Molluscs in Archaeology: Methods, Approaches and Applications 178570608X, 9781785706080

The subject of ‘Molluscs in Archaeology’ has not been dealt with collectively for several decades as most previous volum

1,276 85 45MB

English Pages 448 [449] Year 2017

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Molluscs in Archaeology: Methods, Approaches and Applications
 178570608X, 9781785706080

Table of contents :
Cover
Title
Copyright
Dedication
Contents
Preface
Contributors
Acknowledgements
Molluscs in archaeology: an introduction Michael J. Allen and Bas Payne
PART 1: PALAEO-ENVIRONMENTS; ENVIRONMENT AND LAND-USE
TERRESTRIAL HABITATS, CONTEXTS AND LANDSCAPES
1. Land snails in archaeology - Michael J. Allen
2. The geoarchaeology of context: sampling for land snails (on archaeological sites and colluvium) - Michael J. Allen
3. Numerical approaches to land snail palaeoecology - Matt Law
4. Molluscs and the palaeo-environment of coastal blown sand and dunes - Thomas Walker
5. Molluscs from dune-machair systems in the Western Isles: archaeological site formation processes and environmental change - Matt Law and Nigel Thew
6. Caves and molluscs - Chris O. Hunt and Evan A. Hill
WETLANDS AND FRESH- AND BRACKISH-WATER
7. Molluscs from the floodplain alluvial sediments in the Thames Valley - Mark Robinson
8. Wetlands: freshwater and slum communities - Terry O’Connor
PART 2: PALAEO-ENVIRONMENTAL RECONSTRUCTION: EUROPE, THE MEDITERRANEAN AND NEAR EAST
9. The southern English chalklands: molluscan evidence for the nature of the post-glacial woodland cover - Michael J. Allen
10. (Some thoughts on) using molluscs for landscape reconstruction and ecology in Malta - Michael J. Allen and Bri Eastabook
11. Molluscan remains from early to middle Holocene sites in the Iron Gates reach of the Danube, southeast Europe - Catriona Pickard, Adina Boroneanț and Clive Bonsall
12. Land mollusc middens - Victoria K. Taylor and Martin Bell
PART 3: MARINE AND FOOD AND DIET
13. Marine molluscs from archaeological contexts: how they can inform interpretations of former economies and environments - Liz Somerville, Janice Light and Michael J. Allen
14. Oysters in archaeology - Jessica Winder
15. Shell middens - Karen Hardy
16. The collection, processing and curation of archaeological marine shells - Greg Campbell
PART 4: ARTEFACTS
17. Shell ornaments, icons and other artefacts from the eastern Mediterranean and Levant - Janet Ridout-Sharpe
18. Molluscan shells as raw materials for artefact production - Katherine Szabó
19. How strong is the evidence for purple dye extraction from the muricid gastropod Nucella lapillus (L. 1758), from archaeological sites in Britain and Ireland? - Janice Light and Thomas Walker
20. Marine shell artefacts: cautionary tales of natural wear and tear as compared to resourceful anthropogenic modification processes - Janice Light
PART 5: SCIENCE AND SHELLS
21. Bivalves and radiocarbon - Ricardo Fernandes and Alexander Dreves
22. Radiocarbon dating of marine and terrestrial shell - Katerina Douka
23. Stable isotope ecology of terrestrial gastropod shells - André Carlo Colonese
Index

Citation preview

Molluscs in Archaeology

Studying Scientific Archaeology Studying Scientific Archaeology is a new series of titles from Oxbow Books. The series will produce books on a wide variety of scientific topics in archaeology aimed at students at all levels. These will examine the methods, procedures and reasoning behind various scientific approaches to archaeological data and present case studies or extended examples to demonstrate how data is used and interpretations are arrived at. In particular we aim that they should demonstrate how scientific analyses contribute to our wider understanding of past human behaviour, technology and economy. The series title reflects an inclusivity in the volumes in the sense of encouraging readers in practical research rather than just presenting collected papers as statements of work completed. Our aim is that these titles will come to feature as recommended reading for university courses, providing a sound basis for the appreciation and application of scientific archaeology.

Already published in this series French, C. 2015. A handbook of geoarchaeological approaches for investigating landscapes and settlement sites Hardy, K. & Kubiak-Martens, L. (eds), 2016. Wild Harvest: plants in the hominin and preagrarian human worlds Allen, M. J. 2017. Molluscs in Archaeology: methods, approaches and applications

SERIES EDITORS: Michael J. Allen and Terry O’Connor

Studying Scientific Archaeology No. 3

Molluscs in Archaeology: methods, approaches and applications

edited by Michael J. Allen

Oxford & Philadelphia In conjunction with the Conchological Society of Great Britain & Ireland and Allen Environmental Archaeology

Allen Environmental Archaeology

Published in the United Kingdom in 2017 by OXBOW BOOKS The Old Music Hall, 106-108 Cowley Road, Oxford, OX4 1JE and in the United States by OXBOW BOOKS 1950 Lawrence Road, Havertown, PA 19083 © Oxbow Books and the individual contributors 2017 Paperback Edition: ISBN 978-1-78570-608-0 Digital Edition: ISBN 978-1-78570-609-7 (epub) A CIP record for this book is available from the British Library Library of Congress Control Number: 2017940287

All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopying, recording or by any information storage and retrieval system, without permission from the publisher in writing. Printed in the UK by Hobbs the Printers Typeset in the UK by Frabjous Books For a complete list of Oxbow titles, please contact: United Kingdom Oxbow Books Telephone (01865) 241249 Email: [email protected] www.oxbowbooks.com

United States of America Oxbow Books Telephone (800) 791-9354, Fax (610) 853-9146 Email: [email protected] www.casemateacademic.com/oxbow

Oxbow Books is part of the Casemate Group Front cover: Cornu (Helix) aspersum apex; photo © M. J. Allen 2017 Back cover (l–r): Dupotetia dupotetiana on shrubby vegetation, Moulouya valley, Morocco; adult mussels attached to rocks at Fistral bay, Cornwall; Bronze Age burial from Ban Non Wat (Central Thailand).

SPONSORS

for Prof Les Allen, BSc, DSc, PhD, ARCS, DIC, CPhys, FInstP, FOSA, 1935–2016 To my father, an amateur archaeologist, who nurtured, encouraged and fostered my own interest, and then career in archaeology, ensuring that I continued to publish and attempt to maintain high academic standards. I’m only sorry that I didn’t finish this in time for you to see.

Founded in 1876, the Conchological Society of Great Britain and Ireland is one of the oldest existing societies devoted to the study of molluscs and their conservation. The Society achieves this through meetings, workshops, publications and distribution recording schemes. It publishes its academic journal, the Journal of Conchology, twice yearly and Mollusc World, a colourful magazine three times a year. These both form part of the annual subscription. Occasional identification guides as well as Special Publications in the format of the Journal are issued. The Society’s active programme of events includes lectures, identification workshops, day- and weekend conferences and field meetings, enabling members to work together on projects in a social and learning environment. Our field meetings generate records for our marine and nonmarine recording schemes that have now been operating for over 100 years. Meetings are open to members of affiliated organisations and we welcome non-members. The Society encourages discussion and provides a forum for all those working with, and interested in, molluscs, including those dealing with archaeological and palaeo-environmental assemblages, as this volume clearly demonstrates. Further information can be found on its website (www.conchsoc.org) or by contacting them at [email protected]. The Society is a registered charity (no. 208205)

Allen Environmental Archaeology Allen Environmental Archaeology has been operating for 10 years, with Mike, a leading geoarchaeologist and environmental archaeologist with over 35 years of experience, specialising in land snails, soils, sediments, hillwash and co-ordinating palaeo-environmental programmes. Mike has undertaken land snail analysis across southern England, for over 35 years including working in Stonehenge, Avebury, Dorchester Environs, Cranborne Chase, and the South Downs National park, and in the past has specialised in the examination of colluvial deposits. Allen Environmental Archaeology co-ordinates environmental archaeology sampling, processing, analysis and publications programmes for a number of projects and archaeological organisations, providing a one-stop shop for many field archaeologists. Mike has a long publication record, as well as being series editor for the Prehistoric Society Research Papers, and Oxbow’s Studying Scientific Archaeology Series and on the editorial board of Oxbow’s Insights Series.

Contents

Preface.......................................................................................................................................... x Contributors............................................................................................................................... xi Acknowledgements................................................................................................................xiv Molluscs in archaeology: an introduction............................................................................. 1 Michael J. Allen and Bas Payne

Part 1: Palaeo-environments; environment and land-use Terrestrial habitats, contexts and landscapes ...............................................................5 1. Land snails in archaeology................................................................................................6 Michael J. Allen 2. The geoarchaeology of context: sampling for land snails (on archaeological sites and colluvium)........................................................................30 Michael J. Allen 3. Numerical approaches to land snail palaeoecology....................................................48 Matt Law 4. Molluscs and the palaeo-environment of coastal blown sand and dunes..............65 Thomas Walker 5. Molluscs from dune-machair systems in the Western Isles: archaeological site formation processes and environmental change........................82 Matt Law and Nigel Thew 6. Caves and molluscs.........................................................................................................100 Chris O. Hunt and Evan A. Hill Wetlands and fresh- and brackish-water . ...................................................................111 7. Molluscs from the floodplain alluvial sediments in the Thames Valley................112 Mark Robinson 8. Wetlands: freshwater and slum communities ...........................................................127 Terry O’Connor

viii

Contents

Part 2: Palaeo-environmental reconstruction: Europe, the Mediterranean and Near East 9. The southern English chalklands: molluscan evidence for the nature of the post-glacial woodland cover..............................................................................144 Michael J. Allen 10. (Some thoughts on) using molluscs for landscape reconstruction and ecology in Malta......................................................................................................165 Michael J. Allen and Bri Eastabook 11. Molluscan remains from early to middle Holocene sites in the Iron Gates reach of the Danube, southeast Europe.......................................................................179 Catriona Pickard, Adina Boroneanț and Clive Bonsall 12. Land mollusc middens...................................................................................................195 Victoria K. Taylor and Martin Bell

Part 3: Marine and food and diet 13. Marine molluscs from archaeological contexts: how they can inform interpretations of former economies and environments..........................................214 Liz Somerville, Janice Light and Michael J. Allen 14. Oysters in archaeology...................................................................................................238 Jessica Winder 15. Shell middens...................................................................................................................259 Karen Hardy 16. The collection, processing and curation of archaeological marine shells..............273 Greg Campbell

Part 4: Artefacts 17. Shell ornaments, icons and other artefacts from the eastern Mediterranean and Levant.............................................................................................290 Janet Ridout-Sharpe 18. Molluscan shells as raw materials for artefact production......................................308 Katherine Szabó 19. How strong is the evidence for purple dye extraction from the muricid gastropod Nucella lapillus (L. 1758), from archaeological sites in Britain and Ireland?...........................................................................................326 Janice Light and Thomas Walker

Contents

ix

20. Marine shell artefacts: cautionary tales of natural wear and tear as compared to resourceful anthropogenic modification processes............................342 Janice Light

Part 5: Science and shells 21. Bivalves and radiocarbon .............................................................................................364 Ricardo Fernandes and Alexander Dreves 22. Radiocarbon dating of marine and terrestrial shell...................................................381 Katerina Douka 23. Stable isotope ecology of terrestrial gastropod shells...............................................400 André Carlo Colonese Index........................................................................................................................................414

Preface

Whilst teaching environmental archaeology and geoarchaeology to students and amateur archaeologists, and conducting workshops for professional archaeologists/ colleagues on molluscs in archaeology, it was apparent that there was no single book addressing molluscs in archaeology as a whole to which I could refer anyone. Instead the bibliography included whole books on snails or shells, and chapters in text books, as well as papers in books and academic journals. Hence the concept of this book, the premise of which was trailed at a conference of the same name held at the Natural History Museum on 26 April 2014, and jointly organised by the Conchological Society of Great Britain & Ireland and the Association for Environmental Archaeology. The subject and concept was warmly embraced by the Conchological Society during my presidency. Their help and support for this book is seen not only in their significant financial contribution towards its publication, but also in the fact that its contributors include three Vice Presidents (Allen, Light and Payne), three members of council (Eastabrook, Somerville and Walker), one former member of council (Ridout-Sharpe), two members of the Society (Law and O’Connor), and one former member (Taylor). The conference clearly demonstrated a need to introduce the diverse topics embraced under the umbrella of molluscs in archaeology under a single cover; and this is that book. I hope it will go some way to addressing that need. Mike Allen January 2017

Contributors

Michael J. Allen [Vice President Conchological Society of Great Britain & Ireland] Allen Environmental Archaeology, Redroof, Green Road, Codford, Wiltshire, BA12 0NW

Adina Boroneanț ‘Vasile Pârvan’ Institute of Archaeology, Romanian Academy, 11 Henri Coandă St., 010667 Bucharest, Romania [email protected]

and

Greg Campbell The Naive Chemist, 150 Essex Road, Southsea, Hampshire, PO4 8DJ [email protected]

Department of Archaeology, Anthropology and Forensic Science, Faculty of Science and Technology, Bournemouth University, Dorset, BH12 5BB [email protected] Martin Bell Department of Archaeology, University of Reading, Whiteknights, Box 227, Reading, RG6 6AB [email protected] Clive Bonsall School of History, Classics & Archaeology, University of Edinburgh, William Robertson Wing, Old Medical School, Teviot Place, Edinburgh, EH8 9AG [email protected]

André Carlo Colonese BioArCh, Department of Archaeology, University of York, Biology S. Block, York, YO10 5DD [email protected] Katerina Douka Research Laboratory for Archaeology and the History of Art , University of Oxford, Dyson Perrins Building, South Parks Road, Oxford, OX1 3QY and William Golding Junior Research Fellow, Brasenose College, University of Oxford, Oxford [email protected]

xii

Contributors

Alexander Dreves Leibniz-Laboratory for Radiometric Dating and Stable Isotope Research, University of Kiel, Max-Eyth-Strasse 11–13, 24105 Kiel, Germany [email protected] Bri Eastabrook 19 Townsend Street, Cheltenham, Gloucestershire, GL51 9HA [email protected]

Chris Hunt School of Natural Sciences and Psychology, Liverpool John Moores University, Life Sciences Building, Byrom Street, Liverpool, L3 3AF [email protected] Matt Law College of Liberal Arts, Bath Spa University, Newton St Loe, Bath, BA2 9BN [email protected]

and

Janice Light [Vice President Conchological Society of Great Britain & Ireland] The Old Workshop, Winterborne Kingston, Dorset, DT11 9AX [email protected]

Leibniz-Laboratory for Radiometric Dating and Stable Isotope Research, University of Kiel, Max-Eyth-Str. 11–13, 24105 Kiel, Germany [email protected]

Terry O’Connor Department of Archaeology, University of York, Heslington, York, YO10 5DD [email protected]

Ricardo Fernandes McDonald Institute for Archaeological Research, University of Cambridge, Downing Street, Cambridge, CB2 3ER

Karen Hardy ICREA (Catalan Institution for Research and Advanced Studies), Pg. Lluís Companys 23, 08010 Barcelona, Catalonia, Spain and Departament de Prehistòria, Facultat de Filosofia i Lletres, Universitat Autònoma de Barcelona, 08193 Bellaterra, Barcelona, Spain [email protected] Evan Hill School of Natural & Built Environment, Queen’s University Belfast, Belfast, Northern Ireland [email protected]

Bas Payne [Vice President Conchological Society of Great Britain & Ireland] The Mill House, Clifford Bridge, Drewsteignton, Exeter, EX6 6QE  (formerly of English Heritage) [email protected] Catriona Pickard School of History, Classics and Archaeology, University of Edinburgh, William Robertson Wing, Old Medical School, Teviot Place, Edinburgh, EH8 9AG [email protected]

Contributors Janet Ridout-Sharpe 66 Radnor Road, Wallingford, Oxon, OX10 0PH [email protected] Mark Robinson Oxford University Museum of Natural History, Parks Road, Oxford, OX1 3PW [email protected] Liz Somerville Formerly, School of Life Sciences, University of Sussex Falmer, Brighton, East Sussex, BN1 9RH [email protected] Katherine Szabó Principal Research Fellow, Centre for Archaeological Science, University of Wollongong, Wollongong NSW2500, Australia [email protected] Victoria K. Taylor Formerly, Department of Archaeology, University of Reading, Whiteknights, Box 227, Reading, RG6 6AB [email protected]

Nigel Thew Section d’Archéologie et Paléontologie, Office de la Culture, République et Canton du Jura, Hôtel des Halles - CP 64, 2900 Porrentruy 2, Switzerland [email protected] Thomas Walker Department of Archaeology, University of Reading, Whiteknights, Box 227, Reading, RG6 6AB [email protected] Jessica Winder 37 Greenwood House, Sherren Avenue, Charlton Down, Dorchester, Dorset, DT2 9UG [email protected]

xiii

Acknowledgements

The Conchological Society of Great Britain & Ireland championed the concept of this book from its inception; it co-hosted the Molluscs in Archaeology conference in 2014, significantly financially supported this publication, and many of its members have contributed to this book, or acted as advisers or referees. Of these Jan Light has to be singled out for special thanks for her immeasurable help, support and assistance, especially on the marine shells. Throughout the production Bas Payne and Terry O’Connor provided sage advice and wisdom, the latter provided much useful editorial advice; all much welcomed. Additional figures were drawn for the contributors by Abby George. I would like to thank Tom Walker for checking and discussing information, and Ben Urmston and Emma Firth (AC Archaeology) for scanning large snail histograms. The final production would not have been possible without Julie Gardiner’s advice, and assistance with checking the digital suitability, rectifying and resizing all the digital figures. Thanks as always to our typesetter Julie Blackmore, who continues to be not just excellent but quick and efficient, and has produced pleasing page layout. Apart from the significant contributions from the Conchological Society and Allen Environmental Archaeology, this book is sponsored with contributions from the Marc Fitch Fund, the Association for Environmental Archaeology, and the Research Laboratory for Archaeology and the History of Art, University of Oxford. Additionally, I am humbled by the very significant donations provided by several private anonymous donors, which assisted in enabling the book to be indexed, facilitated more colour, and the keen cover price. The conference which trailed the concept of this book was held at the Natural History Museum in April 2014 and organised jointly by the Conchological Society of Great Britain & Ireland and the Association for Environmental Archaeology, for which thanks for hosting, organising and helping on the day go to Richard Thomas (chair AEA), John Tweddle and Andreia Salvador (Natural History Museum), the late Ron Boyce, Rosemary Hill (Conchological Society), Colin Forrestal, Evan Hill, Jan Oldham, Suzi Richer and Eric Tourigny.

Molluscs in archaeology: an introduction Michael J. Allen and Bas Payne

The purpose of archaeology is to find out more about what people did in the past, and how and why this changed; an important part of finding out about past human behaviour is to understand more about the environment in which people lived, and how and why that changed. Unfortunately, for most of our past, only physical evidence survives – and in many parts of the world, little of that remains. Mollusc shells, however, together with bones, pollen, and charcoal, are preserved better than most other animal and plant remains, and provide an important source of information. Molluscs are an ancient invertebrate group – over 550 million years old. There are something like 85,000 described species, which have evolved and become adapted to a very wide range of habitats and ways of life, in the sea, in freshwater and on the land, being absent only from the coldest and driest environments. Most lay down shells around their bodies, made mainly of calcium carbonate (CaCO3), which protect the soft animal inside from drying out, from damage, and from predators and parasites; and it is these shells that survive in the palaeontological and archaeological record. There are five main classes of living shelled molluscs: Bivalves have paired shells – eg, mussels, oysters and scallops. They live in marine and freshwater environments, mostly, partly or entirely buried in soft sediments, and are nearly all filter feeders. They are important in helping to keep water clean. Gastropods have a single shell, usually coiled – eg, whelks, winkles, garden snails and limpets (whose shells are simple low cones, though their distant ancestors were coiled). They live on land as well as in salt and freshwater, and are much more diverse than bivalves, including herbivores (as any gardener is very aware), detritivores, carnivores and parasites. The other three classes – chitons, Monoplacophora and tusk shells, are less abundant and less diverse, and live only in marine environments. Archaeologically, molluscs provide three main kinds of evidence. First, and probably most important, is evidence that helps us to reconstruct past environments, especially terrestrial environments in drier areas. Although pollen is probably the best-known source of evidence for past terrestrial environments and environmental change, it relies on waterlogged conditions such as peat bogs, mires, and pond and lake sediments for its survival. In drier areas, long pollen sequences are scarce or non-existent. Fortunately

2

Michael J. Allen and Bas Payne

land snails, even though usually thin-shelled, survive in large numbers in hillwash, other deeper stratified deposits and many archaeological contexts, if they are neutral or alkaline, as in South-eastern England. This evidence, therefore, fills a void left by the lack of pollen survival. The evidence snails provide differs in one important respect: pollen, being tiny and airborne, provides direct information about vegetation and individual species, while snails indicate the structure of the vegetation regime. Pollen also tells us about conditions in a fairly large area upwind of the bog or lake; land snails tell us about much more local conditions. Second, people often collect and eat molluscs; their shells accumulate in archaeo­ logical sites, sometimes in enough quantity to form middens, both on coasts and, where land snails were collected in large numbers, inland too. They tell us about human diet. Detailed study, as different contributions in this volume show, can tell us which times of the year they were collected and the site was occupied, and whether the molluscs were overfished or over-collected, how far people were going to collect them, and whether the local environment was changing. In areas with acid soils, shell middens may create locally alkaline soil conditions, preserving bones and teeth which might otherwise be destroyed. Finally shells, especially marine shells, provide raw material for tools and ornaments; these, as in the case of Spondylus bracelets in the European Neolithic, are often found far from where the mollusc lived, telling us about past trade networks. There seems to be no text book embracing molluscs in archaeology as a whole: many books on the biology, ecology or identification of molluscs are designed for biologists, and ecologists collecting modern specimens, not the battered, broken and faded fragments recovered from archaeological contexts. Even Claassen (1998), which does cover both marine and land shells, dedicates fewer than 10 pages (10

15–20

37

12

18

37

Good but often (>70%) mechanically taken at 10cm intervals, usually regardless of context boundaries Good & taken at varying sample interval taking heed of deposit formation, context & horizons boundaries

95

80

70

(I suspect that as a snail specialist that I am given the better snail samples, so this probably under-represents field sampling)

Samples Land snails are small, and shell fragments to 0.5 mm are identifiable and quantifiable, so it is imperative that they are recovered from soil samples processed appropriately, rather than hand-picked on site. Hand-collected shells have limited palaeo-environmental value; they tend to be just the larger and more robust species (Fig. 2.1) which are not representative of the palaeo-fauna. They are often just five or six of the 118 snail and slug species in the British fauna. Samples must be large enough to recover, ideally, at least 100 shells or more … but not so many as to be overwhelming. On average, therefore, John Evans would recommend a sample size of 1000 g (Evans 1972, 41), though I tend to use 1500 g as my standard. Evans suggested removal of large stones by eye. As a standard I sieve air-dried soil through a 16 mm mesh sieve providing consistent removal of the larger stones. In stony fills, especially primary ditch fills, this may require samples in the field of 2, 3 or 5 kg to achieve that volume. If large shells are hand collected they may not be entirely valueless. They are often quick and easy to identify and quantify as they are generally just the larger species (eg, Cepaea, Helix/Cornu etc) which may be slightly under-represented in soil samples. Washing the small volume of soil from the interstices of the shells may recover smaller species from which some palaeo-environmental comment may be possible; as

32

Michael J. Allen

Figure 2.1. Comparison between assemblages; left, processed soil sample, and right, hand-picked (after Evans 1972, fig. 2, and Sparks 1961, fig. 1; illustration: Abby George) demonstrated by the hand-picked assemblages from Round-the-Down (Allen 1995a; see Allen, Chapter 1), and Lewes Priory (Allen 1997), both early 1970s excavations in East Sussex. They are no substitute for proper soil samples.

Samples and sampling It is important to obtain snail-specific samples, as the archaeologists and analysts should be addressing site-specific, time-specific and feature-specific questions which will allow themes relating to the wider landscape to be addressed. Sequences of contiguous samples, taken at the appropriate interval for deposit formation, are required from at least one, if not more, locations on site to enable a history of environmental change to be defined. Bulk samples (typically 10–40 litres) taken routinely from archaeological deposits for charred plant and charcoal remains are usually taken at too crude a depth interval to provide valuable information. Bulk sample flotation processing for the recovery of land

2. The geoarchaeology of context: sampling for land snails

33

snails is rarely compatible with standard practices which require flot and all residue to be retained to at least 0.5 mm (Evans 1972, 44–45; Davies 2008, 5–6; and see below). Although standard bulk flotation flot meshes of 250 μm or 300 μm are compatible, the fact that residues are often discarded, or are retained on only 1 mm mesh, can result in loss of a significant and biased proportion of the snail assemblage. Even the 0.5 mm to 1 mm fraction often contains a significant portion of the assemblage and I have recorded as much as 64% of the assemblage in this fraction! Often grab samples are taken by archaeologists for the analyst for snails, and done so asking rather simplistic and not always very helpful questions. Sequences of samples are much more useful (see buried soils, ditch fills and colluvium), together with careful consideration of the taphonomy of the deposits – this may benefit from discussion between the archaeologists and the geoarchaeologist/palaeo-environmental archaeologist/molluscan analysts. Individual spot samples can have some value. If no long sequences are available, a series of carefully taken and selected samples may enable a chronological set of samples to be examined, providing a potted chronology of the sites land-use and environmental history. This may be suitable in defining and characterising local site histories, but not for detailed analysis of land-use changes. The amount of ecological information that can be gleaned from single spot samples is minimal in comparison with sequences of samples, and often does little more than provide general characterisation.

Taking your samples Sampling is not just a mechanical operation of collecting and bagging soil – it needs careful consideration of the deposit formation. A few hours on site may result in weeks or months of subsequent analysis. Samples need to be taken with care. The identifiable snail fragments are as small at 0.5 mm (look at your ruler!). Any dirt, clods of soil or clay adhering to sampling implements could contain shells and contaminate samples. Using a dirty spade or, in fact, even a clean spade, is generally unacceptable. Sampling must be undertaken, carefully, clinically, and with clean tools (trowel, hand shovel, scoop etc), and these need to be thoroughly scraped clean between samples, if not actually washed with water if the deposit is moist and really sticky. Sloppiness and errors in the field cannot be rectified, nor necessarily identified, in the lab. Minutes of failure in the field potentially wastes not just hours in the lab, but weeks or months of analysis! Time expended on site ensuring detailed description and understanding of the deposits, making good records and photographs, is never wasted. Corners cut at this stage cannot be rectified, will be regretted, and will be detrimental to the time expended later, and results gained.

Ten steps to sampling Details of how to take a column of samples is given by Evans but a quick 10-step guide list is presented below. Ideally a column of contiguous samples should be removed

34

Michael J. Allen

Figure 2.2. Strawberry Hill, Wiltshire. Column of 32 samples covering 3.1 m of colluvium; note complementary samples for soil micromorphology from stone-free lenses (photo: author)

from the exposed section (Fig 2.2). If the column is 20 cm wide then samples of 10 cm thickness allow reasonably easy recovery of samples of 1–2 kg. Sample interval, however, must be dictated by the speed and nature of deposit formation. Samples should never cross horizon boundaries. Too often now samples are taken routinely, systematically, mechanically and unthinkingly at 10 cm intervals, and then provided to the specialist. This denies many of the key aspects laboured by Evans in his excellent but now 40-year-old book (1972), that the geoarchaeological record is as much a part of the interpretation as the snails themselves. A note on nomenclature: Often a ‘column of samples’ for land snails is taken through a deposit or through a feature. Unfortunately this can be confused with a monolith of undisturbed sediment, which archaeologists often call a ‘column sample’. On a number of occasions a column of samples has been requested, only to be provided as a monolith (ie, column sample). Consequently, I advocate never using the term ‘column sample’ but only using monolith, undisturbed sample (kubiena sample) and ‘column of samples’ so as to avoid this potential, and unfortunate, confusion.

2. The geoarchaeology of context: sampling for land snails

35

1. Review the project aims and site with the field archaeologist, and select your key, dated or phased sequences, ensuring that each profile has already been appropriately recorded (drawn, photographed) by them 2. Clean the section without smearing; break off soil lumps, rather than trowel a smooth polished surface, to expose pedological structure, and then photograph 3. Inset a datum nail and run a tape from the datum to the base of the profile. Make your own sketch section and describe in detail by depth, using pedological notation (eg, Hodgson 1997 and Munsell colour chart) ie, depth, colour, texture, structure, stoniness (size, shape and abundance), inclusions, archaeology boundary, interpretation 4. String out a sample column – typically 20 cm wide. Run a tape from a datum at the top of the column 5. On the basis of the sediment interpretation based on your descriptions and observations (and those of the archaeologists) decide on appropriate sample intervals (which may vary through the profile). If in doubt take samples at closer intervals: they can always be amalgamated in the lab, but not separated. 6. With measurements from top down, start sampling from the bottom upwards, taking extreme care to keep within horizons/context boundaries. Work neatly and systematically, removing large stones from the sample as you go. In general aim to acquire samples of similar size throughout the column, regardless of the thickness of the samples, or sampled deposit (obvious exceptions are stony primary fills or obviously very rich shelliferous layers) 7. Label each bag with sample and depth. Seal and set aside 8. On completion ensure the sample column, and any additional samples, are marked on a section drawing. Or at the very least that the datum point/s are clearly recorded on the site section drawing 9. Photograph the sampled column 10. List all the samples for your, and the archaeologist’s, benefit

Processing samples The processing of land snails is described by both Evans (1972) and Davies (2008), and has been published elsewhere and is summarised below. Processing small 1–2 kg samples for land snails is a relatively easy process but does require fastidiously clean (laboratory) work environment, precision laboratory test sieves (Endecott style), and facilities to dry sieves where the light contents will not be disturbed, or contaminated (ie, a drying cabinet). Samples are registered into laboratory records (accessioning system), and set out to air-dry; samples can lose up to c. 20% of weight on air-drying. Air-drying makes samples more archive stable, but importantly provides a dried constant. Although when sampling (see above) large stones are removed, I sieve all air-dry samples through a 16 mm mesh to provide metrical consistency between samples. This is also useful in rock-rubble deposits such as the primary fill of ditches, ensuring that more of the processed material potentially contains snail shells, rather than large chalk stones. Pre-weighed samples of constant weight are placed in a bucket and soaked in warm water and stirred. The immediately floating shells are decanted onto a 0.5 mm mesh sieves. Washing and decanting is repeated until no more shells float, and/or the water goes clear. It can be beneficial then to tip the sample through a small stack of 0.5 mm

36

Michael J. Allen

and 4 mm mesh sieves to remove all the large stones (which are set aside to dry), and some of the easily removed >0.5 mm mud fraction. The 0.5 mm sieve is flushed with water and shaken to wash through as much soil as possible before being returned to the bucket with warm water. A disaggregant (eg, hydrogen peroxide (H2O2), decon 90, or sodium pyrophosphate for very organic samples, or less favourably anhydrous sodium hydroxide, or potassium hydroxide) can then be added and the sample stirred and left to soak ideally for 12 or 24 hours, stirring occasionally. (Note that Calgon no longer works as well as a disaggregant as the active ingredient sodium hexametaphosphate is no longer included. Also, the use of any disaggregant which may contain phosphates should be avoided as they are difficult to extract from waste water and when water is discharged back into the water system, such as a river/lake, this can cause eutrophication. It also can cause some surface shell damage). All new flot should continue to be decanted onto 0.5 mm mesh sieves, with the bucket refilled and decanted five or six times, or until the water is clear and no further material floats. The residue (ideally just clean stones), should then be washed/flushed through a stack of 0.5 mm, 1 mm, and 2 mm mesh sieves; each residue being carefully and individually washed through the stack of sieves before being set aside to drain, and then dry. A rapid scan of the flots will determine the likelihood of the samples to contain anything like enough shell; bearing in mind that up to 60% of the assemblage may still reside in residues. Formal assessment of the flot can be undertaken by the specialist, prior to selection of samples for sorting, extraction and analysis. Flots and residues are then fully sorted and shell extracted under magnification of ×10 to ×30/45 using a stereo-binocular microscope. Only apical (or apertural) and representative diagnostic fragments (including slug plates) need be recovered. This then enables identification, quantification, tabulation and the application of any statistical analysis (see Law, Chapter 3) such as species diversity (eg, Shannon Index (H′), Brillouin Index (HB) (see Allen 2003; 2007a), or delta indices (∆2 and ∆4) (see Enwistle & Bowden 1991), or other tests (see Thomas 1985; Magurran 1988).

Sampling: addressing the archaeological deposits Columns of contiguous samples should be taken, where possible through a range of deposits to build up, principally, a chronological sequence of the site. Some deposits may be more relevant and useful to sample, others are almost pointless. Some indication of the taphonomy, value and consideration for typical key, and commonly encountered, deposits are given below; the two most significant are buried soils and ditch fills (Table 2.2). The point has already been made but cannot be over emphasised that samples must always be context specific, and never cross context boundaries. They must be large enough to yield 100–200 specimens. Sample thickness is dependent on the speed of deposition of the deposit: rapid deposition (primary fill) may require samples of 15 cm, 20 cm or more thickness, while a slow laminated silt may require sampling at 2 cm or 4 cm. The deposit represents time, and you are sampling a portion of that time.

2. The geoarchaeology of context: sampling for land snails

37

Table 2.2. Simplified summary guide of aims and sampling strategy for a variety of commonly encountered contexts, deposit types and features Typical key aims Buried soils Ditches Large Pits

High quality pre-monument environment information Pre, contemporaneous and post monument and site environment

Small pits

Context specific micro-environments and postpit land-use environment –

Postholes



Treehollows etc Banks & mounds

Detailed contemporaneous land-use and environment, but commonly prior to the archaeological activity –

Middens Colluvium Alluvium

– Long stratified sequences of land-use Long stratified sequences of land-use

Typical sampling strategy Close-interval contiguous column Combination of principally a contiguous column augmented with ‘spot’ samples Contiguous column through ‘natural’ colluvial infills Generally poor contexts and only context-specific sampling Rarely worth sampling, unless very large postpits Sequence and spot samples from specific contexts Basal soil, and turves only (see buried soils) Generally poor contexts contiguous column contiguous column

Buried soils Buried soils usually represent long periods of time for which their burial can often be well-dated. They are however, complex in their formation, and may contain considerable pedogenic, biological, physical and chemical reworking which have implications for shell survival, breakage and movement within the profile (Carter 1990). Soils can have clear horizons, not layers, which are formed in situ, and the recognition of this of soil formation (pedogensis) is imperative. On burial the soil may ‘collapse’ or become squashed as voids collapse, but the soil fauna may still survive for a while; some of the soil may be worm-worked into the base the bank, barrow or rampart, as Macphail demonstrated under the Iron Age rampart at Balksbury, Hampshire (Macphail 1986). Soils need to be carefully examined, understood and appropriately, and often carefully close-sampled. Single ‘bulk’ samples encompassing the whole profile can combine different snail assemblages covering thousands of years in which there may be significant change. A bulked sample across this will provide a subfossil assemblage which is nonsense in terms of any former snail population or fauna. Buried soils, more than any other type of context routinely encountered on site, need more careful consideration and sampling. It is important to define in the field, and appropriately sample, these soil horizons and recognise any post-burial alteration. Selecting the sampling location of an exposed

38

Michael J. Allen

buried soil should try and include the best preserved profile, avoid cut features, and include, where possible, a Holocene treehollow or subsoil hollow beneath the soil. A bank has little value. It is normally comprised of the parent material (geology) and will essentially contain no shells. The origin of those few it does contain cannot be determined; they may relate to any or all of soil or archaeological deposits scooped up during construction, or shells fallen in crevices, cracks, voids and rootholes, in the bank or from soils forming over the bank. Typically the Ah horizon (turf), A horizon (topsoil), B horizon (rest of the profile) and B/C or B/R horizon at the base of the soil should be defined, and samples of 4–5 cm, or even 2 cm taken contiguously through them. Typically samples are narrower in the turf (say 2 cm), and larger in the main part of the profile (B horizon). The cruder the sampling interval (say 5–8 cm) through these long-developed sequences, the more likely there will be conflation of very different mollusc faunas.

Ditches One of the most commonly sampled, but also complex, types of context are ditch fills. The taphonomy of the assemblages is related to the tripartite ditch infilling sequence as defined by Evans (1972, 321–8) and Limbrey (1975, 290–300). These provide the key sedimentary units as defined by product (Evans) and process (Limbrey), and summarised by Allen (1995b). Not all ditches contain all three fills and the ditches are not necessarily uniform in the distribution and nature of such fills along their length. The recognition of primary, secondary, and tertiary fills has significant archaeological and interpretative implications and is a convenient method of grouping together a number of contexts which belong to the same episode or phase of ditch infill (but are not necessarily directly tied to the phases of archaeological activity). In addition to this I have added an ‘initial deposit’ and formally included the stabilisation horizon and buried soil separately rather than just as the end of the secondary fill. The taphonomy of the snail assemblage can be considered by each of these defined units (Fig 2.3). Initial Wash: this thin parent-material derived deposit (ie, typically a calcareous mud) represents the first rainwash. It principally includes shells washed from the soil through which the ditch was cut. Fallen turves may also lie on or in this deposit (see Crabtree 1971; Bell 1990; Bell et al. 1996). The deposit is often only a few centimetres thick, so sampling can be challenging, and may require the overburden to be removed or excavated in order to sample this deposit in plan. After taking a contiguous column of samples, it is often possible, with the archaeologist’s agreement, to excavate a portion of the ditch fill to expose this thin deposit for sampling. Primary Fill: the majority of this is clastic material derived from the weathering of the sides – resulting in typically in chalky or stony rubble at the base of the ditch. Initially just occurring at the sides of the ditch, before infilling, often the majority of the ditch.

2. The geoarchaeology of context: sampling for land snails Process

39

Product

INITIAL FILL: a thin parent material-derived wash on the very base of the ditch, & possibly accumulating at the bottom of the sides. It may also include or contain fallen turves. Duration: probably the first 1–2 winters. Rain wash of a) some soil material through which Thin parent-material derived wash. the ditch was cut, and b) of the exposed parent material of the ditch sides. PRIMARY FILL: Weathering of the ditch sides. This can be as much as 70% of the ditch infilling where loose & unstable geologies are concerned. In some cases this may superficially look like the ‘natural’ Duration: These fills can occur in about 30–60 years. Weathering & frost shattering of the newly cut Weathering results in the accumulation of ‘clean’ ditch sides, the soil & turf line, & weathered angular, rock debris largely in the corners (flatnatural (regolith). May include the results of bottomed ditches) or the base (V-shaped ditches). immediate settling & erosion of any adjacent Turves & (top) soil may fall into the ditch & come to bank. rest of its sides or base, appearing as relatively well i) Trample in final stages of digging defined dark humic patches. The primary fill may ii) Initial rainwash of cut geology (chalky marl) have a laminated appearance. Primary fills can be iii) Weathering & frost shattering of the ditch thin bands or occupy as much as 60–70% of the sides ditch. iv) Falling of topsoil/turves derived from the soil i) Dark line, perhaps soil surface through which the ditch was cut ii) Fine wash iii) Angular rock debris, screes, accumulating from the corners of the ditch SECONDARY FILL: Weathering back of the poorly vegetated ditch sides Duration: Decades to centuries. Gentle gradual & continued weathering of the Finer, more humic sediments which tend to become ditch sides & surrounding soil when covered in less stony higher up the ditch profile & may show patchy plant cover. Natural accumulation of ditch traces of bedding. silts, often occupying the main body of the The rate of infill may slow, enabling vegetation feature. Slows with time slows as the angle of colonisation to increase to such an extent that a slope decreases & the infilling material becomes stable soil horizon may develop which, in turn may colonised with patchy vegetation. show evidence of worm sorting into stone-free or small stone lenses. STABILISATION HORIZON/BURIED SOIL: Slowing of sedimentation rates, stabilisation of the ditch sides & herbaceous vegetation grown in the ditch & sides, leading to pedogenesis & soil formation. Duration: Centuries. Speed of sedimentation decreases & almost A less stony, & possibly almost stone-free, horizon stops, allowing typically herbaceous vegetation at the top of the secondary fill. Possibly slightly to take hold, & soil formation to occur. darker in colour (humic), with weak ped structure, & often more dense artefacts content due to the longevity of the deposit formation. TERTIARY FILL: final ditch fill, generally derived from ploughing & plough wash or deliberate backfill Duration: Centuries to millennia. Homogeneous colluvial material, resulting from Finer colluvial & soil material with few large stones cultivation or deliberate infill of the surviving, and common small stones. shallow, ditch & associated with farming, clearance, or landscaping activities.

40

Michael J. Allen

Figure 2.3. Schematic drawing of ditch infilling process, summarising process, product and origin of subfossil land snail assemblage based on Evans (1972, fig. 123) with modifications and additions. Horizontal arrows indicate immediate range micro-environmental the snails reflect (illustration: Abby George) Most of the shells here are derived from the soil through which the ditch was cut (palaeofauna), and turves fallen into the ditch (see Bell et al. 1996, fig. 4.6). Only a few snails are likely to inhabit the short-lived inhospitable ditch micro-environment, and those that do originate from the soil surface. The assemblages largely relate to the pre-ditch environments. The stony nature and rapid mode of infilling requires large samples up to 5 kg and at intervals of perhaps 15–20 cm thickness or more. Secondary Fill: this a mixed fill, usually containing the majority of the artefacts as it relates to the period of site use and occupation. Fills are finer, less stony, and sometimes

2. The geoarchaeology of context: sampling for land snails

41

weakly banded. Snails represent both the ditch micro-environment, with some patchy herbaceous vegetation, but mainly the immediate ditch environs. These deposits form gradually and incrementally and sample intervals of 10 cm are common; and generally 8–12 cm. Stabilisation/Soil Formation: soil formation commonly occurs at the top of the secondary fill as the ditch sides stabilise, weathering and infilling slows, and the ditch becomes more vegetated. The snails are more representative of the wider landscape, with the exception of any more mesic long vegetation within the ditch itself. The slowing of deposition rates and longer time span it represents should be reflected in the sample intervals, typically of 4 or 5 cm or less, dependent on how well pronounced is the horizon. Tertiary Fill: a ploughwash, formed after the abandonment of the ditch as a feature, or monument, or of any accompanying settlement. The finer, generally poorly sorted, homogeneous colluvial fills are representative of the wider local landscape, and samples are commonly 10 cm thick. A completed sequence of samples, which may include spot samples from primary fill in the corner of the ditch profile and a column of contiguous samples through the deepest and most representative deposits, should aim to be cover the entire infill history. It does not need to sample every layer, but sample ‘time’.

Pits Generally only large pits with colluvial infills are worth sampling. Many of the contexts often represent re-deposited or dumped material, rather than natural sedimentation. They may contain material derived from unknown, or undefined, locations (Shackley 1976). The origin of these snails is unknown; they do not reflect the environment of the pit, and they may originate from several different micro-environments conflated into a single discard deposit, and thus are not useful for palaeo-environmental reconstruction. Open pits will create a localised micro-environment that may encourage specific species to inhabit and thus not necessarily, reflect that of the wider environs. Deposits to concentrate upon are any basal or primary fills, and the upper natural colluvial infilling (‘tertiary fill’); an example of useful sequence through such are shown in the large Mesolithic postpits at Stonehenge (Allen 1995c), and of progressive opening of the landscape in the top of the Neolithic pit at Easton Lane, Hampshire (Allen 1989; Fig. 2.4). Spot samples through the colluvial infill of the very large rotted post in the Neolithic postpit Greyhound Yard, Dorchester, were able to indicate a locally open woodland environment after the monument’s demise (Allen 1993).

Postholes In general postholes are poor and not useful contexts to sample for snails. Unless these are very large postholes, or postpits, like that at Stonehenge (Allen 1995c) or Greyhound

42

Michael J. Allen

Figure 2.4. Summary of the land snail evidence from Easton Lane pit 1017, showing progressive opening of the local landscape (after Allen 1989, fig. 108; illustration: Abby George)

Yard (Allen 1993), there is no indication of when the shells got into the features. They could derive from the soils through which the posthole was cut, from the environment during the life of the extant post, or when the post had rotted or was removed.

2. The geoarchaeology of context: sampling for land snails

43

Treehollows and tree-throw hollows Differentiation between a treehollow and tree-throw hollow is important. They are very different features with quite different fills and fill histories. A treehollow, or subsoil hollow, is the natural hollow created by subsurface weathering of the parent material by the roots. These create subcircular or sub-oval features, commonly with weakly defined edges and bases and mixed, but often homogeneous fills comprising chalky and ‘colluvial’ fills (see Allen 1995d, fig. 69; Allen 2000, figs 42 & 43). In contrast a treethrow hollow is a pit or open feature created by a fallen or toppled tree; with a large root plate ripped from ground, and tipped vertically with a large void behind. These create crescentic shaped features, with contrasting fills (see Macphail 1987; Macphail & Goldberg 1990).

Banks, mounds and dumps Banks and mounds are principally parent material (geology) and soil thrown up to create upstanding relief. They are very poor contexts, and not worth sampling. Only intact turves on the old land surface or those within these structures are worth sampling (see buried soils).

Quarries, ponds, scoops and hollows A range of larger negative features such as quarries, ponds, scoops and hollows will have essentially a primary weathering fill and tertiary colluvial fill. As such they can largely be dealt with like shortened ditch fill sequences.

Middens In general middens are poor contexts of dumped accumulated material. Buried soil beneath middens may be of value. This may be especially true of shell middens which may result in preservation of snail shells in otherwise poor preservational, only weakly calcareous, environments. Although shell middens may also preserve land snails within them, temptation should be resisted to sample them as the palaeo-environmental value of these shells is usually generally poor.

Colluvium, ploughwash and lynchets Colluvium (or hillwash) occurs at plateau edge, footslope locations and in dry valleys (Bell 1981, fig 5.1, Allen 1988, fig. 6.5; French 2015, fig. 6), and as ploughwash in lynchets (Fowler & Evans 1967, fig 3c, Evans 1972, fig. 119). These are usually off-site deposits, but are more routinely encountered in either commercial archaeology or landscape

44

Michael J. Allen

research projects, rather than ‘set-piece’ excavations. The majority of the deposits are unsorted homogeneous hillwash, with occasion rills or gravel fans (see Allen 1988; 1992), and sampling can be relatively uniform at 10 cm intervals, but at finer intervals when stasis or stabilisation horizons are identified (see for instance Strawberry Hill, Wiltshire, Fig. 2.2). Occasionally the dryland edge of floodplains contain predominantly colluvial deposits which are darkened by alluviation and may contain only some indications of the floodplain presence (see Bourne Valley, Allen 2007b), rather than full floodplain sequences (see Chapter 7, Robinson). Lynchets are colluvial accumulations of ploughwash at the edge of the field. Sometimes they may have a boundary (ie, wall or hedge), in other cases the field edge can be defined by a negative lynchet (Evans & Fowler 1967, fig. 3c). It is unusual to have a basal buried soil (contra Evans 1972, fig. 119), and the Bishopstone lynchet, East Sussex (Bell 1975, 252–64), and the Stepleton lynchet on Hambledon Hill, Dorset (Bell et al. 2008, 432–3) provide good examples. Like colluvial sequences, samples through the body of the lynchet are typically taken at 10 cm intervals; ideally contiguous, but that may depend on the size of the lynchet and the research question posed. Mollusc diagrams from dated lynchets can be seen from Hambledon Hill (Bell et al. 2008, fig. 5.9), and the Bourne Valley, Eastbourne, East Sussex (Allen 2007b, figs 18 & 21). Negative lynchets may look like ditches and are shallow colluvium-filled scarps at the head of the lynchet; the fills of which usually post-date tillage.

A sampling strategy and beyond A programme of land snail samples needs to be devised and set within that for other palaeo-environmental proxy data such as pollen, phytoliths, waterlogged remains and charred plant and charcoal. The complete suite of land snail samples should encompass the key archaeological periods of interest, and provide information of the environment prior to key archaeological activity. Some features or layers may raise context-specific questions, which need to be evaluated within the context of the overall aims of the project. Sometimes this can be undertaken on site, but it can be re-considered on completion of sampling and during reviewing the samples after the completion of fieldwork. Although an outline generic sampling programme can be provided prior to fieldwork, all sampling needs to be reactive, and re-designed in accordance with the excavation discoveries. On large excavations which may cover several areas, or be undertaken over several months, it may be necessary for the specialist to visit the fieldwork on several occasions to ensure full and appropriate sampling. Samples are not sacred. A site can always be oversampled, the samples reviewed, prioritised and some set aside or even discarded, but it is much more difficult to recompense for sequences, features or contexts not sampled. Some contexts may no longer exist. The balance of samples taken needs to take account of the resolution of land-use history required, and the amount of research capital (finance) available; land snail analysis is time consuming – about eight samples can be processed per day, and just a few extracted and identified per day. Whole programmes of analysis are commonly weeks to months in extent.

2. The geoarchaeology of context: sampling for land snails

45

At the completion of fieldwork, a programme of sample processing will allow the flots to be scanned and assessed. Features, sequences or samples with too few snails for statistically viable analysis can then be reviewed, and selection of samples for analysis made. In some cases, even samples with slightly too few snails might be worth pursuing if they form apart of stratified sequence, as some of the snail histograms in the previous chapter show. The completion of analysis, any statistical examination, and graphic representation should then allow the analyst to address the aims defined at the outset, whether they be context- or feature-specific, or form part of the wider land-use history of the site. Comparison of these interpretations with published data and grey literature data (if available) then allow the land-use history and archaeological narratives to address wider regional and period-related narratives, if not individually, certainly collectively (see Allen, Chapter 9).

Acknowledgements I’d like to thank various colleagues for sharing information with me, a large number of field staff who have assisted and been involved in sampling snails from sites, and Alan Clapham, Chris Hunt and Terry O’Connor for discussing (amongst other things) their experiences with a variety of disaggregants.

References Allen, M. J. 1988. Archaeological and environmental aspects of colluviation in South-East England. In Groenmann-van Waateringe, W. & Robinson, M. (eds), Man-Made Soils, 69–92. Oxford: British Archaeological Report S410 Allen, M. J. 1989. Land snails. In Fasham, P. J., Farwell, D. J. & Whinney, R. J. B., The Archaeological Site at Easton Lane, Winchester, 134–140. Winchester: Hampshire Field Club and Archaeological Society Monograph 6 Allen, M. J. 1992. Products of erosion and the prehistoric land-use of the Wessex chalk. In Bell, M. G. & Boardman. J. (eds), Past and Present Soil Erosion: archaeological and geographical perspectives 37–52. Oxford: Oxbow Books Allen, M. J. 1993. The land snails. In Woodward, P., Davies, S. & Graham, A., Excavations at Greyhound Yard Dorchester 1981–4 340–345. Dorchester: Dorset Natural History & Archaeological Society Monograph 12 Allen, M. J. 1995a. Land-use history of Round-the-Down; the molluscan evidence, 13–16. In Butler, C., The excavation of a Bronze Age round barrow at Round-the-Down , near Lewes, East Sussex, Sussex Archaeological Collections 133, 7–18 Allen, M. J. 1995b. Ditch and feature fills. In Cleal, R. M. J., Walker, K. E. & Montague, R., Stonehenge in its Landscape; twentieth century excavations, 4–6. London: English Heritage Archaeological Report 10 Allen, M. J. 1995c. Before Stonehenge. In Cleal, R. M. J., Walker, K. E. & Montague, R., Stonehenge in its Landscape: twentieth-century Excavations, 41–63. London: English Heritage Archaeological Report 10 Allen, M. J. 1995d. Land molluscs. In Wainwright, G. J. & Davies, S. M., Balksbury Camp Hampshire Excavations 1973 and 1981, 92–100. London: English Heritage Archaeological Report 4

46

Michael J. Allen

Allen, M. J. 1997. Land molluscs. In Lyne, M., Lewes Priory, Excavations by Richard Lewis 1969–82, 163–67. Lewes: Lewes Priory Trust Allen, M. J. 2000. Wood, farm and field: landuse history of Twyford Down – land snail evidence, the pre-Bronze Age environment. In Walker, K. E. & Farewell, D. E., Twyford Down Hampshire; investigations on the M3 motorway corridor from Bar end to Compton, 1990–93, 139–142. Winchester: Hampshire Field Club Monograph 9 Allen, M. J. 2003. Appendix 1: definition of the Shannon, Brillouin and delta indices 233–234. In French, C., Lewis, H., Allen, M. J., Scaife, R. G. & Green, M., Archaeological and palaeoenvironmental investigations of the Upper Allen valley, Cranborne Chase, Dorset (1998–2000); a new model of earlier Holocene landscape development, Proceedings of the Prehistoric Society 69, 201–234 Allen, M. J. 2007a. Molluscan diversity. In French, C., Lewis, H., Allen, M. J., Green, M. Scaife, R. G. & Gardiner, J. 2007. Prehistoric Landscape Development and Human Impact in the Upper Allen Valley, Cranborne Chase, Dorset, 263–271. Cambridge: McDonald Institute Monograph Allen, M. J. 2007b. Evidence of the prehistoric and medieval environment of Old Town, Eastbourne: studies of hillwash in the Bourne valley, Star Brewery Site, Sussex Archaeological Collections 145, 33–66 Allen, M. J. 2017. Land snails in archaeology. In Allen, M. J. (ed.), Molluscs in Archaeology, 6–29. Oxford: Oxbow Books Allen, M. J. 2017. The southern English chalklands: molluscan evidence for the nature of the postglacial woodland cover. In Allen, M. J. (ed.), Molluscs in Archaeology, 144–164. Oxford: Oxbow Books Bell, M. 1977. Excavations at Bishopstone, Sussex Archaeological Collections 115 Bell, M. 1981. Valley sediments and environmental change. In Jones, M & Dimbleby, G. W. (eds), Environment of Man: the Iron Age to the Anglo-Saxon period, 75–91. Oxford: British Archaeological Report 87 Bell, M. G. 1983. Valley sediments as evidence of prehistoric land-use on the South Downs, Proceedings of the Prehistoric Society 49, 119–150 Bell, M. 1990. Sedimentation rates in primary fills of chalk features. In Robinson, D. (ed.), Experiment and Reconstruction in Environmental Archaeology, 237–248. Oxford: Oxbow Books Bell, M., Fowler, P. J. & Hillson, S. W. (eds), 1996. The Experimental Earthwork Project 1960–92. York: Council for British Archaeology Research Report 100 Carter, S. P. 1990. The stratification and taphonomy of shells in calcareous soils: implications for land snail analysis in archaeology, Journal of Archaeological Science 17, 495–507 Crabtree, K. 191. Overton Down experimental earthwork, Wiltshire 1968, Proceedings of the Bristol Spelaeological Society 12, 237–244 Davies, P. 2008. Snails: archaeology and landscape change. Oxford: Oxbow Books Entwistle, R. & Bowden, M. 1991. Cranborne Chase; the molluscan evidence. In Barrett, J., Bradley, R. & Hall, M. (eds), Papers on the Prehistoric Archaeology of Cranborne Chase, 20–48. Oxford: Oxbow Monograph 11 Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press Fowler, P. J. & Evans, J. G. 1967. Plough-marks, lynchets and early fields, Antiquity 41, 289–301 Law, M. 2017. Numerical approaches to land snail palaeoecology. In Allen, M. J. (ed.), Molluscs in Archaeology, 48–64. Oxford: Oxbow Books Limbrey, S. 1975. Soil Science and Archaeology. London: Academic Press Macphail, R. I. 1986. Palaeosols in archaeology: their role in understanding Flandrian pedogenesis. In Wright, V. P. (ed.), Palaesols: the recognition and interpretation, 263–290. Oxford: Blackwell Scientific Publications Macphail, R. I. 1987. The soil micromorphology of tree subsoil hollows, Circaea 5, 14–17

2. The geoarchaeology of context: sampling for land snails

47

Macphail, R. I. & Goldberg, P. 1990. The micromorphology of tree subsoil hollows: their significance to soil science and archaeology, In Douglas, L. (ed.), Soil Micromorphology: a basic and applied science, 431–440. Amsterdam: Elsevier Magurran, A. E. 1988. Ecological Diversity and its Measurement. London: Chapman and Hall Robinson, M. 2017. Molluscs from the floodplain alluvial sediments in the Thames Valley. In Allen, M. J. (ed.), Molluscs in Archaeology, 112–126. Oxford: Oxbow Books Shackley, M. L. 1976. The Danebury project: an experiment in site sediment recording. In Davidson, D. A. & Shackley, M. L. (eds), Geoarchaeology, 9–21. London: Duckworth Thomas, K. D. 1985. Land snail analysis in archaeology: theory and practice. In Fieller, N. R. J., Gilbertson, D. D. & Ralph, N. G. A. (eds), Palaeoenvironmental Investigations: research design, methods and data analysis, 131–157. Oxford: British Archaeological Report S266

3. Numerical approaches to land snail palaeoecology Matt Law

This chapter outlines statistical methods for dealing with molluscan assemblages. Approaches to data presentation are discussed, assessing the merit of presenting absolute and relative abundance of shells; and detailing methods for representing change through time in vertical sequences of samples. Ordination methods which may be used to explore lateral variation between samples are discussed, along with the use of various measures of diversity and their implications for interpreting taphonomy. Finally, metrical variation within and between collections of single species are explored as palaeo-environmental proxies, and methods for extrapolating climate data from land snails evaluated.

Counting ancient snails: should we bother? Quantitative palaeoecology is a surprisingly contentious topic. The month before I wrote this, I commissioned the analysis of some ostracods from an archaeological site. The very eminent ostracodologist concerned told me that his approach would only discuss relative abundance as ‘all that quantitative stuff is worthless’. In many fields of Quaternary palaeoecology though, statistical approaches have become a key aspect of interpretation, for example the use of transfer functions to match variations in species assemblages to variations in environmental factors, which is quite commonplace in the analysis of proxies such as pollen, diatoms and testate amoebae (Birks 2003). Land snail palaeoecology has somewhat lagged behind other biological proxies in this respect, despite a brief flourishing of interest in the 1990s from John Evans and his students (see below for details). This is perhaps no surprise, as Evans himself came to caution that numerical methods often mask finer details (Evans 2004, 366). Those finer details come out in counts of individual snails, and as Thomas (1985) has explored, the choices we make in presenting that information can also reveal or mask important details (see below). As I hope to demonstrate, going beyond the counts into statistics is by no means worthless, and the application of different tests may reveal much about the data. It is always worth bearing in mind the limitations of the tests

3. Numerical approaches to land snail palaeoecology

49

we apply and the datasets we might apply them to however. As Pielou argued (1975, 34), working back from statistical to ecological hypotheses is a very difficult task. In sub-fossil assemblages, where the contextual information is much less complete, the difficulties of drawing palaeo-environmental, palaeoecological and archaeological inferences are even greater. Birks (1986, 744) reminds us that ‘numerical methods are no substitute for ecological knowledge; they are, however, invaluable in identifying patterns worthy of ecological attention’. As George Box explained, ‘essentially all models are wrong, but some are useful’ (Box & Draper 1987, 424). In this chapter I will explore how we count snails; how we may represent and explore change through time; how we can explore the diversity of taxa within our samples and what that might tell us about the taphonomy of the sample; how we can explore variation between samples; how we might be able to use changes in the physical appearance of species to detect ecological change, and finally how we can detect broad climatic changes using snail assemblages. At each point I will attempt to explain the appropriate methods and evaluate their potential in archaeomalacological analysis.

When is a snail not a snail? Or, why I claim there are 973 Cochlicella acuta in this sample Our problems begin with how we should count our snails. Across palaeontology and bioarchaeology the concept of minimum number of individuals (MNI) has arisen to describe the smallest number of individuals of one species that could account for the number of bits of individuals of that species in a sample (Reitz & Wing 1999, 194; White & Folkens 2005, 339). In the case of vertebrates, with lots of bones that could easily become disarticulated and intermingled, it is easy to see how this could lead to an underestimation of the actual number of individuals. With the exception of those slug species that contain internal calcareous granules and snails with hard opercula, most gastropods only have one hard bit that will be preserved, the shell. Our ideal situation then would be to be presented with a sample containing only pristine, intact shells (and none of those ghastly slug granules). In practice, however, a number of processes (eg, chemical weathering, biological attack, physical disturbance in the ground, and overzealous rubbing of difficult clay balls during sample processing) are likely to result in at least some of our shells being fragmentary. The traditional solution is to count each shell apex (Evans 1972, 44–5), although Evans did note that sometimes apertural fragments of a given species may be more diagnostic and in these cases these may be counted instead; and also that it is always worth looking out for non-apical fragments of species otherwise not represented in the sample. Giovas (2009) notes that even by the low standards of MNI, this approach may lead to significant underestimation. Noting that there are various points that only occur once on the shell (called non-repeating elements), she recommends counting these and using the highest number for MNI. In practice this is best limited to the apex, the umbilicus, and the body whorl with aperture. Where opercula are present, it is worth counting these as well, as differences between the count are a good reminder

50

Matt Law

of taphonomic considerations (ie, why do I have two Bithynia apices and 43 opercula? – see O’Connor, Chapter 8, which discusses an example of this situation). We next need to consider how we will present our numbers. We might be tempted to present the frequency of each species in our sample as a percentage of the total, however this might imply that there is a relationship between the occurrence of each species that may not exist, for example that an increase in one species caused a decrease in another (Thomas 1985, 134; Ložek 1986, 734; Davies 2008, 8). Perhaps more seriously, distortions may result from the super-abundance of one or a few species, low overall numbers of individuals, or the introduction or disappearance of a single species (Thomas 1985, 134; Ložek 1986, 734; Davies 2008, 8). Percentage values do, however, give a sense of general fluctuations through time and are not without value, especially if combined with absolute values (Thomas 1985, 135). It is sadly necessary to mention at this point that whether we present absolute counts or percentages, we do need to say which one we are presenting.

Change through time Snail counts from vertical sequences of samples have traditionally been presented in molluscan diagrams (see Fig. 3.1 – this example is plotted using individual taxa, but often ecological groups are used instead), a graphical representation of molluscan biostratigraphy and archaeological lithostratigraphy. Molluscan assemblage zones may then be derived from phases of change, giving us a starting point for interpretative discussion. In land snail analysis, this has tended to be a subjective exercise. A more objective approach is favoured in palynology, using dendrograms derived from Cluster Analysis. This is a method for finding similarities in a dataset. Hierarchical methods either divide the data into progressively smaller clusters (these are called divisive methods) or work backwards and agglomerate the data from small clusters of the most similar items back to large clusters (agglomerative methods) (Smith 2014, 183). The idea behind divisive methods is that large differences within the data should prevail over smaller differences (Dale & Dale 2002, 262). Agglomerative methods meanwhile consider ‘local’ similarity more important than large differences (Dale & Dale 2002, 263). It has been stated that agglomerative methods are most suitable for biostratigraphic zonation (Grimm 1987, 14). Methods used for biostratigraphic zones need to be stratigraphically constrained, so that clusters contain only stratigraphically adjacent samples. Once we have a dendrogram, it is time to think about zones. Something of a subjective element creeps back here, essentially we pick a point just past a split between multiple branches and draw a horizontal line across the tree there (in the mid-point between the horizontal lines of the tree). Each single branch that crosses the horizontal line is at the centre of a zone. There is a less subjective, statistical way of doing this using what is called a Monte Carlo model which compares the reduction in variance between zones drawn at different points on the tree with the reduction obtained by zoning a randomised dataset. The statistical principle is that the zones should be the smallest statistically significant units of the dendrogram.

Diversity within samples We may wish to have some kind of comparative measure of the range of taxa within our samples. Diversity indices, borrowed from ecology, seek to do this. Diversity within a population (known as ‘alpha’ diversity – Ringrose 1993, 280; Reitz & Wing 1999, 102 – diversity may be described on scales from alpha to epsilon, or regional, diversity) takes into account both the number of taxa present (the ‘variety component’) and the relative frequency of each taxon (the ‘evenness component’) (CruzUribe 1988, 179), and so gives us more information about the population than simply looking at the number of taxa alone (it should be noted that population here refers to a statistical population, which is a biological community of multiple species, rather than a biological population of a single species). Diversity of species within a sample can be expressed in a number of terms. Richness is the number of taxa in a community or region. As the sampling of any community is increased, the probability of adding rare species also increases (Reitz & Wing 1999, 102), therefore diversity measures are particularly sensitive to sample size (Reitz & Wing 1999, 107). Equitability (V′) (sometimes referred to as evenness) is the degree to which species are equally abundant, the idea being that an assemblage in which all taxa are equally abundant is more diverse than one in which some taxa are well-represented and others relatively scarce (Reitz & Wing 1999, 105; Orton 2000, 172). A third possible constituent of diversity is heterogeneity (Orton 2000, 172).

51 Figure 3.1. Biostratigraphic diagram showing absolute values of Mollusca, from samples through a Romano-British ditch fill sequence at Weston College, Weston-super-Mare, UK. The diagram has been divided into four zones using single-linkage stratigraphically constrained Cluster Analysis based on Euclidean similarity. The zones appear to relate to different rates of sedimentation through the sequence. Note that Peringia ulvae, which is super-abundant, is presented at a different scale

3. Numerical approaches to land snail palaeoecology

52

Matt Law

Figure 3.2. A variety of diversity indices applied to samples from the Norse period ploughsoils at Bornais, Western Isles, UK. The Shannon and Brillouin indices can be seen to be close, suggesting that the samples are reasonably complete The basic principle of diversity is that if all species within a population are present in the same frequency, then that population is more diverse than one where they are not; and also if two populations share this property, then the one with the most species is most diverse (Ringrose 1993, 280). Studies of modern molluscan communities have found that greater diversity is associated with ‘complex’ habitats which can accommodate more species, such as woodlands or fens, and that diversity indices can vary according to minor topographic, calcium or pH variation across small distances (Davies 2008, 10). As Thomas (1985, 143–144) demonstrates, diversity indices may be difficult to interpret – for example low diversity in an autochthonous assemblage may be the result of relatively unstable habitats which do not favour molluscs or very stable habitats which are favoured by a small number of competitively superior species. Diversity indices may not be appropriate for comparisons between sites, or even samples at the same site, where there might be taphonomically-induced bias (Ringrose 1993, 283). Our archaeological assemblages are usually samples from an original target population. The number of species in the sample can be denoted as s, and the number of species in the original target population as S. The true value of S is unknown, and the value of s is likely to increase along with n, the number of individuals in a sample, as sample size increases (Ringrose 1993, 279–80). A number of different techniques may be applied. Some of the more commonly seen in palaeo-ecological analyses are outlined below and presented for a set of samples in Figure 3.2.

Shannon index (H’ = information content of the sample) This is also known (incorrectly according to Magurran 1988, 34) as the Shannon-Weaver function. It is calculated by the formula

3. Numerical approaches to land snail palaeoecology

53

H’ = -Ʃpi ln pi where pi is the proportion of individuals found in the ith species (that is to say in any given species). In a sample the true value of pi is unknown, but it is estimated as ni/N (Magurran 1988, 35); N being the total number of individuals, ni being the number of individuals in the ith species. In this index, samples with an even distribution between taxa have a higher diversity than samples with the same number of taxa but with disproportionately high abundance of a few taxa (Reitz & Wing 1999, 105). More taxonomic categories lead to greater diversity values when samples show the same degree of equitability in abundance. Equitability may be calculated by the formula V’=H’/lnS

where H’= Shannon index and S = the number of species in the community (Reitz & Wing 1999, 105). Equitability close to 1.0 shows equal abundance of taxa. Ringrose (1993, 281) comments that in actual fact the Shannon index depends very heavily on the most abundant species, and that low abundance species-rich assemblages may have similar values to others if there are similar proportions of more abundant species. Gordon and Ellis (1985, 155) note that the measure was derived from Information Theory and rests on the assumption that organisms act as ‘channels’ for the free flow of ‘information’. In actual fact, they argue, if information can be equated with food, most organisms will act to prevent its movement. Evans and Williams (1991, 117) note, however, that because the Shannon index takes into account the total number of species and the evenness of individuals amongst the species, it is appropriate for archaeological data as it is applicable to samples rather than whole populations.

Simpson’s Index (D) Simpson’s Index of Concentration is a measure of dominance, that is to say that it is heavily dependent on the abundances of the commonest species rather than providing a measure of species richness (Magurran 1988, 39). It measures the ‘evenness’ of the community from 0 to 1. It is calculated by D=∑(ni(ni-1)/N(N-1))

As D increases, diversity decreases, and so the index is usually expressed as 1–D or 1/D (Magurran 1988, 39). Simpson’s index is not very sensitive to species richness.

Brillouin index The Brillouin index (HB) is appropriate when the randomness of a sample cannot be guaranteed (Magurran 1988, 37), and is calculated by

54

Matt Law

Differences between the Shannon and Brillouin indices arise because the Brillouin index describes only the known collection, whereas the Shannon index estimates the diversity of the unsampled as well as the sampled portion of the community (Magurran 1988. 37). Davies and Grimes (1999, 1061) applied a test of statistical significance, Student’s t-test, to the difference between the mean Shannon and Brillouin indices for molluscan samples from a relic water meadow system, finding a statistically significant difference, with p (the probability that the difference between the two means was due to chance) = 0.0001 for both indices. Walker (2014) also used these two indices to explore the completeness of samples. The Brillouin index will approach the Shannon index for an infinite sample, so that difference between the two indices provides an estimation of the completeness of a sample. Walker found a ‘break point’ difference between the two indices of 0.1, demonstrating that samples with more than 180 shells are more likely to be adequate. Nonetheless, he stresses that smaller samples are still useful guides to habitat types.

Fisher’s alpha Kenward (1978, 21) suggests that the most suitable diversity index for death assemblages is Fisher’s alpha, derived implicitly from the formula S=α lne(1 + N/α)

where S is the number of species, N the number of individuals and α the index of diversity. Fisher’s alpha is relatively independent of sample sizes, and is well suited to assemblages that are subject to many random variables (Kenward 1978, 23). An index of diversity gives only the relationship between number of species and individuals. The way individuals are divided up between species (equitability) is important. More information is included in a rank order curve (Kenward 1978, 26), which is based on absolute numbers, but numbers per unit of weight or percentage of the total assemblage can be used as well (Kenward 1978, 18). The rank order curve presents all the species in an assemblage in order of abundance (Evans & Williams 1991, 115). Using percentages allows a comparison between assemblages of different sizes. The flatter the rank order curve, the richer the population is in species and the more diverse the probable origin (Kenward 1978, 18–19). Rank order curves may also show a clear ‘break of slope’ that may differentiate an autochthonous component from allochthonous background. Even without knowing the ecology of individual species, general observations about the environment they were living in can be made (Evans & Williams 1991, 115).

Variation between samples If we want to analyse the variation between samples, especially samples taken across a site laterally rather than in a vertical sequence, we can use multivariate statistics. These can be used to discover structure or patterns within a dataset, to highlight

3. Numerical approaches to land snail palaeoecology

55

relationships between species or samples, to summarise large datasets, to reduce noise and classify outliers, or classify samples based on their contents (Smith 2014, 182). Ordination techniques are based on the presence/absence of taxa within a sample, and also upon the abundances of species; and help describe relationships between species compositions and environmental factors (Davies 1992, 156). They are thus descriptive techniques, which help inform hypotheses about variations in the data rather than explaining the cause of the variation (Davies 1998, 48). Indirect ordination techniques such as Principal Component Analysis, Correspondence Analysis and Detrended Correspondence Analysis are suited to data from sub-fossil assemblages as they do not require measurement of the environmental factors which may control sample variation (Davies 2008, 48), whereas direct ordination techniques such as Discriminant Analysis and Canonical Correspondence Analysis measure the influence of a known variable (Smith 2014, 182). Although Evans and Williams (1991) and Evans et al. (1992) used Cluster Analysis, Smith (1994) and Moine et al. (2005) Correspondence Analysis, and Davies et al. (1996) Principal Component Analysis, Detrended Correspondence Analysis has been most commonly applied to land snail data. This organises data on the basis of species abundances, and so is readily applicable to archaeological situations. It presents a simplification of the different environmental factors or gradients which act on a species, by plotting each gradient along one axis (Davies 1992, 157). The eigenvalue of each axis quantifies how much of the variation that axis accounts for (Davies 1992, 157), with axis 1 lying as far from all the points as a straight line can, and axis 2 being perpendicular to axis 1, but positioned to account for as much of the remaining variation as possible (Peacock & Gerber 2008, 134). The process can continue over a large number of axes, however in most cases the first two will account for most of the variation (Peacock & Gerber 2008, 134). Taxa will be either positive or negative along each axis, and species that respond similarly to an environmental gradient will be grouped close together, as the factor loading process for each of the axes assigns each species a score which represents a quantification of its preferred ecological optimum along the environmental gradient that axis represents (Dale & Dale 2002, 268). Organising the spread of species on the graph into groups of similar behaviour is a subjective exercise (Davies 1992, 158). The detrending procedure has the advantage over Correspondence Analysis that it corrects for artificial patterns that may occur if samples are taken along an ecological gradient by making the axes independent as well as uncorrelated (Hill & Gauch 1980, 48; Peacock & Gerber 2008, 134), it also corrects the problem in Correspondence Analysis that the distance at the ends of the ordination axis tend to be compressed relative to the middle, meaning that the relative distance between species along the axes in Correspondence Analysis does not have a consistent meaning (Hill & Gouch 1980, 50; Dale & Dale 2002, 270). Before attempting multivariate analysis, we need to ‘clean’ our dataset. Multivariate techniques are often sensitive to low snail counts and rare taxa that are only present in a few samples. Many archaeobotanists routinely eliminate samples with a seed count of less than 30–50, as well as taxa that only occur in less than 5–10% of samples (Smith 2014, 189–90). In my molluscan analyses, I have tended to eliminate samples containing fewer

56

Matt Law

than 50 snails, obviously intrusive or residual shells (usually determined by radically different preservation) or taxa that are not likely to be autochthonous (marine snails in some terrestrial contexts, for example). Working with Correspondence Analysis, Rousseau (1987, 296) introduced coding of results before analysis to limit numerical variation between taxa. This required transforming the data values to abundance classes on a logarithmic scale (base 2), giving 13 classes: 0: absence of species 1:1 to 21 = 1–2 individuals 2: (21+ 1) to 22 = 3–4 individuals 3: (22+ 1) to 23 = 5–8 individuals Up to 12: (211+ 1) to 212 = 2049–4096 individuals.

In each case the number of individuals is replaced by the number of the corresponding class (Rousseau 1987, 296). Figure 3.3 presents the results of Detrended Correspondence Analysis of samples containing more than 50 shells from the Norse site of Bornais, South Uist, Western Isles UK (Law & Thew in prep.) using absolute counts, and Figure 3.4 presents Detrended Correspondence Analysis of the same results coded following Rousseau (1987). One of the key strengths of ordination techniques is that they group the taxa within a sample without making any a priori assumptions about the ecological tolerances of each species at that particular place and time. In addition to exploring the ecology of species within samples, ordination techniques may be used to explore ecological variation between samples from the same site. At Kingsmead Bridge, Wiltshire, Davies (1998) used Detrended Correspondence Analysis to do this with a vertical sequence of samples, which grouped the samples along what was interpreted as a moisture gradient, with a trend from wet conditions up to slightly drier conditions and then back to dampness. Figure 3.5 shows Detrended Correspondence Analysis of samples in the Bornais dataset, while Figure 3.6 shows results for the same dataset following coding into base 2. Schembri and Hunt (2009) used a different ordination technique, Non-Metric Multidimensional Scaling (NMDS) based on a Bray-Curtis similarity matrix, on a series of samples from the Brochtorff Circle at Xagħra, Malta. NMDS differs from other ordination techniques as it does not show how species are responding to particular ecological gradients. It may be preferable to DCA in cases where factors other than ecological gradients (such as biogeography or species extinction) are likely to be significant, for example on islands. Unlike other ordination techniques, NMDS does not calculate many different axes, rather it fits the variation to a small number of userselected axes. It is not an eigenvalue method, so axes 1 and 2 do not account for differing amounts of variation. A matrix of pairwise similarity between samples is constructed using one of a number of statistical difference measures (Euclidean and Bray-Curtis are two common examples), and then plotted. Samples that are close together are thus more similar. Stress values indicate the relationship between the position of samples on the similarity matrix and their position on the ordination. Higher stress values (above 0.20) should be treated with caution. Figure 3.7 shows samples from Bornais ordered

3. Numerical approaches to land snail palaeoecology

57

Figure 3.3. Detrended Correspondence Analysis of snails from stratigraphic block GAD at Norse Mound 2A, Bornais, Western Isles. Axis 1 accounts for most of the variation, and may be inversely related to humidity at ground level. Cochlicella acuta is likely to be affected by its very prominent occurrence in the samples

Figure 3.4. Detrended Correspondence Analysis of the same dataset, coded following Rousseau (1987). The distribution of species along Axis 1 is reasonably similar to that in Fig. 3.3, although the distance along the axis is smaller and it now accounts for less of the variation by NMDS using a Bray-Curtis matrix. NMDS can also be applied to investigate the preservation of snails in sets of samples. Using samples from Quaternary sand dunes on Lanzarote Island in the Canary Islands, Yanes et al. (2008), sought to quantify the taphonomic characteristics of the shell

58

Matt Law

Figure 3.5. Detrended Correspondence Analysis of samples from Mound 2A, Bornais, South Uist. In this case, Axis 1 is inversely related to the dominance of Cochlicella acuta, with samples to the right containing relatively higher numbers of Pupilla muscorum, Vallonia costata and Vallonia cf. excentrica, and those to the left containing high numbers of C. acuta, to the detriment of the former three species

Figure 3.6. Detrended Correspondence Analysis of the same dataset after coding following Rousseau (1987). Again the relationship between Cochlicella and Pupilla-Vallonia is emphasised assemblage and assess the compositional fidelity of the assemblage. They counted the number of adult and juvenile shells (specimens with dimensions twice lower than average values and a more depressed globular shape were judged to be juveniles). The NMDS, this time calculated using the Manhattan distance metric, was based on

3. Numerical approaches to land snail palaeoecology

59

Figure 3.7. Non-Metric Multidimensional Scaling of samples from Mound 2A, Bornais, South Uist, calculated in two dimensions using a Bray-Curtis similarity matrix. Stress = 0.04. The closest samples, 10223 and 7760, have similar numbers of snails and taxa and Pupilla-Vallonia and Cochlicella groups. 7752 has one more snail than 10223, and two fewer species, but a far greater dominance of Cochlicella over the Pupilla-Vallonia group

taphonomic variables such as proportions of juvenile and adult fragments, juveniles and adults with colour preservation, and adults with carbonate coatings. Analysis of Similarities (ANOSIM) was used to evaluate whether non-random differences in preservation could be demonstrated. These revealed that within-bed variation in preservation was generally lower than between-bed variation. The Spearman rank correlation coefficient was used to measure the relationship between shell density in different beds and palaeoecological and taphonomic variables. To test compositional fidelity, all samples were divided into two groups using Cluster Analysis with the Manhattan distance and group-averaging linkage. The less well-preserved group was associated with high proportions of carbonate coatings and low proportions of colour preservation, whereas the better preserved group were associated with moderate colour preservation and low proportions of carbonate preservation. The two groups were then testing using NMDS with a Bray-Curtis matrix and ANOSIM, which showed that there were no consistent differences in species composition between well-preserved and poorly preserved samples. ANOSIM is a variant of Multi-Response Permutation Procedures (MRPP), which test the null hypothesis that there is no difference between two or more groups of data (McCune & Grace 2002, 188). Kuchta (2009) also used MRPP to test whether samples from Kulas Quarry and Hwy-JJ in the Upper Mississippi Valley, USA, represent distinct groups using a presence/ absence matrix rather than absolute counts. MRPP works by comparing the observed, average within-group distances of a dataset (observed delta) with an expected value (expected delta), which is the average distances for all

60

Matt Law

possible partitions of the dataset (Kuchta 2009, 136). The reported p-value represents the proportion of the expected-delta that is less than the observed-delta (Kuchta 2009, 136). The measured distances between groups were significant comparing the alluvial samples of Kulas Quarry with the colluvial silt samples of Hwy-JJ, although a factor in this may have been the presence of rare species at Kulas Quarry. Measured distances were not significant between samples from within Kulas Quarry. Evans et al. (1992, 69) report the use of rank order curves and correlation coefficients (Pearson’s r, which is an expression of the dependence between two variables – in this case the percentage abundance of the two taxa) to explore relationships between pairs of species at Ergolding in Bavaria. This revealed two types of behaviour: mutually exclusive, demonstrated by Zonitoides nitidus and Vallonia pulchella, and negatively correlated, demonstrated by Vallonia costata and Succinea oblonga. This type of relatively simple statistical test has not been widely used in archaeological snail analyses, but may reveal much about environmental differences either laterally across a site or through time in vertical sequences of samples. A number of techniques have been used to investigate change between samples. These methods have the advantage of not relying on uniformitarianist assumptions about the ecological needs of individual species at different places and times, and so go some way towards countering a theoretical problem in environmental archaeology. They do not, however, offer an interpretation of the patterns they reveal, and may be responsive to factors other than those we wish to explore. A degree of data cleaning, such as the coding proposed by Rousseau (1987), may help here.

Variation within one species of snail Studies of modern fauna have shown that many snails respond to small-scale changes in vegetation or surface, and Thomas (1978) notes that such small-scale factors can lead to morphological differences between populations of the same species. Metrical analyses of well-dated assemblages may therefore be desirable. This kind of variation, known as phenotypic plasticity, has been observed in Cepaea spp., whose colouring may be a response to environmental factors (Currey & Cain 1968) and Cepaea nemoralis, whose size and shape (including relative size of the aperture) may be environmentally determined (Thomas 1978). Goodfriend (1986) reviews a number of species worldwide that show phenotypic plasticity which may have archaeological relevance shells of many species (generally larger species, in particular Helicellids) may be larger with higher rainfall, larger on high-calcium soils and smaller with higher population density. Despite pleas from Thomas (1978; 1985, reiterated by Evans et al. 1992), this remains an under-investigated area. Evans (2005) measured adult Pomatias elegans from Ascottunder-Wychwood, Oxfordshire, and compared them to measurements from Brook, Kent, and Blashenwell, Dorset. Prehistoric snails at the three sites were larger than modern snails at Brook and Blashenwell, leading Evans to conclude that P. elegans thrived at Ascott. It should be noted that there is sexual dimorphism within P. elegans, females being larger than males, although at Blashenwell it was (justifiably!) assumed that both were represented in the assemblage (Preece 1980, 354).

