,410 +

This trench was excavated in 6 quadrates, one and two being at the easternmost end of the trench. In the easterly third of the trench the layers were found to be generally comparable to those in Trench 4 and the whole range down to the clay of Layer 12 was present overlying the rubbly loam of Layer 13 natural. Layers 6 and 8 are compact midden in that part of the trench. In profile it can be seen that the top of Layers 6 and 8 slopes down gradually in a westerly direction in the manner of the contours of a small mound (Fig. 25 and Plate 40). The bottom of Layer

natural terrain surface, after which

the midden layers continued developing. Layer 12 about midway in the trench, where the shell sample was taken, was at a depth of about 1.6m and was visibly in a position representing a point of maximum depth of the gully.

This reflects the fact that this trench includes, in

part, a section of the south-western part of it gully beginning to slope events which apparently also indicated that the south-westerly direction Report).

on the

constructed stone feature was found, elsewhere (see

described and discussed as ‘Feature 41’. In Trench 40 at Culverwell (see ‘Method of Research’ and ‘Site Description’) each artefact was recorded with a_ twodimensional co-ordinante, i.e. horizontally and _ vertically, within quadrants which sub-divided the trench. Arbitrary depths of about 2cms of shell midden were excavated at a time. A preliminary excavation of the trench had suggested that this part of the site may have been a reasonably undisturbed workarea of the site (Fig. 23).

Right at the south-western end of the trench the diffused midden of Layer 8 overlies a very stiff, almost black clay with very few artefacts and only minute powdery bits of shell but a lot of fine visible specks of charcoal. Stratigraphically, it is in the same position as Layer 9 elsewhere in the trench but it is unlike that layer, being much stiffer and darker coloured and also almost devoid of artefacts. It probably relates to the dark clay (‘Context 31’) of Trernch 31 and overlies Layers 12 and

16

In Trench 40 it was found that only a few artefacts had edge damage and the assemblage included a very wide range of sizes of debitage, including many fine pieces. This suggests that movement by water was not extensive, at least not in this part of the Culverwell site, even though it must have played some role in the site formation processes.

Barton and Bergman (1982, 244) have experimentally shown that the effect of trampling on archaeological deposits is related to the weight of the person and the method of depositioning of the archaeological material. Shells and artefacts are pushed into an underlying deposit of clay when the shell layer is very thin; when it is thick, and especially if the clay is dry, there is little movement in the upper levels of the midden but some in the lower levels (see also Stein, 1992). The movement can also be sideways, away from the place of impact of the foot. In wet clay, such as in the gully at Culverwell, there is downward movement, depending on the degree of fluidity of the clay and the weight of the person walking on the midden. After a lapse of time, the weight of overlying deposits will also cause some sinking of lower levels. Bioturbation, involving particularly earthworms and rodents, can be responsible for upward movement of material. In Trench 40 a number of ‘refits’ of flakes to cores or flakes to flakes have been possible (Fig. 38; Plate 56). The average vertical distance between the different components were 2-6 cms. and the horizontal distance in some cases less than a metre. Refits have also been found within the shell-midden layers of Trench 41 in the gully, but there no exact recording of the position of artefacts were done. It can, therefore, be suggested that post-depositional movement within the midden at Culverwell has not been excessive.

17

1.4.

DESCRIPTION OF ALL FEATURES

It is proposed to give in this Section of the Report a description of each feature and of the juxtaposition of the

ground in the (probable) living and working areas adjacent the floor and particularly near the hearths (e.g. Trenches 1, 3

features to each other;

and 40;

in the next section

all the features

Figs. 2 , 3, 23; Plates 18, 31, 50). Where the midden

is thick, the lowermost layers of the deposits are very compact, particularly under the floor and in the gully (see also 1.3. ‘Stratigraphy’). Away from the floor and the gully the midden gradually becomes less thick and, as it reaches the periphery or outer limits of the occupation area, it is no

and their possible interpretations will be further discussed. The layout of the site as a whole and its features can be seen in Fig.3 and in the aerial photograph, Plate 2.

GENERAL DESCRIPTION

more than just a thin spread,

3 or 4 cms thick, of shells and

debris on the natural pre-Mesolithic surface.

That part of the site which has been researched, covers an area of approximately 1326 square metres i.e. 34 metres north-south and 39 metres east-west. The whole of this area is defined

by

a spread

of shell

debris,

burnt

and

On the peripheries of the site, there is also an absence of compact midden at the base of the Mesolithic deposits. The midden is reasonably well protected and sealed where it is under the limestone floor but elsewhere a few burrows of animals have been located at a number of places within the midden.

unburnt

stone and Mesolithic artefacts, which becomes progressively more sparse towards the peripheries of the above (measured) area (see e.g. Plate 47). Within this area of midden are also several constructed features. It is not possible to say definitely that these boundaries mark the actual extent of the ancient living site as it is likely that at least some of the material on the peripheries are there as a result of colluviation. Also, it is also likely that outside this area various activities did take place which left little or no detritus. The biggest problem in determining the extent of the site is in the fact that only Field 2154 was available for extensive research. Field 2153 was only available for a few weeks over two years and there were conditional restrictions which ruled out intensive research. Part of the site possibly extends under the road and other adjacent fields were not at all available for research. However, taking these points into consideration, it is likely that originally the size of the habitation area or main activity area associated with the structures roughly approximated the measurements given above.

The midden contains Patella

revealed

midden, up to a

that

the

the waste products from a matrix of dark clayey

thickest

part

of this

occasional

Nucella

shells show signs of

direct contact with fire. Vast quantitites of very fine fragments of shell occur throughout the midden deposits; some of the fragments are so minute that they can best be described as ‘specks’ or ‘powder’, giving most of the shell layers a generally grey colour when first excavated. It is possible that this powdery detritus represents the remains of molluscs which have been in direct contact with fire. Throughout the site the midden contains large quantities of lithic material - many being artefacts or the debitage in producing them and also large quantities of natural, burnt or

is

very roughly sub-circular or oval _ shaped, lying approximately along a north-north west to south-south east axis. The midden now appears as one large feature and will be described as such (but see discussion below). Excavation

with

few of the complete, or semi-complete,

molluscan remains, artefacts, burnt and unburnt pebbles or pieces of pebbles and limestone, charcoal, ochre, some bone

as far as can be determined,

lineata,

It was often noticed, during excavation work on the large midden that quite frequently several limpets shells are found one inside the other, almost as though they had been stacked like that. The empty limpet shells also very frequently contain pieces of charcoal despite the fact that

(For further description and discussion of the midden see Chapters 1.3. Stratigraphy and also 1.8. and 2.11. the contributions by K. Thomas and M. Stewart in this Report). The Culverwell site is defined by a large midden of

The outline of the midden,

Monodonta

species of Littorina,

lapillus and other molluscs. A limited analysis of samples were taken from the midden in Trench 4 at the beginning of the Research Project and subsequently numerous samples were taken in other areas for various studies; these have provided clear indications regarding the ratio's of the different molluscs present (see 1.11. K. Thomas and also Appendix 1 this Report for full analysis). A small quantity of bone has also been found in the midden layers of the site and these too are reported on by Thomas. The sampling results demonstrated that the midden content is not everywhere homogenous but there are minor variations from place to place.

MIDDEN

and other living debris, such as_ producing tools, mixed within earth.

and

mostly shells of

unburnt, stone fragments.

Portland chert, Purbeck chert and

oolitic limestone predominate. material’ and 1.8. the study undertaken by M. Stewart).

large

(See 1.7. ‘Archaeological of the unstruck pebbles,

metre thick, lies on the western side of the

site in an area where the rubbish was thrown into a gully (see below) so that the top of the midden at least became level with the surrounding terrain outside the gully, or was just slightly raised above it. These thick deposits of midden within the hollow is straddled, in part, by the central part of a limestone platform or floor (Plate 46 and.see below). Elsewhere it is evident that much of the debris was left on the

When work first commenced on the site, the midden contents of 12 large marked buckets, taken from the thick part of the

midden, were studied and broken down into their different components. The average weight per bucket was 8.165 kg (18 Ibs) of which 44% was shell and the rest stones, artefacts and earth. This figure agrees reasonably well with the comparative weights (dry weights) of limpet, periwinkle,

18

topshell and stone in the samples, referred to this Report Section 2.11. .

by Thomas,

The precise south-western limits of this gully have not been traced. The geophysical survey (Section 1.6. by Gale) suggests that it does not continue beyond Trenches 32 and 41, as also suggested by the excavation work in those trenches ( Fig. 3). It has also not been possible to research the northern parts of the gully fully as this is in Field 2153 where the opportunities for excavation were limited.

Using the above described method of recording the extent of the midden,

a total of 3135 buckets of waste were counted

for the area excavated up to 1997 = 25,080 kg of midden debris = 11,380.05 kg of shells. The weights for the whole site should be at least treble this as large areas remain

The whole of this gully extends downwards for an average of about 0.5 — 1.0 m into the natural original rubbly loam (Layer 13 — see 1. 3. ‘Stratigraphy’ Fig. 4). Prior to the site occupation, the bottom of the gully acquired a dark brown or dark yellow-brown clay covering, Layer 12, overlying the rubbly loam. The clay contains some artefacts but only the top few centimeters of the loam contain artefacts and charcoal, in both cases derived from the midden layers above.

unexcavated, particularly under the floor, in Field 2153 and

possibly also some areas now under the road.

THE AND

PRE-MESOLITHIC GULLIES

TERRAIN

: HOLLOWS

The stratigraphical profile of the site clearly indicates that the western part of the site lies in a shallow basin-like depressed area where the ground in general starts sloping down southward towards the sea and also south-westwards in the direction of where the stream of Culver Well is (see e.g. Fig.4 and Plates 4, 43). This means that in some areas of the site a two-directional gradual slope is noticeable. In addition to this depression, the site contains a large gully which transects much of the sloping terrain on the western side and a number of smaller gullies which are all in the southern and southeastern part of the site.

During the site occupation the whole gully was filled with shell and artefact debris. The midden layers in the gully, i.e. on top of the clay and loam, as seen for instance in Trenches 4 and 32 have a clayey ingredient to varying degrees but it does not have any clearly recognisable layers or lenses of loam or rubble which could have been derived from the sides of the gully during the site occupation. This suggests a relatively short period for the filling of the gully.

The Large Gully A number of features have been found at or near the bottom of this gully in various areas of it and will be separately described below. A large number of oval or round beach pebbles have been found in clusters in the gully on the surface of the basal clay or within the top few centimeters (Plate 15). Two conjoining flakes of a distinctive rust-brown

A large gully anciently cut diagonally across part of the sloping natural terrain on which the site was later established ( Figs. 3, 6(6), 25; Plates 14,15, 41). It is more visible at ( or

near) the western perimeter of the occupation site. It is very sinuous and irregularly shaped, but where it appears in the

chert was found

excavated areas, it lies ona general north- north-east to south-south-west line. As far as ascertained, it is on average

near each other in Layer

12, at the lowest

level of the gully in Trench 41 (Fig. 38).

about 4-5m wide ( e.g. between Trenches 32 and 41. Fig. 3). In Trench 32 large cryoturbated rocks are exposed in a very jagged edge (Plate 14).

No certain evidence has been found that this gully had in any way been modified in shape or size by human intervention, except in the north-east for the construction of

Excavation has shown that the gully is within generally sloping terrain and begins approximately in the area of Trenches 19 and 25 near the northern limits of the occupation site, is under the floor in all of Area A and partly under the floor in Areas B and C, so that its eastern edge falls just short of the eastern edges of Trench 3 and Trench 6 (See Figs. 3 and 4 for position of trenches and features). Here it then turns slightly in a more westerly direction so that an easterly edge of it is visible in Trench 32. Part of the other edge would therefore be under the floor in Area A and under the remaining baulk between Area’s A and C.

Hearth 4, as mentioned above.

Depression and

gullies in the south

There is evidence of aseries of small shallow gullies or possibly one large one in _ the southern part of the site, just south and south-east of the remnants of a wall referred to below

as Feature 2 (Fig. 3;

Plates11

and

12). The

terrain

here dips down noticeably. It was not possible to investigate these gullies fully because of the proximity of the fence and the road.

In the southern part of Trench 32 it appears to become less pronounced and this may therefore mark a point where it begins to level out. This would give it a length of approximately 18m. A short stretch of the westerly edge of the gully was encountered in Trench 41 (Plate 40). The north-eastern limits of the big gully coincides nearly with the edge of a shallow depression which had contained the large hearth 4 in the north of the occupation site ( Fig. 9; Plate 30 and see below, description of that feature). The edge here appears to be very abrupt, so much so that it has the appearance of having been artifically excavated, at least at this particular point.

THE FLOOR OR SLABS (Feature 1)

PLATFORM

OF

LIMESTONE

The reasons for prefering the description ‘floor' rather than pavement or platform for this artificial feature are discussed in the next Section of this Report. The floor consists of slabs of limestone, carefully laid down

on a north-northwest to south-southeast alignment, beginning about halfway into Field 2153 and terminating, as far as

19

known, in Field 2154 adjacent what is now the Bill Road (Figs. 3, 6(1-6), 7,10; Plates 2,3,5,7,

here very shallow and there is a strong possibility that some of the original stones may have been removed by ploughing in medieval times. (A number of limestone slabs were indeed found haphazardly distributed within the ploughsoil of Layers 1 and 2 within the lynchet between Fields 2153 and 2154). However, the floor was still largely intact in the southern part of Trench 19 but about midway in that trench there was a gap from east to west in the floor, a strip of roughly a metre which was free of paving stones. There were no inications as to whether this was an intentional feature or whether the stones of the floor had become dislodged in this area due to the thiness of the topsoil. North of that the floor again continued unbroken for about another metre but further north the stones became more randomly distributed until they were completely absent in Trenches 20 and 23 in Field 2153.

18). It is rectangularly

shaped with the sides generally reasonably straight and parallel, except at a few points, as further indcated below. It has an average width of 3.6m and is, as far as it was possible

to ascertain, 31.09m long, 113.73 sq. metres.

giving an area of no less than

The stones used for the floor are of an average size of 25-30 sq cms and an average thickness of 5 to 6 cms and all have at least one level,

flat surface,

facing upwards;

they are in

plentiful supply on the island and was originally derived from the oolitic limestone "slat" beds (see 2.9. ‘Geological Background’). They are neatly laid horizontally adjacent each other, with only a few gaps. Quite a number of big flat stones have been found, during excavation work, in the hollows and gully to the west and south of the floor, suggesting a fair amount of natural slippage during and after the site occupation. Also, it was evident that a number of slabs of stone have also been dislodged by the plough through the centuries, as they were found lying loose on the surface of the field, concentrated in the area of the floor,

A small sondage was cut into the lynchet dividing field 2153 from 2152 as it was argued that the feature here would have been well protected by the thick overburden of soil had it been there, but the stones of the floor were not present,

confirming the northern limits of the feature. The spread of the midden and artefacts also decreased in this area and ceased all-together in the north of this field. (It was unfortunately not possible to obtain permission to excavate more extensively in Field 2153 nor in Field 2152).

before work on the site commenced.

During excavation, a number were also found loose in a tumbled and haphazard fashion in the brown loam above the area of the floor.

Here in the north of the site the floor of flat limestone slabs is laid down almost directly on the natural rubbly loam of Layer 13, and only a very thin spread of midden material is present below the stones and above them (Fig. 5; Plate 9).

The floor is everywhere associated with the shell midden. In sections cut through the midden in Trenches3, 4 and 6 it was noted that the slabs were placed directly on top of shell debris (Layers 6 or 8 midden of Trenches 4, 6, see 1.3.‘Stratigraphy’; Figs. 4, 5, 12, 17; Plates 3, 9, 46, 47) so

In Areas A and C (see location of trenches Figs. 2, 3) the floor was laid down on the thickest part of the shell-midden within the gully, as already described. In the south of the site, just north of Feature 2 in Area D, the floor is on top of a thinner deposit of midden but there is also midden 5-7cms thick present on top of the floor.

that there are no other deposits of earth or clay intervening between the Mesolithic debris and these stones. In all areas some midden material is found between the stones. In some areas midden material is fairly thick on top of the floor, e.g in Trenches 11,12 and 27 within area D.

The stone floor consists mainly of one thickness of slabs, but in some places two or more stones were placed on top of each other or partly on top of each other, possibly to cover too large gaps between stones or to fill up depressions or to replace original stones which may have slipped on the muddy midden or have become buried under rubbish.

At

further

south

or,

alternatively,

whether

a point about midway

metres

in Area A (Trench

3), about 3

from the eastern edge of the floor, and parallel with

that edge, a series of roughly square-

shaped

stones were

placed in the floor in a compact formation, next to each other

in a double straight line, forming an integral part of the floor, over a distance of about 3.3 metres ( Fig. 10; Plate 7). At about the same point on the eastern edge of the floor, a small square area or gap with straight sides and measuring a little less than a metre square, lacked systematically placed stones. The gap had a neat regular aspect.

The indications are that the floor originally terminated at or near a point in the south where it is associated with one or more other stone features ( Feature 2, see below; Plate 6) on the edge of a small gully, as described above. A number of slabs of stone were found in the small gully, but it was not clear whether they indicated that the floor had originally continued

so

During work on the site it became apparent that the sides of the floor were slightly out of alignment at a number of places(Plate 37). These discrepancies created the impression that the floor was laid down in different stages or in sections without the builders ensuring that the extension's sides were in a completely straight line with the previous edges.

these

stones were in that position due to slippage of the floor or perhaps constituted tumble from some or other superstructure associated with Feature 2. The uncertainty is due to the fact that excavation is only possible up to the edge of the field adjacent the Bill road.

The first change occurs immediately adjacent the where Hearth 4 (see below) was found, just north medieval lynchet (Area B of the excavations, Figs. Plates 30). The edge of the floor here turns slightly

The original northern limits of the floor appear to fall within the area of Trenches 20, 22, 24 and 23 (field 2153 see Figs. 2 and 3 and Plate 9), excavated in 1971 and 1972, or between

from

the rest of the floor

northerly direction.

Trench 19 and these trenches. The deposits above the floor is

20

in Field

2153,

towards

place of the 7, 8; away

a more

The second, more pronounced, change occurs at a point under the highest part of the medieval lynchet (Fig. 3; Sondage 3, north of Trench 3) where it was well protected from later disturbances by the thick overlay of deposits. The eastern edge of the floor here goes off at a slight tangent in a north-easterly direction, the floor thus becoming slightly wider by about half a metre. The third change in the

The westernmost end of this feature coincides with th end of

the

floor

on

that

side

of the

site.

The

assemblage

of

Mesolithic artefacts associated with the feature over most of

its length, is very substantial and of the same nature as found on the rest of the site.

Excavation work over the whole length of Feature 2, made it

construction of the floor occurs in Area D, where the floor

very noticeable that, although the feature is continuous

appears to widen out into a structure with a semi-circular aspect (see Feature 3 below).

generally forms a straight line, there are at least three definite parts of this feature which, although in alignment with each other,

It was also noticed that at each of these points of change there is a larger than average slab of limestone (see below for

big stones were subsequently lifted and various possible interpretations of these features relating to the floor are discussed below. below

records;

aspect

of

the

alignment

will

be

further

The base of some of the stones of the whole feature are within a layer of shell midden material with a lot of whole shells but in the south-western area most of the stones of the feature lie on top of the stones of the Floor. Some of the bigger stones and their peaks are surrounded by a_ clay deposit, similar in appearance to that of Layer 5 as found in Trench 4 and in Keeley's section (see ‘Stratigraphy’ and Fig. 13, 14; Plate12). The implications of this stratigraphy are discussed in the next section of the Report.

the Floor

underneath it are thin in some areas, particularly in the north and south of the site, where it ranges from a maximum of

about 10 - 14cms. The most complete example of deposits were found

this

discussed below.

It is important to point out here that the nature of the deposits above and below the floor is not everywhere identical, as discussed in detail in the chapter 1.3. on ‘Stratigraphy’. The midden deposits on top of the floor and

which

differ from each other in the method of construction

and in the nature of the stones used for construction. There are no gaps or breaks in the alignment of stones and the parts all link up perfectly with each other. Collectively, the whole of the stone alignment is referred to as Feature 2 but the individual parts are referred to as Feature 2a,b and c in the

description of Stone X in Area A; Unnumbered stone at junction Area A and B; Stones S and W in Area D). These

Deposits above and

and

above the floor, can be seen in the soil

profile of the northern and north-western face of Trenches 4 and 18, Sondage 3 and the whole length of that face in Area A, i.e. under the apex of the lynchet dividing Field 2153 and 2154 (Fig. 3, 4; Plates 32, 33, 44).

The westerly edge of this beginning a place in Trench with the western edge of the continuous, in a fairly straight from this edge of the floor to a Trench 38 where it appears

Dr. Keeley's soil samples

were taken at a point in that face in Trench 18 and the column was also taken here. The deposits occuring under floor is best exemplified in the sections of Trench 4 and sondage through Trench 6 and consists of midden with layer of clay below in the gully area and lower down natural rubbly loam over the whole site (Fig. 21;

soil the the a the

stone alignment has as its 14 which coincides exactly floor and from there it is line in an easterly direction point between Trench 35 and to either terminate, changes

character or else turns south-eastward under the road; it is

definitely present in Trench 35 but not in Trench 37, the most easterly trench excavated (see Site Plans Fig. 3). This gives the feature a total length of at least 30 meters (100 ft).

Plates 46,

47). In Trench 38 (between Trenches 35 and 37) a number of very large stones were found, by the southernmost edge of that trench (adjacent the road); they were fairly closely spaced and were possibly part of Feature 2, but this could not be ascertained with any certainty. There is a possibility that

STONE ALIGNMENT IN SOUTHERN PART OF THE SITE (Feature 2 as sub-divided into a., b. and c.). Towards the southern end of the field (towards the road), the midden deposits on top, below and adjacent the floor were found to gradually become a great deal less thick, averaging no more than about 10cms. in any one place (see discussion 1. 3. ‘Stratigraphy’). Here a substantial alignment of big, irregularly shaped nodules of both Portland and Purbeck stone have been placed at an approximate 90° angle to the

Feature

2 here

may

turn

more

southward,

i.e. in a more

south-southeast direction under the road or that the stones in Trench 38 are associated with another feature of which no details are known due to the presence of the road (see also Feature 2c below).

edge of the floor, described above (Figs. 13, 14; Plates 5, 6,

During excavation work in Trench 37, in the easternmost part

10, 11,12).

of the field (Figs.2, 3), far less Mesolithic artefacts were found than elsewhere on the site or in the rest of Feature 2, the only noteworthy find being a big flat stone of which the under surface had some abrasions on it as though used for pounding something on. This led to the conclusion that the original outer perimeter of the main occupation activity zone of the site lay somewhere in this part of the field and no further trenches were opened in this area.

The whole alignment is of irregular width but is on average 1.5m (5ft) wide. The stones forming the feature are generally much larger than those of the floor and their tops are higher than the level of the floor. The over-all general character of this linear feature therefore differs from that of the floor of which the stones are almost all flat, regularly shaped limestone slabs, horizontally placed, unlike the stones in Feature 2. Work has shown that this feature as a whole forms an L-shaped structure with the floor and visually appears to be closely associated with it.

Work everywhere on Feature 2 was complicated by the proximity of the fence, pavement and road and the requirement, for safety reasons, to leave an adequate baulk

21

between the excavations and the modern features. _ For this reason it was not possible to ascertain without doubt what the nature of this feature is and what exactly its association is with other features on the site. Ideally more work should also have been done to research the chronological relationship of this feature with the rest of the site. However,

the site will be discussed separately later in the Report as it is also necessary first to consider the details of Feature 3 and Stone "W", which appeared likely to have had some connection with either Feature 2 or the floor, or perhaps both (see below and next Chapter).

in the next

section of this Report the evidence is discussed to support arguments for its contemporaneity with one or more phases of occupation by the Mesolithic site dwellers.

FEATURE FEATURE

Feature 2a: The western part of Feature 2, aproximately 3.5 meters (12ft) long, coincides exactly with the width of the large floor, discussed above. The majority of the stones used in Feature 2a are very bulky, i.e. bigger and more irregularly shaped than the stones of the floor and also of the rest of Feature 2; they are invariably placed closely to each The majority, but not all, of these stones are

chunks of tufaceous or porous fossiliferous oolitic limestone from the fossil beds of the island (see geological background). Furthermore, within the whole length of Feature 2a , the alignment consists of a clear double line of

but for the next stretch of about

CIRCULAR

Stratigraphical Relationship of Features 1,2 and 3 Sondages 10 and 10A: To obtain some indications of the stratigraphical relationship of the features in the southern part of the

site to

each

other,

Sondages

10

and

1IOA

were

excavated, diagonally through a part of Feature 2a and the Floor, on a line starting from near Stone "W" and going through underneath Stone "3X" to the southern end of Feature 2a (Fig. 14; Plate 13). Some of the stones of Feature 2a were removed, but Stone "3X" (see above) was left in situ on top of a column of the deposits underneath it. After initial excavation, the Sondages were extended southwards, by removing more of the stones in 1988 and re-examined. At all levels only Mesolithic artefacts were found, including three picks and a round pebble in the midden underneath the original double line of stones and almost below Stone "3X".

Feature 2b: The regular alignment of Feature 2 continues as a single line in an easterly direction from Feature 2a, as above,

OR

The hollow has an estimated diameter of at least between 2 to 3 meters (6 - 10ft), as far as it is possible to judge. That part of the circumference which has been excavated, is marked by stones, some quite large and some placed next to each other diagonally into the soil.

big stones, about 60 cms wide, neatly and closely placed, 1.e. diagonally across the width of the floor and this feature has a fairly substantial appearance. One of the big slabs of limestone within the northerly part of the double line is aproximately rectangular in shape and the top surface is smooth and horizontal, and it is larger than the average stone of Feature 2 (this stone is referred to as 3X on site plans (Fig. ** and for discussion see below). This stone is about 2m south-east of another big rectangular stone W which is approximately in the middle of the floor. Feature 2a appears to terminate near another big roughly rectangular stone (Stone "S" on the plans) where Feature 2b apparently begins.

defined

SEMI-CIRCULAR

During work in the southern part of the site it became apparent that partly under, or near, the junction of the floor and the easterly end of the linear stone structure of Feature 2a and the westerly end of Feature 2b there was a semicircular (or part of a circular) hollow under the stones of the Floor and Feature 2 (see Sondage 10: Fig.13,14; Plate 12,13) That part of the circumference of this feature which was initially found, lies outside the stones of Feature 2a but limited excavation here suggested that it continues under Feature 2a and its exact shape and extent was not, therefore determined. Stratigraphically it is underneath the stones of Feature 2 in such a way that it is partly under Feature 2a and partly under feature 2b (see discussion below on Stratigraphical relationships between the features).

The above description of Feature 2 applies to the whole length of it, but , as pointed out there were noticeably three different sections within it.

other (Plate 10).

3:

12 meters

(approx 38ft) the stones are generally less bulky and are mostly of fairly flat limestone slabs; they are also not as closely packed as in Feature 2a. The effect of a double line is also not so noticeable here. For these reasons, this part of the alignment was recorded as Feature 2b in order to give this part an element of separate distinction. Possible explanations for the differences between parts a. and b. of Feature 2, as will be discussed fully in the following section of the Report.

The sequence of features and deposits in the sondages were found to be as follows, in descending order:

1.Yellow-brown Loam: The tops of the majority of stones of Feature 2a are covered by a yellow-brown loam, similiar to Layer 5 of Trench 4 (see 1. 3. ‘Stratigraphy’). Some shells are present.

Feature 2c: The stones of Feature 2 appear to be present within the whole of Trenches 35 and 38 but not in the adjacent Trench 37. Included in Trench 35 are several quite large stones but, generally, the stones have a more ‘organised’ appearance in the way they are placed and are also more horizontally placed than in section 2 of this long feature. In the adjacent Trench 38 the placement of stones were only present in the southern part of the trench, i.e. as far as could be ascertained, they appeared to make a turn southwards or else terminated here.

li. Feature 2a:: Some of the stones of the feature have their lower parts in a darker brown clay which appears to slope down towards the south where there may be another series of gullies (described above). However, some of the stones of the feature are within the top of a layer of midden material.

The interpretation of the stratigraphical and_ spatial relationship of Feature 2 and its different parts to the rest of

The dark brown clay, as it slopes downwards in the southernmost part of the site, dips under another layer of

22

brown clay with very few stones and almost devoid of finds, which is similar to Layer 4 of Trench 4 (see 1.3. ‘Stratigraphy’). This clay here is covered by a very thin lens of stoney clay, reminiscent of Layer 3 of Trench 4.

this evidence

iii. Shell midden (Midden 1 in the section drawing Fig 14, is similar in texture and appearance to Layer 6/7 of Trench 4; 8-15cms thick. This is below the double line of stones of

These four hearths vary in diameter from a metre to about 1.5 to 2 metres. They are all on the eastern side of the floor, i.e. on the side away from the present-day prevailing winds, coming from a south-westerly or westerly direction (see Fig. 3 for their positions). Also, they are all in close proximity to the edge of the floor and certainly less than a metre away. Round the perimeters of the hearths, in every case, a larger than average percentage of picks, heavy pounders, choppingtools, and abraded pebbles have been found (see discussions on artefacts) this suggests that these were the tools most often used in the preparation of food, e.g. breaking up or pounding roots or other tough edible plants, prior to cooking.

Feature

2a

and

also

under

Stone

'3X';

also

above

the Floor,

apparently have their

being tops

acontinuation of within

the

as indicating the presence

of

under the baulks which separate the trenches.

and

Feature 1,

layer of midden,

except in one small place where there is a lens of yellowbrown clay on top of the feature. Mesolithic artefacts are associated everywhere.

Hearth One

This hearth was adjacent the floor of limestone slabs in Trench 1 and partly in Trench 3 and right on the eastern edge of the slopng terrain , but outside the gully, which is partly under the floor as discussed above ( Figs. 11,12; Plates 3, 18 - 25). The hearth was contained within its own small welldefined depression, which appeared to have been arteficially scooped out; it was roughly oval shaped with sloping sides and a diameter of about 1.5 metres, going down through dark grey, compact midden so that its base was in the tough clay of Layer 12, which is here immediately below the midden (see ‘Stratigraphy’). This small depression within the clay layer had a maximum depth of about 20 cms but the total thickness of the black midden material constituting the hearth within the depression was about 40cms. The base of this hearth included a lens of pure charcoal varying between 2cms to S5cms in thickness. The base of the hearth was,

The yellow-brown clay is also present outside the edge of Feature 1, where it partly covers another stone feature at a slightly lower level, as in v. below. v. Closely spaced, horizontally placed limestone slabs, some of them placed at an angle and at the level of the feature above referred to as Feature 3. Some of these stones of this feature are wholly or partly within a layer of compact midden material (Midden 2). vi.

regarded

ancient hearths. Four definite hearths were found but there are indications that other smaller ones remain unexcavated

between stones of Feature 1 (floor, the next layer down). The midden contains Mesolithic artefacts, including picks. In the southern and western extension of the Sondage, under the southern part of Feature 2a and where the stones of the Floor appear to be on the edge of one of the small gullies, mentioned above, the midden merges into a thin deposit of yellow-brown clay.

iv. Stones

was

Shell midden 2,

compact and hard, many burnt stones,

average thickness 5 large stones.

to 8 cms. covers yet another level of

vii.Some more slabs of limestone appear to be least partly below Midden 2. It was not possible these lower stones to a specific feature or to whether they were, in fact, part of Feature 3 or another feature (Feature 47).

present at to relate determine part of yet

therefore,

below the level of the floor but, because

of the

sloping terrain, the top was almost level with the floor. Several large natural slabs of limestone were scattered round the edge of the hearth. The clay within the depression was hard baked and quite red

viii. Under some of the stones of Feature 3 is clay with some whole shells and burnt stones.

when

first

found,

so

much

so

that

it

almost

had

the

consistency of ceramics (Munsell Soil Chart Reading, wet 2.5YR 3/5 dark red to 2.5YR 4/6 red). This suggested intensive use of this particular hearth.

