The Use and reuse of stone circles: Fieldwork at five Scottish monuments and its implications 1785702432, 9781785702433

Citation preview

THE USE AND REUSE OF

STONE CIRCLES

US D T UT E T H D TE ET EU E E T T DR TE TU T U UU UT TS T U E monuments of Orkney and the Western Isles. Several decades of fieldwork have shown how these major structures are likely to be of early date and recognised that U U E UU T T T U TE N U UTEDF DT U UES USE in Northern Britain were reused during the later Bronze Age, the Iron Age and the ED A HE T F E T HET SEU E A U UT S O when were the last stone circles built? How did they differ from earlier constructions? How were they related to henge monuments, especially those of Bronze Age date? How frequently were these places reused, and did this secondary activity change the character of those sites? This major new assessment first presents the results of fieldwork undertaken at the Scottish recumbent stone circle of Hillhead; the stone E T S ET U TE RU UT E U T S at Kintore; and the small ring cairn at Laikenbuie in Inverness-shire. Part 2 brings together the results of these five projects and puts forward a chronology for the T UES UT HE ED S T UT E R H EU S E D U T U Bronze Age examples. It considers the reuse of stone circles, long after they were built, and discusses four neighbouring stone circles in Aberdeenshire which display both similarites and contrasts in their architecture, use of raw materials, associated artefacts and structural sequences. Finally, a reassessment and reinterpretation of Croftmoraig and its sequence is presented: the new interpretation drawing attention to ways of thinking about these monuments which have still to fulfil their potential.

www.oxbowbooks.com

E D SE

ISBN 978≠ 1≠ 78570≠ 243≠ 3

D E TSEU D

is a Postdoctoral Researcher at the Institute of Archaeology, University of Oxford and Assistant Editor of the Proceedings of the Prehistoric T UDF E HE H E E UE U E A ET EUR U EUR Bronze Age and Iron Age society, and maritime and intertidal archaeology.

U

is one of the world’s leading prehistorians and Emeritus Professor at the University of Reading where he taught for 40 years. His research interests are many and varied, with a focus on the Neolithic and Bronze Age of Britain and northwest Europe. Key areas of interest are Prehistoric landscapes, Rock Art and ritual and T TE UT F

Fieldwork at Five Scottish Monuments and its Implications

The Use and Reuse of Stone Circles Fieldwork at Five Scottish Monuments and its Implications

Edited by

Richard Bradley and Courtney Nimura

Oxford and Philadelphia

Croftmoraig stone circle viewed from the northeast (Aaron Watson).

For Aubrey Burl il miglior fabbro

Published in the United Kingdom in 2016 by OXBOW BOOKS 10 Hythe Bridge Street, Oxford OX1 2EW and in the United States by OXBOW BOOKS 1950 Lawrence Road, Havertown, PA 19083 © Oxbow Books and the individual authors 2016 Paperback Edition: ISBN 978-1-78570-243-3 Digital Edition: ISBN 978-1-78570-244-0 A CIP record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Names: Bradley, Richard, 1946- | Nimura, Courtney. Title: The use and reuse of stone circles : fieldwork at five Scottish monuments and its implications / edited by Richard Bradley and Courtney Nimura. Description: Oxford ; Philadelphia : Oxbow Books, 2016. | Includes bibliographical references and index. Identifiers: LCCN 2016013175 (print) | LCCN 2016014062 (ebook) | ISBN 9781785702433 (paperback) | ISBN 9781785702440 (digital) | ISBN 9781785702440 (epub) | ISBN 9781785702457 (mobi) | ISBN 9781785702464 (pdf ) Subjects: LCSH: Stone circles--Scotland. | Excavations (Archaeology)--Scotland. | Archaeology--Fieldwork--Scotland. | Scotland-Antiquities. Classification: LCC GN805 .U87 2016 (print) | LCC GN805 (ebook) | DDC 936.1/1--dc23 LC record available at http://lccn.loc.gov/2016013175

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 Malta by Melita Press For a complete list of Oxbow titles, please contact: United Kingdom Oxbow Books Telephone (01865) 241249 Fax (01865) 794449 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: The Croftmoraig stone circle viewed from the southeast with the summit of Schiehallion in the background (Aaron Watson).

Contents Preface. The Contents of the Volume Richard Bradley Summary Acknowledgments List of Figures List of Tables Abbreviations

vii vii xi xiii xvi xvi

Part One  Excavations at Five Scottish Monuments 1 The Development of the Project Richard Bradley

2

2

Excavations at Hillhead, Tarland, Aberdeenshire: a recumbent stone circle and its history Richard Bradley and Amanda Clarke

7

3

Excavations at Waulkmill, Tarland, Aberdeenshire: a Neolithic pit, Roman Iron Age burials and an earlier prehistoric stone circle Richard Bradley, Amanda Clarke and Fraser Hunter

27

4 Croftmoraig Stone Circle, Perth and Kinross: a reinterpretation in the light of fresh excavation Richard Bradley

56

5

The Hill of Tuach, Kintore, Aberdeenshire: the excavation of a small stone circle and henge Richard Bradley and Amanda Clarke

74

6

Laikenbuie, Auldearn, Inverness-shire: excavation of an Early Iron Age ring cairn and other features Ronnie Scott and Annette Jack

102

Part Two  The Excavated Monuments in their Wider Contexts 7 After the Great Stone Circles Richard Bradley

112

8 Histories of Reuse Richard Bradley

122

9 The Extent of Variation: four stone circles in Cromar in the light of recent fieldwork Richard Bradley

134

10 Croftmoraig: the anatomy of a stone circle Richard Bradley

141

References Index

153 163

preface

The Contents of the Volume

Richard Bradley

The study of stone circles has played a major role in British and Irish archaeology, but the results of this work have seldom been drawn together in book form, although these evocative structures have featured in photographic essays about megalithic architecture intended for the general reader. The last systematic accounts of these sites were by John Barnatt in 1989 and Aubrey Burl in 2000. The most recent investigation is by Colin Richards (2013) and is limited to the large monuments of Orkney and the Western Isles. It is the only one of these books which reflects the current state of knowledge and in a sense The Use and Reuse of Stone Circles is both a companion volume and its sequel. Colin Richards’s book focuses on large monuments which are likely to be of early date and represents a new development in studies of these monuments. Another is the recognition that smaller settings of monoliths had an extended history and were still being built for some time after the period considered in the pioneering accounts of Burl and Barnatt. A further development is the realisation that many of the structures in Northern Britain were reused during the later Bronze Age, the Iron Age and the early medieval period. It happened at a time when comparable monuments had gone out of use in lowland Britain (Bradley 2011, 169–74). This development is most apparent from fieldwork undertaken over the last two decades in Scotland; there may have been a similar sequence in Wales, northern England and Ireland, but the question has still to be explored systematically. The Scottish evidence was originally investigated through three campaigns of survey and excavation concerned with Clava Cairns, recumbent stone circles and henge monuments respectively. The results of these projects were published as monographs in 2000, 2005 and 2011 (Bradley 2000; 2005; 2011).

Taken together, this work raised a series of problems that demanded further investigation. When were the last stone circles built? How did they differ from earlier constructions? How were they related to henge monuments, especially those of Bronze Age date? How frequently were these places reused, and did this secondary activity change the character of those sites? How much variation existed between monuments supposedly of the same types? Why were certain stone circles selected as the sites for later Bronze Age and Early Iron Age roundhouses? How were the last stone monuments related to the settlements of the same periods? Why were a few of them associated with Roman Iron Age burials, when they were located so far outside the northern frontier of Britannia? And what was the relationship between these sites and Pictish symbol stones? These were the main questions investigated by excavation in Inverness-shire, Aberdeenshire and Perthshire. This book is divided into two sections. The first presents the reports of five excavations, most of them conducted on a modest scale (Fig. 0.1). The second treats the results of these projects together and suggests a new framework for understanding Chalcolithic and Early Bronze Age stone circles as well as their subsequent reuse. Summary Chapter 1 sets out the background to this research. It reviews the reasons why these excavations took place and their relationship to previously published work. Part 1 presents the results of fieldwork and is divided into five chapters. It begins with a study of the Hillhead recumbent stone circle (Chapter 2). This was a new discovery and the monument, which had been damaged by forestry, proved to be structurally complex and

viii

Preface

Figure 0.1. Locations of the five excavations published in this study unexpectedly well-preserved. Key features included its unusual siting in the ancient landscape, its relationship to surface finds from the surrounding area, including the possible remains of a Beaker burial, and its remarkable structural history. It began as an unusually massive ring cairn with a stepped interior. It was supplemented by one of the largest but least monumental recumbent stone circles in Scotland and was aligned on a conspicuous landmark 15 km away. Charcoal from the old land surface provides rare dating evidence for a monument of this type. It was built at the end of the third millennium bc. In a secondary phase the interior of the ring cairn was filled with boulders and the perimeter wall was covered by similar material. This deposit was cut by three features dug through the exact centre of the monument. Two seem to have been fireplaces or hearths, whilst the third was an in situ cremation pyre marked by an area of burnt subsoil and a quantity of cremated bone which dated to the Late Bronze Age. After they had been sealed by a further deposit of rubble, the interior of the older ring cairn was covered with broken quartz. The reason for giving this project so much prominence is its complex sequence which includes a number of elements that have not been observed since the nineteenth century. Chapter 3 concerns a stone circle which would once have been visible from Hillhead; it was demolished about

1835. Excavation at Waulkmill in 2012 showed that it had incorporated another ring cairn, with a well-defined inner court associated with pieces of worked quartz. The central area was cut by a large pit, tentatively ascribed to the later Bronze Age, but the monument came back into use as the site of an unusually complex Roman Iron Age cemetery. Further burials were found close to the site a century ago. Taken together, there is evidence for at least four inhumation graves, two or three of them associated with gaming sets or bronze personal ornaments. A couple were in cists defined by drystone walls, while a third was in an oak coffin and associated with the skull of a cow or bull. The recent excavation also found two cremations with radiocarbon dates in the Roman Iron Age. The burials investigated in 2012 were directly associated with the stone circle. Two inhumations were located beside the standing stones, and the fillings of the graves incorporated flakes taken from the monument. There was also a cremation burial in the exact centre of the circle, covered by a setting of rounded cobbles. Chapter 4 publishes the results of re-excavation at Croftmoraig stone circle (also known as Croft Moraig), originally investigated by Stuart Piggott and Derek Simpson in 1965. The work was undertaken to test a new interpretation of the sequence and chronology of the site proposed by the writer and Alison Sheridan in 2005. The

Preface main aims of this work were to reinterpret the excavated structures in the light of the project archive and to collect samples for AMS dating. This was important as the original excavators (followed by subsequent writers) suggested that it was one of the oldest stone circles in Britain and replaced an even earlier timber circle. The new work shows that the post setting was actually a Middle Bronze Age ring ditch house, erected within an existing stone circle. In a subsequent phase the position of the wooden building was marked by an oval setting of standing stones, and between 1400 and 1300 bc the entire monument was enclosed by a stone wall. Like other monuments of its kind, Croftmoraig saw the deposition of cremated bone during a secondary phase between 1250 and 1000 bc. The wider significance of the revised sequence is considered in Chapter 10. Chapter 5 considers the site of a small stone circle enclosed by an earthwork ‘henge’ on the Hill of Tuach at Kintore. The site is interesting as it combines the attributes of the last stone circles, as defined by Aubrey Burl, with those of the latest henge monuments in Scotland. It was additionally important in view of the large number of developer-funded excavations in the surrounding area which provided evidence of other earthworks and a series of Bronze Age and Iron Age settlements. The site at Tuach had been damaged by excavation in 1855, but it was possible to recover the plan of a slightly oval setting of six monoliths aligned towards the southwest where it faced the flank of an older long barrow. Combined with the results of the nineteenth century project, the excavation suggested that the stone setting enclosed the position of at least a dozen cremation burials associated with Collared Urns, Cordoned Urns and, most probably, three bronze artefacts. They were in two distinct zones, one against the inner edge of the upright stones and the other at the centre of the site. They are dated between 1900 and 1600 bc. The ‘henge’ that enclosed the monoliths proved to be unexpectedly massive. It shared the alignment of the stone circle and probably had a blocked entrance. Its bank was revetted on the inside by a kerb and has a terminus post quem of 1050–850 bc. The older monument was enclosed during a period when part of the surrounding area was occupied by roundhouses of the same date. Chapter 6 concerns a small ring cairn at Laikenbuie in Inverness-shire. It was located on the edge of an extensive distribution of clearance cairns, although they remain largely undated. The ring cairn shares several structural features with the Bronze Age monuments – its alignment, the disposition of the kerbstones, and

ix

the use of coloured raw materials – but dates from the Early Iron Age. It seems to mark the end of a long-lived architectural tradition. If Part 1 considers the local settings of these five monuments, Part 2 discusses the wider implications of these projects in relation to some broader themes. It relates the excavated sites to other structures in Britain and Ireland. Chapter 7 brings together the results of these five projects. It puts forward a chronology for the construction and primary use of stone circles, particularly the Chalcolithic and Bronze Age examples which have received less attention in recent years. There were important changes over time as monuments became smaller and played an increasing role in the treatment of the dead. Certain features persisted over a very long period, including their association with cremation burials or pyres, the grading of the monoliths by height, their orientations, and the choice of stones according to their colours and textures. Chapter 8 considers the reuse of stone circles long after they were built. This takes many forms, but four elements seem to have been especially significant. There are signs that many monuments were reused for the deposition of cremated bone during the Middle and Late Bronze Ages. A few monuments, like Hillhead, also contained pyres. Some of these sites saw the erection of further stone settings (for instance Croftmoraig), and others were enclosed (the Hill of Tuach and Croftmoraig). A second development saw the building of roundhouses inside the older circuit at a time when the monoliths remained in place. This occurred between the Middle Bronze Age, where it is evidenced at Croftmoraig, and the Early Iron Age, when it happened at Strichen. There are other cases in which timber buildings of similar form were constructed inside, or near to, older monuments. Thirdly, the excavation at Waulkmill identified a stone circle that was reused as a cemetery during the Roman Iron Age. This was not an isolated instance and the chapter discusses other stone monuments that underwent a similar transformation. Lastly, there is a little evidence that prehistoric standing stones and related structures took on new significance during the Pictish period, and this is also discussed. Chapter 9 considers four neighbouring stone circles in Aberdeenshire, three of which have been excavated by the same team since 1999. The unexpected discovery of a recumbent stone circle at Hillhead, where the earthwork monument had been identified as an Iron Age roundhouse, makes it possible to compare a series of monuments

x

Preface

(within sight of one another), all of which are assigned to the same ‘type’ – the recumbent stone circle. They do share features in common, but at the level of detail that only excavation can provide they also contrast in their architecture, their use of raw materials, associated artefacts and structural sequences. The fourth site is the Blue Cairn whose surface remains are also considered. What lessons can be learned from so much variation across one small area, and what do these differences imply about the ways in which they were used? Chapter 10 returns to Croftmoraig: a site which can be studied in exceptional detail in the light of two excavations, field survey, petrological analysis, and the archive from the 1965 fieldwork. It sheds important light on a series of neglected issues in the study of stone circles. The monument was built on a natural mound which had been sculpted for the purpose. In its surface was a glacial erratic which would have drawn attention

to the place because of its exceptional form and colour. These features were directly in line with the midsummer sunset behind Schiehallion, one of the highest and most conspicuous mountains in the southern Highlands. That phenomenon can still be observed today. The successive stone settings were laid out around that central point, yet the existence of the erratic is hardly mentioned in the report on the first excavation. New work on the site archive also sheds light on the grading of the stones by height and the orientation of the inner setting of monoliths towards the midwinter sunset. All these features are emphasised by the selection of distinctive raw materials for use in different parts of the monument, and even by the distribution of cup marks at Croftmoraig. Thus this chapter is a study of the anatomy of a wellknown stone circle, and the new interpretation of Croftmoraig draws attention to ways of thinking about these monuments which have still to fulfil their potential.