3. Numerical approaches to land snail palaeoecology

61

A somewhat different approach to morphometric analysis of snail shells is to establish whether they were collected by humans or are the result of natural processes. Bernáldez-Sanchez et al. (2014) compared measurements of archaeological examples of Theba pisana pisana from the southwest Iberian peninsula to modern examples collected from aestivation (inactivity during summer) sites (fence posts and the branches of trees) and snails collected for consumption by people who habitually collect them for food. Snails for human consumption had a minimum height of 10.64 mm and a minimum width of 14.95 mm. As might be expected, however, context remains important and caution should be exercised – the average size of archaeological shells at Hospital de las Cinco Llagas was below that of the snails for human consumption, but the shells were deliberately buried in a pit and seem likely to have been food waste. Morphological differences in the shells of individual species between sites and over time are certainly worth further investigation, especially as there is a relatively rich literature covering modern snails. Inferring human exploitation from the size of snails should only be done with caution, however, as modern preferences may not apply to societies or individuals in the past.

Snails and climate change Some of the early work on snails used stratigraphical studies to study how climate has changed, leading to the creation of molluscan zonation schemes in southern Britain and some areas of continental Europe (Davies 2008, 43), the ‘zones’ being stratigraphic associations of taxa which change in response to climatic variation. A numerical methodology for climatic reconstruction known as the Mutual Climatic Range method was developed in the late 1980s using Coleoptera, which has subsequently been applied to Mollusca (Moine et al. 2002). The method works by looking at the modern distribution of all the predator and scavenger species in the assemblage and the climatic range for each species calculated using meteorological data. Predators and scavengers are used rather than herbivores because they can spread independently of plants when the climate changes (the idea being that beetles spread faster than plants). The two most important variables governing beetles seem to temperature of the warmest month (TMAX) and temperature range between the coldest and warmest months (TRANGE). By knowing TMAX and TRANGE for each species, its geographic distribution may be plotted on a climate space chart with TMAX on the x-axis and TRANGE on the y-axis. A climatic envelope for each species is thus obtained. For a fossil assemblage, the mutual climatic range can be determined from a computer-generated plot of the climatic parameters of each species in the assemblage, providing the best estimate of the mutual climatic conditions, in terms of TMAX, TRANGE and TMIN (temperature of the coldest month) in which the species within the assemblage could have coexisted (Bradley 1999, 352). Moine et al. (2002, 171) note that parameters such as humidity and other biological factors that might determine distribution remain to be detected, however they found the technique works well for terrestrial molluscs. The present-day geographical range of most of the British molluscan fauna extends to countries that are beyond the range of Holocene temperature variation, however, so the technique will not be applicable to recent shell assemblages.

62

Matt Law

Conclusions The intention of this review has been to attempt to show that all that quantitative stuff is far from worthless. Statistical methods may be brought to bear on problems of quantifying diversity within samples (and, perhaps more promisingly, in doing so investigating the likely completeness of the sample), in exploring variation between samples and the behaviour of the taxa they contain without making a priori assumptions about their ecological preferences, and in using the differences between samples to assess differences in taphonomy. Metrical differences within a single species might be informative, as modern studies have shown a good deal of environmentally-determined phenotypic plasticity. This type of work remains underdeveloped in archaeology, and would certainly warrant further investigation. Finally, the climatic range overlap of species within samples may provide insights into past climatic conditions, although other factors that may have more influence on the distribution of an individual species are undetectable. Numerical methods do not provide answers in themselves, and are certainly no substitute for ecological knowledge, however they have the ability to highlight areas of variation in land snail assemblages that may be worthy of exploration.

Acknowledgements Thanks are due to Mike Allen for inviting me to contribute this chapter. Much of the material presented here has been influenced by conversations with colleagues, especially Ken Thomas, Nigel Thew and Niall Sharples. I am especially grateful to Tom Walker for allowing me to share part of his unpublished doctoral work. Avon Archaeology Ltd gave me permission to use the data from Weston College, which will not be brought to publication.

References Bernáldez-Sanchez, E. & Garcia-Viñas, E. 2014. Deposits of terrestrial snails: natural or anthropogenic processes? In Szabó, K. Dupont, C. Dimitrijević, V. Gómez Gastélum, V. & Serrand, N. (eds), Archaeomalacology: shells in the archaeological record, 235–244. Oxford: British Archaeological Report S2666 Birks, H. J. B. 1986. Numerical zonation, comparison and correlation of Quaternary pollenstratigraphical data. In Bergland, B. E. (ed.), Handbook of Holocene Palaeoecology and Palaeohydrology, 235–244. Chichester: Wiley Birks, H. J. B. 2003. Quantitative palaeoenvironmental reconstruction from Holocene biological data. In Mackay, A. Battarbee, R. Birks, J. & Oldfield, F. (eds), Global Change in the Holocene, 107–123. London: Arnold Box, G. E. P. & Draper, N. R. 1987. Empirical Model Building and Response Surfaces. New York: Wiley Bradley, R. S. 1999. Palaeoclimatology: reconstructing climates of the Quaternary. Second edition. San Diego: Elsevier Cruz-Uribe, K. 1988. The use and meaning of species diversity and richness in archaeological faunas, Journal of Archaeological Science 15, 179–196

3. Numerical approaches to land snail palaeoecology

63

Curry, J. D. & Cain, C. J. 1968. Studies on Cepaea. IV. Climate and selection of banding morphs in Cepaea from the Climatic Optimum to the present day, Philosophical Transactions of the Royal Society of London B253, 483–496 Dale, A. L. & Dale, B. 2002. Application of ecologically based statistical treatments to micropalaeontology. In Haslett, S. K. (ed.), Quaternary Environmental Micropalaeontology, 259–286. London: Arnold Davies, P. 1992. Sub-fossil Mollusca from overbank alluvium and other wet-ground contexts in Wessex. Unpublished PhD thesis, Cardiff University Davies, P. 1998. Numerical analysis of subfossil wet-ground molluscan taxocenes from overbank alluvium at Kingsmead Bridge, Wiltshire, Journal of Archaeological Science 25, 39–52 Davies, P. 2008. Snails: archaeology and landscape change. Oxford: Oxbow Books Davies, P. Gale, C. H. & Lees, M. 1996. Quantitative studies of modern wet-ground molluscan faunas from Bossington, Hampshire, Journal of Biogeography 23, 371–377 Davies, P. & Grimes, C. J. 1999. Small-scale spatial variation of pasture molluscan faunas within a relic watermeadow system at Wylye, Wiltshire, U.K, Journal of Biogeography 26(5), 1057–1063 Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press Evans, J. G. 2004. Land snails as a guide to the environments of wind-blown sand: the case of Lauria cylindracea and Pupilla muscorum. In Gibson, A. & Sheridan, S. (eds), From Sickles to Circles: Britain and Ireland at the time of Stonehenge, 366–379. Stroud: Tempus Evans, J. G. 2005. The snails. In Benson, D. & Whittle, A. (eds), Building Memories: the Neolithic Cotswold long barrow at Ascott-under-Wychwood, Oxfordshire. 55–70. Oxford: Oxbow Books Evans, J. G. & Williams, D. 1991. Land mollusca from the M3 archaeological sites – a review. In Fasham, P. J. and Whinney. R. J. B. (eds), Archaeology and the M3: the watching brief, the AngloSaxon settlement at Abbots Worthy and retrospective sections. Winchester: Hampshire Field Club Monograph 7 Evans, J. G., Davies, P. Mount R. & Williams, D. 1992. Molluscan taxocenes from Holocene overbank alluvium in Southern Central England. In Needham, S. & Macklin, M. G. (eds), Alluvial Archaeology in Britain, 65–74. Oxford: Oxbow Books Giovas, C. M. 2009. The shell game: analytic problems in archaeological mollusc quantification, Journal of Archaeological Science 36(7), 1557–1564 Goodfriend, G. A. 1986. Variation in land snail form and size and its causes: a review. Systematic Zoology 35, 204–223 Gordon, D. & Ellis, C. 1985. Species composition parameters and life tables: their application to detect environmental change in fossil land molluscan assemblages. In Fieller, N. R. J. Gilbertson, D. D. & Ralph, N. G. A. (eds), Palaeoenvironmental Investigations: research design, methods and data Analysis, 153–164. Symposia of the Association for Environmental Archaeology 5A. Oxford: British Archaeological Report S266 Grimm, E. C. 1987. CONISS: a FORTRAN 77 program for stratigraphically constrained cluster analysis by the method of incremental sum of squares, Computers & Geosciences 13(1), 13–35 Hill, M. O. & Gauch Jr, H. G. 1980. Detrended Correspondence Analysis – an improved ordination technique. Vegetatio 42, 47–58 Kenward, H. K. 1978. The Analysis of Archaeological Insect Assemblages: a new approach. York: York Archaeological Trust Kuchta, M. A. 2009. The Paleoenvironmental Significance of Terrestrial Gastropod Fossils from the Upper Mississippi Valley in Minnesota and Wisconsin. Unpublished PhD thesis: University of Wisconsin – Madison Ložek, V. 1986. Mollusca analysis. In Berglund, B. E. (ed.), Handbook of Holocene Palaeoecology and Palaeohydrology, 729–740. Chichester: Wiley

64

Matt Law

McCune, B. & Grace, J. B. 2002. Analysis of Ecological Communities. Gleneden Beach, USA: MjM Software Magurran, A. E. 1988. Ecological Diversity and its Measurement. London: Croom Helm Moine, O. Rousseau, D.-D Jolly, D. & Vianey-Liaud, M. 2002. Paleoclimatic reconstruction using mutual climatic range on terrestrial mollusks, Quaternary Research 57, 162–172 Moine, O. Rousseau, D. & Antoine, P. 2005. Terrestrial molluscan records of Weichselian Lower to Middle Pleniglacial climatic changes from the Nussloch loess series (Rhine Valley, Germany): the impact of local factors, Boreas 34, 363–380 Orton, C. 2000. Sampling in Archaeology. Cambridge Manuals in Archaeology. Cambridge: University Press Peacock, E. & Gerber, J. 2008. Using land snails and freshwater mussels to chart human transformation of the landscape: an example from North Mississippi, USA. In Reitz E. J., Scarry, C. M. & Scudder, S. J. (eds), Case Studies in Environmental Archaeology, 123–142. Second edition. New York: Springer Pielou, E. C. 1975. Ecological Diversity. London: Wiley Preece, R. C. 1980. The biostratigraphy and dating of the tufa deposit at the Mesolithic site at Blashenwell, Dorset, England, Journal of Archaeological Science 7, 345–362 Reitz, E. J. & Wing, E. A. 1999. Zooarchaeology. Cambridge Manuals in Archaeology. Cambridge: University Press Ringrose, T. 1993. Diversity indices and archaeology. In Andresen, J. Madsen, T. & Scollar, I. (eds), Computing the Past: computer applications and quantitative methods in archaeology, 279–285. (CAA 92). Aarhus: University Press Rousseau, D.-D. 1987. Paleoclimatology of the Achenheim series (middle and upper Pleistocene, Alsace, France): a malacological analysis, Palaeogeography, Palaeoclimatology, Palaeoecology 59, 293–314 Schembri, P. & Hunt, C. 2009. Sampling strategy and methodology. In Malone, C. Stoddart, S. Bonanno, A. & Trump, D. (eds), Mortuary Customs in Prehistoric Malta: excavations at the Brochtorff Circle at Xagħra (1987–94), 23–27. Cambridge: McDonald Institute Monographs Smith, H. 1994. Middening in the Outer Hebrides: an ethnoarchaeological investigation. Unpublished PhD thesis: University of Sheffield Smith, A. 2014. The use of multivariate statistics within archaeobotany. In Marston, J. M. d’ Alpoim Guedes, J. & Warinner, C. (eds), Method and Theory in Paleoethnobotany, 181–204. Boulder: University Press of Colorado Thomas, K. D. 1978. Population studies on molluscs in relation to environmental archaeology. In Brothwell, D. Thomas, K. D. & Clutton-Brock, J. (eds), Research Problems in Zooarchaeology, 9–13. London: Institute of Archaeology Thomas, K. D. 1985. Land snail analysis in archaeology: theory and practice. In Fieller, N. R. J. Gilbertson, D. D. & Ralph, N. G. A. (eds), Palaeobiological Investigations: research design, methods and data analysis, 131–156. Oxford: British Archaeological Report S266 Walker, T. 2014. Shifting sand: the palaeoenvironment and archaeology of blown sand in Cornwall. Unpublished PhD thesis, University of Reading White, T. D. & Folkens, P. A. 2005. The Human Bone Manual. Burlington: Academic Press Yanes, Y. Tomašových, A. Kowalewski, M. Castillo, C. Aguirre, J. Alonso, M. R. & Ibáñez, M. 2008. Taphonomy and compositional fidelity of Quaternary fossil assemblages of terrestrial gastropods from carbonate-rich environments of the Canary Islands, Lethaia 41, 235–256

4. Molluscs and the palaeo-environment of coastal blown sand and dunes Thomas Walker

Coastal sand dunes are common in Britain (Fig. 4.1), especially on Atlantic-facing coasts but also around the North Sea coasts of Scotland and northern England (Doody 2008; 2009). In total these cover about 70,000 ha, of which 71% by area are in Scotland (May 2003, 332). The dunes and areas of blown sand are likely to have been visited regularly by humans during the Holocene. Sandy bays permit access routes for sea transport and to food resources both on land and from in-shore waters. Many archaeological sites remain remarkably well preserved within or under blown sand, but others have undoubtedly been lost to coastal erosion or sea level rise. Relatively few of these coastal sites have included mollusc studies in palaeo-environmental analysis; those which have are indicated in Figure 4.1, and are concentrated mainly in the Outer Hebrides of Scotland and the southwest of England and Wales. Dune formation is thought to have commenced during the Late Glacial Maximum about 18,000 years ago (Goode & Taylor 1988, 29), although most Pleistocene dunes are below modern sea level. In some areas of Britain Pleistocene dunes exist onshore, usually overlying a raised beach, as at Brean Down in Somerset (ApSimon et al. 1961; Bell 1990) and at Newquay and Godrevy in Cornwall (James 1994). These glacial dunes are generally only exposed in cliff sections and are not apparent on the land surface. Holocene sand dunes in Britain are often considered to have accumulated no earlier than around 4000 BC (Carter 1988, 303), once sea levels were sufficiently close to present day levels for sand to be driven onshore by wind and wave action (Shennan & Horton 2002, 518). Some reports suggest older dates for blown sand deposition. Gilbertson et al. (1999) found that carbonate sands on the machair of Benbecula in the Outer Hebrides commenced development from c. 7000–6300 BC and on Vatersay from c. 4800 BC, while on the Bay of Skail on the Orkney Islands blown sand is dated to 5040‒4855 cal BC (Leinert et al. 2000). At Sefton on the Lancashire coast the first dune complexes are thought to have developed around 3800 BC (Pye & Neal 1993). Evans (1972, 296) proposed that sand deposition at Northton in Harris, Outer Hebrides, could not be earlier than about 3000 BC, while a somewhat later date of c. 2000 BC has been suggested for Cornwall (Caseldine 1980, 7; Turk 1984, 271; Lewis 1992). On Sanday in the Orkney Islands the first blown sands accumulated around 3000 BC but with

66

Thomas Walker

Figure 4.1. Coastal sand dunes in the British Isles. The sites which include molluscs in palaeoenvironmental analysis are named

4. Molluscs and the palaeo-environment of coastal blown sand and dunes

67

established dunes not developing until 1000 BC (Sommerville et al. 2007). Some British dunes are argued to be of much more recent origin, commencing in medieval or later times (Saye 2003, 42).

Archaeological importance of coastal dune sites Aeolian sands are important from an archaeological perspective, as they preserve occupation horizons in situ, often in remarkable detail. Blown sands, being derived from beach deposits, are normally highly calcareous due the presence of fragmented marine molluscs and calcareous algae, permitting preservation of materials which would be lost on more acidic sites. This especially applies to environmental evidence such as molluscs which can be found in large numbers in many dune and blown sand deposits. The blown sand which buries the sites may accumulate slowly or rapidly but

Figure 4.2. Schematic diagram to illustrate the key contexts in which archaeological and palaeoenvironmental evidence is found in relation to coastal dunes (redrawn from Bell & Brown 2008, 22)

68

Thomas Walker

stratigraphy is often well preserved and less likely to be disturbed by later ploughing or other human activity than on sites further inland. Repeated episodes of sand blow may cover sequential phases of activity or occupation, the latter representing periods of stability. Bell and Brown (2008) have outlined the key contexts in which archaeological and palaeo-environmental evidence is preserved on dunes (Fig. 4.2). In brief, human presence may be represented by the remains of buildings, field walls and trackways, for example, while their activities can be seen by preserved plough marks or animal footprints, charcoal from burning, midden material or artefacts. Major sand blow episodes in different parts of Britain are likely to be broadly coeval, especially at times of rapid climate change, but less severe episodes will more probably impact only in local areas. Comparison of blown sand sites can aid in establishing chronologies of coastal environmental change, but only when there has been a robust programme of context dating in association with material culture and/or palaeoenvironmental evidence.

Non-marine molluscs in dune assemblages Dunes by their very nature are a mobile environment, with movement of sand subject to wind action (Fig. 4.3). Foredunes, those closest to the sea shore, are often bare of vegetation or have marram or couch grass, the former frequently planted to stabilise the dunes. In habitats further from the shore there is greater likelihood of increasing vegetation, which not only improves stability of the dunes but provides shade for species which are less tolerant of open sunlight. Hindshore dunes may contain pasture used for livestock grazing, with arable use less frequent; manuring to enrich fertility may be required, and seaweed is a readily available source. Further inland dune slacks often have rich flora, but only on mature and stable dunes is it possible for woodland or larger scrub vegetation to become established. Of relevance to archaeology is the presence of buried soils and old land surfaces either beneath the dune or interspersed between layers of blown sand. In most dunes the instability of the sand does not permit sufficient time for true soils to develop, and those found within dune sequences are more likely to derive from colluvial movement from nearby higher ground or by alluviation from adjacent rivers. Taphonomic and diagenetic factors must be considered; these may significantly alter the assemblage after death of the shells (Bobrowsky 1984; Preece 2005). While many of these factors, such as preservation of shells (Carter 1990), or downward displacement by earthworms (Darwin 1882; Atkinson 1957), are the same as those from inland sites, there are special considerations in dune assemblages that are not encountered in more stable environments. Sand is highly mobile, and the consequences of wind conflation and deflation need to be considered. Winnowing tends to displace lighter sand grains and smaller shells very readily, leaving heavier shells in situ (Bell 1987, 6) or only transported short distances (Quick 1926), while conflation often leads to allochthonous assemblages which may not represent local environments when the shells were living. Care therefore has to be taken to assess whether the molluscs found in any particular

4. Molluscs and the palaeo-environment of coastal blown sand and dunes

69

Figure 4.3. The structure of coastal dunes

sediment are autochthonous. Animal interference may also disturb assemblages; rabbit burrows and other small animals regularly lead to redistribution of shells long after deposition (Claassen 1998, 79); rabbit warrens were frequently situated on dunes and the resulting bioturbation can alter not only mollusc assemblages but also the archaeology that is being investigated. Molluscs can only thrive in relatively stable environments, and numbers found in any particular horizon can be used to determine stability. Soils which later become buried are likely to be undisturbed for a sufficient time for a wide range of species to become established (provided conditions are not too acidic), often in high numbers, and are likely to be autochthonous. In contrast, if there are few specimens and/or species in a sand which should be suitable for molluscan life, then it is likely that the deposits are unstable, with episodes of conflation and deflation; those shells that are present may be allochthonous. Sand movement can be dramatic. For example, during the winter storms of 2014/15 several beaches in Cornwall were completely stripped of sand. On other occasions large quantities of sand may bury previously stable surfaces to great depth in a very short length of time, perhaps the result of a single storm. Mollusc assemblages associated with dunes consist largely of open country species, with xerophiles, especially Vallonia excentrica, Pupilla muscorum, Vertigo pygmaea, Helicella itala and Cochlicella acuta, predominating. Shade requiring species tend to be poorly represented, but when found in good numbers may indicate a palaeosol, although several species frequently associated with shade, such as Aegopinella nitidula and Clausilia bidentata, are commonly found living in the base of grass tussocks. The method of sample collection is also clearly important. A minimum of 150 shells identifiable at least to genus level is recommended to ensure that the subfossil molluscs are representative of the original living assemblage (Evans 1972, 83; Davies 2008, 5). This will normally be achieved with bulk samples, where between 500 g and 2 kg is usually adequate. Mollusc columns obtained from dunes follow standard procedures (Evans 1972, 41; see also Allen, Chapter 2), although frequently with one difference. It is often not possible to sample from the base of the mollusc column upwards as sand is insufficiently compact; sampling is normally from the top downwards, and great

70

Thomas Walker

care must be taken to ensure that there is no contamination from sediment falling from higher levels into the material being sampled. Some dune systems are very deep or extend below the water table and coring may be the only appropriate method of sample collection. Hand augers typically have a gauge of 2.5–5.0 cm, while percussion coring uses gauges of 5 cm or larger. Especially with smaller gauges the quantity of sediment for each subsample will generally be considerably smaller than with bulk samples, with consequent reduction in the likelihood of obtaining sufficient mollusc numbers to achieve a species spread representative of the local environment. While it may still be possible to obtain an overall impression of the environment from small samples this will be less accurate than with larger assemblages.

Molluscs and archaeology of blown sand sites in the British Isles The majority of reports which discuss the role of molluscs in environmental reconstruction on archaeological sites have been concerned with chalk downlands. Studies examining molluscs in coastal dunes are less frequent and are concentrated on the Scottish Islands (Outer Hebrides and Orkney) and the southwest of Britain (Pembrokeshire, Somerset, Devon and Cornwall). (Fig. 4.1) Prior to 1960 reports of molluscs on blown sand sites associated with archaeology were mostly limited to the north coast of Cornwall, some merely listing the shells present in different horizons, as at Perranporth (Bullen 1902a), Harlyn Bay (Bullen 1902b) and Daymer Bay (Bullen 1909), although others attempted to match assemblages to the palaeo-environment. At Towan Head, Newquay, woodland was present in the lowest Holocene levels, and is one of the few Cornish sites where the shade species Pomatias elegans has been found (Kennard & Warren 1903; Woodward 1908). At Daymer Bay a mollusc sequence studied by Arkell (1943) showed blown sand only, with no evidence of any early woodland. Reliable chronology of these sequences was necessarily limited as they pre-date the advent of scientific dating methods. A recent study investigating a submerged forest in Daymer Bay, with wood dated 2470–2290 cal BC (Howie & Gwynn 2013), showed a molluscan fauna typical of damp woodland, including Pomatias elegans and Discus rotundatus, with absence of open country species apart from Vallonia excentrica. In southwest England radiocarbon dates were first associated with molluscan analysis in 1986 at Bantham, Devon, where a hearth-like feature containing Cochlicella acuta was dated to the post-Roman/early medieval period (Bell 1986). An extensive programme of radiocarbon dating correlated with mollusc analysis was used to provide chronology at Brean Down, Somerset, showing open country with some shade/woodland in the Neolithic, the latter disappearing by the Late Bronze Age, and with the landscape becoming progressively more open up to the 16th/17th centuries (Bell & Johnson 1990; summarised in Davies 2008, 134). Dating of dune excavations became standard in Cornwall during the 1990s although there are few reports of this being associated with molluscan palaeo-environmental data.

4. Molluscs and the palaeo-environment of coastal blown sand and dunes

71

Multiple radiocarbon dates were obtained from the middle Bronze Age settlement at Trethellan Farm, Newquay (Nowakowski 1991), but the mollusc analyses from the site contained Cernuella virgata and Candidula intersecta, both post-Roman introductions to Britain, proving that the sand was not contemporary with the archaeology, and was probably from a later colluvial deposit (Milles 1991a). At Gwithian palaeo-environmental data was not part of extensive excavations in the mid-20th century (Nowakowski et al. 2007), but several later studies have examined the molluscan fauna (Spencer 1975; Milles 1991b; Davies 2007); a single Bronze Age radiocarbon date was obtained during excavations in 1961, and a further 19, together with two OSL dates, in 2005 (Hamilton et al. 2008). The Gwithian Bronze Age and post-Roman environment shows alternating horizons of stabilisation (with plough soils) and rapid sand accumulation. Another Cornish site, the Romano-British settlement at Atlantic Road, Newquay, has incorporated shell studies (Davies 2008) with radiocarbon dating showing a dune succession from bare to well-vegetated open country conditions. At Gunwalloe on the Lizard Peninsula well vegetated dunes providing light shade were present during the Bronze Age but an early medieval site was buried by sand, probably contributing to abandonment of the area around the 12th–13th centuries AD (Walker 2013). Holocene mollusc analysis of dunes in Wales associated with archaeology is limited to Pembrokeshire. A Beaker to Roman period settlement at Stackpole Warren revealed Bronze Age woodland clearance with progression to open country (Evans & Hyde 1990). Dune grassland, often unstable, was present at Brownslade, a possible Bronze Age barrow with an overlying early medieval cemetery (Bell & Brown 2011) and a similar landscape was found at Freshwater East (Walker 2011). All these sites included radiocarbon dating (Stackpole Warren: Benson et al. 1990; Freshwater East: Schlee 2009; Brownslade: Groom et al. 2011), although only at Stackpole and Freshwater East were any directly related to mollusc analyses. Excavations in the western and northern Scottish islands have used mollusc analysis much more widely in the establishment of the palaeo-environmental history, and permit a chronological successional appearance of several mollusc species on sand dunes to be established (Evans 1979): 1. On the old land surface or initial blown sand xerophile species (Vallonia, Helicella, Pupilla) are often absent, but species indicative of woodland (Carychium, Discus) may be present, and can indicate the presence of woodland prior to sand deposition. 2. In some sites open country species, mainly Vallonia and Pupilla, are present in basal contexts and suggest a degree of open country, but still with some tree and/or scrub cover. 3. A zone of open country species but with Pomatias elegans, as found in Cornwall, implies deforestation with disturbed ground; P. elegans is not present on Scottish or Irish archaeological sites and has never been found in these areas (apart from one 20th century record in Galway, Ireland (National Biodiversity Network)). 4. The introduction of open country species in early blown sand, normally with a regular sequence of appearance: a) Vallonia costata, Pupilla muscorum and Vertigo pygmaea, followed later by Vallonia excentrica. b) Helicella itala, followed by Cochlicella acuta. c) ‘Recent’ introduced helicids, such as Cernuella and Candidula.

72

Thomas Walker

The Neolithic to Iron Age settlement at Northton, on the isle of Harris in the Outer Hebrides, first enabled Evans to propose this succession (1971a; 1972), where the Neolithic woodland environment was initially cleared before briefly recurring in the Bronze Age; sand accumulation was rapid in the Iron Age and has become stable in more recent times. Radiocarbon dating was directly linked to molluscan data (Burleigh et al. 1973). The succession was confirmed at the Beaker to Bronze Age settlement at Sligeanach, South Uist (Evans 2004; Evans et al. 2012). Recent well dated studies on several islands in the Outer Hebrides (Hornish Point, South Uist; Baleshare and Newtonferry, North Uist) (Barber 2003) used molluscs to help establish environmental succession in settlements dating from the Early Bronze Age to the medieval period (Thew 2003). On the machair close to the coast open country conditions generally prevailed but with intermittent marsh formation and with numerous blown sand deposits. Evidence of early woodland was lacking. Many middens contained marine molluscs which informed about the diet at the time of settlement and helped to establish that seaweed was used as a fertiliser. Excavations of Mesolithic middens on Oronsay, Inner Hebrides, (Mellars 1987) included analysis of land snails (Paul 1976) which demonstrated several soil horizons with intervening blown sand; early woodland reverted to open dunes after abandonment of the middens. Radiocarbon dating was related directly to mollusc analyses in two middens. The Orkney Islands contain several sites where molluscs aided environmental interpretation. Original woodland with ponding and open marshy areas was present at the Neolithic village of Skara Brae (Spencer 1975) and in the surrounding area of the Bay of Skail (Leinert et al. 2000) as well as at nearby Bronze Age to Viking sites at Buckquoy and Birsay Bay (Evans & Spencer 1976‒1977; Rackham et al. 1989); deforestation during the Bronze Age resulted in open country with significant blown sand accumulations. Mollusc studies at the Knap of Howar, Papa Westray, (Spencer 1975; Vaughan 1976; Evans & Vaughan 1983) showed a woodland landscape during the Neolithic while both there and at Tofts Ness on Sanday (Milles 1991b; 1994; 2007) blown sand prevailed during later periods. In Ireland molluscs were reported from old land surfaces or middens showing some areas to have early woodland, but few reports contain further details, and none is linked to archaeological investigations or with scientific dating (Kennard & Woodward 1917; Vaughan 1976; Stelfox & Welch 1980).

Example sites The palaeo-environment revealed by mollusc analyses at several dune sites are discussed in detail in previous texts (eg, Evans 1972, 291–297; Spencer 1975; Evans 1979; Davies 2008, 130–141) and in other chapters in this publication, and will not be duplicated here. Two new examples from Cornwall which demonstrate the value of mollusc analyses will be examined, one using samples obtained by coring, the other from a conventional mollusc column. Radiocarbon dates were obtained on molluscs

4. Molluscs and the palaeo-environment of coastal blown sand and dunes

73

at both sites; there is no limestone in the area and it is considered probable that the dates are reliable.

Red River valley, Gwithian (grid ref.: SW 58886.42064) Gwithian, on the east side of St Ives Bay on the north coast of Cornwall, is rich in archaeology from the Mesolithic to the late medieval period. Extensive excavations led by Professor Charles Thomas from the late 1940s to the 1960s revealed a Bronze Age settlement and a post-Roman industrial site (summarised in Nowakowski et al. 2007). In 2012 a coring transect was performed to obtain samples for palaeo-environmental analysis, especially in the valley where sediment depth and a high water table precluded excavation. One of the cores, close to the present course of the Red River and about 300 m southwest of the Bronze Age site but only 150 m from the post-Roman site, demonstrates some of the benefits of mollusc analysis. Percussion cores with a 5 cm bore were obtained to the bedrock at 980 cm, with the basal 370 cm being mainly acidic marsh deposits devoid of shells. The upper 610 cm were all predominantly sandy. Molluscs were obtained in almost every sample, often in high numbers despite all the sample weights being below 500 g. In the mollusc diagram (Fig. 4.4) the graphs are all shown with the mollusc numbers normalised to 500 g sample weights. The lowest shelly layers have few specimens, consistent with deposits of unstable blown sand; above this, at 490–515 cm depth, there is a dramatic increase in shell numbers indicating stability of the sediment, which is likely to be a buried soil representing an occupation horizon. A radiocarbon date on Cochlicella acuta showed this deposit to date to the Late Bronze Age: 1090–910 cal BC (2837±28 BP; OxA-28959), a time when there was a major settlement with farming, fishing and craft industries. There is a mixture of molluscs across all habitat types, the dominant species being Lauria cylindracea, accounting for 36% of specimens present; this species was absent in adjacent cores 15 m south and 20 m north of this core. This mollusc requires dry, shady, places and is often associated with stone walls (Evans 1972, 151); is suggested that the core location is close to a previously unknown field wall, perhaps adjacent to the river which at the time was less than 15 m away. Davies (2008, 141) emphasises that Lauria may be associated with middening, but there are no other fauna at this level to support that possibility. The settlement was abandoned towards the end of Late Bronze Age, around 900 cal BC (Sturgess 2007; Nowakowski 2009), probably being overwhelmed by blown sand, and the few molluscs in the core above the Late Bronze Age occupation horizon suggests instability entirely consistent with regular conflation and deflation of sand. The area was next occupied in the post-Roman period, during the 5th–8th centuries AD, and the core shows a deep midden deposit in the sand from 207–435 cm. There are moderate numbers of land shells, mostly open ground species. The evidence of middening is the presence of mussel (Mytilus sp.), limpet (Patella sp.) and dog whelk (Nucella lapillus) fragments at almost every level, with occasional barnacles and fish scales. While middens have been found elsewhere at Gwithian the one at this location is newly described, based on the mollusc analysis.

Figure 4.4. Mollusc diagram for the Gwithian core (absolute numbers of shells). The actual number of shells in each sample is shown, together with the numbers normalised to 500 g samples

74 Thomas Walker

Figure 4.5. Mollusc diagram (relative numbers of shells) for Strap Rocks, Gwithian, Cornwall

4. Molluscs and the palaeo-environment of coastal blown sand and dunes 75

76

Thomas Walker

The midden is covered (172–207 cm) by another mixed assemblage with numerous wet ground species, including Anisus leucostoma, Gyraulus laevis, Succineidae and Vertigo antivertigo, consistent with marshland covering the midden. This is, however, soon overlain by a thick layer of unstable blown sand, extending up to the modern topsoil with stable grass tussocks, rich in molluscs mainly of open country species; the only ‘shade’ molluscs are Nesovitrea hammonis and Vitrea crystallina, both capable of living in the shade found around the base of the tussocks.

Strap Rocks, Gwithian (grid ref.: SW 58022.41629) A line of stones was observed in 2011 eroding from the cliff, south of Godrevy Bay, 1 km southwest of the main Gwithian archaeological site. A mollusc column from the sand above the stones was taken to establish whether the stones constituted a wall, and if so, its date and the palaeo-environment at the time of use and burial. While probably of Bronze Age origin there was considerable Industrial Age and later mining activity in the area which may have accounted for any walls. The stones below 105 cm proved to be laid, providing good evidence that they were part of a wall. The mollusc Xerocrassa geyeri, a species which became extinct in Britain during prehistory (see below) was present in the lowest levels, thereby excluding an Industrial Age feature; a subsequent radiocarbon date of X. geyeri adjacent to the wall (120–130 cm depth) gave a date of 2140–1940 cal BC (3650±29 BP; OxA-28970) – the Early Bronze Age. Molluscs below 110 cm (Fig. 4.5) show an assemblage of species of which over 50% are shade-loving, including Discus rotundatus. Above 110 cm open country species strongly predominate with X. geyeri at 100–105 cm dated to 1640–1500 cal BC (3290±29 BP; OxA-28971). The conclusion is that during the time the wall was in use the environment was largely shaded, probably wood or scrub, which became buried by sand during the Middle Bronze Age. Blown sand then persisted until the present day. This is the first clear evidence of woodland/scrub in the Gwithian area during the Bronze Age. While anthropogenic causes are often ascribed to reduction of woodland cover during the Neolithic/Bronze Age, burial by blown sand may be a major contributing factor, as suggested on the main Gwithian site and at Towan Head, Newquay (Spencer 1975), and also at Northton in the Outer Hebrides (Burleigh et al. 1973). The presence of Xerocrassa geyeri in these lower sediments is interesting, being dated as 1640–1500 cal BC. It was also present in the core described above in the deposits, dated to 1090–910 cal BC, the Late Bronze Age. This mollusc has been found in many early postglacial sites in southern Britain (eg, Burchell & Davis 1957; Kerney 1963; Evans 1971b) where it became extinct shortly after the commencement of the Holocene, being unable to tolerate climate warming. Only at Gwithian has it been found in Bronze Age levels (Spencer 1975, described as Candidula cf. intersecta; Milles 1991b), and this has now been confirmed with radiocarbon dating. There are two possibilities to explain this: either there was a very unusual micro-habitat at Gwithian in which this mollusc could survive for several millennia after it became extinct elsewhere, or it has yet to be found in Bronze Age sediments in other parts of Britain. These two studies show that the sequential appearance of mollusc introductions

4. Molluscs and the palaeo-environment of coastal blown sand and dunes

77

proposed by Evans does not apply in all parts of Britain. At both Cochlicella acuta was found in deeper deposits than Helicella itala. A similar succession was found at Gunwalloe on the south Cornish coast (Walker 2013) and it seems that the sequences previously observed in Scotland may not hold for some areas in southwest Britain.

Summary of advantages and disadvantages of blown sand sequences Mollusc analyses offer great potential in palaeo-environmental reconstruction of coastal blown sand sites. The generally alkaline nature of the deposits leads to excellent preservation of the shells, usually in very good physical condition, often when pollen is poorly preserved. The extraction of specimens is generally easy, although the need to acquire samples from the top of a column downwards makes potential contamination more of a problem than with base-up sampling. In practice samples should to be taken before the sand has been exposed too long and has become dried and unstable. Preparation of the samples is straightforward, as sand is readily sieved and disaggregation is not necessary. If a large proportion of the sand grains are larger than 500 μm then there will be considerable quantities of sieve residue from which to extract shells which may be time consuming. In archaeological contexts the environmental information gained often covers long chronological sequences, although, as with molluscs from silty sediments, there may be loss of shells from leaching. Care does need to be taken to assess disturbances of the stratigraphy. Bioturbation due to worms, rabbits or other animals is a problem encountered on many sites and can account for shells being intrusive, and therefore in erroneous chronological contexts. A further problem largely peculiar to blown sand sites is the instability of the sand, which is readily moved by wind, particularly of finegrained sediments. Deflation of the sand may lead to loss of long sequences of deposited sand resulting in a chronological hiatus; this is further complicated when effects of winnowing leave heavier shells in situ, leading to false proportions of different taxa. Equally, conflation may deposit sand containing shells from allochthonous sediments, again resulting in unreliable assemblages both with regard to taxa and chronology. The range of environments revealed by blown sand deposits tends to be rather limited, as by their very nature dunes are generally open country with little in the way of established shade vegetation. It is only when the dunes become stable and mature that taller vegetation provides sufficient shade for a much wider range of mollusc species. Archaeological evidence of occupation and land use is generally very well preserved within dunes, and under even shallow deposits of blown sand mollusc analyses provide a significant adjunct to our understanding of human activity in and around these sites.

References Allen, M. J. 2017. The geoarchaeology of context: sampling for land snails (on archaeological sites and colluvium). In Allen, M. J. (ed.), Molluscs in Archaeology, 30–47. Oxford: Oxbow Books ApSimon, A. M., Donovan, D. T. & Taylor, H. 1961. The stratigraphy and archaeology of the

78

Thomas Walker

late-glacial and post-glacial deposits at Brean Down, Somerset, Proceedings of the University of Bristol Speleogical Society 9, 36‒137 Arkell, W. J. 1943. The Pleistocene rocks at Trebetherick Point, north Cornwall: their interpretation and correlation, Proceedings of the Geologists’ Association 54, 141‒170 Atkinson, R. J. C. 1957. Worms and weathering, Antiquity 31, 219‒233 Barber, J. (ed.), 2003. Bronze Age Farms and Iron Age Farm Mounds of the Outer Hebrides, Scottish Archaeological Internet Report 3: www.sair.org.uk Bell, M. 1986. Sediment sample from the hearth, 50. In Griffith, F. M., Salvage observations at the Dark Age site at Bantham Ham, Thurlstone, in 1982, Devon Archaeological Society Proceedings 44, 39–58 Bell, M. 1987. Recent molluscan studies in the south west. In Balaam, N., Levitan, B. & Straker, V. (eds), Studies in Palaeoeconomy and Environment in South West England, 1–8. Oxford: British Archaeological Report 181 Bell, M. 1990. Brean Down excavations 1983–1987. London: English Heritage Bell, M. & Brown, A. 2008. Southern Regional Review of Geoarchaeology: windblown deposits. Portsmouth: English Heritage Research Department; report 005/2009 Bell, M. & Brown, A. 2011. Land molluscs from Brownslade. In Groom, P., Schlee, D. Hughes, G., Crane, P., Ludlow, N. & Murphy, K., Two early medieval cemeteries in Pembrokeshire: Brownslade Barrow and West Angle Bay, Archaeologia Cambrensis 160, 163‒167 Bell, M. & Johnson, S. 1990. Non-marine Mollusca. In Bell, M., Brean Down Excavations 1983–1987, 246–250. London: English Heritage Benson, D. G., Evans, J. G., Williams, G. H., Darvill, T. & David, A. 1990. Excavations at Stackpole Warren, Dyfed, Proceedings of the Prehistoric Society 56, 217‒245 Bobrowsky, P. T. 1984. The history and science of gastropods in archaeology, American Antiquity 49, 77‒93 Bullen, R. A. 1902a. Notes on Holocene Mollusca from North Cornwall, Proceedings of the Malacological Society 5, 185‒188 Bullen, R. A. 1902b. Harlyn Bay and the Discoveries of its Prehistoric Remains. London: Swan Sonnenshein Bullen, R. A. 1909. Holocene and recent non-marine Mollusca from the neighbourhood of Perranzabuloe, Proceedings of the Malacological Society 8, 247‒250 Burchell, J. P. T. & Davis, A. G. 1957. The molluscan fauna of some early post-glacial deposits in north Lincolnshire and Kent, Journal of Conchology 24, 164‒170 Burleigh, R., Evans, J. G. & Simpson, D. D. A. 1973. Radiocarbon dates for Northton, Outer Hebrides, Antiquity 47, 61‒64 Carter, R. W. G. 1988. Coastal Environments. London: Academic Press Carter, S. P. 1990. The stratification and taphonomy of shells in calcareous soils: implications for land snail analysis in archaeology, Journal of Archaeological Science 17, 495‒507 Caseldine, C. J. 1980. Environmental change in Cornwall during the last 13000 years, Cornish Archaeology 19, 3‒16 Claassen, C. 1998. Shells. Cambridge: University Press Darwin, C. 1882. The Formation of Vegetable Mould through the Action of Worms. London: Murray Davies, P. 2007. Gwithian 2005 – Land snail assessment. In Nowakowski, J. A. (ed.), Excavations of a Bronze Age Landscape and a Post-Roman Industrial Settlement 1953–1961, Gwithian, Cornwall. Assessments of Individual Key Datasets 2003–2006; vol. 1. Truro: Historic Environment Service, Cornwall County Council; report 2007R017, 64‒68 Davies, P. 2008. Snails: archaeology and landscape change. Oxford: Oxbow Books Doody, P. 2008. Distribution of sand dunes in Great Britain. Online: www.marbef.org/wiki/ distribution_of_sand_dunes_in_great_britain.

4. Molluscs and the palaeo-environment of coastal blown sand and dunes

79

Doody, P. 2009. Sand dune – Country report, Ireland. Online: www.coastalwiki.org/wiki/sand_ dune_-_country_report,_ireland Evans, J. G. 1971a. Habitat change on the calcareous soils of Britain: the impact of Neolithic man. In Simpson. D. D. A. (ed.), Economy and Settlement in Neolithic and Early Bronze Age Britain and Europe, 27–73. Leicester: University Press Evans, J. G. 1971b. Durrington Walls: the pre-henge environment. In Wainwright, G. J. & Longworth, I. H. (eds), Durrington Walls: excavations 1966–1968, 329–336. London: Report of the Research Committee of the Society of Antiquaries of London 29 Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press Evans, J. G. 1979. The palaeo-environment of coastal blown-sand deposits in western and northern Britain, Scottish Archaeological Forum 9, 16‒26 Evans, J. G. 2004. Land snails as a guide to the environments of wind-blown sand: the case of Lauria cylindracea and Pupilla muscorum. In Gibson, A. & Sheridan, A. (eds), From Sickles to Circles. Britain and Ireland at the time of Stonehenge, 366–379. Stroud: Tempus Evans, J. G. & Hyde, L. M. 1990. Land Mollusca, 229‒234. In Benson, D. G., Evans, J. G, Williams, G. H., Darvill, T. & David, A. Excavations at Stackpole Warren, Dyfed, Proceedings of the Prehistoric Society 56, 179–245 Evans, J. G., Law, M. & Thew, N. 2012. Stability and flux in the dune environment. In Parker Pearson, M. (ed.), From Machair to Mountains: archaeological survey and excavation in South Uist, 250–253. Oxford: Oxbow Books Evans, J. G. & Spencer, P. 1976‒7. The Mollusca and environment, Buckquoy, Orkney. In Ritchie, A., Excavation of Pictish and Viking-age farmsteads at Buckuoy, Orkney, Proceedings of the Society of Antiquaries of Scotland 108, 215‒219 Evans, J. G. & Vaughan, M. 1983. The molluscs from Knap of Howar, Orkney. In Ritchie, A., Excavation of a Neolithic farmstead at Knap of Howar, Papa Westray, Orkney, Proceedings of the Society of Antiquaries of Scotland 113, 106‒114 Gilbertson, D. D., Schwenninger, J.-L., Kemp, R. A. & Rhodes, E. J. 1999. Sand-drift and soil formation along an exposed north Atlantic coastline: 14,000 years of diverse geomorphological, climatic and human impacts, Journal of Archaeological Science 26, 439‒469 Goode, A. J. J. & Taylor, R. T. 1988. Geology of the Country around Penzance. Memoir for 1:50,000 geological sheets 351 and 358 (England and Wales). London: HMSO Groom, P., Schlee, D., Hughes, G., Crane, P., Ludlow, N. & Murphy, K. 2011. Two early medieval cemeteries in Pembrokeshire: Brownslade Barrow and West Angle Bay, Archaeologia Cambrensis 160, 133‒203 Hamilton, D., Marshall, P., Roberts, H. M., Bronk Ramsey, C. & Cook, G. 2008. Gwithian: scientific dating, Appendix I. In Nowakowski, J. A., Quinell, H., Sturgess, J., Thomas C. & Thorpe, C., Return to Gwithian: shifting the sands of time, Cornish Archaeology 46, 61‒70 Howie, F. M. P. & Gwynn, A. 2013. The sub-fossil assemblage from a Holocene calcareous palaeosol in Daymer Bay, north Cornwall, Geoscience in South-West England 13, 191‒201 James, H. C. L. 1994. Late Quaternary Coastal Landforms and Associated Sediments of West Cornwall. Unpublished PhD thesis, University of Reading Kennard, A. S. & Warren, S. H. 1903. The blown sands and associated deposits of Towan Head, near Newquay, Cornwall, Geological Magazine ser. IV, 10, 19‒25 Kennard, A. S . & Woodward, B. B. 1917. The post-Pliocene non-marine Mollusca of Ireland, Proceedings of the Geologists’ Association 28, 109‒290 Kerney, M. P. 1963. Late-glacial deposits on the chalk of South-east England, Philosophical Transactions of the Royal Society of London B246, 203‒254 Leinert, A. C. de la V., Keen, D. H., Jones, R. L., Wells, J. M. & Smith, D. E. 2000. Mid-Holocene environmental changes in the Bay of Skaill, Mainland, Orkney, Scotland: an integrated

80

Thomas Walker

geomorphological, sedimentological and stratigraphical study, Journal of Quaternary Science 15, 509‒528 Lewis, D. 1992. The sands of time: Cornwall’s Hayle to Gwithian Towans. In Carter, R. W. G., Curtis, T. G. F. & Sheehy-Skeffington, M. J. (eds), Coastal Dunes; geomorphology, ecology and management for conservation 463–473. Rotterdam: Balkema May, V. J. 2003. Sandy beaches and dunes. In May, V. J. & Hansom, J. D. (eds), Coastal Geomorphology of Great Britain. Peterborough: Joint Nature Conservation Committee; Geological Conservation Review Series 28, 327‒413 Mellars, P. 1987. Excavations on Oronsay. Edinburgh: University Press Milles, A. 1991a. Molluscan analysis. In Nowakowski, J. A., Trethellan Farm, Newquay: the excavation of a lowland Bronze Age settlement and Iron Age cemetery, Cornish Archaeology 30, 160 Milles, A. 1991b. The Molluscan Biostratigraphy and Archaeology of Holocene Blown-sand in the British Isles. Unpublished PhD thesis, University of Wales at Cardiff Milles, A. 1994. Taphonomy of Mollusca from Tofts Ness, Sanday, Orkney. In Luff, R. M. & Rowley-Conwy, P. (eds), Whither Environmental Archaeology, 112–125. Oxford: Oxbow Books Milles, A. 2007. Landsnails. In Dockrill, S. J. (ed.), Investigation in Sanday, Orkney. Vol 2: Tofts Ness, Sanday. An Island Landscape Through Three Thousand Years of Prehistory, 228–239. Kirkwall: Orcadian National Biodiversity Network Online: www.nbn.org.uk Nowakowski, J. A. 1991. Trethellan Farm, Newquay: the excavation of a lowland Bronze Age settlement and Iron Age cemetery, Cornish Archaeology 30, 5‒242 Nowakowski, J. A. 2009. Living in the sands – Bronze Age Gwithian, Cornwall, revisited. In Allen, M. J., Sharples, N. & O’Connor, T. P. (eds), Land and People; papers in memory of John G. Evans, 115–125. Prehistoric Society Research Papers 2. Oxford: Oxbow Books/Prehistoric Society Nowakowski, J. A., Quinnell, H., Sturgess, J., Thomas, A. C. & Thorpe, C. 2007. Return to Gwithian: shifting the sands of time, Cornish Archaeology 46, 13‒76 Paul, C. R. C. 1976. The non-marine Mollusca of Colonsay and Oronsay, Journal of Conchology 29, 107‒110 Preece, R. C. 2005. Non-marine Mollusca and archaeology. In Brothwell, D. R. & Pollard, A. M. (eds), Handbook of Archaeological Sciences, 135–145. Chichester: Wiley Pye, K. & Neal, A. 1993. Late Holocene dune formation on the Sefton coast, northwest England. In Pye, K. (ed.), The Dynamics and Environmental Context of Aeolian Sedimentary Systems, 201–217. London: Geological Society, Special Publication 72 Quick, H. E. 1926. Shell pockets of Oxwich dunes, Journal of Conchology 18, 57–60 Rackham, D. J., Spencer, P. J. & Cavanagh, L. M. 1989. Environmental survey. In Morris, C. D. (ed.), The Birsay Bay Project, vol. 1, 44–53 + microfiche. Durham: University of Durham Saye, S. E. 2003. Morphology and Sedimentology of Coastal Sand Dune Systems in England and Wales. Unpublished PhD thesis, University of London Schlee, D. 2009. The Pembrokeshire Cemeteries Project. Excavations at Porthclew Chapel, Freshwater East, Pembrokeshire, 2008. Interim report. Llandeilo: Dyfed Archaeological Trust. Shennan, I. & Horton, B. 2002. Holocene land- and sea-level changes in Great Britain, Journal of Quaternary Science 17, 511‒526 Sommerville, A. A., Hansom, J. D., Housley, R. A. & Sanderson, D. C. W. 2007. Optically stimulated luminescence (OSL) dating of coastal aeolian sand accumulation in Sanday, Orkney Islands, Scotland, The Holocene 17, 627‒637 Spencer, P. J. 1975. Habitat change in coastal sand-dune areas: the molluscan evidence. In Evans, J. G., Limbrey, S. & Cleere, H. (eds), The Effect of Man on the Landscape: the Highland zone, 96–103. London: Council for British Archaeology Research Report 11

4. Molluscs and the palaeo-environment of coastal blown sand and dunes

81

Stelfox, A. W. & Welch, R. J. 1980. A history of the land and freshwater Mollusca of Ulster, Proceedings of the Royal Irish Academy 80B, 125‒152 Sturgess, J. 2007. Bronze Age Gwithian. In Nowakowski, J. A., Quinnell, H., Sturgess, J., Thomas C. & Thorpe, C., Return to Gwithian: shifting the sands of time, Cornish Archaeology 44, 23‒33 Thew, N. 2003. The molluscan assemblage. In Barber J. (ed.), Bronze Age farms and Iron Age farm mounds of the Outer Hebrides 121–123, 163–177. Scottish Archaeological Internet Report 3: www. sair.org.uk Turk, S. M. 1984. Non-marine conchology of Cornwall and the Isles of Scilly, Journal of Conchology 31, 263‒280 Vaughan, M. P. 1976. Environmental Change in Areas of Blown Sand on the Western Coasts of the British Isles. Unpublished BSc dissertation, University of Wales at Cardiff. Walker, T. M. 2011. Non-marine molluscs [from Porthclew Chapel, Freshwater East]. Unpublished report for Dyfed Archaeological Trust Walker, T. M. 2013. Mollusc analysis. In Wood, I. (ed.), Gunwalloe through the Ages: middle Bronze Age to the 12th century AD, Lizard Peninsula, Cornwall, 71–76. London: National Trust Woodward, B. B. 1908. Notes on the drift and underlying deposits at Newquay, Cornwall, Geological Magazine ser. V, 5, 10–18, 80‒87

5. Molluscs from dune-machair systems in the Western Isles: archaeological site formation processes and environmental change Matt Law and Nigel Thew

The term machair refers to low plains of calcareous wind-blown shell sand, usually associated with active dune systems, which flank the north and west coasts of Scotland and the western coast of Ireland. In a landscape otherwise dominated by heavy, acidic, peaty ground, the machair provides relatively fertile, free-draining soils. This explains why machair areas played such an important role in agricultural systems and were the focus for settlement activity for both prehistoric and later societies (Figs 5.1 and 5.2). Land snails are abundant on the machair and shells preserve exceptionally well within the highly calcareous sands. The analysis of molluscs from archaeological contexts in the machair began in the 1960s during doctoral research by John G. Evans (1972), and continued during the 1970s accompanied by two students, Penny Spencer and Michael Vaughan. This research largely concentrated on the Western Isles (Na h-Eileanan Siar) and the Orkneys. In the 1980s, four more sites from the Western Isles were investigated by Nigel Thew (2003). Themes emerging from this work were explored by the present authors in a 2010 conference paper, published as Law and Thew 2015. Since 2010 we have been working on material from more recent excavations in the Western Isles. This chapter expands on Law and Thew 2015 and summarises all the work done up till now in these islands. We discuss the taphonomic processes that have affected machair assemblages, evidence for environmental and climatic change revealed by molluscan faunas, and what assemblages and indicator species show us about site formation processes, activity areas and past land-use practices associated with machair sites. We also detail how biostratigraphic changes caused by the arrival and spread of new snail species can be used as a relative dating tool. The sites discussed in the text are shown in Figure 5.1. The dune-machair systems tend to be unstable as they are prone to wind erosion during winter storms (Smith 1994, 18). Sand deposition can at times be particularly rapid, thus burying and preserving archaeological sites and associated land surfaces. At Hornish Point and Baleshare, for example, deposits up to 2–3 m thick were deposited within time intervals too short to be resolved by radiocarbon dating (Fig. 5.3; Barber 2011, 50). By contrast, strong winds and wave action can rapidly expose and erode

5. Molluscs from dune-machair systems in the Western Isles

83

Figure 5.1. Map showing the location of sites in the Western Isles mentioned in the text (contains Ordnance Survey data (based on OS Open Data; Crown © and database right 2017)

84

Matt Law and Nigel Thew

Figure 5.2. Hay bundles in late summer on the machair at Bornais, South Uist (photo: ML) buried deposits, with, for example, some 20 m of coastline being removed by a storm at Baleshare in 2005 (Chapman et al. 2009, 2308). In the Western Isles the machair is mainly found on the west side of the islands, although small patches also occur on the east coast, like at Rosinish (Fig. 5.1). The sands are largely formed from the comminuted shells of marine molluscs, foraminifera and crustaceans, together with quartz grains and other material derived from glacial deposits. They are believed to have originated as an offshore sand bank that was subsequently blown inland due to sea-level change, while subsequent sand movement and accumulation seems to have resulted from instability caused by human activities and/or during periods of cooler, more stormy climate (Edwards et al. 2005, 435; Dawson et al. 2011).

5. Molluscs from dune-machair systems in the Western Isles

85

The Western Isles The Western Isles are situated between 56 and 58° N, some 50–80 km from the northwest coast of mainland Scotland (Fig. 5.1; Smith 1994, 14; Hall 1996, 5). The archipelago consists of 119 named islands, of which only 14 are now permanently inhabited (Boyd & Boyd 1996, 8), stretching 213 km from The Butt of Lewis to Barra Head. The Hebridean climate is cool, wet and windy, although the North Atlantic Drift gives the islands relatively mild winters (Manley 1979, 48). The islands straddle mid-latitude atmospheric circulation patterns and the North Atlantic storm track (Dawson et al. 2004, 282; Dawson et al. 2011, 31). Geologically, the islands are largely formed from an eroded platform of Precambrian Lewisian gneiss (Smith 1994, 15; Hall 1996, 5), which encourages the formation of acidic podzols. Younger rocks include the Permo-Triassic sandstones and conglomerates of the Stornoway Basin. Soil types within the islands broadly belong within three main zones: the calcareous machair of the western coasts, acidic peaty lowlands known as the ‘blacklands’, and acidic moorland on higher ground (Smith & Mulville 2004, 49).

Archaeological background The only Neolithic sites known from the machair belt were found at Northton, Harris and at Udal, North Uist (Fig. 5.1). In both cases the original settlements were subsequently buried by windblown sand (Sharples et al. 2004, 33), which at Northton overlies both the Neolithic I and Neolithic II levels (Evans 1971). The Beaker period (2400–1700 BC) saw a significant expansion of settlement on the machair (Parker Pearson et al. 2011, 67), and many of these sites continued to be occupied during the Early Bronze Age. During the Middle Bronze Age settlement activity may rather have centred on the blacklands as the machair seems to have been largely abandoned, although the earliest excavated level at Baleshare on North Uist may date from this time (Sharples et al. 2004, 35). Major settlement on the machair resumed during the Late Bronze Age, with three substantial sites being established at Machair Mheadhanach, Sligeanach and Cladh Hallan on South Uist around the start of the 1st millennium BC. Many of these sites continued to be occupied during the Early Iron Age (Sharples et al. 2004, 37), while many new sites appeared around 200 BC (Parker Pearson et al. 2011). Norse settlements are fairly evenly distributed across the machair, although they are often located on top of Iron Age settlement mounds, such as at Bornais, South Uist (Sharples et al. 2004, 41). The majority of Norse period sites continued to be occupied during the Medieval period, although subsequently much of the machair appears to have been abandoned during the 14th–15th centuries. This abandonment has been attributed to a combination of climatic deterioration during the Little Ice Age and to a new political system after the islands came under the control of the Scottish crown, leading to restrictions being placed on traditional connections with Ireland, Western Britain and Scandinavia (Sharples et al. 2004, 43).

86

Matt Law and Nigel Thew

Figure 5.3. A section through Iron Age levels covered by wind-blown sand at Baleshare, North Uist (© Historic Scotland, licensor www.rcahms.gov.uk)

Taphonomic processes affecting molluscan assemblages The unstable nature of the machair engenders rather specific archaeological site formation processes. Extensive in situ deposits can often be well preserved beneath windblown sands, but, conversely, layer sequences are frequently interrupted by sedimentary hiatuses that represent varying lengths of time, while horizontal layers are often cut by erosional features. Wind erosion leads to deflation, while the removal of sand causes heavier archaeological material to become concentrated above erosion surfaces as secondary lag deposits (Butzer 1982, 53; Goldberg & Macphail 2006, 122). Layers that incorporate material from several different periods are known as conflation deposits (Barber 2011, 47). Buried land surfaces are often more organic and thus represent periods of greater surface stability. Associated mollusc faunas can thus be considered to be largely autochthonous, although some allochthonous shells may have been introduced by processes such as manuring or flooding. By contrast, rapidly accumulating deposits of windblown sand are likely to include assemblages that have been subject to a high degree of spatial and temporal mixing. Most of the numerical variation between molluscan assemblages from dune-machair systems can be considered to reflect differences in the size of the original populations

5. Molluscs from dune-machair systems in the Western Isles

87

relative to the rate of layer accumulation, although redeposited shells can obviously somewhat blur this trend. In general, greater stability and a denser vegetation cover will lead to more diverse and often more abundant faunas, although molluscan biodiversity is normally fairly low. By contrast, windblown sands are typically associated with rather poor faunas and just a few taxa. Interpretation of land snail assemblages therefore depends upon detecting patterns of variation among faunas of rather restricted diversity that reflect past environments with varying degrees of stability, vegetation cover and dampness, as well as different rates of sand accumulation (Thew 2003, 163). An obvious problem with the interpretation of the molluscan faunas is their relationship to the contexts in which they are found. The basal parts of windblown sand deposits, for example, often include shells indicative of more stable conditions that have been eroded from underlying layers, while the summit of these deposits may include shells brought down from overlying soil horizons by earthworms. Layers interpreted as ploughsoils may include shells indicative of stable, shaded conditions, because ploughing was not very intensive before the post-medieval period and thus allowed many plants and their associated molluscs to persist between episodes of ploughing, while permitting pre-existing shells to survive reasonably intact. Finally, the application of fertiliser that included seaweed and domestic waste to the relatively nutrient poor soils of the machair, introduced allochthonous elements such as small marine shells and molluscs that were feeding on the midden material. Nevertheless, despite these taphonomic problems, most published sequences yield a fairly coherent palaeoenvironmental history (Davies 2008, 131). Moreover, it is possible to make anomalies easier to detect by taking several spatially separated sample columns and series of spot samples across a site (Thew 2003).

Environmental change through time The earliest analysed machair deposits are from Northton (Fig. 5.4; Evans 1971), which begins with a decalcified buried soil (Neolithic I) over boulder clay, succeeded by a sterile layer of decalcified windblown sand, before the snail sequence began within a deposit of clean, calcareous windblown sand. The presence of Aegopinella nitidula, Carychium tridentatum and Vertigo pusilla (Fig. 5.5) points to a shaded habitat that Evans interpreted as being indicative of open woodland, although the significant proportion of open ground species such as Pupilla muscorum, Vallonia costata and Vertigo pygmaea (Fig. 5.7) and the nature of the sediments rather suggest unstable conditions of accumulating windblown sand and low herbaceous vegetation. The shells from shade-loving taxa must therefore have been blown in from nearby patches of open woodland/scrub with an understory of tall herbs. Subsequently, the sequence indicates variable conditions, with episodes of greater stability associated with short-turf grassland and some tall herbs (Beaker I level), or a fairly damp mix of tall herbs and lower herbaceous vegetation (Neolithic II level), or even fairly moist conditions with tall herbs and shorter herbaceous vegetation plus some open scrub/woodland nearby (Beaker II/EBA + the overlying Iron Age II level + the succeeding layer). The sequence is interrupted in several places by layers of clean windblown sand associated with significantly poorer faunas. By

88

Matt Law and Nigel Thew

Figure 5.4. Percentage diagram showing the principal snail groups from the sequence at Northton, Harris (redrawn from Evans 1971) contrast, the final deposit of windblown sand that underlies the Iron Age II level has an extremely rich fauna that includes a mix of elements typical of rather moist deciduous open woodland with an understory of tall herbs, and of stable short-turf grassland, so most of these shells must have been redeposited from biotopes nearby. The important points concerning the Northton sequence are that on the one hand there was moist open scrub/woodland near the site from the Neolithic period until sometime during the Iron Age, while on the other hand woodland clearance and the deposition of windblown sand began during the Neolithic period, with final clearance during the Iron Age. The molluscan sequence also has several episodes with much damper conditions, shown by the presence of marsh snails (Carychium minimum, Galba truncatula, Oxyloma pfeifferi, Vertigo angustior, Zonitoides nitidus; Fig. 5.8), which may well have been linked with episodes of flooding (see climate section below). Beaker period and Early Bronze Age occupation levels were sampled at Rosinish (Vaughan 1976), Sligeanach (Fig. 5.6; Evans et al. 2012) and Ensay (Spencer 1974). The first two sites were associated with plough-mark horizons. All three have fairly organic sandy layers, with or without occupation material, which include faunas typical of either stable short turf grassland or of more shaded conditions with a mix of grasses and tall herbs (presence of Carychium tridentatum, Euconulus fulvus, Punctum pygmaeum and Vitrea contracta, Fig. 5.5), perhaps suggestive of land being allowed to lie fallow. These layers were interrupted by deposits of clean windblown sand. At Ensay the

5. Molluscs from dune-machair systems in the Western Isles

89

Figure 5.5. Shade-demanding species from machair deposits in the Western Isles. Clockwise from top left: Carychium tridentatum, Lauria cylindracea, Vertigo pusilla, Punctum pygmaeum, Oxychilus alliarius (juvenile), and Nesovitrea hammonis (photos: ML)

lowest Beaker horizon has several taxa which indicate that open scrub woodland with an understory of tall herbs must have been present in the vicinity, while subsequent levels continue to point to the nearby presence of a fairly shaded biotope until at least the Early Bronze Age. The Ensay sequence also includes marsh snails indicative of flooding. It is therefore possible that open scrub woodland persisted for longer in damper areas near sites like Northton and Ensay, but disappeared from drier locations like Rosinish and Sligeanach. During the Late Bronze Age a wave of new settlement sites includes those at Baleshare (Thew 2003) and Cladh Hallan (Law & Thew in prep.). Both are associated with ploughing horizons and have faunas typical of open grassy conditions. The snails from Baleshare are typical of stable short turf grassland, with the additional presence of shade-loving taxa indicative of tall herbs during periods when surfaces were allowed to lie fallow, including the ploughed area. Similar, generally open environments are associated with the succeeding Iron Age levels at Baleshare as well as with Iron Age layers at Cladh Hallan, Hornish Point, Balelone, Cnip and Horgabost (Thew 2003; Cerón-Carrasco 2006). The faunas are typical of fairly stable short-turf grassland, which sometimes gave way to more shaded conditions with tall herbs linked with episodes of fallowing or in some cases the abandonment of parts or the totality of the site. The

90

Matt Law and Nigel Thew

fairly organic layers linked with periods of surface stability are interrupted by clean windblown sands with much poorer molluscan faunas. A new phenomenon in Late Bronze Age and Iron Age levels is the spreading of both fresh midden material and seaweed (Law & Thew 2015, 130; see ‘Site formation processes and land-use practices’ below) to stabilise land surfaces and provide fertiliser for light ploughing. Indeed, from the Late Bronze Age onwards, layers in machair sites appear to have accumulated much more rapidly as a result of both middening and windblown sand. During the Late Bronze Age and the earlier Iron Age the only site with true woodland mollusc faunas is Northton. Nevertheless, the presence of isolated specimens of shade-requiring species in Iron Age levels at Horgabost, Baleshare, Cladh Hallan and Rosinish implies that patches of woodland may well have persisted in the vicinity of these sites into the Iron Age. These snails must have been living in suitable biotopes nearby in order to take advantage of the more shaded conditions afforded by tall herbs during periods when patches of land were allowed to lie fallow. All of these sites except Horgabost have low numbers of marsh snails indicative of flooding, again suggesting that small patches of woodland may have survived for longer in machair locations with damper conditions. Even so, the arrival of largely stone-built structures in machair settlements during the Iron Age may well have been a response to the loss of most native woodland, in agreement with the disappearance of woodland snails from the Northton sequence. The molluscan faunas from Norse and Medieval levels at Rosinish, Bornais (Law & Thew unpubl.), Udal (Spencer 1974) and Newtonferry (Thew 2003) show generally open conditions, associated with periods of greater or lesser surface stability. The only records of shade-requiring molluscs from these sites are single specimens from Bornais and Udal, indicating that scrub woodland had largely disappeared from the machair and adjoining areas. Nevertheless, the regular appearance of suites of shade-loving species such as Carychium tridentatum, Euconulus fulvus, Nesovitrea hammonis, Punctum pygmaeum, Vitrea contracta and Vitrina pellucida show that tall herbs continued to colonise surfaces around settlements, presumably during periods of fallowing or abandonment. The practice of depositing fresh midden material and seaweed to stabilise fragile land surfaces and fertilise nutrient-poor calcareous soils continued until the early 20th century (Smith 2012, 390–1).

Site formation processes and land-use practices The habitation area: structures Although Clausilia bidentata can occasionally be found in tall, fairly damp, stable grassland, or among grass-covered rocks in Scotland (Paul 1976; 1992), in the machair sites of the Western Isles, which tend to be fairly dry and prone to sand accumulation, C. bidentata was probably rupestral and its presence reflects the former existence of walled structures or possibly bushes/trees. Peaks in Lauria cylindracea, which often lives in rupestral habitats (Paul 1976; 1992), may also be linked with former stone-

5. Molluscs from dune-machair systems in the Western Isles

91

built structures, like the Beaker levels associated with large boulders from collapsed buildings at Ensay (Evans 2004, 368), and in some Iron Age contexts at Baleshare and Hornish Point (Thew 2003, 170). The presence of Oxychilus alliarius or O. cellarius within walled structures (cf. Howmore in Smith 1994, 135–6) may suggest that they previously contained concentrations of organic material (see below) and have been used as a domestic dwelling, a byre, or a stable.

The infield area: grazing, ploughing and fallow land The molluscan assemblages from machair sites show that in the area surrounding habitation structures, land surfaces were often used for grazing or ploughing, or sometimes allowed to lie fallow. Grazing encourages snail taxa that prefer open ground with low herbaceous vegetation. Sheep graze closer than cows and do not tear the grass up by the roots, encouraging stability. In stable, fairly moist, short-turf grassland with patches of tall herbs, peaks in Vallonia excentrica are often accompanied by fairly abundant Pupilla muscorum and Vallonia costata, as well as smaller numbers of species like Euconulus fulvus, Nesovitrea hammonis, Oxychilus alliarius, Punctum pygmaeum, Vertigo pygmaea and Vitrea contracta (Figs 5.5 and 5.7). Lauria cylindracea may also be fairly common (Evans 2004). Assemblages of this type have been found at Beaker period to Early Bronze Age sites like Northton, Ensay and Sligeanach and less commonly at Late Bronze Age to Iron Age sites like Baleshare and Hornish Point.

Figure 5.6. Absolute counts of snail species from Column 9076, Sligeanach (reproduced from Evans et al. 2012)

92

Matt Law and Nigel Thew

Ard marks have been documented at a number of sites, including Beaker period levels at Sligeanach (Fig. 5.6) and Rosinish, Late Bronze Age layers at Baleshare and Norse levels at Bornais. The marks are fairly widely spaced and shallow, showing that ploughing was not intensive. Even today ploughing on the machair is rather shallow so as not to damage the root mat and thus encourage erosion (Smith 2012, 392). As buried ploughsoils are often rather thick, homogeneous and organic they must have formed over a significant period. The snail fauna from the ploughsoil at Sligeanach includes a significant component of shade-loving species, which suggests rather stable, fairly moist and dense grassland with the presence of tall herbs. By contrast, assemblages from the ploughsoils at Baleshare, Hornish Point and Bornais are dominated by taxa typical of short-turf grassland, with only small numbers of shells from shade-loving species and others that suggest an input of organic material from middening. It seems that a significant change in the treatment of ploughsoils may have taken place between the Early and the Late Bronze Age in the Western Isles. In the Neolithic to Early Bronze Age it appears that episodes of ploughing were interspersed with significant fallow periods, permitting shade-loving species and Lauria cylindracea to flourish (cf. Evans et al. 2012, 251), whereas from the Late Bronze Age onwards ploughsoils seem to have been fertilised with organic waste and seaweed (see below). Nevertheless, at Late Bronze Age to Iron Age Baleshare and Iron Age Hornish Point the practice of allowing land to lie fallow seems to have persisted (Thew 2003, 168). During periods of total or partial site abandonment surfaces could be colonised by denser, taller vegetation, allowing mollusc species that require more shade and dampness to move in. This can be observed within the sequence at Northton following the Beaker II and Iron Age I occupations (Fig. 5.4), when species typical of stable grassland were joined by Carychium tridentatum, Vertigo substriata, Aegopinella nitidula, Aegopinella pura and Vertigo pusilla, together with peaks in Lauria cylindracea.

Middening and the use of seaweed Deposits of domestic organic waste from both plants and animals tend to attract omnivorous and carnivorous species associated with damp conditions, such as Oxychilus alliarius and O. cellarius that are known to feed on rotting meat (Evans 1972, 188). Although O. alliarius occurs in small numbers in fairly moist and dense stable grassland in northern Scotland (Paul 1976), in these locations it is always accompanied by shade-loving taxa (see above), whereas high frequencies of this species only occur in natural habitats where there is a regular supply of rotting flesh (Young & Evans 1991), or quantities of decomposing dung as a result of heavy grazing (Evans & Vaughan 1983, 109). Peaks in O. alliarius within faunas indicative of less stable conditions and from contexts that contain significant quantities of archaeological waste, point to the spread of midden material around a settlement, as documented at sites like Baleshare, Balelone, Hornish Point, Newtonferry, Bornais and Horgabost. This midden material was either dumped to form discernible layers or added to ploughsoils. In the former case these spreads would have served both to dispose of domestic rubbish and to stabilise

5. Molluscs from dune-machair systems in the Western Isles

93

the rather delicate surface of the machair. In the latter instance, the midden waste was applied to act as a fertiliser for the rather dry, nutrient poor, sandy plough-soils. The earliest evidence for O. alliarius responding to the spread of midden material seems to be from Late Bronze Age layers at Baleshare, both as a fertiliser for a ploughsoil (Block 22: 1890–1460 cal BC*, 3360±80 BP, GU-2556; 1780–1390 cal BC*, 3285±85 BP, GU-1966; see endnote 1) and to stabilise fragile surfaces (Block 27: 1240–970 cal BC*, 2910±50 BP, GU-19173; Thew 2003, 300). The use of seaweed as animal fodder, for fuel, as a fertiliser, or to stabilise the machair surface, is well-attested in the Western Isles from both archaeological and ethnographic sources (Pain & Thew 2003; Cerón-Carrasco 2005, 32; Smith 2012, 390–1). Although marine shells form a natural component of the machair sand, these are normally fragmented and heavily abraded due to wind blast, while larger, edible marine shells that can be present in huge quantities in shell middens clearly represent food waste. By contrast, the presence of well-preserved specimens of small marine gastropods can only be interpreted as shells that were brought in with gathered seaweed. The earliest examples from the Western Isles come from Late Bronze Age contexts at Baleshare (Pain & Thew 2003, 174, 176). No small marine gastropods have so far been recovered from Neolithic, Beaker or Early Bronze Age levels. Small marine gastropods become much more common in Iron Age contexts at sites like Hornish Point, Baleshare, Balelone and Cladh Hallan, as well as in later sites like Bornais and Newtonferry (Pain & Thew 2003). In the Iron Age levels at Baleshare and Hornish Point, as well as in the Norse to Medieval levels at Bornais and Newtonferry, ploughsoils and other contexts contained both small marine gastropods and specimens of Oxychilus alliarius, confirming that seaweed was being spread with midden material as both a fertiliser and to protect fragile land-surfaces from wind erosion.

Climate and its influence on natural site formation processes All of the snail taxa recovered from machair deposits have geographical ranges that extend northwards of the Western Isles, so the rather limited temperature changes that took place during the mid to late Holocene (cf. Magny 1995; 2013) are unlikely to have influenced their survival. Instead, the occurrence of land snails in the area covered by the dune-machair systems of the Western Isles has been largely determined by environmental changes linked to human activities and as an indirect consequence of climate change. Increased sand deposition and instability within dune-machair systems can be correlated with periods of cooler, stormier climate (Dawson et al. 2011). Windblown sands tend to have rather poor molluscan faunas of very limited diversity, pointing to rapidly accumulating sand and rather sparse vegetation of mostly marram grass. Notable layers of clean windblown sand are present within the majority of site sequences investigated in the Western Isles, the earliest of which occurs between the Neolithic I and II horizons at Northton, Harris (Evans 1971). Windblown sands dating from the Beaker Period at Sligeanach (contexts 106, 107, 104; Evans et al. 2012), may correlate with

94

Matt Law and Nigel Thew

Figure 5.7. Open ground taxa typical of grassland from machair deposits in the Western Isles. Clockwise from top left: Cochlicella acuta, Helicella itala, Pupilla muscorum, Vertigo pygmaea, Vallonia excentrica, and Vallonia costata (photos: ML) deposits below the Beaker I horizon at Northton and at 70–50 cm in Section 1 at Rosinish (Vaughan 1976), while a slightly later episode of sand deposition at Sligeanach (context 19) may correlate with one at 35–20 cm in Rosinish Section 1. Correlating between sites is always problematic, however, as episodes of sand accumulation may have been rather localised, and the dating of deposits from most sites is far from precise. The use of OSL dating may help to resolve this problem. Nevertheless, it is clear that there are windblown sand deposits dating from the Neolithic and Beaker periods at several sites in the Western Isles. Similar deposits of windblown sand can also be observed in the complex Iron Age sequences at Baleshare (see Fig. 5.3), Hornish Point, Balelone and Rosinish, as well as in Norse to Medieval deposits at Bornais, Udal and Newtonferry. Molluscan evidence for freshwater flooding within machair sequences may also be an indicator of periods with a harsher climate. Low frequencies of freshwater marsh and aquatic species have been recovered from machair deposits at several sites in the Western Isles (Fig. 5.8), including Ensay, Baleshare, Hornish Point, Balelone, Bornais, Cladh Hallan, Cnip, Udal and Newtonferry, while marsh snails were regularly present within the sequence at Northton. These shells were probably introduced as ‘floaters’ during seasonal (winter) or sporadic flooding from marshes and freshwater lochs that commonly lie to the east of the dune-machair system (Ritchie 1979), although patches of marshy ground may also have been present in hollows between the dunes. Alternating sands and organic silts at Borve on Benbecula, for example, have marsh and aquatic molluscs (Evans 1979) typical of a small shallow lake at the edge of the

5. Molluscs from dune-machair systems in the Western Isles machair. If deposits with marsh snails can be shown to be contemporary at several sites, this may indicate periods with a cooler, wetter climate. Marsh species present between 305 and 265 cm at Northton (Carychium minimum, Galba truncatula and Oxyloma pfeifferi, Vertigo angustior and Zonitoides nitidus; Evans 1971), for example, may correlate with marsh snails present between 111 cm and 35 cm at Ensay (Spencer 1974, table 16), perhaps suggesting a climatic downturn during the latter part of the Beaker period and the Early Bronze Age. Marsh snails are also regularly present in Iron Age levels at Baleshare, Hornish Point and Cladh Hallan, as well as at Iron Age Balelone and Cnip. It is also important to understand past patterns of flooding within the machair as this may well have influenced both site location and agricultural practices, and thus siteformation processes, with some sites or parts of them, only being used seasonally (Thew 2003, 168–169).

95

Figure 5.8. Marsh species from machair deposits in the Western Isles. Clockwise from top left: Carychium minimum, Vertigo substriata and Galba truncatula (photos: ML)

Relative dating and biostratigraphy In the machair systems of Scotland and Ireland the arrival and expansion of Helicella itala and Cochlicella acuta (Fig. 5.7), together with a concomitant decline in previously common open ground species, provide clear biostratigraphic markers that can act as a relative dating tool (Evans 1972, 295; Thew 1989; 2003, 163; Law & Thew 2015, 126). In the Western Isles, until the Iron Age molluscan assemblages were very often dominated by the open ground species Vallonia costata, V. excentrica and Pupilla muscorum. The best available dating evidence for the arrival of Helicella itala and Cochlicella acuta comes from Baleshare and Hornish Point. H. itala was present in the British Isles from the late glacial onwards (Evans 1972, 180–182), but only appeared in these islands between the end of the Late Bronze Age at Baleshare around 700 BC1 (1020–800 cal BC*, 2740±60 BP, GU-1965) and the first part of the Early Iron Age (sensu Parker Pearson & Sharples 1999, and Parker Pearson 2012, 20) at Hornish Point between c. 550/500 and 400 BC (calibrated dates of 830–550*, 800–540* and 800–530 cal BC*; 2585±35 BP, SUERC-3207; 2520±35 BP, SUERC-3212, 2502±35 BP, SUERC-3220), rapidly establishing itself as a significant but by no means dominant faunal component. H. itala was also present in significant numbers from the start of the Iron Age occupation at Baleshare around

96

Matt Law and Nigel Thew

300/250 BC (760–380 cal BC*, 2390±55 BP, GU-1961) and at Balelone from c. 350–250 BC (780–390 cal BC*, 2440±80 BP, GU-1803). Cochlicella acuta was a later arrival in the British Isles (Evans 1972, 183), first colonising sand dunes in southwestern Britain during the Early Bronze Age (cf. Spencer 1975) and then progressing northwards up the west coast (cf. Kerney 1999). In the Western Isles it seems to have first appeared in small numbers around 200/150 BC within Early Iron Age levels at Hornish Point (550–200 cal BC*, 2325±50 BP, GU-2021), but remained rather rare until abandonment around 150/100 BC (350–30 cal BC, 2110±40 BP, SUERC-96) and was even scarcer in Iron Age levels at Baleshare until the site was abandoned around AD 50 (180 cal BC– cal AD 230, 1970±80 BP, GU-2554). Small numbers of both H. itala and C. acuta were present in Iron Age levels at Cladh Hallan and at Rosinish (Vaughan 1976), while H. itala was present in Middle Iron Age layers at Cnip, and there were rare C. acuta in the Iron Age deposits at Horgabost. At Northton H. itala first appeared in moderately large numbers within a deposit of clean windblown sand that overlies a fairly thick layer of pale brown sand that covers the Iron Age I level. It was joined soon afterwards within this windblown sand by low frequencies of C. acuta, which then expanded rapidly to dominate the Iron Age II level and subsequent deposits (Evans 1971, 59). C. acuta was similarly dominant in 9th–11th century Norse period levels (X and IX) at Udal (Spencer 1974; Crawford & Switsur 1977, 131–3) and Newtonferry (James & Ridout 2003), although at the latter site it was already dominant in windblown sands beneath the Norse levels. Helicella itala and Cochlicella acuta thus both arrived in the Western isles during the Early Iron Age, with H. itala appearing several centuries earlier, but they only began to dominate dune-machair grassland habitats during the second half of the Middle Iron Age or the first part of the Late Iron Age, between c. AD 100/200 and 500/600. It seems clear that they were able to out-compete and largely replace native xerophile open ground species within dune-machair habitats, possibly due to their ability to climb the stems of marram grass to avoid burial during times of significant sand movement. At Northton, for example, they both first appeared within a layer of clean windblown sand (Evans 1971, 59). At some point after the Iron Age Lauria cylindracea seems to have widened its ecological range in the Western Isles, partly replacing Pupilla muscorum, Vallonia costata and Vallonia excentrica in grassland habitats (Thew 2003, 163, 167; Evans 2004). It is scarce in the Norse to Medieval period contexts at Bornais (9th–13th/14th centuries), while at Newtonferry it is similarly rare in most Norse to Medieval levels (9th–12th centuries), but becomes frequent in relatively damp Medieval midden deposits with significant numbers of Oxychilus alliarius. Lauria is also common from the 11th/12th century onwards in the sequence at Udal, but only in faunas typical of dense, fairly moist, stable grassland, while Pupilla, Vallonia costata and V. excentrica remain fairly common. It thus appears that Lauria’s ecological expansion into grassland habitats took place during the late medieval to post-medieval period, firstly into fairly moist, stable grassland and then later into somewhat drier grassland habitats. Today this species is often rather abundant among stable grazed machair grassland, but it can also sometimes be fairly common among marram grass with a loose sandy substrate (Evans & Vaughan 1983, 109–110; Evans 2004, 372).

5. Molluscs from dune-machair systems in the Western Isles

97

Conclusions The calcareous shell sands of the machair were an important focus for settlement and agriculture from the Bronze Age until the historical period. Within these sands, well-preserved assemblages of snail shells are often abundant but of low diversity. Molluscan faunas from archaeological contexts can provide valuable information about site formation processes, the functioning of various structures and land-use/agricultural practices within the infield area. Changes in molluscan faunas through time reflect environmental changes related to vegetation and climate.