The presence of more than one feature of slabs of stone separated by different layers of clay and midden are further discussed below.

Charcoal from sampling for the archaeomagnetic useful (see 2.10. for this problem

HEARTHS Large quantities of charcoal dust appears to be distributed throughout the length, breath and thickness of the midden, but particularly in the areas where it is thickest, round the floor and in the large 'gully' where the midden has a dark grey or black appearance. Pieces of charcoal are frequently found, often inside shells or between two or more limpets.

this hearth provided two dates and the TL dates also came from here. Samples for dating from here proved too weak to be ‘Dating’, where the possible explanations are discussed ).

The area immediately round the hearth was one of three on the site from which small specimens of bone came (the other two find- spots being the pit and the gully in Trench 41 — see discussion below

and 2.11.

K. Thomas,

this Report)

. The

bone from this hearth was particularly from the north-eastern edge of it. Most of these pieces were of a very fragmentary nature, as though they had been pounded (but see ‘artefacts’ for some which may have been worked to form points (Plate 57). Several exceptionally large chopping-tools, unstruck heavy pebbles, picks, very large steep scrapers and

During the researches a number of well-defined roughly circular areas of dense black midden was encountered which appears originally to have been contained within small depressions, going down through midden material and with their bases scooped out of the underlying clay of Layer 12 ;

23

pounders were found in the same area as that from which the bone came (see 1.7.‘Archaeological material’ and Plate

the yellow loam.

20,21,22).

The base of this large hearth too was in a shallow depression within Layer 12, but this shallow depression itself was contained within clay on the edge of the large gully of the site ( Layer 12), i.e. a depression within a depression. The edge of the small depression containing the hearth sloped gently upward everywhere except in the north-east where the edges of the large and small depressions coincided and sloped up abruptly and stopped. Here there was a fairly clear near-vertical edge, suggesting that the small depression had been artificially shaped, at least in part.

A

large

assemblage

of flake

tools

and

some

pierced shell beads also came from the general area round this hearth. A complete

charcoaled wooden

object, with a well-defined

outline, about 25cms (10 in) long by 5cms (2in) maximum width, was found, partly in the hearth and partly on its edge. The charcoal was, unfortunately, very powdery and the object could not be lifted intact. The charcoal was carefully spooned out and a plaster cast of the resultant hollow was made; the cast was not completely successful but the shape vaguely resembled a short club (Plates 24, 25).

The midden which had accumulated in the hearth is on average about 60 - 75 cms thick and is also very black and densely packed. Just outside this hearth, in the northern

One small post-hole was found near the edge of the hearth, going down through the midden material into the clay below. It had a diameter of only 3.5cms. Hearths 2 and

part of Trench 25, the midden

decreasing in thickness, and was there more grey rather than black and directly on the rubbly loam of the natural deposits on the site outside the gully, as discussed above. The midden seemed to be almost completely absent further north in adjacent trenches, suggesting that this hearth was on the north-easterly limits of the occupation site.

3

Hearth 2 was found in Trench 6, immediately adjacent the eastern edge of the floor and south of the baulk which separates this trench from Trench 1 in which Hearth 1 is, discussed above. There is therefore a possibility that it may merely be the edge of Hearth 1 or else a smaller part of the big hearth. It was also contained in a shallow depression, about

13-18cms

The possibility was considered that this large hearth, in reality, represents not only one fire but a number of smaller ones adjacent each other;

Hearth 3 was also in a shallow depression, varying from about 11-15cms in depth, immediately adjacent the nearcentral part of the floor in Trench 11 (for position see Fig. 3). It did not contain quite as thick and densely packed midden material as the other hearths.

By the south-eastern edge of the hearth (southern extension of Baulk 25) immediately under the apex of the lynchet, a pile of 11 picks and pick rough-outs were found, lying haphazardly directly on top of each other and all touching each other ( Fig.6 (1), 29; Plate 31). One of the picks was much larger than the average found on the site, being about 30 cms (1 ft) long and being beautifully worked all round; six of the other picks were completely retouched but three were in varying stages of completion, with some only slightly retouched. Included in the pile was one completely natural piece of tabular limestone; although roughly of the

Hearth 4

This hearth was the biggest one found on the site, having a diameter of at least 2 metres (Figs. 6(1), 7, 8; Plate30). It is mostly sited in Field 2153 above the lynchet with Field 2154, the

south-western

corner

of

Trench

25,

and

because of limited time, it was not

possible to explore this possibility. Alternatively, there may originally have been more than one fire, and they may have become amalgamated in the final stages of their usage. However, for convenience sake this hearth will be regarded as one feature.

deep, which continued under the baulk and

was therefore not fully explored. There were also several picks associated with this hearth. The midden in this hearth was also black and compact.

under

was considerably thinner, i.e

the

southern part of Baulk 25, but it is partly under the lynchet (unexcavated at this point) and appeared to extend up to or near to the north-western corner of Trench 21 (see Site Plans

same

shape

as a pick, it was

not

at all retouched,

and

so

demonstrating quite clearly that the site occupants made the picks from natural pieces of limestone which frequently required very little work to turn them into a usable pick.

Figs. 2, 3).

Due to the fact that the whole of Field 2153 was available for only a limited time (in 1975 and 1979) for exploration and back-filling, it was unfortunately not possible to examine this hearth and its surroundings fully; it was discovered during the last week or two of work in this part of the site. ( A column of midden/hearth material was left intact, just in case research here again could become possible in the future).

This collection of completed and semi-completed picks gives a good insight into working techniques of the site residents and will be further discussed in section 1.7. of this report. Apart from this cache of picks, a further total of 18 picks, also in varying stages of completion, were found on this side of the hearth, as well as 6 natural tabular ‘pick shaped’ limestone pieces, 3 chopping tools, 2 pounders, a number of cores and several round pebbles of red sandstone, which may have been intended for use as pounders. Everything foun round this midden was kept together to demonstrate a complete assemblage related to activity round a hearth.

The hearth is also near the eastern edge of the floor but it is also by the apparent north-eastern edge of the large gully, which here appears to slope more gently and begins dipping under the floor (as far as could be ascertained). Immediately by this hearth the line of the floor slightly changes direction and about three metres further north the floor lies directly on

As this hearth is virtually under the apex of the lynchet, the full range of deposits as in Keeley's Samples (Layers 1 - 6; see 1.3. ‘Stratigraphy’ and Appendix) were visible in the

24

artificially built-up round the edge. In some places, this clay was baked red (see Munsell Soil Chart readings below and compare Hearth 1). On this clay, and partly embedded in it,

southern face of Trench 25 above the hearth and its surrounding area. The midden in and round the hearth was generally very dark and compact and showed no definable layering, except for becoming even more compact towards the base.

were

Samples for archaeomagnetic dating were taken round the pit edge but the results were not conclusive (see “Dating’).

A small paved area was revealed in Trench 25 on the eastern edge of the large Hearth 4 (Fig. 7, 8; Plate 30) . It was buried

The pit contents was excavated in quarters and the last quarter was left intact as a record for posterity. Excavation of the midden contents of the pit proceeded in ‘spits' (stages) of about 11 to 14 cms (roughly 4 - 5 in.) at a time. The contents were found to be fairly homogenous except towards the bottom of the pit where the shells became very concreted indeed and was difficult to excavate with the trowel, being very compact.

under about .4m of compact midden and was, therefore at the

base of the midden, at a fractionally higher level than the hearth which was contained within a depression. The paving consisted of a single layer of flat limestone slabs laid down closely together on top of some shallow midden material and clay (cf Layer 12 of Trench 4) to cover an area of approximately 1.6 metres (5ft) | square; it was not clear whether it continued under the baulk (lynchet) separating Trench 25 from Trench 21 as it was not possible to remove the lynchet here. One of the flat stones of the pavement was found slid into the adjacent hearth. Two picks lay on top of the pavement.

Approximately two-thirds of a metre down into the midden material, a large limestone slab, roughly 30 cms x 30cms (lft square) was found lying on the shell debris and covered completely by it. Several large slabs of limestone, of approximately the same dimensions, were also found lying flat at the bottom of the pit. Two of the stones proved to be part of one stone originally. They were all covered by several centimetres of very hard concreted midden material, which was also between some of them. Directly under the stones and partly between some of them, was a stiff pale grey clay, mixed with crushed shells, giving the impression that at least some of the stones had partly sunk into the clay.

THE PIT A circular pit with a diameter of about a metre and a similiar depth Gust over 3 ft) was found in Trench 2, approximately 6.5 metres (23 ft) east of the eastern edge of the floor (Figs. 19, 20; Plates 26 - 29). The midden becomes decreasingly thin in all the trenches east and south away from the pit, but to the north, only 3m (about 10ft) away in Trench 40 under the lynchet the midden was much thicker. The postMesolithic layers above the pit were also very thin.

At all depths within the midden fill of the pit large quantities of Mesolithic

Mesolithic

artefacts,

which

artefacts

were

found,

e.g.

microliths,

picks,

shell beads and a number of pieces of bone, some of which were possible worked bone points (Plate 57 and see 1.7. artefact analysis). A charred hazelnut was also found. Pieces of red clay from round the top edge of the pit were also found in the midden. Otherwise the midden consisted of the usual debris of molluscs and burnt and unburnt stones.

This feature is just outside the base of the medieval negative (downslope) lynchet (1.3. ‘Stratigraphy’ and Keeley Appendix). The relict soil level (Layer 4) was present, but only thinly, above this feature. A deposit of about 15 to 17 cms (6-7 in.) of yellow-brown clay loam mixed with broken and whole shells (cf Layer 5, Trench 4 and Keeley's sample VII) formed the base of the post-Mesolithic layers. This layer contained

one of the slabs was

right at the edge of the pit and slightly tilting into it.

Small Pavement by Hearth 4

also

several big slabs of limestone;

Charcoal was seen to be present round the top edge inside it at all levels and also at the base. Tests were carried out for charcoal round the edge of various points, and round the stones at the bottom to be positive.

increased

downwards towards the interface between this layer and the next lower layer, i.e. the midden proper (cf Layer 6, Trench 4), which covered the top of the pit and the whole area round it. Although the midden layer was not very thick, only about 6-7 cms., it was very compact, particularly at the base of the layer. Round the pit the midden material lay directly on top of the yellow rubbly loam which was referred to in Trench 4 (and elsewhere) as Layer 13, ie. the original landsurfae.

Munsell folows:

Midden brown.

Soil chart readings

plus loam

round

(wet)

the pit:

of the pit, with HCL the pit, at and found

for the pit, were

1OYR

44/50

as

yellowish

Natural loam from under midden level, north-east corner of

During excavation, the presence of the top of the pit was indicated by a dark grey circle within the otherwise paler grey of the midden. It was found that the pit had been dug through the deposit of midden but it continued down into the solid stoney yellow loam of the natural bedrock below. The feature was full to the top with very dark, very compact midden material and many artefacts.

surrounding

deposits

of Trench

2:

10

YR

49/56

yellow

brown.

Clay round the top edge: 10 YR 35/3 darkbrown to SYR 3/35 reddish brown with pieces of 2.5YR 3/5 dark red and 2.5YR 4/6 red. Side of pit, loam, varies in patches from 2.5Y 7/4 pale yellow to about : 2.5Y 55/40 light olive brown.

All round the immediate top outer edge of the pit only was a thin band of brown clay which was slightly raised above the level of the adjacent midden and appeared to have been

Midden material in pit 10 YR 3/1.5 very dark grey to 10 YR3/2_ very dark greyish brown near the bottom of the pit.

25

diameter of 32 cms. north to south and 40 cms east to west. Immediately under the big stone and within the space between the packed stones forming the circle was a layer of dark prismatic yellow brown loamy clay, for a depth of 2 to

Clay between stones at bottom (Spit 8) 55/1 grey.

STONES

X, Y AND Z

2.5

Stones X, Y and Z were found in the vicinity of a point in the north of the floor where the alignment of the floor’s sides slightly changes direction (Figs. 3, 6(4 & 5), 16,17,18; Plates 32 -37). At this point the change is associated with a number of bigger than average slabs of limestone which are roughly triangular or rectangular in shape as opposed to the majority of smaller slabs of irregular shape forming the floor. It is not certain whether the three stones constitute three separate features or whether they all form part of one feature with special significance (see discussion below); they will therefore, initially, be described separately.

shell midden

midden

been

section

1.7.

on

‘Archaeological

damaged

but are smooth,

indcating that this too was

under the shell midden, was a dark brown

Stones Y and Z:

In 1975 extended the floor negative

the trench in which Feature X was found, was and the whole of the lynchet at this point on top of was excavated. | Under the lower edge of the lynchet, also under a spread of midden on top of the

floor, another stone was found which was recorded as Stone

Y. This stone is 0.75 metres (2ft 6in)) south from Feature X on the floor (see Fig. 6 (4.)) It is also adjacent the double line of stones on the floor and the small square stoneless gap of about a metre’ in the edge of the floor (see above for Floor description). Between Stones X and Y the direction of the floor changed slightly eastward.

Stone Y was quadrilateral and had three sides of roughly 21 cms and one of 30 cms so that the point formed by the longer sides pointed south-east (Fig. 17). Although it protruded approximately 10 to 12 cms above the level of the floor, it was so firmly embedded between the stones of the floor that

space between them in the middle; these stones were slanting

diagonally into the midden below at an angle; they were closely spaced, either touching or with only small gaps with

also

clay loam like that of Layer 12 of Trench 4. A few more of the stones of the floor round the feature were lifted and they too were all found to lie directly on the midden. There were no signs of previous disturbance of the feature or floor round it.

After completing the recording work, the triangular stone was lifted. It was found to be so well embedded into the floor that two or three of the floor stones had to be moved. Underneath it several closely packed overlapping stone slabs were visible, forming a roughly circular feature with an open

area

36,

6 - 7.5cms thick;

midden material round the stone fell into a narrow gap formed by the uneven edge of the stone and the floor at the south-east corner of it.

oval

( Plate

The midden round and under the three items was found to be

it was noticed that some of the

a roughly

in the circular hole), the first item to become

also an old ornament which had been in use. The edges of the hole pierced into the shell near the hinge end are also smooth.

against the edge of the triangular stone on the southeasternside in such a way that it appeared as though it was pointing into one of the sides of the triangle (see photographs). The stones of the floor fitted tightly round both the stones of this feature, i.e. 1t appeared as though it had intentionally and carefully been built into the floor.

define

circular

One of the surfaces of the pebble found in the hole is not completely flat, the two parts of this one face forming a slight angle to each other. The more protruding part of this face is much smoother and also has a sheen to it, as though it had been used for some or other grinding or rubbing activity. This means that the pebble was one which had been in use and not a new one brought to the site shortly before being included in the feature. The edges of the scallop shell has

a large oval cobble, 30 cms long, was found lying

i.e. to

this

material’): aflake axe of chert, 11 cms long, lying flat in a south-east corner between stones right by the large oval pebble; and a smooth flat oval limestone pebble, also 11 cms long, in the north-east corner, of which only the edge first showed up as it was planted into the midden material so that it stood vertically on its edge (i.e. not lying on either of its flat surfaces).

After the base of the triangular stone had been cleared of

them,

round

visible was a pierced complete scallop shell (Pecten sp.), 7cms diameter, in the north-west corner lying in the midden material so that its hinge pointed in the northerly direction, Gradually two more items became visible through the

Below the thick deposits of Layers 1 to 4, a large triangular stone was uncovered in the north-west corner of Sondage 3 (Area A) approximately 62 cms from the edge of the floor at that point and forming an integral part of the floor, with its top protruding partly into Layer 4 (Fig. 17). It was recorded as Stone X (forming Feature X together with its associated finds, to be described). The three sides of the stone measured roughly 30 - 34 - 37 cms and it was 18 cms thick. The floor appeared intact and undisturbed at this point. There was a spread of approximately 5 - 10cms of shell midden on top of some areas of the floor round the big stone.

between

lots of shell fragments;

After excavating about 3 cms of this midden (i.e. within the

Stone X

While excavating the feature,

with

patch of clay was the usual midden material which occurs everywhere under the floor. Below the yellow clay was a compact layer of midden material.

Attention was first drawn to this aspect of the floor in 1974 when a small trench was cut through the negative slope of the medieval lynchet up to the apex of the lynchet on top of the floor. At this point there was about a metre of undisturbed deposits on top of the floor; the soil profile was complete and included the ancient buried soil horizon of Layer 4, as discussed above ( 1.3. Stratigraphy’; see also Keeley in Appendix. ). There was also a spread of midden on top of the floor at this point, varying in depth from 10 12cms in depth.

midden,

cms.,

a

26

again several smaller slabs had to be removed in order to

depressions, with their bases supported on the stones round the circumference of the holes. The holes did not contain anything of significance.

enable the big stone to be lifted, i.e. the floor was laid down

partly over Stone Y. When

the big

was revealed, material,

as

stone was

lifted, a small

FEATURE 41: TRENCH 41

stone-lined feature

also full of clay but surrounded by midden was

the

case

with

Feature

X.

This

The stratigraphy of the deposits within this trench has been more fully described earlier in this Report (1.3. ‘Stratigraphy’; see also description/discussion of the gully).

hole

contained nothing unusual and the only finds were a blade segment,

6

small

flakes,

1 large

flake

and

2 nodules

of

natural limestone (see Discussion below).

This

revealed a

section right at the edge of the deep gully on the side of the site leading to the Culverwell stream.

Stone Z is 0.60 metres (24 in) due west from Feature X, i.e.Feature X and Stones Y and Z form a triangle to each other. This stone was also under the thickest part of the lynchet and could therefore also be seen to be undisturbed.

In the centre of this trench (Quadrats 3 and 4 and partly into Q1 and 2) an arrangement of stones was encountered with the top of the stones covered by the base of Layer 8 midden but its base within the top of Layer 9, i.e. at the interface between the two layers (Figs.26; Plate 39). From subsequent work in this area, it appeared that this feature was within the gully and associated with the earliest occupation debris thrown into the gully, midden layers 10-11 being absent here. The profile of the deposits in the adjacent (southeasterly) trench section, also appeared to indicate that it was also at or near the periphery of a small heap of midden rubbish which had formed here (Fig. 25; Plate 40). When the mound of rubbish began to slump down, it would have covered the feature.

Stone Z is also foursided, with two sides of about 26 cms and

two sides of about 52/53 cms (10 and a half and 19 in.) thus also forming a point. The stone was lying so that its point was in the direction of Feature X and Stone Y. It was also partly underneath stones of the floor but not as tightly packed in as the two features described above.

A stone-lined hole was also found under Stone Z when it was lifted. The stones lining the hole, were again nearvertcally placed. When the clay in the centre of this feature was excavated, it was found to be 14 to 15 cms (5 - 6in.) deep and also surrounded by midden material. Again nothing unusual (for this site) was found, the only finds being a small single-platform micro-core, 5 waste pieces of chert and 6 flakes (one inside a shell).

STONES W, 3X,

trench, as far as was possible to ascertain,

At the northern end of the feature lay a large smooth beachrolled cobble, roughly oval or kidney-shaped, lying with its longest axis in a direction of approximately SSE to NNW and measuring approximately 31 by 18cms and an average of S5cms thick. It was associated with an oval shaped arrangement of irregularly shaped pieces of limestone of

RANDS

In the south of the site, near the point where Feature 2a and the floor cross each other to form an angle, the eastern side

various sizes,

of the floor slightly changes direction so that it is slightly out of alignment with the other side. Four larger than average stones were revealed here in positions which appeared to be related to the junction between the floor and Feature 2.

Every stage of the work excavating the feature was recorded, drawn and photographed. The stones were numbered prior to liftng and the midden deposits underneath them excavated. It was found that all the stones had one face which was flatter than the other and in every case it was the flat face which had been placed downward on the old midden surface. However, nothing visibly of significance was found under the stones when they were Ifted.

A roughly square flat stone was revealed here on the floor, and was recorded as Stone ‘W’ (Fig. 3; Plate 11). It was .4 x .4m approx in size (16in x 16 in) and about 1.5 meters (approx 5ft) from another equally large stone ‘R’ and approximately 1.5. metres north north-east of another large stone ‘3X’ of about the same dimensions,

In the SSE

stretching

corner

towards the south.

of the trench,

1.8m

away

from

this

feature, another large beach cobble of similar dimensions was subsequently found lying partly into the top of the clayey-loamy deposit of Layer 12 and partly on it. This cobble was tilted at an angle, i.e. it was thrown into the gully at atime when the bottom was muddy. This cobble was not associated with any other specific features or finds.

so that these three

stones appear to relate to a cluster of features in this area (Fig 3). The midden deposit here is quite thin (12 to 15cms), some of it being on top of the floor and some below the stones of the floor. Stone ‘S’ was a thick triangular stone, with the longest side measuring approximately 40cms long, and lying near the junction of the floor with Feature 2a and 2b (Plate 38).

The finds from the midden deposits at all levels in this Trench were (as elsewhere) predominantly of Mesolithic character, although a number of artefacts were found with a Late palaeolithic aspect and a few very tiny pieces of pottery were also recovered. A quantity of pieces of bone, some very small but some sizeable and more so than from

The circumference of these stones lay directly on top of the floor but some shells were found underneath them and were also lodged in irregular parts of the stones. The midden areas round these stones were carefully excavated and they were then lifted. | Underneath them were revealed small depressions in the midden of about 8 cms depth, going into

elsewhere on the site,

were found in the midden

deposits,

some at the same level as the Feature 41 (Thomas this Report 2.11.). Several big pieces of ochre, including one shaped ‘crayon’, also came from this trench. These aspects are further discussed below (1.7. ‘Archaeological material’).

the yellow loam underneath, and about 13 cms diameter with

the circumferences of the holes marked by diagonally placed stones. The big stones were neatly placed to cover the

27

FEATURES IN TRENCH 4

WORK AREAS: TRENCH 40

Within the lowest layer of midden in Trench 4 (Layer 11) three clusters of limestone slabs were found, all lying west of the Floor’s edge on the sloping surface leading down to the gully’ of Trench 31. They were recorded as Features A2, A3 and B2 according to the sections of the trench they were found in. All three had the appearance of having been placed intentionally, i.e. not being random piles of stone thrown into the midden.

In Trench 40 and Baulk 40 as an experiment in a special project excavation involved three-dimensional recording of the position of all finds and features within the midden level (Figs.2,

3, 23, 24; Plate 50 ).

This was

time consuming,

because of the vast quantity of artefacts, and the project was not repeated in other trenches. However, in all trenches this method was followed for anything which at the time was regarded as being of potential significance. The dimensions of Trench 40 were 2m by 4.8m and _ its southern edge was only 3m from the southeastern side of the pit. The trench was divided into smaller quadrats to facilitate recording . The westernmost edge of the trench was 4m from the edge of the limestone floor and it was, therefore, approximately centred between these two features.

Feature A2 consisted of only 3 stones placed together to form roughly a triangular feature. All three stones were nearly vertically planted into the midden and underlying yellow clayey loam. Feature A3 consisted of tabular limestone slabs placed to form a roughly circular feature with a circumference of about .6m (2ft). Some of these stones too had been vertically planted into the midden and underlying loam while others were horizontally placed on the loam.

The trench was well stratified below the post-Mesolithic occupation layers and protected by the medieval lynchet, as discussed in the section on the site stratigraphy. During the 1987 excavation season the top of the shell midden (referred to as Layer 6), was exposed over the whole of the trench and a large number of artefacts appeared to be concentrated in a number of clusters, some artefacts overlying or abutting each other. This led to the decision to excavate and record the midden layer in a detailed manner in order to determine whether there was any evidence of a reasonably intact industrial activity surface.

Feature B2 was similar to the two features described above but had about 16 slabs of stones in it, some also near vertical within the midden (Plate 8). The usual artefacts, burnt stones and shell debris were associated with the features, and no

archaeological material of an unusual nature was found.

POST-HOLES One definite post-hole was found in Trench 2 on the edge of the pit at the same level as the top of the pit. It had a diameter of about 5cms (2in) and penetrated from a midden layer into the underlying hard yellow loam round the pit. It was therefore not very substantial.

north-west of the edge of the floor, i.e. at a point where the

The trench was sub-divided into 18 quadrants and the midden of Layer 6 excavated in Spits of, as near as possible, 2cms in depth. All artefacts were recorded by measurements taken from the Datum Post and three areas of particular concentrations were noticed. Four conjoined flakes were found in quadrant C2, 3 being from Spit 1 and one from Spit 2; subsequently, in looking through the finds of the previous Layer 5, two further refits were found belonging to the same group, i.e. indicating a post-depositional upward movement of aproximately 6cms as well as some downward movement (Plate 56). The implications of the findings from this trench are further discussed below.

negative lynchet of the field system covered it. Unfortunately, an animal burrow could be seen also to have penetrated from layers above the midden down into the midden at this point and it could not be ascertained whether the post-hole was a feature relevant only to the midden layer. A Beaker-type of barbed-and-tanged arrowhead was found at the base of this hole.

Baulk 40 was almost immediately adjacent the eastern edge of the limestone floor in Area A and also less than two metres from the northern edge of Hearth 1. It did not produce any evidence which can be considered as indicative of an in situ’ working or industrial area (see also discussion 3.13. ‘Activity zones’).

Another post-hole was found in Trench 1 within clay on the northern edge of Hearth 1 (see above). This too was of approximately the same measurements but appeared to penetrate the clay at a slight angle. A post-hole

was

found

within the midden

in

Trench

4,

FEATURES OF UNCERTAIN

Two post-holes were found in the northern half of Trench 32 within the gully, west of the floor (Plates 16, 17). The westernmost one penetrated through the shell-midden from a point within the shell deposit right down into the underlying rubbly loam whereas the easternmost one only penetrated through about 12cms (Sin) of midden. Both had a diameter of about 5cms (2in). About 2m further south of these two holes another one was found of approximately the same size, also penetrating through the midden into the natural rubbly loam of the gully.

AGE AND NATURE

Trench 41: Towards the southern part of this Trench a series of 8 striations or shallow furrows were found going into the top of Layer 6 midden down from the base of Layer 4 (Diagram below and Plate 42). (Layer 5 was not noticeable in this Trench). Although the furrows were directly into the midden, their infill was clay, reflecting the fact that they were made into the midden at the junction of the two layers. They ran roughly along a NNE to SSW axis. There were shallow narrow ridges of irregular dimensions between the furrows.

28

( ‘ { (

fans

-

pen

ae

el

eS

ie

eS

a

ey,

f

Scale 1:10

2.160m

Diagram of marks on top of the midden in Trench 41.

parallel with the lynchet. Nothing of it remained except for yellow-brown lines of discolouration on the lighter loam; these marks were completely on the level, i.e. not on a downward slope (see discussion of this point below). Nothing of the walls were found but it is almost certain that the marks were left by a wooden structure; stone walls would not have left pronounced discolouration on the soil and it is most likely that beams laid down horizontally flat on the (then) soil surface would have left exactly the type of marks which were encountered. Four postholes were associated with the structure: three on the southern side and one actually

The furrows varied in length and their depths from the top of the levels between them were as follows (cms):

2.5

6. 1.0

7. 4.0

8. n/a

There were no furrows present anywhere else in the trench and they did not continue up to neither the northernmost or southernmost edge of the trench. There were no further indications of the furrows as_ excavation proceeded downward in the midden deposits. This feature is further

within the structure.

found layer There layers

discussed below.

POST-MESOLITHIC

FEATURES

AND FINDS

The most substantial evidence for the post-Mesolithic usage of the site is, of course, the medieval strip lynchets. Ploughing in medieval times and until fairly recently would have been with ploughs which did not do the same in-depth damage to archaeological remains as what is done by present-day equipment. The lynchets and their effect on the stratigraphy of the site have already been fully discussed (1.3. Stratigraphy).

excavated, these postholes were

to be very shallow indeed, penetrating only through 2 and not reaching down into any of the lower layers. were no signs that this structure had disturbed earlier or any of the Mesolithic features below it.

Finds found with and around this feature included a small number of sherds of thirteenth century green glazed types but also quite a lot of much later, e.g. eighteenth and nineteenth century as well as Victorian. The dating is therefore not conclusive and the interpretation of this feature in relationship to the rest of the site, will be discussed

below.

At Culverwell sherds of late medieval or seventeenth century pottery and much more recent material, including Victorian and Edwardian ceramic fragments, pipe stems, etc. are numerous in the top plough soil and, to a lesser extent, in the lighter yellow-brown clay loam of Layer 2 (see chapter on Stratigraphy). Very few of these recent items occur in Layer 3 and only isolated specimens are found in Layer 4 and below, except in areas where there appears to have been some more recent disturbance of early deposits.

Marks left by the foundation walls or beams of some or other elementary field or agriculatural building, possibly medieval or more recent, were found near the edge of the positive lynchet of Field 2153, on the top of Layer 2 yellow-brown clay loam during the excavation of Trench 19 in that upper field. This structure measured 5.8m by 2.3m (19ft 2in x 7ft 8in) and was divided into four sections;

When

the longest side was

29

A number of very minute pieces of red clay have been found mixed in with the midden material in the pit down to approximately Spit 3 (roughly a depth of 30 to 32 cms); the majority of these appeared to be fragments from the red clay round the edge of the pit but some pieces were identifiable as pottery ( Section 1.7. this Report). They could have been Iron Age, but this is not certain as the pieces are too tiny to make a positive identification. A number of similar tiny pieces of pottery have been found in several trenches in the midden where it was associated with a vast quantity of typical Mesolithic artefacts. The sherds are all dark brown to dark grey and could be either Iron Age or late Neolithic to early Bronze Age. No sherds have ever been found which could be conjoined

and, in number

or size, not enough

of

them have been found to make even one small pot. When floor,

the northern section face of Trench 4, west of the was being excavated, a post hole, tree-fall hole or

animal burrow was found to have penetrated from Layer 3 down into the midden to the top of Layer 6. Near the base of the hole a barb-and-tanged arrowhead of Beaker period type was found. No other evidence for activity of this period was found round this feature. Another similar arrowhead was found in Layer 3 during our excavations in Area D near the Bill Road. In Trench 31, about 5m from the point where the arrowhead came from (see above), a tiny piece of gold wire, just under a centimetre long, was found in Layer 5 as well as three tiny pieces of possible Bronze Age pottery. Two other very small pieces of pottery were found much lower down in the midden of the same trench. A number of small pieces of pottery were found in Trench 41.