Acknowledgments

Most of the fieldwork reported here took place over a concentrated period between 2011 and 2013 and involved many of the same people. Ronnie Scott and Nigel Healy had worked at Laikenbuie between 2003 and 2006 and had taken part in earlier projects on recumbent stone circles and henge monuments. Ronnie Scott, Aaron Watson, Rosemary Stewart and Alex Brown were also involved in the investigation of rock art on the Ben Lawers estate, not far from Croftmoraig. Amanda Clarke co-directed three of the excavations – at Tuach, Waulkmill and Hillhead – with Richard Bradley. For permission to undertake this work we must thank Malcolm Allen (Tuach), Athel Price (Croftmoraig), Simon Power, Rab Blackall, the McRobert Trust (Waulkmill and Hillhead), and Peter Muskus (Laikenbuie). The Hill of Tuach and Croftmoraig are Scheduled Ancient Monuments and we are grateful to Historic Scotland for allowing work to take place there. We are grateful to Martin Brann and Oliver Lewis for their help and support. The project at Tuach was funded by the British Academy and that at Hillhead by Aberdeenshire Council. The Waulkmill excavation received invaluable help in kind from the McRobert Trust and the work at Croftmoraig was funded from prize money from the Prehistoric Society. Radiocarbon dating was funded by Historic Scotland (Croftmoraig), National Museums Scotland (Croftmoraig and Tuach), Aberdeenshire Council (Hillhead and Waulkmill) and the Nairn Discretionary Fund of Highland Council (Laikenbuie). Aberdeenshire Council also funded the conservation of the metalwork from Waulkmill and the publication of colour photographs in this book. Unless otherwise stated, the illustrations in this account are by Sarah Lambert Gates (Chapters 3–5), or Alessandro Guaggenti who is responsible for the drawings in the remaining chapters. We must also thank

Alan Braby, Marion O’Neil and Craig Williams for their contributions. Nearly all the photographs are by Aaron Watson or Richard Bradley – other images are credited in the captions and we must thank the people who provided them. Many people contributed their expertise to the analysis of the excavated material and we owe a special debt of gratitude to Fraser Hunter, Fiona Shapland and Alison Sheridan for providing their contributions when they had far more pressing concerns. We received valuable advice from many colleagues who visited us in the field or commented on drafts of this volume: (in alphabetical order) Trevor Cowie, Peter Craig, George Currie, Hilary and Charly Murray, Brendan O’Connor, Rick Schulting, Dougie Scott, Alison Sheridan and Adam Welfare. The early stages of the work were discussed with Moira Greig and the late Ian Shepherd and its subsequent development has greatly benefitted from discussion with Bruce Mann, David Strachan and Rod McCullagh. Many of the team who participated in these projects are local to the area and their involvement made the research all the more involving. They include: (again in alphabetical order) Diane Collinson, Maria Cowie, Sheila Duthie, Nigel Healy, Moyra Simon, Colin Mitchell, Irvine Ross, Jane Summers and Sheila Young. Ronnie Scott played a vital part in the work. Apart from working on all the sites, he stored the finds from these projects and undertook his own excavation at Laikenbuie. Among those who came from a greater distance were Elise Fraser, Alice Rogers, Jon Tierney and Aaron Watson. Nothing would have happened without the skill, enthusiasm and hard work of all these people. Last but not least we must thank Mabel Lawson who lives at Hillhead for sharing her recollections of her childhood there when she played inside a stone circle whose remains were nearly lost two years ago. Richard Bradley, Amanda Clarke and Courtney Nimura

List of Figures

0.1. 1.1. 1.2. 2.1. 2.2. 2.3. 2.4. 2.5. 2.6. 2.7. 2.8. 2.9. 2.10. 2.11. 2.12. 2.13. 2.14. 2.15. 2.16. 2.17. 2.18. 2.19.

Locations of the five excavations published in this study Varieties of circular monuments in northern and northeastern Scotland The distributions of Clava Cairns, recumbent stone circles and henge monuments in northeast Scotland and along the inner Moray Firth The locations of the three excavated monuments in the Howe of Cromar in relation to the areas investigated by fieldwalking The siting of the four stone circles in the Howe of Cromar, showing the extent of high ground and the area that can be seen from Hillhead Outline plan of Hillhead showing the principal features, the extent of excavation and the locations of the published sections Kite photo taken from the east showing forestry trenches and exposed rubble at the start of the 2013 excavation Trench 1 showing the positions of the internal wall, the external kerb and the fallen monolith Photogrammetric survey plans of the rubble of the ring cairn in Trench 1 Section of the ring cairn in Trench 1 at Hillhead Section 1 through the eastern perimeter of the ring cairn The primary wall on the east side of the ring cairn Detail of the inner kerb in Trench 4 showing a deposit of flat slabs interpreted as a shallow step The inner kerb in Trench 4 showing how it has been pushed outwards by the mass of the ring cairn Section 2 through the western perimeter of the ring cairn The remains of the ring cairn in Trench 3 Section 2 of the ring cairn on the west side of the circuit The outer kerb of the ring cairn viewed from the east Detail of the outer kerb on the eastern side of the monument, showing the remains of two kerbstones and the scarp left where others were removed Trench 1 showing a fallen monolith Photogrammetric survey plan of the rubble of the ring cairn in Trench 3 viewed from the north The hollow in between the terminals interpreted as the position of the recumbent stone

2.20. The stone-packed hollow interpreted as the socket from which the northern flanker had been removed 2.21. A broken monolith interpreted as part of the southern flanker 2.22. A broken monolith interpreted as part of the northern flanker 2.23. Part of the void left marking the socket for the southern flanker, masked by the surviving part of the recumbent stone 2.24. Secondary features in the centre of the monument 2.25. The area of heat-affected subsoil in Trench 1, and the slab-lined hearth or fireplace which contained a dense deposit of charcoal 2.26. The density of broken quartz inside the monument at Hillhead 2.27. The location of the Hillhead monument in relation to the finds of worked flint and quartz revealed by tree planting 2.28. Beaker sherds and a barbed and tanged arrowhead found together in a forestry trench 16 m south of the monument 3.1. The position of the monument at Waulkmill seen from the Tomnaverie recumbent stone circle 3.2. The relocated monolith at Waulkmill, with the position of the stone circle under excavation in the background 3.3. The site of the stone circle under excavation seen from the north 3.4. The location of Waulkmill stone circle in relation to the local topography 3.5. The miniature cauldron, penannular brooch, iron dagger and gaming pieces found in the sandpit at the end of the nineteenth century 3.6. An earlier twentieth century postcard of Tarland 3.7. Plan of the excavated features at Waulkmill 3.8. Reconstructions of Early Neolithic Carinated Bowl Pots 1 and 2 from Waulkmill 3.9. The surviving remains of the stone circle 3.10. The partly excavated foundation trench for the inner kerb, with a recut pit in the centre capped by a setting of boulders 3.11. Plan and section of the circular trench and central pit, with details of the secondary recut and the setting of quartzite cobbles

xiv

List of Figures

3.12. The southern horizon from the Waulkmill monument with Mount Keen in the distance on the right 3.13. Plan and elevation of Grave 1 3.14. Vertical view of Grave 1 after excavation 3.15. Outline plan of Grave 2 3.16. Detailed plan of the lower level of Grave 2 3.17. Grave 2 showing the setting of boulders towards the left flanking the remains of the wooden coffin 3.18. The first indication of a coffin stain within Grave 2 3.19. Two spiral rings and the stain left by the animal skull in Grave 2 3.20. The mica schist counter and brooch from Grave 1, and the rings from Grave 2 3.21. The penannular brooch from Grave 1 3.22. The silver penannular brooch found in 1898 3.23. The dagger found in the late nineteenth century 3.24. The gaming counters from 1898 3.25. The stone ‘gaming pieces’ found in 2012 at Waulkmill 4.1. The location of Croftmoraig in relation to local round barrows and the Rivers Tay and Lyon 4.2. The Croftmoraig stone circle viewed from the southeast with the summit of Schiehallion in the background 4.3. The layout of the different circuits at Croftmoraig in relation to the positions of the portal stones and the decorated slab 4.4. The stone circle viewed from the south 4.5. The published plan of the 1965 excavation 4.6. General view of the 1965 excavation from the south 4.7. A detail of the ‘ditch’ and postholes excavated in 1965 4.8. The relationship between Monolith 20, the ring ditch and associated postholes 4.9. The traces of the ring ditch roundhouse together with the positions of standing and fallen monoliths 4.10. The published plan of the 1965 excavation, with the positions of the trenches excavated in 2012 4.11. Sections of the monument during the 1965 excavation 4.12. Plans and sections of Trenches 1 and 2 4.13. The stone foundation for the enclosure wall viewed from the interior of the stone circle with Monolith 1 in the foreground 4.14. The rubble foundation of the enclosure wall, set in a shallow trench on the west side of the circuit 4.15. Trench 2 at the 2012 excavation showing an unexcavated baulk from the 1965 project and the position of the central erratic in the foreground 4.16. The glacial erratic at the centre of the stone circle showing where large flakes had been detached in the past 4.17. The centre of the stone circle during the first project,

4.18. 4.19. 4.20. 4.21. 4.22. 4.23. 4.24. 5.1. 5.2. 5.3. 5.4. 5.5. 5.6. 5.7. 5.8. 5.9. 5.10. 5.11. 5.12. 5.13. 5.14. 5.15. 5.16. 5.17. 5.18. 5.19. 5.20. 5.21. 5.22. 5.23.

showing the glacial erratic and evidence that people had dug around it in the past Plans and sections of Trenches 3 and 4 Trench 3 showing the positions of two of standing stones and the baulks left between the trenches excavated in 1965 Trench 4 showing the remains of the perimeter wall on the south-east side of the enclosure (a) The phasing favoured by Piggott and Simpson 1971; (b) the revised phasing favoured in this study Stone blocks in the enclosure wall on the south-east side of the enclosure Stone blocks in the enclosure wall The relationship between Monolith 4 in the outer circle and the northern post trench for the timber porch The location of the Hill of Tuach in relation to the positions of nearby monuments A reconstruction of the original monument at Tuach, drawn from earlier surveys The surviving part of the Midmill long cairn The position of the Tuach stone circle seen from the Midmill long cairn A spread of granite flakes overlying the filling of the 1855 excavation The surviving remains of the stone circle and henge monument, showing a profile of the earthwork and the extent of the 2011 excavation General view of excavation in between the growing trees Section of the bank and ditch The enclosure ditch looking into the interior of the monument General view of the enclosure earthwork seen from the interior of the monument The surviving part of Stone 4 overlying the filling of the enclosure ditch on the southwest perimeter of the monument Plan of the excavated features at Tuach Excavated pits in the interior of the enclosure at Tuach Urn 4 (left) on its discovery, and Urn 5 (right) showing the cap of redeposited clay over the inverted vessel Urn 1 Urn 2 Urn 4 Urn 5 The fragmentary razor found in 2011 inside Urn 5 The razor sheath Bone toggle associated with Urn 1 excavated in 1855 The flint artefacts from the excavation Pollen analysis for the buried soil

List of Figures 5.24. The layout of the Early Bronze Age stone circle (above); The layout of the site in the Late Bronze Age (below) 6.1. Survey of the cairnfield, showing the positions of the excavated monuments 6.2. General view of the cairnfield from the north with the boundary banks visible towards the right 6.3. Rubble overlying the earthen mound (3) 6.4. Section of the mound (3) and possible tree hole 6.5. The remains of the ring cairn before excavation, looking south 6.6. Outline plan of the ring cairn and elevation of the kerb on its southern perimeter 6.7. The surface remains of the ring cairn (1) and two profiles across the monument 6.8. The extent of excavation and details of the excavated features 6.9. A deposit of rubble outside the kerb of the ring cairn on the northeast side of the monument 6.10. Detail of the kerbstones on the southern perimeter of the monument 6.11. The outer perimeter on the southern side of the monument with the packing of rubble behind it 6.12. Section of the monument showing the deposit of rubble retained by the boulder kerb 6.13. Excavating the interior of the ring cairn, with the kerbstones to the left 6.14. Section of the ring cairn 6.15. The group of white stones found together just outside the kerb on the east side of the monument 6.16. Six flaked discs or ‘pot lids’ found on the surface of the ring cairn 7.1. Two chambered cairns with southwestern alignments, external platforms and stone circles 7.2. Outline plans of Balnuaran of Clava southwest cairn and Tomnaverie summarising the links between the monoliths and the kerbstones according to materials, textures and colours 7.3. The structural sequences at Balnuaran of Clava northeast cairn and Tomnaverie 7.4. Four Scottish monuments with blocked entrances or other features focusing on the south or southwest 8.1. Secondary cremation pyres at the Hillhead and Old Keig stone circles 8.2. Outline plan of Moncreiffe with details of the two trenches containing burnt material and cremated bone 8.3. The evidence for later timber buildings inside the stone circles at Croftmoraig and Strichen