Acknowledgements Our sincere thanks go to Mike Allen for his patient and helpful editorial advice. Access to Hebridean material was made possible by Jacqui Mulville, Niall Sharples, Mike Parker Pearson, Helen Smith, Mike Church, Rosie Bishop, Kevin Colls and John Hunter.

Endnote 1

Most radiocarbon dates from Baleshare and Hornish Point were on seashells, which are often considered to be too old due to the Marine Reservoir Effect (Barber 2003, 220–221). Three calibrated dates on animal bones from the same stratigraphical blocks are, however, 100–200 years older than those on seashells, while three dates on charred barley grains from the same units are 160–200 years younger. More recent dates on seashells and barley grains from the same blocks at Hornish Point again point to the latter being 200–300 years younger (Ascough et al. 2005), suggesting that an MRE correction for dates on seashells for these sites may be roughly minus 200 years. In practice the MRE seems not to be fixed but to vary by both period and location (Cook et al. 2015). The calibrated results quoted with an * are on seashells and require an MRE correction of minus c. 200 years. Those without an * are on charred barley seeds and require no correction.

References Ascough, P. L., Cook, G. T., Dugmore, A. J., Scott, E. M. & Freeman, S. P. H. T. 2005. Influence of mollusc species on marine DELTA R determinations, Radiocarbon 47, 433–440 Barber, J. 2003. Radiocarbon dating. In Barber, J. (ed.), Bronze Age Farms and Iron Age Farm Mounds of the Outer Hebrides, 215–221. Scottish Archaeological Internet Reports 3. Edinburgh: Society of Antiquaries of Scotland Barber, J. 2011. Characterising archaeology in machair. In Griffiths, D. & Ashmore, P. (eds), Aeolian Archaeology: the archaeology of sand landscapes in Scotland, 37–54. Scottish Archaeology Internet Reports 48. Edinburgh: Society of Antiquaries of Scotland Boyd, J. M. & Boyd, I. L. 1996. The Hebrides: a habitable land. Edinburgh: Birlinn Butzer, K. W. 1982. Archaeology as Human Ecology. Cambridge: University Press Cerón-Carrasco, R. N. 2005. ‘Of Fish and Men’ (‘De iasg agus dhaoine’): a study of the utilization of marine resources as recovered from selected Hebridean archaeological sites. Oxford: British Archaeological Report 400

98

Matt Law and Nigel Thew

Cerón-Carrasco, R. N. 2006. The marine Mollusca with notes on the Echinoidea remains and terrestrial snails. In Armit, I. (ed.), Anatomy of an Iron Age Roundhouse: the Cnip wheelhouse excavations, Lewis, 180–182. Edinburgh: Society of Antiquaries of Scotland Chapman, H., Adcock, J. & Gater, J. 2009. An approach to mapping buried prehistoric palaeosols of the Atlantic seaboard in Northwest Europe using GPR, geoarchaeology and GIS and the implications for heritage management, Journal of Archaeological Science 36(10), 2308–2313 Cook, G. T., Ascough, P. L., Bonsall, C., Hamilton, W. D., Russell, N., Sayle, K. L., Scott, E. M. & Bownes, J. M. 2015. Best practice methodology for 14C calibration of marine and mixed terrestrial/marine samples, Quaternary Geochronology 27, 164–171 Crawford, I. & Switsur, R., 1977. Sandscaping and C14: the Udal, N. Uist, Antiquity 51(202), 124–136 Davies, P. 2008. Snails: archaeology and landscape change. Oxford: Oxbow Books Dawson, S., Smith, D. E., Jordan, J. & Dawson, A. G. 2004. Late Holocene coastal sand movements in the Outer Hebrides, N.W. Scotland, Marine Geology 210(1–4), 281–306 Dawson, S., Dawson, A. G. & Jordan, J. 2011. North Atlantic climate change and Late Holocene windstorm activity in the Outer Hebrides, Scotland. In Griffiths, D. & Ashmore, P. (eds), Aeolian Archaeology: the archaeology of sand landscapes in Scotland, 25–36. Scottish Archaeology Internet Reports 48. Edinburgh: Society of Antiquaries of Scotland Edwards, K. J., Whittington, G. & Ritchie, W. 2005. The possible role of humans in the early stages of machair evolution: palaeoenvironmental investigations in the Outer Hebrides, Scotland, Journal of Archaeological Science 32, 435–449 Evans, J. G. 1971. Habitat change in the calcareous soils of Britain: the impact of Neolithic man. In Simpson, D. D. A. (ed.), Economy and Settlement in Neolithic and Early Bronze Age Britain and Europe, 27–74. Leicester: Leicester University Press Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press Evans, J. G. 1979. The palaeo-environment of coastal blown-sand deposits in western and northern Britain, Scottish Archaeological Forum 9, 16–26 Evans, J. G. 2004. Land snails as a guide to the environments of wind-blown sand: the case of Lauria cylindracea and Pupilla muscorum. In Gibson, A. & Sheridan, S. (eds), From Sickles to Circles: Britain and Ireland at the time of Stonehenge, 366–379, Stroud: Tempus Evans, J. G., Law, M. & Thew, N. 2012. Stability and flux in the dune environment. In Parker Pearson, M. (ed.), From Machair to Mountains: archaeological survey and excavation in South Uist, 250–253. Oxford: Oxbow Books Evans, J. G. & Vaughan, M. 1983. The molluscs from Knap of Howar, Orkney, 106–114. In Ritchie, A. (ed.), Excavation of a Neolithic farmstead at Knap of Howar, Papa Westray, Orkney, Proceedings of the Society of Antiquaries of Scotland 113, 40–121 Goldberg, P. & Macphail, R. I. 2006. Theoretical and Practical Geoarchaeology. Oxford: Blackwell Hall, A. 1996. Quaternary geomorphology in the Outer Hebrides. In Gilbertson, D., Kent, M. & Grattan, J. (eds), The Outer Hebrides: the last 14,000 years, 5–12. Sheffield: Sheffield Academic Press James, H. F. & Rideout, J. S. 2003. Excavations at Newtonferry. In Barber, J. (ed.), Bronze Age Farms and Iron Age Farm Mounds of the Outer Hebrides, 109–113. Scottish Archaeology Internet Reports 3. Edinburgh: Society of Antiquaries of Scotland Kerney, M. 1999. Atlas of the Land and Freshwater Molluscs of Britain and Ireland. Colchester: Harley Books Law, M. & Thew, N. 2015. Land snails, sand dunes and archaeology in the Outer Hebrides, Journal of the North Atlantic, Special Volume 9, 125–133 Magny, M. 1995. Une histoire du climat, des derniers mammouths au siècle de l’automobile. Paris: Errance

5. Molluscs from dune-machair systems in the Western Isles

99

Magny, M. 2013. Orbital, ice-sheet, and possible solar forcing of Holocene lake-level fluctuations in west-central Europe: a comment on Bleicher, The Holocene 23, 1202–1212 Manley, G. 1979. The climatic environment of the Outer Hebrides, Proceedings of the Royal Society of Edinburgh 77B, 47–59 Pain, C. & Thew, N. 2003. The microscopic marine Mollusca. In Barber, J. (ed.), Bronze Age Farms and Iron Age Farm Mound of the Outer Hebrides, 173–177. Scottish Archaeology Internet Reports 3. Edinburgh: Society of Antiquaries of Scotland Parker Pearson, M. 2012. The machair survey. In Parker Pearson, M. (ed.), From Machair to Mountains: archaeological survey and excavation in South Uist, 12–73. Oxford: Oxbow Books Parker Pearson, M., Mulville, J., Sharples, N. & Smith, H. 2011. Archaeological remains on Uist’s machair: threats and potential. In Griffiths, D. & Ashmore, P. (eds), Aeolian Archaeology: the archaeology of sand landscapes in Scotland, 55–86. Scottish Archaeological Internet Reports 48. Edinburgh: Society of Antiquaries of Scotland Parker Pearson, M. & Sharples, N. 1999. Between Land and Sea, Excavations at Dun Vulan, South Uist, (SEARCH 3). Sheffield: Sheffield Academic Press Paul, C. R. C. 1976. The non-marine Mollusca of Colonsay and Oronsay, Journal of Conchology 29, 107–110 Paul, C. R. C. 1992. The non-marine Mollusca of Ulva, Inner Hebrides, Journal of Conchology 34, 175–178 Ritchie, W. 1979. Machair development and chronology in the Uists and adjacent islands, Proceedings of the Royal Society of Edinburgh 77B, 107–122 Sharples, N., Parker Pearson, M. & Symonds, J. 2004. The archaeological landscape of South Uist. In Housley, R. A. & Coles, G. (eds), Atlantic Connections and Adaptations: economies, environments and subsistence in lands bordering the North Atlantic, 28–47. Symposia of the Association for Environmental Archaeology. Oxford: Oxbow Books Smith, H. 1994. Middening in the Outer Hebrides: an ethnoarchaeological investigation. Unpublished PhD thesis, University of Sheffield Smith, H. 2012. The ethnohistory of Hebridean agriculture. In Parker Pearson, M. (ed.) From Machair to Mountains: archaeological survey and excavation in South Uist, 379–400. Oxford: Oxbow Books Smith, H. & Mulville, J. 2004. Resource management in the Outer Hebrides: an assessment of the faunal and floral evidence from archaeological investigations. In Housley, R. A. & Coles, G. (eds), Atlantic Connections and Adaptations: economies, environments and subsistence in lands bordering the North Atlantic, 48–64. Oxford: Oxbow Books Spencer, P. 1974. Environmental Change in the Coastal Sand Dune Belt of the British Isles: snail analysis of sites in Cornwall, The Orkneys, and the Outer Hebrides. Unpublished BSc dissertation, University College, Cardiff Spencer, P. 1975. Habitat change in coastal sand-dune areas: the molluscan evidence. In Evans, J. G. Limbrey, S. & Cleere, H. (eds), The Effect of Man on the Landscape: the highland zone, 96–103. London: Council for British Archaeology Research Report 21 Thew, N. 1989. Cochlicella acuta (Müller) and Helicella itala (Linné) in northern coastal shell-sand deposits, Conchologists’ Newsletter 110, 209–210 Thew, N. 2003. The molluscan assemblage. In Barber, J. (ed.), Bronze Age Farms and Iron Age Farm Mounds of the Outer Hebrides, 163–177, results tables 280–325. Scottish Archaeological Internet Reports 3. Edinburgh: Society of Antiquaries of Scotland Vaughan, M. P. 1976. Environmental Change in Areas of Blown Sand on the Western Coasts of the British Isles. Unpublished BA dissertation, University College, Cardiff Young, M. S. & Evans, J. G. 1991. Modern land mollusc communities from Flat Holm, South Glamorgan, Journal of Conchology 34, 63–70

6. Caves and molluscs Chris O. Hunt and Evan A. Hill

Sediments laid down in the shallower parts of cave systems often contain shells of land molluscs. These may offer valuable clues to changing environments and to human activity. To understand mollusc assemblages from cave fills, we must consider their taphonomy, from the taxonomic composition of assemblages, the nature of damage to and preservation of the shells and the nature of the sediments in which they are preserved. Examples are given of the information that can be gained from mollusc assemblages from cave fills.

Ecology of land molluscs in caves Caves are usually defined as subterranean cavities big enough to accommodate a human body. Rock shelters are in many ways similar to caves, differing in that they are usually located in undercuts in cliffs. Both caves and rock shelters occur in many bedrock lithologies, but the commonest and the most important archaeologically are solution cavities in limestone, which is composed predominantly of calcium carbonate, or in dolomite which is mostly calcium magnesium carbonate. The high calcium carbonate content in these lithologies particularly favours both the formation of mollusc shells and their preservation in what are usually rather calcareous infill sediments. It is worth noting that in favourable circumstances, caves and rock shelters can form in almost any reasonably structurally-strong lithology, but in acid rocks such as granites such sites tend to have acidic fills in which mollusc shell is rarely preserved. Most caves and rock shelters with archaeological significance lie well above the water table, are accessible to walking, scrambling or a short climb and have the dimensions of a small room, or larger. The size of a void below the ground surface does not, however, matter greatly to land molluscs, many of which inhabit ranges of at most a few square metres. Land molluscs have very variable tolerances to subterranean habitats (Fig. 6.1; Weigand 2014). Globally, a very few species may be troglobites – obligate cave dwellers – but these are confined to deep caves with little archaeological significance. A few species are regarded as troglophiles because they can maintain breeding populations and thus can be autochthonous within caves. Typically troglophile land molluscs are carnivorous,

6. Caves and molluscs

101

detritivore, fungivore or carrion feeders that require damp, sheltered habitats and are facultative shallow burrowers (Hunt 1993; Weigand 2014). Many of these taxa are also reported to live within cavities in rock-rubble (Evans 1972, 309–311; Evans 1976; Evans & Jones 1973), making the point that void size is less critical to molluscs than it is to people. A few troglophile molluscs may be vegetarians grazing algal films and plants growing within the photic zone. Carnivorous taxa regarded as troglophile such as Oxychilus cellarius, O. draparnaudi, Discus rotundatus and Vitrea spp. are widespread in northwest European caves (Hunt 1993; Weigand 2014) although analysis of Weigand’s data might suggest that of these, only Oxychilus cellarius is a true troglophile (Fig. 6.1). In southern Europe and North Africa, carnivorous burrowers such as Rumina decollata and the shelled slug Parmacella oliveri are regarded as troglophiles, as are algal-grazing Clausiliids such as Barcania kaltenbachi (Hunt et al. 2011). Trogloxenes do not have troglophile adaptions and so cannot complete their life cycles in caves, but use caves at times, for instance as sheltered places to hibernate, as has been documented for the Cepaea spp. and some Helix spp., or if not hibernaculas, as refugia from extremes of drought or cold (Weigand 2014). Many shelter-demanding molluscs are likely to be at least occasional trogloxenes. In the British Isles, Aegopinella nitidula, Azeca goodalli, Candidula intersecta, Cepaea hortensis, Clausilia bidentata, Cochlicopa lubrica, Discus rotundatus, Helix pomatia, Lauria cylindracea, Nesovitrea hammonis, Oxychilus cellarius, O. draparnaudi, O. helveticus and Vitrea crystallina have all been found alive in caves (Dixon 1974; Hazleton 1975; 1977; Jefferson 1989; Hunt 1993) while 59 land snail taxa have been reported from German caves (Weigand 2014), but this does not mean that all of these species regularly lived where they were found. The ecology of Candidula intersecta, for instance, would suggest that this was not a troglophile or trogloxene species and that its presence in a cave was not the result of its normal behaviour. The majority of land molluscs that are found in cave deposits are, however, unlikely to maintain even temporary populations in caves (Hunt 1993; Weigand 2014) and are thus allochthonous in the cave environment. These will almost certainly have arrived in caves by a variety of processes, discussed below.

Transport of land molluscs into caves Little has been written on land mollusc taphonomy in caves, but a number of vectors and processes are capable of transporting molluscs into caves (Hunt 1993; Girod 2011). Humans, on occasion, have transported substantial numbers of land snails into caves as part of foraging behaviour, for instance the huge land snail midden at Hang Boi Cave in Vietnam (Rabett et al. 2011) and the land snail-rich late glacial and early Holocene horizons at the Haua Fteah in Libya (Hunt et al. 2011) and Taforalt, Morocco, (Chapter 12, Taylor & Bell). It is also suggested (Girod 2011) that people may have brought vegetation into caves and thus inadvertently introduced land snails adhering to the plants. Other vectors include badgers, moles and rodents, which all carry land snails into their burrows in cave fills to eat them undisturbed (Hunt 1993). Birds such as thrushes will carry snails to anvil sites on rocks adjacent to cave mouths (Teichert & Seventy 1947;

102

Chris O. Hunt and Evan A. Hill

Rees 1965; Jenkinson et al. 1982; Girod 2011) from where fragments may be propelled or fall into the cave. Inadvertent transport of land molluscs on fur, skin and plumage has been recorded by mammals, birds, amphibians and even bumble bees (Rees 1952; 1965). Vitrea spp. for example produce sticky secretions which enable them to adhere to, and be carried long distances by migrating birds and other organisms (Rees 1965). A variety of geomorphic processes carry land molluscs into caves. These include processes involving running water, either in streams or as surface wash, massmovement processes such as mudflows, and fall processes (Hunt 1993; Girod 2011). Molluscs may be carried live into caves by these processes, or they may be recently dead (for instance land snails carried in streams into which they had been washed or fallen), or long-dead and recycled from soils or Quaternary sediments (Hill 2015; Hunt 1993; Hunt et al. 2015). In addition to these vectors and processes, it is very likely that the normal foraging behaviour of land molluscs will bring them into caves on a regular basis. This is strongly suggested by the distance-decay occurrence patterns of many land mollusc taxa recorded in German caves (Fig. 6.1; Wiegand 2014).

Recognition of taphonomic pathways In archaeological assemblages from caves, taphonomic pathways can be recognised by

1. The location, sedimentary and stratigraphic context of assemblages: The action of watermediated geomorphic processes such as riverine flow, wash and mudflow entering a cave will tend to lead to mollusc assemblages which occur in deposits formed by these processes. With waterlain sediments, particularly, there is likely to be assemblage change with distance into the cave. In these situations, allochthonous taxa are likely to be found in deposits originating in the catchment of the cave and to be most common close to inlets and entrances, generally declining in abundance into the cave. Autochthonous taxa, in contrast, are likely in very low-energy deposits which formed from atmospheric deposition and granular disintegration of or biogenic deposition from the cave roof, and in geomorphic locations where high-energy processes from outside the cave were not implicit in the formation of the sediments. Allochthonous taxa are otherwise likely to reflect the activities of animals and people, which will lead in some cases to the formation of distinctive deposits. For instance rodent nests are sometimes recognisable as more or less ovoidal bodies rich in characteristically-damaged large snail shells (Hunt 1993). Shell-middens left by human activities are sometimes very recognisable (eg, Rabett et al. 2011; Chapter 12, Taylor & Bell), although lower-intensity human use of land snails not leading to discrete midden deposits may be more difficult to identify with certainty without other taphonomic indicators (Girod 2011). 2. Patterns of occurrence and damage caused by geomorphic transport processes or vectors: Transport by geomorphic processes may lead to disintegration of shells, with abrasion and rounding of whole shells and fragments. This is particularly likely with riverine transport. Wash and riverine processes may also result in sorting and winnowing of assemblages, where heavy elements may not be carried as far as light or buoyant elements (Briggs et al. 1990). In riverine settings, gasses trapped in the apices of elongate shells of, for instance the Pupillidae or Clausiliidae, may cause these to float and be carried long distances undamaged and this might be expected in cave deposits of fluvial

Figure 6.1. Occurrence of land molluscs in different zones in caves in Hesse. Data from Weigand (2014)

6. Caves and molluscs 103

104

Chris O. Hunt and Evan A. Hill

Figure 6.2. Rodent-predated shells of Eobania vermiculata (Müller) origin. Patterns of damage may also identify animal and human vectors. Rodents and shrews tend to predate large land snails such as Cepaea, Helix and Arianta spp. by biting through the last whorl on the side of the shell opposite to the aperture, to extract the soft parts of the animal without having to deal with the foot (Fig. 6.2; Hunt 1993). Birds such as thrushes smash shells on anvil-stones, leading to characteristically shattered shell fragments (Hunt 1993). Burning is sometimes present on land snail shells as the result of roasting at the edge of fires (Rizner et al. 2009). Crushing of shells is also recorded as a way of extracting the soft parts for consumption (Bar 1977). Removal of the apices to enable removal of the soft parts was recorded at the Grotta di Uluzzo by Borzatti von Löwenstern (1964). Radmilli (1960), Hutterer et al. (2011; 2014) and Hill et al. (2015), all report humans piercing shells with stone artefacts or thorns to enable extraction of the soft parts (Fig. 6.3). Similar piercing producing indistinguishable perforations is recorded ethnographically from Tunisia, where people use their canines to pierce the shell to enable sucking out the soft parts (Ismail Saafi pers. comm. 2016). 3. The known habitat tolerances and preferences of species. The degree of ecological coherence

6. Caves and molluscs

105

Figure 6.3. Pierced Trochoidea cretica (Férrusac) from Late Quaternary deposits in the Haua Fteah (Libya)

of assemblages can be a good indicator of parts of an assemblage originating elsewhere than the deposition site and thus most probably outside the cave. For instance, in Pinhole Cave, Creswell, taxa of exposed habitats such as Pupilla muscorum and Vallonia excentrica are perhaps unlikely to have been part of troglophile or trogloxene faunas and are thus likely allochthonous to the cave. These occur mixed with sheltered-habitat ‘interglacial’ taxa such as Helicigona lapidica as well as possible troglophiles such as Discus rotundatus in what is an ecologically mixed assemblage. This probably arrived in the cave as a series of mudflows, which may have relocated some material from older Quaternary deposits as well as some contemporary material (Hunt 1989).

Applications of land molluscs in cave archaeology Land snails provide important information to the cave archaeologist. This may include evidence for dating, subsistence associated practices, non-subsistence behaviour and ancient environments. Some instances of each of these are given below.

106

Chris O. Hunt and Evan A. Hill

Dating Before the advent of modern AMS radiocarbon dating, stratified land snail faunas provided biostratigraphic evidence for the age of cave deposits and their contained archaeology (eg, Gale et al. 1984; Gale & Hunt 1985). This was done by correlating the appearance of species in the cave fill with the biostratigraphic scheme established by Kerney (1977). As dating techniques have become more refined, land snails are becoming amenable to dating by a variety of means. Radiocarbon dating of land snails requires careful evaluation of the radiocarbon ecology of the animals (Hill 2015; Chapter 22, Douka), but in favourable circumstances can lead to highly-resolved dating of cave sequences (Hunt et al. 2015). Other dating techniques are also possible including Electron Spin Resonance (Molodkov 2001; Blackwell 2006) and aminostratigraphy (Rabett et al. 2011), but Uranium-series dating is problematic because of the open-system behaviour of mollusc shell (Hellstrom & Pickering 2015).

Subsistence and associated practices Land snails are poor in fats but rich in other nutrients (Lubell 2004) and as such are likely to have been supplementary rather than staple foods (Lubell et al. 1976) and thus taken when preferred foods are in short supply. In rare cases land snails may be very abundant and form a true shell midden (eg, Hang Boi Cave: Rabett et al. 2011). Elsewhere, it is likely that part of the scatter of land molluscs present in many cave sites will reflect food refuse. Typically, the larger species in a fauna seem to have been targeted, as would be expected from optimal foraging theory (eg, Gutiérrez Zugasti 2011; Hunt et al. 2011; Lloveras et al. 2011; Rabett et al. 2011; Taylor et al. 2011). Recognition of food refuse is more certain if characteristic patterns of damage are recognised, such as piercing or apex removal, or if a large percentage of the shells show signs of heating, but this is not always the case. Scavenging molluscs may also be associated with food waste in caves. In some European sites, small carnivorous species such as Discus rotundatus and Oxychilus cellarius are typical scavengers and found in association with human occupations (Gale & Hunt 1985; Hunt 1989; 1993; Colonese et al. 2007), suggesting the discard of organic material attractive to these species.

Non-subsistence behaviour Land snail shells are generally small in size and structurally weak, so their use in nonsubsistence contexts is rare. We know of no reports of land snail tools, for instance. Pierced operculae of the Pomatiidae were probably used as beads several times in the past, with operculae of pomatid land snails from Neolithic and late Classical sites on Mount Carmel, Israel and from the Middle Stone Age site at Porc-Epic Cave, Ethiopia (Mienis 1990; 2003; Assefa et al. 2008).

6. Caves and molluscs

107

Ancient environments Land snails are environmentally constrained, so in suitable circumstances land snail assemblages may provide information about the nature of past cave and external environments. To disentangle these, attention must be paid to taphonomic factors. Land snail palaeoecology of cave sediments has been carried out fairly infrequently. In South Africa, land snail assemblages suggest climate regimes similar to present in some Middle Stone Age layers at Blombos Cave (Langjeans et al. 2012). In Spain, land mollusc assemblages indicated open rocky ground in deposits of the Epipalaeolithic midden at Nerja Cave (Jordá 2011). A sequence from open landscapes in the Younger Dryas and earliest Holocene to heavily vegetated landscapes during the later Early and mid Holocene was demonstrated on land snail evidence at Kirkhead Cave, UK (Gale & Hunt 1985). Vegetation close to the Haua Fteah (Libya) was apparently similar to present-day scrub during the Neolithic, but became considerably degraded during Classical times (Hunt et al. 2011). Very open highly-eroded landscapes are also suggested for the late Roman Period in Gozo, Maltese Islands (Hunt & Schembri forthcoming). A new development in environmental reconstruction is isotope studies of land snails (eg, Colonese et al. 2007; 2010a; 2010b; 2011) looking at 18O and δ13C signatures to reconstruct, respectively, palaeotemperature records during the lifetime of specimens and water retention efficiency of vegetation in a molluscs diet.

Conclusions Land snails offer considerable potential for archaeological work in caves. Not only were they a component of subsistence through much of later prehistory, but they also offer important information about the palaeoecology and environmental context of cave sites. Future research on the dating of mollusc shell from cave sediments is likely to provide important and robust geochronometers and while isotope work is still in its very early stages, it is already showing promise.

References Assefa, Z., Lam, Y. M. & Mienis, H. K. 2008. Symbolic use of terrestrial gastropod opercula during the Middle Stone Age at Porc‐Epic Cave, Ethiopia, Current Anthropology 49(4), 746–756 Bar, Z. 1977. Human consumption of land snails in Israel, Basteria 41, 53–58 Blackwell, B. A. B. 2006. Electron spin resonance (ESR) dating in Karst Environments, Acta Carsologica 35(2), 123–153 Borzatti von Löwenstern, E. 1964. La Grotta di Uluzzo C (Campagna di scavi 1964),Rivista di Scienze Preistoriche 19, 41–52 Briggs, D. J., Gilbertson, D. D. & Harris, A. L. 1990. Molluscan taphonomy in a braided river and its implications for studies of Quaternary cold-stage river deposits, Journal of Biogeography 17, 623–637

108

Chris O. Hunt and Evan A. Hill

Colonese, A., Zanchetta, G. & Fallick, A. 2010b. Stable isotope composition of Helix ligata (Müller, 1774) from Late Pleistocene-Holocene archaeological record from Grotta della Serratura (Southern Italy), Global and Plenetary Change 71, 249–257 Colonese, A. C., Zanchetta, G., Fallick, A. E., Martini, F., Manganelli, G. & Lo Vetro, D. 2007. Stable isotope composition of Late Glacial land snail shells from Grotta del Romito (Southern Italy): palaeoclimatic implications, Palaeogeography, Palaeoclimatology, Palaeoecology 254, 550–560 Colonese, A. C., Zanchetta, G., Dotsika, E., Drysdale, R. N., Fallick, A. E., Grifoni Cremonesi, R. & Manganelli, G. 2010a. Early-middle Holocene land snail shell stable isotope record from Grotta di Latronico 3 (southern Italy), Journal of Quaternary Science 25, 1347–1359 Colonese, A. C., Zanchetta, G., Drysdale, R. N., Fallick, A. E., Manganelli, G., Vetro, Lo, D., Martini, F. & Di Giuseppe, Z. 2011. Stable isotope composition of Late Pleistocene-Holocene Eobania vermiculata (Muller, 1774) shells from the Central Mediterranean basin: Data from Grotta della Oriente (Favignana, Sicily), Quaternary International 244, 76–87 Dixon, J. M. 1974. Biospeleology in North-West England. In Waltham A. C. (ed.), The Limestones and Caves of North-West England, 149–181. Newton Abbot: David & Charles Douka, K. 2017. Radiocarbon dating of marine and terrestrial shell. In Allen, M. J. (ed.), Molluscs in Archaeology, 382–400. Oxford: Oxbow Books Evans, J. G. 1976. Subfossil land snail faunas from rock-rubble habitat. In Davidson, D. A. & Shackley, M. L. (eds), Geoarchaeology, 397–399. London: Duckworth Evans, J. G. & Jones, H. 1973. Subfossil and modern land-snail faunas from rock rubble habitats, Journal of Conchology 28, 103–130 Gale, S. J. & Hunt, C. O. 1985. The stratigraphy of Kirkhead Cave, an Upper Palaeolithic site in Northern England, Proceedings of the Prehistoric Society 51, 283–304 Gale, S. J., Hunt, C. O. & Southgate, G. A. 1984. Kirkhead Cave: biostratigraphy and magnetostratigraphy, Archaeometry 26, 192–198 Girod, A. 2011. Land snails from Late Glacial and Early Holocene Italian sites, Quaternary International 244, 105–116 Gutiérrez Zugasti, F. I. 2011. Early Holocene land snail exploitation in northern Spain: the case of La Fragua Cave, Environmental Archaeology 16(1), 36–48 Hazleton, M. 1975. The biology of the Mendip Caves. In Smith, D. I. (ed.), Limestones and Caves of Mendip, 313–351. Newton Abbot: David & Charles Hazleton, M 1977. Life underground-biospeleology. In Ford, T. D. (ed.), Limestones and Caves of the Peak District, 231–261. Norwich: Geobooks Hellstrom, J. & Pickering, R. 2015. Recent advances and future prospects of the U–Th and U–Pb chronometers applicable to archaeology, Journal of Archaeological Science 56, 32–40 Hill, E. 2015. The Radiocarbon Dating of Terrestrial Molluscs in Northeast Libya. Unpublished PhD Thesis, Queen’s University Belfast Hill, E. A., Hunt, C. O., Lucarini, G., Mutri, G. Farr, L. & Barker, G. 2015. Land gastropod piercing during the Late Pleistocene and Early Holocene in the Haua Fteah, Libya, Journal of Archaeological Science: Reports 4, 320–325 doi:10.1016/j.jasrep.2015.09.003 Hunt, C. O. 1989. Molluscs from A. L. Armstrong’s excavations in Pin Hole Cave, Creswell Crags, Cave Science 16, 97–100 Hunt, C. O. 1993. Mollusc taphonomy in caves: a conceptual model, Cave Science 20, 45–49 Hunt, C. O. & Schembri, P. J. Forthcoming. Historic-period terrestrial environments and soil erosion in the Maltese Islands: evidence from mollusc assemblages from a cave-fill at Ghajn il-Kbira, near Victoria, Gozo, Ancient Near Eastern Studies Hunt, C. O., Gilbertson, D. D., Hill, E. A. & Simpson, D. 2015. Sedimentation, re-sedimentation and chronologies in archaeologically important caves: problems and prospects, Journal of Archaeological Science 56, 109–116

6. Caves and molluscs

109

Hunt, C. O., Reynolds, T. G., el-Rishi, H. A., Buzian, A., Hill, E. & Barker, G. W. 2011. Resource pressure and environmental change on the North African littoral: Epipalaeolithic to Roman gastropods from Cyrenaica, Libya, Quaternary International 244, 15–26 Hutterer, R, Mikdad, A. & Ripken, T. E. J. 2011. Species composition and human exploitation of terrestrial gastropods from Taghit Haddouch, an Early Holocene archaeological site in NE Morocco, Archiv für Molluskenkunde 140(1), 57–75 Hutterer, R., Linstädter, J., Eiwanger, J. & Mikdad, A. 2014. Human manipulation of terrestrial gastropods in Neolithic culture groups of NE Morocco, Quaternary International 320, 83–91 Jefferson, G. T. 1989. Cave biology in South Wales. In Ford, T. D. (ed.), Limestones and Caves of Wales, 56–69. Cambridge: University Press Jenkinson, R. D. S., Bramwell, D., Briggs, D. J., Gilbertson, D. D., Griffin, C. M., Watts, C. J. & Wilkinson, M. 1982. Death of a Wolf. Nottingham: Creswell Crags Visitor Centre Report 3 Jordá, J. A., Avezuela, B., Emili Aura J. & Martín-Escorza, C. 2011.The gastropod fauna of the Epipalaeolithic shell midden in the Vestibulo chamber of Nerja Cave (Málaga, southern Spain), Quaternary International 244, 27–36 Kerney, M. P. 1977. A proposed zonation-scheme for Late-glacial and Postglacial deposits using land molluscs, Journal of Archaeological Science 4, 387–390 Langjeans, G. H. J., Dusseldorp, G. L & Henshilwood C. S. 2012. Terrestrial gastropods from Blombos Cave, South Africa: research potential, South African Archaeological Bulletin 67, 195, 120–144 Lloveras, L., Nadal, J., Garcia Argüelles, P., Maria Fullola, J. & Estrada, A. 2011. The land snail midden from Balma del Gai (Barcelona, Spain) and the evolution of terrestrial gastropod consumption during the late Palaeolithic and Mesolithic in eastern Iberia, Quaternary International 244, 37–44 Lubell, D. 2004. Are land snails a signature for the Mesolithic-Neolithic transition in the circumMediterranean? In Budja, E. (ed.), The Neolithization of Eurasia – paradigms, models and concepts involved, Documenta Praehistorica 27, 1–24 Lubell, D., Hassan, F. A., Gautier, A. & Ballais, J.-L. 1976. The Capsian escargotières, Science 191, 910–920 Mienis, H. K. 1990. Landsnails from a Neolithic site in Nahal Oren, Israel, The Papustyla 90, 5, 8–9 Mienis, H. K. 2003. Molluscs from the excavation of Horvat Raqit, Carmel. In Dar, S. (ed.), Raqit – Marinus Estate on the Carmel, Israel, 55–56. Tel Aviv: Eretz-Geographic Research and Publications/Israel Exploration Society Molodkov, A. 2001 ESR dating evidence for early man at a Lower Palaeolithic cave-site in the Northern Caucasus as derived from terrestrial mollusc shells. Quaternary Science Reviews 20, 1051–1055 Rabett, R., Appleby, J., Blyth, A., Farr, L., Gallou, A., Griffiths, T., Hawkes, J., Marcus, D., Marlow, L., Morley, M., Tâń, Nguyêń Cao, Son, Nguyêń Van, Penkman, K., Reynolds, T., Stimpson, C. & Szabó, K. 2011. Inland shell midden site-formation: Investigation into a late Pleistocene to early Holocene midden from Tràng An, Northern Vietnam, Quaternary International 239, 153–169 Radmilli, A. M. 1960. Considerazioni sul Mesolitico italiano. Annali dell’Università di Ferrara, n.s., sez. XV 1–3, 29–48. Ferrara: Ferrara University Rees, W. J. 1952. The role of amphibia in the dispersal of molluscs, British Journal of Herpetology 1, 125–129 Rees, W. J. 1965. The aerial dispersal of molluscs, Proceedings of the Malacological Society, 36, 269–282 Rizner, M., Vukosavljevic, N. & Miracle, P. 2009. The paleoecological and paleodietary significance of edible land snails (Helix sp.) across the Pleistocene-Holocene transition on the eastern Adriatic coast. In McCartan, S., Schulting, R., Warren, G. & Woodman, P. (eds), Mesolithic Horizons, 527–532. Oxford: Oxbow Books

110

Chris O. Hunt and Evan A. Hill

Taylor, V. K. & Bell, M. 2017. Land Mollusc middens. In Allen, M. J. (ed.), Molluscs in Archaeology, 195–212. Oxford: Oxbow Books Taylor, V. K., Barton, R. N. E., Bell, M., Bouzouggar, A., Collcutt, S., Black, S. & Hogue, J. T. 2011. The Epipalaeolithic (Iberomaurusian) at Grotte des Pigeons (Taforalt), Morocco: a preliminary study of the land Mollusca, Quaternary International 244, 5–14 Teichert, C. & Seventy, D. L. 1947. Deposits of shells transported by birds, American Journal of Science 245, 322–328 Weigand, A. M. 2014. Next stop underground. Variable degrees and variety of reasons for cave penetration of terrestrial gastropods, Acta Carsologica 43(1), 175–183

Wetlands and fresh- and brackish-water

7. Molluscs from the floodplain alluvial sediments in the Thames Valley Mark Robinson

Molluscan shells can be very conspicuous in the alluvial clays and clay loams which are often to be found covering floodplains. They attracted the attention of some of the pioneers of the study of molluscs from Holocene deposits in Britain in the early 20th century and by the second half of the 20th century at least limited studies were being undertaken on overbank alluvial sediments by Quaternary geologists (Davies 2008, 101). However, the initial work by J. G. Evans (1972) on land snails, which greatly advanced techniques and showed the value of molluscan analysis on prehistoric sites in Britain, did not include alluvium. Probably the first major regional study of molluscs from overbank alluvium in archaeology in Britain was undertaken by Robinson (eg, Lambrick & Robinson 1979; Thomas et al. 1986) on the floodplain of the Upper Thames Valley, that is the Thames Valley above the Goring Gap (Fig. 7.1). This was in part due to the scale of quarrying and construction work on the floodplain which both revealed archaeology beneath or interstratified within the alluvium and provided convenient sections through alluvial sequences. It became clear that the floodplain had only been experiencing alluviation for the past 2000 years or so and that much of the floodplain was relatively dry during the mid Holocene (Robinson & Lambrick 1984). A synthesis of the molluscan results from Roman to medieval alluvium suggested that the floodplain was predominantly grassland but that it was possible to differentiate the molluscan assemblages of pasture from those of hay meadow (Robinson 1988). Subsequently Evans investigated molluscs from more localised and in some instances more complex environments in the Thames drainage system including the infill to a palaeochannel of the Thames at Runnymede (Evans 1991a) and the fills of two streams of the River Kennet at Anslow’s Cottages (Evans 1992). On both these sites, the upper fills of the channels would have been continuous with the overbank alluvium of the adjacent floodplains. Evans (1991b) refined the interpretation of the wet-ground molluscs from Anslow’s Cottages by defining ‘taxocenes’ of recurring assemblages of shells and relating them to the environment from which they were derived. Evans and his students (Evans et al. 1992; Davies 2003) further developed the interpretation of molluscan taxocenes from Holocene overbank alluvium in southern England. The problem of some past environments having no modern analogues was addressed and a series of taxocenes was related to floodplain environments. Davies (2008, 101–118) thoroughly reviewed

7. Molluscs from the floodplain alluvial sediments in the Thames Valley

113

Figure 7.1. The Upper Thames Valley and the location of the sites the study of molluscs from overbank alluvium. Concentrating particularly on the floodplains of the River Wylye and River Itchen, correspondence analysis was used to help define the species associations from alluvial deposits that were regarded as applicable to central southern England (Davies 2008, 102). Work on mollusca from the overbank alluvium of the floodplain of the Upper Thames Valley (Fig. 7.1) has continued since 1988 but, probably more importantly, a greater understanding has developed of the changing floodplain environment (Robinson 1992; Booth et al. 2007; Lambrick & Robinson 2009). Six assemblages of molluscs from the floodplain of the Upper Thames Valley plus one from the Middle Thames Valley are considered in relation to floodplain environment (Fig. 7.2). Results are then presented from an alluvial sequence on the site of the Drayton Cursus (Figs 7.3–7.4), which experienced major environmental change. The nomenclature for molluscs follows Anderson (2005).

Floodplains and overbank alluvium Floodplains are areas of relatively flat land extending from a river or stream which

114

Mark Robinson

experience episodes of inundation when the flow of water exceeds the capacity of the river or stream. Episodes of flooding can vary between active floodplains, some experiencing flooding each year while others perhaps flood only once every few decades. Individual floods do not necessarily cover the full area of a floodplain and there can also be considerable variation between the lengths of time an area of floodplain is submerged. Floodplains have often been created by the migration of river channels flowing across a relatively soft geology, eroding material on the outside of bends and depositing sands and gravels on the inside of bends. A fluviatile terrace is eventually formed as the bends migrate downstream. This process can give a distinct step in the topography at the edge of a floodplain. A terrace can cease to be active floodplain if the base level of the river which created it is lowered. The subsequent downcutting isolates the terrace on the valley side. Floodplains can also be left dry for long periods, sometimes millennia, if there is a decline in peak discharge, which is related to such factors as changes in precipitation levels or human activity affecting the rate of run-off in the catchment. The molluscan fauna of a floodplain is obviously greatly influenced by a combination of the character of flooding, how readily water drains from the floodplain and the type of vegetational cover. In its broadest usage, the term alluvium covers any sediment lain from moving water. However, in this chapter the term will be limited to overbank alluvium; the sands, silts and clays deposited on river beds and within ditches are excluded. Overbank alluvium tends to level up the surface of a floodplain but under some circumstances, deposition of coarser particles closer to the bank can create relatively well-drained levees alongside the channel with backswamps beyond. Alluviation requires a source of fine sediment. This can either be sediment from within the channel system or material derived from wider erosion within the catchment. Alluviation rates are rarely constant. Both climate and human activity in the catchment influence whether alluvium occurs and the rate of sedimentation. A pattern shown by some rivers in central and southern England is of rapid channel migration and fluviatile deposition of gravel terraces during cold (glacial) periods but greater channel stability and some overbank alluviation under temperate (interglacial) conditions (Robinson 1992). These river systems also show a relationship between alluviation and the scale of agriculture in the catchment during the Holocene.

Molluscan analysis If molluscs are to be studied from an alluvial sequence, it is important for the sequence to be dated. In archaeology, this can often be achieved by the stratigraphic relationship with datable archaeological contexts as, for example, in the initial work on the Holocene alluvial sequence of the Upper Thames Valley (Robinson & Lambrick 1984). Many examples were found of archaeological features such as ditches, causeways, settlements and burial monuments interstratified in the alluvial sequences. Radiocarbon dating can be effective on organic inclusions in anaerobic alluvial deposits although caution must be taken to ensure that the material being dated does not contain intrusive roots, tissue

7. Molluscs from the floodplain alluvial sediments in the Thames Valley

115

from submerged aquatic plants (due to the hard-water effect) or residual items that have been washed into the deposit. Identifiable seeds from terrestrial and emergent aquatic plants often prove the most reliable if there are not wooden structures such as preserved wattle-work. On rare occasions dendrochronological dating is possible on timbers in alluvium. OSL (optical stimulated luminescence) dating and magnetic dating can also prove useful (Rees-Jones 1995; Clark 2003) and do not depend on the alluvium being waterlogged. It is necessary to consider the taphonomy of mollusc shells from floodplain contexts before any useful palaeoecological interpretation can be made. Floodplains themselves are often rich habitats for molluscs. Floodplain habitats can range from the very simple, for example closely-grazed open pasture, through to the complex such as where tall vegetation provides a more three-dimensional aspect to the habitat. Floodplains tend to experience great seasonal variation in wetness. A floodplain can retain shallow pools of water throughout the late winter and spring but be dry by the late summer, resulting in very different seasonal faunas (see below). Dead shells are readily carried by floodwaters which not only move shells around on the floodplain but can introduce considerable quantities of shells of aquatic molluscs from a river onto its floodplain where they become incorporated into the alluvial sequence. Davies (2008, 101) notes that high-magnitude rapid alluviation events tend to result in a prevalence of aquatic molluscs from the river whereas low-magnitude seasonal alluviation results in a prevalence of terrestrial species which lived on the floodplain. Mollusc shells are only preserved in soils and sediments with a high (calcareous) pH (Evans 1972, 23). Such conditions are usually the result of the presence of calcium carbonate. Whether alluvium is calcareous is dependent on the pH of the water from which it was derived rather than the underlying geology of the floodplain. Sometimes the calcareous sediment deposited by a river and calcareous floodwaters percolating through it are insufficient to counter the effects of rainwater leaching of shells and, probably a very important factor on some floodplains, acids released by the decay of plant material. Where conditions are not sufficiently calcareous for good preservation, shells, which are composed of calcium carbonate in the mineral form of aragonite, can be left in a very fragmentary condition or lost whereas the opercula of species such as Bithynia, where the calcium carbonate is in the form of calcite, are more resistant to leaching. Sampling of alluvium is best undertaken from an exposed section through the deposit. While samples can also be taken by coring, it is not so easy to determine the stratigraphic relationships. Samples of 1.0 kg generally yield sufficient shells but sometimes the concentration of shells is so high that 0.25 kg is enough. Sampling intervals as close as 40 mm can be effective but it is often necessary to make a compromise between the degree of resolution achieved and the resources available for the study. The procedure for the analysis of samples from alluvial sediment follows Evans (1972, 44–5) except that if the samples have a high clay content, freezing and thawing the samples helps to break down the clay structure. From personal experience, this technique is both more time-efficient and does less damage to shells than trying to pass untreated samples through a sieve mesh.

116

Mark Robinson

Figure 7.2a. Mollusca from seven sites on the Thames floodplain

7. Molluscs from the floodplain alluvial sediments in the Thames Valley

117

Figure 7.2b. Mollusca from seven sites on the Thames floodplain (continued) Following the identification of the shells and the calculation of the minimum number of individuals for each taxon in each sample, the results can be presented in a stratigraphic histogram showing percentage or absolute abundance in a histogram as described by Davies (2008, 7–9) (see Fig. 7.2). Sometimes groups of species are plotted together in the histogram, for example taxa which are thought to have been transported by floodwaters rather than living on the floodplain.

Floodplain environments in the Thames Valley and their molluscan assemblages A series of molluscan assemblages will be considered in relation to the environments under which the shells were accumulating. It is hoped this will enable similar elements to be recognised in assemblages elsewhere where different species fill somewhat similar ecological niches. Where appropriate, the assemblages will be related to the taxocenes of Davis (2008, 102). The results are displayed in Figure 7.2a and Figure 7.2b.

118

Mark Robinson

Rapid alluviation on a floodplain with some pools of standing water (Farmoor, Oxfordshire, Lambrick & Robinson 1979, 102–3, Sample Column II, 7–15)

The alluvial clay of this sample was probably Roman and overlay a middle Iron Age surface within a settlement-related enclosure in the floodplain of the Upper Thames. The moving-water species Bithynia tentaculata and Valvata cristata comprised about a third of the assemblage and had probably been derived from the River Thames. The remainder of the assemblage largely comprised lymnaeids such as Radix balthica and planorbids such as Gyraulus crista and Planorbis planorbis. They could either have originated in the Thames or overflowed from a nearby cut-off palaeochannel which probably held standing water for the remainder of the year. Three members of these families, Lymnaea palustris, Galba truncatula and Anisus spirorbis s.l., were well represented. They are all, to varying degrees, amphibious and it is argued that they were able to flourish in shallow temporary pools left by receding floodwater in the spring. Shells of terrestrial molluscs were entirely absent even though the vegetation was interpreted as flood-pasture (Robinson 1988). However, it is likely that the grass had to grow through a thin layer of mud each spring. The assemblage corresponded to Taxocene 8 of Davies (2008, 102).

Alluviation on a floodplain with pools of water (Port Meadow, Oxfordshire, Robinson 1988, 104, Port Meadow alluvium)

The alluvial clay of this sample was medieval. Flowing water species were absent, perhaps because the sedimentation rate was lower than at Farmoor and the locality was distant from any flowing channels of the Thames. The assemblage was dominated by Galba truncatula and Anisus spirorbis s.l., the former being an amphibious species and the latter a slum aquatic species which can tolerate the drying out of its habitat. Fully terrestrial species were almost entirely absent, comprising little more than 1% of the total assemblage. Port Meadow is an expanse of closely-grazed floodplain near Oxford which has been common pasture since at least early medieval times. Although the site of the sample now tends to retain water after the winter floods have receded, pools of shallow water lingering throughout spring in some years, in summer the ground often dries hard and cracks. When the current fauna of Port Meadow was investigated in July 1982, no terrestrial snails were found but G. truncatula and A. spirorbis were observed emerging from cracks in the ground on rainy nights (some then entering pitfall traps intended for Coleoptera). The interpretation was made that the site retained pools for long enough for these two species to maintain breeding populations but conditions were too hostile for terrestrial taxa: too wet in winter for dry-ground taxa but too exposed and insolated in summer for the taxa usually associated with damp riverside grassland. Another potentially amphibious member of this community which has been recorded from alluvium on other sites is Lymnaea palustris (Robinson 1988, 107). This assemblage can perhaps be regarded as intermediate between Taxocenes 4 and 5 of Davies (2008, 102).

7. Molluscs from the floodplain alluvial sediments in the Thames Valley

119

Alluviation on a floodplain with pools of standing water and, in summer, tall herbaceous vegetation (Pixey Mead, Oxon, Robinson 1988, 104, Pixey Mead wet alluvium)

This sample was from the top of a medieval alluvial sequence of a site which has been documented as hay meadow since the Middle Ages. Flowing-water molluscs were only represented by a single specimen of Bithynia sp. The most abundant shells were of Galba truncatula and Anisus spirorbis was also present. However, terrestrial individuals outnumbered the amphibious snails, Succineidae being the most abundant followed by slugs from the family Limacidae, Carychium minimum, Cochlicopa sp., Vallonia sp. including V. pulchella and Trochulus hispidus. All but C. minimum were found living on Pixey Mead in July 1982. The higher ratio of G. truncatula to A. leucostoma compared with Port Meadow was perhaps a reflection of less standing water remaining after floods. However, it was probably not simply the slightly less wet conditions which favoured the terrestrial taxa. The herbaceous vegetation of a hay meadow would have provided shade and a more humid microclimate at ground level during the summer. The terrestrial component of the fauna corresponds to Taxocene 2 of Davies (2008, 102). Similar assemblages are widespread in alluvium of medieval date on the floodplain of the Upper Thames and its tributaries, very often in localities where there is independent evidence for the medieval occurrence of hay meadow (Robinson 1988). Sometimes the terrestrial fauna is more diverse, with the addition of Punctum pygmaeum, Pupilla muscorum, Vertigo antivertigo and V. pygmaea. The occurrence of P. muscorum is thought to reflect drier conditions in the winter while the occurrence of V. antivertigo was perhaps due to damper conditions during the summer.

Alluvium from short-duration flooding on a free-draining site (Wallingford, Thomas et al. 1986, 180–181, 0.32–0.44)

A sample of medieval or possibly late Roman alluvium exposed in the bank of the River Thames contained a range of shells of aquatic molluscs including Theodoxus fluviatilis, Bithynia spp., Valvata piscinalis and Gyraulus albus. However, the amphibious taxa Galba truncatula and Anisus spirorbis were absent. Terrestrial taxa comprised over three quarters of the assemblage. Three species predominated: Cochlicopa sp., Vallonia pulchella and Trochulus hispidus. Carychium minimum was absent. The occurrence of a diverse riverine fauna was probably a reflection of the proximity of the sampling point to the river while G. truncatula and A. spirorbis were probably absent because floodwaters dried rapidly from the site. The terrestrial component of the assemblage would be regarded as typical of short-turfed damp grassland and corresponds to Taxocene 1 of Davis (2008, 102). Further up the sequence the occurrence of Pupilla muscorum suggested a transition to the somewhat drier conditions favoured by Taxocene 6 of Davies (2008, 102).

120

Mark Robinson

A floodplain where alluviation had ceased and conditions had become drier (Mingies Ditch, Oxfordshire, Allen & Robinson 1993, 231–232, Column 781, 0–0.01)

The floodplain of the River Windrush was sufficiently dry to support a permanent middle Iron Age settlement although the site experienced alluviation in the early medieval period. The pre-Iron Age soil of the site was earthworm-sorted and almost entirely shell-free. However, at the very base (10 mm) of the soil profile was a layer of shells and shell fragments which had been taken down by earthworm activity and were resting on the interface between the soil and the underlying limestone river gravels of the floodplain terrace. The majority of the shells were from flowing water taxa, including Bithynia tentaculata, Valvata piscinalis and Ancylus fluviatilis. However, there were also shells of molluscs of shaded habitats such as Carychium sp., Cochlodina laminata, Discus rotundatus and Oxychilus cellarius. It is suggested that the origin of the soil was fine alluvial sedimentation over the floodplain terrace but as alluviation declined, so woodland became established. Once alluviation and flooding ceased, soil development resulted in the shells migrating to the bottom of the profile and de-calcification of the A horizon of the soil.

A dry floodplain without flooding (King’s Weir, Oxfordshire, Bowler & Robinson 1980, 7, Layer 19)

Much of the floodplain of the Upper Thames Valley experienced little or no flooding during the early and mid Holocene. Non-calcareous brown earth soils developed in which shells do not survive. However, the construction in the early Bronze Age of a round barrow which had some limestone gravel in its core resulted in the preservation of the shells of the contemporaneous fauna on the floodplain surface (a fact that was not fully appreciated when the site was reported upon 35 years ago). The snail fauna was characteristic of dry short-turfed grassland, the most abundant shells being of Vallonia excentrica. Cochlicopa sp., Pupilla muscorum and Trochulus hispidus were also present but shells of aquatic and wet grassland taxa were absent. A rich woodland assemblage of later Mesolithic or Neolithic date was recovered from a tree-throw hole on the Upper Thames floodplain at Yarnton, Oxfordshire (Robinson unpublished). In both cases, the floodplain was effectively an environment similar to the higher gravel terraces although subsequently the sites at King’s Weir and Yarnton experienced alluviation from the Thames.

Alluviation on a permanently wet floodplain (Eton Rowing Course, Buckinghamshire, Allen et al. 2013, 83–85, Sample 665)

Unlike the Upper Thames, the Thames of the Middle Thames Valley had a tendency in the earlier Holocene to form levees, which sufficiently impeded drainage of floodwaters to create backswamps which remained wet throughout the year. An assemblage from

7. Molluscs from the floodplain alluvial sediments in the Thames Valley

121

an early Mesolithic sequence at the site of the Eton Rowing Course, Dorney, contained relatively few shells of fully aquatic taxa. The amphibious species Galba truncatula and Anisus spirorbis were well represented along with species of marsh habitats, particularly Carychium minimum and Zonitoides nitidus. However, the diversity of the marsh component to the assemblage was low.

Other assemblages It is not intended to give details of the assemblages found in the more localised contexts which occur on floodplains, for example cut-off palaeochannels or man-made archaeological features. However, several points are worth making. Alluvial deposits can often be found filling the hollows left by archaeological features, for example ditches. Such deposits have the advantage for study that the archaeological context constrains the dating of the alluvium and the sequence is thicker than on the ground surface of the floodplain. However, it must be appreciated that the fauna which lived in the hollow above the archaeological features is likely to reflect wetter conditions and possibly taller vegetation than prevailed on the remainder of the floodplain. By their very nature, floodplains tend to comprise water-lain sediments. If the shells of aquatic molluscs are present in those sediments, they are likely to have become incorporated into the fills of archaeological features that cut them. This can present a problem for interpretation if the site was occupied in a period when flooding was not occurring. For example, shells of Bithynia tentaculata and Valvata piscinalis were abundant in the fills of some of the Iron Age archaeological features at Claydon Pike, Glos, but they had been derived from the floodplain gravels rather than from contemporaneous flooding (Robinson 2007, CD 4.4, Claydon Pike: The Middle Iron Age Environment). Sometimes a floodplain can experience flooding from water which does not have a heavy sediment load but is carrying shells of flowing-water aquatic molluscs. At the site of the Eton Rowing Lake, Dorney, Buckinghamshire, in the Middle Thames Valley, there was Iron Age and Roman settlement on an island of higher ground on the floodplain. Shells of flowing-water molluscs were absent from the Iron Age to early Roman archaeological features whereas shells of Bithynia tentaculata were present in mid-Roman contexts around the periphery of the settlement, showing that the height reached by flooding had risen in the Roman period (Robinson unpublished).

An alluvial sequence (Drayton, Robinson 1988, 105, table 2; Bradley et al. 2003, 163–178)

It has been observed that great environmental change can occur with time in Holocene floodplain environments and that alluviation can give a stratigraphic sequence containing molluscan shells which reflect environmental change during the period of sedimentation. The sequence at Drayton illustrates this well (Fig. 7.3). Archaeomagnetic dating of the clay alluvium at Drayton showed that overbank alluviation did not begin

122

Mark Robinson

Figure 7.3. The sample locations from the alluvial sequence above the bank of the Drayton Cursus until around c. AD 1 (Barclay et al. 2003, 165). Until then, the floodplain gravel only had a covering of 0.08 m of a non-calcareous sandy clay loam. There was Neolithic and Beaker activity on the site including the construction of a cursus monument and episodes of tree-clearance which resulted in the creation of tree-throw holes. Shells were absent from the prehistoric soil, even where it was sealed beneath a thin layer of limestone gravel which formed the bank to the cursus. However, the disturbance to the underlying gravel in the tree-throw holes left the soil within them sufficiently calcareous that shells were preserved, albeit in fragmentary condition. The alluvium was sampled for shells from a column above the cursus bank (Figs 7.3–7.4). The first 0.21 m of alluvium, which spanned the Roman period, did not contain shells. The rate of sedimentation slowed after about AD 400 (from a depth of 0.44 m in Fig. 7.4) and shells were present. The rate of alluviation increased from about AD 800 (from 0.38 m) and deposition was largely complete between about 1400 and 1500 (Barclay et al. 2003,

7. Molluscs from the floodplain alluvial sediments in the Thames Valley

Figure 7.4. Mollusca from Drayton

123

124

Mark Robinson

170–171). Above a depth of 0.31 m, the alluvium had a higher silt component and its content included Chalk microfossils, suggesting a greater proportion of the sediment was derived from the Berkshire Downs rather than the Cotswolds. The molluscs from the tree-throw hole were mostly shade-loving taxa which commonly occur now in woodland habitats. The most numerous shells were of Clausilia bidentata and Discus rotundatus. The preservation of shells was poor, so Carychium tridentatum, at 8% of the assemblage, was probably under-represented. The only snail more usually associated with open habitats was a single individual of Vallonia costata but it does occur at low frequency in woodland (Evans 1972, 156–158). The molluscs gave no hint of wet or even damp conditions. The onset of alluviation in the Late Iron Age obviously implies that conditions had become much wetter than in the Neolithic. The first samples in the alluvial sequence to contain shells (from –0.44 to –0.31 m) had a prominent component of aquatic molluscs likely to have been derived from permanent or almost permanent bodies of water. They comprised around 18% of the assemblages. Flowing water species such as Bithynia tentaculata were present but the majority were snails of stagnant water such as Radix balthica which had perhaps been flushed from lengths of channel which only experienced seasonal flow. Almost all the remaining shells were from three amphibious taxa: Lymnaea palustris, Anisus spirorbis and Galba truncatula. As was suggested for Farmoor (p. 118), it is believed these species flourished in pools of water which lingered after late-winter floods had receded. Terrestrial molluscs, including those of marshy habitats, only comprised 0.5% of the assemblages and, as argued for Port Meadow (p. 118) was believed to be a reflection of closely-grazed pasture. The three amphibious taxa predominated between –0.31 and –0.17 m while the proportion of terrestrial molluscs remained extremely low. However, the percentage of aquatic taxa dropped to 3% of the assemblages. Possibly alluviation was resulting in the silting of nearby minor channels on the floodplain which had been their main source though, of course, alluvial sediments continued to reach the site from the main channel of the Thames. Between –0.17 m and the modern ground surface, the proportions of both aquatic and amphibious molluscs declined. While this was in part a reflection of a considerable increase in the percentage of terrestrial molluscs unrelated to the degree of wetness of the site, conditions do appear to have become drier. The two more aquatic of the amphibious taxa, Lymnaea palustris and Anisus spirorbis, declined in abundance much more rapidly than Galba truncatula, the most ‘terrestrial’ of the amphibious group. The percentage of G. truncatula also declined over the three samples from 38% of the total molluscs to 16%. The proportion of terrestrial molluscs rose from 56% to 83% in the top sample. The terrestrial species would be appropriate as a fauna of damp grassland. Vallonia pulchella was the most numerous. The occurrence of Carychium minimum suggested that the grassland was not closely grazed while the presence of a member of the Succineidae was suggestive of somewhat marshy areas, as might be expected from the presence of the amphibious species. Species of dry grassland such as Vallonia excentrica were absent. It has been argued (p. 119) that such a fauna on the Thames floodplain is characteristic of grassland managed as hay meadow.

7. Molluscs from the floodplain alluvial sediments in the Thames Valley

125

Conclusions The Mollusca from the alluvium on the Thames floodplain and the palaeosols sealed beneath it track the major environmental changes which were occurring largely as a result of human activity during the second half of the Holocene. Neolithic woodland gave way in some areas to dry grassland in the early Bronze Age but the molluscan faunas suggested the floodplain was experiencing little or no flooding. Increasing wetness and the onset of widespread flooding in the Iron Age was followed by alluviation from the Roman period onwards. The fauna was initially one of wet muddy pasture on which pools of floodwaters lingered but this gave way in many areas during the Saxon or medieval periods to the more diverse fauna of alluvial hay meadow. Many of these changes were shown by the sequence from Drayton. The differences between the molluscan faunas from managed alluvial pasture and traditionally-managed hay meadow are particularly useful for the study of changing land use. With the rise of towns in the region, hay meadow became one of the main uses of the floodplain of the Upper Thames Valley and remained so into the 20th century (Robinson 1992). Such a transition cannot readily be detected by pollen analysis (Lambrick & Robinson 1988) and although it could be revealed by waterlogged seeds or waterlogged insect remains appropriate deposits are rarely found. Mention has already been made of the possibility of encountering molluscan assemblages from habitats which no longer occur. Modern floodplain faunas are not particularly well known. The ecological surveys that are being undertaken in order to help interpret archaeological results are making a useful contribution to malacology. It was noted (Robinson 1988, 109–10) that the restricted fauna of some closely-grazed alluvial pastures, which comprises only three species, all amphibious, had not previously been described.

References Allen, T., Barclay, A., Cromarty, A.-M., Anderson-Whymark, H., Parker, A., Robinson, M. & Jones, G. 2013. Opening the Wood, Making the Land. The Archaeology of a Middle Thames Landscape: the Eton College Rowing Course Project and the Maidenhead, Windsor and Eton Flood Alleviation Scheme. Volume 1: Mesolithic to Early Bronze Age. Thames Valley Landscapes Monograph 38. Oxford: Oxford Archaeology Allen, T. G. & Robinson, M. A. 1993. The prehistoric landscape and Iron Age enclosed settlement at Mingies Ditch, Hardwick-with-Yelford, Oxfordshire. Thames Valley Landscapes: the Windrush Valley 2. Oxford: Oxford University Committee for Archaeology Anderson, R. 2005. An annotated list of the non-marine Mollusca of Britain and Ireland, Journal of Conchology 38, 607–639 Barclay, A., Lambrick, G., Moore, J. & Robinson, M. 2003. Lines in the Landscape: cursus monuments in the Upper Thames Valley. Oxford Archaeology Thames Valley Landscapes Monograph 15. Oxford: Oxford Archaeology Booth, P., Dodd, A., Robinson, M. & Smith, A. 2007. The Thames Through Time, the archaeology of the Gravel Terraces of the Upper and Middle Thames, the early historical period: AD 1–1000, Oxford Archaeology Thames Valley Landscapes Monograph 27. Oxford: Oxford Archaeology

126

Mark Robinson

Bowler, D. & Robinson, M. 1980. Three round barrows at King’s Weir, Wytham, Oxon. Oxoniensia 45, 1–8 Clark, A. J. 2003. Archaeomagnetic dating. In A. Barclay, G. Lambrick, G., Moore, J. & Robinson, M., Lines in the Landscape: cursus monuments in the Upper Thames Valley, 185–187, Oxford Archaeology Thames Valley Landscapes Monograph 15. Oxford: Oxford Archaeology Davies, P. 2003. The interpretation of Mollusca from Holocene overbank alluvial deposits: progress and future directions. In Howard, A. J., Macklin, M. & Passmore, D. (eds), Alluvial Archaeology in Europe, 291–302. Lisse: Balkema Davies, P. 2008. Snails: archaeology and landscape change. Oxford: Oxbow Books Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press Evans, J. G. 1991a. The land and freshwater Mollusca. In Needham, S. P., Excavation and Salvage at Runnymede Bridge, 1978, 263–274. London: British Museum Evans, J. G. 1991b. An approach to the interpretation of dry-ground and wet-ground molluscan taxocenes from central-southern England. In Harris, D. R. & Thomas. K. D. (eds), Modelling Ecological Change, 75–89. London: Institute of Archaeology Evans, J. G. 1992. Mollusca. In Butterworth, C. A. & Lobb, S. J., Excavations in the Burghfield Area, Berkshire, 130–143. Salisbury: Wessex Archaeology Report 1 Evans, J. G., Davies, P., Mount, R. & Williams, D. 1992. Molluscan taxocenes from Holocene overbank alluvium in central southern England. In Needham, S. P. & Macklin, M. G. (eds), Alluvial Archaeology in Britain, 65–74. Oxford: Oxbow Monograph 27 Lambrick, G. H. & Robinson, M. A. 1979. Iron Age and Roman Riverside Settlements at Farmoor, Oxfordshire. London: Council for British Archaeology Research Report 32 Lambrick, G. H. & Robinson, M. A. 1988. The development of floodplain grassland in the Upper Thames Valley. In Jones, M. K. (ed.), Archaeology and the Flora of the British Isles, 55–75. Oxford: University Committee for Archaeology Monograph 14 Lambrick, G. & Robinson, M. 2009. The Thames Through Time. The Archaeology of the Gravel Terraces of the Upper and Middle Thames. Late Prehistory: 1500 BC–AD 50. Thames Valley Landscapes Monograph 29. Oxford: Oxford Archaeology Rees-Jones, J. 1995. Optical Dating of Selected British Archaeological Sediments. Unpublished D. Phil thesis, University of Oxford Robinson, M. A. 1988. Molluscan evidence for pasture and meadowland on the floodplain of the Upper Thames basin. In Murphy, P. & French, C. (eds), The Exploitation of Wetlands, 101–112. Oxford: British Archaeological Report 186 Robinson, M. A. 1992. Environment, archaeology and alluvium on the river gravels of the South Midlands. In Needham, S. P. & Macklin, M. G. (eds), Alluvial Archaeology in Britain, 197–208. Oxford: Oxbow Monograph 27 Robinson, M. 2007. The invertebrate and waterlogged plant remains. In Miles, D., Palmer, S., Smith, A. & Perpetua Jones, G., Iron Age and Roman Settlement in the Upper Thames Valley, CD Section 4: Claydon Pike environmental reports, 4.4. Thames Valley Landscapes Monograph 26. Oxford: Oxford Archaeology Robinson, M. A. & Lambrick, G. H. 1984. Holocene alluviation and hydrology in the Upper Thames basin, Nature 308, 809–14 Thomas, R., Robinson, M. A., Barret, J. & Wilson, R. 1986. A late Bronze Age riverside settlement site at Wallingford, Oxon, Archaeological Journal 143, 174–200

8. Wetlands: freshwater and slum communities Terry O’Connor

‘The nomenclature is confused in a variety of ways. Thus what would be called a pond in lowland country may be dignified as a tarn, loch, lough or llyn in the hills. A “lake” in many parts of England may be a stream; a “pool” may be a pond, a good-sized lake or a slow-flowing piece of a river; a “ditch” in Ireland sometimes means a fence or bank, and you may either fall into a “dyke” or climb over it; “spring” may mean a wood, and the dialect dictionary would no doubt reveal many other possibilities of perplexity, most of which would, as I hope to show, be solved by a knowledge of the Mollusca present.’ Boycott (1936, 117–118)

Where archaeological investigations are closely involved with sites of former human activity, the freshwater sediments that we might encounter are likely to derive from short-lived, often man-made, environments, rather than conventional lentic or lotic freshwater systems. Where the context is not man-made, we are often dealing with river margins or small natural ponds. This poses something of a challenge when it comes to the interpretation of freshwater mollusc assemblages. Environmental conditions in the ditch, flooded pit, swamp or whatever, are likely to have been quite testing, possibly fluctuating in temperature and oxygenation, possibly intermittently drying out, probably subject to frequent perturbation (Fig. 8.1). If the freshwater feature lacked continuous connection to a larger, more stable, freshwater body, there is also the question of how the mollusc community colonised in the first place, and therefore of what selection mechanisms may have acted to modify the species composition even before local environmental conditions acted. Davies (2006) sets out at length the difficulties in using modern analogues in the interpretation of molluscan assemblages, and that challenge recurs throughout this chapter. Faced with these confounding factors, it may seem to be a reasonable response simply to avoid basing any palaeo-environmental interpretation on freshwater mollusca from small former freshwater bodies, especially those likely to have been temporary in nature or frequently perturbed. It would certainly be quite difficult to base any such interpretation on a comparison of analogue faunas, because a modern analogue that is

128

Terry O’Connor

a good match for local environmental conditions is unlikely to have been subject to the same stochastic factors of colonisation. In a previous paper on this subject (O’Connor 1988), it was argued that comparison of archaeological mollusc assemblages with the communities reported in Boycott’s classic paper (Boycott 1936) showed little similarity. That said, we know from modern experience with garden ponds that the same few taxa will rapidly colonise a small freshwater body, appearing as if by magic and often multiplying to dominate the invertebrate fauna. In much of the UK, Planorbarius corneus somehow finds its way into ponds with ease, often with Radix peregra (NB I use this taxon sensu lato throughout, given the contentious taxonomy of Radix). There are certain taxa, therefore, that we might expect to find, including those such as R. peregra that can tolerate damp terrestrial environments for sufficient lengths of time to facilitate their colonisation of new water bodies (Fig. 8.2). On the other hand, taxa that require welloxygenated water, or that have particular requirements with regard to dissolved calcium or aquatic vegetation, might be unlikely to persist in flooded ditches or pits. Even if an initial flooding event were to introduce a population of, for example, Segmentina nitida to a small ditch or pond, persistence of that population would be highly unlikely given that species’ particular requirements and sensitivity to perturbation. The diversity and equitability of the assemblage may also be informative, distinguishing situations dominated by one or two species from others with a more equitable distribution of numbers (O’Connor & Evans 2005, 30). High-diversity assemblages are likely to represent either initial colonisation events, through translocation of a sample of the fauna of a larger water body, or environmental conditions that were relatively stable and of high biotic productivity. Interpretation needs to proceed case by case, taking into account the probable source populations of freshwater mollusca and lithological or other evidence for perturbation events. Despite the predictions of island biogeography models, species richness is only poorly predicted by pond size (Oertli et al. 2002). Clearly, interpreting mollusc assemblages from small freshwater sites is problematic. However, these small freshwater features were part of the landscape of former human activity, and therefore require some investigation if our archaeological study is to be as complete as possible. Perhaps the least complicated examples are those small freshwater bodies that had relatively little direct human impact. These may be encountered during off-site environmental sampling, for example in pollen cores or as silt or marl lenses within sand-dune systems. A good example of the former was the early- to mid-Holocene sequence reported by O’Connor and Bunting (2009) from Quoyloo Meadow, in the archaeologically important West Mainland of Orkney. Freshwater mollusca were recovered in varying concentrations throughout a 2.6 m core taken by Jane Bunting in 1994. The majority of the core represented sediments accumulating pre-Ulmus decline, and the base is dated to the Early Holocene in part on biostratigraphic grounds and in part based on an occurrence at 163 cm of Saksunavatn tephra (c. 8300 cal BC). Mollusc numbers are low throughout, with only Radix peregra and the bivalve Pisidum casertanum showing a near-continuous record. Despite the evident persistence of this small freshwater body, the mollusc fauna does not diversify to any great extent. In total, six Pisidium species are present (P. casertanum; P. personatum; P. milium; P. lilljeborgi; P.