Throughout the site a number of animal burrows have been noticed, starting from the various post-mesolithic levels downwards. Nowhere were they so extensive that the Stratigraphy of the site had become confused or extensively disturbed (see 1.3. ‘Stratigraphy’).

30

1.5. ASSESSMENT

OF

ALL

FEATURES sea and also the nearby spring. This appears to have been the place used systematically as the dumping ground during the site occupation, or at least during part of the occupation. This gully and thick midden starts near the north-eastern limits of the floor and continues in a south-westerly direction and is partly under the floor, which was later put down on it when the terrain became somewhat levelled off by the infilling of the lower areas with rubbish.

(For description of all features see above; refer to plans and photographs throughout)

MIDDEN - GENERAL

ASSESSMENT

The midden at Culverwell has formed the subject of several graduate and post-graduate studies and summaries of these are included in this Report (see 2.11. Thomas). These give a very detailed analysis of the various constituents and characteristics of the midden.

After the floor was created on the crest of the midden, the main

bulk of rubbish apparently continued to be cleared from the floor and dumped

The association of middens with Mesolithic coastal dwellers has for many years been extensively investigated in Denmark, where they are referred to as K6kkenmidden", i.e kitchen

to this

veritable

‘shell midden

culture’,

subject which provides endless opportunities for projects. Middens provide excellent opportunities for the environment, economy and living styles of inhabitants and these aspects at Culverwell warrant discussion in this Report.

This does raise an intriguing question: which factors at Culverwell determined when debris should be scooped up and thrown into the hollow and when was the rubbish to be left in situ where produced? Was this task just haphazardly carried out or was it organised? Maybe determined by the weather or the urgency of other activities, such as food gathering and preparation? Although impossible to answer, it may be interesting to consider some of the aspects of the situation regarding dumping.

a

The accumulation of midden material must have commenced during the early phases of the site occupation. It is not possible to argue whether the inhabitants from the beginning had the intention of filling up the gully so that they could utilize it for laying down a floor or pavement on it and so producing a welldrained living area or whether this is an idea which occurred to them later on. (This question is further discussed below in

research studying the site separate

It is likely that the large midden at Culverwell originated as several mounds of debris placed over the site; these mounds became linked up as a result of the human activity during the site occupation (e.g. walking on the peaks of mounds) and also by virtue of additional dumping of rubbish in the gaps between

relationship to the

made only to ‘a midden’ with regard to the present appearance of the feature. It must also be accepted that the midden was probably higher in some places than it is now. It also has to be accepted that even during the site occupation some shell debris may have become re-distributed beyond the confines of the main activity areas as a result of the activities of the wild animals

and also the elements.

mean that nowhere would absolutely homogenous.

the

contents

of the

This

Either way, this could mean

that

It is of interest to note here that the most compact hard layers of midden material do occur in the gully whereas this compactness is not so much in evidence in the large level activity areas outside the floor (Fig. 5). This effect could, however, be due to damp or other natural events within the low-lying feature. In order to elucidate this problem, it would be necessary to take extensive samples of the midden from

would

midden

Floor).

some rubbish was regularly thrown or pushed into the gully from adjacent areas until it was full and some areas left intact. This model could presuppose that most of the earliest debris from the site is in the bottom of the gully, but this may not be strictly the case as different activity zones may have existed on the site right from the beginning of occupation. Middens could have lain on the surface for lengthy periods of time before being pushed into the gully.

the small mounds. For the sake of this Report, reference will be

inhabitants,

outside the western

Meehan (1982: 114-117) points out that the Anbarra people of Australia had regular clean-outs of the camp-site with their main dumping grounds a short distance away from the huts or daily living areas. Nonetheless, the whole area of the camp sites did become covered with shell-debris.

virtually world-wide (Palmer 1977). Shell middens on living sites certainly became a major feature of sites in coastal areas during the Mesolithic period; changes in the environment and the sea levels were probably the main factors responsible for this phenomenon, as it now became possible to live on the shore for much of the year and a good food supply was amply and easily available. Bailey (1978:37) also pointed out the possible relevance of intensification of economic activity and population pressure. Meighan (1969) pointed out that the shell midden phenomenon corresponded with the extinction and decline of large Pleistocene animals, thus encouraging the search for other food resources. The extensive studies on the shell-gathering practices of living coastal tribes in Australia, have made possible comparisons with prehistoric shell-mounds (Meehan 1982). It is obvious that diverse factors in different contributed

from this floor,

edge of it in the area of the gully (e.g. Trenches 31, 32, 4 and 41); other activity areas continued in use east of the floor and round the other hearths and pit.

middens (e.g. Troels-Smith 1964), but they have, in fact, been identified and studied, as being characteristic of this period,

areas

away

be

At Culverwell the thickest part of the midden appears to be that which is within the western part of the occupation zone within a gully which cuts through ground sloping down towards the

every small area of the site, and even this method may not be

31

this part of the gully. However, the abrupt manner in which the gully terminates at its north-eastern limits, adjacent the big Hearth 4 (described above), suggests that in that part of the site the natural feature was possibly to some extent, at least, altered for use as part of a depression to hold a hearth.

conclusive as other factors may have created discrepancies in the midden content or texture. It is immediately apparent that large percentages of the surviving complete or semi-complete shells have not been extensively burnt in an open fire despite the fact that the midden is full of very large quantities of burnt stones, indicating fires (see M. Stewart this Report). It may be possible to explain this contradictory evidence by pointing out that the midden is also full of very large quantities of very minute fragments of shells, some no more than specks; in any case, shells which have been severely burnt by direct fire or intense heat would be so disintegrated (or even pulverized) that they would not be visible as complete or semi-complete specimens in the midden. Also, it can be argued that large quantities of the molluscs must have been prepared for eating by means other than exposure to direct fire. Limpets can be eaten raw as the flesh is easily removed but periwinkles and

This gully would initially have had the effect of diverting muddy water coming down the slope of the hill away from the more level terrain east of it. The accumulation of clay in the bottom of the gully and the species of freshwater molluscs found within this deposit, suggest that this gully was periodically, or for part of each year, rather damp and marshy (see 2.11. Thomas). As the midden deposits in the baulk between Trench 4 and Trench 31, in part, at least, lens of very stiff black clay (Context 31;

‘Stratigraphy’), it supports the view that the deposition of mud was intermittent and also that some mud was periodically deposited during the period of the shell-midden accumulation. The evidence, therefore, suggests that water did not stand in the gully for very long periods and there was no permanent ponding but boggy conditions occurred periodically during the site’s occupation. The molluscs which are present in clay samples are all of extremely fragile species which could only survive in calcareous soils.

topshells have to be broken, roasted or boiled to get the flesh

out of the shell. the

shells

It leaves one alternative: a large percentage of

which

are

unbroken

or

unburnt,

were

boiled,

probably in skin bags strung over a fire. This means that the presence of numerous bits of charcoal within empty shells represents rake-out from the fires which had become lodged between shells after the flesh had been eaten. Meehan (1982: ch. 7) does indicate that the groups of dune-dwellers which she studied in Australia did sometimes eat molluscs raw and, by choice, also used both methods of cooking, i.e. in open direct fires and in water in containers over a fire. The method of boiling food

at Culverwell

may,

therefore have

been

one

The evidence from most parts of the gully appears to indicate that prior to the Mesolithic occupation rubbly loam (Layer 13) and clay (Layer 12) were present in the gully; at least some of this material was probably weathered directly from the side of the gully and the rest brought in by rain. However, there is a general absence of layers or lenses of rubbly and loamy material which could have been weathered out from the sides of the gully into the deposits of the midden. This suggests that there were no long episodes during the Mesolithic occupation when the gully (as far as demonstrated by the excavated areas west of the Floor) lay unused and exposed to the elements for any length of time. An example of the speed with which weathering influences can become noticeable, have been demonstrated at the Overton Down _ experimental earthworks (Jewell and Dimbleby 1966).

of

choice too or brought on by other crcumstanes, e.g. weather.

There is no indication of the circumstances which led to the stacking of molluscs in small heaps, noticeable in the midden. This stacking could have been done either before the molluscs were cooked or afterwards.

GULLIES, As indicated,

HOLLOWS the whole

WITHIN SLOPING TERRAIN habitation

site is on terrain

which

gradually slopes down in both a southerly direction to the sea and also in a westerly direction to the stream of Culver Well. These slopes are purely natural contours and relate to the proximity of the sea and the strream; the latter may have been nearer the habitation site during its occupation as suggested by the presence of the thick, very stiff clay encountered in Trench

However, the stratigraphical evidence from Trenches 31 and 41 (see ‘Stratigraphy’) indicates that at the westernmost limits of the gully, the midden debris was deposited in a more sporadic fashion. The clay lens (‘Context 31’) in Trench 31 indicates that the deposition of midden material here was briefly interrupted by the deposition or arrival of mud; the interruption could have been of a very brief duration only, possibly amounting to even a single short wet episode when only enough mud moved down the hill to accumulate in a low spot only, bypassing the floor and the tops of any higher adjacent mounds of occupation debris.

41 on the western edge of the site, as described above.

The large gully No certain indications have been found in the areas where it was excavated as to whether this feature was in its entirety natural in origin. The gradual sloping western edge of the feature, visible in most of Trench 4, Trench 31

contain a see 1.3.

In Trench 41 the evidence indicates that midden debris (Layer 9 with a fairly stiff clay content) was deposited in the whole of the gully bottom and the stone Feature 41 was erected at that level; more midden (representing the lower part of the compact Layer 8 over the whole of this trench area) was deposited for a while and then this particular dumping spot was not used again but immediately adjacent it a mound of

and Trench 41,

almost certainly rules out all arguments that it was extensively made or modified by humans. In Trench 32 a part of the eastern edge was found which was very jagged and consisted of cryoturbated rocks firmly embedded, in a completely random fashion, in the natural loamy deposits beneath the occupation debris; this too indicates a natural origin for at least

32

rubbish continued to develop. The final stages of filling along the westernmost limits of the gully in this spot (only as far as known, shown by excavation) did not occur until later in the site occupation or, possibly, even shortly after the site was deserted and natural agencies caused the adjacent midden material to slump and _ level off. (In part of the trench, it may therefore be a situation of inverted stratigraphy, although it is not very apparent in view of the fact that the predominant material everywhere in the trench is midden material of the

having some religious significance for the site inhabitants, with regards to the gully, cannot be overlooked (see also discussion this chapter under 'Ritual'). The finding of two conjoining flakes right at the bottom of the gully in Trench 41 indicates that at least some of the deposits within this feature has not been substantially moved since they were deposited.

same nature as elsewhere, except for the fact that the

There were no certain indications of what the clusters of stones, found in the gully in Trench 4 adjacent the western edge of the Floor could have been used for even though they had the appearance of not being fortuitous in origin (Plate 8). In Trench 32 a similiar cluster of stones surrounded a sub-

midden Layer concentrated).

8 here

is not

very

compact

and

not

Features in the gully of Trenches 4 and 32

top of

very

It is unlikely, however, that this part of the gully was open for very long after the site occupation as the deposits within it at the south-western edge does not indicate that more weathering of the sides of the gully took place here than it did elsewhere. Even though some scant Neolithic or later archaeological evidence have been found on the site (also in this trench, see below and see ‘Archaeological material’ and “Dating’), the deposits themselves and their contents are not adequately different from those over the rest of the site to postulate a later prehistoric date for the backfilling of the western half of the gully. It does, however, seem likely that this edge of the gully was more exposed to environmental conditions than the opposite (easternmost) parts of the gully, at least partly accounting for the thick deposit of very stiff black clay found covering the edge or ‘lip’ of the gully on this side. It is possible that some of this clay pre-dated the Mesolithic occupation.

circular small hollow, excavated into the natural clay.

Structurally the features resemble small hearths but no more charcoal were found in them than is present throughout the midden in general. There were no signs of fire. If these stone clusters do represent hearths and food were placed on the stones, then it would appear that they were not extensively used - perhaps even used on a single occasion only for casual cooking. If this is the correct interpretation, it would mean that the cooking places on the site were not exclusively confined to the areas east of the floor,

even though

there is

strong evidence that the main intensively-used hearths were confined to positions adjacent the easterly edge of the floor. However,

Feature A2 of only 3 stones in Trench 4 is less well-

defined than the other stone arrangements of Feature A3 (circular diameter about .6m) and B2 of 16 stones, both in Trench 4, as well as the feature of

stones in Trench 32, also

forming a rough circle. A3 and B2 may have been for the same purpose, but it is possible that A2 may be unrelated. There is,

The terrain near the edge of the gully would have been an ideal area for encampment when the Msolithic people had first arrived on the site as they could throw their rubbish into it. As it became full of midden debris, rainwater would have quickly filtered down through the debris in it and then would have drained away from any structures or living areas by the gully. When the floor was placed on top of the midden, straddling the gully, an ideal drainage system came into force to keep any activity areas or structures on top of the floor dry. This gully therefore served two purposes: a convenient dumping ground for rubbish and also as a means to enhance the drainage of the site.

however, no strong evidence

for their interpretion.

It is possible that the features marked positions where packed stones had supported posts. In that case these stone-circled features could be seen in the context of nearby post-holes found on the site. However,

the diameter of these features are

too large for the size of posts which could be reasonably expected on the site and they may not be the remains of actual post-holes but me rely something which was associated with posts.

In Trench 32 right on the edge of the gully, one definite welldefined post-hole was found within the midden deposits as well as one possible post-hole, i.e. the circular stone arrangements are here associated with at least two possible post-holes (Plates 16, 17). In this case, the posts must have been quite substantial, with an estimated diameter of at least 20cms (8 in) each and could have been associated with some or other shelter. This would be compatible with the fact that these features are in approximately a straight line adjacent the westernmost edge of the limestone floor. The walls of any structure in this position would have given some shelter against wind for people sitting on the floor. Whatever, these stone circles do suggest that some activities did also take place on the

The finding of concentrations of a relatively large number (i.e. relative to finds elsewhere on the site) of fairly flat, oval or round beach pebbles lying together within the top few centimetres of the mud in the gully (Trenches 4, 31, 32 and 41;

Plate 15, 17) could raise some interesting questions. Natural and unstruck whole pebbles like these are commonly found within the debris of the midden over the whole site but it is only in the gully (as far as excavated) where obvious concentrations of them have been found. The simple explanation could be that they were brought from the beach for practical purposes and were thrown into the gully as a matter of convenience when they were no longer required for their original purpose. However, even though it is not possible to formulate hypotheses about this particular situation ( and also as there may be other concentrations within the unexcavated areas under the floor), the possibility of round/oval pebbles

western, gully, In Trench

33

side of the floor.

4 one

small

post-hole,

marked

by

a discoloured

circle, was found about 100cms_ north of a small stone arrangement. As it contained a post-Mesolithic type arrowhead, it is possible that that hole could have been a later intrusive feature, even though it appeared to start from the top of the midden deposits. However, it was immediately adjacent an animal burrow which penetrated (visibly) through the soil of the medieval lynchet. It is therefore also possible that the posthole could have been Mesolithic but that the arrowhead was displaced from higher levels into the post-hole by the

site? As itis, the feature of laid-down limestone slabs, appears to terminate in the northern area of the excavations in Field 2153 (as far as could be ascertained by work, see above) and there was also there a noticeable decrease in the presence of other evidence for occupation.

burrowing animals. However, it’s relationship to the rest of the

another layer of different nature intervening between midden and stones (see 1.3. ‘Stratigraphy’). 11.The greater part of the floor lies on top of Layer 8 of the midden as in Trenches 4 and 6 where it can be seen to seal off that layer. ii1. midden material as in Layers 6 and 7 (Trench 4) which is definitely Mesolithic were deposited on top of part of the floor, adjacent the floor and all over the site after the floor was laid down. iv. There is midden material and Mesolithic artefacts amongst the stones of the floor and in some places the stones of this feature are covered with a deposit of midden. v. both midden and floor are covered in most areas by Layer 5 (as pointed out by

(b) A feature of post-Mesolithic date: i. The deliberately laid down stone floor or paving lies directly on top of the indisputable Mesolithic

Mesolithic features must be regarded as dubious.

Depression

and gullies in the south of the site

No satisfactory interpretation of the series of small gullies, south of Feature 2, can be given as they could unfortunately not be fully explored due to the proximity of the road. These smaller narrow gullies were found within sloping ground here and they here have the appearance of having been at least partly artificially made or modified. These gullies could have been made or used for the purpose of diverting water away from the site area limited by the wall of Feature 2. It would add to the hypotheses that the whole structure of the site was well planned.

THE LIMESTONE FLOOR

Keeley),

in turn is below a

in most

areas,

without

relict soil of yellow

loam

(Layer 4, see stratigraphy); vi. the main hearths are closely aligned to the eastern edge of the stone floor and a direct relationship therefore appears to be clearly indicated; vii. the large triangular stone or 'ritual' feature with the three items of definite Mesolithic nature (see Feature X description), is an integral structural part of the floor and must therefore beyond any doubt be contemporary. viii. Even though the postMesolithic layers are not fully represented in all parts of the site, all these deposits would have been present anciently over much, or all, of the site prior to their disturbance by medieval ploughing; this is demonstrated by the fact that the northern soil profile of Area A under the lynchet is complete where a soil column has been studied (see 1.3. ‘Stratigraphy. and Keeley, Appendix). ix. the industry associated with the floor is the same as that associated with the well-dated midden, pit and the other features on the site and Mesolithic artefacts are often found in direct contact with the slabs of stones, between them

OR PLATFORM

Alternative Explanations In view of the unusual nature and large dimensions

which

shell-midden

of the

horizontal limestone feature, described above, other explanations for it have been considered, other than those as a

‘floor’ or ‘platform’. A collapsed wall: The stones of a collapsing wall, of whatever nature, would not fall neatly adjacent each other in a horizontal fashion but would be mostly scattered and the base of the wall would retain a linear aspect (see e.g. description of Feature 2, this Report). The stones would not fall to form a feature with relatively straight edges (see e.g. Plates 2, 3, 37). The straight edges of the ‘floor’ are at approximately 70° angle to the ploughing direction of the medieval lynchets and it cannot therefore be argued that the regular appearance of the feature could have been the result of plough spread at a much later date. A collapsed wall would have resulted in a spread along a roughly SSW to NNE line.

and on top of them; x. the artefacts under the floor, on top of

the floor and adjacent the floor are identical in nature, i.e. of the same Mesolithic character. xi. even though some nondescript and, sometimes, unidentifiable tiny sherds of Neolithic, Bronze or Iron Age character have been found

(a) A road of Mesolithic date: This alternative can be shown to be untenable for the following reasons: i. the ‘road’ would only be approximately 31.09m (approx 102ft) long (there are no visible signs of it south of the Bill Road despite ploughing there throughout the centuries) and the width of 3.66m (12 ft) would therefore be disproportionately large; ii. roads are usually constructed in order to link one structure or feature

scattered haphazardly on the site (see 1. 7. Artefacts discussion), not enough have been found to reconstruct even a quarter of a pot of any one particular period; these later finds are isolated scatter and not associated with any one specific part of the site or feature. There is therefore nothing to suggest any substantial post-Mesolithic occupation of the site other than possibly casual transitory activities such as walking over the terrain previously occupied by the Mesolithic people. The few bits of post-Mesolithic material could have become incorporated into the Mesolithic deposits as a result of soil

with another; in view of the width of such a road, it would be

movement

justifiable to argue that those features, so linked, must have

There is no evidence for substantial medieval activities on the site, other than ploughing, in particular not below Layers 2 an 3. Xiii. The whole site is well-stratified with minimal evidence for post-depositional disturbance.

been themselves substantial and important;

no

such structures

have been revealed by the excavations. Feature 2 in the south (see below) may originally have been significant or important to the site inhabitants,

but even

so, the question

would

from

animals,

humans

and natural

elements.

xii.

still

arise: where did the ‘road’ lead to in the northern part of the

It is reasonable therefore to accept the large spread of stone

34

slabs as being associated with some or other aspects of a habitation site, and as a structure of Mesolithic date in its entirety. Comparisons with similar features on other sites of approximately this period will be made in section 3.18. this Report. Assessment

of evidence

for

reference

to a ‘floor’

that the stone-paved area served also as a dry surface for sitting on while using

stone-paved area was intended primarily to serve these purposes. There would have been no need for such a wide and long platform just for this. In Trench 21 a small paved area was found which definitely served as a place for sitting on immediately adjacent the eastern side of the big hearth which was found in that trench (see discussion below).

rather

than a ‘platform’ (Feature 1)

It is necessary here to explain the preference for the use of the word ‘floor’ rather than ‘platform’ in this Report in connection with the large horisontal limestone structure at Culverwell.

It can, therefore,

In archaeolgy the word is not only used to indicate specifically of a house

or other building,

but

it is also

the

standard terminology to indicate a reasonably defined and level surface on which people lived or worked and carried out activities, e.g. ‘living floor’, ’ knapping floor’, ‘threshing floor’ or

‘work floor’;

be argued that

the feature of

horizontally

placed stones is more reminiscent of a large artificially structured ‘floor’ where a large number of diverse activities associated with living could have taken place and which was fairly central to the habitation site. The length of the feature is coincidental with almost the whole length of the site (as far as known) and it is only raised in some area. The use of the word does not necessarily imply that the structure (in part or as a whole) was used as the actual base or floor of a superstructure or hut or series of huts/dwellings; the evidence for this requires seperate assessment. The size of the floor suggests very strongly that it was an important multi-purpose aspect of the site.

Newell (1980) uses the word ‘platform’ in connection with Mesolithic dwelling structures. It is admitted that equally valid arguments can be put forward for that usage and some specialists may postulate that there are not very strong reasons for preferring the word ‘floor’ at Culverwell. It is, therefore, used here with some reservations.

the floor

these hearths. However, it is unlikely that the

The surface of the feature would, in any case, have served the

purpose of keeping the site residents comfortably away from too much direct contact with the underlying deposits. It would have served the purpose of providing a_ reasonably dry living/working surface with water rapidly draining away beneath it into the shell midden and then into the gully.

this surface can be artificial in origin or

natural, maybe with some minimal modification or attention such as sweeping, and it may or may not have had a covering or superstructure over it. The word ‘platform’ is more narrowly defined and usually evokes the image of a surface, specifically raised or prominent above the surrounding terrain e.g. on which to carry out specific activities. It may be reasonable to argue that the

Evidence for dwelling? The evidence is overwhelmingly strong that Culverwell was a habitation site: the vast quantities of molluscs (food debris), the evidence for fire and cooking, vast quantities of artefacts including tools used for cooking, artefacts (e.g. picks) knapped

limestone structure at Culverwell is, in effect, a raised surface,

although it is on top of the clayey soil in the north, for the rest of its length itis on the midden and therefore raised above the natural original landsurface. As the midden height round the Feature 1 was raised by the additional dumping of debris, the level stone area would have been at the same elevation as the surrounding terrain. The term ‘platform’ would have been ambiguous.

on

the

site,

ochre,

beads

of

shell

and

stone

and

possible

evidence for ritual confirm this status (see 1.7. ‘Archaeological Finds’). It would therefore be reasonable to argue that somewhere,

somehow

on the site there must have been huts,

tents or some other form of dwelling.

On most early prehistoric sites a series of post-holes arranged in an accepted orderly fashion serves as the main evidence for the outline of a hut or dwelling (in the absence of other evidence). Such a precise series of post-holes have not been identified at Culverwell and other definite evidence for dwelling structures is (so far) lacking. It may be possible that complete excavation of the midden under the floor could

In some areas it would not have been substantially higher than adjacent working/living places: where it lies on top of the midden within the gully (where the midden appears to have been levelled off), the top of the stone surface would have been level (in height) with the adjoining levels of the work areas, hearths, etc. In the north of the site the stones of the ‘floor’ are lying level directly on the natural and original soil surface and the surface is only raised the thickness of a slab of limestone above the yellow rubbly loam. In the south of the site it is also hardly raised where the midden under the floor is not very thick.

produce some evidence for further post-holes, but,

if found,

it

could be argued that they either pre-date the floor or that they had penetrated into the midden between stones; their direct relationship to the floor would remain unproven. It is also likely that the evidence for at least some post-holes have disappeared, except for those which have been substantially marked by packing with stones. This could be due to the absence of cohesive qualities in the uncompacted areas or layers of midden: as soon as a post is pulled up, the hole would fill up with midden material of an identical nature so that the erstwhile hole would soon become unrecognisable.

It is likely that the large paved area was at least sometimes used as a special ‘platform’, e.g. the site inhabitants could have used it as a_ surface to stand on when disposing of rubbish into the gully. At Culverwell, a series of hearths have

been found on the east side of the floor and it is highly likely

35

The post-holes in Trench 32 were found in the gully and continued into the lower clayey layers where damper conditions may have prevailed, for at least some days during each year; this could account for their preservation. The one post-hole next to the pit east of the floor and the one next to

the floor.

Evidence for the domestic usage of the floor is indicated by the series of definite hearths which have been revealed immediately adjacent the eastern edge of the floor (see details above and discussions below); this suggests that the floor would have enabled the inhabitants to sit in comfort while using the hearths. It would be reasonable to argue that, during inclement weather or at night, they would have preferred staying within whatever shelters they had, very near the hearths, which so would have provided warmth as well as easily accessible cooking facilities.

the eastern edge of the floor, next to hearth one, also penetrated

into the natural clayey-loam. It is unlikely that the latter two were associated with any substantial structure(s); they may have been posts for hanging a skin bag or food on. It should also be considered that perfectly adequate and sturdy huts or shelters can be constructed without using the technique of planting posts or stakes into the earth and so leaving visible traces of superstructures (c.f. American Indian tepees and Mongolian yurts where the weight of the structures provide the balance to preserve them. This has also been argued in an archaeological context regarding, e.g. Friesheim 2, by Grgn 1995:51. See also this Report on the hypothetical construction of a hut.). On the floor at Culverwell the simple method of constructing a shelter or hut by planting posts into the underlying midden deposits, would not have been practical as the midden would not have provide firm enough footings for poles: there is a constant small element of movement in a midden and poles would have collapsed very quickly unless they were also supported by placing stones round their bases. In time to come, even piles of stones would have collapsed or moved so that the evidence for the use of stones in this manner would not necessarily be recoverable during an archaeological excavation.

However, any reconstruction of what the shelters or huts (if any) on the floor at Culverwell could have looked like, will have to be hypothetical, using the limited evidence cited above and following the models of other known archaeological and ethnographical structures. This is discussed in section 3.16. in this Report. Floor constructed in Phases

It would be logical to argue that the floor would probably have been laid down in phases as the midden deposits in the gully increased in size and became levelled off. It has been pointed out above that it is most unlikely that the whole of the large midden had developed in one single stage in any one place, and it is equally unlikely that the floor on it was laid down all at one time. This argument seems to be supported by the fact that the sides of the floor are not precisely parallel everywhere. The points of these changes are marked by bigger

The best way of securing posts would have been to build a shallow wall of stones against which the posts could have leaned and could have been

secured, thus creating a form

than average stones (Stone X in Area A, the unnumbered stone

at the junction of Area and B and Stone W in Area D, e.g. Plate 37).

of

lean-to structure. It has already (above) been pointed out that a straight line of stones is clearly visible on the floor in Area A at a point which coincides with a neat square gap in the stone placement at the eastern edge of the floor ( see illustration). It is possible that this line of stones on the floor could be the remains of such a wall for propping up poles to hold skins or branches on the windward side of the floor. The gap in the floor may also have served as some sort of an entrance or perhaps curtained off ‘lobby’ to a covered-over structure on the floor at this point.

Although it could be a coincidence that the bigger stones occurred at the points of change only, it would seem unlikely; it is possible that the placing of the big stones were deliberate and could: a. have served some practical purpose, such as ‘markers’ to indicate to the builders where to stop the next phase of construction; or b. the big stones were in the nature of

‘foundation’ or ‘offering’ stones to commemorate or record the beginning of a new stage in the laying down of the floor. These stones were subsequently lifted, as described above, and the contents of the holes below them are further discussed below.

Another situation should be considered: the loose, randomly distributed stones found under the lynchet above the floor (as described above),

Drainage of the Floor

could in time have become dislodged from

the floor by natural agencies if they had originally formed part of some or other super-structure(s) on the floor. If there were, indeed, originally huts, shelters or some other type of structures directly associated with the floor, it is logical to argue that these higher standing structures would have collapsed or disintegrated not too long after the site was abandoned by the Mesolithic inhabitants, leaving only the floor intact. As the loose stones were not found under the lynchet in excavated sections of the lynchet in areas outside the floor (e.g. not in the

As the site was on clayey loam, it inevitably would have become very muddy and uncomfortable during rainy weather and some form of drainage would have been required to make the site habitable. The midden(s) would have been ideal for this, combined with the facts that the site was on a slope and adjacent a gully.

Craig (1966: ch.2) made an extensive study of the living patterns of coastal dwellers in British Honduras. He showed that middens were sometimes deliberately constructed on or

excavation of Trench 40 or Baulk 40), it can be assumed that

they were indeed associated with the floor. This interpretation could also explain the presence of randomly dispersed flat slabs of stones found in the gully adjacent the western edge of

near the beach

order

36

to

from

provide

sand,

drainage

shells,

for

stones

a

shelter

and other debris

which

would

in

be

constructed on top of the midden. It appears to have been most effective. It has been noticed on the Culverwell site after rain, during work on the site,

that the areas with

The thin midden above the floor and partly below Feature 2 in the south would have accumulated fairly rapidly during the site occupation, due to daily activity on the site, and also due to transportation by rain water from the higher parts of the midden in the north of the site. Midden material and clay would in time have infiltrated between the stones of the linear feature, even if the stones of that feature had been directly placed on the floor. The clay round some of the larger stones need not represent a substantial (if at all) time lapse, merely an environmental situation as any mud from higher levels during the site occupation would naturally have accumulated here against the wall. This is particularly relevant as the rest of the living floor was afforded excellent drainage by the thick midden which underlies most of the floor, particularly in the gully areas. It is possible that this clay could represent as little as one wet season (observations on site since excavation of the

thick midden

dries out much quicker than the rest of the site and does not become so muddy. It has also been observed how water from the higher trenches percolates through the midden down to lower trenches. Examples of the utilisation of a midden surface for living on during the Mesolithic period are known from some coastal areas of India (Dr. Goss, University of

Calcutta, pers. comm.

during INQUA conference to the site). Even though the site of Culverwell was not directly on the beach, the need for drainage would still have applied because of the exposed position of the site.