8.4.

xv

Plans of a recumbent stone circle and a large roundhouse at the Candle Stane 8.5. Distribution of earlier prehistoric monuments with evidence of Roman Iron Age reuse 8.6. Outline plans of the settlements at Old and New Kinord in relation to the surviving structure at Melgum 8.7. The sockets of the Waulkmill stone circle and Graves 1 and 2 in relation to the Melgum settlement 8.8. Distribution of prehistoric monoliths and other structures apparently reused in the Pictish period 8.9. The Pictish symbol stone from Dingwall showing later motifs apparently superimposed on a distribution of cup marks 9.1. Distribution of ring cairns with and without stone circles in north-east Scotland (above); Outline plans of the recumbent stone circle at Tilliefourie and the ring cairn at Milestone (below) 9.2. The stone circles in the Howe of Cromar discussed in this chapter 9.3. Outline plans of four stone circles in the Howe of Cromar 9.4. Outline plans of two monuments in northeast Scotland with large open courts at their centre 9.5. Two pairs of stone monuments, which may have been used in sequence 10.1. Croftmoraig stone circle viewed from the northeast 10.2. The small stone circle at Killin close to the west end of Loch Tay 10.3. The siting of Croftmoraig in relation to Loch Tay and Schiehallion 10.4. The stone circle viewed from the northeast, emphasising the profile of the mound on which it was built 10.5. The Machuim stone circle built on an artificial mound on the north side of Loch Tay 10.6. The profile of Schiehallion which appears on the horizon viewed from the stone circle 10.7. A sequence of photographs showing the position of the midsummer sun as it sets behind Schiehallion 10.8. The position of the midwinter sunset as viewed from Croftmoraig 10.9. Detail of the banded erratic at the centre of the stone circle 10.10. The heights of the monoliths at Croftmoraig based on the surviving structure and photographs in the 1965 excavation archive

List of Tables

2.1. 2.2. 3.1. 3.2. 3.3.

Identified human bone fragments from Context 1007 Identified human bone fragments from Context 1021 Identified human bone fragments from Context 1031 Pre-Roman and Roman Iron Age penannular brooches from Scotland divided by different site types Catalogue of the Waulkmill gaming counters from 1898 (Group 1)

3.4 5.1. 5.2. 5.3

Catalogue of finds from the Waulkmill excavation 2012 (Group 2) Identified human bone fragments from the Collared Urn (Urn 4) Identified human bone fragments from the Cordoned Urn (Urn 5) Charcoal identifications

Abbreviations

Canmore www.canmore.rcahms.gov.uk NESARS North East Scotland Archaeological Research Society NMRS National Monuments Record of Scotland NMS National Museums Scotland RCAHMS Royal Commission on the Ancient and Historical Monuments of Scotland

part one Excavations at Five Scottish Monuments

chapter one

The Development of the Project

Richard Bradley

Between 1994 and 2008, together with several colleagues, I was involved in a series of excavations at Chalcolithic and Early Bronze Age monuments in northern and north-eastern Scotland. The forms of these structures were closely related to one another, and so at times were their histories (Fig. 1.1). The project took place in Inverness-shire, Caithness and Aberdeenshire and considered three traditions of prehistoric architecture: Clava Cairns, recumbent stone circles and henges (Fig. 1.2). This activity involved four seasons of excavation and survey around Balnuaran of Clava, two seasons at Tomnaverie and three at Broomend of Crichie. The results of this work were supplemented by smaller projects at Cothiemuir Wood, Aikey Brae and Pullyhour. The publications arising from this fieldwork also included the reports on excavations undertaken by other people at Newton of Petty, Strichen, Lairg and Migdale. Three volumes were published, each of them focusing on one of the main architectural traditions. The Good Stones discussed Clava Cairns (Bradley 2000); The Moon and the Bonfire considered recumbent stone circles (Bradley 2005); and Stages and Screens offered an interpretation of henge monuments (Bradley 2011). None of these monographs was completely selfcontained, for the three traditions overlapped. Both Clava Cairns and recumbent stone circles feature rings of uprights which are graded by height towards the south or southwest. Each type encloses a cairn and is associated with the use of coloured stones, deposits of quartz, and pecked decoration (usually cup marks). Nor are they entirely separate from the henges found in the same areas. Some share their characteristic orientation, and the comparatively large example at Broomend of Crichie contained a setting of monoliths. It was close to a recumbent stone circle and

the two structures were connected to one another by an avenue. In the same way, there had been a standing stone in the entrance of the Migdale henge. In fact the overlap between these structures goes even further, for earlier examples of all three kinds were reused during the Middle and Late Bronze Ages and were associated with cremation burials. Such observations required more investigation. Three specific problems remained after the results of these projects were published. They are considered now. Other important issues emerged during the course of this work. Henges and stone settings The first issue was the relationship between the earthworks known as henges and the stone settings sometimes found inside them. The problem was addressed at Broomend of Crichie, but that excavation raised almost as many questions as it answered. Although the monument had been described as an enclosed stone circle, there were two difficulties to be addressed. The first is that the earthwork was later in date than the setting of monoliths inside it. The other was that the setting was not a circle in the first place – it was an arc of upright stones flanking the position of a shaft grave, whose position had been marked by an unusually tall pillar. At the same time, excavations at Pullyhour, and to some extent at Lairg, suggested the existence of a distinctive group of circular embanked and ditched enclosures which were smaller and later in date than Broomend of Crichie. This echoed Aubrey Burl’s suggestion, based on a small body of excavated evidence, that in Scotland stone circles of similar proportions were among the last to be built. In contrast to those in lowland Britain, their chronology might have extended into the

1.  The Development of the Project

3

Figure 1.1. Varieties of circular monuments in northern and north-eastern Scotland. Information from Bradley 2000; Bradley 2009; Welfare 2011.

Middle and Late Bronze Ages (Burl 2000, xiv–xv). Until 2011 each line of argument had been pursued separately. What was needed was new fieldwork on sites where these elements were combined. That requirement was met by two of the structures considered here. The unusual site on the Hill of Tuach (or Tuack) was discussed in Stages and Screens and is located not far from Broomend of Crichie, but it was not possible to excavate there until that volume was complete – in fact the book was published while the new

excavation was in progress. It re-examined a monument first investigated in 1855 which seemed to combine the attributes of later henges, like those at Lairg and Pullyhour, with those of the last stone circles previously identified by Burl. The work had the added advantage that antiquarian accounts suggested that these structures had been associated with burials similar to the urned cremations dated to a late phase at Broomend of Crichie. The other monument which fulfilled this requirement is much better known. This is the stone circle of

4

Richard Bradley

Figure 1.2. The distributions of Clava Cairns, recumbent stone circles and henge monuments in north-east Scotland and along the inner Moray Firth.

Croftmoraig (otherwise Croft Moraig) on Tayside which was first excavated by Stuart Piggott and Derek Simpson in 1965 (Piggott and Simpson 1971). In this case the earlier project had been properly published and there was no doubt that the monument had a lengthy history. For the original excavators it began with an Early Neolithic timber circle, which was replaced in successive phases by an oval setting of monoliths and then by a more substantial stone circle provisionally dated to the Early Bronze Age. The entire monument was enclosed by some kind of rubble bank. As the excavators acknowledged,

that sequence was merely their preferred hypothesis, and other versions were possible. In parallel with the excavation at Broomend of Crichie, Alison Sheridan and the writer suggested an alternative interpretation of Croftmoraig and a new chronology for the site. The stone circle appeared to be the oldest feature, and the timber setting came later. In turn it was replaced by a small oval stone setting which recalled the characteristics of the later stone circles of Tayside. It seemed possible that most of the ‘Neolithic’ pottery published in 1971 in fact dated from the Middle or Late Bronze Age. It was only in the

1.  The Development of the Project final phase that the monument was enclosed by any kind of earthwork (Bradley and Sheridan 2005). The new interpretation was controversial and it is right to acknowledge that it was not accepted by some of the most influential researchers in this field (Gibson 2010a, 69; Welfare 2011, 261). The question was important, for it raised almost the same issues as the excavation at Tuach. Were small stone circles still being built in the later Bronze Age, and could they have been enclosed at that time? Was the outer perimeter of Croftmoraig related to the earthworks described as henges – a suggestion first made by Piggott and Simpson – and was the timber circle necessarily an early feature, as they had supposed? That last question gained a new relevance with the discovery of another timber circle outside the south entrance of Broomend of Crichie. In this case it was the latest prehistoric feature at the site and dated from the end of the Early Bronze Age – the same period as the earthworks at Pullyhour and Lairg. The only way to resolve this problem was to carry out small-scale fieldwork at Croftmoraig with the object of recovering samples suitable for AMS dating: a technique which was not available to the previous excavators. The limits of classification The three monographs published between 2000 and 2011 shared a common feature, for none was confined to new work at a single monument. Stages and Screens published excavations at Balnuaran of Clava, but also included the report on an earlier project at Newton of Petty. The main focus of The Moon and the Bonfire was fieldwork at Tomnaverie, but its results were compared with those of smaller projects at two other recumbent stone circles: Cothiemuir Wood and Aikey Brae. In the same way, the large-scale work at Broomend of Crichie was supplemented by another excavation at Pullyhour, and Stages and Screens also reported earlier projects at Migdale and Lairg. Each of the books compared the results of recent work with those of previously published excavations. There was some evidence of changes over time, particularly in the case of henge monuments, but these projects raised another issue which had been addressed by field survey but not by excavation. How much variation existed between monuments of the same type? That was especially important in the case of Clava Cairns and recumbent stone circles which may have been built simultaneously in adjacent areas: the inner Moray Firth in one case, and north-east Scotland, in the

5

other. Both groups of monuments occurred with some frequency over a limited area, but it was by no means clear whether nearby examples had taken quite the same forms. For example, the recently excavated stone circle at Cothiemuir Wood was very different from its neighbour at Old Keig, excavated by Gordon Childe in the 1930s (Childe 1934). Attempts to compare the structural sequences at the two sites have been controversial. For these reasons it seemed important to undertake excavation at two neighbouring monuments. That was possible in the Howe of Cromar where the recumbent stone circle at Tomnaverie had been excavated on a large scale in 1999 and 2000. In 2012 a smaller project located its neighbour at Waulkmill where the monument had been levelled in the early nineteenth century. The original objective was to compare these structures with one another. Ironically, a third monument of the same kind was damaged during tree planting at Hillhead in 2012 and partly excavated in the following year. As a result, Chapter 9 can compare three neighbouring circles which appear to belong to the same ‘type’. Patterns of later reuse Our excavations between 1994 and 2008 had provided evidence for a widespread but unexpected phenomenon. The passage graves at Balnuaran of Clava contained secondary cremations dating from the Late Bronze Age, and a pit containing cremated bone was excavated through the filling of an older ring cairn at Newton of Petty. At Tomnaverie, the central area of the recumbent stone circle was associated with a large deposit of burnt bone which accumulated a thousand years after the construction of the stone circle. A striking variant of this practice was identified at Aikey Brae. In this case no human remains were discovered, but the perimeter of an older stone circle may have been rebuilt at this time. The direct dating of cremated bone has resulted in the identification of other monuments with similar evidence of reuse. With improved understanding of the ceramic sequence in Scotland the same phenomenon can be recognised at the well-known sites of Old Keig (Childe 1934) and Loanhead of Daviot (Kilbride-Jones 1935). Smaller and later ring cairns at Cairnwell (Rees 1997), Balnuaran of Clava (Bradley 2000, 43–44) and on the Sands of Forvie (Ralston and Sabine 2000, chapter 2) share structural features in common with older monuments, and during the course of this project the connection was reinforced by the excavation of a similar monument at Laikenbuie – the results are presented in Chapter 6.

6

Richard Bradley

There were other developments during the later first millennium bc. A large Iron Age roundhouse was built inside the recumbent stone circle at Strichen (Phillips et al. 2006), and another was erected beside the Candle Stane (Cameron 1999). A third example was known from surface evidence at Loudon Wood (Welfare 2011, 161). It was an open question how far the building of all these structures was influenced by a single body of ideas. Were there consistent phases of reuse at older stone monuments in Scotland? That is particularly relevant to the later Bronze Age. How does such evidence compare with what is known about structures that were newly built at the same time? In the event four of the monuments considered here – the Hill of Tuach, Croftmoraig, Hillhead and possibly Waulkmill – could have seen secondary activity. That was not the end of the sequence in Northern Britain, as the small ring cairn at Laikenbuie was built in the Early Iron Age. Developments during the project Inevitably new questions arose during the course of the fieldwork. It was already known that the reuse of Scottish stone circles continued after the end of the Bronze Age. Occasional Roman Iron Age artefacts had been found in association with older monuments, including a ring cairn at Monquitter (Anderson 1902), and a large structure of the same kind at Balnuaran of Clava was associated with a cremation burial dating from the Pictish period. It had been suggested that early medieval sculptures were erected at the sites of much older monuments, including Broomend of Crichie (Clarke 2007). What was surprising was that one of the monuments excavated during this project had also been reused during the first millennium ad. This was the stone circle at Waulkmill, which was associated with an unusual Roman Iron Age cemetery. In retrospect its discovery could have been predicted as two or three graves of this date had been discovered 350 m from the monument in 1898. Fieldwork in 2012 added another four, one of which was very similar to a burial found in the nineteenth century. The existence of these deposits was unexpected, but not unique, and it raised a new set of problems. It was tempting to associate them with a series of enormous stone roundhouses which also occur in the Howe of Cromar, but they had not been investigated in modern times.