8. Wetlands: freshwater and slum communities

129

hibernicum; P. nitidum), but the last four of these only as the occasional valve or two. At just one point in the core, at 104–112 cm, P. personatum temporarily outnumbers P. casertanum. Projection from the age/depth curve places this event around 5900 BC, firmly pre-Neolithic and near the peak of coryloid pollen. Bunting proposed that hazel woodland was encroaching on the small pond at this time, and O’Connor suggested that increased leaf-fall into the water created poor conditions for mollusca, giving the ‘slum’ specialist P. personatum the opportunity to dominate. Only two gastropods are present: Radix peregra and Galba truncatula. The latter species occurs sparsely through the core and is only abundant at 64–72 cm. This zone is associated in the pollen and plant macrofossil record with a decline in Ulmus and Pinus followed by a rise in Cyperaceae and Calluna, and a sharp rise in magnetic susceptibility. This is consistent with the onset of Neolithic human activity in the area, which is only a few kilometres east of Skaill Bay and the Skara Brae site, and a few kilometres north of the Neolithic ceremonial centre at Stenness. In summary, Quoyloo Meadow is a nice example of a small freshwater body with minimal human interference throughout much of its history that nonetheless never developed a diverse mollusc fauna, presumably because its small size and often shallow water inhibited colonisation by many species, and its relative isolation within Orkney, and Orkney’s isolation from mainland Britain, restricted the number of potential coloniser species. One intriguing point is the complete absence of planorbid gastropods, in particular Anisus leucostoma, which might have been expected in such a habitat and which extends as far North as Orkney today. Given that freshwater mollusca would have had to colonise Orkney after separation of the islands from mainland Scotland, a depauperate fauna is unsurprising, though a modest number of species, including some planorbids, successfully colonised the islands at some time during the Holocene. The Quoyloo Meadow core may indicate that planorbid colonisation of Orkney was post-Neolithic, and so possibly dependent on people or livestock as a vector. When we try to understand the colonisation of past water bodies, good data on past species distributions are important, though not necessarily available. To sum up Quoyloo Meadow, the mollusca integrated well with the pollen evidence to show a small pond or group of ponds that persisted as a minor landscape feature through the Early to mid-Holocene. The mollusca reflect the effects of encroaching hazel woodland at one point in the core, and the clearance and increased erosion that coincides with the onset of the Neolithic. The turloughs reported by Porst and Irvine (2009) make a useful comparison, as these are small freshwater ponds on the karst limestone of western Ireland, biotically rich and distinctive, and subject to little human disturbance. The 2008 survey of five turloughs returned just 12 freshwater mollusc taxa, with a range of two to nine taxa, and around five being typical for any one turlough. A different gastropod taxon numerically predominated in each turlough; lymnaeid species in three of the five, and Bithynia tentaculata and Succinea putris in one each. This inter-site variability may be reflecting subtle differences in microhabitat between turloughs, though one suspects that stochastic factors of colonisation have also played a significant part. Pip (1986) examined a wide range of environmental variables to explain the distribution of molluscan species in freshwater ponds in Eastern Manitoba, Canada. She showed

130

Terry O’Connor

Figure 8.1. Slums ancient and modern. Medieval pits become temporary ponds following heavy rain in York, June 2008

Figure 8.2. Radix peregra. Often the first to colonise unpromising wetland habitats. These are from a small pool in mountains above Llanfairfechan, Gwynedd. Note the damaged apices

that some water quality variables may have been significant, but that ‘… the element of chance was probably important as well’ (Pip 1986, 214). Means of colonisation have been a rich source of speculation. Darwin famously experimented with the role of ducks’ feet as a vector for immature stages of freshwater mollusca (Darwin 1860, 385) and Boag (1986) devised remarkable windtunnel experiments to show that ducks’ plumage had significant potential as a means of molluscan transport. Galba truncatula appears in the archaeological record in contexts that certainly suggest some form of active transport, for example occurring in otherwise terrestrial mollusc assemblages in

8. Wetlands: freshwater and slum communities

131

Iron Age pits at Winklebury Camp, Hampshire (Thomas in Smith 1977) and in RomanoBritish ditches near Folkestone, Kent (this author, unpublished records). More recently, Kappes and Haase (2012) compiled available data for the speed at which freshwater mollusca could potentially colonise in the absence of a vector. Perhaps unsurprisingly, pulmonates are generally faster than prosobranchs or bivalves, with Radix species hurtling along at several tens of metres per day. Compared to Boag’s estimate that a small gastropod could be translocated for up to 10 km by hanging on to a flying mallard (Anas platyrhynchos) for a plausible period of 15 minutes, these figures highlight the importance of vectors in molluscan distribution. Turning to man-made features, the issue of colonisation becomes particularly significant. A ditch or borrow-pit, for example, that becomes a temporary freshwater body may derive its initial water input by flooding from some nearby water body, or by the accumulation of rainfall, or by lateral seepage of groundwater. In the first case, the new ‘pond’ may be seeded with a range of macroinvertebrates, whereas if the latter processes predominate, colonisation will be much slower and more stochastic. Ditches are a common feature of human settlement and agricultural sites, a necessary part of managing the location and movement of surface water. Those ditches may persist for decades or centuries, being put to a range of uses and developing their own distinctive ecology. Periodic recutting or reflooding may be significant factors in that ecological development. Large ditches at Cawood Castle, Yorkshire, probably of medieval date, were investigated by Cath Neal and Emma Tong in 2010 as part of the archaeological investigation of earthworks, including putative fishponds, around Cawood. The samples yielded a number of freshwater mollusc assemblages, modest in size but informative (see Table 8.1). Logistical constraints made it necessary to sample by coring, as at Quoyloo Meadow, thus restricting the sediment sample size. The total of 16 taxa is quite high for an isolated water body, inconsistent with stagnation or marked eutrophication, and indicating a good nutrient and oxygen status. One of the archaeological questions regarding these features was whether they had a connection to any other water body. From the molluscan evidence, that would appear to have been the case. The super-abundance of one taxon in CWU4 may indicate that this sample represents a newly-created or recently-disturbed habitat patch within the system, subject to rapid colonisation by an opportunist species. That possibility could have been resolved had it been possible to see the deposits in section, to see whether CWU4 by chance sampled a lens atypical of the generality of sediments in the ditch. One of the disadvantages of coring is that the larger sedimentary context may be difficult to infer. Of the more abundant taxa, Valvata cristata is typical of slowly-flowing or still water, usually well-vegetated and well-oxygenated, and much the same can be said for V. piscinalis. Gyraulus albus is relatively uninformative, being widely tolerant of almost any freshwater conditions short of desiccation. The other planorbids are consistent with well-oxygenated, rather slow-moving water that does not dry out, and that would seem to describe the prevailing conditions in this system during the deposition of these sediments. The bivalve Pisidium casertanum will live in any soft-bottomed freshwater habitat. The only hints of a terrestrial component are the specimens of Succinea putris,

Terry O’Connor

132

Table 8.1. Samples of mollusc shells recovered from sediment cores taken through large ditch features at Cawood Taxon Bithynia leachii Bithynia tentaculata Valvata cristata Valvata piscinalis Radix peregra Lymnaea stagnalis Anisus leucostoma Planorbis carinatus Bathyomphalus contortus Gyraulus albus Gyraulus crista Planorbarius corneus Hippeutis complanatus Succinea putris Oxyloma elegans Pisidium casertanum

CWB5

CWB6

CWB7

CWL4

CWL5

CWU4

– – 1 – – – – – – – – – – – – –

– 2 12 9 – – 2 1 1 – 1 2 – – – 1

2 – 1 3 – – – – – 3 1 – – – – –

3 1 20 4 2 – 1 2 1 6 – – – – – 3

4 4 9 – – – 2 – – 9 1 – – 1 – 1

– 2 – – – 4 – – – 27 – – 3 1 6 3

Gastropods are quantified as apices unless otherwise indicated, bivalves as individual valves

Oxyloma elegans, and perhaps Radix peregra. These species are commonly found together in wetland environments and on emergent vegetation at the edge of water bodies. In sum, then, the molluscs indicate these earthworks to have held a mesotrophic freshwater body, probably slow-flowing over a muddy substrate (hence Valvata spp.) and with ample vegetation including emergent plants at the margins. Given that the feature was close to a busy settlement, it would have been an easy mistake to assume that it received regular input of refuse (and worse), and was therefore stagnant and foul. The mollusca indicate the contrary, raising interesting questions about the function and maintenance. The possibility that these broad ditch-like features at Cawood may have been used as fishponds was raised during their investigation. The molluscan evidence, indicating maintenance of water conditions, would certainly be more consistent with that function than with a ditch or moat that became a convenient place for refuse disposal. Ditches and other large features present something of a sampling challenge. In the case of Cawood, multiple core samples were available, giving some possibility of representing faunal variability along a length of ditch. Hill-Cottingham (2006, 204) makes the important point, based on her detailed surveys of ditches in Somerset, that ‘… if the whole length of a ditch is not sampled, it is probable that species will be missed’. Repeated sampling of a length of ditch exposed in archaeological excavation is not always possible, either because the ditch is only seen in cross-section at one location or only cored at one point, or for logistical reasons such as the cost of greatly increased sample processing. We are unlikely to be able to meet Hill-Cottingham’s criteria, therefore, but must be aware of the consequences in terms of species visibility. Thus

8. Wetlands: freshwater and slum communities

133

the assemblages from Austin Friars, Leicester, reported by Girling (1981) and further discussed by O’Connor (1988) represent only a part of that medieval ditch complex. Limited excavation trenches sampled the ditch system, and sediments exposed in those trenches were sampled for a range of plant and animal macrofossils. A diverse input of mollusca dominated by Bithynia tentaculata, Valvata piscinalis, V. cristata and, remarkably, Lymnaea auricularia (=Radix auricularia) probably understates the diversity originally present in the ditch system, though the rapid decline to assemblages with little more than V. piscinalis is clear enough even if some of the less frequent taxa in the original fauna have evaded sampling. Robinson (1979) reports much the same range of taxa from Iron Age and Roman features on the Thames floodplain at Farmoor, Oxfordshire, and presumably the same constraints apply regarding sampling the full diversity. In the case of Farmoor, the colonisation of the ditches was probably by overbank flooding from the nearby Thames, whereas at Austin Friars, the high-diversity assemblages indicate a deliberate flooding from the River Soar, a short distance to the east of the site. The same taxa were seen in samples associated with medieval activity around the town of Selby, not far from Cawood (Carrott et al. 1993). One trench sampled alluviation close to Selby Dam, a small stream tributary of the River Ouse, and sediment samples were retrieved during excavation. A useful freshwater mollusc assemblage was reported from a phase associated with local water management to create a fishpond. The range of species reported is similar to that at Cawood: Valvata cristata, V. piscinalis, Bithynia tentaculata, Planorbis planorbis, Bathyomphalus contortus, Gyraulus crista and Pisidium species not further identified. The shells were in good condition and included many juveniles, indicating that the assemblage had undergone minimal transport between death and deposition. As at Cawood, a well-vegetated mesotrophic habitat is indicated, consistent with the putative use as a fishpond. Carrott et al. report the same range of taxa from the fills of the nearby Kirk Dyke, a medieval ditch that was probably connected to Selby Dam. Here, however, mollusca only occurred in the lowest sediments, from the original flooding and colonisation of the Dyke, where the similarity to the Selby Dam assemblage is understandable. As the Dyke became associated with more and more human activity, and seems to have been increasingly used for dumping in all senses, the mollusc diversity falls off rapidly, with none at all found in samples from the upper sediments. The sequence shows that the Kirk Dyke received a faunal input, probably from Selby Dam, when it was first constructed, then received little or no further colonisation as conditions in the ditch deteriorated. A further example from the same region comes from medieval fills of the Hall Garth moat at Beverley (Dobney et al. 1994). The stratigraphic phase that represents early ‘in use’ sedimentation in the moat, probably of mid-14th century date, yielded one quite species-rich assemblage that may therefore be another case of initial colonisation during the deliberate flooding of a ditch. Gyraulus albus, Valvata piscinalis, V. cristata and Bithynia tentaculata numerically predominate, with a number of lymnaeid and planorbid taxa. As silting and infilling of the ditch by refuse progressed, the mollusc assemblages become less species-rich, and more indicative of deoxygenated, slum conditions. A number of taxa persist, notably Gyraulus albus, but diversity falls, and Radix peregra becomes numerically dominant. There may be evidence of an attempt

134

Terry O’Connor

to re-flood and thereby ‘clean’ the ditch. Late in Phase 4, following relining of the ditch in the 16th century, there is a more species-rich assemblage dominated by Valvata cristata, highly suggestive of some translocation of mollusca into the ditch. Possibly the system was drained, repaired, and then re-flooded from a water body that held source populations. Less distinctive, but intriguing, is the occurrence at intervals through the medieval fills of assemblages in which there is a big discrepancy between numbers of shells and numbers of operculae of Bithynia Figure 8.3. Bithynia tentaculata, showing the tentaculata (Fig. 8.3). A marked shortage operculum in place covering the aperture of operculae may indicate phases in which water-flow through the ditch was increased sufficiently to ‘winnow’ the operculae away from the sampling point without significantly translocating the shells. In this example, the mollusca give some indication of the changing use of the ditch during its ‘lifetime’. Taking the Selby and Beverley examples together, they show molluscan assemblages indicative of a wellmaintained system consistent with use as a fishpond (Selby Dam), a ditch that became a sewer (Kirk Dyke) and a moat that had changing conditions, including attempts to ‘refresh’ the water conditions (Hall Garth). The archaeology of river and lake edges often gives evidence of pre-occupation conditions and the initial impact of human settlement and other activities. An informative reduction in freshwater mollusca was seen in riverside sediments sampled at 24–36 Tanner Row, York (Hall & Kenward 1990). These sediments were exposed during deep and spatially-limited excavations which aimed to explore the Roman sequence in what later became the colonia. In samples of late 2nd century AD date, representing the southwest bank of the River Ouse as settlement in that area began, there were good numbers of freshwater mollusca. These are listed in Table 8.2. The species are a mix of obligate freshwater species, such as Theodoxus fluviatilis and species tolerant of riverside mud and emergent vegetation. The presence of a number of specimens of Omphiscola glabra is interesting as this species is in sharp decline across its European range today, and is described as ‘vulnerable’ in the UK (Fig. 8.4). It may be a useful indicator of small, rather swampy, habitats that are inclined to dry out, conditions not unlikely to have occurred in patches around the Ouse flood-plain prior to construction and drainage work on the eventual site of York. O. glabra, in common with some other lymnaeids, undergoes seasonal migration within stream systems, making it resilient in the case of isolated perturbations but susceptible to disruption and discontinuity of migration space (Rondelaud et al. 2005). However, with species that are scarce and local

8. Wetlands: freshwater and slum communities

135

Table 8.2. Freshwater mollusca from late 2nd century AD deposits at 24–36 Tanner Row, York

Figure 8.4. Omphiscola glabra. A species of rather weedy, swampy habitats, now quite rare in the UK

Unio tumidus Theodoxus fluviatilis Valvata cristata Valvata piscinalis Galba truncatula Omphiscola glabra Stagnicola palustris Radix peregra sensu lato Aplexa hypnorum Anisus leucostoma Planorbarius corneus Planorbis planorbis Carychium minimum Oxyloma elegans Succinea putris From Hall and Kenward (1990, 389–391) and this author’s records, with revised taxonomy

today, there is always the possibility that we are seeing only the habitats in which their relict populations have managed to survive, and not the full habitat repertoire of that species. Preece (1998) is not alone in stressing the importance of mollusc distribution records derived from fossil and archaeological sources for species conservation, just as archaeology needs good modern records. Kurzawska and Kara (2015) present mollusca from a similar depositional context in early medieval Pszczew, western Poland. Stratigraphy at the site represents the construction and use of an early medieval ‘stronghold’ on the banks of Lake Pszczewskie. An initial phase with minimal human activity is succeeded by timber constructions and a substantial ditch, interpreted by the excavators as a defensive moat. A 120 cm column was taken through the stratigraphy for molluscan analysis, providing larger samples for analysis than would have been obtained by coring, but with the same disadvantage of sampling the sequence only at one spatial location. Regrettably for this paper, the column did not include fills of the ‘moat’, and it therefore represents sediment accumulation at the water’s edge as human activity intensified, very much like the Tanner Row sequence. Although freshwater mollusca occur throughout the sequence, the general trend is from a predominance of freshwater taxa at the base, representing the pre-activity waterside, to more terrestrial taxa higher in the profile. Equitability is high in the freshwater mollusca at the profile base, and low amongst the assemblages higher in the sequence as terrestrial mollusca, especially Vallonia pulchella, come to dominate. Assemblages associated with the structural phases typically have one abundant taxon (notably Radix peregra) and just a few shells of several others. In sum, the mollusca show the original settlement to have been at the edge of a productive, ecologically healthy water body, with gradual terrestrialisation of the margins as settlement and building

136

Terry O’Connor

proceeded, producing areas of damp grassland. Amongst the richer assemblage from the base of the profile were specimens of Gyraulus riparius. Kurzawska and Kara make the point that this species is rare today. This is certainly the case for Poland and Germany, though it is more common in, for example, Finland (Aho et al. 1981), and its rarity in Central Europe may be a largely modern phenomenon. One complicating factor on which we have only rather partial information is the matter of competitive exclusion. Modern field studies may show a particular habitat association for Species A, in a characteristic community. However, If Species B is absent from its typical community in a rather different habitat, Species A may be able to persist successfully in that habitat, given the absence of some direct or indirect competition from Species B. In modern field studies, certain taxa may be more widespread than in the past, such as Planorbarius corneus, whilst other such as Omphiscola glabra may be much less widespread. It is a matter of speculation rather than evidence that either or both changes in community composition may have affected competitive exclusion, and thus the habitat distribution of some other mollusc species. Boycott (1936) was aware of this possibility, observing on the basis of the repeated observation of a group of ponds that what he calls Planorbis spirorbis (= Anisus spirorbis) lives in ‘bad’, ie, slum, habitats because it is excluded from better habitats by competition from other mollusca. In the absence of other taxa, perhaps through the arbitrary filter of colonisation, Anisus spirorbis can thrive in mesotrophic habitats. The taxa that we encounter in archaeological deposits derived from small ponds, flooded ditches and swampy riversides are quite problematic in interpretation, though with care we can extract some useful information out of them. In the examples reviewed here, certain species recur regularly, and often in abundance. Comparing these taxa with the ecological categorisation of freshwater mollusca originally proposed by Bruce Sparks (1961), his ‘slum’ species occur surprisingly infrequently, other than Pisidium casertanum (Table 8.3). The Pisidium species are difficult to review as they are often not fully identified in archaeological records, a matter that is further discussed below. Of the ‘slum’ gastropods, Galba truncatula occurs the most often, though by no means commonly, and Aplexa hypnorum is distinctly infrequent. The commonly-occurring Radix peregra, Sparks classes as ‘catholic’. As discussed above, the widespread occurrence of this species in small water bodies ancient and modern may have much to do with its capacity for colonisation across terrestrial habitats. Deliberate introduction in recent times may also be a factor: Fysher (1925) mentions that R. peregra was introduced to angling waters in Yorkshire as food for fish. A surprising absence from Sparks’ list is Gyraulus albus, today a widespread species of wide ecological amplitude, and a common species in archaeological samples (Fig. 8.5). Boycott (1936), on whose work Sparks’ (1961) paper drew, discusses G. albus (as Planorbis albus) as a species that will live well enough in productive habitats but also capable of colonising small ponds and weedy streams. He observes of G. albus that ‘It or Planorbis contortus is the most likely Planorbis to be found by itself in inferior loci with permanent water’ (Boycott 1936, 144). In archaeological samples, therefore, G. albus is likely to be present and is not, on its own, a good habitat indicator. That said, the presence of G. albus in an assemblage should not be taken to

8. Wetlands: freshwater and slum communities

137

Table 8.3. The ecological groupings of freshwater Mollusca originally proposed by Sparks (1961), with taxonomy amended to reflect modern usage 1. A slum group (ie, small water bodies subject to drying, stagnation, temperature fluctuations) Galba truncatula Aplexa hypnorum Anisus leucostoma Musculium lacustre Pisidium casertanum Pisidium personatum Pisidium obtusale 2. A catholic group (ie, tolerate wide range of habitats) Lymnaea palustris Radix peregra Gyraulus crista Bathyomphalus contortus Hippeutis complanatus Sphaerium corneum Pisidium milium Pisidium subtruncatum Pisidium nitidum 3. A ditch group (ie, clean flowing water with abundant aquatic plants). Valvata cristata Planorbis planorbis Anisus vortex Anisus vorticulus Segmentina nitida Acroloxus lacustris Pisidium pulchellum 4. A moving water group (ie, streams and larger ponds with water movement) Valvata piscinalis Bithynia spp. Lymnaea stagnalis Physa fontinalis Pisidium amnicum Pisidium henslowanum Pisidium moitessieranum

exclude an interpretation of poor ‘slum’ conditions if the rest of the assemblage indicates thus. From the range and associations of taxa encountered in archaeological assemblages from small wetland features, Sparks’ groupings have some merit, not least in reminding us that certain taxa have very wide ecological tolerance and will persist in almost any habitat if they have successfully colonised in the first place. To sum up, the examples discussed here show that it is possible to extract some archaeologically-relevant information from assemblages of freshwater mollusca from small and/or man-made contexts. Simple comparison with modern analogue faunas is

138

Terry O’Connor

Figure 8.5. Gyraulus albus. A tiny planorbid gastropod, almost ubiquitous and of little palaeoenvironmental value

unlikely to be helpful, in part because past conditions may not have modern analogues and in part because the circumstances of colonisation and competition are seldom factored into such comparisons. In practice, the best use of modern analogues may be to ‘dip’ small ponds and ditches, then ask ‘How did those taxa get here in the first place?’, rather than looking for parallels in species composition. Pond- and stream-dipping by malacophiles is very much to be encouraged, in part in order to widen our knowledge of species’ distributions and community associations with particular habitats, and in part because non-marine molluscs are generally under-recorded at a time when freshwater habitats are undergoing rapid and often detrimental changes. Finally, then, how are we to proceed when faced with freshwater mollusca from archaeological deposits? Having considered whether the samples were obtained by core or during excavation, as discussed above, analysis and interpretation of archaeological assemblages needs to proceed contextually. Where possible take into account whatever other evidence may be available, such as the lithology of the sediment and the occurrence of plant macrofossils or Daphnia ephippia. What indications are there that the assemblage is in situ? Are the shells eroded or battered, and are there fragile elements such as juveniles and operculae? Having made our identifications, diversity and equitability may be as informative as the actual species present. How speciesrich is the assemblage and how evenly are numbers distributed amongst the different taxa? An assemblage that is species-rich but with low equitability may be a mixture of an in situ death assemblage (ie, the abundant taxa, especially if with many juveniles) and translocated specimens from other habitat patches. Species-poor assemblages are likely to represent the more challenging ‘slum’ habitats, unless early colonisers such as Galba truncatula, or Radix peregra predominate, which may indicate an early stage in a temporary feature. Are there bivalves? These are likely to be slower to colonise than pulmonate gastropods, and so indicative of longer-lived freshwater features, and some species have very particular habitat preferences. In fact, the presence of Pisidium species other than P. casertanum or P. personatum could be taken to indicate well-oxygenated, non-slum conditions. One regrettable detail of some of the relevant literature, especially

8. Wetlands: freshwater and slum communities

139

Figure 8.6. Just to make the point about the need for comparative specimens of freshwater bivalves, this is only part of a collection of British Pisidium species

grey-literature ‘environmental reports’, is that Pisidium specimens are not taken to species level. With poorly-preserved and juvenile specimens, that is understandable and probably wise, but where adults are available, knowing whether the assemblage contains only P. personatum or a diversity that includes, for example, P. subtruncatum and P. nitidum could make a big difference to the interpretation. Those identifications would require patience and good comparative material, but would be worthwhile in terms of information (Fig. 8.6). Freshwater Mollusca from small and temporary features associated with human activity can be quite problematic in interpretation. The examples discussed here include samples acquired by coring of moats and putative fishponds, and by sampling of waterside archaeological deposits during excavation. Consideration of the mollusca has added something to the interpretation of the features and deposits, adding weight to the interpretation as fishponds, showing phases of reflooding, showing the consequences of refuse disposal. In all such instances, we have to keep in mind that the mollusca will be just one form of evidence, so our interpretation will benefit from such other evidence as may be available. Remember, too, that well-dated and reliably identified records of freshwater mollusca make a contribution to our knowledge of the present-day fauna and to its conservation. Environmental archaeology is as much about the present as it is about the past.

140

Terry O’Connor

References Aho, J., Ranta, E. & Vuorinen, J. 1981. Species composition of freshwater snail communities in lakes of southern and western Finland, Annales Zoologici Fennici 18, 233–241 Boag, D. A. 1986. Dispersal in pond snails: potential role of waterfowl, Canadian Journal of Zoology, 64(4), 904–909 Boycott, A. E. 1936. The habitats of fresh-water Mollusca in Britain, Journal of Animal Ecology 5, 116–186 Carrott, J., Dobney, K., Hall, A. R., Jaques, D., Kenward, H. K., Large, F. & Milles, A. 1993. An evaluation of biological remains from excavations on land to the rear of Gowthorpe, Finkle Street and Micklegate in Selby town centre (site code Selby 1993). Reports from the Environmental Archaeology Unit, York 93(8). York: Environmental Archaeology Unit Darwin, C. 1860. On the Origin of Species. Second edition. London: John Murray Davies, P. 2006. The present and the past: the interpretation of sub-fossil molluscan assemblages and the relevance of modern studies, with specific reference to wet-ground contexts in the UK. In Lillie, M & Ellis, S. (eds), Wetland Archaeology and Environments: regional issues, global perspectives, 173–185. Oxford: Oxbow Books Dobney, K., Fitter, R., Hall, A., Irving, B., Jaques, D., Johnstone, C., Kenward, H. K., Milles, A. & Shaw, T. 1994. Technical report: Biological remains from the medieval moat at Hall Garth, Beverley, North Humberside. Reports from the Environmental Archaeology Unit, York 94, 60. york: Environmental Archaeology Unit Fysher, G. 1925. Yorkshire Naturalists’ Union at Egton Bridge: Mollusca. Naturalist 50, 187–188 Girling, M. A. 1981. The environmental evidence. In Mellor, J. T. & Pearce, T. (eds), The Austin Friars, Leicester, 169–172. London: Council for British Archaeology Research Report 35 Hill-Cottingham, P. 2006. Freshwater Mollusca in Somerset: risk of loss, Proceedings of the Somerset Archaeological and Natural History Society 149, 203–206 Hall, A. R. & Kenward, H. K. 1990. Environmental Evidence from the Colonia. Archaeology of York 14/6. London: Council for British Archaeology Kappes, H. & Haase, P. 2012. Slow, but steady: dispersal of freshwater molluscs, Aquatic Sciences 74(1), 1–14 Kurzawska, A. & Kara, M. 2015. The contribution of mollusc shells to the reconstruction of environment at the Early Medieval stronghold of Pszczew (Poland), Quaternary International 390, 126–132 O’Connor, T. P. 1988. Slums, puddles and ditches: are molluscs useful indicators? In Murphy, P. & French, C. (eds), The Exploitation of Wetlands, 61–68. Oxford: Oxford: British Archaeological Report 186 O’Connor, T. P. & Bunting, M. J. 2009. Environmental change in an Orkney wetland: plant and molluscan evidence from Quoyloo Meadow. In Allen, M. J., Sharples, N. & O’Connor, T. P. (eds), Land and People: papers in memory of John G. Evans, 161–168. Prehistoric Society Research Paper 2. Oxford: Oxbow Books/Prehistoric Society O’Connor, T. P. & Evans, J. G. 2005. Environmental Archaeology: principles and methods. Stroud: Alan Sutton Oertli, B., Joye, D. A., Castella, E., Juge, R., Cambin, D. & Lachavanne, J. B. 2002. Does size matter? The relationship between pond area and biodiversity, Biological conservation 104(1), 59–70 Pip, E. 1986. A study of pond colonization by freshwater molluscs, Journal of Molluscan Studies 52(3), 214–224 Porst, G. & Irvine, K. 2009. Implications of the spatial variability of macroinvertebrate communities for monitoring of ephemeral lakes. An example from turloughs, Hydrobiologia 636(1), 421–438

8. Wetlands: freshwater and slum communities

141

Preece, R. C. 1998. Molluscan conservation: the importance of the fossil record, Journal of Conchology Special Publication 2, 155–164 Robinson, M. 1979. Molluscs. In Robinson, M. & Lambrick, G. (eds), Iron Age and Roman Riverside Settlements at Farmoor, Oxfordshire, 100–102. London: Council for British Archaeology Research Report 32 Rondelaud, D., Hourdin, P., Vignoles, P. & Dreyfuss, G. 2005. The contamination of wild watercress with Fasciola hepatica in central France depends on the ability of several lymnaeid snails to migrate upstream towards the beds, Parasitology research 95(5), 305–309 Smith, K. 1977. The excavation of Winklebury Camp, Basingstoke, Hampshire, Proceedings of the Prehistoric Society 43, 31–129 Sparks, B. W. 1961. The ecological interpretation of Quaternary non-marine mollusca, Proceedings of the Linnean Society of London 172(1), 71–80

Part 2 Palaeo-environmental reconstruction: Europe, the Mediterranean and Near East

9. The southern English chalklands: molluscan evidence for the nature of the post-glacial woodland cover Michael J. Allen

Mollusc analysis not only helps us refine our understanding of environmental change and land-use history, but can also fundamentally change long held perceptions of land-use development, which consequently have major archaeological implications. Although the environmental and land-use history of the downland has now been largely defined by the land snail evidence, in retrospect, it is important to interpret the land snail evidence without reference to wider assumptions of national vegetation histories (eg, Table 9.1). Such assumptions, as we will see later, have held back interpretations over the past 40 years. By default, land snail evidence has been used to understand the land-use history of chalklands in the absence of the peat bogs mires, and waterlogging, and the consequent general absence of the survival of pollen (see Scaife 1987). This has meant that, until the 1970s to 1980s, the chalklands, containing key archaeological landscapes including Stonehenge, Avebury, Dorchester and Maiden Castle and Cranborne Chase, fell behind in environmental landscape and land-use studies traditionally covered elsewhere by the analyses of pollen and waterlogged plant remains. This was exacerbated by the persistence of the belief that they existed as large tracts of open downland and of thin calcareous azonal rendzina soils (typically 0.3 m thick) which were unlikely to have supported ancient deciduous forest. The common concept of the chalk downland was largely one of a beautiful natural landscape, upon which the ‘impression’ of ancient fields and farmsteads was left as a faint, but sometimes extensive, reminder of prehistoric communities. This chapter provides a new interpretation of prehistoric chalkland land-use and vegetation history. I have published overviews previously (Allen & Scaife 2007; Allen & Gardiner 2009), but the molluscan evidence that drives the change in interpretation is discussed here. The chalklands of southern England were perceived by some in terms of a quintessentially rural idyll, with rolling green downland pasture gradually being turned brown in later historic times under the ever-expanding ploughscape. The conundrum

9. The southern English chalklands

145

Table 9.1. Outline of climatic zonation, basic vegetational change and archaeological events for southern England Pollen Climatic Archaeological Climate & approx. date zone Climatic zone

zone period vegetation calibrated Pollen Archaeological Climate & approx. dateBC/ Godwin/West (BP) zone period vegetation calibrated BC/ Godwin/West (BP) FLANDRIAN Deterioration FLANDRIAN Cold and wet,Deterioration general deterioration. High Roman period rainfall. Decline of deterioration. lime. IncreaseHigh of ash, Iron Age VIII Roman Sub-atlantic Cold and wet, general period birch and beechof lime. Increase of ash, Late Bronze Age rainfall. Decline Iron Age VIII Sub-atlantic ----------------------- Fl. III --------------------------------------------------------------------------------------------1250 cal BC birch and beech Late Bronze Age (c. 2900 BP) Stable ----------------------- Fl. III --------------------------------------------------------------------------------------------1250 cal BC Warm and dry, low rainfall, wind-blown Middle Bronze Age (c. 2900 BP) Stable deposits. Woodland regeneration in southern EarlyBronze BronzeAge Age Warm and dry, low rainfall, wind-blown Middle EnglandWoodland regeneration in southern Final Neolithic deposits. Early Bronze Age Sub-boreal VIIb ----------------------------------------------------------------------------------------------3200 cal BC England Final Neolithic Declining warmth. Landnam and first (4500 Late Neolithic Sub-boreal VIIb ----------------------------------------------------------------------------------------------3200 calBP) BC agriculture. Elm decline:Middle Neolithic Declining warmth. Landnam3800 and BC/(5050BP) first (4500 BP) Late Neolithic EarlyNeolithic Neolithic agriculture. Elm decline:- 3800 BC/(5050BP) Middle ---------------------------------------------------------------------------------------------------------------------------4000 cal BC Early Neolithic Optimum (5200 ---------------------------------------------------------------------------------------------------------------------------4000 calBP) BC Climatic optimum, warm and wet. Increase (5200 BP) Fl. II Optimum Later Mesolithic Atlantic Climatic warm and wet. Increase Fl.VIIa II of 2oC,optimum, poly-climax forest. Increase of alder, Later Mesolithic Atlantic VIIa clearances forest. Increase of alder, of some 2oC, poly-climax ---------------------------------------------------------------------------------------------------------------------------6300 cal BC some clearances (7500 Ameliorating ---------------------------------------------------------------------------------------------------------------------------6300 calBP) BC Continental climate, warm and dry. VI (7500 BP) Ameliorating Assynchronous expansions mixed oak Mesolithic Boreal Continental climate, warm andofdry. VI forest with hazel and successional V Assynchronous expansions of mixed from oak pine Mesolithic Boreal -----------------------VFl. I ----------------------------------------------------------------------------------------------8900 cal BC forest with hazel and successional from pine (9500 Rapid Amelioration ----------------------- Fl. I ----------------------------------------------------------------------------------------------8900 calBP) BC Pre-boreal IV Early Mesolithic Sharp increase in Amelioration warmth at 10,000 BP. (9500 BP) Rapid Birch, juniperin+warmth pine woodland Pre-boreal IV Early Mesolithic Sharp increase at 10,000 BP. ---------------------------------------------------------------------------------------------------------------------------10,000 cal BC Birch, juniper + pine woodland ---------------------------------------------------------------------------------------------------------------------------- 10,000 cal BC This enables pollen zones quoted in many specialist pollen and quaternary geography reports to be equated to the archaeological chronology and in activity 2000b; Allen Gardiner 2009) This enables pollen zones quoted many(Allen specialist pollen and&quaternary geography reports to be equated to the archaeological chronology and activity (Allen 2000b; Allen & Gardiner 2009)

was whether this downland, with its thin azonal rendzina soils ever supported the post-glacial woodland of, initially, Mesolithic (Boreal) pine and hazel (the former can hardly survive, let alone thrive on the thin chalkland soils today), to Neolithic mixed oak forests (see Tables 9.1 and 9.5). Despite the fact that this vegetational succession was seen across most of northwest Europe, archaeologists persisted in considering the downland as open, and subconsciously, but ironically as we will see later, this openness was seen in part to contribute to the Wessex chalkland becoming the centre of Neolithic and Bronze Age culture (Piggott 1954). Surprisingly, initially, many archaeologists and geographers alike considered the chalk Downs to have always been open grassland; even after the 1950s when myxomatosis almost destroyed the rabbit population allowing hawthorn shrubs and other woody vegetation to rapidly take hold (Sumption & Flowerdew 1985; Thomas 1960; 1963).

Michael J. Allen

146

Table 9.2. The archaeological data selected are from Early Neolithic sites (long barrows and causewayed enclosures) Author Archaeologist Childe

Year

Site

Data

Grassland

Scrub

Woodland

1925 1957 1937 1954

– – – –







– –

   –

– –

?









charcoal







snails Snails Snails Snails Snails Snails Snails

– – –   – 

–  –    



Curwen Piggott Ecologist Godwin & Tansley 1941 – Environmental Archaeology Salisbury & Jane 1940 Maiden Castle Conchologist Kennard & Woodward 1929 The Trundle Kennard & Woodward 1930 Whitehawk Kennard 1934 Whitehawk Woodward 1936 Whitehawk Kennard 1936b Whitehawk Kennard 1936a Thickthorn Kennard 1943 Maiden Castle Key:  = refuted,  = present; ? = ambiguous

 ? ? ? ?

‘The problem of the south English chalk downland’ (Piggott 1954, 5) Our understanding of the chalk landscape was tardy in comparison with many other areas of Britain; principally due to poor pollen preservation and the relatively late onset of soil (as opposed to peat) pollen studies (Dimbleby 1957; 1961; 1985). A lack of long environmental histories from the downland encouraged many prehistorians, geographers and naturalists to assume these landscapes to have been essentially unchanging – and to project the modern landscape back into prehistory (Table 9.2). In 1925 when Gordon Childe was writing The Dawn of European Civilisation (1925; with revised editions to 1957), throughout his text on the Neolithic of the Wessex and Sussex Downs there was an unwritten assumption that the Downs were open grassland. Similarly, Cecil Curwen, a Sussex archaeologist, wrote in 1937 in the first edition of his Archaeology of Sussex: ‘in prehistoric times both the western and eastern [South] Downs were in all probability open grassland with a variable amount of scrub’ (p. 13) – clearly projecting current perceptions of landscape into the past. This view remained unchanged in his 1954 edition. However when Stuart Piggott published The Neolithic Cultures of the British Isles in the same year (1954), there was a growing realisation among archaeologists that the chalk, like the rest of the European landscape, may also have been wooded after the last glaciation. But the evidence was elusive: a lack of waterlogged deposits and peat containing pollen data led Piggott (among others) to believe that chalklands constituted a ‘special problem’ (1954, 5). At this time land snail

9. The southern English chalklands

147

evidence was not considered by many to be of great help as it was largely deemed to examine general ‘climatic environments’ rather than vegetation and land-use histories (cf. Evans 1972; Davies 2008). Confusion reigned, with the evidence from charcoal (Maiden Castle, Dorset) being used to suggest that in Neolithic times the chalk was covered in closed woodland (Salisbury & Jane 1940) whilst the famous palaeobotanists and ecologists Godwin and Tansley (1941) contended ‘no sound evidence is presented for any such belief’. Like the archaeologists before them they also seemed to believe that the downland was essentially bare of trees in the Neolithic. Early land snail studies, largely by A. S. Kennard (eg, 1933; 1935; 1936a; 1943), provided an independent and unquestionably local interpretation, even though the discipline was clearly in its infancy (Tables 9.3 & 9.4). Kennard suggested that the snail evidence indicated climatic and vegetation associations local to the sampling sites, and despite the lack of corroborative botanical evidence, he suggested these provided evidence of downland woodland (Table 9.2). Astonishingly, for the time (1934, 130), Kennard wrote from his analysis at the early Neolithic causewayed enclosure at Whitehawk, Brighton, East Sussex: ‘It is obvious that the ecological conditions were very different from those of the present day … The faunule is that of damp woodland or scrub, and these conditions must have existed on the Downs when the Camp was occupied’. He went on to presciently say ‘It would be of great interest if we could date when the damp period [woodland/scrub] ended … We now know that in Wiltshire it had practically ended in the Middle Bronze Age period.’

This latter question is still being posed with as much clarity 50 years later. Subsequently, pioneering work by Godwin (1962) on peat adjacent to the Downs at Wingham and Frogholt in Kent indicated the existence of woodland and subsequent deforestation by human agencies, rather than natural causes. Important studies of peat in river valleys by Anne Thorley at Amberley Wild Brooks, Arundel and Vale of Brooks, Lewes (1971; 1981), and also by Waton at Winnal Moor, Winchester (1982; 1986), and latterly by Waller (Waller & Hamilton 2000), demonstrated the presence of post-glacial woodland on the Downs, and confirmed that the vegetation history followed the general pattern seen elsewhere in Britain (Table 9.1). Archaeologists and ecologists initially denied the presence of woodland on the Downs; not being able to conceive that the thin soils could support this vegetation despite the detailed contrary reporting by conchologists in the archaeologists’ own excavation reports. Kennard and Woodward, from 1929 onwards, had indicated that, in the Early Neolithic, the downland environment associated with long barrows and causewayed enclosures was a) not grassland, and b) that was mesic (damp) environments existed supporting scrub, or c) woodland. These interpretations were from sites spread across the Downs from Maiden Castle and Thickthorn Down in west Dorset, to The Trundle and Whitehawk in West and East Sussex respectively. Despite the strong indications from snails (analysed by Kennard and Woodward, see Tables 9.3 & 9.4) and charcoals (eg, Salisbury & Jane 1940) and the plain, clear prescient words of Kennard (1934, 130),

Michael J. Allen

148

Table 9.3. Kennard’s records (1936; 1943) from Neolithic contexts in Dorset Thickthorn Down long barrow Feature Context Sample Wt (g)

MOLLUSCA Pomatias elegans (Müller) Carychium tridentatum (Risso) Cochlicopa cf. lubrica (Müller) Pupilla muscorum (Linnaeus) Vallonia costata (Müller) Vallonia extrentrica/pulchella Acanthinula aculeata (Müller) Ena montana (Draparnaud) Punctum pygmaeum (Draparnaud) Discus rotundatus (Müller) Arion spp. Vitrina pellucida (Müller) Vitrea crystallina (Müller) Aegopinella pura (Alder) Aegopinella nitidula (Draparnaud) Oxychilus cellarius (Müller) Limacidae Euconulus fulvus (Müller) Cecilioides acicula (Müller) Clausilia bidentata (Ström) Helicella itala (Linnaeus) Trochulus striolatus (Linnaeus) Trochulus hispidus (C. Pfeiffer) Arianta arbustorum (Linnaeus) Helicigona lapicida (Linnaeus) Cepaea nemoralis (Linnaeus) Taxa TOTAL

Buried soil

Neolithic mound

ditch

1

3

4

2

1ry 6

1 – 1 – 5 5 – – – – – – – – – – – – – – – – – – – 18 5 30

– – – 5 6 17 – – – – – – – – – – 1 – – – – – – – – 2 5 31

+ – 1 – 5 3 – – – – – – – – – – – – – – – – – – – 26 5 35

3 – – – – – – – – 2 + – 4 – – 1 – – (2) – – – – 1 – – 7 11

– – – – 1 – – – – – – – – – – – – – – – 1 – – – – – 2 2

posth ole

Beaker ditch

8

secondary 5 7

5 – – 1 1 – – – – – – – – – – – – – – – 1 – – – – 2 5 10

1 24 7 6 11 – 4 + 2 28 + 2 18 4 8 – 1 – – 5 2 – 5 – 2 34 20 164

212 1 4 6 3 1 3 – 1 10 + – 4 1 1 – 1 1 – 1 16 – 7 1 1 4 21 279

Maiden Castle

– – – – – – – – – 11 – – 1 – 14 39 – – – – 4 1 5 – – 3 8 78

Thickthorn Down long barrow and Maiden Castle causewayed enclosure. Thickthorn Down: 1, 3, 4 turf line/buried soil; 2 mound; 6 ditch layer 5 (primary fill); 8 posthole A; 5, 7 ditch layer 2 (secondary fill) Although in the publication Kennard records Carychium mimimum Mull (1936, 94), this is probably C. tridentatum, so is recorded as such in Tables 9.2 and 9.3

both archaeologists, and more significantly leading palaeo-ecologists such as Godwin as late as 1940 and beyond (1940; et seq. see Godwin 1975), failed to acknowledge or accept this. Interpretation and research were probably hindered by this failure. Michael Kerney and his students, when studying land snails to look at late glacial and early Holocene environments from long stratified colluvial valley fills in Kent and Buckinghamshire

9. The southern English chalklands

149

Table 9.4 Woodward and Kennard’s records of land snails from Neolithic deposits in Sussex Published MOLLUSCA Pomatias elegans (Müller) Acicula fusca (Montagu) Carychium tridentatum (Risso) Cochlicopa cf. lubrica (Müller) Vertigo pygmaea (Draparnaud) Pupilla muscorum (Linnaeus) Vallonia costata (Müller) Vallonia excentrica Sterki Vallonia extrentrica/pulchella Acanthinula aculeata (Müller) Merdigera obscura (Müller) Discus rotundatus (Müller) Arion spp. Vitrea crystallina (Müller) Nesovitrea hammonis (Ström) Aegopinella pura (Alder) Aegopinella nitidula (Draparnaud) Oxychilus cellarius (Müller) Limacidae Cochlodina laminata (Montagu) Clausilia bidentata (Ström) Helicella itala (Linnaeus) Trochulus striolatus (C. Pfeiffer) Trochulus hispidus (Linnaeus) Helicodonta obvoluta (Müller) Arianta arbustorum (Linnaeus) Helicigona lapicida (Linnaeus) Cepaea nemoralis (Linnaeus) Cepaea hortensis (Linnaeus) Taxa

Trundle 1929

Whitehawk 1930 1934 1936

 –  – –   – –        

c 1 c c – 8 c 3 2 3 – c c c 5 4 c

c – c r – r r r – r – r r r r – r

c vr vr vr vr c r r – vr – – vr – vr vr vr

 – –          23

c – 4 6 2 4 8 – c 1 c c 25

r – r r r r r – c – c c 21

vr vr – – – vr c – r – c c 20

Key c = common, r = rare; vr = very rare The Trundle and Whitehawk causewayed enclosures (Kennard & Woodward 1929; 1930; Woodward 1934; 1936)

(Kerney et al. 1964; Evans 1966), clearly showed that post-glacial woodland developed on the chalk. From a very early stage in this research (eg, Kerney 1968) he linked this to Scandinavian botanists Blytt and Sernander’s general climate and vegetation history (Table 9.1) promoted by palaeo-ecologists (eg, Godwin 1940), and palynologists across northwest Europe (Lowe & Walker 1982, 132–133). Such still was the confusion, perhaps

150

Michael J. Allen

Table 9.5. Main environmental events on the chalk of Wiltshire, Dorset and Sussex in the lateglacial and post-glacial periods Period

Medieval/RomanoBritish/Iron Age Bronze Age

Late Neolithic Late Neolithic Neolithic Neolithic (Mesolithic) Atlantic (?) Mesolithic Boreal (?) (Upper Palaeolithic) Late-glacial

(Evans & Jones 1979; Entwistle & Bowden 1991)

Environment

Intermittent cultivation & grassland. Formation of ploughwash deposits. Open environment of grassland or arable. Cultivation/grazing intermittent. Formation of windlain material Woodland regeneration. Not at Woodhenge Construction of henge monuments Long period of grassland, probably maintained by grazing. Woodland clearance. Ploughing & possibly other forms of tillage. Ploughmarks at Avebury. Dense woodland. Recorded only at Avebury, but probably at most sites. Open woodland. Evidence of fire & possible influence of Mesolithic man (Evans 1972, 219, 256). Subarctic environment, probably tundra. Formation of periglacial structures & wind-lain material.

Allen

Essentially the creation of open townland, extensive downland pasture an&d increasing tillage and formalised field systems. Opening of land & field leads to colluviation (& wind blow) Extensive tracts of open grassland, managed woodlands, & possible expansion in localised tillage Expansion of open grassland & graze, localised tillage Progressive woodland modification & woodland clearance; pasture being more extensive, localised tillage Mosaic of woodland, varying from dense closed canopy, to light open scrub, & some extensive almost open grassland glades Open woodland, including pine & hazel (Allen 1995), localised influence if fire & Mesolithic activity

From Evans and Jones (1979, 209) modified by Entwistle and Bowden (1991, 41), and as presented here (based on Allen & Gardiner 2009)

largely residing in the archaeological community (see Curwen 1954 vs Piggott 1954), that it was then that one of Kerney’s students took up the baton to specifically research the vegetation (and land-use history) of the downland from archaeological contexts. And so it was that John Evans undertook his PhD (1964–1967) and post-doctoral (1967–1969) research, culminating in the seminal publication of Land Snails in Archaeology (1972). Evans clearly demonstrated the presence of post-glacial woodland, and its local anthropogenic clearance for Neolithic monument building and activities (long barrows and causewayed enclosures). John Evans continued work, sampling widely from archaeological monuments, reinforced the presence, and then clearance, of woodland across the downlands (eg, Evans 1971; Evans & Jones, 1979, 209; see Allen 2006). Thus,

9. The southern English chalklands

151

Figure 9.1. Reconstruction and interpretation of Windmill Hill early Neolithic causewayed enclosure set within a woodland clearing (Whittle et al. 1999), base reproduced by kind permission of Josh Pollard (illustration: Abby George) from the 1970s the existence of an essentially continuous woodland blanket across the Downs (in common with the rest of the northwest Europe) was widely accepted; a woodland that required removal for monument building, for settlement, and for farming (Table 9.5). This clearly seems to have been the case for Neolithic causewayed enclosures dating to the 36th century BC in Sussex (eg, Offham, Bury Hill, The Trundle, and Coombe Hill), and Windmill Hill, Wiltshire (Fishpool 1999; Fig. 9.1). This, however, provided an assumption of an essentially continuous unbroken woodland blanket that had developed more or less uniformly not just across the Downs but the whole of northwest Europe. Even Evans failed really to engage with the type of variations in woodland, but instead concentrated on change and opening of the woodland, clearance, tillage and grazed grassland; see South Street, Beckhampton, Horslip and Easton Down analysed from 1967 – c. 1990 (Fig. 9.2), and even Ascottunder-Wychood, re-appraised in 2003–2004. This assumption too hindered research. Woodland seems to have been just as ‘present’, and its uniform distribution across the whole landscape was taken as unquestioned; the nature, distribution and diversity of that woodland was not really explored. The lack of woodland at an archaeological site was seen to confirm its anthropogenic clearance and removal, its potential absence was not questioned. The lack of evidence for early woodland on any archaeological site was taken as sample location bias; after all archaeological excavations were located specifically where there was evidence of human activity, and therefore, woodland clearance! This leads us to briefly examine two prehistoric downland ecologies; the nature of the post-glacial woodland, and downland openness.

152

Michael J. Allen

Figure 9.2. Land snail histograms diagram of buried soils under key Neolithic long barrows (illustration: Abby George)

9. The southern English chalklands

153

Evidence of prehistoric post-glacial woodland – wanting to conform There are clear long records of environmental change from stratified colluvial deposits at Brook, Kent (Kerney et al. 1964; 1980), Pitstone, Buckinghamshire (Evans 1966), and Asham, East Sussex (Ellis 1985; 1986) for example; the land snail sequences of which were clearly related to phases of Boreal, Atlantic and sub-Atlantic vegetational succession. Here, however, I wish to concentrate on Neolithic land snail sequences recovered from archaeological sites, and more specifically from buried soils beneath Neolithic long barrows as these represent not only the first monuments, but their construction also seems to represent a short time period of about a one to one-and-a-half centuries (Bayliss & Whittle 2007). These land snail analyses we may think are now common, but the evidence is still relatively sparse. Many major long barrows were excavated prior to land snail analysis (Nutbane, Hampshire, and Fussell’s Lodge, Wiltshire), and others such as North Marden, West Sussex whilst excavated more recently, were so heavily ploughed that no buried soil survived. The long barrows rapidly considered here are (Table 9.2): South Street, Beckhampton Road, and Horslip, Wiltshire (Ashbee et al. 1979) West Kennet, Wiltshire (Evans 1967; 1972) Easton Down, Hampshire (Whittle et al. 1993) Amesbury 42, Wiltshire (writer) Alfriston, East Sussex (Drewett 1974) Ascott-under-Wychwood, Oxfordshire (Benson & Whittle 2007)

West Kennet, Avebury environs, Wiltshire Only three samples were taken from a small cutting exposing the buried soil (Evans 1972, 263–264). The basal sample was dominated by ‘other shade loving species’ and Pomatias elegans with Vallonia costata and V. excentrica; a very mixed assemblage with c. 30% shade-loving species and c. 25% open country. Evans suggested that the sequence of three samples showed the development ‘of dry open grassland from one previously shaded’ but the ‘previously shaded’ environment may well not have been dense deciduous Atlantic forest, but open woodland with a grassy sward. Only three samples have ever been examined from the buried soil, and again clearly this site offers much more potential.

Easton Down, Avebury environs, Wiltshire The pre-barrow environmental history indicates woodland, followed by clearance, cultivation, and ending in grassland (Rouse & Evans 1993, 211), which like South Street suggests significant pre-barrow activity in terms of both events and duration (Evans in Ashbee et al. 1979). The soils, even in a shallow subsoil hollow are typical humic rendzinas (Macphail in Whittle et al. 1993, 218–219), and the buried soil itself is a typical humic rendzina, from which pollen suggests open downland with hazel scrub (Cruse in Whittle et al. 1993, 219–221).

154

Michael J. Allen

Sequences of 11 and 26 samples were taken through the buried soil. The body or base of the soil and subsoil hollow, however, is dominated by land snail assemblages characteristic of relatively closed woodlands with, in part, a deciduous leaf litter floor. The only open country species of any significance was V. costata – an open country species occasionally found in woods (Evans 1972, 156). The land snail assemblages were high diversity and dominated by Discus rotundatus, and Carychium tridentatum with Vitrea spp., Aegopinella spp, and Oxcyhilus cellarius. These were accompanied by low levels of Cochlodina laminata, Ena montana, Merdigera obscura, Clauslia bidentata, Acanthinula aculeata and included Vertigo pusilla and Lauria cylindraea. The only other significant species was the omnipresent Trochulus hispidus.

Alfriston, East Sussex A very poorly preserved plough-damaged buried soil produced three depauperate land snail assemblages (16–21 shells) from which little can be said (Thomas 1979), except that they were generally mixed and contained both open country species (including the xerophile Helicella itala) and only two, common, shade-loving species (Discus rotundatus and Clausilia bidentata); hardly a dense closed deciduous woodland.

Ascott-under-Wychwood, Oxfordshire The buried soil was a relict argillic brown earth (brown forest soil) from which a sequence of five samples was taken (Evans 2007). A woodland fauna was recorded in the lower four samples, which was superseded by an open grassland one in the last, uppermost, sample. The assemblages are moderately species-diverse with relatively uniform representation of each species present (Evans 2007, fig. 3.1, table 3.1); again suggesting good woodland cover, possibly little leaf litter, and open habitats. Again the only significant open country species present is V. costata, and Evans records the environment here as open woodland (Evans 1972, 254, fig. 88). This woodland follows from, and succeeds, an earlier clearly closed woodland from a Mesolithic tree hollow.

Horslip, Avebury environs, Wiltshire A very shallow, almost stone-free, black rendzina buried soil only 125–150 mm thick was present producing an essentially open country grassland fauna; the lowest samples (A/C horizons) of which have hints of a former more shady regime (Evans in Ashbee et al. 1979, 275–276).

Beckhampton Road, Avebury environs, Wiltshire The buried soil over the coombe deposits was essentially a grey calcareous rendzina

9. The southern English chalklands

155

with, on the coombe rock and brickearth, a relict Bt horizon of an argillic brown earth. Three of four spot samples from the buried soil were dominated by shade-loving species with one open country assemblage (Evans in Ashbee et al. 1979, 278–298; Evans 1972, 248). Although the shade-loving assemblages were relatively diverse with 13 species present, they were accompanied by open country species (Vallonia costata and V. excentrica, Helicella itala, Pupilla muscorum), and again suggest a woodland with more light, and openings, and possibly some grassy or herbacous swards.

South Street, Avebury environs, Wiltshire A rendzina soils exists over a subsoil hollow; it is a complex soil with ard marks at the base scoring the chalky periglacial drift. The soil has a history of late glacial cold stage (tundra) to a clearly defined open woodland (Evans 1972, 257–261, fig. 80; Evans in Ashbee et al. 1979, 282–285; Evans 1971, 40–52). Evans struggles with the openness here, knowing that a more closed woodland existed on the chalk in the Avebury area (1972, 257), and suggested that the open country species imply either ‘a cover of light woodland or that the deposits are contaminated with late glacial shells’. Although shade-loving species (dominated by D. rotundatus, C. tridentatum, with Aegopinella spp., and Oxychilus spp, small occurrences of Acanthinula aculeata, Clausilia bidentata. Vitrea contracta, Vitrina pellucida, Euconulus fulvus and the ancient woodland rarities Vertigo pusilla and Ena montana) represent the woodland, the significant presence of V. costata and V. excentrica with Helicella itala, Pupilla muscorum and Vertigo pygmaea, clearly represent a significant openness, perhaps short grazed small glades.

Amesbury 42, Stonehenge environs, Wiltshire The thin grey to brown silty rendzina 240 mm thick under the relict bank was sampled in a small test pit by the writer. The land snail assemblages were almost complete devoid of shade-loving species, and were dominated by the Vallonia species with Helicalla itala, and Pupilla muscorum; clearly open dry, probably short grassland, environment. Even samples from the ditch showed no evidence of shady habitats until well into the secondary fill (ie, much later in prehistory).

The Neolithic post-glacial woodland Many palaeo-geographers considered post-glacial vegetational succession in terms of a simple, and more or less uniform, history from a cold and open landscape, through mixed oak forest, followed by the clearance of the woods to create the present open downland. That sequence has been well rehearsed in vegetation (or climatic) terms (eg, Table 9.1). Without going into more detail, what is clearly evident here, is that the postglacial woodland cover of the 38th–37th centuries BC was not a uniform dense, Atlantic

156

Michael J. Allen

Figure 9.3. Vera’s model, consisting of the three phases of Open Park, Scrub and Grove (Vera 2000), to which a fourth Break-up has been added by Kirby (2003) to represent the transition from woodland grove back to open habitats ‘Park’ (from Allen & Gardiner 2009; illustration: Abby George) closed-canopy forest, but a much greater mosaic of woodland. Some was closed-canopy ancient woodland with leaf litter, other areas seem to be more scrubby, whilst others represent significant openness varying from glades within woodland (South Street), to very open established grassland by the 37th century BC (eg, Amesbury 42). The very nature of the variable but varying woodland habitat would also effect the distribution of edible plants and berries, and of animals (herds of deer and wild cattle etc), and thus may directly or indirectly affect the location of human activities, or choices to occupy particular parts of the woodland landscape.We may ponder further on this woodland mosaic and diversity. The post-glacial openness is discussed below, but woodlands were clearly spatially diverse (the ecological norm). Those areas that were more open, allowed more light, and were more accessible, might have been preferentially browsed by deer, and in particular cattle (cf. Vera 1997; 2000; Fig. 9.3). This might then create a greater division, or contrast, between a denser woodland with a closed-canopy that was not browsed, and a more open woodland in which browsing and opening occurred; essentially creating an almost bipartite woodland. A greater range and variety of human activities are more likely to have been performed in, and focussed upon, the more open woodland. Clearing and preparing areas for activities, even for animal slaughter and butchery from the Mesolithic, may have been concentrated in these more open areas, and thus the concentration of human activities, establishing a sense of place, and of repeated visiting or meeting and ultimately monument building, may essentially fossilise an ecological pattern of diversity. How much is the distribution of early Neolithic (or even Mesolithic) activity predicated upon the nature of the vegetation

9. The southern English chalklands

157

and type of woodland, and how much is more one of human decision, whether based on topographical, hydrological, social or personal considerations? Or put another way how much does the distribution of human activity reflect the former woodland ecology?

How open was the downland? The accepted model of post-glacial vegetation history follows Tansley’s ecological developmental progression (1939) through vegetation seres to climax woodland. Evans’ general land-use scheme based on land snails (Table 9.5), is undoubtedly applicable at the landscape or regional scale, as he intended. But our unquestioning assumption that woodland existed, and then was cleared, prevents us from being able to challenge or modify this model on a local, site by site basis. It is perhaps necessary to prove the existence of closed woodland, rather than to assume it. Prehistoric populations may have deliberately exploited niche habitats and more open ground. The presence of extensive post-glacial woodland on the chalk was challenged as a result of pollen analysis from mires on the Yorkshire Wolds (Bush 1988; 1989; Bush & Flenley 1986); but the application of this model to the southern chalk Downs was vigorously refuted by Thomas (1989). In the past 25 years work, largely by Rob Scaife, has provided a series of important, and key pollen sequences, such as Mesolithic contexts at Stonehenge (Scaife 1995), Neolithic deposits at Gatcombe Withy Beds, Isle of Wight (Scaife 1980; Tomalin & Scaife 1979) and alluvial sequences at Durrington (Scaife 2004) and Cranborne Chase (Scaife in French et al. 2003; 2005; 2007). Combined with this, a significantly increased level of land snail analysis has started to look at spatial resolution and mapping of past environments within landscapes rather than considering single sites or sample points to be wholly representative of their landscapes. This has been facilitated by commercial archaeology (eg, Dorchester By-pass) and personal research involving very detailed molluscan analysis. Studies of defined chalkland research areas have thus enabled us to move from a two dimensional history to the postulation of a three dimensional mapped history of vegetation and land-use (Allen 2000a), as at Dorchester (Allen 1997a), Stonehenge (Allen et al. 1990; Allen 1997b) and Cranborne Chase (Allen 2002; Allen in French et al. 2003; 2005; Allen 2007). If we examine chalkland landscapes by attempting to prove the existence of woodland, rather than assuming it, then the examination of selected, well-studied areas is informative. In the Dorchester landscape to the north of Maiden Castle, analysis of over 260 snail samples from nine prehistoric sites (Allen 1997a), with over 75,000 snail identifications, failed to prove conclusively the existence of prehistoric closed woodland. Perhaps the sampling was biased in favour of archaeological sites where woodland was cleared. But it is also possible that a full post-glacial woodland maximum did not exist. Yet at Maiden Castle itself John Evans (Evans & Rouse 1991) clearly demonstrated the presence of an ancient woodland. So – contrary to Evans’s suggestion that this was representative of the landscape – perhaps Maiden Castle survived as a wooded hilltop on open downland, as Danebury does today? Mesolithic posts (around 7000 cal BC) were erected 250 m north of Stonehenge in an open pine and hazel woodland, and the

158

Michael J. Allen

first earthen, monument (c. 3700 cal BC) of Amesbury 42 long barrow, as we have seen, was constructed in a pre-existing, long-established and expansive open downland. On Cranborne Chase, an even more extensive land snail programme of over 400 samples from 22 sites and over 175,000 snail identifications in an area smaller than that studied at Dorchester has also failed to find evidence for extensive closed post-glacial woodland, a finding confirmed by pollen analysis from peats in the chalkland valley at Wimborne St Giles (Scaife in French et al. 2007, figs 2.19 & 2.20). Evidently woodland development was retarded in some locations. It seems increasingly likely that natural openings or glades comprising grassland, scrub and some trees formed a part of the natural ecological variation and patchiness in a large and extensive forest. Some of these relatively open areas may have been as large as an English parish. These would invite vegetation diversity, the woodland fringes providing a niche for fruiting trees and berries. These fruits would have encouraged a range of herbivores to feed and browse and thus maintain the glade. Such a location would also invite Mesolithic communities to hunt and gather, and possibly even maintain these glades artificially and perpetuate them. It is perhaps not too far-fetched to suggest that these areas attracted some of the first large human populations and that in consequence they survive today as zones of more concentrated monuments and sites. It is no coincidence that three of the apparently most densely populated parts of the prehistoric downland, all containing major earlier Neolithic monuments (Dorchester, Cranborne and Stonehenge), all seem to have been open, not closed, wooded downland. This concept has been explored further for the southern chalkland (Allen 2000b; French et al. 2007), see below. Ironically, these ideas of open park woodland propagated initially by Bush (1988) and Bush and Flenley (1986) for the Yorkshire Wolds were independently postulated for areas of the Dorset Downs from 1988 to 1997 (see Allen 1997a, 278). Reassuringly these ideas conform directly to ecological models developed by Vera (1997; 2000), that landscape mosaics existed in which ‘half-open’ and park-like landscape existed in lowland areas, and were maintained by wild herbivores. We postulate that precisely these areas were exploited by Mesolithic populations because of their openness, wild fruit and berries, and the concentration of herbivores (cf. Allen 2002). Evidence of open park woodland and these man and nature interactions are explored further in Allen (1997b; Allen & Gardiner 2009), Scaife in French et al. (2003) and French et al. (2005; 2007). Work by environmental archaeologists over the past 25 years has radically altered this perception. This essay provides a new narrative for the downland. It includes evidence for pine as a former constituent of the post-glacial vegetation. We now know that post-glacial forest did not form a uniform blanket, and we understand better the consequences of the varied tree cover for prehistoric societies. We attempt here to summarise the model developed by John Evans for chalklands (Evans & Jones 1979; Table 9.5), and present some new thoughts on both the natural changes, and the impressive consequence of human activities which essentially in prehistory created the chalkland landscape we recognise today.

9. The southern English chalklands

159

Figure 9.4. Archaeological and ecological contrasts between the Wessex chalk and the South Downs (from Allen & Gardiner 2009)

Concluding thoughts Land snail evidence, and the careful re-examination of classic published land snail analyses has allowed us to, slightly more critically or objectively, re-appraise some of the downland vegetation history. The Downs were not a simple uniform landscape over which a single development history can be charted. Two major conclusion can be drawn. 1. Although widely accepted that the Downs were covered with a post-glacial woodland, it is now clear that this cover was not continuous, ubiquitous nor complete in either space nor over time. Areas of the Wessex Downs had large areas of post-glacial grassland or open land rather than woodland. These have been identified in the environs of Dorchester, Dorset, Cranborne Chase, Stonehenge, and possible the lowland area around Avebury (Fig. 9.4), and these are, not co-incidentally the locations of major centres of Mesolithic activity and flint scatters and/or Neolithic activity and monuments. Previously I have suggested that these areas became foci of human activity (Allen & Gardiner 2009); being more open they encouraged the growth of vegetation such as hazel and fruiting berries, which in turn attracted wild animals such deer, cattle and such like. Consequently, these areas had both food and other resources, ideal for human exploitation from almost the earliest post-glacial times. 2. In some cases, previously, just the identification of the existence of a previous woodland was all that seemed required; and the nature of that woodland cover was not really interrogated, examined or discussed in detail. It is clear that the woodland was very variable, and represents a mosaic over space and time – an ecological norm. Areas of closed-canopy woodland with vegetation-poor ground cover and rich leaf litter, and elsewhere more open woodland with herbaceous vegetation and open woodland with scrub refugia, rupestral habitats and open grassland may have existed.

What is clear about these two conclusions is that they may have a significant, if not profound, effect upon human activity, occupation and settlement (see Allen & Gardiner

160

Michael J. Allen

2009). Not only may the distribution of open areas, or more open woodland vs closed denser woodland effect the distribution of human activity at both a local site scale, but also at, potentially a regional scale (Fig 9.4), and also may affect the range of activities that could be, and were, undertaken. We must remember that the presence of woodland does not preclude many activities; flint mining, for instance can easily occur within woodland, as can grazing and browsing, specially of cattle. Archaeologists’ perception too may be challenged. Instead of considering the removal of woodland to create a space and feeling of openness and huge areas for graze, in some cases it may be to enable activities, browse, or even monument building within the woodland (Allen & Gardiner 2012); after all the woodland ecology is far richer and more diverse in terms of food and economic resources than that of open chalk grassland.

References Allen, M. J. 1995. Before Stonehenge. In Cleal, R. M. J., Walker, K. E. & Montague, R., Stonehenge in its Landscape: twentieth-century Excavations, 41–63. London: English Heritage Archaeological Report 10 Allen, M. J. 1997a. Landscape, land-use and farming. In Smith, R. J. C., Healy, F., Allen, M. J., Morris, E. L., Barnes, I. & Woodward, P. J., Excavations along the Route of the Dorchester By-pass, Dorset, 1986–8, 277–283. Salisbury: Wessex Archaeology Report 11 Allen, M. J. 1997b. Environment and land-use; the economic development of the communities who built Stonehenge; an economy to support the stones. In Cunliffe, B. & Renfrew, C. (eds), Science and Stonehenge, 115–144. Oxford: Proceedings of the British Academy Allen, M. J. 2000a. High resolution mapping of Neolithic and Bronze Age landscapes and landuse; the combination of multiple palaeo-environmental analysis and topographic modelling. In Fairbairn, A. S. (ed.), Plants in Neolithic Britain and Beyond, 9–26. Neolithic Studies Group Seminar Papers 5. Oxford: Oxbow Books Allen, M. J. 2000b. Soils, pollen and lots of snails. In Green, M. G., A Landscape Revealed; 10,000 years on a chalkland farm, 39–46, Stroud: Tempus Allen, M. J. 2002. The Chalkland Landscape of Cranborne Chase: a prehistoric human ecology, Landscapes 3, 55–69 Allen, M. J. 2006. Professor John Gwynne Evans, aka ‘snails’ Evans, Journal of Conchology 39, 111–117 Allen, M. J. 2007. Land use and landscape development: the molluscan evidence. In French, C., Lewis, H., Allen, M. J., Green, M. Scaife, R. G. & Gardiner, J. 2007. Prehistoric Landscape Development and Human Impact in the Upper Allen Valley, Cranborne Chase, Dorset, 151–189. Cambridge: McDonald Institute Monograph Allen, M. J., Entwistle, R. & Richards, J. 1990. Molluscan studies. In Richards, J. C., The Stonehenge Environs Project, 253–258. London: English Heritage Archaeological Report 16 Allen, M. J. & Gardiner, J. 2009. If you go down to the woods today; a re-evaluation of the chalkland postglacial woodland; implications for prehistoric communities. In Allen, M. J., Sharples, N. & O’Connor, T. P. (eds), Land and People; papers in memory of John G. Evans, 49–66. Prehistoric Society Research Papers 2. Oxford: Oxbow Books/Prehistoric Society Allen, M. J. & Gardiner, J. 2012. Not out of the woods yet: some reflections on Neolithic ecological relationships with woodland. In Jones, A. M., Pollard, J., Allen, M. J. & Gardiner, J. (eds), Image, Memory and Monumentality: archaeological engagements with the material world, 93–107. Prehistoric Society Research Papers 5. Oxford: Oxbow Books/Prehistoric Society

9. The southern English chalklands

161

Allen, M. J. & Scaife, R. 2007. A new downland prehistory: long-term environmental change on the southern English chalklands. In Fleming, A. & Hingley, R. (eds), Prehistoric and Roman Landscapes; landscape history after Hoskins, 16–32. Macclesfield: Windgather Press Ashbee, P., Smith, I. F. & Evans, J. G. 1979. Excavation of three Long Barrows near Avebury, Wiltshire, Proceedings of the Prehistoric Society 45, 207–300 Bayliss, A. & Whittle, A. (eds), 2007. Histories of the dead: building chronologies for five southern British long barrows, Cambridge Archaeological Journal 17(1) (Supplement) Benson, D. & Whittle, A. (eds), Building Memories; the Neolithic Cotswold long barrow at Ascottunder-Wychwood, Oxfordshire. Oxford: Oxbow Books Bush, M. B. 1988. Early Mesolithic disturbance: a force on the landscape, Journal of Archaeological Science 15, 453–462 Bush, M. B. 1989. On the antiquity of British grasslands: a response to Thomas, Journal of Archaeological Science 16, 555–560 Bush, M. B. & Flenley, J. R. 1986. The age of the British chalk grasslands, Nature 395, 484–485 Childe, V. G. 1925. The Dawn of European Civilisation. London: Routledge & Kegan Paul Curwen, E. C. 1937. The Archaeology of Sussex. London: Methuen Curwen, E. C. 1954. The Archaeology of Sussex. London: Methuen. Second edition Davies, P. 2008. Snails: archaeology and landscape change. Oxford: Oxbow Books Dimbleby, G. W. 1957. Pollen analysis of terrestrial soils, New Phytologist 56, 12–28 Dimbleby, G. W. 1961. Soil pollen analysis, Journal of Soil Science 12, 1–11 Dimbleby, G. W. 1985. The Palynology of Archaeological Sites. London: Academic Press Drewett, P. 1974. The excavation of an Oval Barrow Mound of the third millennium bc at Alfriston, East Sussex, 1974, Proceedings of the Prehistoric Society 41, 119–152 Ellis, C. 1985. Flandrian molluscan biostratigraphy and its application to dry valley deposits in East Sussex. In Fieller, N. R. J., Gilbertson, D. D. & Ralph, N .G. A. (eds), Palaeobiological Investigations: research design, methods and data analysis, 157–166. Oxford: British Archaeological Report S226 Ellis, C. 1986. The postglacial molluscan succession of the South Downs dry valleys. In Sieveking, G. de G. & Hart, M. B. (eds), The Scientific Study of Flint and Chert, 175–194. Cambridge: University Press Entwistle, R. & Bowden, M. 1991. Cranborne Chase: the molluscan evidence. In Barrett, J. Bradley, R. & Hall, M., Papers on the Prehistoric Archaeology of Cranborne Chase, 20–48. Oxford: Oxbow Monograph 11 Evans, J. G. 1966. Late-glacial and post-glacial subaerial deposits Pitstone, Buckinghamshire, Proceedings of the Geologists’ Association 77, 347–363 Evans, J. G. 1967. The Stratification of Mollusca in the Chalk Soils and their Relation to Archaeology, Unpublished PhD thesis, Institute of Archaeology, University of London Evans, J. G. 1971. Habitat change on the calcareous soils of Britain: the impact of Neolithic man. In Simpson, D. D. A. (ed.), Economy and Settlement in Neolithic and Early Bronze Age Britain and Europe, 27–73. Leicester: Leicester University Press Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press Evans, J. G. 2007. The snails. In Benson, D. & Whittle, A., Building Memories; the Neolithic Cotswold long barrow at Ascott-under-Wychwood, Oxfordshire, 55–70. Cardiff Studies in Archaeology. Oxford: Oxbow Books Evans, J. G. & Jones, H. 1979. Mount Pleasant and Woodhenge: the land Mollusca. In Wainwright, G. J., Mount Pleasant, Dorset: Excavations 1970–1971, 190–213. London: Report of the Research Committee of the Society of Antiquaries of London 37 Evans, J. G. & Rouse, A. 1991. The land Mollusca. In Sharples, N., Maiden Castle; excavations and field survey 1985–6, 118–125. London: English Heritage Archaeological Report 19

162

Michael J. Allen

Fishpool, M. 1999. Land Mollusca. In Whittle, A., Pollard, J. & Grigson, C., The Harmony of Symbols; the Windmill Hill causewayed enclosure, 127–128. Oxford: Oxbow Books French, C., Lewis, H., Allen, M. J., Scaife, R. G. & Green, M. 2003. Archaeological and palaeoenvironmental investigations of the upper Allen valley, Cranborne Chase, Dorset (1998–2000): a new model of earlier Holocene landscape development, Proceedings of the Prehistoric Society. 69, 201–234 French, C., Lewis, H., Scaife, R. & Allen, M. 2005. New perspectives on Holocene landscape development in the southern English chalklands: the upper Allen valley, Cranborne Chase, Dorset, Geoarchaeology 20, 109–134 French, C., Lewis, H., Allen, M. J., Green, M. Scaife, R. G. & Gardiner, J. 2007. Prehistoric Landscape Development and Human Impact in the Upper Allen Valley, Cranborne Chase, Dorset. Cambridge: McDonald Institute Monograph Godwin, H. 1940. Pollen analysis and forest history of England and Wales, New Phytologist 39, 370–400 Godwin, H. 1962. Vegetational history of the Kentish Chalk downs as seen at Wingham and Frogholt, Veröffentlichungen des Geobotanischen Institutes Rubel, Zurich 37, 83–99 Godwin, H. 1975. The History of the British Flora. Cambridge: University Press Godwin, H. & Tansley, A. G. 1941. Prehistoric charcoals as evidence of former vegetation, soil and climate, Journal of Ecology 29, 117–126 Kennard, A. S. 1933. Report on the marine Mollusca, 235–241. In Stone, J. F. S., Excavations at Easton Down, Winterslow, 1931–32, Wiltshire Archaeological Magazine 46, 224–242 Kennard, A. S. 1934. Report on the Mollusca, 129–131. In Curwen, E. C., Excavations in Whitehawk Neolithic Camp, Brighton, 1932–33, Antiquaries Journal 14, 99–133 Kennard, A. S. 1935. Report on the non-marine Mollusca from the Stonehenge excavations of 1920–6, Antiquaries Journal 15, 432–434 Kennard, A. S. 1936a, 94–5. Report on the non-marine Mollusca. In Drew, C. D. & Piggott. S., The excavation of long barrow 163a on Thickthorn Down, Dorset, Proceedings of the Prehistoric Society 2, 77–96 Kennard, A. S. 1936b. The Mollusca, 90–92. In Curwen, E. C., Excavations in Whitehawk Camp, Brighton, third season 1935, Sussex Archaeological Collections 77, 60–92 Kennard, A. S. 1943. Report on the Mollusca. In Wheeler, R. E. M., Maiden Castle, Dorset, 372–374. London: Reports the Research Committee of the Society of Antiquaries of London 12 Kennard, A. S. & Woodward, B. B. 1929. The Mollusca, 69–70. In Curwen, E. C., Excavations in The Trundle, Goodwood, 1928, Sussex Archaeological Collections 70, 33–85 Kennard, A. S. & Woodward, B. B. 1930. The Mollusca, 83–5. In Ross Williamson, P. R., Excavations in Whitehawk Neolithic camp, near Brighton, Sussex Archaeological Collections 71, 56–96 Kerney, M. P. 1968. Britain’s fauna of land Mollusc and its relation to the Post-Glacial Thermal Optimum, Symposium of the Zoological Society of London 22, 273–291 Kerney, M. P., Brown, E. H. & Chandler, T. J. 1964. The late-glacial and post-glacial history of the chalk escarpment near Brook, Kent, Philosophical Transactions of the Royal Society, London B248, 135–204 Kerney, M. P., Preece, R. C. & Turner, C. 1980. Molluscan and plant biostratigraphy of some Late Devensian and Flandrian deposits in Kent, Philosophical Transactions of the Royal Society, London B291, 1–43 Kirby, K. J. 2003. What Might a British Forest Landscape Driven by Large Herbivores Look Like? Peterborough: English Nature Research Report 530 Lowe, J. J. & Walker, M. J. C. 1984. Reconstructing Quaternary Environments. London: Longman Piggott, S. 1954. The Neolithic Cultures of the British Isles; a study of the stone-using agricultural communities of Britain in the second millennium B.C. Cambridge: University Press

9. The southern English chalklands

163

Rouse, A. J. & Evans, J. G. 1993. Land molluscs, 211–217. In Whittle, A., Rouse, A. J. & Evans, J. G., A Neolithic downland monument in its environment: excavations at the Easton Down Long Barrow, Bishops Cannings, North Wiltshire, Proceedings of the Prehistoric Society 59, 197–239 Salisbury, E. J. & Jane, F. W. 1940. Charcoals from Maiden Castle and their significance in relation to the vegetation and climatic conditions in prehistoric times, Journal of Ecology 28, 310–325 Scaife, R. G. 1980. Late-Devensian and Flandrian Palaeoecological Studies in the Isle of Wight. Unpublished PhD thesis, King’s College London Scaife, R. G. 1987. A review of later quaternary plant microfossil and macrofossil research in southern England; with special reference to environmental archaeological evidence. In Keeley, H. C. M. (ed.), Environmental Archaeology; a regional review, vol. 2, 125–203. Historic Buildings and Monuments Commission for England, Occasional Paper 1. London: Historic Buildings and Monuments Commission for England Scaife, R. G. 1995. Boreal and sub-Boreal chalk landscape; pollen evidence’. In Cleal, R. M. J., Walker, K. E. & Montague, R., Stonehenge in its Landscape; twentieth-century excavations, 51–55. London: English Heritage Archaeological Report 10 Scaife, R. G. 2004. Avon valley floodplain sediments: the pre-Roman vegetational history. In Cleal, R. M. J, Allen, M. J. & Newman, C., An archaeological and environmental study of the Neolithic and Later Prehistoric landscape of the Avon Valley and Durrington Walls Environs, 228–234, Wiltshire Archaeological and Natural History Magazine 97, 218–248 Sumption, K. J. & Flowerdew, J. R. 1985. The ecological effects of the decline in rabbits Oryctolagus coniculus L. due to myxomatosis, Mammal Review 15, 151–186 Tansley, A. G. 1939. The British Islands and their Vegetation. Cambridge: University Press Thomas, A. S. 1960. Changes in vegetation since the advent of myxomatosis, Journal of Ecology 48, 287–305 Thomas, A. S. 1963. Further changes in vegetation since the advent of myxomatosis, Journal of Ecology 51, 151–183 Thomas, K. D. 1979. Appendix V. Land Mollusca and the environment of the Alfriston Barrow, 148–150. In Drewett, P., The excavation of an Oval Barrow Mound of the third millennium BC at Alfriston, East Sussex, 1974, Proceedings of the Prehistoric Society 41, 119–152 Thomas, K. D. 1982. Neolithic enclosures and woodland habitats on the South Downs in Sussex, England. In Bell, M. G. & Limbrey, S. (eds), Archaeological Aspects of Woodland Ecology, 147–170. Oxford: British Archaeological Report S146 Thomas, K. D. 1989. Vegetation of the British Chalklands in the Flandrian period: a response to Bush, Journal of Archaeological Science 16, 549–553 Thorley, A. 1971. Vegetational history of the Vale of Brooks, Institute of British Geographers’ Conference Proceedings Part 5, 47–50 Thorley, A. 1981. Pollen analytical evidence relating to the vegetational history of the Chalk, Journal of Biogeography 8, 93–106 Tomalin, D. J. & Scaife, R. G. 1979. A Neolithic flint assemblage and associated palynological sequence at Gatcombe, Isle of Wight, Proceedings Hampshire Field Club and Archaeological Society 36, 25–33 Vera, F. W. M. 1997. Metaforen voor de wildernis. Eik, hazelaar, rund en paard. Unpublished PhD thesis, Wageningen University, Netherlands Vera, F. W. M. 2000. Grazing Ecology and Forest History. Wallingford: CABI Waller, M. & Hamilton, S. D. 2000. Vegetation history of the English chalklands: a mid-Holocene pollen sequence from the Caburn, East Sussex, Journal of Quaternary. Science 15, 253–272 Waton, P. V. 1982. Man’s impact on the chalklands: some new pollen evidence. In Bell, M. & Limbrey, S. (eds), Archaeological Aspects of Woodland Ecology, 75–91. Oxford: British Archaeological Report S146

164

Michael J. Allen

Waton, P. V. 1986. Palynological evidence for early and permanent woodland on the chalk of Central Hampshire. In Sieveking, G. de G. & Hart, M. B. (eds), The Scientific Study of Flint and Chert, 169–174. Cambridge: University Press Whittle, A., Pollard, J. & Grigson, C. 1999. The Harmony of Symbols; the Windmill Hill causewayed enclosure. Oxford: Oxbow Books Whittle, A., Rouse, A. J. & Evans, J. G. 1993. A Neolithic downland monument in its environment: excavations at the Easton Down Long Barrow, Bishops Cannings, North Wiltshire, Proceedings of the Prehistoric Society 59, 197–239 Woodward, B. B. 1936. The Mollusca. In Curwen, E. C., Excavations in Whitehawk Camp, Brighton, third season, 1935, Sussex Archaeological Collections 77, 90–92

10. (Some thoughts on) using molluscs for landscape reconstruction and ecology in Malta Michael J. Allen and Bri Eastabrook

This chapter outlines the application of land mollusc analysis to the palaeo-environ­ mental interpretation of Malta; an island archipelago with abundant archaeology, a strongly calcareous bedrock, but little evidence of the detailed natural history observation seen in Britain and northwest Europe. This is not the place to discuss the results of detailed analysis, but several topics are explored including the application and appropriate aims for palaeo-environmental analysis in this karst landscape, and the boundaries/limitation of palaeoecological studies. This review may be relevant to the application of land snail analysis to other landscapes where the molluscan ecology is poorly, or less well, understood, than in Britain for instance. We also explore the existence of a former more wooded landscape as defined by palaeo-molluscan studies and, as compared with comparative studies in Britain and northwest Europe.

Malta – geology and soils The archipelago of Malta consists of the main island of Malta, with Gozo and the small island of Comino comprising about 316 km2. The geology is predominantly OligoMiocene Limestone, mainly Lower Coralline (140 m thick) and Globigerina with a lesser Lower Coralline Limestone (Bowen Jones et al. 1961; Schembri 1993) with deep stratified ‘valley loams’ in the valleys behind Mellieħa, St Paul’s and Salina bays (House et al. 1961). Much of the island is notably almost bare of soils, and where they do occur they are thin azonal rendzina-form or terra rossa soils such as calcareous rendzinas (carbonate raw soils), xerorendzinas and terra soils or near terra rossa soils (Lang 1960; 1961; Kubiena 1953). They are often thin (Vella 2001), variably stony, to stone-free, loamy soils. Although the parent material is mainly limestones, this weathers slowly resulting in only weakly calcareous soils, rather than the more strongly calcareous soils supported by the chalk of northwest Europe. The palaeo-environmental consequence of this is that shell preservation (or at least land snail shell preservation), may be poor and patchy (as in the limestones of the UK), but where Ca2CO3 levels are high enough,

166

Michael J. Allen and Bri Eastabrook

shell preservation can be very good. It is notable that, although snails are abundant in Malta generally (Mandahl-Barth 1988; Schembri 1992) preservation (or recovery) is often restricted to the larger more robust shells (eg, Pike 1971). However, in Malta a number of even the smaller species have thick, robust shells, to combat the heat and drying out.

Archaeology The Maltese Isles have been inhabited sporadically throughout prehistory (eg, Trump 2008; Zammit & Mallia 2008; Malone et al. 2009; Bonanno 2011) and have been the subject of archaeological curiosity since the dawn of the discipline (see Trump 1961). Increasingly it is being recognised that an understanding of how the landscape of Malta has changed through time is required to enable a better understanding of human activity. Previous palaeo-environmental studies have primarily focused on the use of badly preserved pollen from marine deposits (eg, Hunt 1997) or marine Mollusca (eg, Carroll et al. 2012). The conclusions drawn have, therefore, been geographically limited yet the islands provide the perfect sealed and contained environment and land-use record within which to utilise land snails as an palaeo-environmental proxy.

Palaeo-environmental enquiry Today the islands are almost bare of soil, trees and substantive woody vegetation, yet one presumes that wood was more generally available in Maltese prehistory. Many structures suggest timber roofs, some involving large timbers, and hearths requiring fuel (and charcoal) are common on prehistoric sites. The poor soils and lack of woodland today suggest huge changes have occurred in both, the elucidation of which traditional environmental archaeology is well equipped to tackle. Unlike Britain and northwest Europe, there remains considerable debate about the presence, nature, and distribution of ancient ‘natural’ woodland on the islands (Bonanno 2011). The karst landscape and resulting archaeological contexts and sedimentary deposits provide an environmental challenge. Pollen survives poorly in calcareous soil (Dimbleby & Evans 1974) and many of the deposits on sites are not suitable for palynological analysis. The terra rossa soils, and the weakly calcareous nature of most deposits, make land snail shell preservation patchy and often context-specific; rarely are there deeply stratified deposits such as ditch infills or colluvium which contain long and stratified sequences of well-preserved land snails. Further, where land snails are preserved (eg, Tas-Silġ, Schembri et al. 2000, and Xagħra Circle, Schembri et al. 2009; Fig. 10.1), the comprehension of the palaeo-ecology can be rather basic in comparison to, for instance, the UK. Within the Maltese Isles there has been far less focus on the use of land snails as a proxy, possibly due to the attention given to the use of gastropods (marine and terrestrial) as a source of food on the Islands (see Lubell 2004; Colonese et al. 2011). The Mediterranean, and the Maltese islands, have been understudied compared to northern Europe in terms of palaeo-environmental history. Certainly the Maltese islands do not have the long record of natural history recording of places such as the UK, where

10. (Some thoughts on) using molluscs for landscape reconstruction and ecology

167

Figure 10.1. A map showing the location of the main archaeological sites, and palaeoenvironmental studies and modern ecological studies (illustration: Abby George)

accounts of species presence and habitats have been recorded from Victorian period onwards and collated as long ago as the 1930s (Boycott 1934). Attempts to determine the palaeo-environment have often fallen short due to a number of issues. These include the intense human alteration of the landscape over millennia (eg, Schembri et al. 2009; Djamali et al. 2012), lack of pollen preservation in the limestone terrain, and a tendency to focus on marine mollusc records (Rackham 2003). These factors have led to fragmented records which are often disputed due to concerns over contamination. Thankfully, in recent years, various authors have recognised the need for a greater understanding of the past vegetation history of the Islands (eg, Carroll et al. 2012; Marriner et al. 2012), see Figure 10.1. Msida; Salina Bay; Marija Bay Burmarrad Marsa Xagħra Tas-Silġ

Carroll et al. (2010) Djamali et al. (2012); Marriner et al. (2012) Fenech (2007) Schembri et al. (2009) Schembri et al. (2000)

168

Michael J. Allen and Bri Eastabrook

Only relatively recently has a full molluscan gazetteer been compiled (Beckmann 1992) and full comparanda with detailed species-specific habitat ecology, and more detailed distribution has followed (Guisti et al. 1995). Papers by, for instance, Holyoak (1986), Beckmann and Gittenberger (1987) and more recently Cilia et al. (2012), have significantly increased species-specific ecological and habitat information. Consequently the use of land snails to address the Maltese land-use and environment, and their application in archaeology has been slightly later, if not slower, than Britain. Even in Britain such analysis only became more widely applied with the publication of John G. Evans seminal textbook (1972), and by that time most of the major archaeological excavations and research by Trump at Skorba (1961; 1966) and Evans (1971) had been completed if not wholly published. The analysis of land-snails as a proxy palaeo-environmental indicator is common in the pollen-poor calcareous and karst geologies of northwest Europe and particularly southern England, where it has been a common and standard palaeo-environmental analysis on archaeological sites for over 40 years (Evans 1972). Elsewhere in Europe and the Mediterranean it has not been used a regularly in archaeological and palaeoenvironmental studies. In Malta, although antiquarian analyses are known (Despott 1917; 1918; Soós 1933; Trenchmann 1938; and Baldacchino 1937; 1939 cited by Giusti et al. 1995, 90), analysis has only been used on a relatively few recent archaeological investigations, and here the work of P. J. Schembri is key to both the deployment of that analysis, and also to the interpretations of the Maltese molluscan palaeo-fauna (eg Schmebri 1992; Hunt & Schembri 1999; Schembri et al. 1990). The first analysis on ‘recent’ excavations was the limited listing of shells from the Xemxija tombs (Pike 1971, table 3) in influential John D. Evans’s work (Evans 1971). More recently, the exemplary work at Brochtorff Circle, Xagħra (Schembri et al. 2009), clearly indicates the palaeo-environmental potential of land snail studies. Ongoing work at Tas-Silġ further exemplify this (Schembri et al. 2000; Hunt forthcoming). Studies at the Brochtorff Circle, Xagħra (by Schembri and Hunt) were largely from bulk samples which were then ordered by phase, and no contiguous sequences were examined, thus chronology and environmental history was developed by comparison of samples from sequential phases, albeit from different location and micro-environments.

Ecology and neoecology Palaeo-environmental reconstruction from molluscan analysis in the Atlantic zones is largely based on determining the vegetative state of the local environs; principally differentiating between the shady nature of microhabitats. As such, analyses have been used to define, via local moisture regimes, changes from woodland to open environs indicating anthropogenic woodland clearance. But more specifically analyses can indicate a range of different vegetative types and land-uses as well as defining long grassland (rough graze), short grassland (pasture) and arable etc. The Maltese and Mediterranean fauna are more xerophillic, and the range of habitat groups both narrower and less specific, than those in northwest Europe. Nevertheless, a series of

10. (Some thoughts on) using molluscs for landscape reconstruction and ecology

169

palaeo-environmental niches has recently been defined by Hunt and Schembri (in Malone et al. 2009), making the deployment of palaeo-molluscan analysis possible and useful. Individual species can be categorised into various eco-types: soil and stones, leaf litter, open, subterranean and eurytopic, and the non-troglophile assemblage groups used to define the following surface conditions (cf. Schembri et al. 2009, table 2.5): • • • • • • • •

Exposed steppe Limestone pavement Relatively well-vegetated steppe Vegetated steppe Well vegetated steppe or garrigue Well vegetated steppe or shrubland Sheltered, vegetated Field

Other species groups may indicate rocky rubble or troglophile conditions. As a consequence of the recent work by Hunt and Schembri (in Malone et al. 2009), palaeo-molluscan studies in Malta are now more readily applicable in archaeological, geoarchaeological and palaeo-geographical studies. In Malta, during palaeo-environmental analysis and interpretation, the molluscan species are allocated habitat preferences (Guisti et al. 1995; Hunt & Schembri 1999; Fenech 2007; Schembri et al. 2009) and these are; leaf litter (including mesic), ubiquitous (= catholic or intermediate), open country (and xerophiles), and subterranean (ie, burrowing). This, and species specific ecologies (cf. Eastabrook 2013, table 3.1), do not always allow much nuanced interpretation within major environment classes. The differentiation of grassland, grazed garigue, pasture, bare soil and tillage for instance, has not been entirely satisfactorily isolated (should they ever have existed), despite the fact there are many very specific ecological niches, and many plants inhabit highly localised niches (Grove & Rackham 2003), and recent work has highlighted the molluscan ecology seen within the island (Schembri 2003). The few studies which have utilised Mollusca as sources of palaeo-environmental information have relied on fairly limited information about the species present and their habitat preferences (Schembri et al. 2009), providing a degree of inherent uncertainty. More fundamentally, there are no good ancient woodland palaeo-faunas, and recent limited modern ecological studies have tried to go some way to addressing this (Eastabrook 2013). Importantly land snails are also known to be sensitive to human impacts and so are often used as indicator species for disturbance, especially in woodland context (cf. Pilāte 2003), and thus may be ideal for examining the presence and nature of early prehistoric woodland.