The midden and floor on top of it, therefore, demonstrates

that the site inhabitants had a strong awareness of the problems which faced them and they worked out a relatively ‘advanced’ technological method of adapting to their circumstances.

features has shown

a thick deposit of mud to accumulate). The stratigraphy revealed in Sondage 10 indicates that the clay associated with Feature 2 slopes down by about .5m so that it also lies directly on top of Feature 3 at a lower level (Fig. 13; Plate 12). This indicates that the clay was deposited shortly after the erection of Feature 3 and also after the formation of the midden on top of the floor (Feature 1). This suggests that these features were separated from each other by either a short time gap or else some or other natural event which deposited the mud. The time gap cannot be long as the same mud covers the top of Feature 3 (see below) which therefore must have been exposed at the time of this deposition.

Cleaning of the Floor

It is possible that at least some parts of the floor was deliberately kept clean of refuse, being an activity area in which too much _ rubbish would not have been acceptable. There is no evidence for this hypothesis, but this could perhaps partly explain the paucity of midden material on top of the floor in some areas, even keeping in mind the likelihood of post-occupational displacement of some midden material, as already discussed in the chapter on Stratigraphy.

LINEAR STONE THE SITE

FEATURE(S)

IN

THE

SOUTH

OF There can be several explanations for the structural differences between parts a.,

The stratigraphical relationship between the stone features in the south of the site is very complex and this is, in fact, the only part of the site where there could be possible evidence for structures of more than one period, even though the time gaps cannot be substantial as the features appear to be inter-related and are all associated with identical Mesolithic assemblages (see also Feature 3 below).

one is a linear feature, with for convenience referred

b. and c. of Feature 2 : i. the stones for the

two parts of the feature may have been simultaneously collected from two different sources on the island for reasons of availability or environment or ii. the different types of stones may have been deliberately chosen to be different to act as some form of demarcation between features serving different purposes, with no need for the stones in Feature 2b to be so bulky; iii. The parts could have been constructed at different times when a combination of these factors applied.

These features revealed by excavation in the southern part of the site (see description above) need to be assessed individually but also in relationship to each other as they appear to be inter-related. The main divisions,

that, in fact, a single storm can cause quite

Feature 2a: In view of the more substantial nature of Feature 2a at the westernmost end of Feature 2 and the presence of the double line of stones, it could feasibly be argued that another feature, perhaps even a shelter or hut of stone, was erected here. If the hut or shelter was of a lean-to variety (as perhaps used on the floor - see above and see hypothetical reconstruction 3.16. this Report) then a simple and feasible explanation could be that the double row of stones provided footings or support for the shelter.

three possible subto as Feature 2.

Stratigraphically this feature (with its sub-divisions a,b and c; e.g. Plate 6) is related to the Mesolithic midden deposits all over the site. The artefacts associated everywhere with this feature are typically Mesolithic and with no _ obvious differences from the rest of the assemblage from the site. Featture 2a neatly straddles the floor (Feature 1) and is of exactly the same length as the width of the floor; therefore, whoever erected Feature 2a did so deliberately to take cognisance of the existence of the floor; the floor must therefore have been visible at least in part, or, in the very least,

Features 2b and 2c: These features appear to contemporary with the rest of Feature 2, even features too may have been constructed in stages. 2b is too insubstantial to have supported, or have a structure of considerable substance and was more than a freestanding wall of no great height.

the existence of the floor must have been known to the people who erected Feature 2a.

Feature 2c is more enigmatic as its lay-out and construction

37

be relatively though these Particularly been part of, probably no

thick, even though much more compact than the midden above the floor. There are therefore more than one option in interpreting Feature 3: It was either i. an ‘attachment’ or

suggests some attempt at more precision than in 2b. Unfortunately, it could not be ascertained exactly how it terminated and there is a possibility (not certainty) that it turned in a south-easterly direction, perhaps even forming or incorporating the corner of another structure which could not be researched because of the presence of the modern fence and road.

However,

as the amount

of artefacts

and

‘extension’ to the southern

end

of the Floor;

iit. A detached

feature relatively contemporary to the floor, or else iii. that it is completely separate and pre-dates the Floor in this part of the site, but, almost certainly, not by a long time gap. This lower feature is also clearly associated with relatively thin

midden

began to decrease in this area and were almost completely absent in the next easterly trench (38 Figs. 2,3), there must be a distinct possibility that the apparent end (or turning point) of Feature 2c in reality marks the limits of the actively occupied Mesolithic habitation zone on this side of the site. Only random scatters of material would have encroached beyond this point.

deposits of midden material, above and below it.

In the lower levels of Sondage 10A, cut through Feature 2a (see above and Fig.14 ) evidence for another stone structure was

As the site faces the sea, it must have been very exposed to winds coming into land and it can be assumed that the site inhabitants must have felt the need for some form of windbreak. The most obvious place to have placed such a windbreak would have been at the southernmost perimeter of the main activity zone, precisely where the stone alignment of Feature 2 is. The fact that the combined lengths of Feature 2a, b and 2c continues well beyond the easterly limit of the thick shell distribution (e.g. in Trenches 40 and 2, see below), suggests that this was intentional to keep out the worst effects of the prevalent sea wind, which then possibly (as now) came from due south. In this position Feature 2 would have served to protect the whole site to some degree.

also

revealed

but,

without

demolishing

Feature

2a, it was

not possible to ascertain

between

the different features nor, in fact, was

the

whole

of

the connection it possible to

determine how many features were involved: the stones found in the Sondage could be connected with either the Floor, Feature

2a or Feature

3, another

undefined

lower

feature

or

even possibly with all of these features. The stratigraphy revealed in the sondages excavated through the different features in this area did not enable firm conclusions to be reached, but there appears to be a definite possibility that two (older?) structures are buried beneath Features 1 and 2, i.e. there could be up to four features here one above the other other with intervening thin layers of midden material. The shape of Feature 3, outlined by stones, suggests that a circular or sub-circular hut or shelter or perhaps a storage

Even though Feature 2a was more substantial and perhaps might have supported an actual constructed shelter or hut, the feature would still have had the effect of acting as a windbreak to other areas of the site. The same arguments can be used relating to Feature 2c: the stones here have a more ‘regular’ and compact appearance than in Feature 2b and it is, of course, possible that Feature 2 may in the east have terminated at a point where it incorporated some other structure of which the function did not require it to be too close to the Floor; unfortunately the answer to this problem will never be found as the adjacent modern pavement and Bill road prevents further investigations in that area. Whatever structure was there, it too would have acted as a form of windbreak by virtue of the fact that it would have protected the site against winds from the direction of the sea.

place, or some such structure, had at some stage existed here and it is possible that it was demolished, in part at least, to

make way for Feature 2a. Stones ‘R’, ‘W’, ‘3X’ and ‘S’ in the south of the site and their relationship to Feature 2 and the Floor (see details above)

Although it is possible that the presence may be merely coincidental, these stones it could be argued that they appear to aspect of the site usage in the southern X,Y and Z,

discussed below).

of these large stones are so prominent that mark some specific area (CompareStones

Stone *3X’

relates structurally

to Feature 2a, without any doubt, as it is actually incorporated into the double line of stones within that feature (Plate 11). It

FEATURE 3

could have been

a seat or shelf/platform, but no doubt, there

could be other feasible explanations for the presence of this stone. It could have been structurally important, e.g. to give additional rigidity to whatever the structure connected with it, was. It could have been placed there to mark the southernmost end of the habitation area. It could have been a step, up from the site area to the south or it could have been used as a step to give access to, perhaps, a storage space or device which was above it. As it is, no indications for its purpose or usage have been found.

This is the circular or semi-circular stone feature in the southernmost part of the site where it is partly visible under Feature 2, as described above (Plate 13). The feature consists of a stone arrangement enclosing a slight depression. Its exact relationship to Feature 2 is not clear. As it also became visible during the excavation work on the southernmost end of the floor where it slopes down (as far as known), the possibility of a relationship to that feature also requires consideration but, unfortunately, could also not be established. Without removing Feature 2 at this point, it was not possible to ascertain whether Feature 3 is actually attached to the floor or below it.

It is more difficult to be certain of the relationship of Stone “W’ to Stone ‘R’ and the other features in this part of the site. As they are physically connected to the floor, it is feasible to argue that they served some purpose in connection with the

The midden deposit associated with Feature 3 is also not very

38

floor and/or its adjacent features, if indeed they served any special practical purpose . The strategic positions of Stones “W’ and ‘S’ just before the beginning of Feature 2 and Feature 3 may have served a specific practical purpose e.g. a. to mark the end of the floor (Feature 1) or, b. to mark the beginning of Feature 2a or, c. to act as threshold or entrance stones for one or the other of these two features. or d. to serve as some sort of foundation or commemorative stones (see below for extra large stones on the floor and regarding Feature X). Shallow

fired clay and containing large quantities of burnt stones. A study of burnt stones from the hearths have been undertaken and is discussed in this Report (see 1.8. M. Stewart).

The unusually large numbers of heavy choppers/chopping tools, pounders and picks found concentrated round the hearths are a clear indication that these tools were used very frequently for the preparation of food on the habitation site. It, at least, does not support the suggestion that picks were tools for use primarily in quarrying and that Portland was therefore primarily a quarry site (pace Care 1979).

depression and small gullies south of Feature 2

Immediately south of the stone structures comprising Features 2 is a shallow depression with a number of small gullies within it (see above). Water standing periodically or occasionally in this depression could have caused the thin layers of clay present within the thin layers of midden in this part of the site and between the stones of Feature 2. At least some of the gulliess had the appearance of having been artificially created, and the purpose of this could have been to drain off water from the depression and so also Keeping the stone features (Features 2 and 3) moderately dry.

Hearth 1

This hearth is very near the floor and appears to have a direct relationship with it. It provided charcoal samples for two radiocarbon dates and was also used for the TL dating, giving a good average age for this particular feature of c5,400 5,200bc. (see 2.10. Dating).

Stratigraphically

the base of the hearth is within clay below

the level of the floor.

Two alternatives are, therefore, possible

There appears to be a direct spatial link between the placement of the floor and the hearths: the fact that the main hearths (as far as known) are all on the eastern side of the floor, may indicate that the winds anciently also came from the same direction as today, i.e. coming from mainly a south-western

for the history of the formation of the hearth: The depression containing the hearth could have been excavated, through original, early, layers of the surrounding midden (cf Layer 8 of Trench 4) adjacent the edge of the floor and the ‘gully’ into the clay below, after the floor had been laid down on that part of the midden which is_ within the gully. Alternatively, its original usage could have been related to a_ phase in the occupation of the site before the floor was laid down, but usage of the hearth continued over a period after the floor had

direction; the intention could, therefore, have been that people

been laid down,

sitting or sleeping on the floor would have had the wind blowing the smoke away from them. It is logical to argue that they would have sited their fires so that the main living and sleeping areas could be free of smoke. This would tend to support the argument that there may well have been shelters or huts on the floor (even though evidence for this is not

the top of the floor. As the edges of the hearth were fairly well defined and appeared to be everywhere just clear of the floor,

ASSESSMENT OF

HEARTHS

associated with compact midden at that level, it can be argued that the two features were at least contemporary in part if not in whole.

Although the hearths were adjacent the floor, they were not so close that they would have caused a risk of fire to any As

the top of the hearth became level with

about 0.75 metres away, the first alternative seems to be the most likely one. Either way, as both floor and hearth were

adequate, see above).

structures on the Floor.

until

One small post-hole was found by the edge of the hearth, but it is highly unlikely that it represents any substantial structure and was almost certainly a single post planted there to hang something up for drying. It is, of course, possible that there may originally have been more post-holes, even though the evidence for them have not survived. Forked branches could have been used to support a skin bag for boiling water or cooking a broth in, as (it is thought) was the case with the smaller hearths at the Magdalenian site of Giinnersdorf near

the hearths are all in a straight row,

almost equidistance from the eastern edge of the floor, it certainly does appear as though the site arrangement was not haphazard. It can be envisaged that at Culverwell the fires were used for cooking during the day and may have been left burning throughout cold nights in order to provide heat, i.e. the hearths almost certainly served two functions. All this would imply that the site was carefully planned to provide the maximum protection and convenience for the inhabitants. This aspect requires further discussion (see 3.14 this Report).

Koblenz

(Bosinski

1981:

45

— 57;

see

also

Stewart,

this

Report). The baked condition of the clay at the base of the hearth indicates that it had been extensively used. The radiocarbon date of 7101 + 97 BP was obtained from several combined samples taken from this hearth. It may be argued that this date suggests that this hearth came into use after the midden under the adjacent floor (represented by the slightly earlier date of 7150 + 135 BP)) had formed; this could make sense as the inference would be that the hearth served the site inhabitants

The stratigraphical and spatial relationship between the floor and the hearths suggest that these features are, in part at least, contemporary with each other and that the floor and hearths were central to daily on-site activities.

The presence of hearths were in each case indicated by welldefined areas of a very black appearance and areas of red well-

39

once they could sit on the floor to use it. However, caution is required as the difference in dates may not reflect the situation in real terms and could be merely due to the sampling or dating techniques.

could

explained,

Hearth

2 was

found

in Trench

6, adjacent the

edge of the Floor and immediately on the other side of the baulk separating that trench from Trench 1 and Hearth 1. There is, therefore a possibility that it may merely be part of a bigger Hearth 1; alternatively it could be a smaller separate hearth sited adjacent the bigger one, suggesting that sometimes more than one hearth may have been in use at the same time near to each other. This hearth too was associated with cooking activities associated with picks, of which several were found by it. Hearth 3 is the southernmost hearth found on the site, also sited adjacent the Floor, but here it was

or work

here,

particularly

when

the

soil was

damp and unpleasant, while using the hearth. The two picks found lying side-by-side on a stone of the pavement, indicate that at least one of the activities carried out here (i.e. procuring or preparing food), involved the usage of that type of tool and, also, as they had the appearance of still being in situ where they were left, they give a small measure of insight into the lifestyle of the site inhabitants.

Hearths 2 and 3 As

sit here

This paved area was situated on the eastern edge of the large hearth, which itself is east of the Floor’ discussed above and it

would therefore be reasonable to associate its usage with that of the hearth.

It is,

further, at the same level as the base of the

hearth and the two features were therefore almost certainly in use at the same time. It is constructed in a similar fashion as the Floor’, except that the slabs of limestone were placed directly on the rubbly loam which is here the natural surface on which occupation took place. It is a very much smaller version of the large paved feature but as it is in a similar tradition, and a close connection is indicated.

associated

with only a_ relatively thin deposit of midden, unlike the hearths further north. It may therefore not have been in use for a very long time.

The thick midden which developed round and over the hearth, subsequently also covered this paved area so that it definitely pre-dated the later development of the midden in this part of the site. This feature is therefore of chronostratigraphic significance and is also of importance with regard to conclusions relating to the large floor. Small paved areas associated with hearths are also known from other Mesolithic

Hearth 4

The soil profile above this hearth and the relationship between this hearth and the floor, could suggest that this may have been the earliest hearth on the site (Fig. 9). It is near the northernmost end of the floor and, significantly, by a point which is associated with a phase when the floor was directly laid onto the clayey loam, i.e. before the midden had accumulated enough in the gully to allow for the construction of the floor on top of the crest of the midden. It may, therefore, have served the site occupants before and during the time when the first stage of the floor was being laid down, directly on the clay, and while the midden was still developing in the gully prior to the later development stages of the floor. Unfortunately no dates are available for this hearth and it could not be fully excavated due to limited time and restrictions imposed by the then-owner.

sites, e.g. from Mezano Site 1993; see also below 3.19.).

14,

Torun,

Poland

(Marciniak

Distribution of burnt stones and charcoal

In view of the fact that only a small number of definite hearths were found on the site, it does lead one to query the very extensive distribution of large quantities of burnt stones and charcoal all over the site (see 1.8. M. Stewart). The answer could be that anciently many other random fires had existed on the site without being actually recognisable as contained within

a hollow

or any

structured feature;

the evidence

for

A large collection of completed and semi-completed picks, as well as the unused tabular pieces for their production, were found in immediate proximity to each other by the western edge of this hearth (Plate 51 and see section 1.7.). This is of great significance: firstly, the making of picks actually

them had disintegrated and become incorporated into the rest

occurred on the site, at least sometimes

both these interpretations are correct to some extent; whatever,

of

are found

adjacent the hearths;

all over the site, and

natural pieces could often have been even retouching them, particularly those pointed end (even though not counted inventory; see also M. Stewart 1.8. this

some

midden

debris.

Alternatively,

the

site

inhabitants

habitually raked out fires and the debris became distributed over the site even though the bulk of the material at least sometimes were thrown into the gully. The likelihood is that

secondly, the series of items clearly demonstrate the technique of producing picks of various shapes and sizes from the same type of raw material. A very large number of the pick-shaped natural pieces

the

the

answer,

it is very

evident

prominent role in the lives of

that hearths

played

the site inhabitants,

a very

as can be

expected.

of these

used as picks without which naturally have a as picks in the artefact Report).

THE PIT AND ITS INTERPRETATION The thinness of the layers here on the site, above the pit, can at least in part be explained by the fact that it is near the perimeter of the site, as suggested by the fact that very little evidence of occupation was found in a number of trenches

SMALL PAVED AREA ADJACENT HEARTH 4 The small paved area on the eastern edge of this hearth is of great interest (Fig. 8.; Plate 30). An obvious explanation for the pavement adjacent the hearth is that the site inhabitants

further east and

south

possibility

much

covered

that

this

feature,

of it. However,

of the had

been

layers eroded

there is also a strong

which away

had or

originally had

been

truncated by medieval ploughing as the feature lies just outside

40

natural rubbly loam must have been quite strenuous (see 1991).

the protected space of the lynchet, in an area where the soil is very shallow. The situation in the trench containing the pit can be compared with the deposits in the immediately adjacent trench, Trench 40 (See below) which is protected by the lynchet and where deep deposits can be seen to have originally covered the Mesolithic features in this part of the site too. Here a reasonably intact and well stratified workfloor with many artefacts have been revealed within a layer of compact midden and formed the subject of a special investigation. This suggests very strongly that the top of the pit, only a couple of metres away, was, originally, also well stratified below the same Mesolithic layers.

Enigmatic pits, of used

for

a.

functions,

are

associated

with

other

Storage Pit

If this pit was used as a storage pit, no definite evidence has been

found of the material

stored in it., unless the few tiny

pieces of bone found, can be regarded as suggestive of usage. This point need not necessarily invalidate arguments for storage, as the material stored could have been such as to leave

no trace or else could at some stage have been completely removed by the site inhabitants before infilling took place. The rest of the debris found in the pit contains the same range of midden debris as found all over the site.

It is questionable whether food such as molluscs, plants or even meat would have survived for long in the pit ; damp and the

Pit Contents

action

The pit was completely filled up with the same midden

gully

(see

Thomas

this

Report

for

bone

b.

in it, (with

the

exception

of these

would

have

made

this

Water Tank

As fresh water would have been available at Culver Well stream very nearby, the need for water storage may, however, be questionable. The only need for stored water may have occurred during very inclement weather.

few

c. Evidence for fire: Cooking of Food

The use of the pit as a feature associated in some way or another with the cooking of food and/or boiling of water appears to be an acceptable alternative. By using the term ‘cooking’ no specific method of preparation need to be implied (see Stewart, this Report). Some aspects of the evidence for this too is ambiguous.

There is no evidence from which any deductions can be made as to how long the pit remained in use. In view of the depth it must

etc.

that the site inhabitants found that the pit was naturally adequately waterproof and that water seepage was slow enough to permit water storage without a lining.

Usage of the Pit

workmanship,

worms,

have survived Figs. 19, 20) no evidence for such a use of clay was found within the pit. It could, of course, be quite possible

minute pieces of possible pottery near the top), and as the top of the pit was covered by a midden layer, it is unlikely that the backfilling of the pit took place in a period long after the Mesolithic occupation This is further discussed below.

of the pit and its careful

insects,

If the pit was a water tank, one would perhaps have expected the site occupants to have made some attempt at lining the pit sides with clay, as the most easily obtainable material. No evidence for such a lining has been found. Although a clay band had been smeared round the top outer edge of the pit (and

A number of very minute pieces of red clay was found mixed in with the midden material in the pit down to approximately Spit 3 (roughly a depth of 30 - 32 cms); the majority of these were clearly fragments from the red clay round the edge of the pit, but 2 pieces were identifiable as pottery as they appeared to have rims. They could have been Iron Age, but this is not certain as the pieces are too tiny to make a positive identification. The meagre evidence for post-Mesolithic activity on the site is discussed fully elsewhere in this Report (section 1.3. and 3.20.). Because of the slow development of soil on top of the limestone of Portland, it is quite likely that for many centuries after the Mesolithic occupation, later material would have been trodden into the pit. As the midden within the pit was very compact, had almost exclusively only artefacts

of

impractical although a cover of stone would at least have provided protection against scavenging animals.

debris

as found all over the site and identical in content. It did contain a few pieces of bone, as has also been found by Hearth

Mesolithic

obvious

As the level of the stones, midden material

The work level in Trench 40 was located within compact midden material but very near the bottom of it; the rubbly loam under the midden material round the pit and into which it was dug, is the same as that below the midden/work level in Trench 40. This suggests that both the work level and the pit belong to an early stage of the site’s occupation in this particular area.

in the

various sizes but usually quite small, and

Mesolithic dwelling sites known from Europe (e.g. Marciniak 1993 in Poland and Mellars 1987 Caisteal nan Gillean and Cnoc Coig, Oronsay). The Culverwell pit requires further consideration as it is much more substantial. Only three possible functions for a pit of these dimensions come to mind: a. Storage pit; b. water tank and c. cooking-pit:

and clay round the top edge of the pit in Trench 2 corresponds exactly with the level of the working area in Trench 40 (see below), the two can be regarded as relatively contemporary and, probably, even contemporary in real terms.

One and analysis).

no

Kent

The stones at the bottom of the pit are not fire-crackled and therefore could not have been exposed to direct flames. The concreted nature of the shell deposit immediately on top of the

be assumed

that it was intended for use over at least a moderately prolonged period as the effort in hacking it out of the hard

4]

stones could suggest the presence of heat in the pit, but few of the surviving whole shells showed signs of direct heat. It is, of course, possible that the large quantity of totally fragmented, almost powdered shell is indeed evidence of the result of direct heat (see also discussion above for the midden). The pale grey colour of the clay at the bottom of the pit could be due to exposure to low heat in a closed environment - it should have been originally dark brown as that is the natural colour of the clay on the site (This phenomenon is found in pottery kilns; Dr. I.W.

Cornwall,

tribes

of Queensland

1972:

183).

personal communication).

of the pit,

even

if the

evidence

parcels of food on the bottom of the pit;

for the

in mind

(Coon

that these

town

of that name

on the River

Rhein,

not far from

The In-filling of the Pit At which stage did the whole of the pit acquire its in-fill? There are only three alternatives: i. the pit filled up in stages while still remaining in use during the site’s occupation; ii. the contents of the pit represents back-fill, raked into it at some late or final stage of the site occupation shortly before the site was

or seaweed was of the pit or that and small stones with the wrapped

the whole lot

be kept

facilities

Koblenz (Bosinski 1981: 45-50). The depths of more than 20 pits there were on average 0.2 to 0.3 metres deep and their diameters were about the same. Although much smaller than the Culverwell pit, they had some features in common with Culverwell: stones at the bottom, stones round the top edge and clay baked red. The archaeologists reached the conclusion that some of the pits were used by lighting a slow fire on the stones at the bottom and then stretching a skin over the pit so that it hung into the pit while being secured by stones all round the edge. However, it can be argued that this technique may not be very practical if a pit is much deeper as the weight of a big bag with food in it hanging down, may be too heavy to suspend successfully.

bottom of the pit is not so clear. Archaeomagnetic readings taken from this clay did not produce successful results (2.10. Dating). A number of big stones were found on top of this clay or just outside the clay band and were probably somehow associated with the activities round the pit. These stones too did not show visible signs of fire crackling despite of the heat demonstrated by the colour and baked texture of the clay, but they need not necessarily have become fired if they were used generally adjacent the pit for food preparation prior to cooking. A feasible explanation could be that grass placed on the limestone slabs at the bottom the food was wrapped in grass or seaweed were heated in a nearby hearth and placed

however,

cooking

Some of the earliest known examples of pits used in connection with water heating and cooking were discovered 1970 - 1976 at the Magdalenien site of Giinnersdorf, near the

Important evidence for the association of the pit with fire is provided by the fact that the slightly raised clay band round the top outer edge of the pit had become red and baked hard, a condition which would have required direct heat (Munsell readings for the pit edge varying: SYR 3.335 reddish brown, 2.5 YR 3/5 dark red and 2.5 YR 4/6 red). There must therefore have been a definite situation relating to sustained the top

It must

similiar

people were probably cooking meat regularly, unlike at Culverwell where meat does not appear to have been that important, but there should be no problems in cooking of whatever type by this method.

small

fire round

use

abandoned;

iii. it is material which was moved

by natural or

human agencies into the pit after the site occupation ceased, maybe even in post-Mesolithic times.

was

had

then covered wth hot soil or midden material so that the food cooked slowly in an enclosed space in their own juices (see Stewart, this Report). Even though this technique is most

There is no evidence which gives clear indications as to which of these alternatives is the correct one. If the pit became filled

commonly used

in long

with meat, it can

also be used

with

molluscs

after the site occupation,

there

should

have

been

a

larger percentage of later material present in the in-fill. Ifa small mound of Mesolithic rubbish had accumulated adjacent the pit, this mound would have started to slump naturally

or vegetables.

It can also be suggested that the molluscs (particularly winkles) and edible plants at Culverwell were probably frequently prepared as a broth by _ techniques of applying indirect heat, e.g. by heating water in the pit with hot stones and then submerging a skin bag with food in the hot water or else cooking the food in the bag by utilising steam coming from the pit (see M. Stewart this Report who also discusses other possibilities for the use of hot stones in cooking). These techniques would not necessarily require the use of a deep cooking-pit, but such a pit could have had the advantage that it could have been used to provide some protection against strong wind, with the bags hanging down into the pit.

fairly soon after the site became

deserted; some of the debris

would have been distributed to adjacent areas by natural agencies while some of it would have ended up in the pit. When first discovered, the sides of the pit were perfectly vertical and were not severely weathered. The experimental excavation of earthworks (in a chalk environment) at Overton Down has demonstrated that up to 1.2 m of weathering material can accumulate at the bottom of a feature in only 4 years (Jewell and Dimbleby 1966). It would have been reasonable to expect some signs of weathering if the Culverwell pit had been exposed to the elements for any great length of time after its construction and usage, even though the Culverwell feature was cut into a very durable deposit of rubbly loam. There was no visible presence of material derived from the sides of the pit within the midden fill. Also, the deposits within the pit were very compacted and well stratified below layers of deposit associated with the Mesolithic

Cooking-pits are well-known in an ethnic context. The Polynesians and aborigines of New Guinea use pits which are virtually of the same dimensions as the Culverwell pit: they cook food by placing it on or between hot stones and pack turves or earth on top to cause the food to smoulder slowly over a few hours (Harrer 1964: 137-8). The semi-sedentary

42

period (see 1.3. ’Stratigrahy’ ). This suggests that even if the pit had become full of redistributed debris after the site occupation, this must have happened not too long afterwards.

2. The holes under the stones were neatly lined with smaller stones and these holes were made into the midden under the big stones.

There is no good evidence to indicate which of the first two alternatives is the most likely. Meehan (1982: 114- 7) shows that the Anbarra regularly cleaned their cooking areas. There would be few or no_ logical reasons why the site inhabitants would back-fill the pit at a time when they were intending shortly to desert the site.

3. The smaller stones under the capstones had the appearance of having been deliberately packed in the midden in order to support the big capstones. As the holes were made directly into the midden deposits, it is reasonably certain that the holes and midden are contemporary in real terms and that the making of the holes preceded the placement of the capstones.

A few other points may be worth noting: the hard compact midden deposit at the bottom of the pit on top of the stones suggests that, even if some of the debris was removed daily or at regular intervals, a residue always remained, i.e. the pit was not completely cleaned out while in use. Secondly, the pit contents is homogenous throughout and this too suggests an in situ. accumulation as opposed to secondary infilling (c.f. Mellars' arguments 1987: 228). Thirdly, the big stone found at a depth of about two-thirds of a metre down in the in-fill of the pit, could have been casually thrown in during the in-filling work, but it can equally be argued that it was placed there deliberately when the pit was half full so that the site inhabitants could continue using it in some way or another. It could be suggested that, for some reason or other, the pit may no longer have been required during the final stage(s) of the site occupation.

4.The placement of the three big stones exactly at a point of change in the alignment of the floor, supports the evidence that they are contemporary with the floor; the position of the stones at this point therefore must be of some significance. Although there seems to be a spatial relationship between Stone (Feature) X and Stones Y and Z, this is not certain. One can argue that they could have been three unrelated individual features and the apparent relationship between them more coincidental than real. On the other hand, as the three features

are not identical (e.g. only one had the big oval pebble adjacent the big capstone and only one had a significant assemblage of finds in the hole) one can argue that the three taken together formed different elements of one feature in which the concept of a triangle had some significance and they were therefore complementary to each other and there was therefore no need for them to be identical. However,

Even though it cannot be proved beyond doubt what the pit was

used

for, it is a prominent

feature

of the

site and

there can be no proof for

this hypothesis and it may therefore be more appropriate look first at the three features individually.

was

to

certainly important to the life-style of the site's inhabitants. Feature X

ASSESSMENT AND Z

OF

EVIDENCE

FOR

FEATURES

X,Y There can be no doubt whatsoever about the fact that Feature X was Mesolithic in date. It was found well protected under the

As pointed out above, the edges of the floor are slightly out of alignment at a number of points, and it has been argued that this could indicate that the floor was possibly laid down in at least three stages (Plate 37).

lynchet,

showed no signs of disturbance at any time and all the different layers 1 to 5 (see 1.3. stratigraphy) were represented in the lynchet above it. The large triangular stone and the large oval beach cobble were first revealed and aroused interest as to the possibility of an unusual situation. All stages of the work were carried out by the director and two volunteers of long experience; the whole excavation process was photographed,

It is, further, of interest that larger

than average slabs of limestone are incorporated into the floor at each of these positions of change, giving the impression of deliberate action or intent.

drawn

STONES X,Y AND Z are the northernmost of these big stones, marking such a change in the floor's direction and all three were associated the midden; all three least one case - that unusual for the site,

under a metre or more of soil; the soil profile above it

and

recorded

and

discussed

with

all the

volunteers,

ensuring the highest level of accuracy in recording of the findings. It would have been quite impossible for anybody to have interfered with any aspect of the feature without it being very obvious.

with small holes under them, made into holes were lined with stones and in at of Stone X - various items, which are were found in the hole. A thin lens of

The whole of the feature formed an integral part of the floor; the ordinary floor stones round it were found intact in such a way that they could not have been disturbed at any time without it being noticeable, i.e. the laying of the floor round the feature was completed after the feature was created. The calcified condition of the scallop shell and the lime concretions on the round pebble are indicative of the fact that they have been buried within the midden deposits of the site and are not new items. The shell also shows abrasion signs of having

clay was found filling each hole under the big capstone, probably as a result of muddy rainwater infiltrating into the small cavities round the big stones. These stones, at these positions on the floor, therefore have the appearance of being specific features: 1. All three big stones X,Y and Z are indisputably part of the floor and had been placed in position before the floor round them were completed as they are closely packed in by stones of the floor.

been in use prior to its burial in the feature.

shows

43

some

signs

of usage

The pebble, also,

prior to burial

and

therefore,

without doubt, belongs to the site’s Mesolithic assemblage.

a slightly abraded edge, showing that it had been used prior to being buried with the other two items.