An opportunity to remedy this occurred in the following year when a large circular structure at Hillhead was damaged by forestry. It was intervisible with Waulkmill. As a result the structure was excavated and is preserved as a public monument. Unfortunately, the identification of this feature as a large roundhouse was made when its remains were under trees and has turned out to be incorrect. Against all expectation this feature proved to be the site of yet another recumbent stone circle. There was no indication of any Iron Age activity. Parts of the monument were intact and even retained structural elements that had not been observed at stone circles since the nineteenth century. Its chronology is established by a series of radiocarbon dates, and the site is also notable for including an in situ cremation pyre and 12,250 pieces of broken quartz. The latter had its counterpart at Croftmoraig where hardly any of this material survives. This unexpected discovery has had an influence over the way in which the new programme of research is being published. When the true character of the site became apparent, reports on Croftmoraig, Tuach, Waulkmill and Laikenbuie had been drafted as journal articles. Now it was possible to embark on a more ambitious analysis, comparing the character of five different monuments. Moreover, the identification of a stone circle at Hillhead made it possible to discuss the similarities and differences between three excavated monuments in the Howe of Cromar. They were located within sight of one another and may have been used concurrently. At the same time, the detailed sequence identified at Hillhead – a site which had never been investigated before – shed new light on the other monuments. It was clear that it would make more sense to bring these reports together in a single volume rather than submit them to journals that would be issued at different times. One advantage of the new procedure is that the results of these separate projects can be compared directly with one another, and the separate discussions that had been included in different articles have been brought together to make up Part 2 of this study. This account begins with Hillhead. Somewhat ironically, it was the last to be investigated and the only monument whose excavation was not planned at the beginning of this project. Because it was well-preserved and was used over such a long period it provides a vital source of comparison with the other sites.

chapter two

Excavations at Hillhead, Tarland, Aberdeenshire: a recumbent stone circle and its history

Richard Bradley and Amanda Clarke

Cromar is a natural basin 50 km west of Aberdeen, in between the Grampian Mountains and the fertile soils extending to the North Sea. It was described by the Edinburgh antiquary John Stuart (1854, 258) in the earliest account of the archaeology of the region: The district of Cromar ... is about eight miles [12.9 km] in length from north to south, by four [6.4 km] in breadth from east to west, surrounded by hills, which isolate it from the adjoining country. It lies in between the rivers Dee and Don, at an equal distance from each ... On the south, the district is divided from Deeside by a range [of ] hills. A natural opening in this direction conducts the road from Cromar to the south.

At the time that Stuart was writing the low ground included a loch, but it has since been drained. The basin contains a variety of prehistoric monuments, including two long cairns, round cairns, roundhouses, enclosures, souterrains and a possible hillfort. In 2000, much of the cultivated land was examined by fieldwalking (Fig. 2.1). The prehistoric monument at Hillhead overlooks the lower ground which contained two other stone circles (Fig. 2.2). They include Tomnaverie, which was excavated in 1999–2000 (Bradley 2005, Chapter 2), and Waulkmill which forms the subject of Chapter 3. Another example, the Blue Cairn is on the western edge of the basin, but has not seen any modern investigation (Welfare 2011, 15). Hillhead is located in a saddle beside a long-established drove road communicating between Cromar and another basin to the east, the Howe of Cushnie. The stone circle was built at a height of 345 m and was situated on the watershed of the Dee and the Don (NJ 5072 0714). The site commands an extensive view to the south and southwest, extending as far as Lochnagar and the mountains of South Deeside. The stone circles of Tomnaverie and Waulkmill

are also visible from the monument. A remarkable feature is that the site at Hillhead was constructed at the first point where a visitor climbing the slope from the west would glimpse one of the most prominent landmarks in northeast Scotland, the summit of Mither Tap of Bennachie. That happens at the exact location of the ring cairn. The mountain can be identified from inside the monument, but any further down the slope to the southwest it is invisible. The area above the excavated site may have contained other monuments which have been located during fieldwork by Moyra Simon and Jane Summers. They include the sites of several small roundhouses, one of which dates from the Late Bronze Age. Their results are summarised in a later section of this chapter. The circle was first identified by Ken Cooper as a bank which was covered by trees when he saw it. At the time the site was known as Blackhills. He made the entirely reasonable suggestion that it was the remains of a large roundhouse like those at Old and New Kinord on the lower ground 9 km away. He also noted the presence of quartz on the site and the position of a flat stone in the centre of the enclosure. Although he published a note of his observations (Cooper 1998), Hillhead remained little known. No one seems to have been aware of its archaeological significance when it was first forested and it was only when the trees had been removed and trenches had been excavated for a second phase of planting that its potential importance was recognised by the landowner. Following discussion between the McRobert Estate and Aberdeenshire Council it was agreed that the site should be preserved and that excavation should take place there. Before excavation in September 2013 the entire area that had been planted was surveyed by a team from the North East Scotland Archaeological Research Society

8

Richard Bradley and Amanda Clarke

Figure 2.1. The locations of the three excavated monuments in the Howe of Cromar in relation to the areas investigated by fieldwalking (contours at 50 m intervals). Information from Bradley 2005.

Figure 2.2. The siting of the four stone circles in the Howe of Cromar, showing the extent of high ground (contours at 100 m intervals) and the area that can be seen from Hillhead. Analysis by Irvine Ross.

2.  Excavations at Hillhead, Tarland, Aberdeenshire

9

Figure 2.3. Outline plan of Hillhead showing the principal features, the extent of excavation and the locations of the published sections. In all the plans of the five excavated monuments the smaller arrow represents true north and the larger arrow the direction of magnetic north at the time the fieldwork took place. (NESARS) led by Moyra Simon and Jane Summers who examined the network of trenches extending across the surrounding hillside. Their work is reported on p. 23 and resulted in the identification of an extensive distribution of worked quartz. There was a smaller quantity of worked flint, and 16 m south of the ring cairn sherds of early Beaker pottery were found, together with a barbed and tanged arrowhead. They may be all that remains of a disturbed burial. Unburnt bone would not survive on such an acid subsoil. Since the monument was to be preserved, excavation was conducted on a limited scale and most of the stratified deposits were left intact. The aims of the project were:

• To determine the original form of the monument; • To collect dating evidence; and • To advise on its preservation and interpretation to the public.

The excavation investigated two quadrants, one to the northeast and the other to the southwest. In view of the size of the bank, parts of it were excluded (Figs 2.3–2.4). An additional area was opened at the centre of the monument where there was the flat stone that Ken Cooper had already noted. It proved to be part of a group of three which might have been the remains of a cist – in the event it became clear that they had been put there comparatively recently. The outer perimeter

10

Richard Bradley and Amanda Clarke

Figure 2.4. Kite photo taken from the east showing forestry trenches and exposed rubble at the start of the 2013 excavation (Nigel Healy). of the enclosure was exposed in both quadrants, but the bank was sectioned completely at only two points, one on either side of the enclosure. A further small section investigated part of the inner kerb where it was especially well-preserved in the north-eastern sector of the site. Within the excavated areas the full extent of the monument was exposed and recorded by photogrammetry, but only the latest and most vulnerable deposit – a layer of quartz close to the modern ground surface – was excavated in its entirety. After the work in 2013 the excavated area has been refilled, as have the forestry trenches cutting across the monument. An exclusion zone has been established

around the prehistoric enclosure, and this is to be maintained as grazed grassland, allowing visitors to see what remains of this unusual structure. The structure of the monument Despite the damage caused by tree planting, the monument was generally well-preserved. It had suffered from the removal of a number of kerbstones and monoliths, probably in the 1940s, but beneath the surface the cairn was largely undisturbed (Fig. 2.5). The monument seems to have developed over six structural phases.

2.  Excavations at Hillhead, Tarland, Aberdeenshire

11

Figure 2.5. Trench 1 showing the positions of the internal wall, the external kerb and the fallen monolith. The primary ring cairn The basic structure at Hillhead was a substantial ring cairn 26 m in diameter which enclosed an open court 15 m across (Figs 2.6–2.7). It was built out of large stones, many of them angular blocks, up to 60 cm in maximum dimensions, all of which could have been collected in the vicinity. They lay directly on the prehistoric land surface. To the southwest the enclosure seems to have been provided with an entrance which was 2 m wide. The area was badly disturbed, but the terminals of the bank remained largely intact. It was constructed in two stages. Where it was sectioned towards the northeast its earliest component was a stone wall with a rubble core buried beneath the later ring cairn (Figs 2.8–2.9). It was 40 cm high and 2.4 m wide. The upper surface of this structure was poorly preserved. During the excavation it was assumed that the damage was due to twentieth century tree planting, but the radiocarbon dates for charcoal on the old land surface raise the possibility that stones were being removed from the monument during the first millennium ad. On the other hand, several stone circles were reused during the Pictish period and it may be wrong to place too much weight on purely practical

considerations. This question is considered in detail in Chapter 8. It seems possible that the top of this wall was stepped. The idea was suggested by finding parallel rows of boulders at its core, and this interpretation was supported by excavating a small section through the filling of the court to expose the inner kerb where this part of the structure was better preserved (Figs 2.10–2.11). Here it was constructed of vertical slabs which had been pushed outwards by the mass of the cairn. They retained a level of flat stones. There may have been another step, largely destroyed, within the excavated area. It would have been about 10 cm high. A section across the ring cairn on the south-west side of the enclosure found little sign of a similar structure (Figs 2.12–2.14). Instead it identified a bank of rubble about 2.7 m wide and 45 cm high. It consisted of large angular and rounded stones which had been stacked directly on the old ground surface. Here rather more quartz boulders were incorporated in the core of the monument than on other parts of the site. There was no evidence of any revetment within this part of the enclosure, but a large stone-packed socket was visible in section. It suggests that part of the structure had

12

Richard Bradley and Amanda Clarke

Figure 2.6. (Left) Photogrammetric survey plan of the rubble of the ring cairn in Trench 1 viewed from the interior of the monument (Aaron Watson); (right) the same image, emphasising the unexcavated remains of the inner kerb (Craig Williams).

Figure 2.7. Section of the ring cairn in Trench 1 at Hillhead. The remains of the primary wall are towards the left and the secondary addition and outer kerb are seen towards the right (Aaron Watson).

been removed at some time in the past. In contrast to the surrounding area, there was little worked quartz underneath the bank, but a few pieces of charcoal of Corylus and Salix/Populus were found on the old land

surface and submitted for radiocarbon dating. Five samples returned dates in the later third millennium bc and should provide a terminus ante quem for the building of the monument. The dimensions of the rubble bank on

2.  Excavations at Hillhead, Tarland, Aberdeenshire

13

Figure 2.8. Section 1 through the eastern perimeter of the ring cairn.

Figure 2.9. The primary wall on the east side of the ring cairn (Aaron Watson).

Figure 2.10. Detail of the inner kerb in Trench 4, showing a deposit of flat slabs interpreted as a shallow step (Aaron Watson). the south-west side of the ring cairn were virtually the same as those of the internal wall on the opposite side of the perimeter and it seems likely that they formed part of the same circuit. Perhaps this was the oldest component of the site. Towards the northeast the main body of the ring cairn was a secondary feature abutting the outer face of the existing wall (Fig. 2.15). It had no equivalent on the opposite side of the circuit. In its final form the bank was 8 m wide and survived to a maximum height of 65 cm. Its rubble core was delimited by a kerb of upright Figure 2.11 (left). The inner kerb in Trench 4 showing how it has been pushed outwards by the mass of the ring cairn (Richard Bradley).

14

Richard Bradley and Amanda Clarke

Figure 2.12. Section 2 through the western perimeter of the ring cairn.

Figure 2.13. The remains of the ring cairn in Trench 3, showing the rubble in section in the foreground, and the hollowed area in between the terminals which had been partly disturbed by a tree stump (Aaron Watson).

Figure 2.14. Section 2 of the ring cairn on the west side of the circuit. Towards the centre a possible stone hole can be seen in section (Richard Bradley).

2.  Excavations at Hillhead, Tarland, Aberdeenshire

15

Figure 2.15. The outer kerb of the ring cairn viewed from the east showing several large slabs held in place between deposits of boulders, and the trench which remained after others had been removed (Richard Bradley). slabs up to 50 cm high, wedged between larger boulders placed on the ground. Much of this kerb had been robbed, leaving a trench up to 40 cm wide which was filled with a layer of topsoil and rubble (Fig. 2.16). Other kerbstones had fallen outwards and lay on the surface of the hill. Beyond this external kerb there was a shallow band of rubble up to 80 cm across and a single undated posthole. The north-eastern section of the ring cairn included a few large fragments of quartz, and others were found against the outer kerb where they may have eroded from the bank. By contrast, there was very little worked quartz in the buried soil, especially on this side of the monument, suggesting that the scatter of artefacts in the surrounding area must have accumulated after the cairn had been built. That is clear as its distribution extended up the hillside from the position of the stone circle. It seems that the outer kerb acted as a screen rather than a revetment, for the main mass of the cairn was provided by the remains of the existing wall. Nothing similar was found in the south-western sector of the enclosure, suggesting that only part of the original circuit had been modified. Taken together, these observations suggest that in its final form the ring cairn might have presented a slightly elliptical appearance, with a broader and higher structure on the uphill side and a less impressive boundary towards its lowest point. The exposed kerbstones suggest that it was most substantial towards the north and northeast where the surface of the structure may have included shallow steps. The

Figure 2.16. Detail of the outer kerb on the eastern side of the monument, showing the remains of two kerbstones and the scarp left where others were removed (Richard Bradley).

16

Richard Bradley and Amanda Clarke

Figure 2.17. Trench 1. The broken and fallen monolith is to the right and a displaced kerbstone to the left. They are composed of different kinds of stone (Richard Bradley).

field evidence is tenuous, but, if it has been interpreted correctly, the effect would have been like that of an open-air theatre. An audience occupying this part of the monument would have overlooked the lower ground in the Howe of Cromar and would also have seen the conspicuous outlines of Morven and Lochnagar on the horizon. The south-western perimeter of the enclosure would not have obstructed the view, as it was comparatively low. Evidence of a stone circle Even before the excavation commenced it was clear that the surviving cairn had been accompanied by a ring of standing stones. A local resident, Mabel Lawson, has a clear memory of playing there as a child in the 1940s and even remembers when the circle was dismantled to provide material for a wall. We are most grateful to her for this information. The 2013 excavation confirmed her recollection.