Modern neoecology and ancient woodland indicators Because of the perceived lack of some molluscan habitat information a number of ‘woodland’ habitat locations within Malta were selected for modern ecology study of the land snail fauna. A variety of open woodland habitats exist on the archipelago and the small variations between these habitats create faunas with unique molluscan species

170

Michael J. Allen and Bri Eastabrook

Figure 10.2. Ecological molluscan composition of a section of the modern study sites with >25 specimens (after Eastabrook 2013, fig. 3.2; illustration: Abby George)

composition. Eighteen sites were chosen for ecological study (Fig. 10.1), spread across the Island providing good spatial and ecological representation. Locations sampled included open country xerophilic habitats (7), open pine woodland ie, stands of pine (1), open oak woodland, individual stands of oak (2), open field (3) and garden habitats (3), in which 25 studies were undertaken; ten sites yielded enough specimens (>25) to be used for statistical analysis and were reviewed. Analysis of the mollusc faunas from present day leaf litter shows clear patterns between the various habitat groups (and significant differences with the archaeological results from Skorba (compare Fig. 10.2 with 10.5)). The most species-diverse were the maquis sites, with open country species being most dominant. This raises the question that the dominance of open country species need not, therefore, represent a wholly open xerophilic habitat. Here, the range of species, and species diversity, are more significant pointers. Garden sites had low numbers of molluscs but Rumina decollata and Papillifera papillaris were most common. Vitrea subrimata was the only species present in the three field edge locations examined.

10. (Some thoughts on) using molluscs for landscape reconstruction and ecology

171

Figure 10.3. Composition of different habitats in the modern day survey of the 20 locations; site numbers are those from Eastabrook 2013 (after Eastabrook 2013; fig. 3.3; illustration: Abby George)

Maquis locations were mollusc rich, with high diversity, with 19 of the 20 terrestrial species recorded in the survey present; only Lauria cylindracea was consistently absent from this habitat from nine sample locations. In contrast, the oak-dominated study area, all samples had low species diversity, but with a clear dominance of leaf litter species (Fig. 10.3). Importantly where leaf-litter was dominant, the faunas all included the ‘indicator species’ Lauria cylindracea, often accompanied by larger numbers of Truncatellina callicratis and sometimes Oxychilus draparnudi. The former was almost exclusive to the oak woodland, with only one other specimen being found elsewhere in the whole study. If the presence of L. cylindracea really is that specific, then this may be of significant palaeo-ecological interpretative value. A distinct lack of species in pine woodland, however, may be attributable to the more acidic soil surface, and this lack may be a consideration with some low shell-yielding phases within archaeological sequences – taphonomic and depositional factors aside.

Stratified sequences and preliminary palaeo-ecology of the Skorba ‘temple’ environs Traditionally pollen analysis has been the focus of long palaeo-environmental histories, and has only relatively recently been applied successfully in the Maltese Islands (Hunt & Vella 2005; Fenech 2007; Carroll et al. 2012; Djamali et al. 2012). Unfortunately the records are often fragmented, chronologically limited, and difficult to interpret.

Michael J. Allen and Bri Eastabrook

172

Few land snail analyses have attempted to look at stratified sequences such as colluvium, to examine a wider time depth and land-use history. One recent example is the Maltese Temples Project directed by Tim Darvill and Fritz Lüth (Allen et al. 2010) where, following a larger auger and test pitting programme (Fig. 10.4), locations of stratified colluvium or terrace deposits around the Skorba ‘Temple’ were isolated and examined. In one (test pit 5) a full sequence of land snail samples was taken. The test pit sampled was located within a large, very broad, terrace and also lay at the foot of the slight rise behind the megalithic temple (Fig. 10.4 inset). The typical local soils are shallow xerorendzinas about 0.20–0.30 m depth, with terrace soils typically at their deepest here about 0.7–0.8 m deep. The test pit revealed a brown terrace soil/colluvial brown earth with a distinct occupation horizon at about 0.9 m, at which point excavation was ceased so as not to create localised damage to any in situ archaeological deposits that might relate to the Skorba site. Relatively large sherds of pottery were recovered at the base of the ploughsoil and throughout the colluvium. A dense concentration of sherds at the base of the test pit (where potentially in situ occupation was present) seemed to belong to Tarxien phase (c. 3150–2500 BC) and earlier (R. Grima pers. comm.). Depth (cm)

Context

Snail samples (cm)

0–13

1

13–37

2

24–29 29–34 34–39

37–61

3a

39–44 44–49 59–54 54–59

61–90

3b

90– 92+

4

59–64 64–69 69–74 74–79 79–84

Description Brown (10YR 3/3–3/4) very dry very fine well-developed small–very small crumbs, common small & medium stones, abrupt boundary AP – ploughsoil Brown (7.5YR 4/3) silty (clay) loam, very weak large–medium prismatic–subangular blocky structure, some medium stones, some very small stones (large pottery sherds at base & junction) A horizon Dark greyish brown (10YR 4/2) silt, some medium stones, structureless, contains stone slabs, less coarse silt, coarse silt, pottery present, including dense concentration in bottom few centimetres, clear boundary Upper colluvial deposit (or upper Skorba Deposit) – colluvial brown earth/terrace Brown (strong grey hue) (10YR 5/3) silt to silty clay loam, some coarse silt, massive, common medium stones over stones, pseudomycelium (= free CaCO3), pottery present Lower colluvial deposit (or lower Skorba Deposit) – colluvial brown earth/terrace Same matrix as above with abundant & medium large stones, forming a compact surface over whole of pit – unexcavated because thought to be archaeological structure; large sherd at top of the context

Colluvial soils such as these are normally sampled for snails at 10 cm intervals (Evans 1972; Bell 1983), however, here closer interval and larger (2 kg), samples were taken. A sequence of 12 samples was taken contiguously at up to 50 mm intervals through 0.84 m of the exposed colluvium in test pit 5 (Allen et al. 2010).

10. (Some thoughts on) using molluscs for landscape reconstruction and ecology

173

Figure 10.4. Plan of the Skorba Valley showing Skorba ‘Temple’, the auger and test pit profile, and the location of test pit 5 (marked); inset Skorba ‘temple’ (after Allen et al. 2010; photo T. Darvill)

174

Michael J. Allen and Bri Eastabrook

Figure 10.5. Ecological composition of mollusc assemblages from the Skorba sequence (after Eastabrook 2013; fig. 3.5: illustration: Abby George)

Provisional interpretation of the land snails from the assumed temple-period deposits seem to indicate generally open environments, but with a high micro-habitat diversity (Fig. 10.5). No samples included specimens of L. cylindracea, suggesting a lack of (oak) woodland (sensu Eastbrook 2013) local to the temple site during the main phase of use (Temple Period c. 6000–4500 BP). Nor was this a dominant habitat in the Maltese Temple and Bronze Age periods throughout the island (see Eastabrook 2013, fig. 4.4). Despite the relative uniformity of the colluvium, clear distinction between the lower and upper mollusc fauna is apparent (Fig. 10.6) and may reflect changes in local microhabitats and increase in local habitat diversity, possibly in response to less human activity in the area immediately around the site. What is also clear is that the archaeological assemblages, summarised in Figure 10.6, are clearly different from all of the modern assemblages from open, garden, maquis and oak woodland locations (compare Figs 10.3 and 10.6).

10. (Some thoughts on) using molluscs for landscape reconstruction and ecology

175

Figure 10.6 General habitat change at Skorba as represented by the molluscan evidence (after Eastabrook 2013; fig. 3.6; illustration: Abby George)

Evidence of former woodlands What can this study, and previous analysis, tell us about the potential for identifying prehistoric woodland from the Maltese archipelago? Molluscan ecologies of prehistoric woodland are poorly understood. Woodlands were probably open with some ground vegetation, in contrast to the north-western European climax woodland, Quercentum Mixtum, which may have had a flora poor, but leaf-litter rich woodland floor. What is clear is that the oak woodlands examined seem to have a clearly distinctive fauna, lacking many of the open and ubiquitous species common in a number of the other locations (Figs 10.2 and 10.3). Analysis of even small-scale modern ecological surveys can contribute to our understating of the components and composition of the modern flora which may, as Eastabrook has suggested (2013), allow us to better interpret palaeo-faunas, and possibly even to re-examine and re-interpret existing published palaeo-molluscan data. The hints of valuable information from modern mollusc ecology studies indicate the huge potential in this generally understudied and these obvoulsy quite complex Mediterranean woodland ecologies. Clearly the establishment of an archipelago-specific environmental history is key and recent work (Allen et al. 2010), but more significantly that of the FRAGSUS project (//www.qub.ac.uk/sites/FRAGSUS) will be addressing this in the future.

Concluding thoughts These small modern ecological observations only go to emphasise the value of

176

Michael J. Allen and Bri Eastabrook

continued detailed ecological studies, and that even relatively limited investigations can potentially provide information valuable to the palaeo-environmental interpretations of archaeological assemblages. There are number of issues and criticisms that can be levelled at the small ecological forays discussed, nevertheless it indicates that larger and more detailed further modern ecological studies may be valuable. Establishing a good comprehension of contemporary habitat preferences is clearly meaningful to palaeo-ecological investigations, and can potentially increase our understanding of past environments, in parallel with major new programmes of archaeological analyses. The presence, and nature of, Maltese early Holocene woodland remains undefined, and we hope new targeted molluscan and palynological analyses will soon resolve this question to a greater or lesser extent. This study has highlighted how the use of land snails as a proxy could be successfully applied to palaeo-environmental studies not only in Malta, but the majority of the Mediterranean.

Acknowledgements Bri Eastabrook thanks Professor Neil Roberts, Kim Terribile, Martin Kent, Edwin Lanfranco, Professor Patrick Schembri, Katrin Fenech, Jeffery Sciberras and Chris Hunt all for invaluable help with the completion of work for her research dissertation. The Skorba fieldwork was conducted by Tim Darvill (Bournemouth University) and Fritz Lüth (formerly Römisch-Germanische Kommission des Deutschen Archäologischen Instituts, Frankfurt am Main).

References Allen, M. J., Darvill, T., Gale, J., Lüth, F., Magnavita, S. & Rassmann, K. 2010. Maltese Temples Landscape Project First Interim Report (July 2010). Bournemouth and Frankfurt: Bournemouth University School of Conservation Sciences & Römisch-Germanische Kommission des Deutschen Archäologischen Instituts, Frankfurt am Main, in association with Heritage Malta Baldacchino, J. G. 1937. Annual Report on the Working of the Museum Department during 1936–1937. Malta: Government Printing Office Baldacchino, J. G. 1939. Report by Dr. J. G. Baldacchino on the archaeological section from August 1938 to March 139. In J. G. Baldacchino, Annual report on the working of the Museum Department during 1938–1939, vi–ix. Malta: Government Printing Office Beckmann, K-H 1987. Land- und Süßwassermollsken der Maltesischen Inseln, Heldia 1, supplement 1, 1–38 Beckmann, K-H. 1992. Catalogue and bibliography of the land- and freshwater molluscs of the Maltese Island, the Pelagi Islands and the isle of Pantelleria, Heldia 2, suppllement 2, 1–60 Beckmann, K-H. & Gittenberger, E. 1987. The Clausiliidae (Gastropoda) of the Maltese Islands, some additional data, Journal of Conchology 32, 335–338 Bell, M. G. 1983. Valley sediments as evidence of prehistoric land-use on the South Downs, Proceedings of the Prehistoric Society 49, 119–150 Bonanno, A. 2011. The lure of the islands: Malta’s first Neolithic colonisers. In Phoca Cosmetatou, N. (ed.), The First Mediterranean Islanders: initial occupation and survival strategies, 145–156.

10. (Some thoughts on) using molluscs for landscape reconstruction and ecology

177

Oxford: University of Oxford School of Archaeology Monograph 74 Bowen Jones, H., Dewdney, J. C. & Fisher, W. B. (eds), 1961. Malta a Background for Development. Durham: Durham University Press Boycott, A. E. 1934. The habitats of land Mollusca in Britain, Journal of Ecology 22, 1–38 Carroll, F. A., Hunt, C. O., Schembri, P. J. & Bananno, A. 2012. Holocene climate change, vegetation history and human impact in the Central Mediterranean: evidence from the Maltese Islands, Quaternary Science Reviews 52, 24–40 Cilia, D. P., Sciberras, A., Sciberras, J. & Pisani, L. 2012. Terrestrial gastropod of the minor islets of the Maltese archipelago (Mollusca Gastropoda), Biodiversity Journal 3(4), 543–554 Colonese, A. D., Mannino, M. A., Bar-Yosef Mayer, D. E., Fa, D. A., Finlayson, J. C., Lubell, D. & Stiner, M. C. 2011. Marine mollusc exploitation in Mediterranean prehistory: an overview, Quaternary International 239, 86–103 Despott, G. 1917. Excavations conducted at Ghar Dalam (Malta) in the July, 1916. Report of the Meeting of the British Association of the Advancement of Science 86, 294–301 Despott, G. 1918. Excavations conducted at Ghar Dalam (Malta) in the summer of 1917, Journal of the Royal Anthropological Institute of Great Britain and Ireland 48, 214–221 Dimbleby, G. W. & Evans, J. G. 1974. Pollen and land-snail analysis of calcareous soils, Journal of Archaeological Science 1, 117–133 Djamali, M., Gambin, B., Marriner, N., Andrieu-Ponel, V., Gambin, T., Gamdouin, E., Lanfranco, S., Médail, F., Pavon, D., Ponel, P. & Morhange, C. 2012. Vegetation dynamics during the early to mid-holocene transition in NW Malta; human impact versus climatic forcing, Vegetation History & Archaeobotany 22(5), 367–380 Eastabrook, B. 2013. Ancient Forests in Malta: Fact or Fiction? A Land Snail Analysis. Unpublished MRes thesis, Plymouth University Evans, J. D. 1971. The Prehistoric Antiquities of the Maltese Islands; a survey. London: Athlone Press Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press Fenech, K. 2007. Human-induced Changes in the Environment and Landscape of the Maltese Islands from the Neolithic to the 15th century AD – as Inferred from the Scientific Study of Sediments from Marsa, Malta. Oxford: British Archaeological Report S1682 Grove, A. T. & Rackham, O. 2003. The Nature of Mediterranean Europe: an ecological history. Second edition. London: Yale University Press Guisti, F., Manganelli, G & Schembri, P. J. 1995. The Non-marine Molluscs of the Maltese Islands, Mongrafie 15. Torino: Museo Regionale di Scienze Naturali Holyoak, D. T. 1986. Biological species-limits and systematics of the Clausiliidae (Gastropoda) of the Maltese Islands, Journal of Conchology 32, 211–220 House, M. R., Dunham, K. C. & Wigglesworth, J. C. 1961. Geology and structure of the Maltese Islands. In Bowen Jones, H., Dewdney, J. C. & Fisher, W. B. (eds), Malta a Background for Development, 24–33. Durham: Durham University Press Hunt, C. O. 1997. Quaternary deposits in the Maltese Islands: a microcosm of environmental change in Mediterranean Islands, GeoJournal 41(2), 101–109 Hunt, C. O. forthcoming. Palynology of some archaeological deposits from Tas-Silġ. In Bonanno, A. & Vella, N. (eds), Excavations at Tas-Silġ, Malta, Conducted by the Department of Classics and Archaeology, University of Malta (1996–2005). Malta: Midsea Books Hunt, C. O. & Schembri, P. J. 1999. Quaternary environment and biogeography of the Maltese island. In Misfud, A. & Savona Ventura, E. (eds), Facets of Maltese Prehistory, 41–75. Malta: Prehistoric Society of Malta Hunt, C. & Vella, N. C. 2005. A view from the countryside: pollen from a field at Mistra Valley, Malta, Malta Archaeological Review 7, 57–65 Kubiena, W. L. 1853. The Soils of Europe. London: Murby

178

Michael J. Allen and Bri Eastabrook

Lang, D. M. 1960. Soils of Malta and Gozo. Colonial Research Studies 29. London: HMSO Lang, D. M. 1961. Soils of Malta and Gozo. In Bowen-Jones, H., Dewdney, J. C. & Fisher, W. B. (eds), Malta a Background for Development, 83–98. Durham: Durham University Press Lubell, D. 2004. Prehistoric edible land snails in the circum-Mediterranean: the archaeological evidence. In Brugal, J.-J. & Desse, J. (eds), Petits Animaux et Societes Humaines. Du complement alimentaire aux ressources utilitaires. XXIVe rencontres internationales d’archeologie et d’histoire d’Antibes, 77–98. Antibes: Editions APDCA Malone, C., Stoddart, S., Bonanno, A. & Trump, D. (eds), 2009. Mortuary Customs in Prehistoric Malta: excavations at the Brochtorff Circle at Xagħra (1987–1994).Cambridge; University Press Mandahl-Barth, G. 1988. The Shell-bearing Land-snails of Malta. Mdina: Friends of the National Museum of National History Marriner, N., Gambin, T., Djamali, M., Morhange, C. & Spiteri, M. 2012. Geoarchaeology of the Burmarrad ria and early Holocene human impacts in western Malta, Palaeogeography, Palaeoclimatology, Palaeoecology 339–341, 52–65 Pike, G. 1971. The animal bones from the Xemxija tombs. In Evans, J. D. The Prehistoric Antiquities of the Maltese Islands; a survey, 139–141. London: Athlone Press Pilāte, D. 2003. Terrestrial molluscs as indicator species of natural forests. In Heikkilä, R. & Lindholm, T. (eds), Biodiversity and Conservation of Boreal Nature, 216–220. Vantaa: Kainuun ympäristökeskus Rackham, O. 2003. The physical setting. In Abulafia, D. (ed.), The Mediterranean in History, 33–66. Los Angeles: Getty Publications Schembri, P. 1992. Diversity and conservation of the non-marine molluscs of the Maltese Islands. In Giusti, F. & Manganelli, G. (eds), Abstracts of Eleventh International Malocological Congress Siena 1992, 192–195. Siena: University of Siena Schembri, P. J. 1993. Physical geography and ecology of the Maltese Islands; a brief overview, Options Méditeranéennes B7 (Malta: food agriculture, fisheries and the environment), 27–39 Schembri, P. 2003. Current state of knowledge of the Maltese non-marine fauna. In Malta Environment and Planning Authority Annual Report and Accounts 2003, 35–65. Floriana: Malta Environment and Planning Authority Schembri, P. J., Falzon, A., Fenech, K. & Sant, M. J. 2000. The molluscan remains. In Bonanno, A., Frendo, A. & Vella, N. C. (eds), Excavations at Tas-Silġ, Malta. A preliminary report on the 1996–1998 campaigns conducted by the Department of Classics and Archaeology of the University of Malta, Mediterranean Archaeology 13, 102–109 Schembri, P., Pedley, M., Hunt, C. & Stoddart, S. 2009. The environment of Malta and Gozo and of the Xaghra Circle. In Malone, C., Stoddart, S., Bonanno, A. & Trump, D. (eds), Mortuary Customs in Prehistoric Malta: excavations at the Brochtorff Circle at Xagħra (1987–1994), 17–19. Cambridge; University Press Soós, L. 1933. A systematic and zoogeographical contribution to the mollusc fauna of the Maltese islands and Lamedusa, Archiv für Naturgeschichte 2, 305–353 Trechmann, C. T. 1938. Quaternary conditions in Malta, Geological Magazine 75, 1–26 Trump, D. H. 1961. The later prehistory of Malta, Prehistoric Society of Malta 27, 253–262 Trump. D. H. 1966. Skorba: excavation carried out on behalf of the National Museum of Malta 1961–1963. Oxford: Report of the Research Committee of the Society of Antiquaries of London 22 Trump, D. H. 2008. Malta: Prehistory and Temples. Third edition. Malta: Midsea Books. Vella, S. 2001. Soil information in the Maltese Islands. In Zdruli, P., Steduto, P., Lacirignda, C. & Montanarella, L. (eds), Soil resources of southern and eastern Mediterranean countries, Options Méditeranéennes B34, 171–191 Zammit, M. & Mallia, J. 2008. (eds), Ta’Ħaġrat and Skorba: ancient monuments in a modern World. Valetta: Heritage Malta

11. Molluscan remains from early to middle Holocene sites in the Iron Gates reach of the Danube, southeast Europe Catriona Pickard, Adina Boroneanț and Clive Bonsall

Molluscan assemblages recovered from archaeological sites can potentially provide a wealth of information about past environment and can offer an opportunity to investigate the human use of shells and shellfish. However, there are many practical and theoretical problems associated with the recovery, identification, quantification and interpretation of molluscan assemblages. Some of these issues are universal to all molluscan assemblages, others are specific to the conditions and excavation and recovery practices at individual sites. This paper discusses these problems in the context of the late glacial and early to middle Holocene sites located along the Iron Gates stretch of the River Danube in southeast Europe. It also presents a pragmatic interpretation of the Iron Gates molluscan remains, extracting meaning from less than ideal assemblages. Molluscan remains are common finds on archaeological sites throughout the world. Archaeomalacologists study assemblages of aquatic molluscs (both freshwater and marine species) to reveal the past human use of shells and shellfish, and terrestrial molluscs as proxy indicators of palaeo-environments. However, this is not an absolute distinction – past environments can be explored through the analysis of aquatic molluscs, while terrestrial molluscs can attest to cultural and economic activities (eg, Taylor & Bell, Chapter 12). This paper explores the theoretical and practical problems associated with analysing and interpreting molluscan assemblages. Some of these problems are universal to all sub-fossil mollusc assemblages; others are particular to the conditions and recovery practices at sites in the Iron Gates region of southeast Europe, which form the principal exemplar of this review. The Iron Gates reach of the Danube extends for over 200 km and forms the border between Serbia and Romania (Fig. 11.1). An unparalleled record of Mesolithic and Neolithic settlement from over 50 open-air and cave sites has been documented for this stretch of the Danube (Bonsall 2008). Many of the sites were investigated between 1964–71 and 1977–84 during construction of the Iron Gates I and II dams across the Danube, which resulted in flooding (and in many instances, complete submergence) of the sites. The generally calcareous soil conditions of the Iron Gates sites favoured the preservation of abundant human and animal remains, including mollusc shells.

180

Catriona Pickard, Adina Boroneanț and Clive Bonsall

Figure 11.1. Mesolithic and Early Neolithic (14,700–7500 cal BP) in the Iron Gates with sites mentioned in the text indicated

The remains of both terrestrial and aquatic species of mollusc have been reported from many Iron Gates (for summaries see Radovanović 1996; A. Boroneanț 2012). However, few detailed analyses of the molluscan remains have been undertaken and many published accounts focus on the human use of shells – emphasis being placed on aquatic specimens that were humanly modified, such as the shell beads recovered from funerary and settlement contexts (eg, V. Boroneanţ 1990; Cristiani & Borić 2012; Mărgărit et al. forthcoming). Many molluscan remains recovered from the Iron Gates sites were hand-collected during what were essentially salvage excavations. Dry sieving was employed in some of the 1960s excavations, at Cuina Turcului, Icoana and Schela Cladovei for example, although how extensively or systematically this was done is difficult to assess. A notable attempt at systematic recovery of molluscan remains was in excavations at Schela Cladovei between 1992 and 1996 (directed by V. Boroneanț and C. Bonsall), where in addition to hand collection of shells virtually all soil excavated from within and between identifiable archaeological features was wet sieved.

The nature of the molluscan assemblages from the Iron Gates In spite of the importance of aquatic resources in human diet as reflected in the bone

11. Molluscan remains from early to middle Holocene sites

181

chemistry of the sites’ inhabitants, detailed records of molluscan assemblages are not available from most sites. This lack of documentation most likely reflects recovery techniques, curatorial policy and publication emphasis rather than the cultural or economic significance of molluscs to late glacial–early Holocene populations in the region. Table 11.1 is an attempt to summarise the reported occurrences of molluscan remains in Mesolithic and Early Neolithic sites in the Iron Gates, based mainly on published sources. However, it should be noted that published sources do not often give precise numbers of shell finds, therefore only presence/absence information is given in Table 11.1. The molluscan remains recovered in the 1992–1996 excavations at Schela Cladovei were analysed by one of the authors (CP). In the Romanian literature, a distinction is usually made between scoici (which translates as ‘scallops’, but normally refers to the shells of freshwater mussels) and melci (which translates as ‘snails’), but species identifications are rarely made and ‘snails’ can refer to terrestrial, freshwater or marine species.

Issues with methods and problems with proxies The information that drives archaeomalacological reconstruction includes species identification, together with quantification of number, as well as metric analysis of shell size and morphology. The accuracy and reliability of the reconstruction of past environment and interpretation of human activity depends on a range of variables.

Recovery methods Where conditions permit, vertical sequential sampling is the preferred strategy for molluscan analysis (see Allen 2017a, Chapter 1). For palaeo-environmental study, ideally the samples should be taken from exposed sections with sample volume sufficient to recover over 100 shells per sample (see Evans 1972, 41–44; Allen 2017a, 2017b, Chapters 1 and 2). Terrestrial molluscs should be recovered by flotation using a 500 µm mesh for shell collection, whereas the denser shells of aquatic species can be recovered by wet sieving the excavated deposits (see Campbell, Chapter 16). Hand-collected shell assemblages are generally inadequate for a comprehensive reconstruction of palaeo-environment and past human activities because smaller specimens are likely to be overlooked (cf. Davies 2008; Evans 1972; Pickard & Bonsall 2014). However, wet sieving and flotation of excavated sediments was not regularly undertaken on many archaeological sites until the second half of the 20th century. Archaeological identification of molluscs depends on the presence of species-specific characteristics on the shell. Identification to species level is particularly problematic for closely related species of terrestrial molluscs. Attribution to species and in some cases even to genus level is, therefore, non-trivial in degraded archaeological specimens. For example, shells of the family Helicidae were identified at Padina and Schela Cladovei. The form and size of the shells of members of the family may exhibit high variability at intra- and inter-population levels – some of the large Helix species, such as the edible

EN M EN M M EN EN? EN M M M M M ? LM EN EN EN M M EN MN

Cerithium sp. (?fossil) +

Columbella rustica +

Cyclope neritea +

+

+

Dentaliid scaphopod (?fossil) +

+

+

+ +

Glycymeris sp. +

Fissidentalium badense (fossil) +

Glycymeris glycymeris (?fossil) +

Pannonicardium dumicici (?fossil) +

Melanopsis impressa (fossil) +

?

Spondylus sp. +

+

+

+

Turitella turis (?fossil) +

Lithoglyphus naticoides +

+

+

Lithoglyphus opertus +

Theodoxus sp. +

Theodoxus danubialis +

+

+

+

Theodoxus transversalis +

Unio crassus + +

+

Unio cf. crassus + +

+

Unio pictorum

Cytherea sp. (?fossil)

+

Data from V. Boroneanţ 1969; Srejović and Letica 1978; Clason 1980; Stanković 1986a, 1986b; Vasić 1986; Radovanović 1996; V. Boroneanţ et al. 1999; Păunescu 2000; Dimitrijević & Tripković 2006; Dimitrijević 2008, 2014; A. Boroneanț 2012; Cristiani & Borić 2012. The dates of the contexts have been abbreviated as follows: M-Mesolithic, LM-late Mesolithic, EN- Early Neolithic and MN-Middle Neolithic

Alibeg Băile Herculane Climente I Climente II Cuina Turcului Cuina Turcului Icoana Knjepište Lepenski Vir Ostrovul Banului Ostrovul Corbului Ostrovul Mare Padina A & B Răzvrata Schela Cladovei Schela Cladovei Ušće Kameničkog-Potoka Velesnica Veterani Terrace Vlasac Starčevo Vinča-Belo Brdo

Context

Unio tumidus

Table 11.1. Molluscan remains identified at Iron Gates Mesolithic and Early Neolithc sites compared with Starčevo and Vinča-Belo Brdo (both of which are located upstream from the Iron Gates)

?

? ? +

+

+

? + ?

Unio spp.

182 Catriona Pickard, Adina Boroneanț and Clive Bonsall

11. Molluscan remains from early to middle Holocene sites

183

snails Cornu aspersum syn. Helix aspersa (the garden snail) and Helix pomatia (the Roman or edible snail), are very similar in appearance. Therefore, it was not always possible to attribute specimens recovered at Schela Cladovei to species, particularly as the periostracum was often absent. It is likely, based on the size and form of the specimens collected, that most examples belong to either C. aspersum or H. pomatia.

Ecology, community and taphonomy Reconstructing environment from the presence of specific molluscan remains (proxies for past environments) assumes the principle of uniformitarianism, ie, the ecology of modern mollusc populations is the same as the ecology of past populations. Knowledge of modern species ecology can thus be used to construct models of past habitats based on the species identified (and their relative abundance) in archaeological assemblages. Although the ecology of certain Helicidae, ie, the commercially important edible snails, is relatively well understood, studies of the ecology of other modern terrestrial and freshwater mollusc populations in the Iron Gates region are limited. It is crucial, when analysing molluscan assemblages from archaeological sites, to distinguish between cultural vs natural deposition. Molluscs collected from archaeological deposits may be autochthonous and not directly associated with past human activity. Cultural deposits can provide information on past human activity, while endemic molluscs can indicate past environment and habitat variation where appropriate horizontal and/or vertical sampling strategies are employed. Freshwater species recovered from archaeological contexts in the Iron Gates sites may have been intentionally collected by humans or may have been naturally deposited on sites as a consequence of episodic overbank flooding. For example, the freshwater gastropods Theodoxus danubialis and Viviparus acerosus were each represented by a single specimen at Schela Cladovei, and Viviparus viviparus by three specimens at Starčevo (Clason 1980). None of the shells showed definite signs of modification (although one V. viviparus shell from Starčevo may have damage sustained during extraction of the flesh), and so these molluscs or their shells were not necessarily collected intentionally. Similarly, the Helix spp. shells documented at Padina and Schela Cladovei lack evidence for modification or intentioned deposition. Helix spp. are endemic in the region and so could have been deposited naturally on the site. Moreover, the active relationship between species location and human activity can further confuse the interpretation of finds. C. aspersum/H. pomatia are anthropochorus species known to rapidly colonise land disturbed by settlement and agriculture (Madec et al. 2000). Site disturbance – but also subsequent abandonment – may produce the ideal conditions for the proliferation of these species, thus obscuring the relationship with human activity. Several shells of Helix spp. from the 1992–1996 excavations at Schela Cladovei had small round perforations resulting from the action of carnivorous gastropods (eg, perforations resulting from the action of predators); this type of activity is consistent with predation in a living community and is also suggestive of natural deposition of at least some of the molluscs recovered.

184

Catriona Pickard, Adina Boroneanț and Clive Bonsall

Analysis of molluscan assemblages from the Iron Gates sites is further complicated by geomorphic/sedimentological and pedological processes. These include hillwash and bioturbation including root penetration, and tree ‘throw’, which were identified at some sites (Boroneanţ & Bonsall 2013; Borić et al. 2014), which all obscure stratigraphic detail and limit evaluation of temporal changes in species representation, abundance and/or morphology.

Publication Unfortunately, publication of the molluscan remains recovered from sites in the Iron Gates often comprises little more than a list of the species identified. Discussion of the molluscan remains is often restricted to the perforated shells and shell beads recovered from mortuary contexts. A further issue is the persistent use of colloquial or common names with terms such as ‘edible snail’ and ‘gastropod’ being widely used to describe shell finds. This practice results in confusion in the published literature and difficulty in interpreting the archaeological significance of the molluscan assemblages. For example, 138 perforated shells found with burial M38 at Schela Cladovei and described as ‘coquilles d’escargots’ (V. Boroneanț 1990, 122) are in fact shells of the freshwater gastropod, Lithoglyphus naticoides.

What can the molluscan assemblages from the Iron Gates sites tell us about past human activities and environment? In spite of the problems and caveats mentioned above, the molluscan assemblages from the Iron Gates sites can provide useful information about the human use of molluscs and can supplement data obtained from other environmental proxies.

Human use Food/bait

Larger freshwater bivalves such as the Unio spp. and the large terrestrial gastropods Helix spp. were likely harvested for food, and possibly as fish bait. In most regions of the world where mollusc gathering has been ethnographically documented the primary use was for consumption (Waselkov 1987). Helix spp. are widely harvested as food today. Helix pomatia is one of the most important food species, and is farmed commercially in many regions of Europe. Lubell (2004) presents convincing evidence for the widespread use of land snails as human food at late Pleistocene/early Holocene sites across the circum-Mediterranean. However, in spite of the obvious food value of freshwater mussels, a primary role as human food should not be assumed. Ethnographic records indicate that shellfish intended for human consumption are often cooked to facilitate extraction of flesh (Waselkov 1987). There is little systematic evidence for cooking of

11. Molluscan remains from early to middle Holocene sites

185

Figure 11.2. Unio spp. valves from Schela Cladovei. The specimen on the bottom left has a broken posterior ventral edge, possibly indicating shucking to extract the flesh shellfish at any of the Iron Gates sites. By contrast there is convincing evidence for the physical opening of freshwater mussels at Schela Cladovei. The majority of the near complete freshwater mussel valves had a linear fracture along the ventral edge of the shell suggestive of ‘shucking’ to extract flesh (see Fig. 11.2). The type of fracture observed, a break across the shell rather than along growth lines, argues against simple shell dissolution. Similar patterns of damage to the ventral edge of U. crassus valves is documented at Starčevo (Clason 1980). According to Waselkov (1987) it is unusual to shuck shellfish intended for human consumption; processing methods that leave shell fragments in the flesh are generally avoided. Shellfish processed in this manner may, therefore, have been used as fish bait (Deith 1989). Fish are known to have comprised a significant part of Mesolithic and Early Neolithic human diet at the Iron Gates sites (Bonsall 2008), and freshwater mussels could have served as fish bait. Mussels have been described as ideal fish bait, both as a ground bait and for active fishing with hooks, traps and lines, in marine and freshwaters (Fenton 1984). Shucking of mussels does not entirely preclude use as human food, however. Mussels and other shellfish species are commonly shucked in many recent commercial food industries.

Adornment

Shells of freshwater, marine and terrestrial species were used as a raw material for the manufacture of utensils and/or personal adornments. Many perforated shells collected from the Iron Gates sites have an irregular perforation on the body whorl

186

Catriona Pickard, Adina Boroneanț and Clive Bonsall

of the shell opposite the operculum. Harzhauser and colleagues (2007) postulated that this positioning of the perforation anterior to the penultimate whorl ensured that the operculum of the shell faced outwards when the shell was secured to clothing. Perforated Tritia (syn. Cyclope) neritea, L. naticoides and T. danubialis shells used as adornments (eg, Letica 1969; Păunescu 1970; Srejović 1972) were often recovered from funerary contexts (eg, Lepenski Vir, Schela Cladovei and Vlasac). Fifty perforated T. neritea were found in association with burial 21 at Vlasac (Srejović & Letica 1978), while more recent excavations at this site recovered further examples of T. neritea beads (Borić 2006). Burial M38 at Schela Cladovei was associated with over 100 perforated shells of L. naticoides and over 300 cyprinid pharyngeal teeth appliqués (V. Boroneanț 1990; Mărgărit et al. forthcoming). Perforated L. naticoides were found along with beads of T. danubialis and a dentaliid scaphopod at Cuina Turcului (Păunescu 1970). Shell beads used for personal adornments are known from Mesolithic and Neolithic sites across Europe. For example, pierced specimens of marine (Columbella rustica, T. neritea, Nassarius cf. nitidus and Cerithium cf. vulgatum) and freshwater (Lithoglyphus cf. naticoides and Theodoxus danubialis strangulatus) gastropods were recovered from Mesolithic deposits at Pupićina Cave, Croatia (Komšo 2008). Perforated examples used in headgear are also known from the Early Neolithic LBK site of Kleinhadersdorf, Austria (Harzhauser et al. 2007). The ornamental use of T. danubialis has been recorded at Essenbach-Ammerbreite (Brink-Kloke 1990). Shells of the closely related species Theodoxus fluviatilis were among the most common personal adornments found in association with human remains at the late Mesolithic site of Moita do Sebastião in Portugal and are recorded at the roughly contemporaneous sites of Cabeço da Arruda and Cabeço da Amoreira (Roche 1972; Lentacker 1991). Found in the area of the waist and neck of six of the interred, the shells may originally have been strung as necklaces and belts. Almost 3000 perforated T. fluviatilis shells were found in the late Mesolithic grave of a male at Rollmannsberg, Criewen, Germany (Geissler & Wetzel 1999). Thirty specimens of a related freshwater species, Theodoxus gregarius were found along with perforated L. naticoides among mortuary deposits at Öfnet, Germany. According to Rähle (1978) these adornments may have been woven onto garments or possibly worn as hair ornaments.

Raw material Generally, the shells of terrestrial molluscs lack the mechanical strength of freshwater and marine species. Nevertheless, terrestrial molluscs, particularly the larger, more robust species, were sometimes used as a raw material for the manufacture of adornments and utensils (eg, Bar-Yosef Mayer 2013). In the Iron Gates, ‘beads’ made from the shells of terrestrial molluscs are rare, and there is only limited evidence for the use of molluscan shells generally in the manufacture of other kinds of artefacts. Ten of the freshwater mussel (Unio sp.) valves recovered in the 1992–6 excavations at Schela Cladovei were modified with a circular perforation adjacent to the umbone (see Fig 11.3). The perforations are of a uniform size, 6–7 mm in diameter, and the size, shape and location of the perforations are consistent with artificial modification. While

11. Molluscan remains from early to middle Holocene sites

187

the nacreous ‘mother-of-pearl’ interior surface of the mussel shell may have been desired for decorative purposes, perforated Unio valves could have been used as fishing lures or as weights for fishing lines or nets (cf. Stewart 1975). In parts of North America, prehistoric people living along rivers such as the Mississippi utilised crushed freshwater mussel shells as temper for their pottery. In spite of the widespread availability of freshwater mussels along the Danube, and their common occurrence in Early Neolithic contexts in the Iron Gates (Table 11.1), there is very little evidence of shell-tempering of pottery, although Boroneanț (2012) reported the occurrence of a few Starčevo-Criș sherds with a mixture of chaff and shell temper from the sites of Alibeg and Climente I (Table 11.1).

Exchange networks Specimens of marine molluscs (eg, Cerithium sp., T. neritea, Glycymeris sp., Spondylus sp. and at least one species of dentaliid scaphopod) recovered from sites such as Băile Herculane, Climente II, Cuina Turcului, Icoana and Vlasac in the Iron Gates and from Vinča-Belo Brdo, a Middle Neolithic tell site upriver from the Iron Gates section of the Danube, suggest the existence of long distance exchange networks (Table 11.1). Some of the marine shells (and the majority of the scaphopods) collected at Vinča-Belo Brdo are likely fossil specimens obtained from fossiliferous deposits close to the site (Dimitrijević 2014). Any non-fossil specimens must have been transported to the Iron Gates sites from coastal regions of the Mediterranean or the Black Sea, a distance of at least 400 km. There is a clear temporal distinction in the species of shellfish selected for the manufacture of ornaments. Mesolithic finds generally comprise artefacts made from small globose species such as the nerites, while in the Neolithic the bivalves, Spondylus and Glycymeris, were preferred. The movement of shell beads through organised exchange networks is evident from at least the Upper Palaeolithic in Europe and likely had its origins in the Middle Stone Age of Africa or the Near East (Álvarez-Fernández 2006; d’Errico et al. 2009).

Harvesting strategies Harvesting strategies employed to collect shellfish may be correlated with species ecology, indicated by size structure of the shell assemblage, and inferred from the ethnographically observed practices of traditional shell-fishers. Three principal methods of mussel harvesting have been recorded ethnographically: (i) hand collection, (ii) dredging with rakes or nets, and (iii) brailing (Claassen 1998). Hand gathering in shallow water is the most commonly recorded practice. However, rakes or nets may be employed in the gathering of semi-infaunal specimens or in deeper water. Harvesting mussels by eliciting a closure response or ‘brailing’ is a commonplace practice recorded in both traditional and commercial mussel fisheries. Mussels gape when feeding: dragging a brail (eg, a tree branch or a line of hooks) over the shells causes the valves to close and

188

Catriona Pickard, Adina Boroneanț and Clive Bonsall

the shellfish are then lifted from the water (Lund 1995). Although there is little or no evidence for artefacts associated with the collection of shellfish at any of the Iron Gates sites, the technological capability to produce shellfish harvesting equipment is attested by finds at other early to middle Holocene sites. Small points of bone and wood found at Tybrind Vig, Denmark, and Hohen Viecheln, Germany, could have been hafted to form the prongs of dredging rakes Figure 11.3. Unio sp. valve from Schela Cladovei with a (Gehl 1961; Andersen 1995). Fragments of plant-fibre circular perforation near the umbone nets have been preserved in waterlogged contexts at several sites including Friesack, Germany, Kunda, Estonia, and Tybrind Vig, Denmark (Indreko 1948; Gramsch 1992; Andersen 1995). Harvesting of freshwater mussels along the Iron Gates reach of the Danube may have been facilitated by the spawning behaviour observed in modern Unio crassus populations. Mature female U. crassus move into near-shore waters during the early part of the day in the spawning season, most likely related to maximising dispersal of glochidia or larvae. Migration to shallow waters and conspicuous spurting behaviour make this species particularly vulnerable to collection from riverbanks at this time (Vincenti 2005). A size-selective harvesting technique such as hand-collection or dredging with large mesh nets is suggested at Schela Cladovei by the relatively uniform size of the Unio cf. crassus valves recovered from archaeological contexts. Ninety percent of the intact Unio valves recovered in the 1992–6 excavations at Schela Cladovei measured >40 mm in length, with a median value of 47.0 mm. U. crassus populations from Crişul Alb River, western Romania, have been observed to reach sexual maturity at 3–4 years, which corresponds to a shell length of c. 40 mm (Sárkány-Kiss 1997) suggesting that large specimens were preferentially collected.

Palaeo-environmental reconstruction Palaeo-environmental studies in the Iron Gates have been few and largely restricted to the application of pollen analysis, which at Icoana, Lepenski Vir and Vlasac was interpreted as indicating a forested environment during the Mesolithic (Radovanović 1996).

11. Molluscan remains from early to middle Holocene sites

189

Analysis of the ecology of aquatic and terrestrial mollusc species can provide supplementary data (Davies 2008). Helix spp. were collected in large numbers at Padina and Starčevo (Clason 1980) and also at Schela Cladovei. These terrestrial molluscs preferentially inhabit warm, wet, calcium-rich soils such as those found along the alluvial terraces of the Iron Gates stretch of the Danube. Today both H. pomatia and C. aspersum can be found in gardens, shrubby vegetation or areas of open woodland in the region. The majority of the terrestrial taxa identified at Schela Cladovei (Table 11.1) are unlikely to have been of economic or wider social significance and are expected to have been incidental to or post-date human activity on the site. Despite the meticulous use of wet sieving (using a 1 mm mesh) and/or flotation (using a 250 µm collecting mesh) of excavated sediments in the 1992–6 excavations at Schela Cladovei, the remains of terrestrial molluscs are scant (both in terms of taxonomic diversity and overall abundance). Generally, the density of land molluscs is taken as a measure of the suitability of a habitat to molluscan life; low densities and limited taxonomic diversity are equated with a ‘simple’ habitat, ie, characterised by short grasses (Davies 2008). Bar-Yosef Mayer (2013) suggested that the lack of incidental/autocthonous land snails (n=5) at the Neolithic site of Çatalhöyük in southeast Turkey might be a consequence of the mud-brick architecture at the site resulting in a low humidity environment that did not support land snails. Particularly wet or swampy and disturbed or heavily grazed areas may also be unsuitable habitats for land snails (Davies 2008). With the exception of Helix spp., the remains of terrestrial molluscs were recovered in very small numbers at Schela Cladovei. However, it is worth noting that four of the taxa identified (Cepaea vindobonenis, Chrondrula tridens, Xerocrassa sp. and Zebrina detrita) preferentially occupy open calcareous habitats with short, shrubby vegetation. Ceciliodes acicula is a subterranean (ie, burrowing) species; it may not be contemporaneous with other elements of the molluscan assemblage, emphasizing that the death assemblage may not represent the living community (Davies 2008). The open landscape suggested by the land mollusc species recovered at Schela Cladovei is in contrast to the wooded environment indicated by palynological analyses undertaken at sites in the Iron Gates gorge. Arguably, molluscan remains indicate the palaeo-environment in the immediate vicinity of the collection site – long distance movement is generally physiologically and topographically inhibited (Davies 2008) – while generally pollen is indicative of the wider landscape. Freshwater shells recovered from the Iron Gates sites may serve as indicators of the past aquatic environment. Members of the genus Theodoxus generally occupy the faster, shallower stretches of rivers (Fretter & Graham 1962). Four species of Theodoxus have been recorded in the Carpathian Basin: T. danubialis, T. fluviatilis, T. prevostianus and T. transversalis (Sîrbu & Benedek 2005). T. fluviatilis is a recent introduction. A single unmodified example from Schela Cladovei can be identified as T. danubialis; however the poor preservation of the perforated examples prevents species attribution. Lithoglyphus naticoides (83.3%) and T. transversalis (16.7%) are among the most abundant and widespread freshwater gastropods in this region (Sîrbu & Benedek 2005). The present-day predominance of L. naticoides may reflect changing aquatic conditions

190

Catriona Pickard, Adina Boroneanț and Clive Bonsall

following construction of the Iron Gates I and II dams and the consequent reduction in water flow and increased sedimentation rates, which favours the proliferation of L. naticoides and is disadvantageous to oxyphilous genera such as Theodoxus (Popa 2005). Both the nerite T. danubialis and the river snail Viviparus acerosus are species of freshwater mollusc indigenous to the Danube (Bunje 2007; Soes et al. 2009). T. danubialis is a small species that seldom exceeds 10 mm in height; it is generally found in flowing, highly-oxygenated waters and on solid substrates in the Danube and its tributaries (Bunje 2007). V. acerosus is significantly larger and may exceed 50 mm in length. It also prefers hard, calcium-rich substrates and both T. danubialis and V. acerosus can be gathered from shallow waters near riverbanks. At least 19 species and sub-species of native freshwater mussels inhabit the Danube and surrounding river systems today (Sárkány-Kiss et al. 1997; Bódis & Oertel 2005). It is likely that construction of hydroelectric dams on the Danube and other rivers has significantly reduced species diversity in recent years. Freshwater mussels are documented from Mesolithic contexts in at least four sites (Cuina Turcului, Padina, Schela Cladovei and Veterani Terrace) and from Neolithic contexts in at least four sites (Cuina Turcului, Schela Cladovei, Ušće Kameničkog-Potoka and Velesnica) (Table 11.1). They were also recovered from Neolithic contexts at Starčevo and Vinča-Belo Brdo (Dimitrijević 2014). Three species of Unio were identified at the Iron Gates sites; U. crassus, U. pictorum and U. tumidus (U. crassus at Padina and Schela Cladovei, and all three species at Vinča-Belo Brdo). The various species of large freshwater mussels tolerate a wide range of hydrological conditions from fast flowing to stagnant waters and are found over a wide range of water depths. Gills and feeding apparatus must be kept clear of the substrate resulting in high visibility in clear, shallow waters. Prior to the 1940s U. crassus was the most abundant species of Unionidae in Central Europe (Vincenti 2005). A recent survey of the Unionidae population structure on Crişul Alb River, Romania, established that this species still dominates freshwater systems in this region, accounting for 98.5% of unionids recovered (Sárkány-Kiss 1997). U. crassus is a sub-littoral and epifaunal or semi-infaunal species, often encountered partially buried in soft substrates (Bauer 2001). Intra-species variation in morphology and dimensions of freshwater unionids has been observed in different aquatic environments (eg, Green et al. 1989). For example, shell sculpture may reflect aquatic conditions; unsculptured species are associated with coarse substrates, while sculptured species are associated with fine substrates (Hornbach et al. 2010). Sculpture aids attachment and protects against scour but inhibits burrowing. Smooth species may occupy waters with significant variation in water flow leading to disruption of the substrate and dislodging of shells – the smooth shell allowing fast burrowing following dislocation (Watters 1994). Shell obesity (ie, the length to width ratio of the complete bivalve) may also correspond to hydrological conditions – increased obesity is evident in unionids inhabiting waters with higher discharge or faster flow (Claassen 1998). However, the association of shell morphology and sculpture with environment has not been consistently observed. Zieritz and Aldridge (2009) argue that the degree of dorsal arching is the trait most representative of aquatic environment in unionids – greater dorsal arching is seen in shells inhabiting faster moving water, an adaptation likely associated with improved anchoring.

11. Molluscan remains from early to middle Holocene sites

191

Reconstruction of the aquatic environment by assessing the pattern of variation in shell width and length, ie, obesity of unionids, was limited by the absence of intact, hinged or paired, bivalves in the archaeological assemblage at Schela Cladovei. However, the complete single Unio sp. valves from Schela Cladovei have a straight dorsal edge and a pointed posterior suggesting that the shells developed in relatively slow flowing waters.

Conclusions In spite of the many deficiencies associated with the Iron Gates molluscan assemblages, information about the human use of molluscs and palaeo-environment can be recovered. Freshwater species such as river mussels (Unio spp.) and large edible terrestrial snails (Helix spp.) were likely harvested for food, but possibly also as fish-bait. Both freshwater and marine shells were also used as raw material for the manufacture of personal adornments and utensils. Marine species attest to long distance mobility or participation in exchange networks throughout the Mesolithic–Early Neolithic time range. Although the palaeo-environmental information derived from the terrestrial species must be treated cautiously in light of the small number of specimens recovered, the contrasting evidence from palynological remains points to the need for further environmental research at sites in the Iron Gates. The findings of this review emphasise the need to maximise the information that can be derived from molluscan assemblages by including environmental specialists in the planning of excavations, in order to ensure that sampling strategies are adequate. Intra and inter-site comparison of shell assemblages is significantly affected by post excavation sample processing strategies. Flotation and wet sieving of excavated sediments using fine collecting meshes should be adopted as a standard procedure for mollusc recovery on all archaeological sites.

References Allen, M. J. 2017a. Land snails in archaeology. In Allen, M. J. (ed.), Molluscs in Archaeology, 6–29. Oxford: Oxbow Books Allen, M. J. 2017b. The geoarchaeology of context: sampling for land snails (on archaeological sites and colluvium). In Allen, M. J. (ed.), Molluscs in Archaeology, 30–47. Oxford: Oxbow Books Álvarez-Fernández, E. 2006. Personal ornaments made from mollusc shells in Europe during the Upper Palaeolithic and Mesolithic: News and Views. In Çakırlar, C. (ed.), Archaeomalacology Revisited. Non-dietary use of Molluscs in Archaeological Settings, 1–8. Oxford: Oxbow Books Álvarez-Fernández, E. 2010. Shell beads of the last hunter-gathers and earliest farmers in Southwestern Europe, Munibe 61, 129–138 Andersen, S. 1995. Coastal adaptation and marine exploitation in Late Mesolithic Denmark – with special emphasis on the Limfjord region. In Fischer, A. (ed.), Man and Sea in the Mesolithic, 41–66. Oxford: Oxbow Books Bar Yosef-Mayer, D. E. 2013. Mollusc exploitation at Çatalhöyük. In Hodder, I. (ed.), Humans and

192

Catriona Pickard, Adina Boroneanț and Clive Bonsall

Landscapes of Çatalhöyük. Reports from the 2000–2008 Seasons, 329–338. Çatalhöyük Research Project Series Volume 8/BIAA Monograph 47. London: Cotsen Institute of Archaeology Press Bauer, G. 2001. Life-history variation on different taxonomic levels of Naiads. In Bauer, G. & Wächtler, K. (eds), Ecology and Evolution of the Freshwater Mussels Unionoida, 83–91. Ecological Studies 145. Berlin: Springer Bódis, E. & Oertel, N. 2005. Faunistical and ecological research of mussel species in the Hungarian Danube section, Állattani Közlemények 90, 45–61 Bonsall, C. 2008. The Mesolithic of the Iron Gates. In Bailey, G. & Spikins, P. (eds), Mesolithic Europe, 238–279. Cambridge: University Press Borić, D. 2006. New discoveries at the Mesolithic–Early Neolithic site of Vlasac: preliminary notes, Mesolithic Miscellany 18(1), 7–14 Borić, D., French, C. A. I., Stefanović, S., Dimitrijević, V., Cristiani, E., Gurova, M., Antonović, D., Allué, E. & Filipović, D. 2014. Late Mesolithic lifeways and deathways at Vlasac (Serbia), Journal of Field Archaeology 39(1), 4–31 Boroneanţ, A. 2012. Aspecte ale tranziției de la mezolitic la neoliticul timpuriu în zona Porțile de Fier. Cluj-Napoca: Editura Mega Boroneanţ, A. & Bonsall, C. 2013. The 1965–1968 excavations at Schela Cladovei (Romania) revisited. In Starnini, E. (ed.), Unconformist Archaeology: papers in honour of Paolo Biagi, 35–54. Oxford: British Archaeological Report S2528 Boroneanț, V. 1969. Découverte d’objets d’art épipaléolitique dans la zone des Portes de Fer du Danube, Rivista di Scienze Preistoriche 24(2), 283–298 Boroneanţ, V. 1990. Les enterrements de Schela Cladovei, Nouvelles donnes. In Vermeersch, P. & Van Peer, P. (eds), Contributions to the Mesolithic in Europe, 121–125. Leuven: University Press Boroneanţ, V., Bonsall, C., McSweeney, K., Payton, R. W. & Macklin, M. G. 1999. A Mesolithic burial area at Schela Cladovei, Romania. In Thévenin, A. (ed.), L’Europe des Derniers Chasseurs: Épipaléolithique et Mésolithique, 385–390. Paris, Éditions du Comité des Travaux Historiques et Scientifiques Brink-Kloke, H. 1990. Das linienbandkeramische Graberfeld von Essenbach-Ammerbreite, Ldkr. Landshut, Niederbayern, Germania 68, 427–481 Bunje, P. M. E. 2007. Fluvial range expansion, allopatry, and parallel evolution in a Danubian snail lineage, Biological Journal of the Linnean Society 90, 603–617 Campbell, G. 2017. The collection, processing, and curation of archaeological marine shells. In Allen, M. J. (ed.), Molluscs in Archaeology, 273–288. Oxford: Oxbow Books Claassen, C. 1998. Shells. Cambridge: University Press Clason, A. T. 1980. Padina and Starčevo. Game, fish and cattle, Palaeohistoria 22, 142–173 Cristiani, E. & Borić, D. 2012. 8500-year-old Late Mesolithic garment embroidery from Vlasac (Serbia): technological, use-wear and residue analyses, Journal of Archaeological Science 39, 3450–3469 d’Errico, F., Vanhaeren, M., Barton, N., Bouzouggar, A., Mienis, H., Richter, D., Hublin, J.-J. P., McPherron, S. & Lozouet, P. 2009. Additional evidence on the use of personal ornaments in the Middle Paleolithic of North Africa, Proceedings of the National Academy of Sciences of the USA 106(38), 16051–16056 Davies, P. 2008. Snails: archaeology and landscape change. Oxford: Oxbow Books Deith, M. R. 1989. Clams and salmonberries: interpreting seasonality data from shells. In Bonsall, C. (ed.), The Mesolithic in Europe, 73–79. Edinburgh: John Donald Dimitrijević, V. 2008. Lepenski Vir animal bones: what was left in the houses? In Bonsall C., Boroneanţ V. & Radovanović I. (eds), The Iron Gates in Prehistory: new perspectives, 117–130. Oxford: British Archaeological Report S1893

11. Molluscan remains from early to middle Holocene sites

193

Dimitrijević, V. 2014. The provenance and use of fossil scaphopod shells at the Late Neolithic/ Eneolithic Site Vinča – Belo Brdo, Serbia. In Szabó, K., Dupont, C., Dimitrijević, Gómez Gastélum, L. & Serrand, N. (eds), Archaeomalacology: shells in the archaeological record, 33–41. Oxford: British Archaeological Report S2666 Dimitrijević V. & Tripković B. 2006. Spondylus and Glycymeris bracelets: trade reflections at Neolithic Vinča – Belo Brdo, Documenta Praehistorica 33: 237–252 Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press Fenton, A. 1984. Notes on shellfish as food and bait in Scotland. In Gunda, B. (ed.), The Fishing Culture of the World, 121–141. Budapest: Akadémiai Kiadó Fretter, V. & Graham, A. 1962. British Prosobranch Molluscs: their functional anatomy and ecology. Publication of the British Ray Society 144. London: British Ray Society Gehl, O. 1961. Die Säugetiere. In Schuldt, E. (ed.), Hohen Viecheln. Ein Mittelsteinzeitlicher Wohnplatz in Mecklenburg, 43–60. Berlin: Deutsche Akademie für Wissenschaft zu Berlin Geissler, H. & Wetzel, G. 1999. Mittelsteinzeitliche und mittelalterliche Bestattungen vom ‘Rollmannsberg’ bei Criewen, Lkr. Uckermark. In Cziesla, E., Kersting, T. & Pratsch, S. (eds), Den Bogen spannen ... Festschrift für Bernhard Gramsch, 259–280. Weissbach: Beier & Beran Gramsch, B. 1992. Friesack Mesolithic wetlands. In Coles, B. (ed.), The Wetland Revolution in Prehistory, 65–72. Exeter: WARP Green, R. H., Bailey, R. C., Hinch, S. G., Metcalfe, J. L. & Young, V. H. 1989. Use of freshwater mussels (Bivalvia: Unionidae) to monitor the nearshore environment of lakes, Journal of Great Lakes Research 15, 635–644 Harzhauser, M., Lenneis, E. & Neugebauer-Maresch, C. 2007. Freshwater gastropods as Neolithic adornment: size selectiveness and perforation morphology as a result of grinding techniques, Annalen des Naturhistorischen Museums in Wien 108 A, 1–13 Hornbach, D. J., Kurth, V. J. & Hove, M. C. 2010. Variation in freshwater mussel shell sculpture and shape along a river gradient, American Midland Naturalist 164(1), 22–36 Indreko, R. 1948. Die Mittlere Steinzeit in Estland. Kungl Vitterhets Historie och Antikvitets Akademiens Handlingar 60. Stockholm: Nordstedt Komšo, D. 2008. The Mesolithic in Croatia, Opvscvla Archaeologica 30, 55–92 Lentacker, A., 1991. Archeozoologisch Onderzoek van Laat-Prehistorische Vindplaatsen uit Portugal. Unpublished dissertation, University of Ghent Letica, Z. 1969. Vlasac – nouvel habitat de la culture de Lepenski Vir à Djerdap, Archaeologia Iugoslavia 10, 7–11 Lubell, D., 2004. Prehistoric edible land snails in the circum-Mediterranean: the archaeological evidence. In Brugal, J.-P. & Desse, J. (eds), Petits animaux et sociétés humaines. Du complément alimentaire aux ressources utilitaires, 77–98. Antibes: Éditions APDCA Lund, J. 1995. Flatheads and Spooneys: fishing for a living in the Ohio River Valley. Ohio River Series. Lexington: University Press of Kentucky Madec, L., Desbuquois, C. & Coutellec-Vreto, M. A. 2000. Phenotypic plasticity in reproductive traits: importance in the life history of Helix aspersa (Mollusca: Helicidae) in a recently colonized habitat, Biological Journal of the Linnean Society 69, 25–39 Mărgărit, M., Radu, V., Boroneanț, A. & Bonsall, C. Forthcoming. Experimental studies of personal ornaments from the Iron Gates Mesolithic. Archaeological and Anthropological Sciences Păunescu, A. 1970. Epipaleoliticul de la Cuina Turcului-Dubova, Studii și Cercetări de Istorie Veche 21, 3–47 Păunescu, A. 2000. Paleoliticul și Mezoliticul din Spațiul cuprins între Carpați și Dunăre. București: Editura AGIR Pickard, C. & Bonsall, C. 2014. Mesolithic and Neolithic shell middens in western Scotland: a comparative analysis of shellfish exploitation patterns. In Roksandic, M., Mendonça, S., Eggers.

194

Catriona Pickard, Adina Boroneanț and Clive Bonsall

S., Burchell, M. & Klokler, D. (eds), The Cultural Dynamics of Shell Middens and Shell Mounds: a worldwide perspective, 251–266. Albuquerque: University of New Mexico Press Popa, O. 2005. Contributions to the knowledge of the mollusks from the Romanian sector of the Danube, Travaux du Museum National d’Histoire Naturelle ‘Grigore Antipa’ 48, 7–19 Radovanović, I. 1996. Iron Gates Mesolithic. Ann Arbor: University of Michigan Press Rähle, W. 1978. Schmuckschnecken aus mesolithischen Kulturschichten Süddeutschlands und ihre Herkunft (Probstfeld, Falkensteinhöhle, Burghöhle Dietfurt, Zigeunerfels, Große Ofnet). In Taute, W. (ed.), Das Mesolithikum in Süddeutschland Teil 2: Naturwissenschaftliche Untersuchungen, 163–168. Tübingen: Leidorf Roche, J. 1972. Les amas coquilliers (concheiros) mésolithique de Muge (Portugal). In Schwabedissen, H. (ed.), Die Anfänge des Neolithikums vom Orient bis Nordeuropa, 72–107. Köln: Böhlau Sárkány-Kiss, A. 1997. Structure and aspects of dynamic of the unionid associations of the Crişul Alb/Fehér-Körös river at Ineu. In Sárkány-Kiss, A. & Hamar, J. (eds), The Criş/Körös River Valley. A Study of the Geography, Hydrobiology and Ecology of the River System and its Environment, 203–207. Szolnok: Tisza Klub for Environment and Nature Sárkány-Kiss, A., Boloş, F. & Nagy, E. 1997. Freshwater molluscs from the Criş/Körös rivers. In Sárkány-Kiss, A. & Hamar, J. (eds), The Criş/Körös River Valley. A Study of the Geography, Hydrobiology and Ecology of the River System and its Environment, 195–202. Szolnok: Tisza Klub for Environment and Nature Sîrbu, I. & Benedek, A. M. 2005. The genus Theodoxus Mon fort 1810 (Mollusca, Gastropoda, Neritidae) in the Romanian Inner Carpathian Basin, Scientific Annals of the Danube Delta Institute 11, 92–98 Soes, M. D., Glöer, P. & de Winter A. J. 2009. Viviparus acerosus (Bourguignat, 1862) (Gastropoda: Viviparidae), a new exotic snail species for the Dutch fauna, Aquatic Invasions 4, 373–375 Srejović, D. 1972. Europe’s First Monumental Sculpture: new discoveries at Lepenski Vir. London: Thames & Hudson Srejović, D. & Letica, Z. 1978. Vlasac. Mezolitsko naselje u Djerdapu, I–II. Beograd: Serbian Academy of Science Stanković, S. 1986a. Knjepište – une station du groupe de Starčevo, Đerdapske sveske 3, 447–452 Stanković, S. 1986b. Embouchure du ruisseau Kamenički Potok – site du groupe de Starčevo, Đerdapske sveske 3, 467–471 Stefanović, S., & Borić, D. 2008. The newborn infant burials from Lepenski Vir: in pursuit of contextual meanings. In Bonsall, C., Boroneanţ, V. & Radovanović, I. (eds), The Iron Gates in Prehistory, new perspectives, 131–169. Oxford: British Archaeological Report S1893 Stewart, H. 1975. Indian Artefacts of the Northwest Coast. Washington: University Press Taylor, V. & Bell, M. 2017. Land mollusc middens. In Allen, M. J. (ed.), Molluscs in Archaeology, 195–212. Oxford: Oxbow Books Vasić, R. 1986. Compte-rendue des fouilles du site préhistorique Velesnica, Đerdapske Sveske 3, 264–285 Vincenti, H. 2005. Unusual spurting behaviour of the freshwater mussel Unio crassus, Journal of Molluscan Studies 71, 409–410 Waselkov, G. A. 1987. Shellfish gathering and shell midden archaeology, Advances in Archaeological Method and Theory 10, 93–210 Watters, G. T. 1994. Form and function of unionoidean shell sculpture and shape (Bivalvia), American Malacological Bulletin 11, 1–20 Zieritz, A. & Aldridge, D. C. 2009. Identification of ecophenotypic trends within three European freshwater mussel species (Bivalvia: Unionoida) using traditional and modern morphometric techniques, Biological Journal of the Linnean Society 98, 814–825

12. Land mollusc middens Victoria K. Taylor and Martin Bell

Archaeologists have long recognised middens of marine mollusc shells. As demonstrated in other chapters, marine shell middens are an important category of archaeological site, from which we can learn much about hunter-gatherer subsistence strategies and settlement patterns; however, they are not the only places where large accumulations of molluscs are found. A second, less widely recognised, class of shell midden, similar in size and formation to marine shell middens, is composed of terrestrial mollusc species and frequently found inland rather than in coastal locations. These land snail middens are often referred to as escargotières, from the French term denoting a place where snails are raised. They are also less frequently called rammadiya (Lubell 2001) and cendrières (Gobert 1937; Morel 1974), both of which refer to the considerable amounts of ash often associated with snail middens. They are composed of large volumes of land snails; often of one, or a narrow range of, species, within an ashy matrix which also contains mammal bones, charcoal, plant macrofossils and lithics. A good example is Pond’s description of the Capsian escargotières in Algeria as ‘a group of refuse heaps welded into a single mound … composed of snail shells, camp fire ashes, hearth stones, animal bones and tools of bone and flint’ (Pond et al. 1938, 109). In research terms the land mollusc middens present questions similar to those of the coastal marine mollusc middens, such as those along the Atlantic seaboard of Europe (Milner et al. 2007), particularly Denmark (Andersen 2000), and in many other parts of the world (Bailey et al. 2013). In both cases the quantities of shells are enormous but their significance in the diet must be evaluated alongside other, less obvious, animal and plant resources, requiring detailed laboratory analysis. There is also the question of whether such a concentration of food debris is itself indicative of sedentary communities. In the case of some examples of both Atlantic marine mollusc middens and Mediterranean land mollusc middens, theories of sedentism have been strengthened by the occurrence of burials associated with the middens. The assumption is that burial is more likely to occur when settled populations identify with a specific place. This issue is illustrated by the Taforalt case study below.

196

Victoria K. Taylor and Martin Bell

Figure 12.1. Map of land mollusc middens in the Mediterranean and Near East based on Lubell (2004b) with additions

Distribution of land snail middens Land snail middens are a widespread phenomenon occurring throughout southern Europe, the Near East and North Africa (Fig. 12.1), yet few have been excavated and recorded in detail (Lubell 2004b; Rabett et al. 2010). Fortunately the evidence from several sites has recently been reviewed in a special volume of Quaternary International (Lubell & Barton 2011). Some of the most well-known land snail middens are located in North Africa, particularly in the Maghreb (Morocco, Tunisia and Algeria) and in Cyrenaica in Libya. The archaeology of these middens is mainly attributable to two cultural groups; the Iberomaurusian and the Capsian. Those belonging to the former are found in caves and rockshelters, often near the coast, and date to between 17,000 and 11,000 BP (Lubell 1984; 2001). Those belonging to the Capsian tend to be open-air sites

12. Land mollusc middens

197

which occur further inland, mainly in Algeria and Tunisia, and are Holocene, dating between 11,000 and 6000 cal BP (Lubell 2001). Key sites in North Africa include: Taforalt, also known as Grotte des Pigeons (Taylor et al. 2011; Taylor 2014), Ifri n’Ammar (Moser 2003; Hutterer et al. 2014), Ifri Oudadane (Morales et al. 2013) and Taghit Haddouch (Hutterer et al. 2014) in Morocco; Tamar Hat (Saxon et al. 1974) and Aïn Mistehiya (Lubell et al. 1975; 1976) in Algeria; and Haua Fteah in Libya (McBurney 1967; Hill et al. 2015). After North Africa, the French Pyrenees has one of the highest concentrations of land snail middens. Important sites in the region include Grotte de Poeymaü and Mas d’Azil (Bahn 1983a; 1983b). Land snail middens can also be found in other European countries such as Croatia, where Pupićina Cave has been investigated (Miracle 1995; 2001); Italy (Lubell et al. 1995; Bonizzoni et al. 2009; Lubell 2004a); Portugal (Lubell 2004a) and Spain. Some of the most well-known shell middens in Spain are the Asturian middens along the Cantabrian coast (Aparicio 2001; Lubell 2004a). Although known for their marine shell component, many also contain large numbers of land snail shells, as recorded at La Fragua Cave (Gutierrez-Zugasti 2011). Land snail middens can also be found outside of this region, such as at Nerja Cave in Andalucía (Auro-Tortosa et al. 2002) and a chain of sites close to the east coast of Spain (Lloveras et al. 2011; FernándezLópez de Pablo et al. 2011). Smaller accumulations of land snails are also regularly found at archaeological sites in the Zagros Mountains (Eastern Iraq and Western Iran), with ongoing work in the region by the Central Zagros Archaeological Project continuing to produce evidence of small accumulations of Helix salamonica at Neolithic sites including Bestansur, Sheikh-e Abad and Jani (Shillito 2013; Iversen 2015). Recent work by Rabett et al. (2010) has highlighted the presence of a land snail-dominated midden in Hang Boi Cave (Fortune Teller’s Cave) in Trang An Park in Vietnam. Exploitation of the Giant Land Snail (family Achatinidae) is also reported in the Middle Stone Age Bushman Rock Shelter in South Africa, where some were heat affected (Badenhorst & Plug 2012), and there is also possible evidence for their consumption in later Stone Age contexts in Kuumbi Cave, Zanzibar (Shipton et al. 2016). These sites push the known distribution of sites well beyond the circum-Mediterranean. It seems likely that the distribution of evidence for land mollusc consumption will continue to expand as archaeologists become more aware of their potential contributions to the diet. In contexts where there are accumulations of shells without substantial middens, consideration must be given to whether the accumulation results from human activity, or could be a natural death assemblage (Girod 2011). The latter will generally be characterised by a range of growth stages and species, many not edible, and the absence of associated anthropogenic artefacts. Assemblages derived from human consumption are generally fully grown and of one, or a narrow range of, edible species, occur in specific contexts with cultural material, and show evidence of repetitive middenforming behaviour; shells are also often heat-affected. The earliest clearly defined land snail middens are Upper Palaeolithic in date, the best examples of which are those associated with the Late Stone Age Iberomaurusian culture in North Africa such as Taforalt, Ifri n’Ammar and Tamar Hat. The earliest substantial mollusc midden layers are c. 18,800 cal BP (Unit IV) at Tamar Hat (Saxon et al. 1974,

198

Victoria K. Taylor and Martin Bell

50; Hogue & Barton 2016). There is a notable increase in the number and distribution of land snail middens in the early Holocene, with the majority of sites being Mesolithic in date, such as the Azilian middens in northern Spain and the Pyrenees. La Fragua Cave in Spanish Cantabria contains a substantial midden estimated at 15,000 land snail shells, beginning c. 10,900 cal BP (Gutierrez Zugasti 2011). The Capsian middens in North Africa also date to the early Holocene; however, many continue into the Neolithic period with land snail consumption continuing alongside early domestication at sites such as Ifri Oudadane (Lubell et al. 1976). A similar pattern is seen in the Zagros Mountains where land snails are again found in contexts containing early evidence for animal domestication (Shillito 2013; Iversen 2015). There is also evidence for land snail consumption into the Neolithic, Roman and Hellenic periods in Libya indicating that ‘eating of gastropods seems to have been a consistent feature of the coastal Cyrenaican sites through the Holocene’ (Hunt et al. 2011, 24). Consumption of a wide range of land mollusc species continues to this day in Mediterranean countries and beyond. Helix pomatia eaten as escargot is particularly well known and Cornu aspersum (formerly Helix aspersa) is similarly consumed. The first is thought to have been introduced to Britain by the Romans for food (Davies 2010) but there is no obvious reason why that did not also apply to the second which has been considered an accidental Roman introduction. In Portugal Theba pisana is a traditional dish with some 4000 tonnes being consumed annually. Of particular interest is that present day land snail consumption in parts of the Mediterranean seems to be associated with special festivals and gatherings at certain times of year. Examples are snail festivals at Caragol, Spain in May, Graffignano, Italy in August and Digoin, France in August, at all of which vast quantities of snails are consumed (Taylor 2014). Ethnohistoric practices associated with recent snail gathering have been particularly well documented in Crete where consumption is particularly associated with festivals before Easter and in mid-August (N. Galanidou pers. comm.). The Cretan festivals occur at times where snails were particularly abundant and easily gathered. Such events serve to remind us that land molluscs may be seen not just as an everyday item of diet, or something to be eaten when other resources were scarce; indeed the ethnohistoric evidence often identifies them as a delicacy, and a food of particular social significance because of an association, however created, with special events. Miracle (1995) has interpreted the molluscan evidence from Pupićina Cave in terms of feasting associated with burial practice.

Methods of land mollusc midden investigation Investigation of land mollusc middens requires a strategy carefully constructed to facilitate investigation of the key research questions. The approach is designed to obtain, not just molluscs but other evidence, including plant and animal, which will contribute to an understanding of the diet, environment, and way of life of the people concerned. Some previous investigations have been restricted to the small numbers of larger intact shells (Gutierrez-Zugasti 2011; Lloveras et al. 2011; Lubell et al. 1976), or have used

12. Land mollusc middens

199

large mesh sieves for collection. Such an approach introduces a bias towards larger and more robust species. This is particularly problematic since many land mollusc species are fragile, easily crushed in an active occupation area, and tend to be represented by apices and fragments. Hand collection and larger sieve meshes prevent quantification and also mean the loss of small land mollusc species. Those include species too small for consumption but potentially of palaeo-environmental significance; these can be well represented in shell middens both land and marine (Nielsen 2007). The approach recommended, and used in the Taforalt case study outlined below, is an adaptation of the methodology well established in the environmental analysis of land molluscs (Evans 1972; Davies 2008). Samples are taken in a column of specific dimensions (eg, 0.25 × 0.25 m) and at suitable intervals (eg, 50–200 mm), respecting stratigraphic boundaries, through the thickness of the midden. This facilitates investigation of change through time and quantification of the numbers of shells per unit volume as a proportion of the volume of a stratigraphic horizon, or the total midden. In this way the food resource represented by the midden can be quantified and potentially some estimate of its calorific value obtained. The individual samples are weighed, soaked in water and floating material is washed onto a sieve; 0.5 mm mesh is adequate to retrieve tiny and fragmentary shells (Fernández-López de Pablo et al. 2011; Hunt et al. 2011; Rabett et al. 2010; Taylor et al. 2011), but a finer mesh may be desirable where tiny seeds are also present. Where sieving down to 1 mm or 0.5 mm has been done on land snail midden sites this has facilitated recovery of both large, edible species and smaller species naturally present. Material that does not float is washed onto a nest of sieves eg, 4 mm, 2 mm, 1 mm and 0.5 mm and cleaned with a jet of water. Division into size fractions makes sorting easier. The sieves are dried and the material sorted under a binocular microscope, not only for the molluscs but also the other plant and animal resources which may contribute to an understanding of the diet and environment of the site in question. Archaeological investigation of land Mollusca is relatively straightforward in Britain with a relatively small and well-studied fauna, reasonable knowledge of associated habitat and present-day distributions, good published guides and reference collections (eg, Evans 1972; Kerney & Cameron 1979). A substantial advance has been made with the recent publication of a comprehensive, well-illustrated guide to the Mollusca of the whole of Europe (Welter-Schultes 2012). Even so, the extent of molluscan knowledge varies nationally within Europe and is in general greater for northern than southern Europe. Of the Mediterranean countries where land mollusc middens are found, detailed distributional data is available for Portugal, the Balearics, Malta, Albania, Serbia and Crete and some Aegean islands (Welter-Schultes 2012, 7). Beyond Europe in North Africa and the Near East the faunas, and particularly knowledge of their ecological preferences and distributional ranges, are in general very limited compared to Europe, although some areas are better served than others, for instance Israel (Heller 2009) and Turkey (Schutt 2005). In parts of North Africa and the Near East much of the taxonomic work was done during the period of colonial European rule in the 19th and early 20th century and the coverage is patchy, species description are sometimes limited and there has been a tendency to splitting, with the result that what may be the

200

Victoria K. Taylor and Martin Bell

same species can have multiple names. Working in areas where the mollusc fauna is less well known it is likely to be necessary to carry out work on the present day fauna in order to obtain information on species ecological preferences; we have also found this very helpful in identifying landscape contexts which are particularly suitable for the collection of large numbers of land molluscs. Work in areas where the molluscan fauna is less well known also requires the detailed description and illustration of the species named, to facilitate comparison with those found elsewhere and to contribute eventually to an improved taxonomy. Quantification of mollusc shells is generally based on the minimum number of individuals derived from the counts of apices. The results may then be presented as a histogram of species abundance through the midden, similar to the diagrams used for environmental land snail analysis (Evans 1972). If resources allow it is desirable to obtain sequences of samples from more than one part of a midden in order to investigate lateral variation, which, depending on how it grew, may also equate to a temporal sequence. Where many of the shells are intact, morphometric studies of whole shells can be employed (Claasen 1998), to investigate, for instance, changes in size over time which might indicate decreasing shell size as a result of population over-exploitation (Mannino & Thomas 2001; 2002) or environmental changes. Unfortunately in many land snail middens the shells are highly fragmented, limiting a morphometric approach.

Taforalt land snail midden as a case study Taforalt is a large cave site located in the Beni Snassen mountains in northeastern Morocco, close to the Algerian border, 40 km from the Mediterranean Sea. It is well known for its large Iberomaurusian cemetery at the back of the cave (Ferembach 1962; Humphrey et al. 2012). The site also contains thick anthropogenic deposits which were the subject of large scale excavations in the 1950s (Roche 1963) and more recently between 2003 and 2016 by a joint Moroccan and British team led by Professors Abdeljalil Bouzouggar (Rabat) and Nick Barton (Oxford). The sequence has a high precision chronology modelled from a sequence of 52 AMS radiocarbon dates (Barton et al. 2013). The Iberomaurusian occurs in two distinct units, the upper part of the Yellow Series and the Grey Series (Fig. 12.2). The lower of the two units, the Yellow Series, appears to be sediment washed into the cave; this contains lithic artefacts, bones and some shells. The Grey Series deposits formed between 15,000 and 12,600 cal BP, after which the sequence is truncated, so there are no Holocene sediments. The Grey Series is a very different, essentially anthropogenic deposit, containing far more abundant evidence of human activity such as lithics, artefacts and chips, animal bones, charcoal, land snail shells and stones, some heat-affected (Fig. 12.3). The Grey Series layers at Taforalt have always been described as a ‘land snail midden’. The starting point at Taforalt was therefore to test this hypothesis through detailed scientific analysis. All too often, particularly in Mediterranean contexts, the anthropogenic nature of such deposits has been assumed rather than evaluated. Analysis of bulk samples taken from a 0.25 m wide column was undertaken in order to investigate this hypothesis and address wider questions such as methods of

12. Land mollusc middens

201

collection and consumption, contribution to diet and environmental change (Taylor & Bell forthcoming). The Yellow Series deposits below have been analysed from about 20,760 cal BP (see Fig. 12.4). In the Yellow Series, species which are likely to have been eaten account for only 17% of the total molluscs; these are species also present in the overlying midden suggesting small-scale molluscan consumption from 20,760 cal BP. Indeed individual large shells of Otala punctata were observed in the Taforalt Calcareous Group sediments Layer R26 dated c. 90–95,000 BP, although here there were no concentrations of shells and only scattered worked lithics and charcoal (S. Collcutt pers. comm.). More significant were concentrations of land molluscs in ashy hearth deposits in the Lower Laminated Group Layer R22 which is dated c. 80–82,000 BP (ClarkBalzan 2012; Barton et al. 2014). This layer also contained perforated shell beads of the marine mollusc Nassarius gibbosulus which is regarded as among the earliest evidence of human symbolic behaviour worldwide (Bouzouggar et al. 2007; De Errico et al. 2009). The overlying Grey Series midden, is up to 4 m thick, and, judging by the exposed section, may originally have comprised c. 1500 m³ of largely anthropogenic sediment. Rough calculations based on the number of shells in the sampled column indicate that it may originally have contained something like 62 million shells deposited over about 2400 years, ie, maybe 28,000 per year. Inaccurate as these numbers probably are, they give some indication of the significance of molluscan exploitation. Analysis of the molluscan component identified four main species within the Grey Series deposits at Taforalt: Dupotetia dupotetiana, Otala punctata, Alabastrina soluta and Cernuella globuloidea (Fig. 12.4). Helix aspersa (Cornu aspersum) was also present in smaller numbers. All five species are of a size suitable for consumption and together account for over 99% of the total molluscs recovered from the Grey Series. A clear bias can therefore be seen towards large, edible species in the Grey Series. Lithic artefacts, abundant lithic debitage, animal bone, charcoal and charred plant remains were also recovered from the mollusc samples which supports the hypothesis that the majority of land snails in the Grey Series are anthropogenic in origin. Dupotetia dupotetiana is by far the most commonly occurring species, accounting for over 60% of all the molluscan material in the Grey Series and over half of all apices overall. In comparison, Otala punctata and Alabastrina soluta occur in much lower numbers throughout the sequence. Cernuella globuloidea is also much less frequent, accounting for only 9% of the total edible molluscs, the majority of which come from the lower half of the Grey Series. At around 13,680 cal BP there is a steep decline in this particular species which may be the result of environmental changes. At the same time there is an overall increase in mollusc numbers representing a further intensification in the use of molluscan resources. Thus it appears that the use of molluscan resources began in a small way by about 80,000 BP, they became more consistently used after the last glacial maximum, represented in North Africa by a cooler dry period with much dust input, and saw major intensification with the onset of midden formation c. 15,000 cal BP and further intensification from 13,680 cal BP. The intensification represented by the Grey Series midden from 15,000 cal BP sees a remarkable diversity of dietary resources, this includes extensive evidence for the use of plant resources studied by J. Morales, especially sweet acorns and pine nuts. High

202

Victoria K. Taylor and Martin Bell

Figure 12.2. Taforalt: the section of the Grey Series Iberomaurusian midden and the underlying Yellow Series sediments

Figure 12.3. Mollusca and a range of other biological evidence from the Taforalt midden on the sieve (scale 5 cm)

12. Land mollusc middens

203

Figure 12.4. Taforalt midden mollusc diagram showing percentage of edible and non-edible species, on the right are dates cal BP based on the model in Barton et al. (2013)

204

Victoria K. Taylor and Martin Bell

levels of caries in the human burials are also interpreted as indicating a diet with high levels of consumption of starchy plant foods (Humphrey et al. 2014). Animal bone is also frequent in the Grey Series and a marked increase in sedimentation rate is to a large extent of anthropogenic origin. This intensification is a particularly noteworthy manifestation of the Broad Spectrum Revolution which Flannery (1969) identified in the Middle East in the late glacial and initial Holocene but has since been identified at similar dates in many parts of the world. At Taforalt the onset of the Grey Series midden was followed soon after by development of an extensive cemetery at the back of the cave. Evidence of middening, Broad Spectrum resource utilisation and particularly burial have often been taken as indicating increased sedentism. The plants utilised indicate activity from late spring to autumn and the nuts could have been stored and used over winter (Humphrey et al. 2014). The hypothesis of sedentism can only be fully addressed when the whole range of dietary resources and human skeletal evidence from the site has been put together in the monograph currently in preparation (Barton et al. forthcoming). The same applies to evaluation of the relationships between the molluscan and other environmental and palaeoeconomic evidence and wider evidence for environmental changes. It is notable, however, that the major sedimentary transition marked by the onset of the Grey Series midden coincides with the generally warmer Greenland Interstadial 1 (= Late Glacial Interstadial) (Grootes et al. 1993). A marked decline of Cernuella globuloidea occurs at the time of a short-lived cooler episode Greenland Interstadial 1–1d. The period of most intensive mollusc exploitation occurred following this in the later part of the Greenland Interstadial. There is little indication from the Mollusca of a subsequent change which might correspond to the climatic downturn of Greenland Stadial 1 and indeed the dating sequence indicates that the cave deposits have been truncated to below this level at the point sampled, also resulting in the loss of all Holocene stratigraphy. However, wood charcoals from the top of the sequence, nearer the cave entrance, do indicate the onset of a significant cool damp period with dates within Greenland Stadial 1 (S. Collcutt pers. comm.).