Possible Interpretations It can be argued that at least some of the smooth, round pebbles were brought to the site because of their esoteric or symbolic interest (see below section 3.17). The fact that the pebble under the capstone was 'planted' on its edge into the underlying midden, so that it stood upright, also has the appearance of intentional or meaningful design.

The whole feature has the appearance of having been deliberately arranged, intentional in every way and of some major significance although with no apparent functional use (accordng to present-day understanding of the concept of ‘function’ - see Wymer 1991: 55, 'Art and magic’). The items under the stone of Feature X,

The large oval cobble which lay immediately adjacent the capstone also confirms this interest in smooth regularly shaped pebbles and cobbles and the intentional, meaningful design of the whole feature.

and their arrangement,

adds weight to the hypothesis that there could be some or other ritual connotations:

i.Scallop shells are very rare on the Culverwell site; this is the only complete one which has been found, the others being so fragmentary that all the pieces together may not represent more than one other specimen. The mollusc is a deep water species and is therefore not as easily accessible as the other species found in large quantities at Culverwell. Newell et al (1990) carried out a study of all known Mesolithic decorative ornaments in Europe and their study lists only 12 specimens of

It has already been pointed out that there appears to be a spatial relationship between Feature X and the slight change in the floor direction and it would be logical to maintain the argument that the feature therefore somehow acted as a ‘marker’ in relationship to this change or stage in laying the floor down. The idea of a ‘foundation stone’ is one which still persists till today. Apart from the esoteric interest of these stones they also serve a practical purpose in that they clearly mark the place where building has to commence. For the site inhabitants they remain of significance long after building has been completed and marks an important event or stage of development in the life of the community.

Pecten sp., including the Culverwell example, one from Hoédic, 4 from Téviec and six from Cnoc Sligeach, Ireland.

This makes ornaments produced from this mollusc relatively rare; when a complete scallop shell was found by the Culverwell site inhabitants, they must have prized it. The wearing of ornaments invariably always had a deeper significance for primitive people than just the desire for personal decoration; they must be interpreted as signs and symbols to convey information, messages and intended to evoke specific responses (Newell). The burial of such a rare item in the context of a ‘deposit’ would increase and endorse the significance which was attached to it.

It is difficult to escape the conclusion that Feature X had a ritual significance, whatever that may have been, connected with the laying-down of the floor. Stones Y and Z

These stones may belong to features in the same tradition as Feature

ii. and

the axe form is also relatively rare on the Culverwell site on

the

whole

of

Portland,

whereas

picks

are

X

and

they

may,

indeed,

have

formed

part of the

general feature of Stone X.. It must be admitted that the evidence for this is tenuous as no items of significance were found under them. They were also deliberately constructed features, appear to have a spatial relationship with Feature X, although it may be possible to argue that this was fortuituous. It is feasible to argue that the ‘offerings’ under these stones (if there were any) could have consisted of some or other perishable material, e.g. food, of which no trace has remained. Another possibility is that these two features, forming part of

more

commonly found on Portland sites in large numbers (see this Report 1.7. and Palmer 1977). The axe therefore also had a rarity value and possibly a deeper significance for the site inhabitants and they probably deliberately chose it for their offering under the stone. iii. Round smooth pebbles from the beaches are very common on the Culverwell site; they are found in large numbers concentrated particularly in the gully (see Gully above and M. Stewart this Report). The pebble could, therefore, have been chosen for placement in the feature specifically because of its common or ‘popular’ interest to the site inhabitants. Although round pebbles are common on the shore round Portland and in the Raised Beach of the Portland Bill area, it would appear likely that a fair percentage of these pebbles on the site came from the Chesil Beach. The Chesil Beach specimens are often easily identifiable as they are more frequently smoother than the ones from the Raised Beach and more regularly shaped with perfectly flat top and/or bottom surfaces so that they are virtually disc shaped. Wherever they came from, they had to be specially brought to the site. Many of them do not show signs

there had been some redistribution of at least part, but not all, of the original midden on this part of the site, i.e. within the

of utilisation, but the one found associated with this feature has

gully here;

the

same

feature

as Feature

X,

represented

‘token’

offerings

which did not include any tangible items. The same comments which apply to Stones Y and Z could apply also to Stones W, 3X , S and R in the south of the site, already discussed above in relationship to Feature 2.

FEATURES IN TRENCH 41 The stratigraphy of this trench has been fully discussed above ( see 1.3. ‘Stratigraphy’ and 2.11. Thomas, this Report, and see also discussion of the gully) . It appears to confirm that

44

this must not be regarded as unusual for the site in

POST-HOLES

view of the complexities of midden development on an occupation site, as explained above. However, this trench is noteworthy in as far as that the effect of this growth process and post-depositional changes are still visible. The importance of the two conjoining flakes from Layer 12 at the bottom of the gully does indicate that at least some of the original deposits were not substantially displaced after being deposited and caution is therefore required in assessing any post-depositional movement. It would seem that not all the deposits in the gully were subject to movement.

A total of six definite post-holes have been found on the site. With the exception of the post-hole in Trench 4, they are all stratigraphically and spatially associated with Mesolithic features.

The two isolated holes, one by the pit and the other

on the edge of Hearth One, almost certainly relate to fairly insubstantial posts or sticks which could have held some or other object connected with the features next to which they were planted, e.g. a bag containing food or else even the hanging up of a part of a carcass or piece of meat, before, during or after cooking.

The trench is also important as it appears (as far as excavations indicate) to include the limits of the Mesolithic occupation on the westerly side of the site at this particular point, as tentativily also suggested by the geophysical survey (Gale section 1.6.). Rubbish was randomly thrown into the gully from the earliest occupation phases onwards and spread over the clayey bottom of this feature and later the rubbish formed a mound which, in time, partly slumped down into the gully. Other mounds may well have formed nearby in the gully, which eventually merged with some _post-depositional disturbance occurring, but details were not shown by the excavation in this particular trench.

The post-holes in Trench 32 by the western edge of the floor were probably associated with one and the same feature. As the subsoil is very rocky within the gully here, the posts would have been fairly firmly planted. Even though it cannot be proved, it can be argued that it is possible that they could have carried some or other material (e.g. skin, grass, withies, etc) forming part of a shelter extending onto the adjacent floor. No post-holes were noticed on the adjacent floor, but it is likely that they would not have been so easily noticeable there and, also, any associated posts there could have been held up by stones packed round the bases.

The stone feature (Feature 41) with the large cobble, was constructed right at the edge of the (then) limit of occupation activities immediately adjacent the small mound of midden material described above. It lay on the top of the compact midden Layer 9 and was covered by the base of Layer 8.

The post-hole in Trench 4, with a_ barbed-and-tanged arrowhead at its base, was immediately by a place under the lynchet where an animal burrow was also visible. It is, therefore possible that: a. the post was Mesolithic, the top starting within a midden deposit, but subsequently became damaged and obscured by the animal burrying into it; in that case the arrowhead was possibly brought down from a higher level by the animal’s activities into an earlier feature; b. the post was a late prehistoric feature which penetrated into the earlier levels, but, again, the stratigraphy is obscured by the

This feature was not natural and also did not develop randomly: it has the appearance of deliberate arrangement as all the stones were placed with their flatter sides on the surface of the time. Nothing of special note was found under the feature but it could be significant that a very large quantity of rounded, unstruck, beach pebbles have also been found in the bottom of the gully at this point and more pieces of bone in this part of the site than in any other area (see Thomas, this Report for bone analysis). Sizeable pieces of ochre were also found all round the feature. This supports the hypothesis that the feature was not fortuitous and that some or other special actvity may have been associated with it or with the area in which it is.

animal. There is, therefore, no certain answer to the dating of this post-hole, even though the arrowhead is, without doubt, of

a Late Neolithic or Early Bronze Age type (Fig. 39).

TRENCH 40 AND WORKFLOOR In Trench 40 a special project was carried out to ascertain whether there were indications of the presence of an undisturbed workfloor or space ( work undertaken in part

It may have been intended as some sort of practical token to mark the limits of the site or to mark the area where rubbish was ‘buried’ or, again, it may have been a feature with some esoteric connotations for the site inhabiants. It is possible that it had some association with the other large cobble found a little distance away on Layer 12. An animal or human (perhaps infant or foetus) burial was initially considered a possible interpretation but no evidence for that was found. It is, of course, possible that the feature may originally have had other associations which did not survive.

towards

M.

Woods

unpublished

dssertation,

Institute

of

Archaeology). Three-dimensional measurements were taken of the positions of all finds and these were plotted. This trench was chosen for this work because of its proximity to the pit and because it is situated under the thick deposits of the medieval lynchet and it was, therefore, argued that all the deposits with

their fnds stood a reasonable chance of having remained relatively or moderately undisturbed. It was also an attempt to ascertain whether any discernible pattern of distribution of artefacts could perhaps be determined. The trench was situated in an area which is well outside the limits of the gully and excavation confirmed that it was wholly in an area where the terrain was fairly level.

All these features involving large stones and/or cobbles may, therefore, open the prospects for a more extensive discussion on the evidence for ritual or mystic thought processes (see also 3.17. this Report).

It was realised in advance that only signifcance in general terms can be attached to the outcome of this project in a shell

45

midden, in view of the multiplicity of factors which could have, and almost certainly did, cause various degrees of displacement of artefacts during and after the site occupation.

Marks in the top of midden,

Post-depositional disturbance of the rubbish deposited by the Mesolithic inhabitants probably started immediately after deposition as a result of natural and human agencies but it is possible that this disturbance continued for long after the site was deserted by them. It has been discussed above (‘Stratigraphy’) that it is possible that part of the gully in the area of Trench 41 on the westernmost limits of the site were not fully filled in during the Mesolithic occupation phase; this could have some relevance to the fact that at least part of the site was vulnerable to post-Mesolithic activities. As the mound(s) of rubbish in the gully would have been more muddy than in the level areas of the rest of the site, they also would have had a tendancy to be more ‘mobile’; the direction of this mobility would naturally have been towards the lower points of the gully. Any later prehistoric items dropped in this area would also have had the tendancy to sink lower. This would account for the presence of several pieces of coarse pottery in the lower levels of this trench.

The plotting of artefacts indicated that there were several definable clusters of debitage within the midden spread of the lowermost layers in the trench area (Fig. 24). This could suggest that the clusters were possibly associated with particular places where somebody had sat to do the knapping; it must, however, take into consideration that after knapping somebody could have redistributed the debris by sweeping or clearing up. Even natural post-depositional circumstances, e.g. water swirling through the site, could have had the effect of clustering debris, particularly the lighter pieces. However, as this clustering was more noticeable in the lowermost level of the midden in the area,

Trench 41

it would be reasonable to assume that

this could indeed indicate that knapping did take place here near the pit (only about between 1.5 — 3 meters away, depending on the spot in the trench being referred to).

It can also be argued that the presence of a knapping place is indicated by the fact that a number of flakes could be conjoined to a small core from Layer 11 (Plate 56). However, one of these flakes did come from Layer 10, indicating clearly that at least sometimes or in some circumstances some vertical (and probably horizontal) displacement of material did take place after being deposited by the site inhabitants. This appears to confirm that some redistribution of the middens did take place during the occupation of the site and that rubbish was , at least periodically, swept up to form mounds, causing material from a lower level to migrate to a point higher up in the mound and probably also well away from the original spot of placement.

The seven deep straight grooves which were found scarred into the

top

of

the

midden

interpreted ( Plate 42).

in

Trench

41,

can

be

variously

The problem is that without further

excavation south of this area, it is impossible to know whether

the marks continue in that directions or whether the marks found in this trench represent their full extent. They were not present in any of the trenches northwards of this Trench and may therefore have started under the baulk between it and Trenches

4 and 31 ( see Site Plan Figs 2, 3).

None

marks appear to continue right up to the face of the southern trench section and, indeed, the indentations of them all appear

to curve upwards at both visible ends of this feature On the one hand, the findings from Trench 40 confirm that in

trench.

general

interrupted but continued unexcavated areas.

the site has remained reasonably well stratified since

Mesolithic times but, on the other hand,

it

of the

is also a warning

about attaching too much significance to the layering of strata within a_ shell midden (see above discussion of the site stratigraphy).

It

could,

nonetheless,

again

be

after

that

the

a short

in the

grooves

gap

were

into the

The grooves do resemble marks made with a very primitive rip-aard. Similar marks have been found under a long barrow at South Street, Wiltshire, dating to at least the fourth millennium

A large number of chopping-tools, picks and other core- tools were found in situ in Trench 40. This type of tool assemblage is of the same nature as that which came from round Hearth 1 and Hearth 4 (see above) and it is reasonable to argue that Trench 40 may represent the area where the site inhabitants sat while using the nearby pit for whatever purpose they did.

FEATURES OF POST-MESOLITHIC UNCERTAIN AGE All

features

Mesolithic,

found with

on

the

site

the folowing

ASPECT

are regarded

exceptions

about

as

BC, but there, of course, they were undisputable and very much more extensive than the marks at Culverwell (Smith 1984:113). Plough marks of possibly very early dates have been found at the Céide Fields,

Ireland (Waddell,

1998: 29). There it is thought possible that the aard was of a very simple nature, consisting basically of a forked branch with one prong acting as a share and the other as a beam.

OR OF

However, the Culverwell marks could also have been made by repeatedly pulling a pick lashed to a stick, or some other form of digging-stick, over the midden for some reason or other (see experiments with picks on Site 1, Palmer 1969). They could also have been made by dragging a log, or several logs, over the midden; several logs lashed together to form a sleigh would make marks like these if used to transport (for instance) heavy

definitely

which

Co. Mayo,

there

are uncertainty: the grooves found in the top of the midden in Trench 41, the post-hole in Trench 4 (discussed above) and the structure which was found on Field 2153. Each feature will be separately discussed and the evidence for it assessed in this section. These features too have been described above.

slabs of stone.

The nature of the marks can, therefore, not be

decided with certainty.

The biggest problem relating to these marks is their dating. The stratigraphy of this trench, as described above (see 1.3.

46

‘Stratigraphy’ and discussions above) appears to indicate that a small mound of shell and other occupation debris was formed in this area by the Mesolithic occupants, within the gully and just south-west of the limestone floor. The evidence indicates that the top of this mound became levelled off and the sides slumped down into the rest of the gully. It may be arguable that the gully in these exposed areas, particularly right at the westernmost limits of the site, (i.e. not under the floor) were not fully utilised by the Mesolithic people and that it may only have become full and completely levelled off with redispersed Mesolithic material on top of the earlier in situ layers during later times. This is possibly reflected in the radiocarbon dates (see 2.10.’ Dating’ ).

would have been seen to be on a slope and not on a level surface as they were; it is the lynchet formation which created a level surface on the otherwise earlier slope of the hill.. This dates the structure to some time after the lynchets were formed. The structure is quite small and could not have been very heavy or substantial as the marks of the base was only very shallow on the loam; if it had have carried any weight, such as a heavy superstructure, the impressions on the soil and into Layer 2 (or even Layer 3) would have reflected that. It is therefore likely that the structure was of a light agricultural nature: perhaps for the storing of farm equipment during ploughing or harvesting time.

However, as the grooves penetrated into the top of the midden and they were infilled with the dark brown clay which has been described in Trench 4 as a buried soil sealing off the Mesolithic layers (see Stratig raphy Layers 4 and 5), this must mean that the groove marks correspond to either the fnal phase of the midden development in this area or they must have been made very soon after the period of the midden deposition, not long after c5,200 bc. Either way, a very early dating for the marks would

Field systems

The most prominent post-Mesolithic feature on the site is, of course, the medieval strip lynchets. These have already been fully discussed as well as their stratigraphical relationship to the Mesolithic strata (see 1.3. Stratigraphy and above Site description). Medieval and 17" century artefacts occur very frequently in the the top layers of the site but progressively get less in the lower layers and very rarely occur below Layers 4 and 5.

be indicated, not substantially later than the dates

for the Mesolithic occupation. /f they are indeed plough (aard) marks, it would also imply that that tool was already in use in Late Mesolithic times or else that some incipient agricultural activities, usually associated with the Neolithic period, took place earlier than has been previously accepted; this could be rather contentious. A similar controversial problem in Co. Mayo appears to be implied (Waddell, 1998: 29 and ch. 1-2). The assessment of this interesting therefore, evidence.

remain

inconclusive

Structure in Trench 19

in Field

Culverwell feature in

the

absence

of

Features of virtually every period of the past have been found on the Isle of Portland during quarrying and other works and objects remain

Roman,

Saxon,

medieval

standing (Bettey 1970; Morris

and

other

and a few structures 1985; Palmer

1998). It

would therefore indeed be surprising if a few items of postMesolithic date were not incorporated into the Mesolithic deposits of the site as a result of animal activity, natural agencies and even the occasional human activity. However, as the evidence for the later periods on the site are very

must, further

insubstantial indeed, it can be argued that whatever

2153

activities

were involved, they were very transient. The few later prehistoric finds from the site, probably represent items that were dropped when people walked over the field and later were trodden down into the earlier deposits or were taken down by burrowing animals or heavy animals walking over the fields, particularly during wet weather. A few sherds exceeding

of later prehistoric appearance,

1.5cms”,

found

on

the

site

cannot

hardly ever be

taken

as

evidence of substantial activity on the site and their evidence is negligible (see 1.7. Section and also 3. 20). This is a not unusual situation encountered during the excavation of early prehistoric sites, e.g. Woodman (1985) also found a few items of later date on his Mesolithic site at Mount Sandal and Bonsall (1989) specifically commented on this problem at the Williamson's Moss site in relationship to other sites in England. In many parts of England, and in particularly on limestone, soil develops very slowly so that the old Mesolithic soil surface probably remained exposed (or with only very shallow soil cover) for millennia afterwards.

However, again considering its stratigraphical position above or within the top of the yellow loam regarded by Keeley as probably medieval (see Keeley Appendix this Report), it is likely that this structure probably dates to a time soon after or during the time of the formation of the lynchets. As the nature of the marks of the footings of the horizontally laid down structure clearly suggested that the base of the structure was everywhere touching the soil surface equally, it must have been erected on a level surface such as that created by the lynchet development If the structure on the hill-side had prethe nature of the marks

prehistoric,

periods are in the small local museum

This structure was in the clay loam of Layer 2, well above the stratified Mesolithic layers and also above the buried soil of Layer 4 which were subsequently found at this point under the positive medieval lynchet. It had the appearance of being formed by marks of wooden sleepers laid down on the level soil to form the base of some elementary structure, such as a shed. Stratigraphically it therefore cannot be confused with the Mesolithic features and finds which are in situ below Layers 4 and 5. The finds from Layer 2 were of mixed nature and therefore provided no conclusive dating for the feature.

dated the lynchet formation,

of

It is inevitable that some later items would have worked their way down into much earlier strata. An interesting example of

found,

47

the uneven downward Culverwell in Trench sherds were found in other sherds of the down.

distribution of later finds can be seen at 31 where some of the Bronze Age type a post-Mesolithic layer (Layer 5) and two same type in the midden, much lower

more distant past and during the activities which gave rise to the formation of the lynchets. This is an example of the type of disturbnce which could account for the instances of finding Mesolithic material loose on the surface of the field and also can account for the fact that sometimes more recent items are found in the lower Mesolithic layers. However, in general the Mesolithic features and layers are very well stratified and in an excellent condition of preservation.

During the years of work at Culverwell, local cows broke into the field on a number of occasions, once during a severe storm.

On that occasion their feet sank into the soil of two of the baulks between trenches in Area C and D in the central part of the field, to a depth of more than 30cms and penetrated into the archaeological layers where the deposits above them were not too

deep.

(These

baulks

were,

therefore,

ignored

for

stratigraphical work relating to the Mesolithic occupation). That sort of disturbance must have happened fairly often in the

48

1.6. CULVERWELL, PORTLAND. John

D. Gale,

A GEOPHYSICAL SURVEY 1997

School of Conservation Sciences,

Bournemouth University. enable the excavator to maximise the productivity of a further season of excavation by aiding the choice of trench location. In addition survey might also serve to define changes in below ground deposits which might help to further define the archaeology of the site.

Introduction The

Mesolithic

habitation

site at Culver

Well,

Portland

has

been under archaeological investigation by Susann Palmer for a number of years. Recent discoveries at the site have led to the potential for prospection by geophysical analysis, to address land

use

problems,

which

would

assist

the

Director

in

excavation strategy. The principle objective was to try and ascertain if evidence for habitation could be recovered to the south west of the area that had been excavated by the early summer of 1997. The nature of such evidence was likely to take the form of changes in sub-soil deposits possibly associated with a gully or ditch identified by the excavator as possibly being a boundary (man made or natural) which may have defined the limit of the habitation area (Palmer, S pers. comm).

Type of survey and Instrumentation Magnetometry and Resistivity surveys were carried out over the site on a single day in July 1997. Magnetometry was undertaken using a Geoscan FM36 Fluxgate Gradiometer. Resistivity was carried out with a Geoscan RM15 Resitivity meter with a twin probe array configuration. Both machines have automatic data logging facilities which enables easy transfer to computer processing software. Data was downloaded via a field computer into Geoplot v 2.01 geophysical processing software for subsequent processing. A discussion on the suitability and application of both techniques to the detection of archaeological remains can be seen in Anthony Clarks ‘Seeing Beneath the Soil’ (Clark, 1990).

A geophysical survey was undertaken in July 1997 by the author, immediately prior to a season of excavation, which in part would be sited on the basis of the results provided by the survey.

Survey methodology Prior to the commencement of the Geophysical survey it was necessary to establish a grid over the defined survey area. Because of technical and topographical limitations it was only possible to survey a 50m x 10m transect (see fig 1) through the centre of the site. Both the magnetometer and resistivity surveys were completed over the same grid.

Site Location and Topography The site of Culverwell (NGR SY 685693) is located on the south-east facing slope of the lower Purbeck limestone beds, at an elevation of approximately 100 feet. The southern half of the Portland landscape where the site is located, is dominated by stone quarries which have been so important to the areas economy since the Middle Ages. In addition to the nearby quarrying, the site is also located within open-field terraces which form part of a relatively intensive relic landscape , probably dating from the Medieval period. The soils at the site are predominantly shallow brown calcareous earths and rendzinas which are for the most part free draining.

Following the establishment of the surveying grid framework, all grids were subsequently surveyed in 1-metre traverses with a sample interval of 0.5metres. In each case the recording (by automatic data-logger) began in the north-east corner of every grid square with all complete grids producing a total number of 800 readings. On occasion, largely due to obstructions on the ground, it was necessary to introduce dummy readings into the data set. These readings are represented within the data set by abnormally large readings (relative to background levels) which are fixed at a constant value of 2047 nT/ohms. Such high values can lead to the production of ‘spikes’ in the data set which have largely been ‘smoothed’ in post survey processing.

Current Land Use The area directly under examination has been under archaeological investigation for a number of years, and has subsequently been out of agricultural use during this time. However, the survival of what appear to be medieval field terraces surrounding the site, would suggest that the incidence of disturbance to buried archaeological deposits is probable. The presence of stratified Mesolithic deposits at the site however does demonstrate a degree of preservation, which would support evaluation by geo-prospection methods.

Prior to the commencement of the systematic survey of the defined area, (Gale fig 1) it was scanned to establish operating perameters and settings for the instrumentation. Once established these operating perameters and settings remained unchanged for the duration of the survey, and are summarised thus for each technique:

Aim of Survey So far excavation has revealed evidence of extensive Mesolithic activity, and although the limits of the habitation area are in part understood the extent to which the site spreads to the south west is unknown. In an attempt to clarify the south westerly limit of a gully or ditch which the excavator believes may well define and limit the habitation site, geophysical prospection techniques were used to locate it. This would

e

49

Magnetometry — initial scanning revealed approximate background level varience in the range of 3-4 nT with occasional surges in excess of 40 nT. The surges for the most part were isolated and consequently were assessed as being responses to buried ferric material.

The background stability observed in the scanning exercise led to a_ resolution setting of 1nT being chosen as the likely optimum for the detection of archaeological features.

area. This would appear to be entirely due to the previous excavation of Trench’s 32 and 10 in addition to others to the south of these. The characteristic ‘iron spikes’ within these areas, are subsequently likely to be ferric material associated

Culver Well, Portland, Dorset Location of Geophysical grid and Transect A

Traunsect 4A

-___ Geophysical Grid

Ce

° ee

Trench di

:

tren

32

i" Pe

sf

*.

ifm

Trench 16

Fig 1. with the excavations.

e

Resistivity — initial scanning withthe _ resistivity meter revealed minimal variance of readings between 15 — 30 ohms. On this basis the survey proceeded with the power set at 1mA with the gain set at x10.

The principle archaeological feature represented on the plot (Anomaly 1, fig 2) consists of an arc roughly aligned east west, which runs at right angles to the contour of the slope. The anomaly appears to be a well defined linear feature whose character would be consistent with a boundary, in this case most likely to be a relic ditch. With the exception of further ‘iron spikes’ which are likely to be fairly modern in origin, the remainder of the plot would appear to be devoid of any anomalous activity.

Weather and Ground Conditions

The survey was conducted during a prolonged dry spell in July. The moisture levels within the soil/sub-soil profiles were consequently likely to be low. The ground cover at the site was uniformly that of various grasses which had been cut immediately prior to work commencing so as not to interfere with the laying out and surveying of the defined area. The unsurveyed area in the south of the geophysical grid represents dense vegetation growing on top of an old excavation spoil heap (see fig 2 and 3).

Resistivity

Figure 3 represents a grey scale plot of the results derived from the same area as previously described for the Magenetometer survey. As with the plot from the gradiometer the extreme southern end of the Geophysical grid is characterised by the disturbance caused by the previous excavations. The remainder of the surveyed area is less disturbed but shows little

Results

Magnetometry

or

(Gale fig 2)

partially filtered out, as has the effects

no

evidence

for

discrete

anomalies,

which

could

be

considered archaeological or otherwise.

Resistivity Survey of Transect A

Figure 2 represents a grey scale plot of the results derived from the survey of the 50m x 10m grid. The data has been only marginally processed to remove slight survey inconsistencies, which mar interpretation. Excessive ‘spiking’ due to ferric material has been

(Gale fig 3)

In addition to the area surveys of the defined geophysical grid, the excavation Director asked if it would be possible to survey a 25m x 1m transect running along the baulk between the current excavation trenches (for location of trenches see Palmer fig 2). The aim here was to test whether the observable changes in soil deposits within the excavated area, could be observed at the northern end of the site by

of

sensor probe drift.

The extreme southern end of the traverse (areas 1 and 2, fig 2) is much more disturbed than the remainder of the surveyed

50

ET

aaa a aa

Te ane venee

i

|

One

ORR RRA RIAD

:

H;

3

i

Anomaly

RR

sony,

}

eoBenia

RAR

fe ee ee

a

ER

;

Disturbed area }

4g

!

igesuipveved LIPS UPYE ¥e a

Clip Parameters: Minimum

“are: PRET

~%

ge... Maximum

tinits

.

Absolute

White triarigies domoie muataetio arabes

Rn

BE 3.

ateare:

rary

Da

a ec EEL

gue

ee

ee a

—~

eee A enn ee eee eee OR RR ARRARRAARDWE REESE

ae a AAA AAA AA ARR

2

Disturbed areca 2 fexcavated areca}

i

f

?

¢ t

CULVER

WELL,

Portland

Kesistivity Survey

40] 37.0 AAAAAD PARA RADADADA IEEE REESE

33.0 30.9 /

excavated area)

D

248

is? 13.6

:

12.6 OG OS

i

3.5

ohms

iOm

Clip Parameters: ‘

Minimum

~3.G0

Winsurveyved ares

Maximum

3 OG

a

a

Linus

FS

ee

e

2h

ASAABAAA SARA RA Amann?

/

URsturbesd ares

te

td.

Pleyv.

noting the changes observed in a transect of resistivity values which ran alongside the trench sections.

Both the gradiometer and the resistivity meter were able to clearly detect the disturbed ground of previous excavation (to the north-east), and the gradiometer alone has established the location of a linear anomaly (Anomaly 1). The identification of what this linear feature represents is to some extent speculative, but it would appear consistent with a negative or cut feature, most likely a ditch. As to its association with the mesolithic activity to the north-east this is unproven and can only be demonstrated by intrusive methods.

The main difficulty with this proposal concerns the narrow (1m) baulk which separated the excavation trenches along which the survey would take place. The baulk itself (depth of which varied from approximately 30cms — 90cms) would tend to dry out very quickly because of its exposure to the elements on two faces. This factor alone would influence any resistance values obtained when sampling along the baulk, but potentially at least the sample depth achieved by the twin probe configuration would enter that below currently excavated levels.

As to the location of the gully (observable in the southern end of the excavations) its absence could be explained by its own weakness of definition. The gully’s edges are characterised by very shallowly angled sides which would tend to produce a graded change in readings which consequently would be more difficult to detect. In addition the very dry conditions at the time of the survey were not wholly suitable for investigation by resistivity, a fact borne out by the techniques inability to define anomaly 1. Even allowing for more conducive ground conditions, I doubt whether the resistivity meter would produce improved definition on below surface deposits, beyond that already identified through gradiometry.

Readings were eventually taken along the baulk at 0.5m intervals and the results are plotted as a single line trace (Gale fig 4). Interestingly the relative resistivity values produced from this transect do mirror in part the palaeotopography determined by excavations on either side of the baulk. The limestone ‘paving’ that runs through trenches 6 and 7 appears as higher resistivity values between readings 13 and 28. However the gully or ditch which can be seen in section is not visible in the trace plot.

Transect A

ra = os a m a

fr

Reading No.

Me4 Resistividy Transect A

Conclusions The defined aims and objectives of the survey discussed above, required the identification of a positively identified feature in parts of the site through non-intrusive methods. Unfortunately it has been impossible to demonstrate with any degree of certainty that the gully or ditch is present or absent in those parts of the site investigated. What is clear is that both techniques employed, did produce results which highlighted that variability in below ground deposits was being detected.

Acknowledgements I would like to thank Tamsin Goldring , Rob Haslam and Bronwen Russell for their help in undertaking the survey.