On the north-east side of the ring cairn just beyond the outer kerb there was a fallen monolith, 35 cm long, composed of a distinctive piece of magmatite (Fig. 2.17). It had an unusual texture because the rock had melted to a greater extent than the kerbstones composed of the same material. It could have been brought from nearby (we must thank Peter Craig for this information). It is not certain whether the monolith was complete, but its original position seems to be indicated by a shallow hollow 5 cm deep in the natural subsoil. It was accompanied by the disturbed remains of what was probably a ‘packing cairn’ of the kind associated with stone circles like those at Balnuaran of Clava (Bradley 2000, 73–78). On the south-west side of the enclosure, directly opposite the fallen monolith, the bank was broken by what seems to have been an entrance 2 m wide (Fig. 2.18). The outer edge of the enclosure was marked by an oval hollow 2 m across but less than 10 cm deep (Fig. 2.19). It seemed possible that a large horizontal stone had occupied this position and had later been taken away. All that remained was the imprint left by part of its base, but the hollow that remained did not show any sign of compression. Equidistant from this feature there were two holes edged by large boulders, 65 and 75 cm in maximum dimensions respectively (Fig. 2.20). Both of them were 45 cm deep. They were entirely empty but their positions had been masked by surface debris. Both are very like the sockets for the flankers at Tomnaverie (Bradley 2005, 24). Assuming the comparison is valid, it would suggest that similar features had been removed in modern times. If so, the south-western limit of the site included a recumbent stone approximately 3 m in length flanked by two standing stones. There are several arguments in favour of this interpretation. The putative recumbent stone is exactly in line with the fallen monolith outside the kerb to the northeast and also with the centre of the ring cairn. As is common in Aberdeenshire, it is located on the south-west side of the circle. Seen from the highest section of the bank there would have been a view between the flankers towards a cleft on the far horizon 15 km away. It was formed by Lochnagar in the background and by a conspicuous spur running south from Morven. The recumbent stone circle at Tomnaverie was also orientated on Lochnagar. There may be other evidence of the ‘missing’ stones. Close to each of the sockets attributed to the flankers are two substantial pieces of magmatite containing prominent bands of schist; another was recognised in the inner kerb (Figs 2.21–2.22). The two fragments

2.  Excavations at Hillhead, Tarland, Aberdeenshire

17

Figure 2.19. The hollow in between the terminals interpreted as the position of the recumbent stone. Here half of it has been excavated (Richard Bradley).

Figure 2.18. (Above) Photogrammetric survey plan of the rubble of the ring cairn in Trench 3 viewed from the north, showing the positions of part of the recumbent stone and one flanker towards the left, and traces of a second flanker to the right (Aaron Watson); (below) the same image, emphasising the positions of the bank terminals (1, 2), stone sockets (3, 5) and the hollow (4) left by the recumbent stone (Craig Williams). Scale bar 1 m. could have formed a pair, and their proportions raise the possibility that they were the base of one flanker and the upper part of another. In between them was a substantial fragment of a pink rock, identified by Peter Craig as felsite which had fractured along a series of horizontal bedding planes. This kind of rock could have been found locally, but it is the only substantial

Figure 2.20. The stone-packed hollow interpreted as the socket from which the northern flanker had been removed (Richard Bradley). piece from the excavation. It may have formed part of a recumbent stone, broken and largely removed when the cairn was robbed for building material. It overlay one of the features identified as a stone socket (Fig. 2.23). In the centre of the monument were two more fragments of magmatite like those used for the putative flankers. Excavation showed that they did not form part of a cist. In fact they overlay the latest filling of the ring cairn and fitted together to form a monolith which would have been more than 2 m high. This may have been another component of a stone circle, but, of course, its original position is unknown.

18

Richard Bradley and Amanda Clarke

Figure 2.22. A broken monolith interpreted as part of the northern flanker (Richard Bradley).

Figure 2.21. A broken monolith interpreted as part of the southern flanker (Richard Bradley). There is a little evidence of sequence. There is no way of establishing the relationship between the fallen standing stone and the north-western sector of the ring cairn, but the holes left by the putative flankers did not penetrate the old ground surface. If these features have been interpreted correctly, the monoliths must have been bedded in the edge of the ring cairn and would have been placed there after its construction.

Figure 2.23. Part of the void left marking the socket for the southern flanker, masked by the surviving part of the recumbent stone (Richard Bradley).

The secondary ring cairn The next development was the capping of the entire monument by a layer of rounded boulders, most of them between 10 and 20 cm in maximum dimensions. According to Peter Craig, they could have been collected close to the site, but did not come from the same deposit as the larger blocks used to build the original structure. They would have been found further downslope than the other raw material. They were various colours, from white to red, but there was no evidence of patterned stonework of the kind identified at Tomnaverie. At Hillhead, a single deposit covered the full extent of the

primary cairn and filled the court to the top of the rubble wall which was left unaltered. This material sealed a small deposit of Corylus charcoal in the bottom of the court. It provides a terminus post quem for the covering of the primary monument but actually predated the samples from the land surface beneath the cairn. Together with a date of 2461–2207 bc from the buried soil, they suggest a period of activity between about 2580 and 2300 bc before the monument was built. The question is considered further on p. 139.

2.  Excavations at Hillhead, Tarland, Aberdeenshire

19

Figure 2.24. Secondary features in the centre of the monument. Secondary settings of boulders and a cremation pyre At the exact centre of the monument three small features cut through the filling of the court (Fig. 2.24). There may well have been more beyond the area investigated. Two shallow pits, each of them 10 cm deep, were lined by medium-sized boulders and completely filled with charcoal. One provided a radiocarbon date of 1409–1302 bc. Next to them, in the middle of the ring cairn, was a larger area over which the natural ground surface had been scorched. The burnt soil was associated with a deposit of charcoal, but in this case it was mixed with a quantity of cremated human bone which had reached a temperature of over 900°C. The evidence for in situ burning suggests that this was the position of a cremation pyre (Fig. 2.25). More burnt bones were scattered over the land surface around it. The charcoal associated with all these deposits originated from the same two species – alder and oak. According to Fiona Shapland’s report, each of these deposits may have contained the remains of only one person. A sample of burnt bone from the feature identified as a pyre has a radiocarbon date of 1214–1005 bc. The capping of worked quartz The three features just described were covered by a thin layer of unburnt boulders which were smaller than those used to fill the court. It was sealed by a continuous scatter of broken quartz (Fig. 2.26). Most of it was white, but a small proportion was pink in colour, perhaps reflecting the relative frequency of these types on the hillside. Its quantity decreased with depth, but this deposit extended across the whole of the interior of the monument. On the other hand, it was uncommon on top of the bank. It was the latest deposit at Hillhead, and parts of it had been brought to the surface in the course of tree planting.

Figure 2.25. To the left is the area of heat-affected subsoil in Trench 1 interpreted as the position of a pyre, and to the right is a slab-lined hearth or fireplace which contained a dense deposit of charcoal (Richard Bradley).

For that reason it was excavated in its entirety. When the finds were recorded on a two-metre grid it soon became clear that their density increased steadily towards the centre of the monument and the southwest. There were 12,250 pieces in all but only five worked flints in the same context. Unworked quartz is ubiquitous around the monument but many of these fragments had been deliberately fractured. Since they overlay the remains of the pyre the radiocarbon date for the cremated bone from that feature provides a terminus post quem of 1214–1005 bc for this deposit. Dating evidence The earliest dates from Hillhead are between 2460 and 2200 bc and probably predate the building of the monument (for greater clarity the dates quoted in this section have been rounded to the nearest decade). The material of the ring cairn had been placed directly on the old land surface. Samples from this level on the south-west perimeter of the monument provided dates of 2290–1980 bc, but those in the north-east section must be intrusive. A tiny sherd in the same fabric as the Beaker pottery found in fieldwalking also came from this level in Trench 1. There is little to show when the original structure was capped with glacial boulders, but charcoal samples from the surface of the court beside the inner kerb provide a terminus post quem for that phase. There were also samples of charcoal and cremated bone from the features cut through the centre of the monument. In

20

Richard Bradley and Amanda Clarke

Figure 2.26. The density of broken quartz inside the monument at Hillhead. Unstratified material is not included. this case they provide a terminus ante quem of 1500–1300 bc for the filling of the court and a terminus post quem of 1210–1000 bc for the layer of quartz which covers the interior of the ring cairn. Radiocarbon dates There are 12 radiocarbon dates from the excavation, all of them funded by Aberdeenshire Council. Eleven are on single pieces of charcoal belonging to short-lived species. They came from: the old land surface beneath the ring cairn on the north-east and south-west sides of the monument; the surface of the court at the foot of the inner kerb in Trench 4; and a pebble-lined feature in the centre of the monument which had been excavated through the filling of the court. There is also a date for a sample of cremated bone from the secondary feature interpreted as the remains of a pyre.

The surface of the old land surface on the south-west side of the monument (Context 3): SUERC-53890 Corylus charcoal. 3728±35 2027 bc at 95.4% probability SUERC-53891 Corylus charcoal. 3723±35 2024 bc at 95.4% probability SUERC-53892 Corylus charcoal. 3753±35 2037 bc at 95.4% probability SUERC-53893 Corylus charcoal. 3696±35 1977 bc at 95.4% probability SUERC-53894 Salix/Populus charcoal. BP/2461–2207 bc at 95.4% probability

BP/2275– BP/2273– BP/2287– BP/2200– 3855±35

The surface of the old land surface on the north-east side of the monument (below Context 3): SUERC-53899 Alnus charcoal. 1209±35 BP/ad 690–938 at 95.4% probability

2.  Excavations at Hillhead, Tarland, Aberdeenshire SUERC-53900 Alnus charcoal. 1180±35 BP/ad 725–967 at 95.4% probability The surface of the court below the secondary filling (Context 2) in Trench 4: SUERC-53901 Corylus charcoal. 3982±35 BP/2581–2350 bc at 95.4% probability SUERC-53902 Corylus charcoal. 3955±35 BP/2572– 2342 bc at 95.4% probability SUERC-53903 Corylus charcoal. 3935±35 BP/2565– 2299 bc at 95.4% probability Secondary fireplace or hearth in the centre of the monument: SUERC-53895 Alnus charcoal. 31415±35 BP/1499–1302 bc at 95.4% probability Secondary cremation pyre in the centre of the monument: SUERC-53904 Human bone. 2909±35 BP/1214–1005 bc at 95.4% probability Finds Pottery Richard Bradley A single tiny sherd was found on the old land surface where the ring cairn was sectioned on the north-east side of the monument. It is of a fine fabric with a buff exterior surface. The interior surface does not survive, but the core contains small pieces of grog. It can be identified as part of a Beaker vessel, but its date and form are not known. The worked stone Richard Bradley Flint Five pieces of worked flint were scattered over the disturbed surface of the monument. None was securely stratified. Two are chips or spalls and a third is a small fragment of a worked pebble. The only diagnostic items were a small flake scraper which had been completely worked down, and the bulbar segment of a small blade. Quartz and quartzite In contrast to the paucity of worked flint, pieces of smashed or flaked quartz and quartzite were very common. Because these materials occur naturally on the hillside, it is impossible to arrive at an exact figure for the number of artefacts, but their total is estimated at

21

12,250. A few pieces of imported pink quartz were also identified, and this is significant as it did not occur far outside the monument. This collection comes from four different contexts. In chronological order, they were: 1. The old land surface beneath the ring cairn, with 70 pieces; 2. The surface of the bank in the north-eastern sector, with about 120 fragments; 3. A few flakes and shattered fragments of quartz from the features at the centre of the monument; and 4. A massive deposit which covered the interior of the enclosure, increasing in density towards its centre and the south-western sector. Approximately 3050 of these had been disturbed by the plough, whilst another 9000 remained in situ and were recorded on a two-metre grid. Taken together, they were the latest deposit on the site. Since approximately half the area of the ring cairn was investigated, it seems likely that up to 25,000 fragments of quartz had originally covered the interior of the monument. The largest fragments were found at the centre and in the south-western sector of the enclosure and were as much as 15 cm in maximum dimensions. Among the smallest the equivalent figure was 3 mm.

There was nothing to suggest the use of different methods of stone working during successive phases in the development of the monument. That may be because such limited samples were associated with the earlier deposits. The largest fragments were clearly of vein quartz, whilst most of the smaller pieces probably derived from the same source. There was only limited evidence for the use of pebbles of quartz and quartzite. These raw materials were generally of poor quality and frequently contained flaws. An estimated 5% of the collection consisted of small flakes of higher quality quartz and only 1% was imported pink quartz. They would have been more suitable for artefact production, but there was no evidence that any of them had been used. The larger fragments are flakes or irregular chunks with one or more removals. More regular cores were hardly represented and there was little to suggest that any of the raw materials had been systematically worked. The remaining part of the assemblage consisted of irregular chunks, mainly of quartz, and angular shatter which had been broken rather than flaked. In the large collection from the interior of the monument material of all sizes from the largest to the smallest was found together, suggesting that the stone had been fractured on the spot. Apart from two dubious candidates, none of this material had been retouched, and any hammerstones appear to have been removed from the site.

22

Richard Bradley and Amanda Clarke

It is clear that virtually all this material results from the deliberate smashing of raw material rather than tool production. That process would have caused the stone to give off sparks or even to glow, suggesting that the process was intended to create a spectacle. It is not clear whether it happened on more than one occasion, but, after it was over, the entire surface of the cairn would have glittered in strong light (as it does today). It would be wrong to consider the treatment of the raw material as a specific technology, for its aim was essentially aesthetic. The character of the broken material was different from the collection found in field survey outside the enclosure. This is described in a later section (p. 23). On the other hand, it did resemble the small samples of quartz and quartzite from Waulkmill, Tuach and Croftmoraig which are considered in later chapters of this book. Cremated bone Fiona Shapland Cremated human bone was recovered from two secondary features cut through the centre of the ring cairn, where in one case (Context 1021) evidence of in situ burning was recorded. Methodology The cremated human bone from this site was macroscopically examined in the laboratory. Recovered bone was weighed using 0.01 g digital scales. Bone colour and condition, including the extent and nature of bone fragmentation, were recorded for the information this can provide on the cremation process (McKinley 2000). All bone fragments identifiable to an area of the skeleton or specific bone were separated out for analysis. Age-at-death was estimated, where possible, using the criteria of dental development wear (Miles 1962), pubic symphysis deterioration (Brooks and Suchey 1990), sternal rib end deterioration (Loth and Iscan 1989), cranial suture fusion (Acsádi and Nemeskéri 1970) or the progress of epiphyseal fusion (Scheuer and Black 2000). Sex was assessed using the pelvic and cranial criteria laid out by Buikstra and Ubelaker (1994). Any visible evidence for skeletal or dental pathology was recorded. However, the fragmented nature and surface alteration of cremated material means that evidence for most pathological processes is unlikely to be recognisable (McKinley 2000, 413).