Climate and seasonality A number of studies have suggested that the periodicity of mollusc collection coincided with periods when the molluscs today are observed to be particularly active. For instance in the case of Holocene Iberian example at Balma del Gai, Spain (Lloveras et al. 2011) collection in late summer and autumn was suggested and at Arenal de la Virgen and Casa Corona activity in spring and summer was proposed (Fernández-López de Pablo et al. 2011). At La Fragua the most suitable period for collection was suggested as summer and autumn, although vertebrate faunal evidence indicated that the main period of activity was in winter (Gutiérrez Zugasti 2011). The clustering behaviour of molluscs in spring on woody plants noted in lowlands downriver of Taforalt showed how significant numbers of molluscs could have been collected. However, we must exercise caution in extrapolating from modern analogues to the conditions of the early Holocene and especially the late Pleistocene.

12. Land mollusc middens

205

The analysis of stable isotopes from shells provides new ways of addressing issues of past climate and seasonality (Thomas 2015a; 2015b). Oxygen isotope analysis of sequences of land mollusc shells can provide palaeoclimatic sequences (Leng & Lewis 2014). Analyses of isotopic values of modern land snail shell have been used to demonstrate the relationship between mollusc shell isotope signatures and environmental factors (Yanes et al. 2009; Stott 2002; Zanchetta et al. 2005). These data can then be used to provide a baseline from which to compare archaeological samples (Lécolle 1985; Balakrishnan et al. 2005; Colonese et al. 2010; Kehrwald et al. 2010; Stevens et al. 2012; Yanes et al. 2011). This contributes to a multi proxy approach to reconstruction of past environments and palaeoclimate. For further discussion of methodologies and interpretation of data see Prendergast et al. (2015). Carbon and oxygen isotope analysis of incremental bands in the shells of the African land snail Limicolaria kambeul chudeaui from Ethiopia have provided evidence of climatic seasonality (Leng et al. 1998). Land mollusc shells, at least of some taxa, show evidence of periodic banding both on a coarse scale on the surface of the shell and on a very fine microscopic scale in thin section, particularly in the thickened apertures of some species, eg, Dupotetia dupotetiana at Taforalt (F. Katsi pers. comm.). Periodic banding coupled with isotopic analysis could potentially establish the seasonality of land mollusc collection and, when combined with other sources, test hypotheses of sedentism or mobility.

Shell collection and consumption One question which arises on sites with enormous collections of land mollusc shells is how prehistoric communities gathered such numbers. Confronted with this, some writers have even flirted with the notion that the molluscs were farmed (Bahn 1983a; 1983b; Fernández-Armesto 2001), although no convincing evidence has ever been advanced in support of this idea. It may be more realistic to think in terms of nonanalogue ecological communities in the rapidly-changing climatic conditions of the late glacial and early Holocene, creating particularly favourable conditions for molluscan life round parts of the Mediterranean. Nor can we exclude the possibility that people contributed in some ways to the creation of niches in which these Mollusca flourished, just as the Mollusca and other resources contributed to the creation of niches with a broad spectrum of resources in which some groups became more sedentary. Some indication of how large numbers of shells might be collected are provided by a small scale survey of the present day malacofauna around Taforalt. Close to the cave Alabastrina soluta was to be found in micro-caves in the limestone, apparently the result of solution-etching of the rock by generations of molluscs themselves, as recorded elsewhere in the Mediterranean and on Mendip, UK (Danin 1996; Stanton 1986). Survey in the wider Moulouya Valley between Taforalt and the coast recorded large numbers of Dupotetia aestivating on bushes, often in tight clusters, as shown in Figure 12.5. At one location c. 100 Dupotetia dupotetiana individuals were counted on a single bush. This bush was one of at least ten within a 10 m radius indicating that somewhere in the region of 1000 molluscs could be collected from that small area with minimal effort.

206

Victoria K. Taylor and Martin Bell

Consumption of land snails may be relatively easy to recognise where there are large numbers associated with anthropogenic material in middens. More challenging is the interpretation of small collections on sites where they might be assumed to be of natural occurrence. This is perhaps the case with Helix pomatia and Helix aspersa (Cornu aspersum), both introduced to Britain by the Romans (Davies 2010). In Spain there are records of very substantial middens of Cepaea nemoralis at La Fragua Cave in Cantabria (Gutierrez Zugasti 2011) and Balma del Gai (Lloveras et al. 2011). Cepaea has a wide European distribution and this poses the question of whether it has been overlooked as a potential resource in Britain where, in the first half of the Holocene, it is the only land mollusc of sufficient size for consumption. Various writers have discussed the ways in which molluscs were prepared for consumption. Hutterer et al. (2011; 2014) found small intentional perforation marks on a large percentage of shells in the midden at Taghit Haddouch in northeast Morocco, which he concluded were to break the vacuum so that the snail could be sucked from the shell; these perforations are Figure 12.5. Dupotetia dupotetiana on shrubby not recorded before the Neolithic. vegetation, Moulouya valley, Morocco It has often been suggested that cooking was involved (Lubell et al. 1975; Bar 1977; Bahn 1983; Heller 2009). That possibility is strengthened by the abundance of charred plant material and heat-fractured rocks in several sites, especially Taforalt where up to 60% of Grey Series shells were heat-affected. Today the most commonly employed method for cooking snails is immersing them in boiling water (Arrébola Burgos & Álvarez Halcón 2001) which loosens the muscles and enables the flesh to be easily removed from the shell,

12. Land mollusc middens a

207

b

Figure 12.6. Experiments in land mollusc cookery at Reading University showing (a) roasting of Cornu aspersum maximum (Helix aspersa maxima) (and two Cepaea) on hot rocks, (b) calcined remains of snails in hearth a method which Lubell et al. (1975) believe was used by prehistoric North African communities. They may have used skins, or potentially ceramic vessels at Capsian sites such as Aïn Mistehiya, as containers within which water could be boiled using heated rocks known as ‘pot boilers’. Another possibility is that snails were cooked directly by placing them in the fire bed or onto stones heated in the fire (Bonizonni et al. 2009; Heller 2009; Matteson 1959; Pond et al. 1938), or into large pits lined with heated rocks, a technique for cooking a range of foods which is widely attested through ethnographic studies (Linderman 1962; Wandsnider 1997; Meehan 1982). Experiments in cookery of Cornu aspersum (Helix aspersa) maximum at Reading University showed that they can be very rapidly cooked in boiling water by adding hot rocks to a container, although those roasted on hot rocks were, to modern taste at least, more palatable (Fig. 12.6).

Conclusion Land Mollusc middens have, until recently, seldom received the attention from archaeologists given to middens of marine Mollusca. Land and marine middens are both significant environments of deposition preserving a wide range of palaeo-environmental evidence. When the Mollusca themselves are analysed in detail, as at Taforalt, alongside the other sources of biological evidence using a comparative multi-proxy approach, they can make a significant contribution to study of palaeoeconomy, palaeo-environment, sedentism, mobility, and past diet. Land mollusc exploitation is attested at Taforalt from c. 80,000 BP but that was small scale and episodic. Major intensification in the use of Mollusca took place c. 15,000 cal BP during the latter half of the Late Glacial Interstadial. The midden which formed from 15,000–12,600 cal BP contains a remarkable diversity of food resources and is a classic case of Broad Spectrum resource utilisation in the Late Glacial Interstadial. The Palaeolithic land mollusc middens occur particularly in North Africa with scattered occurrences north of the Mediterranean where most of the middens are Holocene (Fig.

208

Victoria K. Taylor and Martin Bell

12.1). They are mainly of Mesolithic date but there are examples up to Roman and Medieval times and land molluscs are still consumed in large numbers particularly in some religious festivals.

Acknowledgements We are grateful to Professor Nick Barton and Professor Abdeljalil Bouzouggar, leaders of the Taforalt project. Our involvement in that project through the Cemeteries and Sedentism project was funded by the Leverhulme Trust (F/08 735/F). We are also grateful to Dr S. Collcutt (who kindly provided comments on a draft), Dr L. Humphrey, Dr S. Black, Dr J. Hogue, Ingrid Brack and Faidra Katsi and the team of Taforalt researchers as a whole.

References Andersen, S. H. 2000. ‘Køkkenmøddinger’ (Shell middens) in Denmark: a survey, Proceedings of the Prehistoric Society 66, 361–384 Aparicio, M. T. 2001. Malacofauna terrestre del Yacimento de Cubío Redondo (Matienzo, Cantabria), Munibe 53, 61–66 Arrébola Burgos, J. R. & Álvarez Halcón, R. M. 2001. La explotación de los caracoles terrestres en España: Aspectos ecolólogicos y socioculturales, Temas de antropología Aragonesa 11, 139–172 Aura Tortosa, J. E., Jordá Pardo, J. F., Pérez Ripoll, M., Rodrigo Garcıa, M. J., Badal Garcıa, E. & Guillem Calatayud, P. 2002. The far south: the Pleistocene–Holocene transition in Nerja Cave (Andalucıa, Spain), Quaternary International 93, 19–30 Badenhorst, S. & Plug, I. 2012. The faunal remains from the Middle Stone Age levels of Bushman Rock Shelter in South Africa, South African Archaeological Bulletin 67(195), 16–31 Bahn, P. G. 1983a. Pyrenean Prehistory. Warminster: Aris & Phillips. Bahn, P. G. 1983b. Late Pleistocene economies of the French Pyrenees. In Bailey, G. N. (ed.), Hunter-gatherer Economy in Prehistory, 168–186. Cambridge: University Press Bailey, G., Hardy, K. & Camara, A. 2013. Shell Energy: mollusc shells as a coastal resource Oxford: Oxbow Books Balakrishnan, M., Yapp, C. J., Meltzer, D. J., & Theler, J. L. 2005. Paleoenvironment of the Folsom archaeological site, New Mexico, USA, approximately 10,500 14C yr BP as inferred from the stable isotope composition of fossil land snail shells, Quaternary Research 63, 6331–6344 Bar, Z. 1977. Human consumption of land snails in Israel, Basteria 41, 53–58 Barton, R. N. E., Bouzouggar, A., Hogue, J. T., Lee, S., Collcutt, S. N. & Ditchfield, P. 2013. Origins of the Iberomaurusion in NW Africa: new AMS radiocarbon dating of the Middle and Later Stone Age deposits at Taforalt Cave, Morocco, Journal of Human Evolution 65(3), 266–281 Barton, R. N. E., Lane, C. S., Albert, P. G., White, D., Collcutt, S. N., Bouzouggar, A., Ditchfield, P., Farr, L., Oh, A., Ottolini, L., Smith, V. C., Van Peer, P. & Kindermann, K. 2014. The role of cryptotephra in refining the chronology of Late Pleistocene human evolution and cultural change in North Africa, Quaternary Science Review 118, 151–169 Barton, R. N. E., Bouzouggar, A., Collcutt, S. & Humphrey, L. (eds), forthcoming. Cemeteries and Sedentism in the Later Stone Age of NW Africa: excavations at Grotte des Pigeons, Taforalt. Mainz: Römisch-Germanisches Zentralmuseum

12. Land mollusc middens

209

Bonizzoni, L., Bruni, S., Girod, A. & Guglielmi, V. 2009. Archaeometric study of the shells of Hellicidae from the Edera Cave (northeastern Italy), Archaeometry 51(1), 151–173 Bouzouggar, A., Barton, R. N. E., Vanhaeren, M., D’Errico, F., Collcutt, S., Higham, T. F. G., Hodge, R., Parfitt, S., Rhodes, E., Schwenninger, J-L., Stringer, C., Turner, E., Ward, S., Moutmir, A. & Stambouli, A. 2007. 82,000-year-old shell beads from North Africa and implications for the origins of modern human behaviour, Proceedings of the National Academy of Sciences USA, 104(24), 9964–9969 Claassen, C. 1998. Shells. Cambridge: University Press Clark-Balzan, L. A., Candy, I., Schwenniger, J-L., Bouzouggar, A., Blockley, S., Nathan, R. & Barton, R. N. E. 2012. Coupled U-series and OSL dating of a Late Pleistocene cave sediment sequence, Morocco, North Africa: significance for constructing Palaeolithic chronologies, Quaternary Geochronology 12, 53–64 Colonese, A. C., Zanchetta, G., Fallick, A. E., Martini, F., Manganelli, G. & Drysdale, R. N. 2010. Stable isotope composition of Helix ligata (Müller, 1774) from Late Pleistocene-Holocene archaeological record from Grotta della Serratura (Southern Italy): paleoclimatic implications, Global and Planetary Change 71, 249–257 Davies, P. 2008. Snails: archaeology and landscape change. Oxford: Oxbow Books Davies, P. 2010. Land and freshwater molluscs. In O’Connor, T. & Sykes, N., Extinctions and Invasions: a social history of British fauna, 175–180, Oxford: Oxbow Books Danin, A. 1996. Patterns of corrosion and abrasion induced by Mediterranean land snails on limestone rocks, Malacological Review 19, 91–98 D’Errico, F., Vanhaeren, M., Barton, N., Bozouggar, J., Mienis, H., Richter, D., Hublin, J. J., McPherron, S. P. & Lozouet, P. 2009. Additional evidence on the use of personal ornaments in the Middle Paleolithic of North Africa, Proceedings National Academy of Sciences, USA 106, 16051–16056 Evans, J. G. 1972. Land Snails in Archaeology. London: Academic Press Ferembach, D., Dastugue, J. & Poitrat-Targowla, M. J. 1962. La Nécropole Epipaléolithique de Taforalt (Maroc Oriental): Étude des squelettes humans. Rabat: Edita Casablanca Fernández-Armesto, F. 2001. Food: a history. London: Macmillan Fernández-López de Pablo, J., Gómez Puche, M. & Martínez-Ortí, A. 2011. Systematic consumption of non-marine gastropods at open-air Mesolithic sites in the Iberian Mediterranean region, Quaternary International 244(1), 45–53 Flannery, K. V. 1969. Origins and ecological effects of early domestication in Iran and the Near East. In Ucko, P. J & Dimbleby, G. W. (eds), The Domestication and Exploitation of Plants and Animals, 73–100. London: Duckworth Gobert, E. G. 1937. Les escargotières: le mot et la chose, Revue Africaine 81, 639–645 Girod, A. 2011. Land snails from Late Glacial and Early Holocene Italian sites, Quaternary International 244(1), 105–116 Grootes, P. M., Stuiver, M., White, W. C., Johnsen, S. & Jouzel, J. 1993. Comparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores, Nature 366, 552–554 Gutiérrez Zugasti, F. I. 2011. Early Holocene land snail exploitation in northern Spain: the case of La Fragua cave. Environmental Archaeology 16(1), 36–48 Heller, J. 2009. Land Snails of the Land of Israel. Sophia: Pensoft Hill, E. A., Hunt, C. O., Lucarini, G., Mutri, G. & Barker, G. 2015. Gastropod consumption during the late Pleistocene and early Holocene in the Haua Fteah, Libya, Journal of Archaeological Science Reports 4, 320–325 Hogue, J. T. & Barton, R. N. E. 2016. New radiocarbon dates for the earliest Later Stone Age microlithic technology in Northwest Africa, Quaternary International 413A, 62–75. Hunt, C. O., Reynolds, T. G., El-Rishi, H. A., Buzaian, A., Hill, E. & Barker, G. W. 2011. Resource

210

Victoria K. Taylor and Martin Bell

pressure and environmental change on the North African littoral: Epipalaeolithic to Roman gastropods from Cyrenaica, Libya, Quaternary International 244(1), 15–26 Humphrey, L., Bello, S., Turner, E., Bouzouggar, A. & Barton, N. 2012. Iberomaurusian funerary behaviour: evidence from Grotte des Pigeons, Taforalt, Morocco, Journal of Human Evolution 62, 261–273 Humphrey, L. de Groot, I., Morales, J., Barton, N., Collcutt, S., Bronk Ramsay, C. & Bouzouggar, A. 2014. Earliest evidence for caries and exploitation of starchy foods in Pleistocene huntergatherers from Morocco, Proceedings of the National Academy of Sciences 111(3), 954–959 Hutterer, R., Mikdad, A. & Ripken, T. E. J. 2011. Species composition and human exploitation of terrestrial gastropods from Taghit Haddouch, an early Holocene archaeological site in NE Morocco, Archiv für Molluskenkunde 140, 57–75 Hutterer, R., Linstädter, J., Mikdad, A. & Mikdad, A. 2014. Human manipulation of terrestrial gastropods in Neolithic culture groups of NE Morocco, Quaternary International 320, 83–91 Iversen, I. 2015. The molluscs of Bestansur, Iraqi Kurdistan: a case study of Early Holocene Helix salomonica exploitation in the Zagros. Environmental Archaeologies of Neolithisation: Old World case studies. Environmental Archaeology Special Issue 20(3), 239–250 Kehrwald, N. M., McCoy, W. D., Thibeault, J., Burns, S. J. & Oches, E. A. 2010. Paleoclimatic implications of the spatial patterns of modern and LGM European land-snail shell δ 18O, Quaternary Research 74, 166–176 Kerney, M. P. & Cameron, R. A. D. 1979. A Field Guide to the Land Snails of Britain and North-West Europe. London: Collins Lécolle, P. 1985. The oxygen isotope composition of land snail shells as a climatic indicator: applications to hydroecology and paleoclimatology, Chemical Geology 58, 157–181 Leng, M. J., Heaton, T. H. E., Lamb, H. F. & Naggs, F. 1998. Carbon and oxygen isotope variations within the shell of an African land snail (Limicolaria kambeul chudeaui Germain): a highresolution record of climate seasonality? Holocene 8, 407–412 Leng, M. J. & Lewis, J. P. 2014. Oxygen isotopes in Molluscan shell: applications in environmental archaeology, Environmental Archaeology 21(3), 295–306 Linderman, F. B. 1962. Plenty-coups, Chief of the Crows. Lincoln: University of Nebraska Press Lloveras, L., Nadal, J., Garcia Argüellas, P., Fullola, J. M. & Estrada, A. 2011. The land snail midden from Balma del Gai (Barcelona, Spain) and the evolution of terrestrial gastropod consumption during the late Palaeolithic and Mesolithic in eastern Iberia, Quaternary International 244(1), 37–44 Lubell, D. 1984. Paleoenvironments and Epipaleolithic economies in the Maghreb (c. 20000 to 5000 BP). In Clark, J. D. & Brandt, S. (eds), From Hunters to Farmers: considerations of the causes and consequences of food production in Africa, 41–56. Berkeley: University of California Press Lubell, D. 2001. Late Pleistocene–early Holocene Maghreb. In Peregrine, P. N. & Ember, M. (eds), Encyclopedia of Prehistory, Volume 1: Africa, 129–149. New York: Kluwer Academic/Plenum Publishers Lubell, D. 2004a. Are land snails a signature for the Mesolithic–Neolithic transition in the circumMediterranean? In Budja, M. (ed.), The Neolithization of Eurasia – Paradigms, Models and Concepts Involved. Documenta Praehistorica 31, 1–24 Lubell, D. 2004b. Prehistoric edible land snails in the circum-Mediterranean: the archaeological evidence. Petits animaux et sociétés humaines, du complément alimentaire aux ressources utilitaires, XXIVe recontres internationales d’archéologie et d’histoire d’Antibes, 77–98. Antibes: Editions APDCA Lubell, D. & Barton, N. (eds), 2011. Gastropods and humans in the Late Palaeolithic and Mesolithic of the West Mediterranean basin, Quaternary International 244, 1–126 Lubell, D., Ballais, J.-L., Gautier, A., Hassan, A. F., Close, A., Chippindale, C., Elmendorf, C. &

12. Land mollusc middens

211

Aumassip, G. 1975. Prehistoric cultural ecology of Capsian escargotières I: preliminary results of an interdisciplinary investigation in the Chéria-Télidjène region 1972–73, Libyca 23, 44–121 Lubell, D., Hassan, F. A., Gautier, A. & Ballais, J-L. 1976. The Capsian Escargotières. Science 191, 910–920 Lubell, D., Sheppard, P. & Jackes, M. 1984. Continuity in the Epipalaeolithic of northern Africa with an emphasis on the Maghreb, 143–191. In Wendorf, F. & Close, A. (eds), Advances in World Archaeology, Vol. 3. New York: Academic Press Lubell, D., Mussi, M., Musacchio, A. Agostini, S. Coubray, S. & Giraudi, C. 1995. Exploitation of seasonal resources in the mountains of Abruzzo (central Italy) Epigravettian to Neolithic. L’Europe des derniers chasseurs, 5e Colloque International UISPP, 465–472 Mannino, M. & Thomas, K. 2001. Intensive Mesolithic exploitation of coastal resources? Evidence from a shell deposit on the island of Portland (Southern England) for the impact of human foraging populations on intertidal rocky shore molluscs, Journal of Archaeological Science 28, 1101–1114 Mannino, M. & Thomas, K. 2002. Depletion of a resource? The impact of prehistoric human foraging on intertidal mollusc communities and its significance for human settlement, mobility and dispersal, World Archaeology 33, 452–474 Matteson, M. R. 1959. Snails in archaeological sites, American Anthropologist 61, 1094–1096 Meehan, B. 1982. Shell Bed to Shell Midden. Canberra: Australian Institute of Aboriginal Studies. McBurney, C. B. M. (ed.), 1967. The Haua Fteah (Cyrenaica) and the Stone Age of the South-East Mediterranean. Cambridge: University Press. Milner, N., Craig, O. & Bailey, G. N. (eds) 2007. Shell Middens in Atlantic Europe. Oxford: Oxbow Books Miracle, P. 1995. Broad-Spectrum Adaptations Re-Examined: hunter-gatherer responses to Late Glacial environmental changes in the Eastern Adriatic. Unpublished PhD Thesis, University of Michigan Miracle, P. T. 2001. Feast or famine? Epipalaeolithic subsistence in the northern Adriatic basin. Documenta Archaeologica 28, 177–197 Morales, J., Pérez-Jordà, G., Peña-Chocarro, L., Zapata, L., Ruíz-Alonso, M., López-Sáez, J. A. & Linstädter, J. 2013. The origins of agriculture in North-West Africa: macro-botanical remains from Epipalaeolithic and Early Neolithic levels of Ifri Oudadane (Morocco), Journal of Archaeological Science 40, 2659–2669 Morel, J. 1974. La Faune de l’escargotière de Dra-Mta-El-Ma-El-Abiod (Sud-Algérien). Ce qu’elle nous apprend de l’alimentation et des conditions de vie des populations du Capsien supérieur. L’Anthropologie 78(2), 299–320 Moser, J. 2003. La grotte d’Ifri n’Ammar. Tome 1: L’Ibéromaurusien, Forschungen zur Allgemeinen und Vergleichenden Archäologie 8, 1–163 Nielsen, N. 2007. Land snails and shell middens: a new approach in Danish archaeological research. In Milner, N., Craig, O. & Bailey, G. N. (eds), Shell Middens of Atlantic Europe, 70–77. Oxford: Oxbow Books Pond, A. W., Chapuis, L., Romer, A. S. & Baker, F. C. 1938. Prehistoric Habitation Sites in the Sahara and North Africa. Logan Museum Bulletin, 5. Beloit: Beloit College Prendergast, A. L., Stevens, R. E., Hill, E. A., Hunt, C., O’Connell, T. C. & Barker, G. W. 2015. Carbon isotope signatures from land snail shells: implications for palaeovegetation reconstruction in the eastern Mediterranean, Quaternary International [10.1016/j.quaint.2014.12.053] Rabett, R., Appleby, J., Blyth, A., Farr, L., Gallou, A., Griffiths, T., Hawkes, J., Marcus, D., Marlow, L., Morley, M., Tan, N.C., Son, N.V., Penkman, K., Reynolds, T., Stimpson, C. & Szabo, K. 2010. Inland shell midden site-formation: Investigation into a late Pleistocene to early Holocene midden from Trang An, Northern Vietnam, Quaternary International 239, 153–169

212

Victoria K. Taylor and Martin Bell

Roche, J. 1963. L’Epipaléolithique Marocaine. Lisbon: Foundation Calouste Gulbenkian Saxon, E. C., Close, A., Cluzel, C., Morse, V. & Shackleton, N. J. 1974. Results of recent excavations at Tamar Hat, Libyca 22, 49–91 Schutt, H. 2005. Turkish Land Snails. Solingen: Nature und Wissenschaft. Shillito, L-M. 2013. Molluscs from Sheikh-e Abad and Jani. In: Matthews, R., Matthews, W. & Mohammadifar, Y. (eds), The Earliest Neolithic of Iran, 2008 Excavations at Sheikh-e Abad and Jani, 201–205. British Institute of Persian Studies Archaeological Monograph 4. Oxford: Oxbow Books. Shipton, C., Crowther, A., Kourampus, N., Prendergast, M. E., Horton, M. & Douka, K. 2016 Reinvestigation of Kuumbi Cave, Zanzibar, revealing Later Stone Age, early Holocene abandonment and Iron Age reoccupation, Azania: Archaeological Research in Africa 51(2), 197–233 Stanton, W. I. 1986. Snail holes (Helixigenic cavities) in hard limestone – an aid to the interpretation of karst landforms, Proceedings of the University of Bristol Spelaeological Society 17(3), 218–226 Stevens, R. E., Metcalfe, S. E., Leng, M. J., Lamb, A. L., Sloane, H. J., Naranjo, E & González, S. 2012. Reconstruction of late Pleistocene climate in the Valsequillo Basin (central Mexico) through isotopic analysis of terrestrial and freshwater snails, Palaeogeography, Palaeoclimatology and Palaeoecology 319–320, 16–27 Stott, L. D. 2002. The influence of diet on the δ13C of shell carbon in the pulmonate snail Helix aspersa, Earth and Planetary Science Letters 195, 249–259 Taylor, V. K. 2014 Land snail middens in Late Pleistocene and Early Holocene in North Africa: a case study from Taforalt. Unpublished PhD thesis, University of Reading Taylor, V. K. & Bell, M. G. forthcoming. Land Mollusca. In Barton, R. N. E, Bouzouggar, A. Collcutt, S. & Humphrey, L. (eds), Cemeteries and Sedentism in the Later Stone Age of NW Africa: excavations at Grotte des Pigeons, Taforalt. Mainz: Römisch-Germanisches Zentralmuseum Taylor, V. K., Barton, R. N. E., Bell, M., Collcutt, S., Black, S. & Hogue, J. T. 2011. The Epipalaeolithic (Iberomaurusian) at Grotte des Pigeons, Taforalt, Morocco: a preliminary study of the land Mollusca, Quaternary International 244, 5–14 Thomas, K. D. 2015a. Molluscs emergent, Part I: themes and trends in the scientific investigation of mollusc shells as resources for archaeological research, Journal of Archaeological Science 56, 133–140 Thomas, K. D. 2015b. Molluscs emergent, Part II: themes and trends in the scientific investigation of mollusc shells as resources for archaeological research, Journal of Archaeological Science 56, 159–167 Wandsnider, L. 1997. The roasted and the boiled: food composition and heat treatment with special emphasis on pit-hearth cooking, Journal of Anthropological Archaeology 16, 1–48 Welter-Schultes, F. W. 2012. European Non-marine Molluscs, a Guide for Species Identification: Bestimmungsbuch Für Europäische Land- und Süsswassermollusken. Göttingen: Planet Poster Editions Yanes, Y., Romanek, C. S., Delgado, A., Brant, H. A., Noakes, J. A., Alonso, M. R. & Ibáñez, M. 2009. Oxygen and carbon stable isotopes of modern land snail shells as environmental indicators from a low-latitude oceanic island, Geochimica Cosmochimica Acta 73, 4077–4099 Yanes, Y., Yapp, C. J., Ibáñez, M., Alonso, M., De-la-Nuez, J., Queseda, M. L., Castillo, C. & Delgado, A. 2011. Pleistocene-Holocene environmental change in the Canary Archipelago as inferred from the stable isotope composition of land snail shells, Quaternary Research 75, 658–669 Zanchetta, G., Leone, G., Fallick, A. E. & Bonadonna, F. P. 2005. Oxygen isotope composition of living land snail shells: data from Italy, Palaeogeography, Palaeoclimatology, Palaeoecology 223, 20–33

Part 3 Marine and food and diet

13. Marine molluscs from archaeological contexts: how they can inform interpretations of former economies and environments Liz Somerville, Janice Light and Michael J. Allen

Marine shells are a relatively frequent find on archaeological sites of all ages, both with access to the coast as well as sites further inland (Light 2003a). Collection by modern humans goes back to Palaeolithic times (eg, Jerardino & Marean 2010) and is a behaviour shared with the Neanderthals (eg, Zilhão et al. 2010). This chapter provides an introduction to marine molluscs and the part that their shells have played in archaeology: this will be considered with respect to excavation, retrieval, analysis and interpretation throughout an archaeological project. The value of marine shells to archaeological site interpretations will be outlined, as well as the roles and contributions that marine molluscs have played in the lives of the inhabitants of former occupation sites of past human communities. While the methods of analysis and approach described are primarily applicable to the UK and northwest Europe, other chapters in the book provide examples from the wider geographical context of the Mediterranean, the Levant and southeast Europe. Gastropods and bivalves are the most common marine shell component to be found at archaeological sites. Molluscs are an important food source right up to the present day and often the shells represent food waste or bait residue. In this chapter we deal principally with marine shell as a by-product of human food, although middens are dealt with more specifically by Hardy (Chapter 15).

Marine molluscs: a brief introduction Current estimates of present day marine mollusc species richness are between 45,000 and 50,000 (Appeltans et al. 2012; Rosenberg 2014). Through adaptive radiation molluscs have colonised the marine realm from the extremes of abyssal depths and mid-ocean ridges to the littoral zone around the world, including marginal habitats such as saltmarshes and mangrove forests. This versatility is made possible by their morphological diversity, modes of life and behaviour. Of interest archaeologically are

13. Marine molluscs from archaeological contexts

215

the two principal classes, Gastropoda and Bivalvia, and of minor interest Cephalopoda (octopus and squid) and Scaphopoda (tusk shells sometimes referred to by the generic name Dentalium). Marine molluscs inhabit the full range of marine environments and habitats from hard rocky substrates through to soft sediments: gravels, sands, silts, muds. They live high on the shore, ranging from the supralittoral zone, into the sublittoral environment extending down to considerable depths. Their varying morphologies (shape, size and, to a lesser extent, colour and robustness) reflect their habitat and mode of life. Marine mollusc animals are foraged, harvested and farmed for food or collected for bait, and the dead shells are collected as a raw material for tools or ornaments. With some species, both animal and shell are exploited from one individual. This versatility, and diversity of usage, enables their occurrences in excavated material to be of value. Some of the most common shells recovered on archeological sites in northwest Europe are oyster (Ostrea edulis), mussels (Mytilus edulis), limpets (Patella spp.) cockles (Cerastoderma spp.) and winkles (Littorina littorea). The basic mollusc types (gastropods and bivalves) are shown in Figure 13.1, and shows appropriate parameters for measuring shells. Shell features can be usefully diagnostic: some gastropods have tooth-like sculpture around the aperture. Bivalve shells bear scars of muscular and other soft part attachment on the inner surfaces of valves and the dentition at the umbo in the region of the ligament is particularly instructive.

Gastropods This class makes up about 80% of named species of mollusc, both marine and nonmarine. All gastropods are free-living, and their diet is very variable; some are grazers, carnivores or filter-feeders, whilst a minority are specialised parasites/symbionts. Some marine gastropods are non-shell-bearing (sea-slugs) but most consist of a single shell, frequently spirally coiled. In some species, for example Cowries, the coiling is not evident as the protoconch spiral is masked by the morphology of the adult shell. Similarly, for limpets which are common seashore gastropods, the juvenile spiral shell is subsumed by rapid expansion of the broadly conical adult shell. Marine gastropods inhabit the widest variety of marine habitats and may be free-living as shallow burrowers in soft sediments, for example the common whelk (Buccinum undatum), or they may live as part of the motile epifauna on rock or other hard lithological substrata.

Bivalves Bivalve shells consist of two components (rarely more) and the two valves articulate about a hinge to form a ‘clam’. Most bivalves are infaunal, living in soft sediments and feed by filtering, whether as suspension-feeders or deposit-feeders. They feed by means of inhalant siphons but some are epifaunal and live attached, sometimes permanently,

216

Liz Somerville, Janice Light and Michael J. Allen

13. Marine molluscs from archaeological contexts

217

Figure 13.1. Upper: the main features of gastropod (left) and bivalve interior (right). Lower: measurements used for whole shells. H = height; W = width; L = length. Top row: Gastropods. Arrow indicates apex. Shell shapes (left to right) are Periwinkle (no siphon); Whelk (siphon); Limpet shown in profile (above) and underside/interior (below) showing position of the muscle scar. Bottom row: Bivalves. Arrow indicates umbo. Shell shapes (left to right) are carpet-shell, showing exterior of left valve; mussel, showing exterior of right valve; clam showing both valves (illustration: Abby George) Note: some early career analysts have in the past have transposed length and width terminology when measuring shells, making their results incompatible with the large collective datasets and with most analyses and reports by other analysts. This is, perhaps, even more worrying where, on the rare occasion, such analyses have been published.

to hard substrates, notably oysters and mussels. One bivalve group is nektonic (ie, they swim and migrate easily), such as the Pectinidae (scallops).

Scaphopods This is a minor marine molluscan class of some 900 species. Known as tusk shells, based on their morphology, they are infaunal or surface dwelling on soft sediments and might be collected from shallow water or from the intertidal zone. They are also washed ashore from sublittoral habitats. They offer no potential as a food species, but can be strung and used for money, known in native American cultures as Wampum (Stearns 1999), and are a useful raw material for such items as beads, belts and headdresses.

Cephalopods This is an exclusively marine molluscan class of bilaterally symmetrical animals with a head, mantle and a set of arms/tentacles. There are 800 species. One subclass contains octopuses, cuttlefish and squid. The former contain no hard parts other than a chitinous ‘beak’, but cuttlefish and squid carry either an internal phragmocone (cuttlefish) or a chitinous gladius or pen (squid) (Fig. 13.2) as well as a ‘beak’. Nautilus has an external shell, is a deep water genus, and shells might occasionally be washed ashore and collected.

Marine shells in archaeology Archaeological sites vary considerably in the amount of marine shell they yield, and the excavation team will already have some idea of the basic information that can be obtained, and the key questions that can be addressed in a routine excavation project. The retrieval of the marine shell and the information their occurrences may provide

218

Liz Somerville, Janice Light and Michael J. Allen

Figure 13.2 Top: Phragmocone of the cuttlefish, Sepia orbignyana (Férussac) and fragment showing porous aragonite interior; Middle: Phragmocone of Sepia officinalis (L.) Note both Sepia examples are oriented anterior to the right; Bottom: Gladius, also known as the ‘pen’, of the squid, Loligo sp. The gladius is principally chitinous. Orientation is anterior to the right (photos: J. Light)

should be considered in advance, ideally with appropriate molluscan specialist input. If it is known or thought that the site will yield significant assemblages, or middens, of marine shell then clear sampling strategies should be devised in advance (eg, see Luff & Rowley-Conwy 1994). Shell middens form the basis of Chapter 15 (Hardy, this volume). Oysters also require particular considerations in retrieval, sampling, processing and analysis (Winder, Chapter 14). There is no single formula for a project or research design. Often, with large assemblages, an assessment will help identify the questions, themes and potential of that assemblage. Assessments can be crucial at the post excavation stage and can help define the appropriate level of analysis that may be warranted. Therefore, prior to carrying out an analysis the specialist and the archaeological team need to agree on project aims, the level of analysis, and a definition of what is worth analysing. Full detailed measuring and recording is not required on all assemblages, and in some cases particular species may be isolated for more detailed treatment than others. Defining these aims at the outset ensures that appropriate techniques are applied to the project under consideration. The resolution of analysis will depend on the size of the assemblage and upon archaeological parameters such as context, phasing and dating.

13. Marine molluscs from archaeological contexts

219

Marine molluscs on archaeological sites Types of marine shell deposit in archaeology During excavation of a site it will become evident that the style of marine shell discard/ deposition varies. The manner in which those shells became incorporated provides early insights into the mechanism of transport, their purposes and uses. Campbell (2017a; Chapter 16) has identified a hierarchy of deposit types of which the four principal ones are given below: • Discrete shell-rich scatters: where heaps or lenses of shell retaining integrity occur as isolated spreads which may overlap nearby scatters. These may represent single discard events and form part of paths, tracks, yard surfaces. • Homogeneous shell-rich masses, such horizons having greater depth than the scatters but little structure within the layer. The shells are likely to have been deposited against walls or in ditches and pits with a period of accumulation. • Middens are widely recognised at archaeological sites where a location has received repeated deposits of shell over a longer interval, although the discarded shell may represent the residue not only of food but of multiple shell-use activities of economic significance. • Unusual, exotic, and non-native species can occur as single items or as a minor component. There may be shells that are not normally associated with food collection. An example would be the perforated shell of the great scallop, Pecten maximus, which holds a special place in the Christian mind as a badge of pilgrimage (Saul 1974).

Shells can, however, also be distributed widely at lower densities across a site and can provide, collectively, assemblages large enough to undertake varying levels of analysis. None of the analyses and interpretation outlined below can be realised without the acquisition of appropriate assemblages. Shells are frequently found in low densities in many contexts, especially ditches and pits, across excavation sites and are commonly collected during hand excavation. Where dumps and caches of shells are encountered, or spreads and middens excavated, a clear sampling and sieving strategy should be devised and deployed with the aim of recovering representative and appropriate-sized assemblages of at least 200–600 shells for detailed analysis (see Campbell, Chapter 16).

Food: collection/harvesting, processing, consumption and discard The majority of, but by no means all, excavated marine shells are related to human nutrition. A project/analysis can investigate the use of the molluscs as a food source, and applied aspects such as changes in species preference over time, harvesting preferences, locations of on-site consumption and discard. The range of the species present may indicate the nature of former coastlines, the persistence of mollusc species locally and exploitation practices in relation to shore type on those coastlines be they rocky, sandy, muddy, estuarine. Another useful indicator of provenance lies in the associated fauna, for example species that infest marine molluscs (epibionts). Some are biogeographically constrained and may allow the analyst to make inferences about possible areas of coast from which marine molluscs were collected (eg, Winder’s studies of oysters, Chapter 14).

220

Liz Somerville, Janice Light and Michael J. Allen

Palaeo-environments Marine shell both from archaeological sites and from natural accumulations can also be used in palaeo-environmental investigations, reviewed recently by Thomas (2015a; 2015b). Shells may accumulate in fine increments, preserving information about harvest season (Laurie 2008), sea temperature and past climate (Fenger et al. 2007), heavy-metal pollution (Labonne et al. 1998; Gillikin et al. 2005) and absolute dating (Demarchi et al. 2011; Russell et al. 1998): see Andrus (2011) for a review. The shells of tiny molluscs (microshells), which typically live attached to algae and interstitially in fine sediments, as well as those marine molluscs whose maximum shell dimension falls below, say, 10mm when adult, are not likely to be food residue. Their occurrence may relate to former environments (sand dunes), or accidental incorporation through exploitation of resources, such as water, reeds, or mud. Inevitably, when seaweed is harvested for soil manuring, the associated fauna represents bycatch (Bell 1981; Law 2013). Considering this from the other perspective, the presence of certain mollusc species which are associated with seaweeds can indicate the use of this resource (eg, Ainis et al. 2014, for the California Channel Islands). At Ardnave (Islay, west Scotland), Evans (1983, 356) noted that the presence of the blue-rayed limpet (Patella pellucida) could indicate presence of the kelp (Laminaria), and that the flat winkle (Littorina littoralis) could likewise indicate the presence of brown wrack (Fucus spp.). Windblown sand is a persistent vector in transporting adventives and in this way shell is incorporated into the machair environments such as those along the western seaboard of Britain and Ireland and the Western Isles (Law 2013; Law & Thew, Chapter 5), and into extensive dune systems, for example at Gwithian, Cornwall (Walker, Chapter 4).

Shells as artefacts and exotic shells Shells themselves may be used for a wide range of personal, artefactual and architectural ornamentation (see Ridout-Sharpe, Chapter 17) or as tools (Szabó, Chapter 18). Such use has a very long history of 100,000 years (eg, Vanhaeren et al. 2006). The analyst should take into account the patterns of wear (Light, Chapter 20), and possible modification by the human hand. Amongst exotic shells, cowries (the Cypraeidae family) are found quite frequently in archaeological contexts in the UK (eg, Jackson 1917; Light 2007; Reese 1991). Larger species may have totemic use, but others exchanged hands as currency (Cypraea moneta, C. annulus; Safer & Gill 1982) and trade may have brought these to the northwest European archaeological sites. Alternatively, exotic shell might represent waste from post-medieval industries such as button-making. Simmonds (1879) lists the remarkable tonnage of imports for this use. Shells can also derive from other industrial activities. Charred shells may occur as the waste product of salt-working, where they were incorporated in intertidal clay, which was processed in salterns for the evaporation of brine. Such sites occur in the English Fens, for example on Morton Fen in Lincolnshire (Murphy 2001). Further, cockle shells (Cerastoderma edule), were used as a raw material in medieval lime kilns, for instance at the Gilberd School excavations, Colchester. The lime was used for various purposes

13. Marine molluscs from archaeological contexts

221

including soil improvement, and as a component of building mortar (Murphy 1992). The production of purple dye from various species of mollusc within the superfamily Muricidae, was a substantial and widespread industry in the Mediterranean in former times and is well-documented (Light & Walker, Chapter 19). A smaller scale industry for which that Classical area is also known, and of which there are still practitioners in the present day, is the production of so-called ‘Cloth of Gold’ (also known as Pinna silk or wool, marine silk); a fine thread with a golden sheen which is spun from the long, fine byssal fibres of the fan mussel (Atrina (Pinna) spp.), (Lovell 1867; Yeats 1872), and see discussions in Treasures of the Sea (Enegren & Meo 2017). Marine shells are also used for shell trumpets, rattles, toys, game-pieces and fish hooks, among many other things.

Analytical approach Individual specialists have their own approach and method of working. Analysis should comprise a sequence of steps and processes in order to extract as much useful information from the shell assemblage. A common standard is imperative as it allows intra-site and inter-site comparisons on a chronological or regional scale, the latter by reference to published research and analysis. It is important that basic information conforms to recognised parameters; that it is comparable with both standard anatomical biological measurements and criteria, and the work of other specialists in the discipline. This enables comparison with modern populations of species under consideration. Consulting relevant publications (eg, Box 1), both during the analytical work and the writing of the report can inform the interpretation of the assemblage. This may be particularly helpful with respect to the selection of suitable statistical tests. Depending on the quantity and quality of the shell assemblage which is submitted for assessment, and the requirements of the excavation team, an analysis of the mollusc shell retrieved during excavation may encompass all or some of the following. The level of analysis may depend upon the nature and dating of the contexts, but more detailed analyses may only be undertaken when shell numbers are high enough (eg, 200–600+, see Campbell 2017a; Chapter 16).

Pre-analysis examination of shells Shells should initially be examined to obtain an impression of the overall condition of the assemblage and any observations noted. It is advisable to have a spreadsheet or database at an early stage in the analysis in order to record all potentially useful information. Any bivalves which are present as closed pairs should be set to one side to avoid disarticulation. The extent to which shells will need to be cleaned of adhering sediment will also become obvious at this stage. Any lumps of aggregated soil should be teased apart to look for other shells or fragments. Other calcareous faunal remains such as crab carapace or bird egg-shell fragments may sometimes be mistakenly assigned to molluscs by archaeologists at the sorting stage. Crabs, being crustaceans,

222

Liz Somerville, Janice Light and Michael J. Allen

do not fall within the remit of a mollusc analysis. Although they are usually a very small component, the analysis of crab shell residue may be within the competence of the mollusc specialist (eg, Light 2005a), but normally this material will need to be re-bagged, labelled and returned to the excavator, together with any other faunal remains (eg, sea-urchin test fragments, bones) and artefacts (eg, pottery) to pass on to the appropriate specialist.

Cleaning

Excavation teams vary in their approach and techniques with regard to cleaning shells, and the methods they choose will be determined by the site conditions, soil type and other variables. In general, the specialist should receive an assemblage where the shells have had excess soil removed, whether by washing, dry brushing or other methods. When submitted for assessment/analysis the shells should be sufficiently clean to allow species identification, which often uses shell sculpture, measurement, and examination of traces and marks or damage to shell exteriors. A certain amount of adherent soil need not impede examination of the shells. Often, with the exception of oysters, simple immersion in water and gentle washing will suffice. Oysters require special treatment and are considered by Winder (Chapter 14). Some further dry-brushing depending on the fragility of shells may be appropriate. Soil needs to be carefully removed from the internal surface of limpets to reveal the orientation of the muscle scar (see Figure 13.1); from the internal surface of bivalves to show the muscle scars; and from the aperture of gastropods to reveal identifying features. Pairs of bivalve shells need to be carefully cleaned and gently separated, they should then be kept together as a pair through the subsequent processing of the shell.

Identification

As with all archaeozoological analyses access to a reliable reference collection is essential. Textbooks, no matter how authoritative or comprehensive, can never be a substitute for reference material of both fresh and archaeological specimens. Archaeological specimens usually lose the outer periostracum and natural colour may be lost with this. When examining shells, the level of preservation may alter the appearance; factors such as fragmentation, abrasion, loss of colour and staining can impede accurate identification. Shells can be heavily stained by the substrate and sediments in which they are preserved, often leaving shells orange (iron rich) or grey. Abrasion, weathering and chemical alteration during preservation may distort shells, making comparator shells (modern and archaeological) invaluable. A further complication can arise where epifauna (epibionts) are attached to shells, see below. Identification may rely not just on overall shape and form, but, especially for partial shells, on details of external sculpture, the hinge (umbo) in bivalves, internal features (muscle scars), and apertural ‘teeth’ in gastropods. To build up a personal collection, marine shells can be collected from the beach (labelling with a note of locality and date of collection is essential). Modern native shell food remains (eg, mussels, cockles, winkles, clams) are also useful. Reference collections are also available in museums and similar institutions, university departments and commercial archaeological companies.

13. Marine molluscs from archaeological contexts

223

Box 1 Useful sources for shell identification There is no substitute for a reference collection Fish, J. D. & Fish, S. 2011. A Student’s Guide to the Seashore. Third edition. Cambridge: Cambridge University Press Graham, A. 1988. Molluscs: prosobranch and pyramidellid gastropods. Second edition. Synopses of the British fauna (New Series). Leiden: Brill Hayward, P. Nelson-Smith, T. & Shields, C. 1996. Collins Pocket Guide: Sea shore of Britain and Europe. London: HarperCollins McMillan, N. F. 1968. British Seashells. London: Frederick Warne Tebble, N. 1966. British Bivalve Seashells. London: British Museum (Natural History) Yonge, C. M. 1966. The Sea Shore. London: Collins

Key websites

The Conchological Society of Great Britain and Ireland. http://www.conchsoc.org The National Museum of Wales (Oliver, P. G. et al. 2015). http://naturalhistory. museumwales.ac.uk/britishbivalves MarLIN. http://www.marlin.ac.uk/ Archaeomalacology Working Group (facebook page). https://www.facebook.com/ groups/250778365036844/

An adjunct to a reference collection is a library of appropriate taxonomic literature. Digital interfaces are increasingly indispensable as a resource. However such sources, both hard copy and digital, are not infallible; mollusc specialists can err when consulting two-dimensional images. A small bibliography of essential publications and websites which are relevant for northwest Europe is provided (Box 1), and a more complete list of analysts’ reference texts is given in Box 3. In addition there is a huge molluscan literature which can be tracked using the internet. It may not always be possible, or necessary to identify to species level. For example, within the cockles, only the fine detail of the shell sculpture distinguishes the three Acanthocardia species found around Britain (Tebble 1966) and this character is not reliably distinctive on archaeological specimens, making it sensible to take the identification here only to genus. Complete valves of the common edible cockle Cerastoderma edule can be distinguished from C. glaucum by overall shape, longer internal grooves extending from the ventral margin and details of the morphology of the hinge (Tebble 1966). Incomplete shells and fragments cannot always be confidently assigned to species. In the case of limpets, the differences between Patella vulgata and P. depressa and P. ulyssiponensis cannot readily be seen in archaeological material. On west and southwest coasts of the British Isles Mytilus galloprovincialis lives alongside M. edulis, and these are also difficult to distinguish. Thus for such instances, identification should be left at genus level and

224

Liz Somerville, Janice Light and Michael J. Allen

in practice it is rarely necessary to take the identification further for the purposes of archaeological analysis.

Epibionts (organisms which live on or parasitise marine molluscs) The shells of marine molluscs play host to a number of other organisms (epibionts). The usefulness of these in archaeomalacology has been explored for the most part for oysters, more so than for other species (see Winder, Chapter 14). Epibionts may leave traces on the shell such as single or multiple holes, calcareous tubes or scars. Preliminary inspection of the assemblage during the first stage of recording (or assessment) will indicate whether undertaking this additional examination of whole shells for epibionts is appropriate. Traces may be visible on incomplete shells and notes about any such observations should be made. Examination often needs to be carried out with magnification. Shells can be damaged by the burrowing polychaete Polydora ciliata, and in larger shells, particularly oysters, by Polydora hoplura; both leaving small entry and exit holes. Each species constructs a distinctive burrow. The burrowing sponge Cliona celata may also affect shells leaving a series of holes and, in extreme cases, producing galleries within the shell. However, single holes in shells are more likely to be evidence of predation by gastropods such as the dog whelk (Nucella lapillus) or naticids (so-called necklace or moon snails), especially if the inside of the hole has a drilled appearance. Because species are constrained by habitat and environment, as well as biogeographical parameters, identification of epibionts may assist in determining locations at varying scales, or environments, of the marine molluscs prior to collection.

Quantification The calculation of Minimum Number of Individuals (MNI) is a tool for quantification in archaeological faunal assemblages for the purposes of data quality, comparability and repeatability. Because many of the shells which have been excavated are partial, it must be explicit at the outset how MNI has been calculated. A standard method which has been used routinely is to count a selected range of Non-Repetitive Elements (NREs). This is popular for its analytic speed and efficiency. Usually it involves gastropod apex frequency, and for bivalves, this involves that portion of the valve which supports the umbo, or hinge. In the case of bivalves the resulting total has to be halved to give an estimate of the number of individuals. This assumes a similar number of right and left valves in the assemblage, which may not be the case. None the less, NRE-based MNI is a frequently used index. Giovas (2009) provides a thorough review of analytical methods and problems in counting shells for analysis and outlines some of the approaches to choice of unit to be counted. Ultimately she proposes that if the shell analyst seeks only general approximations, NRE-based MNI can aid in the analysis of larger numbers of specimens

13. Marine molluscs from archaeological contexts

225

and, perhaps, in learning ultimately, more than might otherwise be possible. However, she also finds that NRE-based MNI calculations yield variable results depending on a number of assemblage factors, such as fragmentation, that may not be detectable except through the analysis of every specimen. While use of NREs is expedient and may be appropriate in some cases, the comparative, assemblage-level MNI totals reveal that use of NREs always underestimates MNI. Some more comprehensive and transparent protocols involving detailed metrical analyses have been proposed by other authors, eg, Campbell (2014) and Harris et al. (2015) The latter authors have suggested that the standard method outlined above can underestimate the relative abundance of some taxa. Applying a case study from the Marshall Islands they have outlined a new protocol (tMNI) that incorporates a wider range of NRE and calculates MNI based on the most frequently occurring NRE for each taxon. In their case study they showed an increase of 167% in relative abundance of gastropods and 3% increase in bivalves which changed the rank order of taxon abundance.

Sizing Where a shell assemblage consists of large numbers of shells and subject to the requirements of the analysis to be carried out, it will not be practical to measure each individual shell (Fig 13.1). An approach should be devised that aims to provide the archaeologists with the appropriate level of information that they require for the assemblage under consideration. Sometimes (as was the case at an archaeological site known as Atlantic Road at Fistral Bay, in Cornwall; Light 2003b) the huge numbers of shells precluded exhaustive measurement. In such cases an approach can be devised to allow an assessment of the number of shells that had been collected by the inhabitants of the archaeological site, using a combination of shell counts and weights of known samples. Where measurement is deemed useful for intra-site/inter-assemblage comparison, or wider empirical comparison with other assemblages then basic standard and uniform records should be made. The majority of bivalves are equivalve ie, the right and left valves have the same dimensions, and it is therefore generally necessary to measure one valve only. Most larger shells (more than c. 20 mm) are measured to the nearest millimetre or 0.5 mm. Accurate measurement of some shells may be best with callipers (eg, gastropods), whereas a sliding measurement board (similar to an osteometric board; Brothwell 1972, figs 32 & 37) can be used for bivalves and limpets, although the length of small limpets can be measured with callipers. Alternatively shells which lie flat can be placed onto graph paper taking care to avoid parallax error when parts of the shell stand away from the paper. For most spiral gastropod shells a maximum height dimension consists of the vertical measurement from shell apex to the lowest basal point of the shell. For limpets a basic anterior-posterior shell dimension (length) can be used but if they are a significant component of marine shell assemblage, shells can be sorted into appropriate size cohorts by shell length at intervals of, for example 5 mm, 55% in the 19–22 mm or 26 mm range, and the sharp cut-off eliminating small specimens may suggest a degree of selection. This was also seen in the much smaller sample of cockles from Bishopstone (Somerville 2010). This analysis uses percentages because of the large variation in the number of whole

13. Marine molluscs from archaeological contexts

229

valves from each of these periods. Nonetheless, there is an apparent trend to increasing size over time (13th century average 22 mm; to 14th and 15th century average 24 mm). Over-exploitation of marine mollusc populations may have measurable and recordable effects on the living populations, and on the resultant catch as inferred from the archaeological assemblage. Reduction in sizes of individual shells is a common effect. Steele and Klein (2008) recorded that the size of limpets for a variety of species has significantly decreased in middens in South Africa between Middle Stone Age (>40–≤127 kya) and later Stone Age (0.7–0.11 kya). However, Mannino and Thomas (2002) point out the difficulty of distinguishing between environmental effects, overexploitation by humans or changes in human foraging strategies as causes of changes in shell size. One important criterion is to be able to determine the biological age of the shells. For many species this requires sectioning of the shell to count growth lines (eg, Claassen 1998), but in some (eg, Phorcus (Monodonta) lineata toothed top shell) it is possible to count the growth lines (varices) visible on well preserved shells and Mannino and Thomas (2001) show that for the Culverwell (Dorset) midden, there was a trend to towards taking younger shells, which may explain the observed decrease in size. Their interpretation is that predation removed animals which would have survived to the older and larger sizes in an unexploited population. Notably Milner et al. (2007) found a marked decrease in the size of limpets at Quoygrew (Orkney) from the 10th century AD to the 11th–13th century deposits. They ascertained age from the visible winter growth checks on the exterior of the shell and found that the older age classes were missing from the later assemblage. Specific selection of the younger (smaller) individuals can, however, be for a number of reasons: larger limpets tend to be very tough and may be specifically avoided. The calorific value of marine molluscs can be calculated and their significance to a community’s diet can be inferred. A large number of taphonomic factors need consideration relating to the proportion of the molluscs recovered and represented on site, including the understanding what proportion of the consumed shells were even brought back to the site, let alone recovered. Nevertheless, the record of the minimum number of individuals (MNI) can be used directly in the calculation of meat weight. This will only be a rough approximation as basing such calculations on MNI requires the use of average figures for meat yield. Winder (1980) gives a useful set of estimates for this for a range of common species. For marine molluscs, such a calculation is often a salutary reminder of just how many would need to be gathered to provide sufficient food as a single food resource. Bailey (1978) supplies a comparison in terms of calories for a single red deer carcass; 52,267 oysters or 156,800 cockles or 31,360 limpets.

Provenance Determining the provenance, or provenances, for the marine mollusc assemblage on site, can provide the evidence to indicate the location of the coastline exploited, and the range of substrates from which the variety of molluscs were collected ie, from littoral rocky, sandy or muddy shores, or fished from shallow (nearshore) or deeper waters.

230

Liz Somerville, Janice Light and Michael J. Allen

Some species are more or less tolerant to fluctuations in the overall level of salinity than others. The lagoon cockle (Cerastoderma glaucum), for instance, is found on lower foreshore of estuaries and able to tolerate fluctuating salinity (McMillan 1968; Barrett & Yonge 1984). As noted above, distinguishing this species from the edible cockle can be problematic. Where archaeological sites are coastal, the present day marine environmental locally has the potential to answer questions regarding the provenance of the marine resources whose waste is excavated during site investigations. At Fistral Bay in Cornwall (Light 2005b) there was a strong correlation between the components of the marine mollusc shell assemblages in the middens from that Romano-British site and the macromollusc species composition on the adjacent shore in the modern day environment. This is good evidence that the archaeological shells were gathered locally and although this would be an intuitive interpretation, ‘ground-truthing’ (a site visit to test that hypothesis) added strong evidence for that interpretation. Similarly, at Ower Farm in Dorset correlations were possible between the midden contents and the local marine fauna, with suggestions for position of the exploited archaeological cockle beds based on the salinity gradients within the harbour (Winder 1991). In addition to that mechanism, and where archaeological sites are not in the vicinity of possible sources of marine resources, the use of epibionts can potentially be instructive. Whilst some epibionts have a contiguous geographical distribution around the coasts of the UK for example, some species have local distributions. A case in point is the burrowing marine polychaete worm, Polydora of which two species infest oysters in northern European waters. Polydora ciliata has a ubiquitous coastal distribution whereas P. hoplura is restricted to the south and south west coasts of England (Winder 1993).

Distribution of site activities and locations of on-site discard Records of the distribution of shells and mollusc species may show that shells as residue are dumped within contexts and varying feature types, as well as within pits and along ditches, and also comprise more random spreads across the site. Consideration of these data in association with period and phase allows patterns of discard distributions to be recognised. This in turn helps to inform the nature of site occupation and activities, and the role of marine mollusc exploitation for the site. As an example, if a distribution of oyster shells shows a marked difference between the flat (right) valves, which represent the primary waste component, and the cupped (left) valves, which form the ‘receptacle’ from which the oyster is consumed, this might differentiate between areas of preparation and consumption of oysters. However during an analysis of oysters from the Shapwick Project, Somerset Light (2007) was able to match several pairs of left and right oyster valves, taken from their sample bags, to the original individuals. The oysters, a total of ten individuals, from late medieval and 18th/19th century phases, were variously excavated from a former moat, a pond fill and a redeposited clay deposit. This may have been the primary discard site after consumption, or the oyster shells had remained associated during transfer to a secondary site.

13. Marine molluscs from archaeological contexts

231

It is important not to study the distribution of shells in isolation; it needs to be examined in the context of the discard patterns of other artefacts (pottery) and ecofacts (bone). This can help determine areas of differing uses across a site, and in particular can help with the interpretation of smaller ‘caches’ of shells particularly if they are proximal to the coast.

Summary Whilst methods of approach to, and analysis of, shell assemblages from archaeological sites have been illustrated citing case studies from the UK, other chapters in this volume provide examples from further afield. The chapter is not intended to act as a manual; it does not set out to outline a definitive method for carrying out an assessment and analysis of a marine shell assemblage. However as a general guide such an exercise might encompass all or some of the following processes: 1 2 3 4 5 6 7 8 9

Identify, quantify and size shells Record and comment on distribution of shells throughout contexts Assess and describe preservation condition Attempt to distinguish between comestibles and other species Attempt an interpretation of likely provenance(s) Assess any selection processes or harvesting strategies based on sizes of edible species Look for evidence of modification/hand-working (linked to 2) Consider reasons for occurrence on site Assess significance of assemblage

Conclusion • • • •

Human utilisation of marine molluscs has a very long history An outline of basic methodology for processing marine shell has been given Use of a reference collection is essential when working with marine shells in archaeology Interpretation of the marine shell found on archaeological sites requires consideration of both anthropogenic and natural reasons for shell accumulations • Anthropogenic shell may derive from collection for food, for bait or from collection for decoration or tools. In all cases important factors to consider are provenance, including habitats of exploitation • Palaeo-environmental information can be derived from archaeological marine shell

Analysts key identification texts Apart from key references for shell identification given in Box 1, the references collated in Box 3 are a specific selection of major works for marine shells, sporadic exotics and of the inventive use for shells which may occur on archaeological sites. This compilation of key and useful texts is aimed at the analyst of the archaeological material.

Liz Somerville, Janice Light and Michael J. Allen

232

Box 3 Analysts’ key identification references (compiled by Janice Light & Kevin Brown) Northeast Atlantic Audibert, C. & Delemmare, J-L. 2009. Guide de Coquillages de France, Atlantique et Manche. Paris: Belin De Bruyne, R. H. 2004. Veldgids Schelpen. Zeist: KNNV Publishing Chambers, P. 2009. British Seashells: a guide for conchologists and beachcombers. Barnsley: Remember When Christensen, J. M. & Dance, S. P. 1978. Seashells: bivalves of the British and Northern European Seas. Harmondsworth: Penguin Lellak, J. 1977. A Concise Guide in Colour: shells of Britain and Europe. London: Hamlyn Nordsieck, F. 1968. Die europäischen Meeres-Gehäuseschnecken (Prosobranchia) Vom Eismeer bis Kapverden und Mittelmeer. Stuttgart: Gustav Fischer Nordsieck, F. 1969. Die europäischen Meeresmuscheln (Bivalvia) Vom Eismeer bis Kapverden, Mittelmeer und Schwarzes Meer. Stuttgart: Gustav Fischer Poppe, G. T. & Goto, Y. 1991. European Seashells. Volume I. (Polyplacophora, Caudofoveata, Solenogastra, Gastropoda). Wiesbaden: Christa Hemmen Poppe, G. T. & Goto, Y. 1993. European Seashells. Volume II. (Scaphopoda, Bivalvia, Cephalopoda) Wiesbaden: Christa Hemmen Saunders, G. 2008. Shell Collecting Made Simple. Ely: Melrose Books

Mediterranean Cossignani, T. & Ardovini, R. 2011. Malacologia Mediterranea: Atlante delle conchiglie del Mediterraneo. Palermo: Naturama Doneddu, M. & Trainito, E. 2005. Conchiglie del Mediterraneo: Guida ai molluschi conchigliati. Il Castello: Cornaredo

West Africa Ardovini, R. & Cossignani, T. 2004. West African Seashells. Ancona: L’Informatore Piccin

Red Sea and Arabian Gulf Bosch, D. & Bosch, E. 1982. Seashells of Oman. London: Longman Bosch, D. T., Dance, S. P., Moolenbeek, R. G. & Oliver, P. G. 1995. Seashells of Eastern Arabia. London: Motivate Publishing

13. Marine molluscs from archaeological contexts

Rusmore-Villaume, M. L. 2008. Seashells of the Egyptian Red Sea: the illustrated handbook. Cairo: American University in Cairo Press Sharabati, D. 1984. Red Sea Shells. London: KPI.

Worldwide guides Abbott, R. T. & Dance, S. P. 1982. Compendium of Seashells. New York: E. P. Dutton De Bruyne, R. H. 2006. The Complete Encyclopedia of Shells: informative text with hundreds of photographs. Godalming: Rebo Dance, S. P. 1974. The Collectors’ Encyclopedia of Seashells. New Jersey: Chartwell Press. Dance, S. P. 1992. Eyewitness Handbook: Shells. London: Dorling Kindersley Eisenberg, J. M. 1981. A collector’s guide to Seashells of the World. New York: McGraw-Hill Oliver, A. P. H. 1975. Shells of the World. London: Hamlyn Robin, A. 2008. Encyclopedia of Marine Gastropods. Yarram, Victoria: IKAN Unterwasser-Archiv Robin, A. 2011. Encyclopedia of Marine Bivalves: including Scaphopods, Poly­ placophora and Cephalopods. Yarram, Victoria: IKAN Unterwasser-Archiv Wye, K. 1996. The Encyclopedia of Shells. London: Grange Books

Family Monographs There are monographs to many mollusc families of which the most useful is likely to be for Cowries: Lorenz, F. & Hubert, A. 2000. A Guide to Worldwide Cowries. Harxeim: ConchBooks

There are many sources in the scientific literature on the artefactual use of shells; below is a selection of popular books on the history and use of shells as artefacts and in personal and architectural adornment Jones, B. 1974. Follies and Grottoes. London: Constable Miles, C. 1963. Indian & Eskimo Artifacts of North America. New York: Bonanza Books Ritchie, C. I. A. 1974. Shell Carving: History and Techniques. Cranbury, New Jersey: A.S. Barnes Safer, J. F. & McLaughlin Gill, F. 1982. Spirals from the Sea: an anthropological look at shells. New York: Clarkson N. Potter Taborin, Y. 2004. Langage sans parole: La parure aux temps prehistoriques. Paris: La maison des roches, Editeur Thomas, I. 2007. The Shell: A World of Decoration & Ornament. London: Thames & Hudson

233

234

Liz Somerville, Janice Light and Michael J. Allen

Acknowledgements In researching, writing and compiling this chapter we would like to thank the advice and support of, in particular, Bas Payne and Terry O’Connor, but also Kevin Brown, Jessica Winder, members of the Conchological Society of Great Britain & Ireland (Council), and four other referees. In addition we would like to thank Nick and Julie for their assistance, wholehearted support, cups of tea and glasses of wine.

Endnote Caveat when measuring marine shells; some early career analysts of archaeological data have confused and transposed the terms length and width when referring to bivalve dimensions (see Fig 13.1). This has resulted in incompatibilities between most other assemblage datasets, between the associated analyses, and between the reporting of those analyses. Where such reports have been published this gives particular cause for concern. On the rare occasions such analyses have been published and have been used for synthesis and comparators over time and for national basis and inter regional comparison, the resultant comparisons have been fundamentally flawed. It is, therefore, important to attempt to ensure that their texts or archives are scrutinised to confirm measurement terminology used. This does not apply, however, where analysis have transposed the terms height and width. The latter term has frequently been used by archaeo-zoologists, and this does not affect statistical comparison.

References Ainis, A. F., Vellanoweth, R. L., Lapeña, Q. G. & Thornber, C. S. 2014. Using non-dietary gastropods in coastal shell middens to infer kelp and seagrass harvesting and paleoenvironmental conditions, Journal of Archaeological Science 49, 343–460 Andrus, C. F. T. 2011. Shell midden sclerochronology, Quaternary Science Reviews 30(21–22), 2892–2905 Appeltans, W., Ahyong, S. T., Anderson, G., Angel, M. V., Artois, T., Bailly, N. et al. 2012. The magnitude of global marine species diversity. Current Biology 22(23), 2189–2202 Bailey, G. N. 1978. Shell middens as indicators of postglacial economies: a territorial perspective. In Mellars, P. (ed.), The Early Postglacial Settlement of Northern Europe: an ecological perspective, 37–63. London: Duckworth Barrett, J. H & Yonge, C. M. 1984. Collins Pocket Guide to the Sea Shore. London: Collins Bell, M. G. 1981. Seaweed as a prehistoric resource. In Brothwell, D. &. Dimbleby, G. (eds), Environmental Aspects of Coasts and Islands, 117–126. Oxford: British Archaeological Report S94 Brothwell, D. R. 1972. Digging up Bones. London: British Museum (Natural History) Brothwell, D. & Brothwell, P. 1998. Food in Antiquity: a survey of the diet of early peoples. Baltimore and London: Johns Hopkins University Press Campbell, G. E. 2014. Size prediction in archaeomalacology: the common mussel, Mytilus edulis L., as an example, Archaeological and Anthropological Sciences 6(3), 305–318 Campbell, G. 2017a. ‘What do I do with all these shells?’ Basic guidance for the recovery, processing and retention of archaeological marine shells, Quaternary International 427(part A), 13–30 Campbell, G. 2017b. The collection, processing, and curation of archaeological marine shells. In Allen, M. J. (ed.), Molluscs in Archaeology, 273–288. Oxford: Oxbow Books Claassen, C. 1998. Shells. Cambridge: Cambridge University Press

13. Marine molluscs from archaeological contexts

235

Davidson, A. 2003 North Atlantic Seafood. Totnes: Prospect Books Demarchi, B., Williams, M. G. M., Milner, N., Russell, N., Bailey, G. N. & Penkman, K. 2011. Amino acid racemisation dating of marine shells: a mound of possibilities, Quaternary International 239, 114–124 Enegren, H. L. & Meo, F. (ed.) 2017. Treasures from the Sea; sea silk and shellfish purple dye in antiquity. Ancient Textiles Series 30. Oxford: Oxbow Books Evans, J. G. 1983 Mollusca and other invertebrates from Ardnave, Islay, 350–358. In Ritchie, G. & Welfare, H., Excavations at Ardnave, Islay. Proceedings of the Society of Antiquaries of Scotland 113, 302–366 Fenger, T., Surge, D., Schöne, B. & Milner, N. 2007. Sclerochronology and geochemical variation in limpet shells (Patella vulgata): a new archive to reconstruct coastal sea surface temperature, Geochemistry, Geophysics, Geosystems 8(7) doi 10.1029/2006GC001488 Gillikin, D. P., Dehairs, F., Baeyens, W., Navez, J., Lorrain, A. & André, L. 2005. Inter- and intra-annual variations of Pb/Ca ratios in clam shells (Mercenaria mercenaria): a record of anthropogenic lead pollution?, Marine Pollution Bulletin 50, 1530–1540 Giovas, C. M. 2009. The shell game: analytic problems in archaeological mollusc quantification, Journal of Archaeological Science 36, 1557–1564 Hardy, K. 2017. Shell middens. In Allen, M. J . (ed.), Molluscs in Archaeology, 259–272. Oxford: Oxbow Books Harris, M., Weisler, M. & Faulkner, P. 2015. A refined protocol for calculating NMI in archaeological molluscan shell assemblages: a Marshall Islands case study, Journal of Archaeological Science 56, 168–179 Jackson, J. W. 1917. Shells as Evidence of the Migrations of Early Culture. London: Longmans, Green & Co. Jerardino, A. & Marean, C. W. 2010. Shellfish gathering, marine paleoecology and modern human behavior: perspectives from cave PP13B, Pinnacle Point, South Africa, Journal of Human Evolution 59 (3–4) 412–424 Labonne, M., ben Othman, D. & Luck, J.-M. 1998. Recent and past anthropogenic impact on a Mediterranean lagoon: lead isotope constraints from mussel shells, Applied Geochemistry 13(7), 885–892 Laurie, E. M. 2008. An Investigation of the Common Cockle (Cerastoderma edule (L.)): collection practices at the kitchen midden Sites of Norsminde and Krabbesholm, Denmark. Oxford: British Archaeological Report S1834 Law, M. 2013. Past agricultural practices in the Western Isles revealed by subfossil mollusc shells. Poster presented at the UK Archaeological Sciences Conference, Cardiff. http://www.bathspa. academia.edu/MattLaw/Talks [accessed 13.4.15] Law, M. & Thew, N. 2017. Molluscs from dune-machair systems in the Western Isles: archaeological site formation processes and environmental change. In Allen, M. J. (ed.), Molluscs in Archaeology, 82–99. Oxford: Oxbow Books Light, J. M. 2003a.The Oyster shells and other molluscs. In Hardy, A., Dodd, A. & Keevill, G. D., Elfrics Abbey: excavations at Eynsham Abbey, Oxfordshire, 1989–92, 1427–1432. Thames Valley Landscapes Volume 16. Oxford: Oxford Archaeology Light, J. M. 2003b. Dog cockle shells as occasional finds in Romano-British shell middens from Newquay, North Cornwall, UK, Environmental Archaeology 8(1), 51–59 Light, J. 2005a. Marine shell; crab. In Sharples, N. (ed.), Norse Farmstead in the Outer Hebrides: Excavations at Mound 3, Bornais, South Uist, 42 & 162–163. Oxford: Oxbow Books Light, J. 2005b. Marine mussel shells – wear is the evidence. In Bar-Yosef Mayer, D. E. (ed.) Archaeomalacology: molluscs in former environments of human behaviour, 56–62. Oxford: Oxbow Books

236

Liz Somerville, Janice Light and Michael J. Allen

Light, J. 2007. Marine molluscs. In Gerrard, C. & Aston, M. (eds.) The Shapwick Project, Somerset. A Rural Landscape Explored, 927–930. Society for Medieval Archaeology Monograph 25. Abingdon: Taylor & Francis Light, J. 2017. Marine shell artefacts: cautionary tales of natural wear and tear as compared to resourceful anthropogenic modification processes. In Allen, M. J. (ed.), Molluscs in Archaeology, 342–361. Oxford: Oxbow Books Light, J. & Walker, T. 2017. How strong is the evidence for purple dye extraction from the muricid gastropod Nucella lapillus (L. 1758), from archaeological sites in Britain and Ireland? In Allen, M. J. (ed.), Molluscs in Archaeology, 326–341. Oxford: Oxbow Books Lovell, M. S. 1884. The Edible Mollusca of Great Britain and Ireland with Recipes for Cooking Them. London: L. Reeve & Co. Luff, R. & Rowley-Conwy, P. 1994. The (dis)integration of environmental archaeology. In Luff, R. & Rowley-Conwy, P. (eds) Whither Environmental Archaeology?, 1–3. Oxbow Monograph 38. Oxford: Oxbow Books Mannino, M. A. & Thomas, K. D. 2001. Intensive Mesolithic exploitation of coastal resources? Evidence from a shell deposit on the Isle of Portland (Southern England) for the impact of human foraging on populations of intertidal rocky shore molluscs, Journal of Archaeological Science 28, 1101–1114 Mannino, M. A. & Thomas, K. D. 2002. Depletion of a resource? The impact of prehistoric human foraging on intertidal mollusc communities and its significance for human settlement, mobility and dispersal, World Archaeology 33, 452–474 Marchis, E. 1930. Les Mollusques marins comestibles. Paris: Société d’Éditions Géographiques, Maritimes et Coloniales McMillan, N. F. 1968. British Shells. London: Frederick Warne. Milner, N., Barrett, J. & Welsh, J. 2007. Marine resource intensification in Viking Age Europe: the molluscan evidence from Quoygrew, Orkney, Journal of Archaeological Science 34, 1461–1472 Murphy P. 1992. Land and freshwater molluscs; marine invertebrates; samples from the medieval lime-kiln. In Crummy, P., Excavations at Culver Street, the Gilberd School and Other Sites in Colchester 1971–85, 276–278 & 288–289. Colchester Archaeological Report 6. Colchester: Colchester Archaeological Trust Murphy, P. 2001. Various reports. In Lane, T. & Morris, E. (eds), A Millennium of Salt-making: prehistoric and Romano-British salt production in the Fenland. 75–89, 147–157, 320–323, 377–385. Lincolnshire Archaeology and Heritage Report Series 4. Heckington: Heritage Trust of Lincolnshire Reese, D. S. 1991. The trade of Indo-Pacific shells into the Mediterranean Basin and Europe, Oxford Journal of Archaeology 10(2), 159–196 Ridout-Sharpe, J. 2017. Shell ornaments, icons and other artefacts from the eastern Mediterranean and Levant. In Allen, M. J. (ed.), Molluscs in Archaeology, 296–307. Oxford: Oxbow Books Rosenberg, G. 2014. A new critical estimate of named species-level diversity of the recent Mollusca, American Malacological Bulletin 32(2), 308–322 Russell, N., Cook, G. T., Ascough, P., Barrett, J. H. & Dugmore, A. 2011. Species specific marine radiocarbon reservoir effect: a comparison of ΔR values between Patella vulgata (limpet) shell carbonate and Gadus morhua (Atlantic cod) bone collagen, Journal of Archaeological Science 38(5), 1008–1015 Safer, J. F. & Gill, F. M. 1982. Spirals from the Sea: an anthropological look at shells. New York: Clarkson N. Potter Saul, M. 1974. Shells, an Illustrated Guide to a Timeless and Fascinating World. London: Hamlyn Simmonds, P. L. 1879. The Commercial Products of the Sea. London: Griffith & Farran

13. Marine molluscs from archaeological contexts

237

Somerville, E. M. 2008. The marine molluscs. In Barber, L. & Priestley-Bell, G., Medieval Adaptation, Settlement and Economy of a Coastal Wetland, the Evidence from around Lydd, Romney Marsh, Kent, 228–238. Oxford: Oxbow Books Somerville, E. M. 2010. Marine molluscs. In Thomas, G., The Later Anglo-Saxon Settlement at Bishopstone: a downland manor in the making, 176–187. Council for British Archaeology Research Report 163. York: Council for British Archaeology. Stearns, R. E. C. 1999 [1887]. Wampum and Dentalium – a study of Native American shell money. Hummelstown: Tucquan Steele, T. E. & Klein, R. G. 2008. Intertidal shellfish use during the Middle and Later Stone Age of South Africa, Archaeofauna 17, 63–76 Szabó, K, 2017. Molluscan shells as raw materials for artefact production. In Allen, M. J. (ed.), Molluscs in Archaeology, 308–325. Oxford: Oxbow Books Tebble, N. 1966. British Bivalve Seashells. London: British Museum (Natural History) Thomas, K. D. 2015a. Molluscs emergent, part I: themes and trends in the scientific investigation of mollusc shells as resources for archaeological research, Journal of Archaeological Science 56, 133–140 Thomas, K. D. 2015b. Molluscs emergent, part II: themes and trends in the scientific investigation of molluscs and their shells as past human resources, Journal of Archaeological Science 56, 159–167 Vanhaeren, M., d’Errico, F., Stringer, C., James, S. L., Todd, J. A. & Mienis, H. K. 2006. Middle Paleolithic shell beads in Israel and Algeria, Science 312, 1785–1788 Walker, T. 2017. Molluscs and the palaeo-environment of coastal blown sand and dunes. In Allen, M. J. (ed.), Molluscs in Archaeology, 65–81. Oxford: Oxbow Books Winder, J. 1980. The marine Mollusca. In Holdsworth, P. (ed.), Excavations at Melbourne Street Southampton 1971–76, 121–127. Council for British Archaeology Research Report 33. London: Council for British Archaeology Winder, J. M. 1991. Marine Mollusca. In Cox, P. W. & Hearne C. M. Redeemed from the Heath; the archaeology of the Wytch Farm Oilfield (1987–90), 212–216. Dorchester: Dorset Natural History & Archaeological Society Monograph 9 Winder, J. M. 1993. Oyster and other marine mollusc shells. In Woodward, P. J., Davies, S. M. and Graham, A. H., Excavations at Greyhound Yard, Dorchester 1981–4, 347–348. Dorchester: Dorset Natural History & Archaeological Society Monograph 12 Winder, J. 2017. Oysters in archaeology. In Allen, M. J. (ed.), Molluscs in Archaeology, 238–258. Oxford: Oxbow Books Yeats, J. 1878. The Natural History of the Raw Materials of Commerce. New York: Scribner, Welford & Armstrong Zilhão, J., Angelucci, D. E., Badal-Garcia, E. d’Errico, F., Danile, F., Davet, L., Douka, K., Higham, T. F. G. Martínez-Sánchez, M. J., Montes-Bernárdez, R., Murcia-Mascarós, S., Pérez-Sirvent, C., Roldán-García, C., Vanhaeren, M., Villaverde, V., Wood, R. & Zapata, J. 2010. Symbolic use of marine shells and mineral pigments by Iberian Neanderthals, Proceedings of the National Academy of Sciences 107(3), 1023–1028

14. Oysters in archaeology Jessica Winder

This chapter provides an overview of the way in which simple methods for studying the macroscopic features of oyster shells, excavated from relatively recent historical deposits in the UK, were developed during the late 20th century. It also shows how the resulting data have been used to make spatial and temporal distinctions between samples and enabled discussion about oyster trade and collection practices. The chapter tentatively suggests the application of advanced techniques for identification of chemical and protein composition in archaeological oyster shells in order to improve our understanding of the exploitation of oysters in the past. Successful employment of these newer methods could perhaps facilitate interdisciplinary research into wider issues such as oyster population identification, effects of global climate change, and the impact of industrialisation on coastal water quality, by providing baseline data for the investigations.

Background to research Shells of the European flat oyster or British Native oyster Ostrea edulis L. (Figs 14.1 & 14.2) record and reflect to an extraordinary degree the chemical, physical, and biological environment in which they grew. Few other edible bivalve molluscs equal this species for variability in shell shape and structure, and for the range of evidence related to epibiont organisms that use the shell as a habitat. The most useful aspect for study of variation in the archaeologically-derived European flat oyster material, is the extent to which shell size, shape and other features are modified not only by factors in the growth environment but also by the effects of human activities associated with its collection, its use as food, and its disposal. Readily observable features in oyster shells from archaeological excavations can provide important evidence for their source location and manner of exploitation. However, this potential is accompanied by methodological and epistemological challenges for the investigator. The methods were first developed when a surge of urban redevelopment in 1970s Britain, with its accompanying archaeological surveys and excavations, unearthed large quantities of historical oyster and other marine mollusc shell food remains.