Bibliography Clark, A. 1990

52

Seeing Beneath the Soil. Batsford, London

1.7. ARCHAEOLOGICAL FINDS FROM CULVERWELL In the processing at present, particular attention is being paid to the immense quantity of industrial debitage, resulting from tool-making on site, and to identifying items showing signs of deliberate re-touch. Items showing signs of retouch are being recorded and counted but all waste products or items only showing signs of usage without special retouch are being weighed; visual inspection leads to an estimation that at least a third of all waste pieces have been used at some stage for some or other activity prior to being discarded. It would be impractical to count waste and utilised pieces because of the immense numbers involved

Introduction

Vast quantities of artefacts of typical Mesolithic types have been found on the site (Figs. 27-39). Surface scatters of these were frst revealed during deep ploughing, as explained earlier in this Report; the scatters were more concentrated in

the immedate area of the present excavations and were subsequently found to coincide with the main areas of activities and features within the habitation site. Excavation showed that these artefacts were, generally, present in all parts of the site and in all layers, even though some types were found to be more prominently associated with some

and, also,

counting

would be meaningless

because

of

the

very large range of sizes involved. Waste pieces generally range from two or three millimetres to ten centimetres, occasionally more (further discussion below) and numbers would therefore not give a meaningful indication of the scale of productivity represented by the debitage.

particular features or areas, as will be discussed.

Archaeological material of all types were more concentrated in the central areas of the site, round the main features and

within the gully. Their decrease coincided with the decrease of midden debris and this was taken to mark, approximately, the limits or peripheries of the activity zone of the settlement site beyond which excavation did not take place. These limits were tested in a number of small sondages outside the areas excavated (Figs. 2,3) and they revealed no shells or artefacts. Artefact distribution up the slope in_ the north of Field 2153 diminished, coinciding with the petering out of the midden deposits and with an area where the Floor appeared to terminate. In an easterly direction, after about

Unstruck pebbles and other stones which must have been intentionally brought to the site from the local beaches or cliff exposures, have been studied by Myfanwy Stewart as part of a larger project and her report is included in this volume (Section 1.8). Many of these stones do not show any signs of utilisation and offering suggestions as to the purpose for

which

they

were

brought

to the

site is, therefore,

a

specialist research project.

12 metres away from the Floor, there was a gradual decrease

of midden deposits and artefacts. To the south of the field the limits of distribution could not be determined with any certainty as excavation was not possible beyond the fence of the field. In Trench 41, the most westerly of the trenches

Lithic Material

All the artefacts so far recorded at Culverell are made from material which can be locally obtained and, almost certainly, were always locally obtained. The stratigraphy and morphology of the Jurassic Portland Beds, the Raised Beach with the deposits above it and the Chesil Beach have been described by many geologists and are discussed in section 2.9. of this Report. These three sources supplied all the material required for the production of artefacts, building material and the necesssities of daily living on the site.

excavated, the occurrence of artefacts in the deeper parts, the

gully,

was prolific but in the deposit of black clay in the

western end of the same trench, found. Artefacts were, therefore,

very few artefacts were collected and recorded

over an area of approximately 30m”. Preliminary assessment of each day’s finds took place on-site and specific tool types and finds of unusual significance were numbered and recorded in ‘Small Finds’ books as and when found. At present the number of these are over 5000. However, as most of the finds required cleaning prior to final identification, they were usually re-assessed away from the site and this work is still continuing, so that final figures could be higher. As full assessment of the total bulk of the material, in particular the debitage, from the site is likely to take some years and can be published separately, it has been decided to incorporate a preliminary assessment in this Report rather than hold up publication of the rest of the Excavation Report. It is highly unlikely that further work on the assemblage will reveal any new aspects or lead to conclusions radically different from what is apparent from the material already studied. It is primarily the figures for the different artefacts types which may have to be adjusted eventually and this does not necessarily exclude the making of valid observations and comparisons at present. A special study of the picks is planned.

More than an estimated 95% of Mesolithic lithic artefacts on the site are made of material from the Jurassic Portland Beds; the flakes and blades are mostly of fine black or grey chert but picks and core-tools are often of coarser Jurassic limestone

with

cherty

inclusions;

an

estimated

3%

of

artefacts are made of Purbeck stone and about 1.5% of Cretaceous flint and, very occasionally, quartsitic sandstone from the beach deposits. (This contrasts with lithic objects from known Neolithic sites on the Isle which are almost exclusively made from Cretaceous flint — see discussion 3.20. this Report). The chert is abundant in the so-called Cherty Series of limestone in siliceous tabular masses or veins which are exposed in the cliffs along the whole length of the Island’s coast (Plates 62, 63). The chert is characteristically grey or black and resembles flint in its general properties and also fractures conchoidally (Palmer 1971). In section Portland Chert often shows numerous spicules of the Jurassic sponge Pachastrella cf antiqua Moore. It varies in texture from very fine, opaque to coarse with gritty, crystalline and fossiliferous inclusion, the first

53

quality being the more frequently used for the finer flake and blade tools and the latter for core tools. Some of the chert, more siliceous, are derived from the Upper Purbeck Beds, a freshwater series, overlying the Portlandian series and exposed in limited areas of the Island. These are sometimes more translucent in appearance than the Portland cherts.

profusion in the solifluction deposits above the Raised Beach. Picks made from the finer cherts are relatively rare. Choppers, chopping tools hammerstones and countersunk pebbles are variously made from rounded nodules of cherty limestone or from quartzite or siltstone pebbles from the beaches.

Artefacts from the living sites tend to patinate black, pale grey to dark grey according to the quality of the piece of material and various (unknown) post-depositional factors. Artefacts which have been in contact with organic deposits tend to be, more often than not, very dark grey/brown to black and often highly glossy and on the Fleet the peaty mud often patinates them to a rich rust-brown. Artefacts

Lithic Artefact Classification

from the lower levels of the site, e.g.

During the past few decades many extensive, but different systems of nomenclature in dealing with lithic material world-wide have been published, particularly abroad (see Brézillon

pebbles

quantities

in

and

cobbles

are

also

from the loam in the

the

Raised

Beach,

found the

in

unlimited

Chesil

Beach

(particularly the Portland end of it) and all the other beaches round the Island. The condition of the artefacts from the Culverwell site (and also Portland Méesolithic Site 1) suggests that chert from all these sources were used for industrial purposes but a more detailed study of this aspect could be worthwhile and could provide a more in-depth insight into the economy and material procurement strategies of the Mesolithic inhabitants.

most,

Dorset,

as “Beer

or

many,

of the

major

sites

in

the

south

and

its usage

is,

therefore,

unavoidable

and

for the

present the only practical solution without re-doing old work. Any controversial aspects relating to classification will be pointed out in this present Report. A summary of the nomenclature and typologies follows and the different artefact types present at Culverwell are represented in the illustrations (see also Appendix 2 this Report).

Artefact Types Distribution Patterns It has become an accepted practice in excavation reports to do a study of the lateral and horizontal distribution of artefacts on archaeological sites, particularly habitation sites. However, because of the nature of a site largely involving a shell-midden (see section 1.3. on ‘Stratigraphy’), it could be dabatable to what extent such a detailed research would be valid or meaningful over the whole area and at all depths in the deposits at Culverwell. Further comments about the general distribution of a few tool types at Culverwell are made in the chapter 3.13. on Activity Zones.

A very small percentage of blade and flake tools were made from flint of which the source would have been pebbles from the beaches, particularly derived from the Raised Beach and the Chesil Beach. Occasionally a piece is found made chert, known

1980

In order to make meaningful comparisons between material from Culverwell and other sites, particularly elsewhere in

There can be very little doubt that pieces of chert must have been broken off from sharp-angled exposures of the material, using hammerstones, along the clffs near the site(s); it is however, unlikely that an actual place where this procurement took place, will ever be identified because of wave action during and since the early Flandrian sea-level phases. It is likely that the initial preparation of large nodules of material took place very near the supply source, possibly on the beach or on rock platforms by the beach. Very few, indeed hardly any, really large cores are found on the habitation sites, confirming that the roughing-out took place elsewhere.

the golden-brown

1980; Jacobi

elsewhere in England.

The cortex, where present, of

beach pebbles tends to be characteristically more battered than that of fresh nodules obtained from cliff exposures. Also, the fracturing of these pebbles are more frequently faulty than in operations which used fresh chert.

of

1972 and Tixier

for England). No consensus of opinion has ever been reached on which is the best and there is no universally accepted system. The system used in this Report is based, in modified form, on the typology devised for the British Mesolithic by Clark (1932,1934a,1934b) and used in Rankine (1956, 1962) and Palmer (1977) for material from the south of England. Although there are various shortcomings in the system, it has also recently been used, again in modified form, by Barton (1992) for the analysis of material from Hengistbury Head and for comparisons with other sites. The nation-wide survey edited by Wymer (1977), to which a large number of people had contributed (including the present author) was also broadly based on the Same system and during the past 60 years has been used for

gully, are often either a dull light grey colour but, quite noteworthy, many from these lower deposits have very little patination.

Chert

1968; Feustel

chert or flint’

from Devon; this had possibly reached the Portland shores by long-shore drift although its excact source is not known (Tingle 1998).

In a midden the effects of post-depositional trampling, redistribution and other activities during the period of use of the site would be largely unavoidable and would distort any findings from scatter graphs to a certain degree. For the

The majority of core-tools were made from the coarser material available on the Island. This applies particularly to the very large number of picks from the site (and from other areas of Portland). Large numbers of semi-finished specimens indicate that they are commonly made from elongated tabular, lentoid or rod-like pieces of Portland or Purbeck limestone or cherty-limestone which occur in

archaeologist,

this

distortion

would

become

further

accentuated by numerous natural factors in the site formation processes, even long after the site was abandoned, especially because of the ‘mobility element’ of a shell midden. Compaction would take place in the lower levels and in the upper levels the material would be moved around by many

54

concentration was in the vicinity of Hearth 4 in the northern

agencies.

part of the site. However, as the cache of picks found there,

included a number which were in varying stages of completion, it merely demonstrates that they were probably

As discussed elsewhere (1.4. and 1.5.) in this Report, threedimensional recording of artefacts did take place in Trench 40 and Baulk 40 but only from Layer 8 midden downwards. Baulk 40 did not produce any useful evidence. It was situated less than a metre from the edge of the floor, which could perhaps lead to the argument that it is unlikely that a knapping area or other specialist activity zone with large quantities of waste would have been sited here as it may have interfered with the activities on the floor.

made

the

d. There is (on preliminary assessment) no

find

does

not

such discernible

pattern for retouched flake and blade tools, being found also

in the gully deposits and everywhere else on the site. This suggests that they were used for a multiple range of activities and were possibly not often saved after the initial usage. They may, therefore, not be usefully indicative of specified activity zones. e. Knapping debitage and other waste material is found in very large quantities all over the site within midden deposits but in the gully they occur in significant bulk as, obvious, discarded waste. Larger pieces of waste material is more common in the deposits within the gully, particularly in the lower layers; this lateral position may, in part, at least be due to downward movement of the heavier artefacts within the soft midden material due to human and natural factors operating after the rubbish was discarded.

be limited

but one came from tbe next Spit down. Either this could also confirm that there should not be too much reliance on the evidence of excavating in arbitrary spits or else it demonstrates that some minor movement within the midden deposits took place, possibly either immediately or shortly after the activity producing or using the artefacts. Long term post-depositional movement may have had the effect of wider re-distribution but did not involve the complete distruption of the midden in all areas.

f. Unstruck or non-knapped pebbles are very common within the midden and in all areas of the site but concentrations of oval or round pebbles particularly occur more frequently at the bottom of the gully or at points where there appears to be some sort of change in the site arrangement, e.g. a small change in floor direction. Some of these pebbles show signs of usage (see e.g. hammerstones, below) but the majority show no signs of their purpose. It maybe that at least some of them were brought to the site in order to be used, but were discarded. Some may perhaps suggest ritual connotations. This situation is fully discussed in section 3.17. and M. Stewart this Report

Several other trenches produced a number of flakes which can be refitted to each other or to cores (Fig. 38). It, therefore, appears unlikely that following the same system of recording and excavating as in Trench 4 over the whole site, would have produced evidence much more significant than is more immediately apparent from the general distribution of artefacts. The following observations may indicate that there certain artefacts

fire;

the square formed between the floor, pit and windbreak, possibly suggesting that they were associated more often with everyday living activities and were not so often discarded in the rubbish heap. This requires further research after the final microlith assessment.

conjoined flakes in one quadrat, C2, three were from Spit 1

areas in which

to the

on the westerly side of the site, most of them being from deposits on the floor and east, south-east of that feature in

(Fig. 24). Of more interest was the fact that flakes were found which can be refitted to one nodule (Plate 56). Although this does indicate that they were worked in the near-vicinity of this area, they also did demonstrate upward and horizontal movement within the midden: of four

some

next

c.The preliminary assessment indicates that very few microliths came from the midden deposits within the gully

The dense midden was excavated in arbitrary spits of approximately 2cms depth each; this method indicated (as was expected) that the artefact distribution pattern varied greatly throughout the midden layer and any meaningful

were, however,

place

demonstrate unequivocally that they were used there. A preliminary study does, however, suggest that picks are more generally found in clusters in the large area adjacent the floor on the easterly side of the site and more randomly within the midden deposits in the gully.

Trench 40 was chosen for this project as excavation of the upper layers of this trench by the end of the 1988 season had suggested that the deposits in the northern part of this trench, where they were buried under the medieval lynchet, had not been too greatly disturbed since the Mesolithic occupation except by, possibly, rodents. Also, the terrain covered by this trench is well away from the floor and gully (which, as described, was the main rubbish dumping zone) and is near the edge of the pit; it was argued that it may, therefore, have been a locality where the site inhabitants sat to take part in some or other activity associated with daily living.

interpretations of these patterns must, therefore,

at that

Preliminary discussion of types of lithic artefacts

were

more common than in others:

Debitage from knapping procedures:

a. Large pebbletools, e.g. choppers and choppingtools, and hammerstones were more commonly found in concentrations in the vicinity of hearths and may therefore often have been used in food preparation activities.

The most substantial proportion of lithic artefacts consists of industrial waste or debitage, i.e.

blades, bladelets, flakes,

cores and irregularly shaped pieces of struck chert, with or without areas of cortication. These are found in such vast numbers that counting them was regarded as not feasible or meaningful and they are being recorded by weight, as giving a better indication of the scope of industrial activity on the

b. Picks are found in large numbers throughout the site and in all areas and layers of the midden but the largest

55

damage tends to show no_ decisive ‘pattern’ of multiple retouching in order to achieve a particular effect. This category, would, of course include a substantial quantity of material which became edge damaged as a result of accidents, e.g. a flake or blade being stood on, and which is often very difficult to distinguish from utilisation damage. The difference generally is that marks produced by the utilisation of an artefact’s edge, i.e. involving at least a

site. The material varies in size from 2mm (so-called chips) to pieces of 3 - 8cms on average, but bigger pieces do occur. An estimated quarter of the waste still retains areas of cortex, i.e. were involved in the primary knapping activity.

A blade is generally taken by most authors to be an artefact where the length is two times the width whereas a flake has a lesser length to width ratio. There can be a problem in classification where a blade has been accidentally or deliberately reduced in length and for this reason some authors prefer to give the length to width ratio as 3:1 (Palmer 1977). In a blade there is also frequently a visible element of laminate working, i.e. the producer wanted a tool with reasonably straight edges and modified his knapping technique (fr. Lamelle).

This category of blades have been this process, as trimming-flakes struck

from

the

waste includes removed and well as the can either be

‘usable’

of the

base

of the

core,

easily obtainable;

of the waste

tends

to

be a

little

more

flakes

and blades

of usable

sizes

were

Retouched flakes and blades

This, rather vaguely classifiable, category of artefact is used for material which shows signs of having been deliberately retouched, by percussion or pressure flaking along one or more edges or part of edges (Figs. 33, 34). Although, strictly speaking, all identifiable specific tools should generally be included

in this classification,

artefacts

where

the term is often

no specific purpose

used

for

can be defined from the

nature of the retouch, but which shows a more visible degree

of planning or deliberate industrial activity than in the previous category of artefact. It is likely that the majority of the tools belonging to this category were general purpose tools and were used for cutting, scraping or some other unspecialised activities and were probably produced on the spur of the moment for an immediate task without the producer working round the edges in a __ thoughtful methodical manner as for, e.g. a specialised scraper.

Utilised waste:

at least half

edge,

struck from usually small (fairly lightweight) chunks brought to the site, were promptly used for the immediate task and were then discarded, with a new piece being used for each subsequent task. Secondary retouch was only employed for more particular tasks.

The cores from the site tend to average between 5-—7cms in height but smaller, micro-cores with laminar scars, do occur; bigger cores up to (estimated) 10-12cms are relatively rare and when they do occur, are found mostly in the bottom layers of the midden in the gully. There could be several explanations for this situation: a. the first settlers on the site produced their tools from bigger cores because of some or other factor in the procurement or production Strategies, e.g. doing the rough-outs also on the site, while later people did this elsewhere, perhaps on the nearby beach; b. the difference could be indicative of technological development traits, maybe with chronological implications; c. bigger cores, being somewhat heavier, had a tendancy to sink down lower in the loose (possibly wet) deposits in the gully; d. for some obscure reason, the site occupants by choice threw bigger material into the gully while smaller material was often left lying around the site (e.g. for the same reason as throwing pebbles into the gully?). This aspect also requires further assessment after the final analysis of the lithic material.

is estimated that

the

Of this category of artefacts, on a rough estimate, about a third shows signs of having been’ extensively utilised, resulting in the blunting of the edges. Chert, being a very hard material, is ideal for cutting activities without any further retouch to the razor-sharp edges created by knapping. The vast number of debitage which were utilised on, possibly, not more than one ocacasion for the completion of one particular task, is compatible with the fact that the site inhabitants had no need to be frugal with the usage of material and artefacts. The chert supply was limitless and

plunging to the apex, or else the basal striking platform can be rejuvenated by the removal of a lateral flake. A flake which removes all or most of the core’s flaking surface is sometimes described as a flanc de nucléus (e.g. Brézillon 1968). At Culverwell (in the present project) no separate records of the different types of trimming flakes have been made but may be included in future research projects.

It

of

consistent than the completely irregular haphazard removal of a number of spalls resulting from an accident. This differentiation must not be regarded as a rule but merely as a working guideline in forming an opinion and may often require further examination, such as microscopic analysis (which in itself can often be quite inconclusive or uninformative).

cores from which flakes or which still retains the scars of core-trimming flakes. These of the plunging variety, i.e.

circumfereance

portion

material,

particularly primary or secondary flakes and blades over approximately 2cms in length, have been utilised, or damaged by utilisation, often on all sides. Smaller pieces with edge damage may, often, have been accidentally produced from bigger pieces. Cores often show signs of having been used as hammerstones after flakes/blades were struck from them.

In this Report the category of retouched flake/blade will not be used for pieces with possible accidental or so-called ‘spontaneous

retouch’;

these

will

be

classed

with

the

previous category of waste material. Picks, Axes and Related

Core-Tool Types

These tools have been found in unusually large numbers at Culverwell and at Portland Mesolithic Site 1 as well as at other known unexcavated sites containing concentrations of

Utilisation is usually indicated by a series of small irregular primary spalls which have flaked off the edges; this edge

56

Adzes or Planes: this is another core-tool type similar to picks but always with a flat undersurface and the working edge is formed by an upper surface intersecting sharply at an

Mesolithic material on the Isle (Figs. 28,29; Plates 48,49 and

Palmer 1977). At the time research on the Isle of Portland started, many archaeologists associated tools of this type invariably with the Neolithic period but the evidence from Portland showed conclusively that these tools were in use much earlier. At the site of Pinnberg, Schleswig Holstein, picks similar to the Portland specimens came increasingly into use from a Late Palaeolithic stage, Younger Dryas/PreBoreal phase (Rust 1958); they can be regarded as being a response to changing environmental conditions confronting previous reindeer-hunting groups after edible plants became part of the diet.

angle with the undersurface,

or sub-rectangular. One end may be extensively retouched to form a sharp angled edge as in present-day chisels or planes. These tools are suitable for working wood but they can differ little from the core-scraper type of tools mentioned already. Not many tools from Culverwell can definitely be regarded as adzes. Axes are relativelty rare at Culverwell, being a tool primarily intended for cutting wood or maybe fracturing bone in a neat manner (other than just smashing it) or laboriously cutting it. However, more than 20 tranchet (transverse) axe trimming flakes have been found on the site, 8 of these being found near each other by the pit (Fig. 37). This scarcity may suggest a limitation on the usage of coretools on the site or could suggest that the axes were used in activity zones away from the site (although no concentrations of them have yet been found on the Island)

The tool is made from an elongated or tabular piece of material, on Portland usually cherty limestone but occasionally examples of quality chert are found. On Portland they vary between 10cms to 17cms, with the majority falling at the lower end of that scale and only rare examples exceeding the higher dimensions. The body of the tool can resemble an axe in form but at least one end is usually tapered or pointed; in section picks are triangular, quadrilateral,

sub-oval,

lentoid

but

also

assymetrical.

A pick can very easily be turned into an axe by removing a transverse or other wide flake from one end of the long coretool in such a manner as to create a broad sharp working edge. Very few tools from Culverwell can be classed as axes, one example being the tool found in the so-called ritual feature (see 1.4., 1.5. and below 3.17; Plate 36). No typical

Sometimes the point shows signs of having been resharpened by directing a transverse blow at a pointed end, producing a characteristic pick- sharpening flake, very much in the style of

a transverse

axe,

also characteristic

of the

Mesolithic

period (Fig. 37).

Pieces of natural material which would be suitable for the manufacture of picks (see description above) or which have had a few preparatory flakes removed from them, were also kept and recorded as “pick material’, especially when found in clusters with fully completed picks. Examination of the picks indicate clearly that very often a piece of tabular material would require only to have one or two flakes removed from an end, which may already be naturally slightly pointed, in order to become a usable tool.

flake axes, such as found in the Ertebglle culture of Denmark

have so far been noticed in the Culverwell assemblage although three were found at Portland Site 1 (Palmer 1977). Axe and pick-trimming flakes

These are flakes removed from the one end of either a pick or axe and they retain some of the tool’s original flaking scars and also part of the original working edge or point (Fig. 37). A sharpening flake rmoved by a transverse blow to the length of an axe, near the working end, produces a specific type of tool, namely a tranchet axe. The same can apply to a pick.

It is suspected that this category of core-tool includes artefacts which were sometimes used for activities other than ‘picking’, but the term will be used here pending much more lengthy and detailed research in the future. Specimens are found with extensive secondary retouch along one or two edges and which would probably have served as large core scrapers or push-planes held in the palm of the hand. It is, of course, possible that these examples may have undergone several phases of secondary retouch, may have initially been designed to be used for picking activities but, when blunt, was made into another type of tool. It is estimated that more than 300 picks have been found at the Culverwell site, but this figure requires confirmation after the final count.

Choppers/chopping-tools/ limpet scoops

A large number of core-tools have been found at Culverwell which are characterised by the removal of a _ few flakes from round or oval beach pebbles or cobbles (Fig. 35). They usually retain the nodular shape and large areas of cortex. Some of these are unifacial and the cutting edge, formed by the removal of one or more flakes worked in only one direction, intersect at an angle with a natural surface. These have been classed in the records as choppers. Tools have been classed as chopping tools when they have bifacial cutting edges with the working edge formed by the intersection of flake scars struck alternately in two directions round part of the perimeter of the cobble. These tools, large

Fabricators can be made in the same basic forms as_ the picks. They tend to be usually small, averaging 7 to 8cms and has a neat, sharp point, but it is recognised that it is not always valid to make a disinction between a small pick and a fabricator.

Also,

some

fabricators

i.e. in section it is rectangular

do not have

and small, have been found in some numbers in the immediate vicinity of the Culverwell hearths, thus indicating

the body-

a link with cooking or food preparation activities.

shape of a pick and can be made on a thick flake, although this type is rare on Portland; this suggests that picks may sometimes have been used in the manufacturing process of other tools. Some of the smallest picks from Portland may, in fact, have been intended for use as fabricators or may have been so used.

Not so common are tools classed (with some strong reservations) as ‘limpet scoops’ (see M. Stewart). These tools are usually rounded or oblong pebbles, occasionally tabular, with a large, deep flake scar at one or both ends.

57

This scar has the effect of forming a scoop-like concave working edge. They can be used to scoop up food products. It is possible that the name is a misnomer and they should perhaps better be referred to merely as ‘food scoops’ or even just “scoops’. It should be kept in mind that limpets very often do not require a special tool for removal; a quick, sharp blow with a pebble or even with the side of the hand can be successful if it takes the animal by surprise. It is likely that in many cases the deep scar is fortuitous as it was actually created by the user aiming a blow at something or other.

Specific flake and blade tools

It is generally felt by academics that a great deal more research requires to be done on these categories of artefacts. It is recognised that on a site these tools could have been used for diverse or multiple activities and that assigning ‘purpose’ to a tool on the basis of morphology or technology may be misleading. A functional analysis, i.e. use-wear or microwear analysis,

for instance, denticulations However,

Perforated or holed pebbles have in the past often been referred to as ‘mace-heads’ but it is widely accepted that this too is probably a misnomer and that the tool was most used as a weight on a

Scrapers:

stick, i.e. a

digging-stick weight (see Plate54,55 and Fig. S5). They are made from rounded or oblong beach pebbles, occasionaly iregularly shaped tabular pieces of stone, and occur in a wide range of sizes but on Portland they average 6 to 10cms in diameter. They are made by boring from alternate faces of the pebble, creating an hour-glass perforation (a another possible use for burins or borers). Although a number of these tools have been

found

on the Isle of Portland,

at

that

These

can

they are tools in

as

‘unstruck tools’

and

as

such,

are

classed

as

such

severely abraded or battered surfaces,

when

pebbles

that this technique

of lithic

Culverwell,

even

the most common if

not

considering

flake/blade tool a_

substantial

the

common,

as indeed

found to be also the case

on Site 1 (Palmer 1977). These scrapers are sometimes made on naturally fractured lumps of stone and the angle of retouch is near-vertical or sometimes even vertical, often with undercutting. Some occasionally occur which can be described as carinated (beaked). Successive removal of trimming flakes from cores causes platform abrasion but sometimes this abrasion is so regular that there is a probability that the core was deliberately retouched round the edge of the platform so as to become useful as a core-scraper (Fig. 36). A distinction between this form of scraper and the previous two types is not always practical.

is

further discussed in M. Stewart’s contribution to this Report. Hammerstones

Scrapers are

scrapers are

or abraded pebbles

be classed

found

end they are classed as end-scrapers (either the distal end or the opposite proximal or bulbar end). The end retouch can form a_ short convex edge or straight transverse edge. At Portland the end-scrapers are less common than the side variety. Convex-scrapers are so classed when a flake has been retouched round the edge in such a way that it results in a tool with circular or semi-circular working edge (sometimes classed as round scrapers). The working edge of a scraper can also be hollow or indented (concave-scrapers) and are sometimes regarded as ideal for use as ‘spokeshaves’. They are not a prominent tool at Culverwell. Scrapers made on thick chunky flakes, i.e. steep-edged

none

a specific category. They have two opposing hollows or depressions, usually quite shallow and seldom more than a centimetre deep, made by a pecking or hammering technique. The circumferences of these tools are sometimes abraded, indicating that they were used as hammers, held firmly between the thumb and forefinger. Hammerstones

been

end they are side-and-end-scrapers; if retouched only at

About 6 or 7 countersunk pebbles have been found at Culverwell. It is possible that some of these were intended to become perforated pebbles but that the perforation was not however, it is more likely

it has

proportion of the retouched and utilised flakes/blades which may have served as scrapers (Fig. 31, 32). A preliminary assessment indicates that scrapers with one or two sides fully or partially retouched (side-scrapers) are the most common. When blade scrapers have been retouched at both sides and

with a completed perforation have (yet) been found at Culverwell. Pendants: Two small artifically holed pebbles and two which appears to have been naturally holed from the site were possibly used as pendants (Fig. S.7).

completed;

that what appears to be a ‘saw’ with or serrations was really used as a scraper.

analysis does not always lead to conclusive evidence (e.g. Moss 1983). As yet no microwear analysis has been undertaken on any of the Culverwell material but this may become a consideration for the future.

Perforated or countersunk pebbles

likely, more often than not,

can be expected to reveal,

have

sometimes round the

edge only but often elsewhere on the face of the pebble as well. The tool surface often displays incipient bulbs of percussion and numerous small, often scaly and overlapping flake scars caused by a vigorous hitting motion. The term ‘abraded’ is used when the tool surface has one or more areas with a finer bruised appearance caused by gentle repeated tapping, i.e. it can be argued that the distinction between the two types is a matter of degree. On Portland they are mostly quartzite or hard siltstone pebbles from the Chesil Beach or other beaches round the Isle.

Many Mesolithic sites are characterised by tiny very neatly retouched

micro-scrapers,

usually

less

than

diameter, but they are not common at Culverwell. common are nosed-scrapers where the working convex but is formed by a constricted or ‘pinched of the sides. They are generally more common Palaeolithic contexts and at Culverwell it informative to note that a few have been recorded bottom layers in the gully.