Cremated bone from Context 1007 The total weight of bone fragments recovered from this context was 1032.5 g, which is consistent with the cremation of a single adult (McKinley 1993, 285) but would indicate that only part of the body was represented. No clearly non-human bone fragments were identified, although the majority of fragments were too small for definite identification. Bone colour and condition: the cremation process The cremated human bone from this burial was predominantly white in colour with very little variation between elements. Most fragments measured less than 5 × 5 mm, and were recovered only through sieving with a 4 mm and then a 2 mm sieve. Fracture lines were curved and irregular, with considerable warping. This indicates that the body was fleshed at the time of cremation, and that a temperature of over 900°C was maintained (Walker et al. 2008, 133). Identified fragments The vast majority (over 97%) of bone fragments from this burial were unidentifiable due to their small size. Only 86 fragments were identified as belonging to an area of skeleton or, in a small proportion of cases, a specific bone (Table 2.1). These identified fragments provide a minimum number of individuals of one, as there were no repeated elements identified. Evidence for age-at-death and sex Estimation of age-at-death was impeded by the fragmented and incomplete nature of the skeleton. No complete teeth were recovered, the state of the pubic symphysis and auricular surface of the pelvis were not recordable, and sternal rib ends were not observable. Where observable, cranial sutures were partially fused. The two observable epiphyses (femur and hand phalanx) were completely fused. All of these factors indicate that this individual was an adult, over 18 years of age. Sex assessment was problematic due to the poor survival of elements; no pelvic criteria could be recorded and only one cranial feature, the orbital margin was, identified. This was sharp and female in morphology, but this single marker does not provide strong enough evidence to assign a sex to this individual. Pathology No pathology was observable on these remains.

2.  Excavations at Hillhead, Tarland, Aberdeenshire Table 2.1. Identified human bone fragments from Context 1007. Area of skeleton Cranium Vertebrae Ribs Upper limb Hands Pelvis Lower limb Feet

No. Specific bones identified fragments 27 Frontal, occipital, temporal, zygomatic, mandible, 2 fragments of dental enamel 5 None 8 Left and right ribs 12 Humerus, radius, ulna 4 1 metacarpal, 2 proximal phalanges, 1 middle phalanx 8 Ossa coxa 21 Femur, tibia, patella, fibula 1 1 proximal phalanx

Cremated bone from Context 1021 This feature was interpreted as the remains of a pyre. The total weight of bone fragments recovered was 736.2 g, which is smaller than would be expected from the cremation of a single adult (McKinley 1993, 285) and it seems possible that the burnt area extended outside the section investigated in 2012. No clearly non-human bone fragments were identified, although the majority of fragments were too small for definite identification. Bone colour and condition: the cremation process The cremated human bone from this burial was predominantly white in colour with very little variation between elements. Most fragments measured less than 4 × 4 mm, and were recovered only through sieving. Fracture lines were curved and irregular, with considerable warping. This indicates that the body was fleshed at the time of cremation, and that a temperature of over 900°C was maintained (Walker et al. 2008, 133). Identified fragments The vast majority (over 95%) of bone fragments from this burial were unidentifiable due to their small size. Only 73 fragments were identified as belonging to an area of skeleton or, in a very small proportion of cases, a specific bone (Table 2.2). These identified fragments provide a minimum number of individuals of one, as there were no repeated elements identified. Evidence for age-at-death and sex Estimation of age-at-death was precluded by the fragmented and incomplete nature of the skeleton. The only evidence of any kind was a partially closed cranial suture line, which may indicate that this individual was an adult. No criteria for sex assessment were observable.

23

Table 2.2. Identified human bone fragments from Context 1021. Area of skeleton Cranium

No. Specific bones identified fragments 22 Frontal, temporal

Vertebrae Ribs Upper limb Hands

3 5 14 1

None None Radius 1 proximal phalanx

Pelvis Lower limb

3 25

Ossa coxa Femur, tibia

Pathology No pathology was observable on these remains. The surroundings of the monument Moyra Simon and Jane Summers The planting of trees had left a series of trenches cutting through the monument and the surrounding area. They were investigated in two ways. The first was to examine the site of the stone circle and the area immediately outside it on a twenty-metre grid, subdivided into 16 five-metre squares (Fig. 2.27). Beyond its limits an area of approximately 12 ha was investigated by transects at 10 m intervals. It meant that one in every five furrows was investigated. In the first stage of the fieldwork surface finds were recorded by grid square and GPS coordinates; GPS locations recorded those recovered from the transects. The work was carried out by members of NESARS and the GPS measurements were taken by Irvine Ross. Apart from the monument excavated in 2013, several features were recorded in the area covered by the survey. At NJ 5106 0725 there were the remains of a cairn, and a hut circle was located at NJ 5099 0707. Alder charcoal associated with this building is dated to 1030–850 bc (SUERC-56971). Other concentrations of stone require further investigation. These structures will be reported when work there is complete. Individual artefacts may have moved when the area was planted but only one piece of worked flint came from inside the monument whilst another 15 were identified in the grid squares to the south and east of the stone circle where they occurred together with a concentration of flaked quartz. They included a barbed and tanged arrowhead found in the disturbed topsoil together with four sherds of Beaker pottery (Fig. 2.28). These artefacts

24

Richard Bradley and Amanda Clarke

Figure 2.27. The location of the Hillhead monument in relation to the finds of worked flint and quartz revealed by tree planting (contours at 50 m intervals). were about 16 m south of the outer boundary of the monument and may be all that survive of the contents of an unmarked burial damaged during a previous campaign of forestry. The surrounding area contained an extensive scatter of worked quartz which reached as far as the summit of the hill but increased in density to the south and northeast of the recumbent stone circle. It was not represented on the lower ground SSW of that monument or on the northeastern limit of the area investigated by fieldwalking. This work also revealed a group of worked flints about 140 m northeast of the principal structure. There was a concentration of worked quartz in the same area. Five other pieces of flint came from peripheral areas on the edge of the area that was investigated. Four hundred pieces of worked quartz came from the area surrounding the stone circle. The raw material was generally of higher quality than the finds from its interior and included small flakes which probably result from artefact production; others could have been utilised. In contrast to the area within the ring cairn, there was less evidence of deliberately broken slabs or boulders of quartz and quartzite, and pink quartz was not distributed far beyond the limits of the monument.

The character of the principal monument Richard Bradley This section draws on the results of excavation and field survey in north-east Scotland to shed more light on the nature of this remarkable monument. Greater detail will be provided in Chapter 9 which discusses the amount of variation between the stone circles excavated in the Howe of Cromar (Tomnaverie, Hillhead and Waulkmill), and the Blue Cairn which was surveyed by Adam Welfare. Although the site had been disturbed in modern times and also in the first millennium ad, there seems little doubt that it was originally a recumbent stone circle similar to those investigated in recent years at Tomnaverie and Cothiemuir Wood (Bradley 2005, chapters 2 and 3). It must have been one of the largest structures of its kind with an overall diameter of 26 m, and seems to have been the second highest of those whose exact positions are known. It shares the same components as other monuments – a ring cairn which was filled with rubble during a secondary phase; an outer stone circle; an alignment towards the southwest; an apparent association with major mountains in the region; and evidence that the central area was reused. Several local monuments illustrate the special importance of quartz. The new

2.  Excavations at Hillhead, Tarland, Aberdeenshire

25

Figure 2.28. Beaker sherds and a barbed and tanged arrowhead found together in a forestry trench 16 m south of the monument (Jane Summers). project suggests that the stone setting at Hillhead was later than the construction of the ring cairn: an additional characteristic shared with other sites. Some features are more unusual. One is the location of Hillhead. It contrasts with that of most comparable sites because it is on such high ground. This is very different from the positions of both its neighbours which were associated with the basin to the west. At the same time, neither of those monuments was in an area with many surface finds; the same is true of Cothiemuir Wood (Bradley 2005, chapter 4). Hillhead, on the other hand, is surrounded by finds of worked quartz which outnumber those found during fieldwalking on more sheltered land in the Howe of Cromar. It seems likely that the deposition of worked quartz on the hillside began after the ring cairn had been built. The quartz artefacts cannot be dated exactly but they were not associated with the Late Bronze Age roundhouse excavated in 2014. The form of the Hillhead monument is most unusual. The original enclosure was defined by a rubble wall approximately 2.4 m wide and at least 45 cm high. It may have extended around the entire perimeter, with an entrance to the southwest. It was only later that the width of that boundary was increased along part of the circuit and an outer kerb was added. The final ring cairn was massive. The point can no longer be established with

any confidence, but the fact that the existence of this site was overlooked when so many other monuments were recorded suggests that the stone circle associated with it may have been inconspicuous. In that case there would have been a striking contrast between the large scale on which the cairn was constructed, and the less prominent role played by a ring of monoliths. It is obvious that the cairn increased in height towards the northeast. If this was a recumbent stone circle, the monoliths would have been tallest at the southwest. Recumbent stone circles were sometimes built on carefully levelled platforms, but Hillhead was located on a slope and this effect was emphasised by the structure of the ring cairn which was built on a much larger scale on its uphill side. The outcome was a secluded space that appeared to be slightly hollowed. This effect may have been enhanced by the stepped interior of the enclosure towards the north and northeast. A similar feature was observed at Hatton of Ardoyne in the nineteenth century (Coles 1901, 241–46), but this may be the only example suggested by a modern excavation. On the other hand, it could be one way of interpreting Childe’s plan (1934, fig. 9) of Old Keig which depicts a line of upright slabs running parallel to the outer kerb of a ring cairn but underneath the centre of a bank of rubble. Other structural features have counterparts elsewhere in north-east Scotland. The enclosure wall is particularly

26

Richard Bradley and Amanda Clarke

distinctive. It lacked the usual kerbs of upright stones but was not unlike a series of monuments in Buchan. Hillhead would have been approximately 20 m in diameter, which is considerably larger than any of those structures. A more appropriate comparison may be with the stone circle at North Strone. The two are strikingly similar to one another in size and ground plan, but in this case there is no evidence of a wall (Welfare 2011, 88). Another comparison might be with the bank of large boulders interpreted as the inner kerb of a ring cairn at Cothiemuir Wood (Bradley 2005, 57). This could have been a free-standing element, and again it might have been supplemented by a deposit of rubble and an outer kerb during a later phase. That is a tentative suggestion, but it is clear that such a sequence did occur at Hillhead. In fact the distinctive character of its outer kerb is closely matched by its counterpart at Cothiemuir Wood which was held in place between two deposits of large boulders. In their final forms the plans of these monuments were similar to one another. Hillhead was 26 m in diameter and its court measured 15 m across. At Cothiemuir Wood the equivalent measurements were 16 m and 10 m respectively. The ratio of the court diameter to that of the ring cairn was 1:1.6 at Cothiemuir and 1:1.7 at Hillhead. There is also a contrast with the well-preserved structure at Tomnaverie which will be considered in more detail in

Chapter 9. Its outer kerb was built by exactly the same method, but, unlike that monument, Hillhead did not show any sign of patterned stonework on the surface of the cairn. There was no indication of radial divisions, nor was there anything to suggest the choice of raw materials according to their colours. On the other hand, the putative recumbent and flankers at Hillhead conform to a more general pattern in north-east Scotland where the large horizontal stone has a different colour and texture from the pillars to either side (Welfare 2011, 117–18). At Hillhead the features taken to be sockets for the flankers cut through the material of the ring cairn, suggesting that the stone circle was erected afterwards. That was also the case at Tomnaverie and Cothiemuir Wood, but here the monoliths were bedded in an external platform abutting the outer kerb. There was little sign of a similar feature at Hillhead. A last anomaly also recalls early observations of these monuments. This is the deposit of broken quartz spread across the interior of the ring cairn. It was the latest component of the monument and has no equivalent in the more extensive excavation at Tomnaverie. Something similar was recognised in the recumbent stone circle at Strichen (Phillips et al. 2006), and it also happened at Croftmoraig where the results of a new excavation are considered in Chapter 4. That is just one of the elements discussed in more detail in the second part of this book.

chapter three

Excavations at Waulkmill, Tarland, Aberdeenshire: a Neolithic pit, Roman Iron Age burials and an earlier prehistoric stone circle

Richard Bradley, Amanda Clarke and Fraser Hunter Introduction The site at Waulkmill was selected for excavation in 2012 because the stone circle recorded there in the nineteenth century was probably the neighbour of Tomnaverie, which is only 2 km away. The aim of the work was to compare these two monuments. At the time it took place nothing was known about an example at Hillhead just 4 km to the northeast; in the event it was investigated one year later. That project has already been considered, and

the relationship between all three sites is discussed as a separate study in Part 2 of this monograph. The exact position of the Waulkmill circle had been lost and it was necessary to establish it by excavation. The site was located on the low ground of the Howe of Cromar on the edge of the modern village of Tarland. It was intervisible with Tomnaverie and Hillhead (Fig. 3.1). Little was known about the monument or its surroundings, but about 700 m to its west a settlement

Figure 3.1. The position of the monument at Waulkmill (arrowed) seen from the Tomnaverie recumbent stone circle (Aaron Watson).

28

Richard Bradley, Amanda Clarke and Fraser Hunter

Figure 3.2. The relocated monolith at Waulkmill, with the position of the stone circle under excavation in the background (Aaron Watson).