14. Oysters in archaeology

239

Archaeologists asked how useful this material might be for site interpretation. Would it be possible to say where the oysters had come from? Were they from natural wild populations, farmed or cultivated? What was their significance in the diet and economy of the local and regional economy? Importantly, conscious of the enormous numbers of shells to be processed and funding limitations, could these questions be answered using simple, cost-effective, and easy-to-learn methods? In the 1970s and 80s, there was already a great deal of interest in archaeological shell deposits, including oyster shells, but research had mainly focused on large early period middens in Britain such as those on Oronsay (Mellars 1987). Work was also being undertaken on shells from Palaeolithic, Mesolithic, Neolithic, post-glacial and Quaternary sites in Australia, Japan, the Americas, and continental Europe, on topics such as midden distributions along prehistoric coastlines, shellfish gathering patterns

Figure 14.1 The two valves of the oyster (Ostrea edulis L.) showing major features and shell orientation with left valve top and right valve below (illustration: Abby George)

240

Jessica Winder

and subsistence strategies, and seasonality of collection using shell growth line and oxygen isotope analysis. Significant contributors to the research included Bailey (1975; 1978; 1983; Bailey & Parkington 1988); Deith (1983a; 1983b; 1985a; 1985b; 1988); Koike (1979; 1981); Meehan (1982); Mellars (1978; 1987); Shackleton (1983; 1988); and TroelsSmith (1967); and at a later stage Milner with other authors (eg, 2001; 2002; 2009; 2013; Milner & Barrett 2012; Milner & Woodman 2002; Milner et al. 2007; Demarchi et al. 2011; 2013; Gutierrez-Zugasti et al. 2011; Surge & Milner 2003). However, the species of oyster under consideration by these authors were often different from the European flat oyster that was being excavated from English urban sites, and the questions being asked of the material and the strategies for investigation such as those outlined in Cherry et al. (1978) were not always applicable or pertinent to the newly-recovered historic pit and stratum oyster shells. Claassen (1998) gives an excellent comprehensive account of the kind of questions that were and are still being asked of archaeological shell material elsewhere. These include enquiries into the taphonomy of the shells and shell assemblages, sampling methods and quantification, palaeo-environmental reconstruction, season of death, and shells as artefacts. The exception to this general research trend was the work of Kent (1988) in Making Dead Oysters Talk: techniques for analysing oysters from archaeological sites, a research project in Maryland, USA, which was being undertaken at much the same time as work along similar lines had started in Britain with the examination of urban deposits of oyster and other marine shells from Saxon sites in Southampton (Winder 1980). In Britain, the oyster shells and other marine molluscs were generally being excavated from smaller deposits, often in urban and rural, inland as well as coastal locations, and dating from only the last 2000 years. The questions being asked of the material were directed specifically at within-site interpretation and an understanding of aspects of diet, trade and economy on the local and regional scale. Methods were devised to account for these differences in aims and the limited available resources. Preliminary investigations of the literature indicated that oyster fishermen and oyster connoisseurs could reputedly tell where an oyster had come from merely by its appearance and taste, showing that characters existed by which those from different locations could be distinguished – at least in the fresh undamaged oysters. Lucilius the Roman poet said ‘When I but see the oyster’s shell, I look and recognise the river, marsh or mud where it was first raised’. What might those characters be? Yonge (1960) gives an invaluable account of the structure, biology and natural history of the oyster (Ostrea edulis Linnaeus), including the types of marine organism that infest and encrust the shells, as a starting point for understanding what could be the most useful shell characteristics to seek in the archaeological shells. One question for the archaeomalacologist was whether any useful distinguishing characters still remained in archaeological shells since many features present in fresh specimens would not survive in long-buried samples. The fleshy parts of the mollusc itself and also soft parts of epibiont organisms such as marine worms, barnacles, sea squirts and algae would readily decompose. Breakage, wear and weathering – before, during, and after burial – may have damaged the shell, smoothing the surface sculpturing, obscuring growth lines, removing foliation, along with the destruction of

14. Oysters in archaeology

241

adherent sessile barnacle plates, and chalky or sand-grain tubes of polychaete worms. Shells fade, delaminate, and disintegrate with time. Diligence is often needed to observe and record the features accurately when the condition of the excavated shell is poor. It is frequently possible to identify former epibiont associations on oyster shells by examining the damage they render or the remnant encrusting material. For example, the holes left by mud-tube bearing marine worms, predatory gastropods and boring sponges can be distinguished from each other (eg, Boekschoten 1966; Carriker & Yochelson 1968); whilst attached organisms like calcareous tube worms and acorn barnacles leave recognisable attachment scars or basal plates. Notably, whilst many excavated shells are worn and relatively featureless, some can remain surprisingly fresh in their appearance and even retain pigmentation or fragments of ligament and periostracum, to the extent that they could easily be mistaken for freshly-dead shells. In some special circumstances oyster shells survive well with minimal damage and preservation of the proteinaceous structures of the ligament and the conchyolin framework that supports the largely inorganic shell. This usually happens in waterlogged deposits. Examples include shells from the extensive late Saxon and early Conquest-period midden found by the old Town Cellars on the edge of Poole Harbour in Dorset (Horsey & Winder 1991; Winder 1992a) and medieval shells recovered from a well in North West Cambridge excavations. Not all samples of examined oyster shell are suitable for analysis. Whereas shell samples from a site can all be recorded in a basic way by making species identifications and counts, facilitation of viable statistical comparisons between samples require that detailed records be taken only from shells in contexts that are securely dated or phased and (as far as can be ascertained unbiased), and with samples comprising larger numbers of near-intact shells. Correct identification, quantification, and understanding of the significance of the varying shell features, allied with knowledge of the limitations of extrapolation and interpretation from the archaeological data, allow the questions posed to be addressed.

Methods The methods devised for recording these features are described in Winder (1992b) and later with numerous accompanying illustrations in Winder (2011) – both supplying detailed instructions for the initial processing and recording methods for macroscopic characteristics in archaeological oyster shells. Standardisation of recording methods is vital, especially when samples recorded by different individuals are being compared. The methods involved quantified recording of objective characters such as numbers, ratios of left to right shell valves, (Figs 14.1 and 14.2), measurements (Fig. 14.3), and details of epibiont infestation in the oyster shells (eg, damage caused by forms of worms and sponges, see Figures 14.4 and 14.5 respectively). Quantified recording of subjective characters noted relative shell thickness, presence of chambering, shape, colour, degree of wear, clumping, attached spat, degree of distortion and man-made marks (eg, Fig. 14.6). A combination of the measurable and objective, together with some subjective

242

Jessica Winder

Figure 14.2 Typical oyster right (upper) and left (lower) valves, showing inner and outer views

and descriptive characters, can be used in analyses. Records of up to 25 features are suggested. For accurate recording of these kinds of features great care needs to be taken with handling and washing the shells, as this could potentially physically damage shells and destroy evidence.

14. Oysters in archaeology

243

Figure 14.3. Photograph of right oyster valve (exterior view) showing measurable dimensions. Note that height is the measurement from the beak or umbo to the ventral margin – but this measurement has commonly been called the width in many archaeological analyses, see Endnote

Outcomes The information gathered from the oyster shells is expressed as a mean frequency of occurrence of each characteristic in the whole sample. These frequencies give each sample a unique description. Statistical analysis of both objective and subjective sample characteristics are used to make spatial intra-site and inter-site comparisons of samples from a local context and feature level, to a wider geographical level; and also to make temporal comparisons both within a site on a phase-by-phase level and across broad historical time or occupation periods. Size comparisons within a site are made using parametric two-sample tests and also non-parametric Kolmogorov-Smirnov or Mann-Whitney U-tests. Infestation frequencies can be simply collated and compared visually with both archaeological material and modern marine invertebrate distributions of nearby coastal localities. Some examples

244

Jessica Winder

of first attempts to collate and analyse oyster shell data are published in Winder (1980; 1985; 1991; 1992a; 1992b; 1993a; 1993b; 1994a; 1994b; 1997a; 1997b; 2000a; 2000b; Winder & Reidy 1996), and Wyles and Winder (2000). Winder (1992b) brought together much of this reported material (from 1980 to 1992) and used the substantial database of over 30,000 oyster shells from 60 archaeological sites (together with other marine molluscs) to demonstrate distinctions in size and infestation between oysters from the south coast of England in Poole Bay and Harbour, the Solent and Southampton Water, and the Thames estuary and East Anglian coastline. Comparisons of size for broadly defined historical periods revealed interesting variations in mean sizes between the Roman, Saxon, medieval, post medieval and modern oyster shells. This appeared to indicate statistically significant temporal differences in the average size of oyster shells. Roman shells were largest but size decreased progressively through successive periods until a recovery to almost Roman dimensions in the modern period. In addition to clear ideas about movements and transport of oysters between different localities in the past, and of site specific information about oyster usage, in brief terms, the following picture emerged about oyster exploitation in Britain. No oyster shells were, at that time, recovered from Iron Age sites. Specimens found at Owslebury in Hampshire are now believed to be incorrectly assigned to that period. Roman sites throughout the UK were renowned for the massive quantities of oysters, but, contrary to assertions in the literature, no physical or documentary evidence was found at that time to indicate that the Romans introduced oyster cultivation as such to Britain. The cultivation techniques used in Italy between AD 0–400, would have been impractical and unnecessary in Britain. Although prehistoric oyster middens have been found in the Scottish isles, oysters appear to have been a largely unexploited resource in the period immediately prior to the Roman invasion of Britain. The claim that oysters were transported around Britain alive in tanks of water during the Roman occupation (Frere 1967) seems also to be highly unlikely and immensely impractical. Since oysters will remain fresh for up to 10 days if kept cool and closely-packed, oysters could have been simply tightly-packed into baskets, barrels, or even British-made pots for transport. Black burnished-ware pottery manufactured on the shore of Poole Harbour, Dorset, was sent as far afield as Hadrian’s Wall (Cox & Hearne 1991) and it would have been easy to fill them with fresh oysters from the adjacent beds before dispatch. The large average oyster size for the period may reflect an abundance of mature specimens, a preference for eating larger oyster meats than we select today, as well as a rapid growth rate. Saxon sites also produced lots of oysters but these were mostly near the coast or within easy reach of the coast by river. Deterioration of the roads at the end of the Roman period and poorer organisation meant that oysters could not be sent far. Average size was found to be slightly but significantly smaller than those from Roman sites. To date there is still no evidence for farming or cultivation of oysters in that period. By the medieval period, oysters were far more widely distributed across the country. They were also noticeably smaller. Their size tended not to be a selection of immature specimens but rather of prevalence of slower growth rate, possibly attributable to temperature changes but also maybe directly resulting from oyster relaying and storage activities. Documentary records exist for the ownership of oyster beds and oyster fishing

14. Oysters in archaeology

245

rights at this time together with evidence for re-laying stock for fattening. Intertidally re-laid oysters cease to grow shell whilst periodically out of water and therefore achieve smaller sizes. Simultaneously they learn to keep the valves tight shut when exposed to the air, preventing desiccation, which means they stay fresh when being transported over greater distances. The increasing numbers of oysters found on coastal sites reflects their easy availability and indicates that they were a staple of the diet there. The smaller numbers of oysters found at inland sites suggests that the cost of transporting oysters made them an occasional and luxury item away from the coast. Not many oyster specimens of post-medieval date were available for the study, so conclusions are few. The examined shells were smaller than in earlier periods. The Modern period was taken as including the 19th century onwards which saw the advent of railways and with them an efficient countrywide distribution of increasingly cheap oysters. It was a boom time for oystermen fishing the natural and re-laid beds, with shoreline holding pits to store the catches and ensure constant availability for marketing in season. Oyster stocks eventually became depleted by overfishing. All attempts to increase oyster stocks by cultivation and introduction of foreign species failed. Rare physical evidence of this type of cultivation experimentation at the Sinah Circle (Fig. 14.7), has been recorded in Langstone Harbour (Adams et al. 2000). The final blow to the incredibly successful oyster industry of the 19th and early 20th centuries came with massive extinctions of beds in the 1920s – thought to result from extreme cold weather and disease. A few natural beds of oysters survived. Oysters became a luxury item on the menu again. A second catastrophe in the form of Bonamia disease devastated remaining stocks in the 1970s. Modern technology came to the rescue of the British oyster industry by breeding oyster spat of both Ostrea edulis and Crassostrea gigas in the laboratory so that beds could be restocked. Oyster farming today with its net bags of lab-reared oysters and floating platforms would not be recognised by our predecessors. Their methods were undoubtedly simpler but harder and we still have much to find out about them.

Models of oyster exploitation The strong database of detailed information about oyster shells enabled the formulation of models in which data recorded from oyster shells could be used to interpret the mode and level of exploitation of this marine resource (Winder 1992b, 281–304). The models identify which types of evidence, from the shells themselves and the context in which they are found, might indicate different types of oyster bed location, and suggest the degree of effort required to take advantage of this natural resource. The models represent a system view considering direct evidence from the oyster shells themselves, all associated data recorded for the natural environment where they were possibly reared, and for the man-made environment in which they were collected, used and discarded (such as other associated marine mollusc species, contextual information, coastal ecosystem data, and historical records) to characterise the whole system from which they were derived and of which they were an integral part. The proposed five

246

Jessica Winder

Figure 14.4. Oyster with: top Polydora ciliata (worm) infestation, and bottom Cliona celata (sponge) infestation

Figure 14.5. Oyster with: top burrow of Polydora hoplura and, bottom: with blisters caused by Polydora hoplura

theoretical models illustrate exemplar points in what is really a transitional series from a representation of a simple collection strategy of sporadic hand-collection of oysters from natural intertidal beds; through gradually increasing intensity of effort to a fullscale cultivation, harvesting and marketing scenario. Each element of data recorded from the oyster shells and the site can potentially contribute to our understanding of the particular type of environment in which the oysters lived, and the level of activity involved in their collection or harvesting. For example, infestation evidence could be used to suggest the locality of the bed, whether the bed was intertidal littoral or shallow sub-littoral, harder or softer substrate, and also the degree of salinity. Size distributions may reflect growth rate, recruitment variability, selection preferences, and survival rates. Certain combinations of shell sample characteristics can be used in an attempt to distinguish between fished and farmed oysters. A natural population might be suggested by a wide range of size and age, irregularity in shell shape, and the presence of attached oysters including spat. Shells from re-laid or cultivated populations might show a narrowing of size and age range, greater regularity in shape, an absence of attached oysters (especially spat), and

14. Oysters in archaeology

247

possibly cultch (deliberately deposited spat-collection material), or an imprint of it, at the heel of the shell. Fluctuating salinity regimes typical of inshore shallow waters where oysters are re-laid can be indicated by chambering of the shell, and by chalky deposits. At the lowest organisational level of exploitation (Model 1; Fig. 14.8) there would be sporadic collection by hand of oysters from natural populations in the intertidal zone on the sea shore, estuaries or creeks. In this scenario the exploitation level would be low and indicated by small quantities of shell, possibly in isolated pockets or separate layers suggesting short-term periodicity of collection. There might be a wide size and age range from random collection, and a high proportion of irregularly shaped and clumped groups of shells of different ages (indicative of a natural population because distorted shapes result from a competition for growing Figure 14.6. Oyster with man-made perforations and notches space when many spat oysters settle on the same object). An example of a Model 1 situation is provided by the 12th–13th century shell midden at Ower Farm on the southern shore of Poole Harbour (Winder 1991) where all the evidence pointed to collection from a small, natural, overcrowded population on a rough substrate including accumulations of empty cockle shells. Model 2 postulates the introduction of special equipment which enables fishing for oysters by dredging inshore shallow sub-littoral natural beds of oysters. This requires greater expenditure of effort and a more organised approach to collection which is probably conducted on a more regular basis because the equipment makes the resource accessible at all times. The average sizes of the shells might be larger than those recovered from the intertidal zone because growth would not have been interrupted by periodic exposure to air. The size range might possibly be narrower if a dredge net had been used. As in Model 1 a high proportion of shells with irregular shape and groups of oysters clumped together might be expected, as typical of an unmanaged bed. An example of a possible Model 2 situation was seen in samples from Greyhound Yard in Dorchester (Winder 1992b; 1993a) where a study of oyster shells from medieval and Roman contexts indicated that they had originated from 30 miles (48 km) away in the shallower waters of Poole Harbour rather than the deeper water of Poole Bay. Model 3 is an extension of Model 2 involving the dredging of deeper off-shore sub-littoral oyster beds. Exploiting the deeper waters at this distance from shore

248

Jessica Winder

Figure 14.7. Sinah Circle structure from Langstone Harbour (from Allen & Gardiner 2000, fig. 36; illustration: Abby George)

would require more effort than expended in Models 1 and 2, better equipment, and more advanced skills. Many of the characteristics expected in Model 2 would also be found in Model 3. Some differences would be likely as in the lower intensity and type of epibiont damage from relatively nutrient-poor deeper water, a different range of associated molluscs, and possibly the shape of the shells from growing on firmer offshore substrates (Winder 1992a; 1992b). A modern example of the Model 3 type of exploitation is seen in the fishing for wild oysters that takes place in Poole Bay for relaying or for direct sale. Archaeological samples that parallel the size characters and infestation patterns of the natural oysters from Poole Bay, include all of the medieval oysters from Paradise Street, Poole (Winder 1992a; 1992b). The first three models describe a trend towards the more systematic and expert recovery of oysters growing naturally in the wild. Model 4 illustrates a further intensification of procedures which are designed to increase stocks and availability while improving quality of oyster meat. This model postulates the introduction of deliberate management of oysters stocks, with foresight, planning, and development

14. Oysters in archaeology

249

Figure 14.8. Model 1: lowest levels of exploitation of oysters (sporadic hand collection from natural beds) showing the relationship between combinations of recorded shell sample characteristics, shell collection preferences, and features of the source bed

of cropping strategies, implying an understanding of the oyster life-cycle. The scenario proposes dredging and culling immature oysters from natural beds for re-laying on sheltered inshore sub-littoral or intertidal oyster beds for fattening – which is the first stage of oyster farming. Typical features of shell samples would include a restricted size and age range from regularised cropping, grading for market, or stock conservation regulations. Clumping and irregular shape would be less common because of separating individuals and removal of organisms by culling, and infestation damage might increase in the relatively nutrient-rich water. A modern example of Model 4 was found in the re-laid oysters from Wych (sic) Channel and South Deep in Poole Harbour of 1987 and the sample from the 1971 Colchester Oyster Feast. Model 5 involves full-scale cultivation and marketing and represents the maximum amount of effort for maximum gain in terms of food for the local market and surpluses for cash or goods trade. It describes a situation where the oyster populations are managed to the fullest extent from spawning to table. It includes dredging activities and relaying as in other models but goes further, with actual cultivation in which spawning is monitored, spat is collected and nurtured in special conditions, grown on, harvested, stored live, graded and marketed. Model 5 type of activity would be comparable to 19th century methods rather than current ones. Since the mid-20th century oyster cultivation practices in Britain have changed a great deal to take into account vulnerability of the natural stocks to disease and the need for greater control of the beds and productivity resulting in increased use of laboratory cultured spat

250

Jessica Winder

oysters of non-British species and new techniques of relaying for growing on. The first four models of oyster exploitation could well describe activities that had developed through time in the region where the Horndon-on-the-Hill (Essex) oysters originated. The oysters from the four features under consideration at the Horndon site differed from each other and it was considered possible that these differences could be interpreted as the result of varying fishing practices.

Oyster shape Differences in oyster shell shape in relation to habitat can contribute to the identification of the source oyster beds for archaeological material. Shape was first investigated in British material by Winder (1992a) who correlated shape in archaeological shells of late Saxon, medieval, and post medieval date from Poole in Dorset, England, with modern shells from known locations – firm cleaner substrate in deep water within Poole Bay and softer muddy substrate in shallower water of Poole Harbour – thus indicating the oyster source and suggesting transfer of stock for historical shell deposits. This work has since been advanced by Campbell (2010) who developed a more effective means of calculating shape in oysters from Roman Winchester and in living populations of oyster from across the Solent (Hampshire). Campbell found shell shape varied between harbours, near-shore and deeper water, probably in response to differing bed currents. Archaeological shells changed abruptly during growth from a range of shapes to a single shape arguing for oyster management in late Roman England.

Principal Component Analysis Experimentally from 1998 onwards a meta-analysis by Principal Component Analysis (PCA) was used to compare the sum total of all recorded characteristics in oyster shell samples from sites at Elms Farm and Great Wakering in Essex, and the Royal Opera House in London (Winder & Gerber-Parfitt 2003), for example, compared with all other oyster shell material gathered to that date. This way of presenting the data gave a unique virtual ‘fingerprint’ identity to each sample and was an attempt to address concerns that comparisons of not only size but also infestation and other characters in oysters needed a statistical base. However, it was found that PCA based on only the epibiont infestation and encrustation characteristics (which closely relate to the natural conditions in which the oyster was growing – such as the depth of water, the substrate and the geographical location), proved most useful in differentiating oysters from different regions. PCA of infestation in Roman oyster samples demonstrated segregation mainly into one group with similar characteristics from east coast sites in Essex and Suffolk, and another from south coast sites in Dorset, Hampshire and Wiltshire. Oysters from The Shires excavation in Leicester (Monckton 1999) and from Pudding Lane in London (Winder 1985) were included in the grouping of samples known to have originated in

14. Oysters in archaeology

251

East Anglia thus indicating that oysters at the two inland sites were obtained from that part of the coast. The same marked differentiation can be seen in PCAs for other periods as well. The organisms that seem primarily to account for this regional differentiation of oysters from the south coast compared with the east coast of England are polychaete worms of the Polydora genus. These worms leave characteristic burrows in the shells. Polydora ciliata (Johnston) seems to be ubiquitous (Fig. 14.4) while the larger species Polydora hoplura Claparède (Fig. 14.5) appears to be restricted to southern waters. PCA seems a promising approach for pinpointing the source of oyster samples.

Caution about interpretations It is important to reflect, albeit briefly, upon the nature of the data being used, and the particular constraints that can arise when using archaeological material rather than recent samples over which there would be a greater level of control in selection. There are many challenges to working with archaeological oyster shells. It is important to know the exact nature of the sample being selected for study to avoid biases – to know how representative are the examined samples of the potentially available pool of archaeological oyster material. For example, in the case of the extensive Saxo-Norman oyster heaps on Poole waterfront (Horsey & Winder 1991; Winder 1992a), and the nearby 12th century middens at Ower (Winder 1991) on the southern shore of the Poole Harbour, neither are likely to have been permanent habitation sites. The shells excavated from these sites are thought to result from processing of the meats prior to marketing, with the shells being discarded on the spot, so the shells in the middens would probably represent the entirety of the catches. On sites such as Elms Farm in Essex near to the head of the Blackwater estuary, famed for its oyster beds, the smaller numbers of shells remaining on site from Roman and early Saxon phases may well indicate that the majority of the catch was being marketed in the shell. Oyster shells are very bulky and can present a disposal problem when fishing for and eating oysters is an important part of community life; so an alternative possibility to consider is that the shells may have been recycled. They can, for example, be returned to the sea bed as cultch on which oyster spat can settle; used to fertilise (lime) the fields; used in the manufacture of lime; crushed for chicken feed, shell-tempered pottery, medicines and cosmetics; used as hardcore, for paths and yard surfaces; and used as mortar for stone work. Another question might be how representative are the shells from an individual site of the original incoming samples to that site – both in quality and quantity? Moorgate and Coleman Street excavations in London uncovered two 11th–12th century domestic rubbish pits with strikingly different oyster shells in each. One contained poor quality oysters of very small and very large size, while the other had all the better quality shells of the optimal mid-size range. It is easy to see how erroneous conclusions could have been drawn if the specimens from only one pit had been selected for analysis. Has there been an excavation bias with only the larger or intact shells being retained? We need to know the criteria for retrieval. And subsequently, what was the rationale

252

Jessica Winder

for selecting samples for analysis? How much reliance can be placed on comparisons of archaeological oyster shells with samples of modern material from known locations? Comparisons of this sort would be very informative. However, there have been substantial losses of natural oyster beds in Britain, plus coastline and sea-level changes, and possible contamination of native oyster beds by interbreeding with imported oysters from home and abroad. Finally, the taphonomic history of the shells, soil conditions and disposal methods will affect the chemical and mechanical wear on the shells. There is randomness to shell survival and recovery as well as to the process of shells being made available for study. All of these factors have to be considered and they place constraints on the interpretations based on the shells. Additionally, there can never be enough samples. Only with this awareness can the data from oyster shells be analysed.

Summing up and future work The elementary nature of the preliminary analyses reflects an original requirement to develop methods for the study of Ostrea edulis shells that were easy to learn and to replicate on a wider scale by on-site non-specialists constrained by limited time, funding and expertise. The advantages of the devised methods are their simplicity and ease of application. The drawbacks are their labour intensiveness, lack of consistency in recording between individuals, and paucity of suitable modern comparative material. Despite these difficulties, the methods have provided a means of addressing questions about the use of oysters and disposal of their shells, of distinguishing farmed from natural wild oysters, and suggesting source locations of oyster beds, by using records of multiple macroscopic characteristics to make spatial and temporal comparisons and statistically identify clusters of associated samples. However, it seems time to consider whether the questions originally posed could now be addressed more effectively with alternative techniques; and whether a different set of questions could be addressed. It seems likely that oyster shells, as with many other species of marine mollusc from archaeological or historic deposits, may still contain protein which is potentially recoverable from within and between the crystals of the shell matrix as well as in any surviving ligament. Archaeological O. edulis shells often retain colour banding caused by organic pigments and occasionally remnants of hinge ligaments; and closed system pigment proteins have been extracted from fossil brachiopod shell crystals (eg, Comfort 1951; Bouniol 1982; Curry 1991; 1999; Fox 1996; and Evans et al. 2009). A number of researchers have also studied shell ligaments showing open system presence of amino acids in various species including the Pearl Oyster Pinctada maxima (eg, Zhang 2007 and De Paula & Silveira 2009). If proteins can survive in ancient mollusc shells, this may allow the calculation of time-since-death by amino acid racemisation in O. edulis as well as by radiocarbon dating. Radiocarbon dating has already been applied to historical deposits of O. edulis shells by Horsey and Winder (1991; 1992) and Reimer (2014) but radiocarbon dating gives wide time margins and calculating the appropriate carbon reservoir correction can be a

14. Oysters in archaeology

253

problem (Chapters 21 & 22 by Fernandes & Dreves, & Douka) Amino acid racemisation (AAR) for dating shells from archaeological deposits is increasingly used, providing shorter time-scale assessments that can be used in conjunction with radiocarbon dating. A substantial body of published work would serve as the basis for any similar studies in O. edulis. It is also possible, as the work of Demarchi et al. (2011) has shown, that AAR use can even detect exposure of shells to heat from fires, which may allow the identification of cooking. Of even greater interest is the field of proteomics using closed or open system protein lacking genetic material, which has the potential to allow the problem of distinguishing between oysters from different populations and localities to be addressed. The ability to derive unique amino acid profiles (that might be termed protein ‘fingerprinting’) using pattern recognition methods on bulk amino acid composition of stable intra-crystalline proteins preserved in biominerals has been demonstrated by Demarchi et al. (2014). The further tool for future archaeological oyster shell study might be trace element analysis. This could lead to the identification of source oyster beds by enabling comparison of the chemical constituents of shells from different geographical locations, whilst also contributing to knowledge of heavy metal accumulation in coastal sediments and waters that have accompanied increasing industrialisation and pollution. The techniques for this have been developed in a variety of marine shell species from archaeological deposits and recent specimens including oysters (eg, Claassen & Sigmann 1993, Markwitz et al. 2003). Significantly Medakovic et al. (2006) showed that the malformed chambers present in the inner nacreous layers of the O. edulis exposed to TBT pollution in the marine environment contained tin. Chambers are one of the macroscopic features routinely recorded by the earlier methodology, and the contents of them could be an important source of information about past environments. Methods that might be applicable to trace element analysis of O. edulis shells have been developed by Schone (2008), Bougeois et al. (2014), Pourang et al. (2014), Ferella et al. (1973), Boyden et al. (1981), and Zacherl et al. (2009). The main achievement from applying simple macroscopic character recording and basic statistical analysis to archaeological oyster shells has been that the resulting detailed information provides a unique descriptive identifier for each sample that can then be used to compare and contrast samples in space and time, and make distinctions between groups of samples, leading to the development of theoretical models of oyster exploitation. Assiduously applied and further developed, the methodology remains a useful tool for understanding how this marine resource has been utilised over at least the last two thousand years in Britain. The use of rapidly developing efficient technologies for amino acid racemisation dating, proteomics studies, and trace element analysis, to investigate deeper structural and constituent aspects of ancient Ostrea edulis shells, could potentially augment the existing foundational database by more directly and objectively addressing questions regarding point of origin; and also supplying baseline information for investigations into oyster population and evolutionary studies, effects of climate change and ocean acidification, and the monitoring of pollution and contamination by metals in an increasingly industrialised world.

254

Jessica Winder

Endnote The axis of an oyster shell that runs from the dorsal to the ventral margin is known as the height dimension in biological terminology. However, the height has frequently but incorrectly been called width in much archaeological research and this is in common usage in many archaeological reports. The difference in nomenclature does not affect the integrity of the analyses undertaken because the so-called width measurement is the same as for height. It is clear to archaeo-zoologists that the two names refer to the same dimension/measurement. Caveat: a problem arises, however, where some early career analysts of archaeological data have confused and transposed the terms length and width when referring to bivalve dimensions. This has resulted in incompatibilities between assemblage datasets, between the associated analyses, and between the reporting of those analyses. Where such reports have been published this gives particular cause for concern.

References Adams, J., Watson, K. & Allen, M. J. 2000. Sinah circle structure. In Allen, M. J. & Gardiner, J., Our Changing Coast: a survey of the intertidal archaeology of Langstone Harbour, Hampshire, 84–86 & 88–123. York: Council for British Archaeology Report 124 Bailey, G. N. 1975. The role of molluscs in coastal economies: the results of midden analysis in Australia, Journal of Archaeological Science 2, 45–62 Bailey, G. N. 1978. Shell middens as indicators of post-glacial economies: a territorial perspective. In Mellars, P. (ed.), The Early Postglacial Settlement of Northern Europe, 37–63. London: Duckworth Bailey, G. N. 1983. Problems of site formation and the interpretation of spatial and temporal discontinuities in the distribution of coastal middens. In Masters, P. M. & Fleming, M. C. (eds), Quaternary Coastlines and Marine Archaeology, 559–582. London: Academic Press Bailey, G. N. & Parkington, J. 1988. The archaeology of prehistoric coastlines: an introduction. In Bailey, G. & Parkington, J. (eds), The Archaeology of Prehistoric Coastlines, 1–10. New Directions in Archaeology. Cambridge: University Press Beaumont, A., Garcia, M. T., Honig, S. & Low, P. 2006. Genetics of Scottish Populations of the native oyster, Ostrea edulis: gene flow, human intervention and conservation, Aquatic Living Resources 19, 389–402 Boekschoten, G. J. 1966. Shell borings of sessile epibiontic organisms as palaeo-ecological guides (with examples from the Dutch Coast), Palaeogeography, Palaeoclimatology, Palaeoecology 2, 333–379 Bougeois, L., de Rafélis, M., Reichart, G-J., de Nooijer, L. J., Nicollin, F. & Dupont-Nivet, G. 2014. A high resolution study of trace elements and stable isotopes in oyster shells to estimate Central Asian Middle Eocene seasonality, Chemical Geology 363, 200–212 Bouniol, P. 1982. L’ornamentation pigmentaire des coquilles de Cerithides actuels et fossiles (S.L.); apport de la technique de l’ultra-violet, Malacologia 22(1–2), 313–317 Boyden, C. R. & Phillips, D. J. H. 1981. Seasonal variation and inherent variability of trace elements in oysters and their implications for indicator studies, Marine Ecology – Progress Series 5, 29–40 Campbell, G. 2010. Oysters ancient and modern: potential shape variation with habitat in flat oysters (Ostrea edulis L.) and its possible use in archaeology, MUNIBE Supplemento – Gehigarria 31, 176–187. Donostia: San Sebastian, DLSS 1055 2010 Carriker, M. R. & Yochelson, E. L. 1968. Recent gastropod boreholes and Ordovician cylindrical borings. Contributions to Palaeontology, Geological Survey Professional Paper 593-B. Washington: U.S. Government Printing Office

14. Oysters in archaeology

255

Cherry, F., Gamble, C. & Shennan, S. 1978. Sampling in Contemporary British Archaeology. Oxford: British Archaeological Report 50 Claassen, C. 1998. Shells. Cambridge: University Press Claassen, C. & Sigmann, S. 1993. Sourcing Busycon artifacts of the eastern United States, American Antiquity 58(2), 333–347 Comfort, A. 1951. The pigmentation of molluscan shells, Biological Review 26, 285–301 Cox, P. W. & Hearne, C. M. 1991. Redeemed from the Heath: the archaeology of the Wytch Farm Oilfield (1987–90), 122–147. Dorchester: Dorset Natural History & Archaeological Society Monograph Series 9 Curry, G. 1991. Fossils in colour, New Scientist (16 November) 1795, 32–34 Curry, G. B. 1999. Original shell colouration in late Pleistocene terebratulid brachiopods from New Zealand, Palaeontologia Electronica 2(2), 1–31 Deith, M. R. 1983a. Molluscan calendars: the use of growth-line analysis to establish seasonality of shellfish collection at the Mesolithic site at Morton Fife, Journal of Archaeological Science 10, 423–440 Deith, M. R. 1983b. Seasonality of shellfish collecting determined by oxygen isotope analysis of marine shells from Asturian sites in Cantabria. In Grigson, C. & Clutton-Brock, J. (eds), Animals and Archaeology 2: shell middens, fishes and birds, 67–76. Oxford: British Archaeological Report S183 Deith, M. R. 1985a. Subsistence strategies at a Mesolithic camp site: evidence from stable isotope analyses of shells, Journal of Archaeological Science 13, 61–78 Deith, M. R. 1985b. Seasonality from shells: an evaluation of two techniques for seasonal dating of marine molluscs. In Fieller, N. R. J, Gilbertson, D. D. & Ralph, N. G. A. (eds), Palaeoenvironmental Investigations: palaeobiology, 119–130. Oxford: British Archaeological Report S226 Deith, M. R. 1988. A molluscan perspective on the role of foraging in Neolithic farming economies. In Bailey, G. & Parkington, J. (eds), The Archaeology of Prehistoric Coastlines, 116–124. New Directions in Archaeology. Cambridge: University Press De Paula, S. M. & Silveira, M. 2009. Studies on molluscan shells: contributions from microscopic and analytical methods, Micron 40, 669–690 Demarchi, B., Williams, M. G. M., Milner, N., Russell, N., Bailey, G. N. & Penkman, K. 2011. Amino acid racemisation dating of marine shells: a mound of possibilities, Quaternary International 239(1–2), 114–124 Demarchi, B., Rogers, K., Fa. D. A., Finlayson, C. J., Milner, N. & Penkman, K. E. H. 2013. Intracrystalline protein diagenesis (IcPD) in Patella vulgata: Part 1: Isolation and testing of the closed system, Quaternary Geochronology 16, 144–157 Demarchi, B., O’Connor, S., de Lima Ponzoni, A., de Almeida Rocha Ponzoni, R., Sheridan, A., Penkman, K. E. H., Hancock, Y. & Wilson, J. C. 2014. An integrated approach to the taxonomic identification of prehistoric shell ornaments. Public Library Of Science ONE 9(6): e99839. doi:10.1371/journal.pone.0099839 Douka, K. Radiocarbon dating of marine and terrestrial shell. In Allen, M. J. (ed.), Molluscs in Archaeology, 382–400. Oxford: Oxbow Books Evans, S., Camara, M. D. & Langdon, C. J. 2009. Heritability of shell pigmentation in the Pacific Oyster, Crassostrea gigas, Aquaculture 286(3–4), 211–216 Ferella, R., Carville, T. & Martinez, J. 1973. Trace metals in oyster shells, Environmental Letters 4(4), 311–316 Fernandes, R. & Dreves, A. 2017. Bivalves and radiocarbon. In Allen, M. J. (ed.), Molluscs in Archaeology, 364–381. Oxford: Oxbow Books Fox, D. L. 1966. Pigmentation of Molluscs. In Wilbur. K. M. & Yonge, C. M. (eds), Physiology of Mollusca Volume 2, 249–274. New York: Academic Press

256

Jessica Winder

Frere, S. 1967. Britannia: a history of Roman Britain, 299. London: Routledge & Keegan Paul Gutierrez-Zugasti, I., Andersen, S., Araujo, A, Dupont, C., Milner, N. & Monge-Soares, A. 2011. Shell midden research in Atlantic Europe: state of the art, research problems and perspectives for the future, Quaternary International 239, 70–85 Horsey, I. P. & Winder, J. M. 1991. Late Saxon and Early-Conquest period oyster middens at Poole, Dorset. In Good, G. L. Jones, R. H. & Ponsford, M. W. (eds), Waterfront Archaeology, 102–104. London: Council for British Archaeology Research Report 74 Horsey, I. P. & Winder, J. M. 1992. The Late-Saxon and Conquest-period oyster middens. In Horsey, I. P., Excavations in Poole 1973–83, 60–61. Dorchester: Dorset Natural History & Archaeological Society Monograph 10 Kähler, G. A., Fisher, F. M. Jr. & Sass, R. L. 1976. The chemical composition and mechanical properties of the hinge ligament in bivalve molluscs, Biological Bulletin 151, 161–181 Kent, B. W. 1988. Making Dead Oysters Talk: techniques for analysing oysters from archaeological sites. Maryland Historical Trust, Historic St Marys City: Jefferson Patterson Park and Museum Koike, H. 1979. Seasonality of shell collecting activity and accumulation rate of shell midden sites in Canto, Japan, Dalyonki Kenkyu (The Quaternary Research) 17, 267–278 Koike, H. 1981. Seasonality of shellfish gathering at Isarago shell-midden. In Zuzuki, K. (ed.), Isarago Shell-midden, 607–615. Tokyo: Education Committee of Minato Ward Markwitz, A., Barry, B. Gauldie, R. W. & Roberts, R. D. 2003. Probing for heavy element impurities in the shell of the Pacific Oyster, Crassostrea gigas, with nuclear microscopy. Nuclear Instruments and Methods in Physics Research, Section B: Beam interactions with materials and atoms 210, September 2003, 418–423 Medakovic, D., Traverso, P., Bottino, C. & Popovic, S. 2006. Shell layers of Ostrea edulis as an environmental indicator of TBT pollution: the contribution of surface techniques, Surface and Interface Analysis 38, 313–316 Meehan, B. 1982. Shell Bed to Shell Midden. Canberra: Australian Institute of Aboriginal Studies Mellars, P. 1978. Excavation and economic analysis of Mesolithic shell middens on the island of Oronsay (Inner Hebrides). In Mellars, P. (ed.), The Early Post-glacial Settlement of Northern Europe, 371–396. London: Duckworth Mellars, P. 1987. Excavations on Oronsay: prehistoric human ecology on a small island. Edinburgh: University Press Milner, N. 2001. At the cutting edge: using thin sectioning to determine season of death of the European Oyster, Ostrea edulis, Journal of Archaeological Science 28, 861–873 Milner, N. 2002. Incremental Growth of the European Oyster, Ostrea edulis: Seasonality information from Danish kitchenmiddens. Oxford: British Archaeological Report S1057 Milner, N. 2009. Mesolithic middens and marine molluscs: procurement and consumption of shellfish at the site of Sand, Scottish Archaeology Internet Publications 31 Milner, N. 2013. Human impacts on oyster resources at the Mesolithic-Neolithic transition in Denmark. In Thompson, V. D. & Waggoner, J. C. (eds), The Archaeology and Historical Ecology of Small Scale Economies, 17–40. Gainesville: University Press of Florida Milner, N. & Barrett, J. H. 2012. The maritime economy: mollusc shell. In Barrett, J. (ed.), Being an Islander: production and identity at Quoygrew, Orkney, AD 900–1600, 105–115. Cambridge: McDonald Institute for Archaeological Research Milner, N. & Woodman, P. 2002. Oysters, cockles and kitchenmiddens: consuming shellfish on Danish middens. In Miracle, P. T. & Milner, N. (eds), Consuming Patterns and Patterns of Consumption, 89–96. Cambridge: McDonald Institute for Archaeological Research Milner, N., Barrett, J. & Welsh, J. 2007. Marine resource intensification in Viking Age Europe: the molluscan evidence from Quoygrew, Orkney, Journal of Archaeological Science 34(9), 1461–1472 Monckton, A. 1999. Oysters. In Connor, A. & Buckley, R. (eds), Roman and Medieval Occupation in

14. Oysters in archaeology

257

Causeway Lane, Leicester: Excavations 1980 and 1991, 337–340. Leicester Archaeology Monograph 5. Leicester: University of Leicester Archaeological Services Pourang, N., Richardson, C. A., Chenery, S. R. N. & Nasrollahzedeh, H. 2014. Assessment of trace elements in the shell layers and soft tissues of the pearl oyster Pinctada radiata using multivariate analyses: a potential proxy for temporal and spatial variations of trace elements, Environmental Monitoring Assessment 186, 2465–2485 Reimer, P. J. 2014. Marine or estuarine radiocarbon reservoir corrections for molluscs? A case study from a medieval site in the south of England, Journal of Archaeological Science 49, 142–146 Schone, B. R. 2008. The curse of physiology – challenges and opportunities in the interpretation of geochemical data from mollusc shells, Geo-Marine Letters 28, 269–285 Shackleton, J. C. 1983. An approach to determining prehistoric shellfish collecting patterns. In Grigson, C. & Clutton-Brock, J. (eds), Animals and Archaeology 2: shell middens, fishes and birds, 77–85. Oxford: British Archaeological Report S183 Shackleton, J. C. 1988. Reconstructing past shorelines as an approach to determining factors affecting shellfish collecting in the prehistoric past. In Bailey, G. & Parkington, J. (eds), The Archaeology of Prehistoric Coastlines, 11–21. New Directions in Archaeology. Cambridge: University Press Surge, D. & Milner, N. 2003. Oyster shells as history books, Shellfish News 16, 5–7 Troels-Smith, J. 1967. The Ertebolle culture and its background, Palaeohistoria 12, 505–528 Watabe. N. 1984. Shell. In Bereiter-Hahn, J. Matolsy, A. G. & Richards, K. S. (eds), Biology of the Integument, Volume 1, Invertebrates, 448–485, Berlin & New York: Springer Winder, J. M. 1980. The marine Mollusca. In Holdsworth, P., Excavation at Melbourne Street, Southampton, 1971–76, 121–127. London: Council for British Archaeology Report 33 Winder, J. M. 1985. Oyster culture. In Milne, G. (ed.), The Port of Roman London, 91–95. London: Batsford Winder, J. M. 1991. Marine Mollusca. In Cox, P. W. & Hearne, C. M. Redeemed from the Heath – the archaeology of the Wytch Farm Oilfield (1987–90), 212–216. Dorchester: Dorset Natural History & Archaeological Society Monograph 9 Winder, J. M. 1992a.The Oysters. In Horsey, I.P. Excavations in Poole 1973–83, 194–200. Dorchester: Dorset Natural History & Archaeological Society Monograph 10 Winder, J. M. 1992b. A Study of the Variation in Oyster Shells from Archaeological Sites and a Discussion of Oyster Exploitation. Unpublished PhD Thesis, University of Southampton Winder, J. M. 1993a. Oyster and other marine mollusc shells. In Woodward, P. J., Davies, S. M. & Graham, A. H., Excavations at the Old Methodist Chapel and Greyhound Yard, Dorchester, 1982–1984, 347–348. Dorchester: Dorset Natural History & Archaeology Society Monograph 12 Winder, J. M. 1993b. Oyster and other marine mollusc shells. In Lucas, R. M., The Romano-British Villa at Halstock, Dorset Excavations 1967–1985, 114–116. Dorchester: Dorset Natural History & Archaeological Society Monograph 13 Winder, J. M. 1994a. The marine mollusc shells. In Watkins, D. W., The Foundry: excavations on Poole Waterfront 1986–87, 84–88. Dorchester: Dorset Natural History & Archaeological Society Monograph 14 Winder, J. M. 1994b. Oyster shells, 110–111. In Papworth, M. (ed.), Lodge Farm, Kingston Lacey Estate, Dorset, Journal of the British Archaeological Association 147, 57–121 Winder, J. M. 1997a. Oyster and other shells. In Hawkes, J. W. & Fasham, P. F., Excavations on the Reading Waterfront Sites, 1979–1988, 105–107. Salisbury: Wessex Archaeology Report 5 Winder, J. M. 1997b. Oyster and other marine molluscs. In Andrews, P. (ed.), Excavations at Hamwic, Volume 2: excavations at Six Dials, 247. York: Council for British Archaeology Research Report 109 Winder, J. M. 2000a. The marine molluscs. In Ellis, P. (ed.), Ludgershall Castle – Excavations

258

Jessica Winder

by Peter Addyman 1964–1972, 241–242. Devizes: Wiltshire Archaeological & Natural History Society Monograph 2 Winder, J. M. 2000b. Oysters – their variations in time and space. In Carr, A. P., Seaward, D. R. & Sterling, P. H. (eds), The Fleet Lagoon and Chesil Beach, 93–94. Revised edition. Dorchester: Chesil Bank & the Fleet Nature Reserve with Dorset County Council Winder, J. M. 2011. Oyster Shells from Archaeological Sites: a brief illustrated guide to basic processing. Online publication at: https://oystersetcetera.files.wordpress.com/2011/03/ oystershellmethodsmanualversion11.pdf Winder, J. M. & Gerber-Parfitt, S. 2003. The oyster shells. In Malcolm, G. & Bowsher, D. with Cowie, R. (eds), Middle Saxon London – Excavations at the Royal Opera House 1989–99, 325–332. London: Museum of London Archaeology Service Monograph 15 Winder, J. M. & Reidy, K. 1996. Marine Mollusca, 168–169. In Medlycott, M., A medieval farm and its landscape – Stebbingford, Felstead 1993, Essex Archaeology and History 27, 102–181 Wyles, S. F. & Winder, J. M. 2000. Marine Mollusca. In Young, C. J. (ed.), Excavations at Carisbrooke Castle, Isle of Wight, 1921–1996, 184–188. Salisbury: Wessex Archaeology Report 18 Yonge, C. M. 1960. Oysters. New Naturalist Special Volume. London: Collins Zacherl, D. C., Morgan, S. G., Swearer, S. E. & Warner, R. R. 2009. A shell of its former self: can Ostrea lurida Carpenter 1863 larval shells reveal information about a recruit’s birth location?, Journal of Shellfish Research 28(1), 23–32 Zhang, G. 2007. Photonic crystal type structure in bivalve ligament of Pinctada maxima, Chinese Science Bulletin 53(8), 1136–1138

15. Shell middens Karen Hardy

Shell middens are found on many coastlines around the world. The accumulation of shells into piles, which can vary in size from a few centimetres to vast landscapechanging mounds, appears to have begun in earnest during the mid-Holocene period. While these accumulations have left us with an abundant resource with which to study ancient coastal human populations, shell middens can be complex to study, in terms of their abundance which can be overwhelming, and at times their size, which can make excavation complex, time-consuming and expensive. This chapter investigates what constitutes a shell midden, examines the different approaches used to study them, and explores possible reasons why people might have accumulated shells into middens and mounds. There is a very wide range of sites that fall into the shell midden category and we explore the ways this specialised branch of excavation and study is used to provide information on past human populations.

Shell middens Shellfish offer readily available resources that are normally edible, abundant, easy to process and accessible all year round. They live in shells formed principally from calcium carbonate which constitute their exoskeleton. Shell does not readily degrade and is not soluble in water and this has resulted in the accumulation of shell waste into piles, known as shell middens. Shell middens are found in their thousands on coastlines around the world, and can range from small scatterings of shells to huge landscapechanging mounds which at times contain multiple burials. The term ‘shell midden’ is generic and encompasses a very wide range of shelly deposits. This means that the accumulation of shells on archaeological sites goes by a range of different names. There are few guidelines about definition of these deposits other than in Denmark where the term kitchen midden (Køkkenmødding) is reserved for sites with a volume of at least 50% marine shell and a size over 10 m2 (Gutiérrez-Zugasti et al. 2011). Other than this there is no clear definition of when a scatter becomes an accumulation, when an accumulation becomes a midden and what the difference between middens, mounds and mega-middens are, except that these last two terms tend to be used to describe

260

Karen Hardy

larger sites. While the names are to some extent the arbitrary decision of the excavator, in general terms, a scatter is more likely to be a small spread while an accumulation may be a small pile. An accumulation becomes a midden largely when multiple piles suggest repeated deposition. This chapter examines where and when shell middens occur, archaeological methods of survey and excavation of middens and highlights some of the theories about why shells were accumulated, and ways shell middens can provide information on past human populations.

Where and when do shell middens occur? Shell middens occur in many places across the world and appear in abundance around the mid-Holocene. In Europe, this has resulted in shell middens becoming particularly correlated with the later hunter gatherer phases, most notably the Mesolithic; in reality, shell middens occur across a much broader time scale ranging from the Middle Palaeolithic up to the present day. The earliest archaeological evidence for shellfishing currently known is at Pinnacle Point, in South Africa. Pinnacle Point is an AMH (anatomically modern human) site and dates to the MIS (marine isotope stage) 6 (Marean 2010), around 162,000 years ago (Jerardino & Marean 2010). At almost the same time, 150,000 years ago, evidence from Bajondillo Cave in south Spain has revealed that Neanderthals also exploited shellfish (Cortés-Sánchez et al. 2011). During the subsequent MIS 5 (130,000–71,000 years ago) stage, evidence for shellfish exploitation becomes widespread along the southern African coastline (Klein & Steele 2013) and on many southern European Neanderthal sites (Cortés-Sánchez et al. 2011). In many cases, the shell accumulations are equal in size to those found in later mid-Holocene periods (Klein & Steele 2013). It is very likely, however, that shellfish were exploited even earlier than the first archaeological evidence for them, and the readily available food they provide has been used as an argument in favour of coastlines as hominin dispersal routes (Erlandson 2001; Bailey et al. 2007; Mellars et al. 2013). The way some higher primates exploit shellfish may reflect early hominin exploitation and would explain why there is little evidence for this in the archaeological record prior to 150,000 years ago. For example, long-tailed macaques exploit a range of species including bivalves, gastropods and crabs which they crack open using stone hammers onto rocks on the shore (Gumert et al. 2009; Gumert & Malaivijitnond 2012; Tan et al. 2015). Once the meat is extracted, the shells are left where they are dropped therefore accumulation does not occur. Around the mid-Holocene, shell middens appear in many places across the world, suggesting an increasing focus on coastal resources at this time. However, sea level rise may also have played a part in the apparently simultaneous appearance of shell middens as earlier coastline are now submerged in many regions (Bailey & Flemming 2008; Benjamin et al. 2011). In some places, there may be other causes for the relatively rapid appearance of shell middens; for example, Bicho et al. (2010) suggest that in central Portugal the development of settlement along the Tagus valley, with its associated Muge shell middens, may have been a response to climatic deterioration linked to the 8.2 kya abrupt Holocene climate change event (see Alley and Ágústsdóttir 2005).

15. Shell middens

261

The focus on coastal resources, reflected by shell middens in the mid-Holocene, is well established in many places across the world; even so, the number of surviving middens, for example in west Scotland, is too small to reflect this. It is very likely that the middens here and elsewhere represent only a very small proportion of those that originally existed (Hardy 2013). It has been argued that marine based resources became subsequently less important with the introduction of farming at the start of the Neolithic (Richards et al. 2003) and there is little doubt that the abundance of evidence for coastal resources diminishes in some places with the onset of farming. However, evidence for coastal and marine resource use is relatively widespread in the Danish and Baltic Neolithic record (Craig et al. 2011; Fischer 2007) while Hardy (2015) has shown that evidence for the exploitation of shellfish and coastal resources does not disappear at any point in the coastal archaeological record in northwest Scotland. The later prehistoric evidence from here suggests multiple small scale collection events, which indicates a change in the use of coastal resources after the Mesolithic, possibly representing a marginalisation of these in place of larger accumulations that may indicate a more significant, possibly seasonal role for coastal resource use, during the Mesolithic. Outside Europe, shell middens continued to develop in places across the world in parallel with farming (for example Korea, see Crawford & Lee 2003), and in some cases, continue today (Hardy et al. 2016). Shell middens vary hugely, they can range from small accumulations, often in caves or rockshelters, to very large mounds which can reach huge, landscape changing proportions. Places where such large accumulations occur include Denmark (Andersen 2000), South Africa (Parkington et al.1988), the Saloum Delta, Senegal (Camara 2010), the sambaqui middens from Brazil (Gaspar et al. 2008), Cape York, Australia (Ulm 2011) and parts of north America including British Colombia (Cannon 2000a; 2000b), the Ohio river valley (Claassen 2013) and Florida (Schwadron 2013). The scale of these shell middens can be vast and in many cases they can contain multiple burials or even cemeteries, for example at Jabuticabeira II, in Brazil there are an estimated 43,000 burials (Gaspar et al. 2008) while at Fadiouth, in Senegal, the shell midden continues to be used as a cemetery today (Fig. 15.1). In some places, high concentrations of smaller shell middens are also found including the Farasan Islands, Red Sea (Bailey et al. 2007), the Mauretanian coastline (Vernet 2013), Tierra del Fuego (Estévez et al. 2001; Orquera & Piana 2009) and Japan (Habu et al. 2011).

Survey and excavation of shell middens Shell middens can be complex and expensive to excavate, and at times their sheer size and abundance is overwhelming. Bailey and Hardy (2013) highlight the importance of understanding taphonomic processes, which can be highly variable, and also the need to develop sampling strategies which reflect the sometimes vast quantities of potential data that is available. The shell matrix is bulky, shells can be wholly or partially fragmented, and therefore post excavation requires space and time while detailed sorting can be slow and expensive. Additionally, due to the calcium carbonate of the shells, many

262

Karen Hardy

Figure 15.1. Active cemetery in the shell midden at Fadiouth, Saloum Delta, Senegal shell middens have excellent survival of organic remains, including bone and plants, therefore extensive flotation strategies can reap rich rewards. Examples of approaches to understand shell middens range from large scale survey and excavation to minute reconstructions of individual events. A method that is currently being developed in the Saloum Delta Senegal, which is renowned for its numerous vast shell middens which date to at least 5000 BP, involves the use of Google Maps to reconstruct midden distribution. The abundant, huge shell middens, together with the enduring economic practices which continue to be largely based on traditional shellfish collecting, form the basis for the UNESCO World Heritage Site here (http:// whc.unesco.org/en/list/1359). Thousands of shell middens exist, many of which are very large. For example the site of Diorom Boumack is 250 × 450 m and 12 m high (Fig. 15.2), while Tioupane stretched for almost a kilometre, though much has been destroyed now (Hardy et al. 2016). The area has little modern infrastructure and field programmes are expensive and logistically challenging. However, the resolution of the maps is very high and work is currently in progress to create a distribution map of the larger shell middens across the entire area. While this cannot replace on-the-ground survey work, it can provide a preliminary indication of the quantities and distribution

15. Shell middens

263

Figure 15.2. Diorom Boumack shell midden, Saloum Delta, Senegal of the larger shell middens. In this complex archaeological landscape, it offers a context within which to develop a strategy and programme of study. However, this method only works in certain places. In the Saloum Delta shell middens are clearly visible against the background of the very consistent mangrove vegetation; an attempt to develop this work northwards into Mauretania where huge numbers of shell middens can be found right along the coastline (Vernet 2013) has so far failed due to the lack of relief in the sandy landscape. At the other end of the spectrum, Estévez and Vila (2006) describe excavation of a shell midden at a minute scale, based on ‘peeling’ of very thin stratigraphic units and consolidated by soil micromorphological analyses (Balbo et al. 2010). This has permitted identification of a detailed sequence of events, at times to the level of individual activities as well as formation processes.

Sampling When faced with hundreds or thousands of shell middens and mounds, sampling strategies need to be designed to obtain an overview of the area as well as a closer understanding of age and why and how the shell middens were constructed. This is the case in the Farasan Islands, Red Sea, where Bailey et al. (2013) separated the middens into three groups based on their size and shape. They then created a distribution map of the middens using general features including size, shell type and artefacts. Finally they selected two sites for trench excavation and cut a section through the middens. This permitted them to examine broad interpretations of the middens while radiocarbon date sequences provided insights into the length of time the middens were in construction. Another method used by Cannon (2000a; 2000b) involves the extraction of cores and bucket auger samples to characterise large shell middens in British Colombia, Canada. In west Scotland, the Scotland’s First Settlers survey project was unexpectedly confronted with numerous shell deposits which were found in many of the recorded coastal caves and rockshelters. A map was created of the shell midden distribution and an attempt to characterise the shell middens was implemented through a programme of selective test pitting and shovel pitting (Hardy & Wickham-Jones 2009). While the presence of midden deposits was recorded, the characterisation and age determination of the deposits was only partly successful due to rockfall in most rockshelter sites, which permitted only a

264

Karen Hardy

Figure 15.3. Active shell midden, Djifere, Saloum Delta. Note the proximity to the shore, the path up the midden to the processing area, and the areas of erosion on the midden

partial perspective of the depth and age of the deposits. Sampling strategies also need to be developed to evaluate the relative results of material found within the sampled sediments. Estévez et al. (2001) provide a useful example of how to evaluate relative proportions of fish and animal bones from sampled shell middens and how to apply this to an understanding of variation in resource exploitation and an understanding of subsistence strategies while Cannon (2000a) uses samples taken from bucket augers to distinguish between long term habitation sites and specialist camp sites.

Taphonomy Shell and shell middens are highly susceptible to taphonomic processes and an understanding of formation processes are integral to shell midden archaeology. Stein (1992) in particular was important in developing the study of archaeological shells as sedimentary particles, subject to natural effects including wind, water and soil. Estévez et al. (2013) have applied this perspective to the study of ethnographic shell middens in Tierra del Fuego where detailed reconstruction of taphonomic and cultural effects, and other analytical techniques, have permitted reconstruction of activity at the midden in minute detail, at times identifying multiple sub-units sometimes at the level of individual activity events. Claassen (1998) provides a detailed analysis on the many taphonomic effects that can occur on individual shells and shell middens. These include the need to ensure that any shell accumulation is evaluated for its potential as a natural versus cultural deposit. Shell accumulations can be significantly disrupted by animals, notably worms, moles, rodents and birds, as well as environmental processes such as wave action. Trampling can also alter the way shell is distributed and fragmented, and in some cases, paths can be seen along or over middens (Fig. 15.3). Shell midden material is a very useful fertiliser and middens or parts of middens can be removed for this purpose. In the Saloum Delta, whole shell middens have been removed and the shell extracted for use in the construction industry, in particular on roads and buildings (Thiam 2013).

15. Shell middens

265

Figure 15.4. Profile of trench A, Sand rockshelter. Note the earlier context 028 (Mesolithic shell midden) overlying the contexts 022 and 027 which had Neolithic dates (from Hardy & Wickham-Jones 2009)

A recent pilot project in the Saloum Delta, Senegal is providing new insights into the way large shell middens develop. Observation of primary deposition of shelly waste in Senegal has demonstrated how this can occur over a wide area (Hardy et al. 2016). Separate accumulations slowly merge and eventually join up to the point where there is an extensive area of midden with both horizontal and vertical deposition occurring. The potential for erosion and consequent mixing of deposits is also evident in these large unconsolidated piles as can be seen in Figure 15.3. An archaeological example of slumped midden material can be found at Sand in Applecross where a polished stone axe was found in a deposit with younger radiocarbon dates than the overlying shell midden (Hardy & Wickham-Jones 2009) (Fig. 15.4). It is important also to remember that shell middens can consist of unconsolidated, clast-supported sediment with shell therefore movement of small material within middens is common (Claassen 1998).

What can shell middens tell us about previous populations? Shell middens have been a subject of interest since the early 19th century initially with the recognition and study of the Danish Køkkenmødding (Kitchen middens) (Andersen & Johansen 1986). This was rapidly followed by studies on shell middens in other places notably in Atlantic Europe including Scotland (Mellars 1978; 1987; Mellars & Payne 1971; Ricards & Mellars 1998), France, Spain and Portugal (Waselkov 1987; GutiérrezZugasti et al. 2011). In Scotland, the ‘Obanian Culture’ was proposed on the basis of shell middens and the artefacts recovered from sites on Oronsay and in and around Oban (Lacaille 1954) though this is no longer recognised as a distinct culture (Bonsall 1996; Hardy & Wickham-Jones 2009). But despite this longevity of study and their visibility and abundance, shell middens are still not well understood. Ethnographic investigations into shellfish collecting and accumulation can shed light on the different ways that shell middens accumulate and the significance of midden deposits in terms of diet and people’s lives. For example, Meehan (1982); Bird and Bliege

266

Karen Hardy

Figure 15.5. Roasting shellfish by the shore, Diakhanor, Saloum Delta, Senegal

Bird (1997; 2002), have noted that collection for direct consumption is often based on the easiest way to obtain edible species. Likewise, the proportion of shell types found in middens does not correlate with parallel proportions of dietary composition due to selective on site processing and portage (Meehan 1982; Bird & Bliege Bird 1997; 2002). Meehan (1982) provides examples of small scale processing, including cooking, and consumption of shellfish, off site which would be unlikely to leave any enduring evidence among the Anbarra in northern Australia. It is widely assumed that shellfish in archaeological contexts represent food. In many cases this may be true; however, Hardy (2013; 2015) suggests that accumulation patterns in some later medieval middens in Scotland may also reflect collection for shellfish as bait. This also corresponds with the extensive historical evidence for use of limpets and mussels in particular, for this purpose here (Coull 1996). The use of shellfish as bait is also documented ethnographically in the North West Coast of America (Moss 1993). In the Saloum Delta, bivalves and gastropods are today collected for the purpose of trade within a traditional economic framework (Fig. 15.5); however, while in some places shellfish continue to be accumulated into large shell middens, in others, the large scale processing of shellfish is conducted immediately above the shoreline, and leaves no long term shell waste deposits or any evidence of processing (Hardy et al. 2016). All the middens visited in the Saloum Delta, both modern and ancient, contain large, single species deposits. In all the modern Senegalese contexts studied, this reflects shellfishing for trade; shellfish are rarely eaten and the dietary protein is predominantly based on fish. It is well known that evidence for shellfish is generally over represented in the archaeological record in relation to other resources (Bailey & Hardy 2013) and these examples highlight the need for care in dietary reconstructions based on the middens. For example, Pickard and Bonsall (2014) found a broad range of species suggesting unselected harvesting for food at four Mesolithic and early Neolithic sites in Scotland which are very likely to represent food, while a contrasting example is found at Wetweather Cave, Geodha Smoo in northern Scotland, where single species deposits of dog whelk, and their breakage patterns, have been interpreted as a specialist extraction site for purple dye (Pollard 2005; see Light & Walker, Chapter 19).

15. Shell middens

267

Figure 15.6. Profile of test pit in later prehistoric shell midden deposit: SFS 68, Allt na Criche, Test Pit 2 SSW facing section. (from Hardy & Wickham-Jones 2009)

The way shells are deposited can also provide some insight into the nature of the occupation; for example, Mesolithic shell middens in Scotland, though small by comparison to the vast shell middens in many other places, tend to comprise uninterrupted layers of some depth, which fits well with the perspective of coastal resource use as part of seasonal exploitation of different resources. The later prehistoric middens which largely comprise thin shell layers interspersed with charcoal-rich layers suggest multiple short term visits (Fig. 15.6). In later periods, single species shell middens appear in the 16th century in parts of west Scotland. These have been tentatively linked to bait collection which may be connected to an increase in deep sea fishing around this time; alternatively, the Little Ice Age also peaked in the 16th century which may have led to poorer harvests and a need to expand food sources, while the numerous religious fast days, which add up to as much as 4 months each year, may have resulted in an increased need for marine resources (Hardy 2013).

Why are shells accumulated into middens? Some shell midden deposits have been considered as having a purpose and a meaning beyond discard. The huge sambaqui shell middens in coastal Brazil which date from 9000 to around 2000 years ago have been interpreted as deliberately constructed landmarks and monuments impregnated with symbolic meaning linked to mortuary ritual and a cult of ancestors (Gaspar et al. 2008) while Claassen (2013) suggests the Ohio river valley is a sacred landscape with ceremonial centres and mortuary camps in the shell middens. While this may be so, in many cases, the reality may be more pragmatic. Observations in the Saloum Delta have highlighted some of the realities of shellfish processing, most notably how heavy and awkward they are to carry, and how rapidly they need to be processed following collection. All the new shell middens we observed here were located on the beach as close to the landing spots as possible (Fig. 15.3). However, some older middens have acquired other purposes and in some cases, meanings. Shell deposits act as insulation against the damp ground and huts are placed

268

Karen Hardy

Figure 15.7. Allt na Uamha, Craig, Loch Torridon; a) the rockshelter and, b) and c) partial vista of the Inner Sound from the rockshelter, with An Corran, Loch a Sguirr and Fearnmore 1 Mesolithic sites in view

on middens for this reason. This eventually leads to the middens becoming named villages. Over time, people are buried in the middens, and eventually the middens also therefore become cemeteries and many archaeological shell middens here contain burials. Baobab trees grow on shell middens due to the calcium carbonate-rich soils. These important trees are widely used for food and for extraction of raw materials (Piqué et al. 2016) while some baobab trees house ancestral spirits, which leads to the middens becoming spiritual places. Finally broad vegetation development leads them to become foraging locations (Guèye 2010). There may be many reasons why people accumulated shell into mounds and middens in the past, but it is clear that practical issues must have played a significant part. It is also important to remember that all shell middens begin as individual shell piles and it can take many thousands of years for them to accumulate into large mounds. It is, therefore, likely that their significance and their use will have changed over time as they developed. In the light of this, sites such as the early prehistoric shell midden site of Allt na Uamha at Craig, on the north coast of Loch Torridon in Scotland raises

15. Shell middens

269

interesting questions as it lies around 85 m above present sea level (Hardy & WickhamJones 2009) up a tricky, steep slope, with an extraordinary vista of the Inner Sound including several known Mesolithic sites, which suggests that the primary purpose of the site is unlikely to be have been shellfishing (Hardy et al. 2016) (Fig. 15.7).

Conclusion There are many aspects to shell midden archaeology; new methods of dating are being developed (Demarchi et al. 2011) while the potential for paleoclimatic and environmental reconstruction (eg, Andrus 2011; Azzoug et al. 2012) remains to be fully developed. This chapter touches on only a few topics, and highlighted some of the ways ethnography can be used as a source of insight and inspiration (Bailey & Hardy 2013). Shell middens are complex archaeological deposits, most often sites in their own right. Understanding their taphonomies and creating sampling strategies that can inform, what can be at times bewilderingly abundant archaeological deposits, together with the sheer volume of material excavated, can create challenges. Once the material is excavated, the need for detailed post excavation analysis is paramount, as organic survival can be better in middens than elsewhere. Shell middens offer an outstanding resource for the study of the past though in places many have been destroyed either through sea level rise or development of coastal areas. Today, sea level rise is risking many of those that remain: further archaeological work is needed and will likely reap rich rewards.

References Alley, R. B. & Ágústsdóttir, A. M. 2005. The 8k event: cause and consequences of a major Holocene abrupt climate change, Quaternary Science Reviews 24(10), 1123–1149 Andersen, S. H. 2000. ‘Køkkenmøddinger’ (shell middens) in Denmark: a survey, Proceedings of the Prehistoric Society 66, 361–384 Andersen, S. H. & Johansen, E. 1986. Ertebølle revisited, Journal of Danish Archaeology 5(1), 31–61. doi: 10.1080/0108464X.1986.10589957 Andrus, C. F. T. 2011. Shell midden sclerochronology, Quaternary Science Reviews 30(21), 2892–2905 Azzoug, M., Carré, M., Chase, B. M., Deme, A., Lazar, A., Lazareth, C. E. & De Morais, L. T. 2012. Positive precipitation-evaporation budget from AD 460 to 1090 in the Saloum Delta (Senegal) indicated by mollusk oxygen isotopes, Global and Planetary Change 98, 54–62. doi:10.1016/j. gloplacha.2012.08.003 Bailey, G., Alsharekh, A., Flemming, N., Lambeck, K., Momber, G., Sinclair, A. & Vita-Finzi, C. 2007. Coastal prehistory in the southern Red Sea Basin, underwater archaeology, and the Farasan Islands, Proceedings of the Seminar for Arabian Studies 37, 1–16. Oxford: Archaeopress Bailey, G. & Hardy, K. 2013. Introduction. In Bailey, G., Hardy, K. & Camara, A. (eds), Shell Energy: mollusc shells as coastal resources, 1–6. Oxford: Oxbow Books Bailey, G., Meredith Williams, M. & Alsharekh, A. 2013. Shell mounds of the Farasan Islands, Saudi Arabia. In Bailey, G., Hardy, K. & Camara, A. (eds), Shell Energy: mollusc shells as coastal resources, 241–254. Oxford: Oxbow Books

270

Karen Hardy

Bailey, G. N. & Flemming, N. C. 2008. Archaeology of the continental shelf: marine resources, submerged landscapes and underwater archaeology, Quaternary Science Reviews 27(23), 2153–2165 Balbo, A. L., Madella, M., Vila, A. & Estévez, J. 2010. Micromorphological perspectives on the stratigraphical excavation of shell middens: a first approximation from the ethnohistorical site Tunel VII, Tierra del Fuego (Argentina), Journal of Archaeological Science 37(6), 1252–1259 Benjamin, J., Bonsall, C., Pickard, C. & Fischer, A. (eds) 2011. Submerged Prehistory. Oxford: Oxbow Books Bicho, N., Umbelino, C., Detry, C. & Pereira, T. 2010. The emergence of Muge Mesolithic shell middens in central Portugal and the 8200 cal yr BP cold event, Journal of Island & Coastal Archaeology 5(1), 86–104 Bird, D. W. & Bliege Bird, R. L. 1997. Contemporary shellfish gathering strategies among the Meriam of the Torres Strait Islands, Australia: testing predictions of a central place foraging model, Journal of Archaeological Science 24(1), 39–63 doi:10.1006/jasc.1995.0095 Bird, D. W. & Bliege Bird, R. L. 2002. Children on the Reef, Human Nature 13(2), 269–297 doi: 10.1007/s12110-002-1010-9 Bonsall, C. 1996. The Obanian Problem: coastal adaptation in the Mesolithic of western Scotland. In Pollard, T. & Morrison, A. (eds), The Early Prehistory of Scotland, 183–197. Edinburgh: University Press Camara, A. 2010 Shell middens from the Saloum Delta, Senegal. In Hardy, K. (ed.), Archaeological Invisibility and Forgotten Knowledge, 53–59. Oxford: British Archaeological Report S2183 Cannon, A. 2000a. Assessing variability in Northwest Coast salmon and herring fisheries: bucket-auger sampling of shell midden sites on the central coast of British Columbia, Journal of Archaeological Science 27(8), 725–737 Cannon, A. 2000b. Settlement and sea-levels on the central coast of British Columbia: evidence from shell midden cores, American Antiquity 65(1), 67–77 Claassen, C. 1998. Shells. Cambridge: University Press Claassen, C. 2013. Freshwater shell mounds of the Ohio River valley, USA. In Bailey, G., Hardy, K. & Camara, A. (eds), Shell Energy: mollusc shells as coastal resources, 59–68. Oxford: Oxbow Books Cortés-Sánchez, M., Morales-Muñiz, A., Simón-Vallejo, M. D., Lozano-Francisco, M. C., VeraPeláez, J. L., Finlayson, C. & Bicho, N. F. 2011. Earliest known use of marine resources by Neanderthals. PLoS ONE 6(9), e24026 Coull, J. R. 1996. The Sea Fisheries of Scotland: a historical geography. Edinburgh: John Donald Craig, O. E., Steele, V. J., Fischer, A., Hartz, S., Andersen, S. H., Donohoe, P. & Heron, C. P. 2011. Ancient lipids reveal continuity in culinary practices across the transition to agriculture in Northern Europe, Proceedings of the National Academy of Sciences 108(44), 17910–17915 Crawford, G. W. & Lee, G. A. 2003. Agricultural origins in the Korean Peninsula, Antiquity 77(295), 87–95 Demarchi, B., Williams, M. G., Milner, N., Russell, N., Bailey, G. & Penkman, K. 2011. Amino acid racemization dating of marine shells: a mound of possibilities, Quaternary International 239(1), 114–124 Erlandson, J. M., 2001. The archaeology of aquatic adaptations: paradigms for a new millennium, Journal of Archaeological Research 9(4), 287–350 Estévez, J. & Vila, A. 2006. Variability in the lithic and faunal record through 10 reoccupations of a XIX century Yamana Hut, Journal of Anthropological Archaeology 25(4), 408–423 Estévez, J., Vila, A., & Pique, R. 2013. Methodological reflections on shell midden archaeology: Issues from Tierra del Fuego Ethnoarchaeology. In Bailey, G., Hardy, K. & Camara, A. (eds), Shell Energy: mollusc shells as coastal resources, 107–122. Oxford: Oxbow Books Estévez, J., Piana, E., Schiavini, A. & Juan-Muns, N. 2001. Archaeological analysis of shell middens

15. Shell middens

271

in the Beagle Channel, Tierra del Fuego Island, International Journal of Osteoarchaeology 11(1–2), 24–33 Fischer, A. 2007. Coastal fishing in Stone Age Denmark – evidence from below and above the present sea level and from human bones. In Milner, N., Bailey, G. & Craig, O. (eds), Shell Middens in Atlantic Europe, 54–69. Oxford: Oxbow Books Gaspar, M. D., DeBlasis, P., Fish, S. K. & Fish, P. R. 2008. Sambaqui (shell mound) societies of coastal Brazil. In The Handbook of South American Archaeology, 319–335. New York: Springer Guèye, M. 2010. Diversité floristique dans la lagune Joal-Fadiouth. Dakar: Edition Emmanuel Seck ENDA.753 Gumert, M. D. & Malaivijitnond, S. 2012. Marine prey processed with stone tools by burmese long-tailed macaques (Macaca fascicularis aurea) in intertidal habitats, American Journal of Physical Anthropology 149(3), 447–457 Gumert, M. D., Kluck, M. & Malaivijitnond, S. 2009. The physical characteristics and usage patterns of stone axe and pounding hammers used by long‐ailed macaques in the Andaman Sea region of Thailand, American Journal of Primatology 71(7), 594–608 Gutiérrez-Zugasti, I., Andersen, S. H., Araújo, A. C., Dupont, C., Milner, N. & Monge-Soares, A. M. 2011. Shell midden research in Atlantic Europe: State of the art, research problems and perspectives for the future, Quaternary International 239(1), 70–85 Habu, J., Matsui, A., Yamamoto, N. & Kanno, T. 2011. Shell midden archaeology in Japan: aquatic food acquisition and long-term change in the Jomon culture, Quaternary International 239(1), 19–27 Hardy, K. 2013. The shell middens of Scotland’s Inner Sound. In Bailey, G., Hardy, K. & Camara, A (eds), Shell Energy: mollusc shells as coastal resources, 123–136. Oxford: Oxbow Books Hardy, K. 2015. Variable use of coastal resources in prehistoric and historic periods in Western Scotland, Journal of Island and Coastal Archaeology. DOI: 10.1080/15564894.2015.1103336 Hardy, K. & Wickham-Jones, C. R. (eds), 2009. Mesolithic and later sites around the Inner Sound, Scotland: the Scotland’s First Settlers project 1998–2004, Scottish Archaeological Internet Reports 31. www.sair.org.uk/sair31 Hardy, K., Camara, A., Pique, R., Dioh, E., Guèye, M., Diaw Diadhiou, H., Faye, M. & Carré, M. 2016. Shellfishing and shell midden construction in the Saloum Delta, Senegal, Journal of Anthropological Archaeology 14, 19–32 Jerardino, A. & Marean, C. W. 2010. Shellfish gathering, marine palaeoecology and modern human behavior: perspectives from Cave PP13B, Pinnacle Point, South Africa, Journal of Human Evolution 59(3–4), 412–424 Klein, R. G. & Steele, T. E. 2013. Archaeological shellfish size and later human evolution in Africa, Proceedings of the National Academy of Sciences 110(27), 10910–10915 Lacaille, A. D. 1954. The Stone Age in Scotland. Oxford: University Press Light, J. & Walker, T. 2017. How strong is the evidence for purple dye extraction from the muricid gastropod Nucella laillus (L. 1758), from archaeological sites in Britain and Ireland? In Allen, M. J. (ed.), Molluscs in Archaeology, 326-341. Oxford: Oxbow Books Marean, C. W. 2010. Pinnacle Point Cave 13B (Western Cape Province, South Africa) in context: the Cape floral kingdom, shellfish, and modern human origins, Journal of Human Evolution 59(3), 425–443 Meehan, B. 1982. Shell Bed to Shell Midden. Canberra: Australian Institute of Aboriginal Studies Mellars, P. 1978. Excavation and economic analysis of Mesolithic shell middens on the island of Oronsay (Inner Hebrides). In Mellars, P. (ed.), The Early Postglacial Settlement of Northern Europe, 371–396. London: Duckworth Mellars, P. (ed.), 1987. Excavations on Oronsay: prehistoric human ecology on a small island. Edinburgh: University Press

272

Karen Hardy

Mellars, P. & Payne, S. 1971. Excavation of two Mesolithic shell middens on the Island of Oronsay (Inner Hebrides), Nature 231, 397–398 Mellars, P., Gori, K.C., Carr, M., Soares, P. A. & Richards, M. B. 2013. Genetic and archaeological perspectives on the initial modern human colonization of southern Asia, Proceedings of the National Academy of Sciences 110(26), 10699–10704 Moss, M. L. 1993. Shellfish, gender, and status on the Northwest Coast: Reconciling archeological, ethnographic, and ethnohistorical records of the Tlingit, American Anthropologist 95(3), 631–652 Orquera, L. A. & Piana, E. L. 2009. Sea nomads of the Beagle Channel in southernmost South America: over six thousand years of coastal adaptation and stability, Journal of Island and Coastal Archaeology 4(1), 61–81. doi: 10.1080/15564890902789882 Parkington, J. E., Poggenpoel, C., Buchanan, W., Robey, T., Manhire, A. & Sealy, J. 1988. Holocene coastal settlement patterns in the western Cape. In Bailey, G. N. & Parkington, J. E. (eds), The Archaeology of Prehistoric Coastlines, 22–41. Cambridge: University Press Pickard, C. & Bonsall, C. 2014. Mesolithic and Neolithic shell middens in western Scotland: a comparative analysis of shellfish exploitation patterns. In Roksandic, M., Mendonça, S., Eggers, S., Burchell, M. & Klokler, D, (eds), The Cultural Dynamics of Shell Middens and Shell Mounds: a worldwide perspective, 251–266. Albuquerque: University of New Mexico Press Piqué, R., Guèye, M., Hardy, K. & Camara, A. 2016. Not just Shellfish: Wild terrestrial resource use among the people of the Saloum Delta, Senegal. In Biagetti, S. & Lugli, F. (eds), The Intangible Elements of Culture in Ethnoarchaeological Research, 217–230. Dordrecht: Springer Pollard, T. (ed.), 2005. The excavation of four caves in the Geodha Smoo near Durness, Sutherland, Scottish Archaeological Internet Reports 18. http://www.sair.org.uk/sair18/ Richards, M. P. & Mellars, P. 1998. Stable isotopes and the seasonality of the Oronsay middens, Antiquity 72, 178–184 Richards, M. P., Schulting, R. J. & Hedges, R. E. M. 2003. Sharp shift in diet at onset of Neolithic, Nature 425, 366 Schwadron, M. 2013. Prehistoric shell landscapes of the Ten Thousand Islands, Florida. In Bailey, G., Hardy, K. & Camara, A. (eds), Shell Energy: mollusc shells as coastal resources, 43–58. Oxford: Oxbow Books Stein, J. K. 1992. Deciphering a Shell Midden. San Diego: Academic Press Tan, A., Tan, S. H., Vyas, D. Malaivijitnond, S. & Gumert, M. D. 2015. There is more than one way to crack an oyster: identifying variation in Burmese long-tailed macaque (Macaca fascicularis aurea) stone-tool use, PLoS ONE 10(5), e0124733. doi:10.1371/journal.pone.0124733 Thiam, M. 2013. Senegambian shell middens and burials – a heritage in danger. In Bailey, G., Hardy, K. & Camara, A. (eds) Shell Energy: mollusc shells as coastal resources, 191–198. Oxford: Oxbow Books Ulm, S. 2011. Coastal foragers on southern shores: marine resource use in northeast Australia since the Late Pleistocene. In Bicho, N., Haws, J. A. & Davis, L. G. (eds), Trekking the Shore. Changing Coastlines and the Antiquity of Coastal Settlement, 441–461. New York: Springer Vernet, R. 2013. L’exploitation ancienne des resources du littoral atlantique mauritanien (7500– 1000 cal. BP). In Daire, M. Y., Dupont, C., Baudry, A., Billard, C., Large, J-M., Lespz, L., Normand, E. & Scarre, C., (eds), Anciens peuplements littoraux et relations Homme/Milieu sur les côtes de l’Europe atlantique, 371–396. Oxford: British Archaeological Report S2570 Waselkov, G. A. 1987. Shellfish gathering and shell midden archaeology. In Schiffer, M. B. (ed.), Advances in Archaeological Method and Theory, 93–210. New York: Academic Press

16. The collection, processing and curation of archaeological marine shells Greg Campbell

Many archaeologists are not very sure what they should do with marine shells from archaeological deposits. Some may assume that these shells can be dismissed because they are uninformative, but this is a survival of antique archaeological thinking which died out across most of the world in the 1980s (Claassen 1998, 5–6). Other chapters in this volume (especially Somerville et al. and Winder, Chapters 13 and 14) and recent reviews (eg, Claassen 1998; Thomas 2015a; 2105b) amply demonstrate that archaeological marine shells store information that archaeologists need. The main archaeological interest is not in the shells themselves, but in what those shells can tell us about the sophistication of past peoples’ understanding of a complex part of their world (Bailey 1978); more complex than many appreciate. Archaeologists often dismiss marine shells as a minor part of the diet (eg, Osborn 1977, 177). However, humans are highly mobile omnivores, so most of our foods are minor, or major for only a season. People in mobile societies often walked many kilometres (hundreds, sometimes) through the landscape in a year, eating a huge variety of foods (including shellfish) at different times. The information needed to discern when and for how long in the year shells were part of the diet, and which parts of the shore were being selected for harvesting, survives cemented into the shells themselves (eg, Andrus 2011; Mannino et al. 2003). People in settled societies often hauled shellfish many kilometres even though they are more shell than food, and they must make the journey and be eaten within days: for these cultures, analysis of the shells is the only way to understand the efficiency of bulk transport, and the role of perishable luxuries. Since past peoples (mobile or sedentary, prehistoric or historic) did not dismiss foods because they were a minor part of their overall diet, archaeologists have no business dismissing them either. Archaeological shellfish can be numerous, often much more numerous than other remains; even single meals can produce enough shells for statistical testing and comparison. Shellfish are fragile, so their fragmentation and dissolution records the formation and post-depositional changes in the strata containing them (Stein 1992). However, this tendency to be numerous and fragile makes archaeological shellfish challenging and frustrating to excavate. Excavators, directors, artefact processors and curators can find themselves pondering a deposit, finds-bag or archive box containing shells, and wondering what best to do with them. The wondering is especially strong

274

Greg Campbell

if the deposit, finds-bag or archive box contains either trifling or colossal numbers of shells. This chapter provides some general advice about how to recover, process, and archive archaeological marine shells. It integrates and extends advice presented in greater detail elsewhere (Campbell 2008a; 2014; 2017).