58

2

cms

in

Also not edge is in’ shape in Upper may be from the

Burins/gravers

tip formed by two retouched and converging edges (Fig. 33 no. 16). The retouch may sometimes be abrupt. They are

This is a tool, of which there are several varieties, classified

suitable for making a small hole in, for instance, a piece of leather, bone, shell or also a stone. Holes can be made by two

according to the number of burin facets and the morphology of the surface on which the point of impact for the removal process is situated (e.g. angled, lateral, retouched,

different techniques: a straight forward-and-backward movement of the hand holding the tool or else by a quarter or

concave,

convex, straight or plain and unretouched) (Fig. 30). Its recognition sometimes present difficulties and it could be overlooked in assemblages. It first appears in large numbers in Upper Palaeolithic assemblages and was used in quantity during the Early Mesolithic (e.g. 225 at Star Carr), becoming rarer in Late Mesolithic and quite rare in Neolithic tool-kits (Brézillon 1968). It is a

tool,

more

often

a flake or blade,

from

which

half circular movement

a

Knives/saws

Most sharp flakes or blades can be used to cut most materials, whether they are specially retouched or not. It is likely that the majority of the flakes/blades classed above as utilised or retouched were used for this purpose. For this reason some researchers do not specifically include these tool-types in lists of Mesolithic artefacts. When a flake/blade has been specifically retouched for use as a knife or a saw, the edge retouch ranges from fine, with small regular nibbly spalls removed from the edge, to bold and serrated (Fig. 33 no. 1-8) These are described by some workers as microdenticulates when made on bladelets. Blades/flakes (of

In the Mesolithic period there are only two or three common types: the first is the angle burin with a single burin facet with the point of impact at one of the edges of a retouched, opposing, lateral edge; the angle can be acute or obtuse. Sometimes this type can have two burin facets, usually one each at two different edges at the same end of the flake or blade. The burinal facet sometimes runs for the whole length of the flake/blade’s edge but sometimes it terminates a short distance from the point of impact on the striking platform (or spall removal surface). In the latter case, the spall sometimes ‘curves outwards’ with a shallow hinge fracture, depending on how far the point of impact was from the edge of the flake/blade. This edge often shows signs of having had several successive resharpening spalls removed at the same point and this feature may be an additional aid in positively identifying the tool. The burin facet can also oppose an unmodified natural lateral edge, e.g. of a primary flake or it can oppose an edge which was created as a result of a break. Another type of burin, reasonably common in the Mesolithic period (but less so than in the early periods), is the diihedral burin where the spall is detached from an edge which

all

being

so

retouched,

are

sometimes

also

classed

as

A number of artefacts with pronounced serrated edges have been found at Culverwell but they are not common. Blunted blades/flakes (Backed blades/flakes)

From Late Palaeolithic times several categories of tools are characterised by edge blunting retouch. This is a technique whereby the angle of retouch to the edge is abrupt, 90° or only slightly less; several spalls may be removed from the edge at the same point to achieve the effect whereby the edge is no longer sharp. In some artefacts the retouch is bidirectional (from both the dorsal and ventral surfaces), when it is sometimes referred to as ‘anvil retouch’.

carinated burins and polyhedric burins are relatively rare in the Mesolithic periods. Burins are not too common at Culverwell but they are also not rare, occuring in most parts common

sizes)

‘serrated flakes/blades’. Only when this specific retouch is present, will an artefact from Culverwell be counted in these categories of tools.

already contains a burin facet; the two burin facets then form an angle to each other. Double burins, nosed _ burins,

the most

It should, however, be

realised that other tools (e.g. a pick or even the rod-like microliths) or an unretouched piece of waste with a fortuitously sharp point can be used for this purpose. A sturdy piece of wood or bone could also be used, depending on the strength of the material to be pierced. This probably accounts for the relative scarcity of the tool on many Mesolithic sites but they do occur at Culverwell, in smaller numbers than at Portland Site 1.

narrow splinter of flint or chert (burin spall) has been removed in such a way that the tool has a working edge sometimes resembling that of a small chisel. There are several varieties of burins in the Upper Palaeolithic (particularly Magdalenian and Late Palaeolithic cultures; Fig. 34 no. 17 shows a type reminiscent of the Late Palaeolithic period).

of the site and in all layers,

of the hand.

It is probable that the majority of these artefacts were used as knives,

the

single-blow varieties.

scrapers,

saws

etc. with the forefinger held on the

blunted edge to prevent injury to the user. These blunted edges can be either straight or convex and the opposing edge is sometimes retouched with ordinary secondary retouch as found in the flake/blade tools described above. A problem in assigning possible usage is presented by the few artefacts where the blade/flake has both edges blunted so that there is no visible manner in which it could have been used.

A problematic type of burin is the ‘core burin’ (particularly on the types of cores with two striking platforms on the same plane); a core with a facetted platform could have the appearance of being a burin when regular flake scars have been removed from edges towards the centre. The scars on micro-cores can often look like burinal scars.

Some flakes/blades are retouched by a blunting technique across an end, either the distal or proximal end (Fig. 33 no’s

Awls/borers

17,18, 20).

This end can be straight, convex or

concave and

transverse or oblique to the length of the tool. These are sometimes referred to as truncated flakes/blades. Blunted

This tool is sometimes also referred to as a piercer (bec) and is characterised by having a narrow, fairly long sharp-pointed

59

Tanged points are numerous at Hengistbury Head and at least one of them was made of Portland chert (Barton 1992; Palmer 1970).

blades occur fairly often at Culverwell but it requires to be considered whether some of them may belong to stray remnants of a Late Palaeolithic industry present on the site (see below). The blunting retouch is also used on several specific tool-types, discussed below. Blade segments are found (including Culverwell). This present

difficulties

in

In the Mesolithic period macrolithic tanged and shouldered points become rare and at Culverwell they tend to occur mostly in the lowermost layers (Layers 10 to 13) and may be indicative of pre-Mesolithic activities on the site. Twenty or more have been found (Fig. 34 no’s 4-16). The tang is not always very pronounced. It is possible that these were lying on the old pre-Mesolithic landsurface when the first Mesolithic inhabitants arrived on the site or else they are examples of the continuation of Late Palaeolithic typological traits into the Mesolithic period. They are associated with quantities of large waste, utilised or retouched flakes or blades, often in the bottom of the gully, e.g. in Trench 41.

on many Mesolithic sites category of artefact could

classifying,

if , indeed,

it should

always be classed as a specific tool-type: if interpreted literally, this nomenclature would apply to all pieces of blades even those accidentally broken off from a blade, but in this Report a specific type of artefact is indicated. This is a centre piece left from a blade after the bulbar end and the tip end have been removed, either by percussion or snapping. One edge is usually retouched and often the opposing edge will have been worked with the blunting technique. They could have been suitable as a composite tool, inserted in a groove in a piece of bone or wood (Fig. 33 no. 4).

Shouldered points are also more commonly found in Late Palaeolithic assemblages and some have also been found on Portland at the Verne and on Mesolithic Site 1 (Palmer op cit.). It is therefore of interest that both shouldered and tanged points appear together on the Isle in the same assemblages, as also at Hengistbury Head (Barton 1992). These points are also made by abrupt retouch but this is done along one edge only, usually at the proximal end of the blade/flake, thus forming a shoulder. There is often also an area of retouch at the distal end, usually opposite the retouched edge of the shoulder.

The so-called Petit tranchet arrowhead is made from a blade segment of which two edges transversely to the original edge of the blade have been retouched after snapping, i.e. the retouch is along the snapped edges, and is done by abrupt or semi-abrupt retouch. A part of the original edge of the blade then becomes the working edge of the tool. The artefact is often trapeze-shaped and can be broadbased or narrowbased.

Microliths

Petit tranchet derivative arrowheads are not common in Mesolithic assemblages and somearchaeologists regard them as beng (always) of Neolithic date. One has been found at Culverwell in Trench 4 but as there had been some animal

The nature of these artefacts are often misunderstood by non-specialist archaeologists. The name is sometimes erroneously taken to apply to any small (micro) piece of flake/blade or tool but it should only be used for a specific tool-type, demonstrating special technological characteristics (Fig. 27; Plate 52).

disturbance in that trench, there could be a small degree of

doubt about its age (Fig. 39 no. 9). They are also made from blade segments but here the original edges of the blade are retouched to produce a trapeze-shaped tool; one of the snapped edges is then sharpened by a transverse blow to form a working edge. In the Neolithic period these tools frequently have fine incursive retouch.

The majority forms of microliths are made by the microburin technique whereby a blade or bladelet has a ‘notch ‘ removed from one edge by abrupt or semi-abrupt retouch visible (usually but not always) on the dorsal face so that the piece fractures obliquely, more frequently near the butt or proximal edge. On the ventral (bulbar) side of the bladelet the fracture characteristcally results in a flat roughly triangular scar, sometimes resembling the shape of a droplet of water. If the impact creating the notch is made from the

More research is required on the chronological development and position of these last two tool-types. Late Palaeolithic

contacts?:

Tanged and shouldered Points

A distinction must here be made between the macrolithic and the microlithic shouldered and tanged forms (Forms G — see below) which occur in Mesolithic assemblages. Tanged points flakes belong

(pointe a pédoncule) to the varieties of

ventral face, the scar will be on the dorsal face, a form which

is not very common. Less frequently the microburin (a waste product) is removed from the distal end of the blade, a technique used often in making quadrangular or trapezoidal microliths. A double-microburin is the resultant waste product when two microliths are made from one bladelet, one from both the distal and proximal end. A’ microblade showing evidence of unsuccessful division, or where the snapped end is transverse instead of oblique, is classed as either a micro-burin miss-hit or sometimes as a’ micro-

on blades or larger points which are

characteristic of the Late Palaeolithic and are found, for instance, on Hamburgian, Ahrensburgian, ‘Cheddarian’ and

‘Creswellian’ sites and have previously been found on Portland (Brézillon 1968; Palmer 1968 and 1971). It is usually the proximal end of a long blade (less often a flake and very rarely the distal end) where abrupt or semi-abrupt retouch on both edges has been carried out so that the artefact has a protrusion or tang formed by two concavities. Often part, or the whole, of an edge has also been retouched and sometimes just the tip of the tool opposite the tang.

burin allied form’. In all these cases,

the resultant microlith

will lack the bulb of percussion.

Strictly speaking, the classification as ‘microlith’ should only apply to tools where the bulb of percussion has been removed by the microburin technique, as described above;

60

Dib

Dia

Dla

some workers, however, do include in this category tools with the general morphology of a microlith but which retain the bulb. While it is recognised that a blade of possibly the same nature and function as a microlith can be fashioned by different techniques, in this Report ambiguity will be avoided by classifying bladelets with the bulb intact as ‘blunted blades’ (see above) rather than as microliths.

above,

and

in addition

would

have

one

&—-

-

H

MHS

—»—,

The first detailed classification of this tool-type was undertaken by Clark (1934 and 1939) but has since been modified by many workers in England and abroad, even where Clark’s basic system is used (see e.g. Brézillon 1968 272-273; Jacobi 1976). The descriptions given by different workers can vary in some details but there appears to be a reasonable agreement about the types found in most Mesolithic assemblages. They fall into two main categories: geometric (triangles, trapezes and rhombic forms) and the rest non-geometric. It is often argued that the first class is more prominent in the Later Mesolithic phases but this issue remains arguable and ecological factors requires to be also considered within the time zone concepts.

or two

edges where microlithic abrupt retouch is present. The form is More common where long scalene triangles are numerous.

After a piece of bladelet has been detached from a longer blade, the remaining longer piece of bladelet (without the bulb of percussion) is then fashioned into a microlith, a tool of which there are a number of varieties, not all common

G3

DS

tools may have been projectiles and examples have been found preserved in organic sediments (e.g. Gramsch 1987) and in the bones of animals (Clark 1975). Some forms would have been suitable for use as sickles for grass cutting with, for example, several scalene triangles arranged in one groove in a wooden shaft.

Krukowski microburins are discussed in detail by Brézillon (1968: 129-130). These artefacts (a specific waste product) can be created accidentally during the retouching process of a microlith or they can be created when an existing microlith or blunted bladelet is re-used to create another form of microlith by the notch technique. The resultant waste piece would then be characterised by having micro-burin facets, as described

D2b

For reasons already explained above, the classification, numerical with descriptions, used by Clark will be used in this Report; where applicable, nomenclature in general use by other authors, will also be given and explained (see e.g. Barton 1992 for Hengistbury Head).

on

all sites (illustration above). Retouching can be unidirectional (one face only, usually showing on the dorsal face) , inverse (on ventral face only) or bidirectional , the latter two being rarer.

Form A - this form of microlith is found in virtually all Mesolithic assemblages and can be regarded as the most

Usage

basic.

From it other,

more complex,

forms can be created

by additional retouch and it is therefore possible (or likely)

There is no real consensus amongst archaeologists regarding the usage of microliths and it is highly likely that each and every varied form may have been usable for more than one different task. In most instances their dimunitive sizes would have made it virtually impossible for them to have been held in the hand during usage and it is reasonable to assume that they were intended for use as composite tools, hafted in grooves made in wood or bone. Some of these composite

that

when

found,

at least some

of

them may

have been

intended for the production of other forms but were not completed (Rankine 1956). However, it is generally accepted that this form was also a_ specific tool-type. The microlith consists of a simple blade, usually with the point intact, and which have areas of blunting retouch along part of one edge. The retouch can be anywhere along one edge,

61

but more often is obliquely at the tip when most authors then refer to this type as obliquely blunted points (pointe oblique). Forms B

are also found in

at Culverwell.

all Mesolithic assemblages,

as

There are several varieties of this form, which

have in common the retouch along of the blade.

(where the microburin facet was) blunted; in the first the base is transversely retouched while in the second form the base is convex or, occasionally, slightly pointed. Only one dubious example has so far been recorded from Culverwell (not illustrated).

Form B1,

at least one whole edge

the simplest form,

Forms F are similar to the previous form in that the base is retouched; Clark (1934 and 1939) referred to these microliths as Horsham points or hollow-based points and he classified them into four different types: Fla symmetrical, flaked from above (i.e. from the dorsal face with the retouch showing on

1s identical to the

A form except that here the whole edge of a blade is blunted.form B2 has two edges blunted and can be lanceolate or rod-shaped..

the ventral face), Flb

Some forms of retouched and edge (Palmer Culverwell they

microliths on Portland have one whole edge only part of an opposing (parallel) straight 1977). In tabulated inventories from will be classed as Form B3..

Portland Site 1, Culverwell.

at the proximal end. In form Cl the basal blunting is oblique to the edge while in form C2 it is transversely. The latter form is not as numerous at Culverwell as at Site 1.

constitutes

more

than

one

Forms

D2a

and D2b

are both

by

crescents,

sometimes

In Form D2a the

recorded

from

three

small

(not

of it, i.e. between the floor and the windbreak in the south.

This would still not answer the question: discarded microliths not thrown into the gully?

are microliths with the proximal

why

were

very

As regards the distribution in the different layers of the occupation midden, it has already been pointed out that, due

distinctive) examples have been found at Culverwell.

Forms El and E2

been

As pointed out above, microliths and microburins appear to be rarer in the midden deposits in the deeper parts of the gully on the western side of the floor on the habitation site. It would be difficult to regard this as purely coincidence and also it cannot have only chronological significance as all the other Mesolithic tool-types are present here as on the rest of the site. It may be feasible to suggest that this distribution had some relationship to activity zones: microliths were used for activities on the floor and in the large open area east

also

Forms D4. These geometric forms are foursided and have two or more sides of blade segments blunted by abrupt retouch to form rhomboids, regular or irregular. They were subdivided by Clark into forms 4-6. They are rare on England;

have

Microlith distribution and numbers at Culverwell

bladelet. Form D2b has the chord of the crescent blunted. Neither of these forms are common at Culverwell. (Form D3 — see above under triangles).

in

1, none

Unclassifiable microliths (U). On most prolific sites microliths are found which evades definite classification. There can be overlaps or variations of the above forms and incomplete or broken microliths present statistical problems; in this Report dubious specimens will not be assigned to any specific type.

it is likely that this form of microlth is formed

sites

at Site

Forms Hare trapezoidal points with two edges retouched and which can be narrow-based (H1) or broad-based (H2). These forms can also be further sub-divided into points with Straight or curved edges and symetric or assymetric. None have so-far been recorded from Culverwell.

a

the removal of a microburin spall from both ends of a

Mesolithic

found

as barbs for a harpoon-

referred to as lunates (segment de cercle). arc is blunted;

was

Culverwell.

third of all microlith forms. Some specimens of this form are blunted on all three sides of the triangle. A problem arises with some triangles which do not run out to a_ retouched point but terminate in the original unretouched tip of the microblade; quite a substantial number of these are found at Culverwell. It is likely that many of these specimens are scalene triangles of which the point has accidentally been broken off or where the retouch has not been completed, but it is also possible that some form a specific sub-class of sub-triangle (Clark’s Form D5 but in Palmer 1977 classed as D3). Triangles are the most likely microlith type to have been used, at least sometimes, like tool, hafted in a wooden spear.

Palmer 1977). Five have been recorded at

Forms G_ shouldered and tanged points. This form of microlith must be distinguished from the macrolithic shouldered and tanged points which are described above and which are common in Late Palaeolithic contexts. The microlith forms are made by the notch technique. Although

Forms D - geometric. These forms are also produced by the microburin technique, involving more extensive abrupt edge retouch. They appear to become more common in Mesolithic Form D1a is an isosceles triangle and is not at all common on the Portland sites. The scalene triangle Form DIb is the most common form of microlith at Culverwell and, assessment,

(i.e.

numbers (e.g. Thatcham, Berks., Iwerne Minster, Dorset and

and also blunting across the base, usually, but not invariably,

to present

flaked from below

from the ventral face with the retouch showing on the dorsal face), F2a assymetrical flaked from above and F2b assymetrical flaked from below. These were microlith types originally regarded as very late Mesolithic and characteristic of the sandy heath areas of the Wealds of Sussex and Surrey. However, they have been found elsewhere, albeit in small

Forms C I and C2 Both these forms have one curved or straight edge blunted

according

symmetrical,

end

62

Table 1:

sons

om

corer

Tren ch

ae oS= rarerefnstraoness ia

ae = ronenees and _

0

ae

1

|

6 |i Oe 7 (6 |2/ | |2] | jaja | ja} | a eee 1 — | eC ‘Wot2 wo} Asia 4013 15 jeoojt 17 OO 19 20 | mu} 22 | 23 |

S__

| 4A ee

|2 ia} | | | | | | |] | 2} | | | | | | ee a | | | |2 | | | | | | |

i

8 6 | | ]3 | 15

— ja}

|

8

ja ls a8

| a 8 | | |8 (2) | jm) | | ila | a | | a a a | | | | | | | } ff} Pf Oe ee 0 Oe OG DO a a a ee ee | | ft a | |

10

= el

ois it

fa s_

s 3

Mb 597° Allied 856 Kmb 358;

*2 mliths D2a white flint found together in T14 L2S1; U=unclassified

63

R=rhomboid

Table 2 Microliths: Totals from different layers of Trench 4 (Layer 1 includes surface finds from the trench. There were no microliths in Layers 12/13)

L}L) LIL) Lv IL 1) 2;3/4!15 | 6 I

]

Table Microliths 1 gives the totals of all microliths found on the site per trench, i.e. the horizontal distribution. Here some differences are noticeable, as already pointed out above: the majorityof microliths came from trenches 1 — 15; these include the trenches in the area of the floor and east of it and microliths can be seen to become less numerous in trenches sited further from the floor or in the area of the gully (compare with site plan). As the deposits on the site have been shown to be well stratified despite possibly some postdepositional movement, the microlith evidence tends to confirm that at least a number of important activities were centred round the floor.

Lyi ylj{L Lye 7 |8 |9 | 1} 1 0} 1 ] ] 2

Al

Although the microlth totals do not appear to be very high, it should be kept in mind that only a small section of the midden under the floor has been excavated (i.e. the Sondage

in Trench 6 — see Site Plans) and excavation of the midden under the floor could conceivably substantially alter the

microlith count. Also, of course, there is the possibility that the daily living conditions at Culverwell may not have required a large number of microliths to be in use; picks may well have been more important for the actvities on and round the site, reflecting an economy which may be based more on the harvesting of molluscs and edible plants. The most common microlith on the site is the scalene triangle Form D1b followed by the non-geometric form A1. The isosceles form of triangle is relatively rare as also crescents, rhombic forms and Horsham points with the blunted

base.

There

are

no

Forms

D2b,

suggesting a rather limited requirement diverse possible usage.

or H

for microliths

of

A total of 597 microburins have been found on the site,

86

miss-hits and 358 Krukowski micro-burins. On the table R = rhomboids. Of interest is the fact that 2 microliths of the rare form D2a were found together in T14 Layer 2 Spitl and both were made of white flint instead of the usual chert; this suggests that they were made at the same time, maybe even by the same person, and also that their position in the midden has not moved since deposition.

F S R Unt | 3] i) fait | tat 2 aaa P P E Te Pe re Kb _ 2 ;1[2;7

Comparisons between the microliths and other lithic material from

Culverwell

with

the material

from

other

follow in another separate section of this Report.

BEE

sites,

will

Fossil molluscs

Approximately twenty pieces of Jurassic fossil molluscs have been found, mostly bivalves. Only one of these have signs of arteficial retouch and there is, therefore, no way of knowing whether these were deliberately or accidentally brought to the site with pieces of limesone intended for other utilitarian purposes.

to the complex nature of a large midden, the exercise of comparing distribution by layers may not be significant. It can be seen from the table (above) that there is no consistent pattern about the vertical distrbution of microliths, microburins and Krukowski burins, in Trench 4 except for the fact that these forms are present in only Layers 1-11 but not Layers 12 and 13 (see 1.3. ‘Stratigraphy’ this Report). As Layers 12 and 13 are only present in the deep gully, this situation also suggests that microliths were not delierately thrown into the gully with other refuse, as pointed out above, at least not in the earliest period of the infilling of the gully.

E1,E2,G

NON-LITHIC

FINDS FROM

CULVERWELL

Shell Beads Shell

beads

have

been

found

on

the

site,

most

from

the

midden in the areas of the floor or adjacent it (Plate 53). None definitely artificially pierced have been found in the

64

deep parts of the midden west of the floor. This suggests strongly that none of the beads were knowingly discarded

responses’. However, even though the scallop pendant was found in the context of a feature of which the ritual context

but that those found,

were lost during the course of daily

is reasonably obvious, the significance of the other beads in

activities on the floor or in the general working/living areas adjacent it.

the general context of the Culverwell midden is not so clear. They do, nonetheless, demonstrate at least a good degree of technological competence and perhaps some ‘aesthetic’ sense on the part of the site inhabitants.

Patella (limpet) shells have been found in fair numbers with holes pierced through them but they do present a problem as various carnivorous gastropods, such as the dog whelk (Nucella lapillus) and the sting winkle (Ocenebra erinacea) do pierce holes in them while they are still alive. Enzymes are secreted to soften the mollusc they are preying on and this enables the predator to bore through the shell from the outside by means of its radula. These holes are characterstically round and even. The limpets can also easily get accidentally holed by post-depositional factors as they formed part of the food chain and these holes often have jagged or frayed edges. For these reasons only those Patella shells where the holes have visibly penetrated the shells from the inside of this creature (i.e. after the animal had been taken

out

of the

shell),

or

where

striations

Ochre

Ochre is a natural earth material containing ferric oxide as colourant but small percentages of other chemicals could also be present. Ochre changes colour when exposed to heat. It can occasionally occur in sedimentary deposits derived from ochreous rock outcops. Many small pieces of ochre have been found on the site — in all areas and at all depths within the miden. They vary from being mere crumbs to pieces of 5 - 10mm. A number of pieces have been found which are larger, 15 to 40mm in size and three or possibly four pieces appear to have been shaped or moulded into conical or elongated crayon-like objects. The largest pieces

of deliberate

careful boring are visible, can with any degree of certainty be regarded as being artificially pierced and therefore possibly beads. Between 9 to 12 shells of this category have been found.

came from Trenches 40, 32 and 41; the first trench is in an

area of the site which may have formed part of a working area and the other two trenches were sited in the area of the gully on the western side of the floor (see discussion of these trenches 1.5. above). The majority of the ochre pieces

The most distinctive group of definitely artificially pierced shells are species of winkles: 17 specimens of Littorina cf obtusata.,

9

of Littorina

cf mariae

and

2

range from burnt umber,

economical

and they were,

of Littorina

therefore,

either

brought to the site by accident with other molluscs or were deliberately brought in order to be made into beads. An estimated 20 or more unpierced specimens of these shells have been found during excavation of the site and may represent examples which were discarded as beads for some or other reason or maybe were just overlooked or lost during the work of making the holes. In most cases the marks left by the tools (burins?) which were used to make the holes are very clear and, in some

cases, quite extensive.

One

but

An intensive study of the ochre from the Culverwell site still requires to be undertaken and the source of ochre on the Island has not been located with any certainty. South of the Verne (the highest part of the Isle in the north) peri-glacial deposits of red ochreous clay were noted before being quarried away at the turn of the century in the vicinity of a natural feature known as Nicodemus Knob (see discussion of Late Palaeolithic finds above). It is possible that there may have been other areas of ochreous rocks or _ peri-glacial deposits elsewhere on the Island which were quarried away without being recorded, particularly in the areas in the north where peri-glacial solifluction deposits which had scoured over ochreous rocks outside the Portland area could have

or two

beads even have signs that more than one attempt had to be made to get the hole pierced through the shell. One Nucella lapillus (dogwhelk), Cerastoderma lamarcki (cockle) and one piece of an Ostrea edulis (oyster) have also been worked as beads. The piece of oyster has been artificially shaped and holed and can probably best be described as a “‘semi-heart shaped’ pendant (see Plate 53).

accumulated.

It is,

imported to the possibly chert.

The most distinctive specimen of a worked shell is the large complete Pecten sp (scallop) shell which was found with an axe and a round pebble in the ritual feature (see 1.4. and 1.5. ‘Features’ above and Plate 36). It is the only one of this species found worked on the site and even unworked only two or three small pieces of this mollusc has been found in the midden;

and terracotta red

some are a paler pink or orange colour. A minority are yellow in colour. Lithic debris have been found with traces of ochre on them in Trenches 32 and 41 but this need not necessarily have any significance but could be fortuitous.

littorea. The harvesting of these small shells for food would not have been

burnt sienna

of course, also possible that ochre was

Island

in

exchange

for

other

material,

Ochre has been used since Upper Palaeolithic times at least, as evidenced clearly by cave paintings all over Europe and into more recent times, e.g. in aboriginal paintings. No such definite evidence for the use of ochre has yet been found in England. However, there is a large volume of evidence from archaeological sites that ochre has probably been extensively used for other purposes (e.g. at Hengistbury Head, Barton 1992) It could have been used, and still is, for purposes such as ritual burying or in ritual ceremonies for body painting and was so valued that regular mining activities took place from prehistoric and proto-historic times to procure it (Flood 1983). No such evidence has been found on Portland, probably because no definite Mesolithic burials have yet been found or identified. Ochre could have been used on

it is therefore rare for Culverwell.

The significance of beads and pendants in a Mesolithic context is not fully understood but has been extensively investigated by Newell and Kielman (1990) and the general conclusion appears to be that these objects were more than just ornaments but were ‘signs, symbols to convey information, messages and intended to evoke specific

65

hides for colouring them when used for clothing, rugs or other covering material on huts or shelters. It has also been suggested that it could have been used on hides as a grease cleansing agent or preservative (Barton 1992; Moss 1983).

Miscellaneous Post-Mesolithic finds

Apart from the pottery pieces, only a few finds came from Stratified layers. A tiny piece of finely twisted gold, 7mm long and about 1.5mm thick was found in the black clay deposit of Trench 31 ( which is also present in the western end of Trench 41 right on the edge of the gully, as described in this Report). It is too small and uncharacteristic to say whether it is Bronze Age or perhaps Roman, possibly from a brooch or bracelet.

POST-MESOLITHIC FINDS Victorian and more recent finds: A large number of seventeenth century and more recent pieces of ceramics and porcelain and small pieces of metal objects and other material were found all over the site in the ploughsoil and in Layer one. These become progressively less in Layers 2 and 3 and are rarely found below Layer 4, a buried soil level (see 1.3. ‘Stratigraphy’).

Five or six very small pieces of glass could be Roman or medieval and a number of very badly corroded tiny iron pieces are reminiscent of parts of a fibula. A small flat bronze bar 3cms long by 3mm wide, with slightly wider ends, could also be part of a Roman piece of jewelry, possibly a_ brooch, buckle or pin (Fig. 39 no. 10). Two beads, the one possibly Roman and the other possibly medieval came from Layers one and 2 of Trenches 38 and 20 respectively.

Pottery

A total of 57 small

pieces of

baked clay or pottery have

been found on the site, some from

Layers 2 to 3 but some

within stratified layers below Layer 4 (Fig. 39). Some are so small and so fragmented that it is not possible to be certain that they are indeed pottery and not baked clay from hearth areas, but it can be estimated that at least half are definitely pottery. Ofthese the majority were found in trenches west,

It is feasible that the presence of all these finds below the upper layers of the site could be due to animal or agricultural activities, as already discussed.

north-west and south-west of the floor, i.e. in the area of the

gully, suggesting that this feature was fairly vulnerable to later disturbance, either by natural agencies, such as water/mud coming down the hill or possibly later grazing or burrowing animals. Possible Early Neolithic pieces are those with shell grit or straw temper and of a brown/black colour. The Later Neolithic or possible Bronze Age pieces are of a finer texture and have brown surfaces with a black core. One tiny piece of reddish-brown ware with little or no grog has a linear motif of a number of rows of comb-impressed dots and may be part of a Beaker. Some darker brown or black pieces are possibly Iron Age or later. Only one definite Roman piece has been found and this is a piece of Samian from Trench 31 but the piece does not show enough of the profile to allow classification. The collection of sherds are comparable to the collection of very tiny pieces of pottery found at Site 1 (Palmer 1969). At Culverwell,

as

at

Site

1,

their

identification

cannot

be

regarded as definite. These pieces of pottery are therefore not characteristic of any one cultural group or period and do not provide any substantial evidence for post-Mesolithic activity on the site. Later Prehistoric lithic finds

Two barbed-and-tanged arrowheads have been found on the site (Fig. 39). One of flint, from Trench 4, came from an area where there was, visibly, a

later intrusion into the midden

layers (see above). The specimen from this trench is of a Late Neolithic or Beaker type. A similiar one made of Portland chert came from Sondage 10 in the southern part of the site in Area D. Neither were associated with other characteristic finds of that period. Sherds from a pot of this period were found at Mesolithic Site 1 (Palmer 1969).

66

1.8. THE USE OF

NON-KNAPPED STONE AT CULVERWELL

Myfanwy Stewart grey with its characteristic oolites almost black. Much of it was very fragmented and friable. Various types of limestones, much of it changed by heat from characteristic cream to beige, pinks and grey, together with sandstones and quartzites in shades of orange, brown, purple and pink, accounted for the bulk of the remaining stone. Sampling was essential because of the huge quantity of material. Aims: 1 To demonstrate that the preliminary observations were correct in that burnt Purbeck chert was the most common non-knapped stone in the midden. 2 To investigate a possible correlation between individual species of shellfish and the burnt stone.

Introduction

The midden was established at Portland Bill on the slopes of silty clay which contain numerous cryoturbated angular pieces of tabular limestone known locally as ’slat’ . This distinctive rock enables the stone, brought on to the site by man, to be more easily recognised. Colluviation may well have taken place bearing in mind the local topography and the age of the site. Furthermore, ploughing has taken place from at least medieval times and thus it cannot be assumed that all material is in situ, particularly in the upper levels. Considerable quantities of non-knapped stone have been excavated from the mesolithic midden in close association with knapped stone, charcoal and _ sea shells. It may be divided into two main categories: A The Fragmented Burnt Stone B The Pebbles

In order to prevent unnecessary disruption to both excavation and the midden itself and also to have comparative material, the unknapped stone content of the samples, that had been taken and processed during the bioarchaeological research program, were analysed and used for the purpose of this study.

A The Fragmented Burnt Stone Literally tons of burnt stone were present in the midden . Visually, the most common type was Purbeck chert, normally a light silvery grey but now transformed by heat to a dark Column

1 Trench

4

1000 900

weight in grams

800

600 500 400 300

otal adel

sample

(layer)

It can be seen that Purbeck chert is by far the most common stone. Its colour ranged from dark grey to almost black and some was so friable it could easily be broken by the fingers, indicating that it had been subjected to a high degree of heat. Much of the limestone was soft and a pinkish-grey colour, indicating

it

had

been

burnt,

too.

Similar

colours



The amount of chert in the samples fluctuates within each level, although there appears to be a peak in level 7, and the first part of layer 8, whereas the quantity of limestone and sandstones remain fairly constant. This may be a reflection of a range of activities being undertaken, where the amount of chert needed for each task varied thus producing differing amounts of waste stone. On the other hand it could simply be the result of one task being carried out on a number of occasions but to a greater or lesser extent.

and

constituencies of the burnt stone were found in all subsequent samples. The small amount of sandstone found in the samples was all fragmented.