Figure 3.3. The site of the stone circle under excavation, seen from the north. The picture emphasizes its location on top of a shallow ridge (Aaron Watson). consisting of three enclosures or roundhouses was identified. Approximately 350 m to its east a number of Roman Iron Age artefacts were discovered in the nineteenth century (Coles 1905, 214–17; Ogston 1931, figs 62, 63). They were associated with at least one grave, but they were too far away from the stone circle for any connection to be inferred.

The Waulkmill stone circle was demolished by the farmer early in the nineteenth century, but one of the uprights was allowed to remain in place (Fig. 3.2). The monument was located where the ground fell away at the western end of an inconspicuous glacial ridge (Fig. 3.3). Until recently the lower ground was boggy and poorly drained, meaning that this feature would have appeared

3.  Excavations at Waulkmill, Tarland, Aberdeenshire

29

Figure 3.4. The location of Waulkmill stone circle in relation to the local topography, the position of a former sandpit where burials were found in the nineteenth century, and the settlement site at Melgum (contours at 50 m intervals). more prominent than it is today. The eastern end of that ridge is cut by a stream, and in between it and the circle there was a quarry. It was here that the Roman Iron Age artefacts were discovered in 1898 and shortly after. Early Ordnance Survey maps supply important information, for over time that quarry expanded towards the west. When the artefacts were found, the part closest to the stream was being worked. This provides an accurate estimate of the original findspot (Fig. 3.4). Nineteenth century discoveries and observations The stone circle Little is known about the Waulkmill stone circle. When Fred Coles (1905, 214) was writing his account of the site in 1905 he was told that it had been destroyed ‘about seventy years ago’; ‘ten or eleven’ monoliths were removed, leaving the one that survives today. More recently it was moved to the edge of the field. No artefacts have been found there, but, in 2000, scatters of worked flint and quartz were recorded from the hillside north of the site (Bradley 2005, 87–92). They were separated from the monument by an area of low-lying ground. Roman Iron Age artefacts from the quarry More is known about the finds from the quarry, but the earliest records raise the possibility that two separate groups of artefacts have been conflated in subsequent accounts.

The first discovery was in 1898 and was documented seven years later in Fred Coles’s account of the Waulkmill stone circle (1905, 214–17). He drew on a contemporary newspaper account and his own examination of the finds. According to the Aberdeenshire Evening Express for August 6th 1898, a gamekeeper came upon four undressed stones, each measuring 2 feet broad and 18 inches high, placed one above the other, and at a depth of 4 feet from the surface … On making further search he discovered eight stone and glass buttons … lying at regular intervals as if they had fallen from the garment worn by the occupant of the tomb. A silver buckle … and two or three pieces of steel or wrought-iron, apparently the remains of some weapon, as well as a number of small bones, were also found. The bones, when lifted, crumbled into dust.

The landowner told Coles that ‘at a later date, some time in 1899, a number of silver articles were found in the same sand-pit … I fear all trace of them is lost’. The artefacts discovered in 1898 were a penannular brooch, a dagger and 13 gaming counters (Fig. 3.5). There was also ‘a piece of almost colourless glass ... cylindrical in form’. In a recent paper Hall and Forsyth (2011) date the grave to the second or third century ad, and Hall considers them in greater detail in a later section of this chapter. Since these artefacts were gaming pieces rather than buttons, the observation that they were ‘lying at regular intervals as if they had fallen from the garment’ needs to be reconsidered. Perhaps they were laid out on a board.

30

Richard Bradley, Amanda Clarke and Fraser Hunter a burial and could date from a later period (as discussed further below). The 2012 excavation

Figure 3.5. The miniature cauldron, penannular brooch, iron dagger and gaming pieces found in the sandpit at the end of the nineteenth century (NMS). Sixteen years later some of the subsequent discoveries were traced to a private collection and published by J. Graham Callander (1915). In his words they included a small cup of cast bronze, the crown of a human molar tooth found in the cup, a disc of translucent blue glass broken in two, and twelve pebbles of brown, grey, and whitish quartzite. The owner said that the glass disc was all that remained of a number of similar objects and other relics found in a sand-pit near Tarland, which were once in his possession but which he had been compelled to give up to the Laird; that he had kept the glass disc, it being broken, and that the cup, tooth and pebbles had been found afterwards in the same sand-pit and he had secured them, their discovery being kept secret. (Callander 1915, 204; our emphasis)

It is possible that these artefacts came from more than one deposit. Those found in 1898 were in a cist containing bones. Since it was bounded by massive stones it would have been easy to remove its contents. The 1899 collection included a tooth, the miniature cup and a second set of gaming counters made of different materials from the other group. They could have accompanied a second burial. Lastly, there is the reference to ‘silver articles’ being found in 1899. It cannot refer to any of the artefacts catalogued by Callander, raising the possibility that a third deposit was discovered in the quarry and subsequently lost. It may not have been associated with

The aim of the 2012 excavation was to relocate the stone circle and, if possible, to establish its plan and chronology. That would allow it to be compared with the well-preserved monument at Tomnaverie. Geophysical survey performs badly on the glacial geology of north-east Scotland, and at Waulkmill the likely position of the stone circle was estimated by considering the local microtopography in relation to the siting of similar monuments in the area. Although the structure was levelled as an obstacle to the plough, the surviving monolith was recorded on the first edition Ordnance Survey map; it is also apparent on a photograph taken in 1915, but has been moved since then (Fig. 3.6). Some of the other stones were probably built into a field wall, as it includes some exceptionally large fragments just south of the supposed location of the monument. The character of the excavation was guided by three considerations. The first was that the stone circles of north-east Scotland occur in a limited range of sizes and usually include about a dozen monoliths. In most cases they command an uninterrupted view towards the south or southwest (Welfare 2011). A second feature is the presence of a ring cairn inside the circle of monoliths. Experience at Tomnaverie and other sites showed the kinds of subsoil feature that would result from the removal of a stone setting. Work at Broomend of Crichie (Bradley 2011, 20) and the Hill of Tuach (this volume, Chapter 5) also documented the ways in which monoliths were broken up during the nineteenth century. At Waulkmill the modern ploughsoil was removed mechanically over an area measuring 25 × 15 m, revealing the glacial till which was a mixture of coarse gravel, clay, boulders and patches of sand (Fig. 3.2). The ploughsoil at the end of the glacial ridge contained an unusually high density of rubble, perhaps from a levelled cairn. Beyond conforming to the estimated position of the monument, it no longer retained any structure. The excavation identified subsoil features dating from three separate phases (Fig. 3.7). In chronological order they were: • The base of an isolated pit associated with Neolithic pottery; • Traces of the inner and outer kerbs of a demolished ring cairn, the sockets of up to four standing stones, and, possibly, a large pit in the centre of the monument. These features should date from the Bronze Age.

3.  Excavations at Waulkmill, Tarland, Aberdeenshire

31

Figure 3.6. An earlier twentieth century postcard of Tarland showing (arrowed) the surviving monolith at Waulkmill in its original position on top of a low ridge (Photographer unknown).

Figure 3.7. Plan of the excavated features at Waulkmill.

32

Richard Bradley, Amanda Clarke and Fraser Hunter • Two graves associated with Roman Iron Age artefacts, and two deposits of cremated bone with radiocarbon dates in the same period.

The Neolithic pit The earliest feature was a steep-sided pit which had been truncated by the plough and survived to a depth of only 20 cm. It had a single layer of filling with a broken nodule of waterworn quartz, 10 cm in maximum dimension, in its centre. The feature contained sherds of Carinated Bowl. No other artefacts were found there. A few shallow postholes were identified nearby. One contained a flint flake, but they remain undated and do not define any recognisable structure. In view of the fieldwalking project undertaken in 2000, it is worth drawing attention to the absence of worked flint or quartz in the pit itself. It is clear that this method may not always locate the sites of this phase. That is particularly true in Cromar where two long cairns are recorded but no surface finds can be assigned with any confidence to the same period. The same disparity has been noted on other Early and Middle Neolithic sites in Scotland (Sheridan and Sharples 1992, 7–8). The pottery from the pit Alison Sheridan One hundred and nineteen sherds, together with a few fragments and crumbs, were recovered from the truncated pit; the overall weight amounts to just over half a kilogram. Small parts of eight or nine pots are present, with most represented by just four or fewer sherds. The largest single sherd, from Pot 2, measures 83 × 51 mm; this is the least incomplete pot, although only 10–15% of its body is present (in the form of 39 sherds plus a few fragments). While there had been considerable spalling of some pots (principally Pots 2–4), due to the laminar nature of the fabric, there was generally only minimal or minor abrasion to the sherds, suggesting that they had not lain around for long before being deposited in the pit. The two sherds from Pot 9 are softer and somewhat more abraded than the rest, but these may well have been burnt. Where lower body sherds exist, these indicate that the pots had been round-based; and sufficient diagnostic material survives to indicate that at least five or six of the pots had been carinated or (in the case of Pot 2) shouldered. The pots vary in size, in wall thickness, and in fabric and finish, although in no case are we dealing with the very thin-walled, very fine pottery

that characterises the earliest Carinated Bowl pottery in Britain. Pot 1 (Fig. 3.8, top) appears to have been the smallest, finest and thinnest-walled vessel. It is a small carinated bowl with an estimated rim diameter of c.145 mm and a projected depth of c.96 mm; wall thickness varies from 8 mm at the neck to 8.5 mm at the carination, narrowing to 4.6 mm at the upper belly. It has an upright, flattened-rounded rim, an upright straight neck, a gentle carination and a medium-depth belly. The surfaces had probably been wet-smoothed but are slightly uneven, with lithic inclusions protruding through the interior and a fingertip depression from where the rim was shaped on the outside; there are hints of horizontal wipe-marks on the exterior. The pot is a medium to blackish-brown throughout. Inclusions consist of angular and sub-angular fragments, up to 2.5 × 1.0 mm in size, of a creamy-buff mineral, possibly feldspar, and possibly also some quartz; there are also occasional small mica platelets, and the overall density of inclusions is c.5%. Blackish staining on the interior may represent the evaporated (rather than burnt-on) former contents of the pot. Pot 2 (Fig. 3.8, bottom) is a slightly larger, shouldered bowl, with an estimated rim diameter of c.230 mm and a projected depth of c.143 mm; wall thickness ranges between 6.5 mm and c.10 mm. It has a rounded, everted rim; a fairly short, curving neck that is either upright or minimally everted; a ledge-like shoulder; and a mediumdepth belly, with the body curving in unevenly below the shoulder. The surfaces are uneven, with numerous lithic inclusions projecting. The surfaces have a slip-like covering that could have resulted from wet-smoothing; there are traces of wipe marks on both surfaces. The exterior is a dark brown, slightly reddish in places and blackish elsewhere; the core is medium brown to blackish; and the interior is dark brown and black. The fabric is hard and laminar, with many sherds existing as spalls. It is also gritty, with abundant subangular fragments (up to c.4.5 × 4.5 mm) mostly of quartz, some accompanied by a black mineral or by mica; the overall density is c.15%. There is probable black staining on the exterior. Pots 3 and 4 (not illustrated) had been large, thickwalled (up to 17.5 mm), coarseware bowls, either carinated or shouldered. A neck sherd from Pot 3 suggests that its diameter at that point had been c.270 mm, while a neck or upper belly sherd from Pot 4 suggests a diameter there of c.300 mm. Several sherds had clearly broken along coil joints. Surfaces are uneven but had been wet-smoothed, with clear wipe marks

3.  Excavations at Waulkmill, Tarland, Aberdeenshire

33

Figure 3.8. Reconstructions of Early Neolithic Carinated Bowl Pots 1 (above) and 2 (below) from Waulkmill (Marion O’Neil). visible on both surfaces of the Pot 3 neck sherd and on some of the Pot 4 sherds. Surface colours are generally reddish-brown, with the core colour either echoing this or (in Pot 4) being a dark grey-brown. The same lithic inclusions as those noted in Pots 1 and 2 are present, but they include some larger, angular fragments (up to 4.5 × 5.5 mm in the case of Pot 3, and 7 × 5 mm in Pot 4); the density in Pot 3 is c.5%, whereas in Pot 4 it is as much as 15–25%. No obvious traces of organic residues are visible on these pots, although there may be hints of thin staining or sooting on the outside of Pot 4. The other pots lie between Pots 1 and 2 on the one hand, and Pots 3 and 4 on the other, in terms of

their size, wall thickness and surface finish, with Pot 6 probably resembling Pot 1 (but perhaps with a slightly splaying neck). Detailed descriptions of each exist in archive form. The same inclusions occur, in different sizes and densities, in the various vessels; one sherd from Pot 5 is notable for its abundance of mica and for a large quartz-with-mica inclusion. With the possible exception of one of the Pot 9 sherds, there is no evidence for decoration on any of the pots; the exception has two very shallow parallel lines on its exterior, but it is unclear whether these had been deliberately made, or represent the burnt-out traces of grass blades. No obvious signs of organic residues were noted on Pots 5–9.