Recovering shells Recovery technique Hand excavation only recovers a biased proportion of the total assemblage (Clarke 1978; Levitan 1982, 26–7; Payne 1972), consequently sieving or wet-sieving to a fairly fine mesh size (ie, 1–2 mm) will almost always be needed to recover archaeological shells that can really answer archaeological questions. Hand-retrieval of shell is only rarely useful, so excavators should always expect to sieve. A need to wet-sieve for shells applies to every type of excavation, whether large or small, commercial, academic or amateur. There are three main reasons for wet-sieving for shells: any one of them is enough to make it mandatory. First, shells, which are sometimes small; shells smaller than a thumbnail are acceptable. Secondly, virtually all archaeological shells are fragmented. Some shell is almost always crushed down to the size of a small fingernail, and even large tough shells sometimes erode to this size; sorting, identifying and counting fragments of this size is fairly painless for the specialist. Hand-retrieval recovers almost none of the fragile types of shell, even at sites where they are seen to form extensive deposits in the ground (eg, Winder 1980, 125). Further, in some deposits it is the contrast between big and small shell fragments that establishes how the deposit formed and altered (eg, Gutiérrez Zugasti 2011). Thirdly, the smaller shells that are more sensitive indicators of the habitats being exploited, and the small parts of some kinds of shell needed to establish the numbers gathered, are often about the size of a broken pencil-lead (eg, Campbell 2008b). It is simply impossible to reliably find and extract shells and fragments this small if they are in the dirt, or unwashed and still caked with dirt (Dupont 2003). It is no real surprise that fine-mesh wet-sieving (to 1 mm) is needed for marine shells. Hand-retrieval is well-known to poorly reflect what was in the ground (Dibble et al. 2005; Orton et al. 1993, 46–7; Payne 1972). Most excavators seem to accept Shaffer’s conclusions (1992) and fine-sieve at least a portion of nearly every appropriate archaeological deposit: it seems that hand-retrieval alone is predominantly limited to European excavators of historic and late prehistoric sites. In the author’s opinion, any planning-authority specifications and permissions for excavation (‘briefs’, ‘written schemes of investigation’, ‘excavation licences’; the terms vary) should specify best practice and appropriate sieving, rather than just hand-retrieval, for the recovery of archaeological shells.

Mesh size Recovery by fine-mesh wet-sieving requires guidance on appropriate size meshes. Sizes

16. The collection, processing and curation of archaeological marine shells

275

should be standardised, since the shell contents of different samples and sites can only be compared if all samples are recovered in the same way (Orton 2000, 165). Even 1 mm mesh does not recover everything identifiable, but is required for sea-urchins (Campbell 2008b). It is convenient to sieve for shells using the same mesh sizes used for artefacts and animal bones. New World archaeologists usually follow Grayson (1984, 169) and sieve samples through a stack of ¼ inch, ⅛ inch, and ⅟₁₆ inch screens, while most Old World archaeologists tend to follow Clason and Prummel (1977) and use 10 mm, 4 mm, and 1 mm mesh. Current guidance in England is 4 mm, 2 mm, and 1 mm mesh (Baker & Worley 2014, 13). However, the comparison between shells and fragments over 10 mm and those less than 10 mm is a good routine measure of fragmentation. The recommendation for extracting shells (see sieving), therefore, is to wet-sieve samples through a stack of 10 mm, 4 mm, 2 mm and 1 mm meshes.

Sampling for shells It is unfeasible (and unnecessary) to wet-sieve the bulk of a site’s deposits for shells alone. Clearly, appropriate samples will have to be taken. Deploying the appropriate sampling strategy maximises the recovery of the excavated material’s real archaeological potential while minimising effort and cost: it can prevent the spectacular waste of excavators’ time, processors’ effort, specialists’ hours, curators’ shelf-space and client’s or taxpayers’ money spent digging, washing, drying, counting, weighing, bagging, boxing, analysing and archiving too many shells, too few shells, or shells excavated in the wrong ways from the wrong deposits. If appropriate samples of shells are being taken from those deposits for analysis, it is safe to discard the remaining shells with the rest of the spoil. The best means of devising an appropriate sampling strategy is to consult the specialist responsible for analysing the shells, during the planning and costing of the project. Every region has its own tradition and background knowledge of its shells, and marine shell specialist provision, while thin, is global. If a quick consultation of the local archaeological literature or local archaeologists does not provide a good candidate, many practicing archaeomalacologists are listed by region on the ICAZ Archaeomalacology Working Group website (at time of writing: http://archaeomalacology.com/). The guidance given below is generic and general; if specific guidance from a local specialist exists, use that instead.

Sample size How big a sample must be is a complicated question (see Orton 2000 for general guidance), complicated further by ‘sample size’ meaning two different things. For the field archaeologist, ‘sample size’ means the volume of deposit that needs to be collected. For specialists, ‘sample size’ means the count in a sample of the items being investigated. Consequently the archaeologist’s sample size should be large enough to fulfil the

276

Greg Campbell

specialist’s sample size requirement. For shells, this is the number of whole shells and those fragments that can be identified as the remains of a single shell (ie, number of identifiable specimens, or NISP: Grayson 1984, 17–26). A sampling strategy is working well if most samples produce between 200 and 600 identifiable shells (Campbell 2017): • the absence of a specific shell type from a sample shows that it is a negligible part of the shells in the deposit (that it is almost certainly less than 2% of the shells in the deposit), but only if that sample contains 200 or more shells (Rolhf & Sokal 1995, table P) • percentages between 10% and 90% are usually reliable to ±4% for samples of 600 shells (van der Veen & Fieller 1982, table 4)

Routinely collecting 200–600 shells sounds a worryingly large number of shells to those who do not routinely wet-sieve, but it is not challenging: archaeological shells are often thumbnail-sized, or smaller for fragile shells (eg, Dupont 2003), so getting these numbers of shells often needs quite small volumes of deposit. Samples should be large enough to approximate to a single meal, or a single basketful of shells, discarded in the past. They also have to be of volume and weight that is practical to manoeuvre. A good initial guess (Campbell 2017) is • a sample 0.25 × 0.25 m by 50 mm deep in the ground if most shells are thumb-sized (mussels, cockles, winkles) • an in-ground sample 0.5 × 0.5 m by 0.1 m deep if most shells are hand-sized (oysters) and • a sample 1.0 × 1.0 m by 0.1 m deep of deposit being the absolute maximum, and not routine

The specialist will usually need to know the volume of space in the ground that held the shells; the volume of excavated dirt (in litres or buckets) is much less important. Since shells were usually discarded onto a surface, forming broad, thin spreads and not thick, square-ish brick-shapes, it is better to sample broad, thin spaces, not brick-like ones. The volumes these samples will take up once dug out of the deposit are hard to predict, and will vary depending on its content of clay, stones and water. Crudely, the thumb-sized shell sample guess, above, will take up less than a bucket; the hand-sized shell sample guess will need a bucket or two; the absolute maximum sample will need a dozen or more (another reason why it should not be routine). There is no ‘magic number’ of bucketfuls or litres of excavated deposit, and there is no ‘magic percent’ of a square-metre excavation unit, that makes a sample; field staff that think so must replace concern with buckets excavated with concern for shells in the ground. However, ‘there is no specifiable amount of matrix that can be deemed statistically adequate for worldwide application’ (Claassen 1991, 258). Guidance for about which deposits are worth sampling, and how to sample them, is given below. Ideally, therefore, some samples should be processed, sorted and counted during excavation to ensure most samples will provide 200–600 identifiable shells. Samples with less than 200 shells are often informative; they simply should not be routine.

16. The collection, processing and curation of archaeological marine shells

277

Intact shells Whole and nearly-whole archaeological shells are uncommon (eg, Dupont 2003), but by far the most informative (Campbell 2014). Even big tough-looking shells (oysters, scallops, conches) are fragile, and ideally need to be lifted individually and packed separately from the sample which contained them (and labelled with the correct sample number), unless the shells are present in dozens (as in some shell middens). To prevent damage to whole shells during sieving, gently dry-brush or wash with soft brushes, and air-dry with only moderate heat: surface encrustations and features are so important that some specialists will insist on cleaning the shells themselves (eg, Winder 2011; Chapter 14).

Integrity Knowing the sample volume excavated to recover the shells is critical for comparing samples (O’Connor & Barrett 2014, 274). Samples must, therefore, have nothing removed from them (no stones, gravel, ceramic artefacts, or bones) while they are being collected in the field: these are ‘whole-earth’ samples (Baker & Worley 2014, 12) or ‘solid samples’/ undisturbed samples (Bowdler 2014, 367). To prevent wet-sieving damage, remove and damp-wash whole and nearly-whole shells, labelling them with the correct sample number and context number.

Types of deposit Deposit types (as defined by Campbell 2017), and how to sample them, are described below in order from most, to least, important.

1) Discrete shell-rich scatters Heaps or lenses of shells less than 2 m across and about 20–100 mm thick, on their own or overlapping adjacent scatters. These are sometimes found on old surfaces or within extensive layers. Shells can also appear as lenses in ditches or pits, or packed together in the tops of postholes. These are probably single discard events, perhaps a single meal or a few days’ gathering, so recognising them in the field and sampling correctly is important as this provides the most direct evidence of the shore being exploited and the harvesting strategy used (Orton 2007; Parkington 2008). Collect a large enough sample to get approximately 200–600 shells (possibly 0.5 × 0.5 × 0.1 m deep), but do not sample shells from adjacent scatters. Record the in-ground sample volume (the size of the hole in the deposit that was excavated to provide the sample), for comparison with other shell-rich samples.

278

Greg Campbell

2) Homogeneous shell-rich masses Larger shell-rich spreads (2–5 m across and over 0.1 m thick) can lack any visible layering or structure, but probably resulted from several discard events, so they are likely to differ subtly with depth. Collect a series of vertically superimposed samples (columns of samples or ‘incremental samples’, like those for terrestrial molluscs: Evans 1972) where the deposit is thickest: keep the surface area constant and each depth increment thick enough that most of these samples produce 200–600 shells (probably 0.5 × 0.5 × 0.1 m deep for hand-sized shells, 0.25 × 0.25 m × 50 mm deep for thumb-sized shells). Do not sample two visibly different deposits with one increment: vary the depth of the increments near the boundary instead. Record the surface area and depth of each incremental sample, so the in-ground deposit volume can be calculated.

3) Middens Very large shell-rich accumulations (over 5 m across and 0.1 m thick). Exactly how big a heap of archaeological shells must be to be called a ‘midden’ varies between regions and researchers; some use volume (Dupont 2003); some do not like using size (Claassen 1998, 11–12 explains why). Some archaeologists assume (or perhaps hope) that middens are formless heaps of re-deposited shell; they are, however, rich and varied repositories of well-preserved cultural remains (especially of the changing relationship between a culture and the sea) that archaeologists have been investigating for over a century, and from all over the world (Balbo et al. 2011; Guttiérez Zugasti et al. 2016; Hardy, Chapter 15). Sampling middens is a rich, subtle skill with many regional traditions (Claassen 1998, 99–104 & Stein 1992 remain essential reading). Some prioritise the changing relationship with the sea over time, and sample mainly by incremental columns (Peacock 1978); others prioritise midden formation history via changing shell preservation (Stein 1992; Stein et al. 2003), and sample mainly with equal volumes from all deposits; some give equal weight to both, and combine both sampling regimes (Bowdler 2014, 368–9). As an absolute minimum, collect incremental columns through the deepest part of the midden (or the longest depositional sequence), and every 3.0 m along the maximum length of the midden (Campbell 2017). Collect columns of samples as for homogeneous shell-rich masses (deposit type 2): keep the surface area constant and each depth increment thick enough that most of these samples produce 200–600 shells (probably 0.5 × 0.5 × 0.1 m deep for hand-sized shells, and 0.25 × 0.25 m × 50 mm deep for thumb-sized shells), but vary the depth increments so they do not cross deposit boundaries, and record each increment’s surface area and depth. This advice applies only when shell middens are unexpectedly encountered during excavation. If the site is known to include a midden, specialist advice while planning and costing an excavation is required, more for than many other deposits, to fulfil its full archaeological potential.

4) Unusual species In most regions, most archaeological shells are from a small number of familiar kinds (in

16. The collection, processing and curation of archaeological marine shells

279

Atlantic Europe, oysters, mussels, cockles and periwinkles). Occasionally, an unfamiliar kind of shell is the most common in a deposit; in southern Britain unusual species are razor-shells, whelks, dog-whelks, limpets, sea-urchins, and kinds that haven’t been spotted yet. If unusual shells form a discrete scatter (deposit type 1) or a homogeneous mass (deposit type 2), sample them as explained above. Otherwise, collect 20–50 litres of excavated deposit (ie, enough to get approximately 200–600 shells); in-ground volume is not required.

5) Shell-bearing deposits from shell-sparse periods In most regions there are cultures or periods in which archaeological shellfish are rare, which can make them especially informative. In Atlantic Europe, marine shells are rare from prehistoric and ‘Dark-Age’ deposits and sites. Sample discrete scatters, large masses, or middens of these periods as explained for deposit types 1, 2, and 3. Most deposits, however, have few shells (one or two visible per litre: a quick trowelling through a bucket of spoil finds ten shells or fewer). For this kind of deposit, collect 40–100 litres of excavated context-specific deposit (to get approximately 200–600 shells). Retain any other shell collected by hand.

6) Shell-bearing deposits from shell-rich periods In most regions, some periods or cultures transported, consumed and discarded considerable numbers of shellfish (eg, Roman, medieval and post-medieval Atlantic Europe). Sample discrete scatters, large masses, or middens of these periods as explained for deposit types 1, 2, and 3. However, many deposits are not visibly rich in shells, but have more than one obvious shell per litre: a quick trowelling through a bucket of spoil finds more than five shells but less than 20; finds-trays often contain numerous shells. Despite shell specialists’ fascination with middens, most field and curatorial archaeologists would recognise that most archaeological shells come from this type of deposit. Collect and retain a bulk sample of 20–50 litres excavated volume (enough to recover approximately 200–600 shells). Do not recover shells solely by hand.

7) Shell-poor deposits from shell-rich periods These have less than one obvious shell per litre; a quick trowelling through a bucket of spoil finds only ten shells or fewer. Since shells are common in these periods, shells in shell-poor deposits are likely intrusive or residual. No sampling specifically for shells is required.

280

Greg Campbell

Extraction Sieving Extracting shells from their various types of deposit is outlined in Figures 16.1 and 16.2. Air-dry and record the weight of every sample from shell-rich deposits (scatters, masses, and middens: deposit types 1, 2, and 3): air-dry weight is often used to understand how the deposit was formed and transformed (taphonomy). Wet-sieve samples through stacks of 10 mm, 4 mm, 2 mm and 1 mm mesh and air-dry the resulting >10 mm, 10–4 mm, 4–2 mm and 2–1 mm residue fractions. The 2 mm mesh can optionally be omitted: air-dry the resulting fine fraction (4-1 mm) and split it into 4–2 mm and 2–1 mm fractions by dry-sieving through 2 mm mesh. Processing clay-rich deposits can badly damage shells during sieving, but this is minimised by alkaline solutions (Berglund & Ralska-Jasiewiczowa 1986, 456), or other deflocculants. Soak clayey deposits overnight with a half-cup of washing soda (commercial-grade sodium carbonate, which is cheap to buy in bulk and safe to store), and then wet-sieve (wearing gloves and waterproof clothing) (Campbell 2017).

Sorting The quality of the sorting is crucial. The sorters’ training and experience is critical, and needs assessment by the specialist: the sorters should feel free to ask the specialist for identifications, and to retain any difficult objects for specialist identification. Many specialists insist on sorting themselves (Casteel 1976). The sorting and its recording differs between shell-rich deposits (scatters, masses and middens: deposit types 1, 2, and 3) and less rich deposits (types 4, 5, and 6). Shell-rich deposits (deposit types 1–3) need more complicated sorting (Fig. 16.1), because the aim is to understand both how the deposit was formed and transformed (taphonomy) as well as how shells were used in the past (shell utilisation). Taphonomy is reconstructed using the relative proportions by weight of all the constituents (not just the shells) in all the sample fractions. The proportions used are either; 1. the more useful weight per litre in the ground (which is why the sizes of the samples in the ground needed recording during fieldwork), or 2. weight-percent of total air-dried weight of the sample (which is why the total sample was air-dried and weighed before sieving).

Air-dry and record the weight of each sample fraction. For the >10 mm and 10–4 mm residue fractions, sort out all constituents (stones, bones, artefacts, charcoal, shellfragments, and whole shells, including any intact shells that were extracted and cleaned before sample sieving) and record the weight of each constituent. Discard the stones, and label and bag the other constituents. Sort the 4–2 mm fraction for five constituents: (1) identifiable bones (probably rodents and fish) and teeth; (2) artefacts other than ceramics; (3) charcoal; (4) identifiable elements of shells (whole shells; umbones of bivalves; apices, apertures and columella

16. The collection, processing and curation of archaeological marine shells

281

Figure 16.1. The complex shell extraction process for samples from discrete scatters, large masses, or middens (shell-rich deposits: types 1–3) bases of gastropods; encrusting organisms; chiton plates; sea urchin fragments); (5) oddities, that are difficult to identify (not just shells). Record the weight of the first four constituents; bag and label all five. Discard the stones and shell-fragments (the ‘shell hash’) unless the project design or specialist requires them to be kept.

282

Greg Campbell

Figure 16.2. The simpler shell extraction process for samples from shell-bearing deposits worth sampling (deposit types 4–6)

Scan the 2–1mm fraction (usually by the specialist): if identifiable elements (constituent 4) are present in some numbers, or if the larger fractions had species with small identifiable elements (sea-urchins, for example), sort the 2–1 mm for those elements, and bag and label them. If sorting is not needed, discard the 2–1 mm fraction. If a 10–4 mm, 4–2 mm or 2–1 mm fraction is big (over 5 litres), halve the fraction using a sample-splitter, or even quarter it, and sort one half (or one quarter). Record the fraction has been halved or quartered, and discard the unsorted residue fraction (unless the project design or specialist says to keep them). Bulk samples (deposit types 4–6) need simpler sorting (Fig. 16.2), because the aim is only the understanding of how shells were used in the past (shell utilisation), not how the deposit was formed and transformed (taphonomy). The air-dried weights of the sample, its residue fractions and their constituents are not needed. Sort the air-dried >10 mm, 10–4 mm and 4–2 mm fractions for identifiable elements of shells (whole shells; umbones of larger bivalves; apices, apertures and columella bases of larger gastropods; encrusting organisms; chiton plates; fragments of sea-urchin), bones, and charcoal and artefacts (but not ceramic artefacts from the 10–4 and 4–2 mm fractions). Discard the stones and shell-hash. Scan the 2–1 mm fraction (usually by the specialist): sort it for identifiable shell elements if they are present in numbers, or if the larger fractions showed species with small identifiable elements; otherwise discard it. Halve or quarter the 10–4 mm, 4–2

16. The collection, processing and curation of archaeological marine shells

283

mm and 2–1 mm fractions (and record this) before sorting if they are big: discard the un-sorted portions unless the project design or specialist require them to be kept.

Shellfish biology and archaeological analysis While analysing archaeological marine shells is outlined by Somerville et al. (Chapter 13), there are three basic principles of shellfish biology that make analysis more complex than generally appreciated: Allometry: every shellfish species grows allometrically (Seed 1980), changing shape with size (as all animals must); the relationship between shape and size is usually rendered linear by log-transformation (Gould 1966; Huxley 1932; Teissier 1948). Therefore proportions (the ratios of any two dimensions) are not constant, they are intimately wedded with size: average proportions will be have to be different whenever average size is different. There are no ‘magic numbers’ for multiplying a surviving dimension to provide original shell size or meat weight. Ecophenotypic plasticity: each shellfish species alters its size-shape allometry (not just its shape) to survive the hydrodynamics of their immediate surroundings (Seed 1980), and the change in viscosity-effects with size (Vogel 1994). A small shell from one spot can have the same proportions as a large shell of the same species from another spot on the same shore (yet another reason why there are no ‘magic numbers’). Patchiness: a shellfish species is not spread uniformly across a shore: it is broken up into a mosaic of small patches, each patch at a different stage in a repeating cycle of boom, bust and recovery (eg, Hawkins & Jones 1992 for gastropods; Seed & Suchanek 1992 for epifaunal bivalves; Ducrotoy et al. 1991 for infaunal bivalves). Immediately adjacent patches on the same shore will have different average sizes, so differences in average size in archaeological shellfish usually reflect slight differences in locale harvested, not intensity of exploitation (eg, Thakar et al. 2017, contra Mannino & Thomas 2002). Exploitation intensity is better reflected by co-variation in the maximum size and the smallest acceptable size (the biggest and smallest shell in a sample). Further, medium- and long-term variation in size, shape and growth rate ‘can also result from other sorts of environmental changes not involving human activity, posing major issues of disentangling human and nonhuman effects’ (Bailey et al. 2008). While there are underlying trends in size gradients (Vermeij 1972) and allometry (eg, Baxter 1983 for gastropods; Seed 1980 for bivalves) which can help with the relative positions of the patches exploited (eg, Cabral & da Silva 2003), the detailed research required to assign archaeological shellfish to quite specific points on the shore is just beginning. Campbell (2014) provides an example for developing accurate predictions in the face of allometry, plasticity and patchiness.

284

Greg Campbell

Curating archaeological shell Storage Marine shells are quite stable once washed or wet-sieved (which removes most of the soluble salts) and air-dried (which reduces water-mediated chemical degradation), requiring archive conditions much like archaeological ceramics: in stable conditions, between 16–21°C and 50–55% relative humidity, and labelled, bagged, and boxed using archive-quality materials (Sturm 2006). The main risk is mechanical damage during handling and moving, so do not over-fill bags with shells, do not over-fill boxes with bags, and do not stack boxes too high (but do stack them in some obvious order). Properly-stored shells sometimes suffer ‘Byne’s disease’ (Tennant & Baird 1985), gradually degrading to powder because of slowly-released volatile acids, sometimes from some paints, adhesives (eg, in some plywood) or building insulations, but usually from the wood used for the shelving or cabinets (especially oak). Storage must be made from non-volatile materials (acid-free boxes, and spruce, beech or powder-coated metal shelves). Individual shells sometimes need consolidation, but most consolidating agents (commercial shellacs and varnishes, paraffin/xylene) use solvents that dissolve or degrade the organic chemical component of the shell, or acidic or acid-releasing resins that degrade the calcium carbonate minerals in the shell. Check with the archiving museum’s conservation staff for current best practice: at the moment the best method is an aqueous solution of polyvinyl acetate (PVA) (‘white glue’), followed by thorough air-drying (Carter 2000; Sturm 2006).

Retention Archives are full or filling up fast, not just in England (Edwards 2013) but around the world (Childs 2006), so what is sent to archive, and retained in them, must be reviewed to ensure archives are ‘avoiding replication, repetition or the retention of materials not germane to future analysis’ (Brown 2011, 23). Hand-recovered archaeological marine shell (the bulk in British archives, and probably elsewhere) is highly biased. Retain hand-retrieved shells only if they are: a: from shell-sparse periods (deposit type 5); b: artefacts, or shells used to produce artefacts; c: ritual or votive offerings (burials, foundation deposits, or ‘structured’ deposits); d: from sites designated for preservation by statute (eg, Scheduled Ancient Monuments, National Historic Places, World Heritage Sites); e: numerous enough in a deposit to be statistically comparable, but not so numerous to be considerably time-averaged (between 200–1000 whole or nearly-whole shells).

Retain shells recovered by sieving: they are much more representative of what was present in the site, even if coarse mesh was used, or few shells were recovered. Following these fairly simple rules might retain a few uninformative shells, but it avoids involving a specialist in the detailed assessment and recording needed for de-accessioning animal bones (Baker & Worley 2014, 24; Rainsford et al. 2016).

16. The collection, processing and curation of archaeological marine shells

285

Conclusions Archaeological marine shells preserve much information about how past peoples lived, but require strong guidance to maximise the value for analysis and curation. Shells forming discrete scatters, large homogeneous masses, and middens require samples of known in-ground volume (often using incremental columns, sometimes at several points), processed to understand both how they formed and the shells harvested to form them. Deposits containing shells, if the shells are unusual, or from periods where shells were rarely exploited, or are fairly rich in shells from periods where shells were regularly exploited, require samples for the shells harvested. Samples designed to gather 200–600 identifiable specimens, wet-sieved to 1mm but keeping the intact shells intact, will produce assemblages worthy of analysis and permanent curation, and could fit easily with sampling for other materials (such as water-flotation for charred plant remains). Archives should retain shell artefacts and production waste, sieved shells, shell artefacts, and hand-recovered shells from shell-sparse periods, from ritual offerings, from designated sites, and from any deposit which has produced 200–1000 shells.

Acknowledgements This chapter is based on previous work strongly supported by Dr Antonieta Jerardino while at ICREA/Pompeu Fabra University, Barcelona (Campbell 2017), which itself grew from a guidance note for archaeological shells for the Sussex Museum Group commissioned and supported by Dr Robert Symmonds, curator at Fishbourne Roman Palace.

References Andrus, C. F. T. 2011. Shell midden sclerochronology, Quaternary Science Reviews 30(21–22), 2892–2905 Bailey, G. N. 1978. Shell middens as indicators of postglacial economies: a territorial perspective. In Mellars, P. (ed.), The Early Postglacial Settlement of Northern Europe, 37–63. London: Duckworth Bailey, G., Barrett, J., Craig, O. & Milner, N. 2008. Historical ecology of the North Sea Basin. In Rick, T. C. & Erlandson, J. M. (eds), Human Impacts on Ancient Marine Ecosystems: a global perspective, 215–242. Berkeley: University of California Press Baker, P. & Worley, F. 2014. Animal Bones and Archaeology: guidelines for best practice. Swindon: English Heritage Balbo, A., Madella, M., Godino, I. B. & Álvarez, I. M. (eds), 2011. Shell midden research: an interdisciplinary agenda for the Quaternary and social sciences, Quaternary International 239(1–2), 1–170 Baxter, J. M. 1983. Allometric relationships of Patella vulgata L. shell characters at three adjacent sites at Sandwick Bay in Orkney, Journal of Natural History 17, 743–755 Berglund, B. E. & Ralska-Jasiewiczowa, M. 1986. Pollen analysis and pollen diagrams. In Berglund, B. E. (ed.), Handbook of Holocene Palaeoecology and Palaeohydrology, 455–484. Chichester: Wiley

286

Greg Campbell

Bowdler, S. 2014. Shell middens and mollusks. In Balme, J. & Paterson, A. (eds), Archaeology in Practice: a student guide to archaeological analyses, 361–384. Chichester: Wiley Brown, D. H. 2011. Archaeological Archives: a guide to best practice in creation, compilation, transfer and curation. Second edition. Reading: Institute for Archaeologists for Archaeological Archives Forum Cabral, J. P. & da Silva, A. C. F. 2003. Morphometric analysis of limpets from an Iron-Age shell midden found in northwest Portugal, Journal of Archaeological Science 30(7), 817–829 Campbell, G. 2008a. Beyond means to meaning: using distributions of shell shapes to reconstruct past collecting strategies, Environmental Archaeology 13(2), 111–121 Campbell, G. E. 2008b. Sorry, wrong Phylum: a neophyte archaeomalacologist’s experiences in analyzing a European Atlantic sea-urchin assemblage, Archaeofauna 17, 77–0 Campbell, G. E. 2014. Size prediction in archaeomalacology: the common mussel, Mytilus edulis L., as an example, Archaeological and Anthropological Science 6, 305–318 Campbell, G. 2017. “What do I do with all these shells?” Basic guidance for the recovery, processing and retention of archaeological marine shells, Quaternary International 427(part A), 13–30 Carter, J. 2000. The conservation of molluscan collections, Biology Curator 18, 17–20 Casteel, R. W. 1976. Comparison of column and whole unit samples for recovering fish remains, World Archaeology 8(2), 192–196 Childs, S. T. 2006. Archaeological collections: valuing and managing an emerging frontier. In Agnew, N. & Bridgland, J. (eds), Of the Past, for the Future: integrating archaeology and conservation, 204–210. Los Angeles: Getty Conservation Institute Claassen, C. W. 1991. Normative thinking and shell-bearing sites, Archaeological Method and Theory 3, 249–298 Claassen, C. 1998. Shells. Cambridge: University Press Clarke, D. V. 1978. Excavation and volunteers: a cautionary tale, World Archaeology 10(1), 63–70 Clason, A. T. & Prummel, W. 1977. Collecting, sieving and archaeozoological research, Journal of Archaeological Science 4(2), 171–175 Dibble, H. L., Raczek, T. P. & McPherron, S. P. 2005. Excavator bias at the site of Pech de l’Az- IV, France, Journal of Field Archaeology 30, 317–328 Ducrotoy, J. P., Rybarczyk, H., Souprayen, J., Bachalet, G., Beukema, J. J., Desprez, M., Dorjes, J., Essink, K., Guillou, J., Michaelis, H., Sylvand, B., Wilson, J. G., Elkaim, B. & Ibanez, F. 1991. A comparison of the population dynamics of the cockle (Cerastoderma edule L.) in northwestern Europe. In Elliott, M. & Ducrotoy, J. P. (eds), Estuaries and Coasts: spatial and temporal intercomparisons, 173–184. Fredensborg: Olsen and Olsen Dupont. C. 2003. Les coquillages alimentaires des dépôts et amas coquilliers du Mésolithique récent/final de la façade atlantique de la France, Préhistoires Méditerranéennes 12, 221–238 Edwards, R. 2013. Archaeological Archives and Museums 2012. Colchester: Society of Museum Archaeologists Evans, J. G. 1972. Land Snails in Archaeology. London: Seminar Press Gould, S. J. 1966. Allometry and size in ontogeny and phylogeny, Biological Reviews 41, 587–640 Grayson, D. K. 1984. Quantitative Zooarchaeology: topics in the analysis of archaeological faunas. London: Academic Press Gutiérrez Zugasti, F. I. 2011. Shell fragmentation as a tool for quantification and identification of taphonomic processes in archaeomalacological analysis: the case of the Cantabrian region (northern Spain), Archaeometry 53(3), 614–630 Gutiérrez Zugasti, I., Cuenca Solana, D., Colonese, A. C. & Fernández-López de Pablo, J. (eds),

16. The collection, processing and curation of archaeological marine shells

287

2016. Time for the tide: New perspectives on hunter-fisher-gatherer exploitation on intertidal resources in Atlantic Europe and Mediterranean regions, Quaternary International 407(B), 1–154 Hardy, K. 2017. Shell middens. In Allen, M. J. (ed.), Molluscs in Archaeology, 259–272. Oxford: Oxbow Books Hawkins, S. J. & Jones, H. D. 1992. Rocky Shores. Marine Field Course Guide 1. London: Immel Huxley, J. S. 1932. Problems of Relative Growth. London: Methuen Levitan, B. 1982. Excavations at West Hill, Uley: the sieving and sampling programme. Western Archaeological Trust Occasional Paper 10. Bristol: Western Archaeological Trust Mannino, M. A. & Thomas, K. D. 2002. Depletion of a resource? The impact of prehistoric human foraging on intertidal mollusc communities and its significance for human settlement, mobility and dispersal, World Archaeology 33, 452–474 Mannino, M. A., Spiro, B. F. & Thomas, K. D. 2003. Sampling shells for seasonality: oxygen isotope analysis on shell carbonates of the inter-tidal gastropod Monodonta lineata (da Costa) from populations across its modern range and from a Mesolithic site in southern Britain, Journal of Archaeological Science 30(6), 667–679 O’Connor, T. & Barrett, J. 2014. Animal bones. In Balme, J. & Paterson, A. (eds), Archaeology in Practice: a student guide to archaeological analysis, 264–299. Chichester: Wiley Orton. C. 2000. Sampling in Archaeology. Cambridge: University Press Orton, J. 2007. The sampling of ephemeral shell scatters in Namaqualand, South Africa, South African Archaeological Bulletin 62, 74–78 Orton, C., Tyers, P. & Vince, A. 1993. Pottery in Archaeology. Cambridge: University Press Osborn, A. 1977. Strandloopers, mermaids, and other fairy tales: ecological determinants of marine resource utilization – the Peruvian case. In Binford, L. R. (ed.), For Theory Building in Archaeology, 157–205. New York: Academic Press Parkington, J. 2008. Limpet sizes in Stone Age archaeological contexts at the Cape, South Africa: changing environment or human impact? In Antczak, A. & Cipriani, R. (eds), Early Human Impacts on Megamolluscs, 169–178. British Archaeological Report S1865. Oxford: Archaeopress Payne, S. 1972. Partial recovery and sample bias: the results of some sieving experiments. In Higgs, E. S. (ed.), Papers in Economic Prehistory, 49–62. Cambridge: University Press Peacock, W. 1978. Probabilistic sampling in shell middens: a case study from Oronsay, Inner Hebrides. In Cherry, J. F., Gamble, C. & Shennan, S. (eds), Sampling in Contemporary British Archaeology, 177–190. Oxford: British Archaeological Report 50 Rainsford, C., O’Connor, T. & Connelly, P. 2016. The embarrassment of riches: rationalising faunal assemblages from large urban sites, International Journal of Osteoarchaeology 26, 221–231 Rohlf, F. J. & Sokal, R. R. 1995. Statistical Tables. New York: W. H. Freeman Seed, R. 1980. Shell growth and form in the Bivalvia. In Rhoads, D. C. & Lutz, R. A. (eds), Skeletal Growth of Aquatic Organisms: biological records of environmental change, 23–67. London: Plenum Press Seed, R. & Suchaneck, T. H. 1992. Population and community ecology of Mytilus. In Gosling, E. (ed.), The Mussel Mytilus: ecology, physiology, genetics and culture, 87–171. Developments in Aquaculture and Fisheries Science 25. Amsterdam: Elsevier Shaffer, B. S. 1992. Quarter-inch screening: understanding biases in recovery of vertebrate faunal remains, American Antiquity 57(1), 129–136 Somerville, E., Light, J. & Allen, M. J. 2017. Marine molluscs from archaeological contexts: how they can inform interpretations of former economies and environments. In Allen, M. J. (ed.), Molluscs in Archaeology, 214–257. Oxford: Oxbow Books Stein, J. 1992. Deciphering a Shell Midden. San Diego: Academic Press

288

Greg Campbell

Stein, J. K., Deo, J. N. & Phillips, L. S. 2003. Big sites – short time: accumulation rates in archaeological sites, Journal of Archaeological Science 30(3), 297–316 Sturm, C. F. 2006. Chapter 5: Archival and curatorial methods. In Sturm, C. F., Pearce, C. A. & Valdés, A. (eds), The Mollusks: a guide to their study, collection and preservation, 45–57. Boca Raton: Universal Publishers Teissier, G. 1948. La relation d’allometrie: sa signification statistique et biologique, Biometrics 4(1), 14–53 Tennant, N. H. & Baird, T. 1985. The deterioration of Mollusca collections: identification of shell efflorescence, Studies in Conservation 30, 73–85 Thakar, H. B., Glassow, M. A. & Blanchette, C. 2017. Reconsidering evidence of human impacts: Implications of within-site variation of growth rates in Mytilus californianus along tidal gradients, Quaternary International 427(part A), 151–159 Thomas, K. D. 2015a. Molluscs emergent, Part I: themes and trends in the scientific investigation of mollusc shells as resources for archaeological research, Journal of Archaeological Science 56, 133–140 Thomas, K. D. 2015b. Molluscs emergent, Part II: themes and trends in the scientific investigation of molluscs and their shells as past human resources, Journal of Archaeological Science 56, 159–167 Veen, M. van der, & Fieller, N. 1982. Sampling seeds, Journal of Archaeological Science 9(3), 287–298 Vermeij, G.J. 1972. Intraspecific shore-level size gradients in intertidal molluscs, Ecology 53(4), 693–700 Vogel, S. 1994. Life in Moving Fluids: the physical biology of flow. Princeton: University Press Winder, J. M. 1980. The marine Mollusca. In Holdsworth, P., Excavations at Melbourne Street, Southampton 1971–76, 121–127. London: Council for British Archaeology Research Report 33 Winder, J. M. 2011. Oyster Shells from Archaeological Sites: a brief illustrated guide to basic processing. Online publication at: https://oystersetcetera.files.wordpress.com/2011/03/ oystershellmethodsmanualversion11.pdf Winder, J. 2017. Oysters in archaeology, 238–258. In Allen, M. J. (ed.), Molluscs in Archaeology. Oxford: Oxbow Books

Part 4 Artefacts

17. Shell ornaments, icons and other artefacts from the eastern Mediterranean and Levant Janet Ridout-Sharpe

Molluscan artefacts are shells that have acquired cultural significance by modification to fashion ornaments, tools, containers and other items, and also unmodified shells determined by their context to have had significance. The study of molluscs as artefacts has been eclipsed in recent years by a switch of emphasis towards molluscs as palaeoenvironmental indicators and tools for the reconstruction of ancient economies. Nevertheless, we can still learn a great deal about the archaeology of human behaviour, technological development and social interactions by the careful study of shell artefacts. This chapter sets out to provide guidelines for recording and analysing molluscan artefacts from archaeological sites in countries bordering the eastern Mediterranean and, using previously published site reports as examples, to show how the results may be used to interpret aspects of social archaeoanthropology. The study area extends east from peninsular Greece to the Levant, encompassing the Aegean islands, Crete and Cyprus, southwestern Anatolia (Turkey) and the Middle Eastern countries of Syria, Lebanon, Israel and Jordan, with occasional excursions outside this area following ancient cultural boundaries rather than modern political ones. Timewise the sites presented in this brief and necessarily selective review range from the late Palaeolithic through the Epipalaeolithic (Mesolithic), Neolithic, Bronze and Iron Ages and beyond. Most molluscan material relates to the earlier periods and after the Bronze Age the use of shell in the eastern Mediterranean and Levant became less important.

Historical background In the early days of archaeological excavation, the main purpose was the recovery of ancient works of art to furnish the museums of Europe and America that funded the work. Little or no attention was paid to shells unless they were obvious components of necklaces or cosmetic containers and the like, and much potentially valuable information was destroyed in a process that was little more than authorised looting. Perhaps the first archaeologist working in this area to become aware of the sig­ nificance of shells was Dame Kathleen Kenyon (1906–1978) who conducted excavations at Jericho and Jerusalem in the mid-20th century. Her work at Jericho in 1952–1958 was

17. Shell ornaments, icons and other artefacts

291

the first to establish the existence of the Late Epipalaeolithic (Natufian) and Pre-Pottery Neolithic (PPNA and PPNB) in the Levant – and she was possibly the first to entrust the examination of shells to an experienced conchologist, the Revd H. E. J. Biggs (1895–1973). Jericho is famous for the discovery of PPNB human skulls plastered to show facial features with cowries inserted in the eye sockets. Biggs (1963) acknowledged the unusual thoroughness of Kenyon’s workmen in recovering a relatively large number of terrestrial, freshwater and marine molluscs. Considering the latter, he pondered the question, ‘For what purpose was this or that shell brought from the shore?’ and he offered ‘some tentative suggestions’. He identified seven species of marine bivalves, 20 species of gastropods and a single scaphopod; both Mediterranean and Red Sea species were represented. The dog cockle Glycymeris violascens (now nummaria) was the most frequent with 320 valves and most of these had been artificially perforated by grinding the umbo. Because of their frequency, he suggested that the circular Glycymeris shells symbolised the moon deity. Of the small dove shell Columbella rustica, he wrote that it ‘is worn by women in Greece today as a necklace and love charm’ and suggested that it held the same function in PPNB Jericho. He also commented on the resemblance of the inturned slit-like aperture of Columbella to the female genitalia and wondered if the scaphopod shell provided the male counterpart. In a subsequent paper, Biggs (1969) considered the non-dietary role of molluscs as raw material for beads and pendants and their possible symbolic or ritual function, although in this respect he emphasised the need for ‘great care and much caution’. He drew attention to the frequent lack of ‘associated precise field data’ with excavated material and its importance for interpretation. In this way the Revd Biggs was ahead of his time and a pioneer in the modern discipline of archaeomalacology. The molluscs recovered from Kenyon’s excavations in Jerusalem in 1961–1967 were not examined or published until very much later (Reese 1995; 2008). The shells were described at length but their interpretation was cautious in the absence of detailed contextual data, resulting in little more than a catalogue. Such catalogues have been the norm for the presentation of shell material in published site reports, often in the form of appendices. These provide a rich source of material for comparing different sites and periods but a major difficulty remains the absence of ‘precise field data’. Advances in the taxonomy and systematics of molluscs, which have accelerated in recent decades with the introduction of DNA analysis, have resulted in changes in the nomenclature of many species, sometimes several times over, since the first catalogues were published and provide pitfalls for the unwary. More important, however, are misidentifications, especially those leading to misinterpretation. For example, Mienis (2007) has cited the inclusion of a North American species among Levantine material: ‘So who actually discovered America?’

Recording and analysis A shell assemblage will invariably contain numerous small, apparently meaningless fragments and unmodified shells and it is advisable to follow a set procedure for

292

Janet Ridout-Sharpe

recording and analysis which can be divided into a number of tasks. Much can be achieved with minimal equipment, but whereas shell artefact analysis may be cheap in terms of apparatus it can be expensive in terms of labour. Analytical procedures have changed over time, making it difficult to compare different assemblages (Claassen 1998, 120) and there is a need for a consistent methodology in order to extract the maximum amount of information from the shells (Karali 1999, 8).

Collecting the material Ideally the archaeomalacologist will be part of the team excavating the site and will supervise the collection and recording of material in the field and ensure that any special deposits of shells are photographed or drawn in situ (Karali 1999, 5). In practice, however, this is rarely the case and the material will have been excavated, bagged and labelled by others, awaiting post-excavation analysis. In older excavations and even today in rescue operations where lack of time or funding cannot support extractive techniques, the material may simply have been collected by hand. This immediately imposes a bias in that only the larger and more obvious pieces will be retrieved, and sampling will be limited to the area of the trench. Sieving will recover a much larger sample which will increase with the volume of soil sieved and the fineness of the sieve mesh. Mesh sizes of 1–10 mm are commonly used, with the finer meshes ( 3) Sample size required Sampling considerations

Reservoir correction Ways to correct

Marine Usually aragonite, some calcitic species 1

Freshwater Mostly aragonite 2

Terrestrial Aragonite 3

Min. 20 mg/max. 200 mg of shell (depending on preservation). Single entity (ie, shell) preferred Feeding practices (suspension vs Water mixing sources Feeding practices detritus feeders) Local geology Local geology Coastal upwelling Old-shell effect Burrowing Old-shell effect Diagenesis (secondary deposits) Diagenesis Diagenesis Global R(t) and local ΔR offset Hard water effect Limestone effect Online database (eg, http://calib.org/marine) AMS measurement of pre-1940s shells Comparison of archaeological charcoal-shell pairs

AMS measurement of modern specimens to calculate 14C uptake

AMS measurement of modern specimens to calculate 14C uptake

this can lead to serious interpretational issues. Providing a guide for archaeologists for submission is, however, not straight forward. Shells vary tremendously between marine and terrestrial, between species, and degrees of preservation. They also vary locally and sometimes they adopt different diets based on bioavailability. That is why there is a need to check the local geology/Carbon uptake with modern freshwater and terrestrial specimens.  Nevertheless, a crude guide is that single specimens are best and ideally a minimum of 20 mg and maximum of 200 mg of shell, depending on preservation (see Table 22.1).

Laboratory preparation of shell for AMS dating Sample purity is paramount in the proper application of radiocarbon dating. This, however, can be fiddly when dealing with archaeological material that is often heterogeneous and altered from use and burial in a variety of depositional environments for various lengths of time. Laboratory pretreatment protocols are specifically designed to target exogenous contaminants, reduce and eliminate them, either physically and/ or chemically. Each treatment starts with mechanical cleaning of the sample and is followed by chemical purification of the extracted carbonaceous components and the removal, at a molecular level, of any intrusive contaminants. In the past years, revised methods for the screening and decontamination of shell

388

Katerina Douka

carbonates have been published (Douka et al. 2010; Russo et al. 2010). These involve the screening, using high-precision XRD and FTIR methodologies to establish shell mineralogy and the absence of diagenetic alterations, and the physical removal of any recrystallised material using density separation of different polymorphs (CarDS). If such methodologies are not available, most laboratories either perform an initial chemical leaching to remove surface residues and/or the Feigl’s staining protocol as a rough way of separating aragonite from (presumably secondary) calcite. Subsequently the sample is converted into carbon dioxide (CO2) by reacting few milligrams of powdered shell with pure ortho-phosphoric acid (H3PO4). Organically-derived CO2 from the shell proteins does not contribute to the final AMS measurement. Approximately 8–10 mg of shell powder will give 1 mg of pure C (12% C content), the average size of an AMS radiocarbon target. Hence, it is possible with the use of AMS to date minute molluscs, other carbonate organisms such as foraminifera, or archaeological shell artefacts with minimal destruction.

Calibration of radiocarbon results Because the production of 14C in the upper atmosphere fluctuates, calculated radio­ carbon years are not the same timescale as conventional calendar years. The conversion of one to the other was first performed by direct radiocarbon dating of tree-rings that were also independently and precisely dated by dendrochronology to a single year (Stuiver & Suess 1966). The first internationally agreed calibration curve was published in 1986 (Stuiver et al. 1986; Pearson & Stuiver 1986) and spanned to the 8th millennium BP. Revised and enhanced versions followed in the next decades (Stuiver & Braziunas 1993; Stuiver et al. 1998; Reimer et al. 2004; 2009), the latter spanning back to 50 ka cal BP. The most recent version, IntCal13 (Reimer et al. 2013) is based upon annually banded biotic records from terrestrial archives for the younger part, notably tree-rings from present to ~13,900 years BP, supplemented by the Lake Suigetsu macrofossil data from 13.9–50 ka cal BP, as well as measurements from speleothems and marine coral archives. With regards to the calibration of marine samples, the IntCal group also constructed the Marine13 curve, a ‘general’ marine calibration curve that assumes constant reservoir (Reimer et al. 2013). However, the variability of the marine reservoir in temporal and spatial terms when compared with the terrestrial and atmospheric 14C records makes a globally applicable marine calibration curve prior to the Holocene questionable (see below in ‘Marine reservoir effects’).

Technical considerations and limitations in the radiocarbon dating of shell Several factors may influence radiocarbon results of molluscan shell, causing errors towards younger or older ages and these are discussed below.

22. Radiocarbon dating of marine and terrestrial shell

389

Marine reservoir effects One of the erroneous primary assumptions in the early days of radiocarbon dating was that the concentration of 14C within an organism was in equilibrium with that of the atmosphere. The dilution of different C pools with respect to the contemporary atmosphere is called a reservoir effect. Complex circulation patterns and variable rates of mixing mean that water may reside in the deep ocean for centuries, becoming ‘old’, ie, highly 14C-depleted. This ‘old’ deep water eventually mixes with modern surficial water, and so the oceanic carbon signal and that of marine organisms living in the ocean, is isotopically heterogeneous, both in geographical and bathymetrical terms (Hutchinson et al. 2004). The observed offset between the true (atmospheric) 14C age and the apparent (marine) 14C age is called the ‘marine radiocarbon reservoir effect’ and is known to vary spatially and temporally (eg, Kennett et al. 1997; Russel 2011). To compensate for it, a correction is applied to radiocarbon ages of material obtaining their carbon directly from the marine environment. A global offset of about 400 radiocarbon years between the atmosphere and the surface oceans, known as R(t), was calculated by Stuiver et al. (1986) and Stuiver and Braziunas (1993) using a simple ocean-atmosphere box diffusion model. The R(t) for the Mediterranean, for example, is estimated at 390±85 14C years BP (Siani et al. 2000), a value comparable to that for the North Atlantic Ocean (30 ka BP) samples for which offsets of several millennia may be recorded. Diagenesis of archaeological shell takes place in near-surface, marine and meteoric environments and involves several processes such as dissolution, recrystallisation, replacement and calcitisation, all of which are often embraced under the term neomorphism. Neomorphism includes all transformations between one mineral and itself or a polymorph, and often requires the presence of aqueous solutions in order to evolve. Some of the earliest cases of diagenesis in archaeology and radiocarbon dating were reported by Chappell and Polach (1972). Calcium carbonate is a chemically active compound that is expected to convert to its most stable form under certain depositional conditions. Hence, it is broadly assumed that the two metastable biogenic polymorphs (HMgC and aragonite) will, in time, recrystallise to the more stable LMgC. Very rare exceptions to this assumption involve the isomineralogical recrystallisation of aragonite structures back to aragonite (Perrin 2004; Webb et al. 2007). However, these cases are attributed to very specific and rarer environmental conditions, aqueous geochemistry and the degree of carbonate supersaturation in the diagenetic fluid. For example, aragonite cement precipitation is favored in shallow freshwaters with high Mg:Ca ratios (Webb et al. 2007) because the presence of Mg2+ interferes with calcite formation. In such conditions, as well as in sediments containing organic matter enriched in acidic amino acids (Stephenson et al. 2008) or CO2 released by bacterial decomposition of organic matter (Peckmann et al. 1999), aragonite cement may grow instead of secondary LMgC (Fig. 22.2).

22. Radiocarbon dating of marine and terrestrial shell

393

Figure 22.2. (a) Scanning Electron Microscopy (SEM) of the internal structure of a modern mussel shell in pristine condition where the nacreous layer clearly illustrates the stacking of aragonite tablets (microstructure ‘b’ in Fig. 22.1) and calcite prisms (microstructure ‘d’ in Fig. 22.1) on the upper left corner of the image. Modified from Blank et al. (2003); (b) SEM image of a Pleistocene-age shell clearly demonstrating the presence of secondary cement connecting adjacent tiles into vertical columnar arrays. In this case, the cement was identified as aragonite and this isomineralogical transformation is attributed to rare groundwater chemical composition and significantly elevated Mg:Ca ratios (modified from Webb et al. 2007). Meteoric diagenesis occurring early in the depositional history of the sample will have a small effect in radiocarbon dating; however, it is virtually impossible to know when diagenesis took place, what carbon species were involved and their effect on the radiocarbon measurement. Rigorous screening and effective pretreatment is essential to detect and reduce diagenetic effects and the risk of erroneous radiocarbon ages.

Case studies Despite a seemingly reluctance of archaeologists and radiocarbon laboratories to date shell due to difficulties involved in the interpretation of the results, this type of material has become more and more popular because it survives better than other forms of organic C. This is especially so with large-scale dating projects, where the issues reported above can be explored and accounted for, or in cases where alternative terrestrial material is not present. The use of shell as a dating material has been widely explored by laboratories and archaeologists in the past decade. For example, in the Palaeolithic, the antiquity of marine resources exploitation by hominins has become an important subject of research which often involves the use of marine shell for dating purposes. Douka (2011) and Douka et al. (2012; 2013) approached the subject through the use of Palaeolithic shell beads as proxies of the first anatomically and behaviorally modern humans arriving into Europe, about 40–45,000 years ago. Given

394

Katerina Douka

Figure 22.3. Examples of two Bronze Age burials from Ban Non Wat (Central Thailand), with rich evidence for the use of molluscan shells – both exotic marine and local freshwater types – for ornamental purposes in the form of bangles and beads (after Higham & Kijngam 2011). Given that collagen does not preserve in the skeletal remains of the site, the shell ornaments were the only means of establishing the absolute age of the burials. Associated charcoal remains were also used to calculate reservoir effects. In this case, due to non-limestone geology, no hard water effect was observed in the freshwater shells, hence the radiocarbon age of the ornaments corresponds closely to the age of the buried skeleton. that Neanderthals were not manufacturing beads, and did not widely use shell for ornamental purposes, the presence of shell beads at Upper Palaeolithic levels can be taken as a marker for incoming Homo sapiens. In some cases, the close correspondence of 14 C ages on marine shell and charcoal samples from the same contexts (eg, Douka et al. 2012; 2013; 2014), have bolstered the reliability of AMS determinations of shell material. Marine shell was used successfully as a dating material to explore the earliest colonisation and marine resource use of island southeast Asia and Sahul, around 42,000 years ago (O’Connor 2007; O’Connor et al. 2011). For later prehistoric periods, eg, the late Mesolithic and Neolithic, marine gastropods were used to define the age of late hunter-gatherers and pastorals and their subsistence patterns living in Sicily during the Pleistocene/Holocene transition (Mannino et al. 2007). Freshwater shells also offer an important alternative in the dating of sites where other types of organic material do not survive or where shell artefacts are used to date specific contexts. In the case of the Neolithic and Bronze Age site of Ban Non Wat, in Thailand, where over 600 human burials were uncovered (Higham & Kijngam 2011),

22. Radiocarbon dating of marine and terrestrial shell

395

freshwater and marine shell was used widely for the manufacture of ornaments, mainly in the form of tiny beads, earrings and large bangles (Fig. 22.3). A comprehensive dating programme focusing on the analyses of ~50 shell artefacts from individual burials produced a series of AMS dates closely corresponding to the mortuary phases identified during excavation (Higham & Higham 2009). At this site, a limestone-free location, the freshwater bivalves used for dating were shown not to be affected by the hard-water problem (ibid.). Recently, the suitability of terrestrial shell for dating an archaeological site was examined in an unpublished PhD thesis (Hill 2014) who used terrestrial molluscs, especially Helix melanostoma, from the Palaeolithic site of Haua Fteah and adjacent region in Cyrenaica, Libya, as a dating material. The site was dated by a large number of AMS determinations of charcoal (mainly) and marine shell, as well as by luminescence and tephra methodologies (Douka et al. 2014). In the case of the terrestrial landsnails, a relatively constant pattern of fractionation was found in modern Helix melanostoma specimens, which were favored for dating. In addition, a constant deriving from the dating of paired samples of charred pine cones and terrestrial molluscan shell from the same contexts was calculated at 410±24 14C years (Douka et al. 2014). This value was subtracted from all radiocarbon determinations prior to calibration. The land snail AMS determinations show a degree of internal variability, which reaches up to 6000 years for samples from the same contexts. At a low-rate sedimentation trap such as Haua Fteah, this can either be attributed to depositional disturbances which cause material of different ages to be found in the same context, or else, it may reflect the likelihood of terrestrial molluscs burrowing and dying in sub-surface sediments.

Final remarks Molluscan shell is a resistant biomaterial, regularly recovered from prehistoric and historic archaeological sites. However, technical considerations in the radiocarbon dating of this material, such as the various reservoir effects and uncertainties over their quantification and temporal variability, and the problem of diagenesis often need further attention. Fortunately, as described above, the majority of these variables can be addressed and overcome when archaeologists work in close collaboration with the radiocarbon specialists. In the presence of suitable samples (based on definitions of ecology, sources of CaCO3, composition and structure of the shell), reliable 14C dates can be obtained on molluscan shell that are not significantly different, if at all, from determinations obtained on other materials from the same context.

References Addadi, L. Joester, D., Nudelman, F. & Weiner, S. 2006. Mollusk shell formation: a source of new concepts for understanding biomineralization processes, Chemistry – A European Journal 12(4), 980–987 Aitken, M. J. 1990. Science-based Dating in Archaeology. Longman Archaeology Series, London: Longman

396

Katerina Douka

Ascough, P., Cook, G. T., Dugmore, A., Scott, M. E. & Freeman, S. P. H. T. 2005. Influence of Mollusk species on marine !R determinations, Radiocarbon 47(3), 433–440 Blank, S., Arnoldi, M., Khoshnavaz, S., Treccani, L., Kuntz, M., Mann, K., Grathwohl, G., Fritz, M. 2003. The nacre protein perlucin nucleates growth of calcium carbonate crystals, Journal of Microscopy 212(3), 280–291 Bøggild, O. B. 1930. The Shell Structure of the Mollusks. Det Kongelige Danske Videnskabernes Selskabs Skrifter, Naturvidenskabelige og Mathematiske Afdeling Roekke 9, 233–326. Copenhagen: Det Kongelige Bronk Ramsey, C. 2008. Radiocarbon dating: revolutions in understanding, Archaeometry 50, 249–275 Carter, J. G. 1980. Environmental and biological controls of bivalve shell mineralogy and microstructure. In Rhoads, D. C., Lutz, R. A. (eds), Skeletal Growth of Aquatic Organisms, 69–113. New York: Plenum Press Carter, J. G. 1990. Skeletal Biomineralisation: patterns, processes and evolutionary trends. New York: Van Nostrand Reinhold Chappell, A. & Polach, H. A. 1972. Some effects on partial recrystallisation on 14C dating late Pleistocene corals and molluscs, Quaternary Research 2, 244– 252 Chateigner, D., Hedegaard, C. & Wenk, H.-R. 2000. Mollusc shell microstructures and crystallographic textures, Journal of Structural Geology 22, 1723–1735 Currey, J. D. & Taylor, J. D. 1974. The mechanical behavior of some Molluskan hard tissues, Journal of Zoology 173(3), 395–406 Dauphin, Y. & Denis, A. 2000. Structure and composition of the aragonitic crossed lamellar layers in six species of Bivalvia and Gastropoda, Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology 126, 367–377 Deevey, E. S., Jr., Gross, M. S., Hutchinson, G. E. & Kraybill, H. L. 1954. The natural 14C contents of materials from hard-water lakes, Proceedings of the National Academy of Sciences of’ USA 40, 285–288 Douka, A. 2011. Investigating the Chronology of the Middle to Upper Palaeolithic Transition in Mediterranean Europe by Improved Radiocarbon Dating of Shell Ornaments. Unpublished DPhil Thesis, University of Oxford Douka K., Hedges R. E. M. & Higham T. F. G. 2010. Improved AMS 14C dating of shell carbonates using high-precision X-Ray Diffraction (XRD) and a novel density separation protocol (CarDS), Radiocarbon 52(2), 735–751 Douka, K., Bergman, C. A., Hedges, R. E. M., Wesselingh, F. P. & Higham, T. F. G. 2013. Chronology of Ksar Akil (Lebanon) and implications for the colonization of Europe by anatomically modern humans, PLoS ONE 8(9): e72931 Douka,  K., Grimaldi, S., Boschian, G., del Lucchese, A. & Higham, T. F. G. 2012. A new chronostratigraphic framework for the Upper Palaeolithic of Riparo Mochi (Italy), Journal of Human Evolution 62(2), 286–299 Douka, K., Grün, R., Jacobs, Z., Lane, C., Farr, L., Hunt, C., Inglis, R. H., Reynolds, T., Albert, P., Aubert, M., Cullen, V., Hill, E., Kinsley, L., Roberts, R. G., Tomlinson, E. L., Wulf, S. & Barker, G. 2014. The chronostratigraphy of the Haua Fteah cave (Cyrenaica, northeast Libya), Journal of Human Evolution 66, 39–63 Dye, T. 1994. Apparent ages of marine shells: implications for archaeological dating in Hawaii, Radiocarbon 36, 51–57 Epstein, S., Buchsbaum, R. Lowenstam, H. & Urey, H. C. 1951. Carbonate-water isotopic temperature scale, Bulletin Geological Society of America 62, 417–426 Evin, J., Marechal, J., Pachiaudi, C. & Puissegue, J. J. 1980. Conditions involved in dating terrestrial shell, Radiocarbon 22(2), 545–555

22. Radiocarbon dating of marine and terrestrial shell

397

Facorellis, Y. & Vardala-Theodorou, E. 2015. Sea surface radiocarbon reservoir age changes in the Aegean Sea from about 11,200 BP to present, Radiocarbon 57(3), 491–503 Furuhashi, T., Miksik, I., Smrz, M., Germann, B., Nebija, D., Lachmann, B. & Noe, C. 2010. Comparison of aragonitic molluscan shell proteins, Comparative Biochemistry and Physiology B 155, 195–200 Geyh, M. A., Scotterer, U. & Grosjean, M. 1998. Temporal changes of the 14C reservoir effect in lakes, Radiocarbon 40(2), 921–931 Gillikin, D. P., Lorrain, A., Bouillon, S., Willenz, P. & Dehairs, F. 2006. Shell carbon isotopic composition of Mytilus edulis shells: relation to metabolism, salinity, δ13CDIC and phytoplankton, Organic Geochemistry 37, 1371–1382 Gillikin, D. P., Lorrain, A., Meng, L. & Dehairs, F. 2007. A large metabolic carbon contribution to the δ13C record in marine aragonitic bivalve shells, Geochimica et Cosmochimica Acta 71, 2936–2946 Gischler, E., Gibson, M. A. & Oschmann, W. 2008. Giant Holocene Freshwater Microbialites, Laguna Bacalar, Quintana Roo, Mexico Sedimentology 55(5), 1293–1309 Goodfriend, G. A. & Hood, D. G. 1983. Carbon isotope analysis of landsnail shells: implications for carbon sources and radiocarbon dating, Radiocarbon 25(3), 810–830 Goodfriend, G. A. & Stipp, J. J. 1983. Limstone and the problem of radiocarbon dating of landsnail shell carbonate, Geology 11, 575–577 Harper, E. M. 1998. Calcite in chamid bivalves, Journal of Molluscan Studies 64, 391–399 Heller, J. & Magaritz, M. 1983. From where do land snails obtain the chemicals to build their shells?, Journal of Molluscan Studies 49(2), 116–121 Henshilwood, C. S., d’Errico, F., Vanhaeren, M., Van Niekerk, K. & Jacobs, Z. 2004. Middle Stone Age shell beads from South Africa, Science 304, 404 Higham, C. F. W. & Higham, T. F. G. 2009. A new chronological framework for prehistoric Southeast Asia based on a Bayesian model from Ban Non Wat, Antiquity 83, 125–145 Higham, C. F. W. & Kijngam, A. (eds), 2012. The Origins of the Civilization of Angkor. Volume V. The Excavation Ban Non Wat: Bronze Age. Bangkok: Fine Arts Department of Thailand Hill, E. A. 2014. The radiocarbon dating of terrestrial molluscs in North East Libya. Unpublished PhD thesis, Queen’s University Belfast Hogg, A. G., Higham, T. F. G. & Dahm, J. 1998. Radiocarbon dating of modern marine and estuarine shellfish, Radiocarbon 40, 975–984 Hutchinson, I., James, T. S., Reimer, P. J., Bornhold, B. D. & Clague, J. J. 2004. Marine and limnic radiocarbon reservoir corrections for studies of late – and postglacial environments in Georgia Basin and Puget Lowland, British Columbia, Canada and Washington, USA, Quaternary Research 61, 193–203 Joordens, J. C. A. & d’Errico, F., et al. 2015. Homo erectus at Trinil in Java used shells for tool production and engraving, Nature 518(7538), 228–231 Keith, M. L., Anderson, G. M. & Eicheler, R. 1964. Carbon and oxygen isotopic composition of mollusk shells from marine and fresh-water environments, Geochimica et Cosmochimica Acta 28, 1757–1786 Keith, M. L. & Anderson, G. M. 1963. Radiocarbon dating: fictitious results with mollusk shells, Science 141(3581), 634–637 Kennett, D. J., Ingram, B. L., Erlandson, J. M. & Walker, P. 1997. Evidence for temporal fluctuations in marine radiocarbon reservoir ages in the Santa Barbara Channel, southern California, Journal of Archaeological Science 24, 1051–1059 Kobayashi, I. 1969. Internal microstructure of the shell of bivalve molluscs, American Zoologist 9, 663–672 Lorrain, A, Paulet, Y-M, Chauvaud, L, Dunbar, R, Mucciarone, D. & Fontugne, M. 2004. 13C

398

Katerina Douka

variation in scallop shells: increasing metabolic carbon contribution with body size?, Geochimica et Cosmochimica Acta 68, 3509–3519 Lougheed, B. C., Filipsson, H. L. and Snowball, I. 2013. Large spatial variations in coastal 14C reservoir age – a case study from the Baltic Sea, Climate Past 9, 1015–1028 Mangerud, J. 1972. Radiocarbon dating of marine shells, including a discussion of apparent age of Recent shells from Norway, Boreas 1, 143–172 Mannino, M. A., Thomas, K. D., Leng, M. J., Piperno, M., Tusa, S. & Tagliacozzo, A. 2007. Marine resources in the Mesolithic and Neolithic at the Grotta dell’Uzzo (Sicily): evidence from isotope analyses of marine shells, Archaeometry 49, 117–133 McConnaughey, T. A. 2003. Sub-equilibrium oxygen-18 and carbon-13 levels in biological carbonates: Carbonate and kinetic models, Coral Reefs 22, 316–327 McConnaughey, T. A., Burdett, J., Whelan, J. F. & Paull, C. K. 1997. Carbon isotopes in biological carbonates: respiration and photosynthesis, Geochimica et Cosmochimica Acta 61, 611–622 O’Connor, S. 2007. New evidence from East Timor contributes to our understanding of earliest modern human colonisation east of the Sunda Shelf, Antiquity 81, 523–535 O’Connor, S., Ono, R. & Clarkson, C. 2011, Pelagic fishing at 42,000 years before the present and the maritime skills of modern humans, Science 334(6059), 1117–1121 Pearson, G. W. & Stuiver, M. 1986. High-precision calibration of the radiocarbon time scale 500–2500 BC, Radiocarbon 28(2B), 839–862 Peckmann, J., Paul, J. & Thiel, V. 1999. Bacterially mediated formation of diagenetic aragonite and native sulfur in Zechstein carbonates (Upper Permian, Central Germany), Sedimentary Geology 126, 205–222 Perrin, C. 2004. Early diagenesis of carbonate biocrystals: isomineralogical changes in aragonite coral skeletons, Bulletin de la Société Geologique de France 175, 95–106 Petchey, F., Ulm, S., David, B., McNiven, I. J., Asmussen, B., Tomkins, H., Richards, T., Rowe, C., Leavesley, M., Mandui, H. & Stanisic, J. 2012. 14C marine reservoir variability in herbivores and deposit-feeding gastropods from an open coastline, Papua New Guinea, Radiocarbon 54, 967–997 Pigati, J. S., Rech, J. A. & Nekola, J. C. 2010, Radiocarbon dating of small terrestrial gastropod shells in North America, Quaternary Geochronology 5, 519–532 Poulain, C., Lorrain, A., Mas, R., Gillikin, D. P., Dehairs, F., Robert, R. & Paulet, Y.-M. 2010. Experimental shift of diet and DIC stable carbon isotopes: influence on shell 13C values in the Manila clam Ruditapes philippinarum, Chemical Geology 272, 75–82 Rick, T. C., Vellanoweth, R. L. & Erlandson, J. M. 2005. Radiocarbon dating and the ‘old shell’ problem: direct dating of artifacts and cultural chronologies in coastal and other aquatic regions, Journal of Archaeological Science 32, 1641–1648 Reimer, P. J., Baillie, M. G. L., Bard, E., Bayliss, A., Warren Beck, J., Bertrand, C. J. H., Blackwell, P. G., Buck, C. E., Burr, G. S., Cutler, K. B., Damon, P. E., Lawrence Edwards, R., Fairbanks, R. G., Friedrich, M., Guilderson, T. P., Hogg, A. G., Hughen, K. A., Kromer, B., McCormac, G., Manning, S., Bronk Ramsey, C., Reimer, R. W., Remmele, S., Southon, J. R., Stuiver, M., Talamo, S. & Taylor, F. W. 2004. IntCal04 terrestrial radiocarbon age calibration 0–26 cal kyr (BP), Radiocarbon 46(1), 1029–1058 Reimer, P. J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Bronk Ramsey, C., Buck, C. E., Burr, G. S., Edwards, R. L., Friedrich, M., Grootes, P. M., Guilderson, T. P., Hajdas, I., Heaton, T. J., Hogg, A. G., Hughen, K. A., Kaiser, K. F., Kromer, B., McCormac,  F.  G., Manning, S. W., Reimer, R. W., Richards, D. A., Southon, J. R., Talamo, S., Turney, C. S. M., van der Plicht, J. & Weyhenmeyer, C. E. 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP, Radiocarbon 51, 1111–1150

22. Radiocarbon dating of marine and terrestrial shell

399

Russell, N. 2011. Marine Radiocarbon Reservoir Effects (MRE) in Archaeology: temporal and spatial changes through the Holocene within the UK coastal environment. Unpublished PhD thesis, University of Glasgow Russo, C. M., Tripp, J. A., Douka, K. & Higham, T. F. G. 2010. A new pretreatment method for molluscan shell using density fractionation of carbonates in Bromoform, Radiocarbon 52(3), 1301–1311 Siani, G., Paterne, M., Arnold, M., Bard, E., Metivier, B., Tisnerat, N. & Bassinot, F. 2000. Radiocarbon reservoir ages in the Mediterranean Sea and Black Sea, Radiocarbon 42(2), 271–280 Stephenson, A. E., DeYoreo, J. J., Wu, L., Wu, K. J., Hoyer, J. & Dove, P. M. 2008. Peptides enhance magnesium signature in calcite: insights into origins of vital effects, Science 322, 724–727 Stuiver, M. & Braziunas, T. F. 1993. Modeling atmospheric 14C influences and 14C ages of marine samples to 10,000 BC, Radiocarbon 35(1), 137–189 Stuiver, M. & Suess, H. E. 1966. On the relationship between radiocarbon dates and true sample ages, Radiocarbon 8, 534–540 Stuiver, M., Pearson, G. & Braziunas, T. 1986. Radiocarbon age calibration of marine samples back to 9000 cal yr BP, Radiocarbon 28(2b), 980–1021 Stuiver, M., Reimer, P. J., Bard, E., Warren Beck, J., Burr, G. S., Hughen, K. A., Kromer, B., McCormac, G., van der Plicht, J. & Spurk, M. 1998. IntCal98: Radiocarbon age calibration 24,000–0 Cal BP, Radiocarbon 40(3), 1041–1084 Szabó, K. 2017. Molluscan shells as raw materials for artefact production. In Allen, M. J. (ed.), Molluscs in Archaeology, 308–325. Oxford: Oxbow Books Tamers, M. A. 1970, Validity of radiocarbon dates on terrestrial snail shells, American Antiquity 35, 94–100 Tanaka, N., Monaghan, M. C. & Rye, D. M. 1986. Contribution of metabolic carbon to mollusc and barnacle shell carbonate, Nature 320, 520–523 Taylor, J. D. & Reid, D. G. 1990. Shell microstructure and mineralogy of the Littorinidae: ecological and evolutionary significance, Hydrobiologia 193, 199–215 Urey, H. C., Epstein, S., Lowenstam, H. A. & McKinney, C. R. 1951. Measurement of palaeotemperatures and temperatures of the Upper Cretaceous of England, Denmark and the southeastern United States, Geological Society of America Bulletin 62, 399–416 Webb, G. E., Price, G. J., Nothdurft, L. D., Deer, L. & Rintoul, L. 2007. Cryptic meteoric diagenesis in freshwater bivalves: Implications for radiocarbon dating, Geology 35, 803–806 Yates, T. J. S. 1986. Selection of non-marine molluscan shells for radiocarbon dating. DPhil dissertation, University College London http://discovery.ucl.ac.uk/1317885/2/295736.pdf

23. Stable isotope ecology of terrestrial gastropod shells André Carlo Colonese

With approximately 35,000 species, terrestrial gastropods are one of the most successful, ubiquitous and diverse animals in land-based ecosystems (Barker 2001). In geological and archaeological contexts, the shells of these animals offer valuable environmental information, accessed through distinct analytical techniques. This chapter provides an introduction to the analysis of carbon and oxygen stable isotopes in terrestrial gastropods shells. It reviews aspects of sampling strategy, preparation and analysis, and discusses some applications on archaeological shells. The main objective is to demonstrate the contribution of the stable isotope ecology of terrestrial gastropods for the study of past climates and environments.

The stable isotope ecology of terrestrial gastropods – synthesis Stable isotopes are atoms that do not decay into other elements. They have the same number of protons and electrons but differing numbers of neutrons. As a consequence, isotopes of the same element have similar chemical properties but distinct atomic masses. Amongst stable isotopes, those of oxygen (O) and carbon (C) are the most commonly used in palaeoclimatic investigations based on carbonates (shell, spelothems). Oxygen exists as three isotopes: 16O, 17O, 18O with natural abundances of 99.76%, 0.04% and 0.20% respectively. Carbon instead has two stable isotopes: 12C and 13C, respectively present in nature at 98.89% and 1.11%. During chemical and physical reactions, such as the precipitation of shell carbonate (CaCO3) by terrestrial gastropods, the mass of the isotopes influences their distribution between reactant (eg, body fluid) and product (eg, shell), with heavier isotopes reacting more slowly than the lighter isotopes. This differential distribution is called isotopic fractionation and depends on factors such as temperature and/or rate of reaction. Stable isotope analysis investigates the ratio of the heavier isotope to the lighter, for example 18O/16O and 13C/12C. However, as the absolute abundance of minor isotopes cannot be determined accurately, researchers usually use the unit δ, defined as per thousand (‰) in relation to standards of known isotopic composition: δ‰ = [(Rsample / Rstandard) – 1)] x 103

23. Stable isotope ecology of terrestrial gastropod shells

401

where R is the isotope ratio of a given element (eg, 18O/16O, 13C/12C) in the sample (eg, shell) and in the standard. Positive δ values indicate enrichment in the heavy isotopes (eg, 18O) in relation to the standard, while negative δ values indicate depletion. The commonly used standard for shell carbonate, plants and soils is the Vienna Pee Dee Belemnite (V-PDB). The common standard for water is the Vienna Standard Mean Ocean Water (V-SMOW).

Source of oxygen and carbon stable isotopes in shell carbonate The oxygen and carbon isotope composition of terrestrial gastropod shells is determined by the isotopic composition of their sources, as well as by subsequent isotopic fractionations within the system between the body fluid and the shell. Most terrestrial gastropods secrete an external aragonite shell (CaCO3) from dissolved bicarbonates (HCO3¯) in body fluids. The water in the body fluid derive predominantly from atmospheric water (eg, rain, water vapour and dew), absorbed during the animal’s activity after rainfall, when relative humidity is high, or in moist habitats (Prior 1985; Cook 2001). The oxygen in the body water then undergoes rapid isotopic exchange with the HCO3¯ (Goodfriend 1992). The main source of carbon in the HCO3¯ is respired CO2 derived from food (Goodfriend 1992), which for the vast majority of terrestrial gastropods are living or decaying plants, along with fungi, animals, soil, algae, mosses, lichens, etc. (Speiser 2001). These processes make the shell carbonate of terrestrial molluscs attractive for exploring stable carbon and oxygen isotope composition in relation to environmental conditions.

Interpreting shell δ18O values The oxygen isotope composition of CaCO3 is a function of the isotopic composition of ambient water and temperature (Epstein et al. 1953). However, in the case of terrestrial gastropods, the oxygen isotope composition of their shells is also affected by other environmental processes and by physiological and ecological conditions. Therefore, although field studies have demonstrated that bulk shell δ18O values can be a valuable source of palaeoclimatic information (palaeorainfall and humidity) environmental interpretations will generally depend on regional-scale processes. For example, several studies have observed that shell δ18O values reflect, to different extents, the δ18O values of the local environmental waters including rain, water vapour and dew, which are absorbed by the animal. Shell δ18O values covary positively with the δ18O values in environmental waters (eg, Goodfriend et al. 1989; Zanchetta et al. 2005; Prendergast et al. 2015a), and this has been recently confirmed by a laboratory study with snails raised with water having distinct δ18O values (Rui & XueFen 2015). However the oxygen isotope compositions of atmospheric water is variably controlled by temperature, altitude, amount of precipitation, rainout, source and trajectory of air mass (Rozanski et al. 1993), therefore positive relationships between bulk shell and atmospheric water δ18O values will vary from region to region (Lécolle 1985; Goodfriend & Magaritz 1987; Goodfriend et al. 1989; Zanchetta et al. 2005). Furthermore, it is also possible that δ18O values are affected by the δ18O values of ingested plants.

402

André Carlo Colonese

The effect of temperature on the oxygen isotopic fractionation of shell carbonate is also less evident and somewhat contingent upon the δ18O values of environmental waters. For example, as said above, the oxygen isotopic composition of CaCO3 is strongly controlled by temperature, with a coefficient of ~-0.2‰ per ºC (Kim & O’Neil 1997). However this relation can be masked by the effect of the temperature on the oxygen isotope composition of precipitation. For example, in European localities the coefficient of this relationship is ~+0.6‰ per ºC (Rozanski et al. 1992). Thus the use of shell δ18O values of terrestrial gastropods to directly derive the temperatures at which the shell is formed present some limitations. Evaporation is an additional factor affecting the δ18O values of terrestrial gastropod shells. Evaporation preferentially removes the light water molecules (H216O) from the body fluids of terrestrial gastropods when they are active, as well as from the environmental waters they absorb/ingest (Goodfriend et al. 1989; Leng et al. 1998; Yapp 1979). Balakrishnan and Yapp (2004) developed a comprehensive model to interpret shell δ18O values which takes into account the effect of evaporation on the animal body fluids. The model reveals that evaporation promoted by low humidity would notably enrich body fluids in 18O and, consequently, increase shell δ18O values. Balakrishnan and Yapp’s (2004) model has been shown to predict shell δ18O values with good accuracy in several environmental contexts (eg, Balakrishnan et al. 2005a; 2005b; Colonese et al. 2011a; 2013a; 2013b; Yanes et al. 2012; 2013b). Beside environmental factors, the stable isotope record of gastropod shells is affected by physiological and ecological controls (Goodfriend 1992). For example, water availability is a crucial requirement for terrestrial gastropods as they are extremely susceptible to dehydration. As a consequence, they tend to adjust their times of activity to the most favourable conditions. As the shell is formed when the mollusc is active, its isotope record will reflect environmental conditions during the activity of the animal. In warm and humid tropical regions, for example, the shells will most likely reflect annual environmental conditions (Baldini et al. 2007; Yanes et al. 2015) as optimal conditions are fairly consistent throughout the year. In temperate and high latitude regions, more variable seasonality of growth would be expected with shell isotopes recording mainly warm, summer conditions with a period of interruption during lower winter temperatures (eg, Yanes 2015). By contrast, many species living around the Mediterranean basin aestivate in summer, thus their shell stable isotopes will predominantly reflect environmental conditions in spring, autumn and potentially also in winter. Insights into the seasonality of shell formation are offered by isotopic analysis along the shell growth increments (Fig. 23.1). Cyclical variations in δ18O values are associated with seasonal changes in the isotopic composition of precipitation coupled to evaporation during the animal’s life cycle. Co-existing species usually display similar shell δ18O values, which reflect their similar isotopic response to local atmospheric conditions. However, some intra-specific variability is common at the population level and possibly arises from several factors, including the microhabitat conditions, duration and extent of activity, life cycle, and size (Goodfriend & Magaritz 1987; Goodfriend & Ellis 2002; Balakrishnan et al. 2005a). In summary shell δ18O values are a valuable proxy of local to regional atmospheric

23. Stable isotope ecology of terrestrial gastropod shells

403

Figure 23.1. Variations in shell δ18O and δ13C values (2-period moving average) of modern Cornu aspersum (garden snail) from Greece. This shell was collected empty in June 2006 from a recently dead individual. The grey bands indicate period of growth interruptions, likely during beginning and end of winter (hibernation). Higher growth rate is observed during higher and lower δ18O and δ13C values respectively, after the hibernation. Although it is not possible to assign a precise calendar to isotopic fluctuations, the trend in δ18O values appears to record changes in precipitation δ18O from a wetter/colder to a drier/warmer season over a one year time interval. The trend is also recorded by δ13C values that could indicate changes in feeding behaviour or activity site. However the negative correlation between δ18O and δ13C values could also indicate a lower metabolic rate during winter, when the snail has visibly reduced its activity

conditions (rainfall, relative humidity) with the potential to also provide quantitative data on the relative amount of precipitation and humidity (Colonese et al. 2010a; 2010b; Yanes et al. 2011b). Empirically derived relationships between shell δ18O values and environments variables are commonly developed in order to interpret shell δ18O values in geological and archaeological records.

Interpreting shell δ13C values Shell carbonate is formed from HCO3¯ in body fluid which is derived from three main sources of carbon: respired CO2 derived from diet, metabolic CO2 from ingestion of carbonate rocks (limestone or soil-rich carbonate), and atmospheric CO2 (eg, Goodfriend & Hood 1983). Although the relative contribution of these sources is far from being systematic, several lines of evidence converge on the model that respired CO2 from diet (mainly plants) is the main source of carbon in shell carbonate, meaning that the stable

404

André Carlo Colonese

carbon isotope composition of the food will largely determine the isotopic composition of the shell (eg, Stott 2002; Metref et al. 2003; Zhang et al. 2014). Plants having different photosynthetic pathways have very distinctive stable carbon isotope compositions due to the physical and chemical processes involved in the assimilation of atmospheric CO2 (O’Leary 1988). For example plants using Calvin (C3 plants) and Hatch–Slack cycle (C4 plants) pathways have average δ13C values of –27‰ and –14‰ respectively, while a third group composed mainly of succulent and desert plants, Crassulacean acid metabolism (CAM plants), have δ13C values usually between those of C3 and C4 plants. These isotopic differences are large enough that they can be detected in shell δ13C values. C3 plants dominate the vegetation composition in temperate environments, and include trees as well as shrubs, while C4 plants are mainly represented by herbaceous plants, particularly inhabiting arid, semiarid and tropical regions (Still et al. 2003). As a general rule, terrestrial gastropods living in environments dominated by C3 plants, such as Europe, have shell δ13C values comparably lower than those inhabiting environments dominated by or containing a large relative proportion of C4 plants (eg, Yanes et al. 2013c). Field and laboratory experiments, the latter using snails raised with isotopically labelled food (eg, plants) and mineral supplements (CaCO3 and CaPO4), demonstrate that correlations between shell and plant δ13C values are strong but not constant. Whilst the δ13C values of body tissues usually reflect the δ13C values of their food, considerable variation has been observed between shell and ingested food δ13C values, indicating that isotopic fractionation between diet and shell may be affected by complementary sources of carbon (carbonate rocks), environmental variables and food preferences (eg, DeNiro & Epstein 1978; Stott 2002; Metref et al. 2003; Zhang et al. 2014). Ingestion of carbonate has been shown to promote considerable deviations in shell δ13C toward high values (Yanes et al. 2008; Zhang et al. 2014), but correlations between shell δ13C values and amount of ingested carbonates are not always straightforward (Romaniello et al. 2008). Furthermore, while its contribution can be remarkable for some species (or in some environments) it may be minimal or negligible for others. For example, metabolic CO2 from carbonates has been estimated to contribute up to ~40% to shell carbonate of some species (Yanes et al. 2008), and particularly in calcareous areas (Goodfriend & Stipp 1983; Yates et al. 2002; Romaniello et al. 2008; Yanes et al. 2013a), but it has been found to be negligible for some small gastropods (eg,