67

The harder cherty limestone is that it has been burnt and, being have been used in a similar way the dip in the Purbeck chert line

not be so pronounced if the count for the cherty limestone was added to that of the chert. The amount of sandstone was very small and no individual piece exceeded 9 grams in weight.

invariably crazed indicating a relatively hard stone, may to the Purbeck chert. If so, in samples 2, 8 and 9 would

Column

2 Trench

4

1000 900 800 700

weight in gms

600 500 400 300

all

7(8)

8(8)

9(8)

ol pen

whe

sees

10(8)

11(9)

Boos

a

12(12)

13(12)

sam ple(layer)

The sample came from an area in Trench 4 where, during the

It can be seen

1990 and 1991 excavations, it was noted that there was much

samples is smaller compared with those taken in column 1 although there is a similar rise in the values for Purbeck chert at the end of layer 7 and the start of 8. The higher proportion of limestone is in agreement with the recorded matrix description during the excavation. The steep rise in limestone in samples 12 and 13 is only to be expected because both come from layer 12 and the slat strata of the ancient land surface may well be approaching.

more burnt limestone to be seen in the southern part of the trench compared with other parts of the midden. The area was close to where column 1 was taken and _ the archaeological layers were similar. They range from layer 6 at the top to layer 12, probably part of the earliest deposition of the midden in this area of the site.

Column

3

Trench

that, in general,

the

stone

41

3000

2500

weight in grams

2000

1500

1000

500

ae



oo:

6(8)

5(7) sam

ple

(layer)

68

10(12)

content

to these

Column 3 was taken from Trench 41, situated approximately

proximity of the ancient limestone ‘natural’.

2

vertical

9, from layer 9, is above the ‘slat’ level, and secondly, almost

variation of the mollusc taxa in the midden. Although Purbeck chert continues to be the most common stone, there is higher percentage of limestone, similar to that seen in column 2 . The abrupt rise in limestone in samples 9 and 10, taken from the deeper layers, is only partly due to the

50% of the limestone is burnt, very fragmented and appears

metres

from

column

2,

column

in

order

4 trench

to

41

assess

the

horizontal

to have

been

utilised.

Sample

10,

Firstly, sample

albeit in layer

12,

also

shows a similar high proportion of burnt, fragmented limestone. The peaks for Purbeck chert appear lower in the midden, namely at layers 8 and 9.

comparison

Purbeck

chert,

limestone

and

samdstone

1200 7

1000 +

weight grams

800 +

600

:

400 4

200 x

11(8)

12(8)

13(8)

14(8) sample

Purbeck

chert

within

a

layer,

whereas

16(8)

(layer)

Column 4 was composed of two non-continuous sets of three samples, again taken from Trench 41, but all from layer 8 in order to assess the horizontal variation within one layer . The graph demonstrates the dumping of differing amounts of the dominating

15(8)

Possible correlation between the burnt stone and the shell content of the midden Ethnological studies have shown that hot stone technology has been employed extensively in a wide range of activities and that its use continues up to the present day. One of the most important of these activities is the heating of liquids and the method was used by groups which did not have or, having

the

limestone and sandstone remain fairly constant.

Conclusion The preliminary observation of the midden content were confirmed to a large extent in that the most commonly used stone was Purbeck chert found in a burnt fragmented state. The limestone was more variable both in its state of burning and quantity. In sample 2 and particularly the lower part of sample 3, the evidence indicates that some limestone was being heated and the quantity, fragmentation and depth suggests that this was not by accident. These samples were taken in comparatively close proximity and this may well indicate that hot limestone was being used for heating water for a specific purpose. Heated limestone slakes when immersed in water, disintegrates and a residual sludge is formed. In view of the fact that large amounts of ochre were

them, did not wish to use, ceramic or metal pots over a fire. The condition of much of the midden stone, both very

found on this part of the site, it is possible that ritual activities

eaten raw. Periwinkles ( here the Monodonta and the Littorina ) have to be dead before they can be removed from their shells and the unbroken condition of these shells show that smashing, the other option was not one that was used. Steaming is a possibility but again hot stones would have been needed in the process.

fragmented and highly burnt, suggests that liquids were heated in this way at Culverwell. Shellfish were obviously an important food resource during the occupation of the midden. Although some of the shell excavated at the bottom of the gulley in Trench 4, for example, is burnt and highly fragmented, much of it in the upper midden is not and accords with most of the midden in other areas. The lime content in limpet shells, particularly, would cause the shells to be very adversely affected by direct heat but whole shells are very common in the midden. Thus it does not seem probable that roasting had been the main method

were undertaken here and that the heating of water for bathing was carried out using the conveniently available limestone. This would have been needed in some quantity and the slaking would be less of a problem than it would be with food preparation, especially if steam production was the objective rather than hot water.

69

used

to cook

the

shellfish,

although

some

can

be

In view of the high proportion of Purbeck chert in the midden, it seemed possible that there was a relationship between the quantity of shells in the midden and the chert especially where Monodonta and Littorina were concerned. The correlation between the Purbeck Chert and the shell fish species will be now be considered The midden is the result of random dumping over time and progressions are not easy to detect. However, the bioarchaeological research study at Culverwell found a Column

1 comparison

Purbeck

Chert

: MNI

relationship in column 1 between the counts of M. lineata and L. littorea which is discussed in the appropriate chapter. Briefly, in the lower part of the midden, M. lineata was dominant over L. littorea but this trend was reversed at the top. As this progression was detected, a trend might be expected in the stone content of the samples but in the sample 1 graph the counts for the two species are considered together because it is considered likely that they were prepared in the same way, namely by boiling.

Littorina littorea

+ Monodonta

6(8)

7(8)

lineata

1000 900

weight P.chert: MNI shell

800

600 500 400 300 200

3(6)

2(6)

4(7)

5(8) sample

comparison

Purbeck

chert:

MNI

shell are counterbalanced with large weights of stone in the upper samples. This might indicate that a wider range of food resources were being processed because, as Thomas and Mannino have demonstrated in the bioarchaeological section of this report, the numbers of shellfish declined in upper layers of the midden, perhaps as a result of over exploitation by man. Corroborative evidence may be found in the considerable number of pounding tools found in the upper levels of the midden which might be associated with the preparation of vegetable foods. Patella,

Littorina

littorea

+

Monodonta

lineata

weight P chert: MNI shell

2

9(9/12)

(layer)

In the column 1 chart there appears to be a relationship between weight of the chert and the M.N.I of the shellfish up to sample 4 (layer 7) with similar rises and falls but the trend is reversed for the remaining particularly in the last 3 samples presumably deposited earlier in the life of the midden. In the first set of samples the weight of chert seems excessive for so small amount of shell especially in view of the much larger weights of shellfish that were apparently processed with smaller amounts of stone in the deeper samples. After sample 5 increasing amounts of fish appear to have been prepared with far less stone whereas small amounts of Column

8(9)

—@—P. chert 8 P a te lla

sample

(layer)

70

13(12) .8

12(12)

11(9)

10(8)

9(8)

8(8)

7(8)

3(6)

2(6)

1(6)

sx--L.Littorina

+M.lineata

Samples 1 to 5 in column 2 show an apparent correlation between the MNI of shellfish and the weight of Purbeck chert. Yet the samples taken further down the column, deeper in the midden, show steeply rising amounts of Purbeck chert which are out of proportion to the shellfish numbers, notably the Patella. However the exceptionally high numbers of M. lineata + L. littorea out-number the weight of the chert in samples 8 and 9(layer 11) and sample 10(layer 12). If this Column

3

comparison

Purbeck

chert:

MNI

weight of chert was sufficient to process that amount of shellfish, then the chert found in the upper samples 1 to 5 is unlikely to have been used solely to process shellfish. Although the amounts of burnt stone are less, a similar picture to that indicated by the column 1 chart is seen with the suggestion that foods other than shellfish are being processed later on in the life of the midden.

Patella,

Littorina

littorea

+ Monodonta

lineata

2000 1800

weight grams P chert:

MNI shell

1600 1400 1200

—@®—P. vedin-

1000

sx

1(L6)

2(L6)

3(L6)

4(L6)

5(L7) sam

7(L8)

6(L8) ple

4 trench

41

horizontal

MONI

Patella

MNIL.littorea+

M

lineata

10(L12)

9(L9)

(layer)

In the column 3 graph there appears to be a greater correlation between the MNI of shellfish and the weight of Purbeck stone although the amount of the latter seems excessive as compared with the numbers of shellfish, column

8(L8)

chert

comparison

particularly in samples

6 and 7 (layer

Purbeck

the rise

chert

present,

8). With

in burnt

so much

limestone

point, referred to above, does not appear to be associated with the preparation of shellfish.

Purbeck

chert:

MNI

Patella

and

L.littorea+

M. lineata

1200

1000

weight grams: MNI shell

at this

800

—¢—

600

~~ s- -M

sample

These two sets of samples, all taken from layer 8, are similar

to those both from the column sampling. However in sample 16 the Patella numbers show a tendency to rise as the Purbeck chert weights fall.

71

Purbeck MNI NI

chert

Patella L.littorea +M.

lineata

the

B THE PEBBLES

“limpet

hammers”

from

the

Orkney

shell

middens

(Mellars 1987,122).

The excavations at Culverwell have resulted in a very large collection of non-knapped pebbles, both whole and fragmented. They consist of stone types absent from the silty clay and fragmented limestone slat that characterises the ‘natural’ on which the midden developed. Although all could have been

obtained

from

local

sources,

the expenditure

Nevertheless, some groups do emerge which may be regarded as ‘types’, albeit in small numbers, whilst other pieces appear to be unique. All the material was examined with a hand lens and, as a result, will be considered under the following categories:

of

energy is clear. The fact that so many pebbles, both used and unused, were left behind when the site was vacated may indicate that the area was associated with unique tasks that were resumed when the inhabitants returned. When considering any aspect of the data it must be remembered

that,

whilst

some

areas

of the

site have

A Utilised material 1 smoothed stone

2 abraded / pecked stone B Unaltered material

been

excavated to a depth of 14 levels to the base of the midden, others have been deliberately left by the Director in order to preserve the ‘ floor ‘ intact. Overall site distributions of all material must be viewed with a degree of caution, although concentrations of finds in some trenches are apparent.

A Utilised material 1 smoothed stone - total 209

Examination of the utilised material demonstrated that 75% of it had been associated with tasks that left areas of smoothing on the individual stones. Almost all were oolitic limestone beach pebbles and no doubt the mesolithic people came purposefully to the area and its shoreline in order to carry out specific tasks. The pebbles vary widely in size from

Processing A total of 678 whole pebbles and 630 fragments have been examined by hand to date. As with the burnt stone, it was not

practical to retain and therefore examine all the enormous quantity of fragments which were excavated. Dimensions, weight, petrology, evidence for burning and signs of usage were recorded. Weights of individual pebbles vary from almost two kilos to less than one gram. 40% of the material appears to have been utilised and, of this, approximately 75% shows evidence of smoothing or rubbing with the remainder being pecked or abraded.

27 cms x 10 cms and weighing over 2 kilos to small, disc-like

ovate shaped specimens of only a few grams. Often the term rubber is used to describe pebbles with this sign of usage although the range in size and weight of the Culverwell assemblage must inevitably point to a commensurate functional range. Areas of smoothing are recognisable by the fact that the individual oolites are worn down and that the stone has paler more matt appearance when turned to the light and viewed close-up. With the aim of putting the material in some sort of order, it was divided into three simple classes on the basis of dimension alone. No functional criteria are implied although there must inevitably be a link between extremes of weight and size. Large - 2.7kilos to 300gms 19 large stones were recorded representing 9.1 % of the total 209 smoothed stones. Flat forms 60% of the large stones appear to have been used only for smoothing and are often comparatively thin and platter- like for their size. Figure S 1(1) illustrates a typical example. The remaining 40% differ either by having completely different forms and/or by showing wear patterns in addition to smoothing. Long composites Fig S 1(2) is an example of a heavy pebble (596gms) with smoothing and abrasion and may be considered as a composite tool. A second example, weighing 443 gms was found in the same trench in an upper level. Another find associated with S1.1 weighed nearly 2 kilos and has a club-like appearance with two faces more flat than rounded ( Plate 1). One flattened side has been smoothed whilst heavy battering has removed an area of stone at the thick end. A third composite tool, weighing over 2 kilos, was found in T41 and is similar in shape to S1.2 although the abrasion at the end is only light. Sub-triangular composite A few pieces excavated at Culverwell are composed of a very distinctive two-toned grey-cream limestone which has a clear

Typology

The question of typology is a constant topic of archaeological discussion. It is recognised that diagnostic criteria may be difficult to select and that after analysis there is a distinct possibility that the term allotted to the type does not reflect its function in ancient times. Inevitably, subjective elements are present in the analysis but, provided that the selective criteria are clearly stated and illustrated, typology remains the best and _ essential method by which archaeological material is put into manageable order. The classification of non-knapped material is not so standardised as knapped although Jacobi, (1980,11) lists 29 items in his type list for such material. Some of the Culverwell pebbles are a combination of types. For example, one pebble has a countersunk rounded cavity on one face, damage more reminiscent of an anvil on the other and a clearly smoothed area on its end. There is the distinct possibility that some pebbles have been used in several activities Beaune (1989,61), in her study of a pebble by members of an African Touareg group, found that over a few days the same quartz pebble was used for a wide variety of tasks by several members of the community. A second problem is that some pieces are the only one of their kind to be excavated at Culverwell and are reported here as individual items rather than ‘types’. Terms such as rubbers, smoothers and crushers are frequently used in archaeological reports but possible associated resources are not usually discussed although an exception is

72

demarcation between the two colours. One example is in this class of large stones, weighing 394 grams, sub-triangular in section, and with smoothing on the largest face Fig S2 (1). The tip is abraded, particularly on the largest of the three faces, showing that the stone was used in more than one activity. Large crusher This large shelly limestone cobble, weighing 833 grams, with a distinct area of smoothing on one rounded face, was found in the vicinity of the hearth situated on the uppermost slope of the site in the north west. The cavity on the top of the stone would be very convenient for the users thumb and it is presumed that this heavy tool was used for crushing and slow pulverising (Plate 2) . ‘Hone’ Excavated in two segments, with a total weight of 310 grams and composed of red sandstone, this piece fits comfortably in the hand with the indentation providing a convenient finger rest. In this position it seems more likely that another object was drawn across its surface thus accounting for the smoothing on both faces Fig S2(2). Distribution Any apparent concentrations must be approached with a degree of caution because: 1 Large sections of the midden in the southern part of the site remain unexcavated in order to preserve the mesolithic ‘floor’ directly above. Thus tool counts in the northern section, which appear as concentrations, could have been equalled or exceeded if this southern section had been dug. 2 A large ditch or gulley runs in a NE to SW direction and particularly affects Trenches 4,32 and 41. These trenches have been dug to a deeper level than most of the others and the midden is much thicker. This inevitably results in a greater number of finds. However certain types are much more common than others e.g. the smoothed pebbles, whilst others are fewer than elsewhere such as the large pounders and the countersunk pebbles. As it does not seem to be simply a matter of dumping rubbish into the gulley, indications of concentrations of certain types are made. Large smoothed stones were recorded in 10 trenches all of which are in the northern half of the site. 36.8% came from Trench 21 which abuts Trench 1 with its hearth. If the score for trench 1 is added, this percentage rises to 43% and indicates a possible work area in the vicinity of the fire. The large crusher was associated with the most northern hearth . Two countersunk pebbles are included in the total for this class because they have limited areas of smoothing but are dealt with below because they also have substantial abraded areas and are considered as individual types in the section below.

section, with evidence of areas of smoothing

on the flat side

and a distinctive nosed beak-like end. Fig S2 (3) illustrates a complete specimen, found associated with a possible cooking pit. The remainder are represented by the beaked end only. One limestone piece similar to this group was associated with a hearth Fig S2 (4). It is rib-like in appearance, lacks the nosed end and is oblong in section but may have been used for similar work, perhaps skin preparation.

Grooved smoother One find in this class is unique to the site Fig S3(1). Composed of fine-grained limestone, it has a pronounced smooth-sided groove in the flattened top and was presumably used in the final stages of finishing some small tool such as a wooden or bone needle or perhaps the base of an arrow flight. Two-toned smoother Only one stone of this type was found at Culverwell. In the two-toned limestone described above, it has a fish-like form

and light smoothing down the “spine” Fig S3(2). Sinuous Smoother Another single example is this sinuous pebble of brown sandstone Fig $3(3). One face is lime-encrusted but the other has clear smoothing. Weighing 196 grams, this substantial tool is unusual in being of sandstone rather than the usual limestone.

Distribution Medium sized smoothed pebbles were recorded in 16 trenches but again mainly in the north of the site. The area around the hearth in Trench 1 had 15% but the most prolific area was around trenches 4, 32 and 41 which accounted for

52.3% of the finds. Small - the remainder. The remaining 146 stones (69.9 %) pebbles often showing smoothing probably had a variety of functions may well have been used for several types have emerged, based on their

are mostly ovate, small on two faces. They and individual pebbles different tasks. Three distinctive shapes alone

with no single function implied, and have been called discs

and sub rectangular tools. Two other pieces are described because they are single finds from the excavation. Discs Some 50 of these have been excavated and they represent 34.2 % of the material in the small range. The amount of visible smoothing and its location shows variation but all are flat sided oolitic limestone pebbles. As the name suggests they tend to be naturally round or ovate with a circumference of between 3cms and 7.5 cms but with 80% falling in the 3cm to 5.5cms range. In thickness the range is much smaller being between 0.5 and 1.5cms (Plate 3). The discs were obviously chosen for their size but also for their petrology because the fine oolitic limestone rubs down to a surface akin to that of pumice or coarse sandpaper and

Medium - 299gms to 100gms and measuring 11cm-5 cm. 44 pebbles are in this group representing 21% of the total smoothed stones. 35 (i.e. almost 80%) are smaller versions of those large flat specimens described above with only areas of smoothing taken as evidence of their usage by man. A few of the smaller specimens are chunkier versions of the discs described below. One other type has emerged in this group and three individual finds occur singly. Nosed pieces With regard to the remaining 9 pebbles, 4 are a distinctive type being pieces of tabular limestone, sub-triangular in

would have been suitable for precise, if not delicate work.

Trench

41

had

a concentration

producing

21

discs

and

Trench 4, immediately to the north, with 9 thus accounting

for 60% of the total. In view of the high amount of ochre, including one very large piece from T41, also recovered from this area, it is possible that these small discs were associated in some way with ochre. However the apparent concentration must be viewed with a degree of caution because, as noted above, this trench has been excavated to a

deeper level than many of the other trenches.

73

Disc

distribution Midisc

25

Ss

B30

B25

BQ

B6

B11

ke

B4+5

9

B3+4

ke

B2+5

Fe

@

T41

Fe

al

T38

Fe

a

T36

A

T34

T19

T17

T15

T13

T11

T9

T7

T5

T3

T1

$11

$8

$2

number

20

trench

Sub-rectangular tools Both of the items in this class are unusual in that they have a distinctive sub-rectangular shape and neither are of the usual limestone

a bone

pebble. A deposit of lime is present on the rough side but the smoothing on the other face is plainly visible. The tool feels like fine grade sandpaper and may have been used for smoothing wood Fig S3 (4). A similar piece, thin and flat but of silky, grey shale, has one face smoother than the other but with the main area of smoothing round the edge of the pebble. Although very few bones

were

found

in the midden,

pelts

could

with clear traces of smoothing.

Like the

discs, it was probably used in a smoothing activity where a certain amount of precision was needed. However the stone must have been held in a more robust fashion that the more delicate discs. As it is a singular form, this might suggest that the tool was used for less common tasks although others of its kind might have been taken away from the site when the inhabitants moved on. Fine polisher Fig S3(7) This small tool is the most delicate of all the non-knapped tools. Made of fine-grained limestone and almost diamond shaped, it has smoothing on one face and _ oblique smoothing/polish on both sides of the tip which has been worn down to be paper thin. Obviously such a tool can only have been used in commensurately delicate tasks and perhaps it was associated with needle production in view of the grooved smoother referred to above.

The first, found in Trench 32, layer 9, is a thin, flat, sandstone

animal

in shape,

have

been brought in and this tool could have been useful for the preparation of animal skins e.g. fat removal. Facetted Spheroids Fig S3(5) 19 spheroids have been found mainly in the upper levels of the midden associated with the upper levels of Culverwell. None occur lower than level 9 and 58% come from levels 1 and 2. They are all composed of a type of cherty limestone which is harder than the oolitic variety. All are spherical in form with mainly clear facetted edges which have been smoothed. This smoothing is probably part of the shaping process, as is a certain amount of pecking, although none of the usual signs of knapping are present, such as positive/negative bulbs of percussion. Two are more rounded in form but still have facets. 45% weigh between 90 and 154 grams, 52% between 10 and 48 grams and thus two groups emerge. The use of these distinctive objects is enigmatic but magic/ritual or simply games are possibilities. Precision smoother Fig S3(6) This is an example of a form sufficiently distinctive for it to be considered as a type although only one specimen was excavated. It is composed of cherty limestone, reminiscent of

Distribution The main distribution of these small smoothed pebbles

is

again the area around Trenches 4, 32 and 41 where 67.8% of

the finds were recorded which includes a high percentage of the “discs”. The hearth area in Trench 1 and Trench 21 to the north had 10.3%. 42% of the spheroids also come from Trench 4 although the remainder are distributed over a wider area of the site. 2 Abraded stone - total 71

71 abraded stones were examined and again have been divided into 3 groups according to weight. As stated above, no functional distinction is implied although the size of a tool must have some bearing on its usage in view of the range of weights. Although the large items are still in the minority, the abraded stones are more evenly represented in each set compared with the smoothed specimens.

74

With regard to typology, many appear to be hammerstones resembling those found on other mesolithic sites and usually associated with flint knapping. However many could have served more than one purpose. Some of the larger specimens are more akin to pounders, some are composite tools as referred to above and a few items were found only as individual pieces. Large - 943 - 300gms 21.1% of the abraded pebbles are in this class Hammerstones 6 finds in this group may be regarded as hammerstones associated with knapping. The weights range from 368gms to 458gms with the two lightest being red Devonian

The part countersunk pebbles do not come in to the large class of pebbles but are considered here for the sake of clarity. They are useful because they show the profiles of the cavities. Fig S5(4) is a good example being 7 cms across with the depth of stone between the two cavities as barely a centimetre. Both cavities show clear evidence of pecking although one may have smoothing, too. This is obviously a well used stone that probably finally broke in spite of it being a hard cherty limestone. Fig S5(3 ) demonstrates the variety in size of these countersunk pebbles. The find is only 5 centimetres across and, as the worked areas are not diametrically opposed, this results in a sinuous profile. Again the pecking is clear. From the same trench and layer comes a 7 cm piece of pebble singular on two counts in that it is sandstone and has only one worked cavity. Furthermore the cavity appears to have been used for some smoothing activity which is logical given the granular nature of sandstone Very faint traces of ochre may be present. A pronounced cracking shows that this type of stone is less robust than the hard cherty limestone . Anvil stone Fig S6(1)

sandstone, the heaviest Purbeck chert and the remaining three

brown/yellow quartzite. The abrasion appears to occur in two forms. In the first type the damage is heavier at either one or both ends and the Purbeck chert example shows this Fig S4(1). Although 458 gms might appear to be too high for this stone to be regarded as a knapping hammer stone, its general ovate shape and areas of damage are typical of hammerstones in general and the size of some of the Portland ‘ picks ‘ at Culverwell would have necessitated the use of a very substantial knapping tool. However it is quite possible for a pebble to be used for general crushing work in addition to knapping and this might be particularly applicable to large abraded tools. In the second type the abrasion is around the edge of the pebble. This can be seen clearly in Fig S4(2), one of the red Devonian sandstone specimens. Although a substantial tool, it was obviously used for more precise work. Pounder This heavy tool lacks the ovate shape and systematic abrasion of the hammerstones Fig S4(3 ). Weighing 943 gms, it is a flat-sided tan quartzite cobble with heavy wear and damage on one end and lighter abrasion on the top. It was found on the most south eastern part of the site. Countersunk pebbles 2 whole and 3 part countersunk pebbles were excavated. 4 were

made

from

limestone,

common

on

the

local

The

beach

a circular

area

area with a diameter of 2 cms. On the

of abrasion

diameter, and, like the previous of abrasion of approximately 6 side. However this area is much previous find and a section of lightly damaged. Furthermore,

one

side,

some

3 cms

areas

on

each

face

of

this

round,

Medium - 299 - 100 gms Hammerstones 43.7 % of the utilised pebbles are in this medium class and most can be regarded as tools that were used in more precise work than the large examples. More than 70% of them are rounded with abrasion at one or both ends. Fig S6(3) shows a typical example. The remainder are generally flatter with the abrasion around the edge e.g. Fig S6(4 ). As might be expected, all the hammerstones from the excavation are of hard stone with flint and chert predominating. However Devonian sandstone is well represented and the two hammerstones illustrated here are of the red type and the less common banded green. The following were each the sole example of the type found during the excavation. Precision hammer This find is again in two-toned limestone Fig S5(5). Weighing 228 gms the area of abrasion on the head is very clear. The slight dip on the reverse side of the pebble makes for a convenient rest for the forefinger and would have given the user more purchase and added precision.

other side the abrasion forms a more rectangular shape approximately 3cms x 2 cms. At the more pointed end there is aclear area of smoothing approximately 5 cms x 3. 5 cms, shown by the dotted area in the drawing, which gives the appearance of the pebble being worn down. This combination of wear patterns suggests a composite tool Fig S5(1). The second whole countersunk pebble, weighing 539 grams, has

abraded

limestone pebble are more shallow than those described above and it appears to be an example of a light anvil stone. However, its resemblance to the countersunk pebbles is clear and this stone may simply be of the same type but used to a lesser degree ‘‘Waisted stone” Fig S6(2) This limestone object, the only example found at Culverwell, was found in an area associated with several unusual finds made from two-toned grey-white stone. It weighs 635 gms and appears to have been deliberately pecked into this axelike shape. Extra pecking accentuates the two cavities and makes the piece easy to grasp but equally would make it easier to encircle it with a fine rope. The find was provisionally described as a weight but could well be a ritual object of some kind The piece stands firmly on its side as shown but with equal stability will stand upright with the flat side as a base (Plate 4). In this position it is reminiscent of a torso but any resemblance to a particular form is obviously a highly subjective matter.

today, whilst the fifth was sandstone. The two whole specimens were of similar size viz. 10cms x 8cms and 12cms x 6.5 and came from the same trench and level. They are of particular interest because they show clear evidence of different use pattern on the individual pieces. Weighing 323 gms, one side of the pear-shaped pebble has a distinct, circular, pitted

circular

in

stone, has an oblong pattern cms x 3 cms on the reverse more amorphous than on the the rectangle is much more the pebble is abraded at one

end and so, whilst both are countersunk and have been used

for more than one task, these two pebbles although similar have been used differently Fig S5(2).

75

the outer edge of the cavity. From the point of view of petrology, weight and size, this piece would be very suitable for removing limpets from rocks. The reason that only one of these was found on the excavation may be that it was customary to leave limpet hammers down on the beach. Small - Less than 100 gms The remaining 43.7 % of the abraded stones are in this small category and most appear to be flint knapping hammer stones. 75% were rounded pebbles with abrasion at one end and 25% were flatter with damage around the perimeter. Fig S7(3) shows a well-used example of the latter in veined red Devonian sandstone. Two finds differ and are the only examples of their type to be excavated at Culverwell. ?Pestle Fig S7(4) This finger-shaped limestone pebble, weighing 36 gms has clear bruising on the tip, mainly confined to one face. Although not found in a trench where countersunk pebbles

Two-toned “limpet scoop” Fig S7(1) Weighing 152 gms, this piece is also in two-toned limestone. It seems likely that the stone was deliberately shaped, perhaps by pecking in view of the rough surface of the indentation, so that the rounded base would provide a ‘handle’. The abrasion is light and occurs only on the tip and four centimetres from the tip down the side. Again there is a slight dip in the stone on the reverse side of the pebble which conveniently supports the forefinger. This tool could not have been used in heavy work, bearing in mind the petrology and light wear. It has a different form and lacks the heavy abrasion of the limpet hammers and scoops from Oronsay described by Mellars (1987 121,122). Furthermore it does not seem logical to bring a limpet hammer so far in from the beach. However, the unusual stone might have given it a special status or it might simply be a favourite tool. Perhaps it was used for scraping limpets out from their shells back at base

when

the limpets

were

cooked.

On

the other hand,

a

small piece of wood or twig would have served equally well and, considering the huge numbers of limpets that must have been consumed, one would have expected more than one scoop to be found during the excavations. Perhaps the twotoned stone is the clue showing that this tool had special significance to the mesolithic people, particularly as a high proportion of “one off’ tools share this petrology. ‘‘Limpet hammer” Fig S7(2) A more convincing argument for a limpet hammer can be made out for this pebble tool weighing 133 gms and composed of hard Purbeck chert. One face has a dipped, circular, pitted area, with the cortex removed,

showing

were

recorded,

it is possible

that

this

tool

was

used

conjunction with them. Light retouching tool Fig S7(5) Weighing 99 gms, this smooth, grey chert pebble has an unusual shape and was found in association with the hearth in Baulk 25. The abrasion on both ends is clear but this tool must have been used very lightly used, perhaps for delicate tasks. The general appearance is reminiscent of a figurine and it is tempting to attribute a possible symbolic significance to this item..

the

dark grey oolites that are typical of this type of stone. The base of the pebble is abraded and heavy wear can be seen on Summary

smoothed

and

abraded

pebbles Olabraded

Msmth. 80

70

60

50

40

30

tr ench

76

$8

B30

t+

B25

o

-

B11

op)

©

Bg

(| Topsoil

7A)

Yellow

oam

Wi

Brown clay

fow finds”

[=] She

net en

Fig. 9. Section through Hearth 4 and the adjacent paved area. P indicates picks.

184

SHELLS

F L O O R

PP STONES OF

aw

i

FEATURE

—~~pe

-

|

FLOOR

ON SHELL

“A

MIDDEN

“\

Ci}

_—

Z

/ d

=< COMPACT BLACK #

=

a

“N\

A




TL- Hole for thermoluminessence

31M

SC

LFT

©

TL

Fig. 11. Plan of Hearth One

aan Se

| | brown ‘| ay - stoneless

rt

|

%

| wi

*

aa ts

*

° a

»

boa.

es

midden

|

* *

Pa

*,

|

.

(Trench 1 and 3) and adjacent area.

| 4 fe

‘ «

foe Es

ew

Fa



»

*

.

e

2

wf

31m lft

Fig. 12. Section through the edge of the floor and Hearth

One.

186

sample H - hammerstone

oO