34

Richard Bradley, Amanda Clarke and Fraser Hunter

This assemblage clearly belongs to the Early Neolithic Carinated Bowl tradition (Sheridan 2007a), but apparently not to its earliest version, which has a higher proportion of thinner-walled, fine-textured, carefully-finished vessels (as seen, for example, in the assemblage from the ‘hall’ at Crathes Warren Field, Aberdeenshire; Sheridan 2009). However, it is clear from the assemblage found in another ‘hall’ close to Crathes, at Balbridie, that ‘modified’ versions of Carinated Bowl pottery had started to emerge soon after the initial appearance of the tradition as a whole (Sheridan 2009). A date for the Waulkmill assemblage as early as the 38th century bc is therefore possible, although a later date within the second quarter of the fourth millennium is equally possible. The Carinated Bowl tradition in general is well-represented in this part of Scotland, particularly along and between the valleys of the Don and the Dee, which appear to have been important foci for early farming communities. Several assemblages from this area have come to light since the publication of a Carinated Bowl pottery Scottish distribution map in 2007 (Sheridan 2007a, fig. 1). Find contexts in this area range from the aforementioned ‘halls’ to smaller houses (as at Garthdee Road, Aberdeen; Murray and Murray 2014) and individual pits or pit groups (as, for example, at Pitdrichie Quarry, Drumlithie, Aberdeenshire: Murray and Murray 2009). The presence of small parts of several pots, with the sherds not heavily abraded, is characteristic of the pit contexts. The contents of the Waulkmill pit could well constitute refuse from a settlement in the immediate vicinity. Worked flint Richard Bradley A single flint flake was found in a posthole which was not associated with any other artefacts. It was part of a pebble that had been worked down to its full extent, using narrow flake technique. It could date from the same period as the Neolithic pottery, but this is by no means certain. The Bronze Age stone circle The second phase is represented by the surviving traces of the stone circle levelled nearly two hundred years ago. Four components could be identified, all of them in the area that contained most of the rubble in the ploughsoil. Up to four possible stone sockets were found, following an arc that suggested the original monument had been approximately 18 m in diameter. Like the other features, they had been truncated by the plough, and three of them

were no more than 25 cm deep. They could be identified by their regular plans and profiles, and by comparison with well-preserved features at Tomnaverie. Another monolith had a double socket which was 40 cm deep, indicating that a stone had been replaced. Large boulders, interpreted as disturbed packing, filled the western half of this feature. The monoliths had obviously been of varied proportions and that affected the extent to which they were bedded in the ground. Of those excavated, only the double socket was large enough to have held the stone that still survives. The stump of another one remained in the ground, but it was not clear whether it had been the base of a monolith or part of the kerb. The structure was built out of pieces of felsite like the standing stone at the edge of the field (Peter Craig pers. comm.). The stone holes followed the same arc as a series of shallower features which were interpreted as the setting of a kerb in which the more substantial boulders were shallowly bedded; it is likely that others were simply laid on the ground surface, as happened at similar sites in the area (Fig. 3.9). These features were irregular in outline and up to 25 cm deep. In contrast to the stone sockets, their edges were difficult to define, and these features were probably disturbed when the site was levelled around 1835. Not all of them need have been associated with the monument, but they appear to have followed its outer limit. If that is correct, then approximately 30% of the area of the original monument was within the excavated area. As expected, it had been built on the end of the glacial ridge where the ground fell away to the north, south and west. To the southwest the spread of rubble in the modern ploughsoil ended just outside the line of the outer kerb, but there was no evidence that the sockets for the monoliths there were any deeper than the others. This point is considered further in Chapter 9. The central area of this structure was better preserved. It was represented by a curving gully, cut steeply into the natural subsoil (Fig. 3.10). It was up to 50 cm wide at the surface and 30 cm deep, but had obviously been truncated by the plough. It defined a circular area 4 m in diameter (Fig. 3.11). Comparison with the excavated site at Newton of Petty suggests that it marked the foundation trench for a kerb of upright slabs defining an open ‘court’ in the centre of a ring cairn (Bradley 2000, chapter 6); its dimensions are very similar to those of examples that remain intact. The kerbstones had been carefully removed, but sufficient evidence remained to suggest that the original structure was associated with large pieces of broken quartz whose distribution focused on the western sector of the ring. They were possibly

3.  Excavations at Waulkmill, Tarland, Aberdeenshire

35

Figure 3.9. The surviving remains of the stone circle. The buckets mark the positions of stone holes and the shallower features in between them are interpreted as the position of the outer kerb. To the right the foundation trench for the inner kerb has been fully excavated, as has the central pit. The rectangular feature top left is Grave 2 (Aaron Watson).

Figure 3.10. The partly excavated foundation trench for the inner kerb, with a recut pit in the centre capped by a setting of boulders (Aaron Watson).

36

Richard Bradley, Amanda Clarke and Fraser Hunter

Figure 3.11. Plan and section of the circular trench and central pit, with details of the secondary recut and the setting of quartzite cobbles. used to hold the kerb in position. Alternatively, they may have been piled up against its inner face. On other sites an open enclosure was filled with rubble when the monument was closed. That could have happened at Waulkmill where the internal space contained patches of larger stones than those in the surrounding area. At the centre of the monument was a pit which extended beyond the limit of the excavation. It was 1.8 m across and 1.05 m deep. Its function is unknown. The bottom 35 cm of the filling consisted of banded silts, perhaps resulting from natural erosion, but above them was redeposited soil and gravel containing many boulders. After these layers had accumulated, the uppermost part of this feature had silted naturally. There was no evidence of any burial and no artefacts were found. Its date remains

uncertain, but there are other sites, including Newton of Petty (Bradley 2000, chapter 6) and Loanhead of Daviot (Kilbride-Jones 1935), where similar pits were excavated in the centre of a ring cairn. They date from the Late Bronze Age and are associated with cremated bone. In this case the only human remains were in a subsequent recut and date from the Roman Iron Age. This evidence is considered in a later section of the chapter. Discussion Although so little remained after nearly two hundred years, it is clear that the stone circle was not much smaller than its neighbour at Tomnaverie: it was about 18 m in diameter, compared with 17.5 m on the latter site. The

3.  Excavations at Waulkmill, Tarland, Aberdeenshire

37

Figure 3.12. The southern horizon from the Waulkmill monument with Mount Keen in the distance on the right (Aaron Watson). sizes of the excavated stone sockets did not suggest that there had been a setting of taller monoliths towards the south or southwest, but this cannot be conclusive. At the same time it is unusual to find a monument in which the standing stones were set into the outer kerb. Normally, they form two concentric circuits separated by a narrow space, but, as Chapter 9 will explain, there is a local parallel for exactly this arrangement at the Blue Cairn, Ladieswell, only 7 km from Waulkmill (Welfare 2011, 15). The most likely reconstruction of this monument is that it took the form of a low circular cairn, with an open space at its centre. It may have had a boulder kerb on the outside, but the inner court was probably defined by a wall of slabs set upright in the ground and was associated with pieces of flaked and broken quartz. The outer kerb incorporated a circle of standing stones set in their own sockets at intervals of approximately 3.5 m. For that reason it is possible to estimate that the monument was originally defined by a circle of up to 15 monoliths; however, it is not known whether they were equally spaced. In 1905, Coles was informed that originally there had been ‘ten or eleven’ standing stones on the site, plus the one that survives today (Coles 1905, 214). There was no direct dating evidence for any part of the monument, but in plan and dimensions it is very similar to a ring cairn at Sundayswells 14 km to the east. That structure lacked a stone circle, but included a court 3 m in diameter defined by a setting of upright

slabs like that postulated at Waulkmill. In the centre of the site was a complete All-Over Corded Beaker (Welfare 2011, 228). The circle at Waulkmill would have commanded an all-round view towards the horizon, but, in common with many of the stone circles in this part of Scotland, the main focus was towards the south and southwest. It shares this characteristic with recumbent stone circles. There is no dead ground immediately outside the monument, but the vista is dominated by an intermediate ridge 3 km away and, in the distance, by the mountains of south Deeside. If the structure was directed towards a specific peak, it could have been the pointed profile of Mount Keen at a distance of 18 km (Fig. 3.12). Worked stone Richard Bradley Quartz and quartzite The foundation trench for the inner court contained 28 substantial fragments of quartz and quartzite. All were angular fragments, but only five of the smallest pieces and one of the larger fragments were struck flakes. The rest had simply been smashed. They fell into three broad size classes. The smaller ones (14 examples) were between 4 cm and 8 cm in maximum dimensions; for another six the equivalent figure was between 9 cm and 11 cm; and

38

Richard Bradley, Amanda Clarke and Fraser Hunter

for the eight largest fragments it was between 12 and 20 cm. There is nothing to suggest that artefacts were being made here, and there may have been a greater concern with the appearance of the quartz when the stone was freshly broken. This material compares with the sample from the 2012 excavation at Croftmoraig which is considered in Chapter 4. The Roman Iron Age burials Inhumation graves At an early stage in the excavation two distinctive features were identified close to the edge of the monument and on the outer edge of the spread of rubble identified in the ploughsoil. They were 2.5 m apart and larger than any of the excavated stone sockets, but contained large flakes of rock very like those excavated on the Hill of Tuach, where a ring of monoliths was demolished during the nineteenth century (the report on this project appears in Chapter 5). It seemed likely that the fragments at Waulkmill resulted from a similar process. At first these features were interpreted as the holes left behind when two of the taller monoliths were uprooted. It raised the possibility that the structure had been a recumbent stone circle similar to that at Tomnaverie. A problem with this interpretation was that only one of these pits was especially close to the perimeter of the prehistoric monument. Another was that their positions were difficult to reconcile with the layout of its inner kerb, which was the best-preserved part of the structure. It was difficult to see how they could have formed parts of two concentric rings. It was only when excavation was further advanced that the possibility arose that they were really graves. Soon it was confirmed by the discovery of a gaming set and a penannular brooch, very like those found in 1898. At that point it became apparent that the remains of the stone circle had been reused during the Roman Iron Age. The first of these features to be excavated was a rectangular pit (Grave 1), 1.9 m long and a metre deep (Fig. 3.13). The filling was of loose gravel and sand, disturbed by roots and burrows, but in the bottom 20–45 cm of this feature the sides were lined by between two and four vertical courses of rounded or roughly squared stones, up to 35 cm in maximum dimension; slabs performed the same role at its western end (Fig. 3.14). They defined a rectangular cist which was at least 1.4 m long and 60 cm wide; the full length is uncertain as the eastern edge of this structure had collapsed before the grave was filled. That could not have happened if the stones had been packed around a coffin. The bottom of

the grave was a layer of smaller, more irregular boulders and clean sand. There was no trace of any bone or wood, but at this level a set of 11 gaming pieces/amulets was found in a compact group in the centre of the cist. They might have been in a bag. Towards the eastern end of the grave a penannular brooch was discovered at the same level. It was strikingly similar to the one from the nearby quarry. The upper part of the cist may have collapsed as there were a number of large boulders in its filling, as well as the flakes mentioned earlier. It must have been reinstated rapidly, as it was dug through a patch of loose sand and its edges were extremely unstable. The second grave (Grave 2) occupied a larger oval pit (Fig. 3.15). At the level of the glacial till its maximum dimensions were 3.5 × 2.9 m (Fig. 3.16). It was dug through more than one layer in the natural subsoil, including a level of small boulders, and bands of humus and gravel (Fig. 3.17). At a depth of 2 m was another setting of boulders, the largest of which had maximum dimensions of 20–30 cm. They included pieces of quartz, and substantial flakes like those associated with the other burial. On its south side this setting had a straight edge that recalled a wall foundation, but elsewhere it merely filled the lower part of the pit and lapped against its sides. It seems to have covered the base of the original cutting over an area measuring 2.0 × 1.4 m. The stones enclosed an approximately rectangular space which was 2.5 m long. It was between 50 and 70 cm wide and 20 cm deep and contained traces of a coffin, the limits of which extended a short distance beyond the distribution of boulders (Fig. 3.18). It seems likely that these stones were packed around the edges of the coffin when it was introduced to the grave. At the level of the stone setting it was possible to identify the discontinuous stain left by the sides of a timber coffin. They seem to have buckled as the filling of the grave settled against them, but enough remained to suggest that it was originally 2.5 m long and about 60 cm wide. Part of the base was still preserved. It was in very poor condition, but samples of the wood were taken where it survived intact. They have been examined by Steven Allen who reports that they come from a single radially faced board, most probably of oak. Human bones were not represented, but Dr Aleks Pluskowski (pers. comm.) was able to identify three heavily mineralised teeth of bos close to the western end of the grave (Fig. 3.19). They may represent all that remains of a skull, traces of which could be identified as a patch of discoloured soil. Beside them were two bronze rings.

3.  Excavations at Waulkmill, Tarland, Aberdeenshire

Figure 3.13. Plan and elevation of Grave 1.

Figure 3.14. Vertical view of Grave 1 after excavation (Aaron Watson).

39

40

Richard Bradley, Amanda Clarke and Fraser Hunter

This feature seems to have been filled in two stages. In the first, soil accumulated around the coffin. The upper part of the pit may have stayed open long enough for the faces of the cutting to be affected by erosion. The secondary filling consisted of loose sediments which had probably weathered from the sides. They included a distinct lens of dark humic material, suggesting that these deposits had time to stabilise before more material accumulated.

Figure 3.15. Outline plan of Grave 2.

Discussion The second grave was obviously much larger than its neighbour, and there were other contrasts between them. Each contained a setting of boulders towards its base, but in one case this took the form of a drystone wall built against the edge of the pit to create a kind of cist – there was no sign of any wood. In the second case the stones were laid around a timber coffin on the bottom of the grave. The first grave was refilled rapidly, but not before one wall of the burial chamber had given way. After that, it was completely covered. Its neighbour had a different history. Again the burial was enclosed, this time in a coffin surrounded by a setting of boulders. Its wooden sides could be traced through the lowest filling, but there was no trace of the lid, although it might have been level with the surface of the stone setting. The upper part of the pit seems to have been left open for some time, perhaps to allow the coffin, or even the body, to be viewed. Its sides were exposed to natural weathering and it filled by natural silting. The positions of both graves appeared to acknowledge the presence of the stone circle. They were just outside its perimeter – the distances were 1.25 m and 3 m respectively – and to some extent their alignments echoed that of its outer kerb, although the relationship was only approximate. Still more important, large flakes which may well have been removed from the stones had been incorporated in the fillings of both these features.

Figure 3.16. Detailed plan of the lower level of Grave 2.

3.  Excavations at Waulkmill, Tarland, Aberdeenshire

41

That happened as the first grave was closed, and here the main concentration was at a higher level than the cist. In Grave 2, however, they were associated with the setting of boulders around the coffin. Here the flakes were in the same deposit as pieces of quartz – a distinctive material that was associated with the earlier ring cairn and may have been taken from that structure. Cremation burials Two deposits containing human remains were discovered in the excavation of the stone circle and were initially thought to date from a period of reuse during the Late Bronze Age, when many monuments in the region received deposits of burnt bone. Nine metres outside the stone circle at Waulkmill was a shallow pit, 40 cm in diameter and only 8 cm deep. It was capped by small pieces of stone. Others lined the sides and bottom of this feature, which contained a cremation burial. What was unexpected was its age. A sample from this deposit returned a date in the Roman Iron Age. The second deposit raised more problems. The existing feature in the centre of the ring cairn was recut by a steep-sided pit, 60 cm in diameter and 55 cm deep. It had been filled by a deposit of soil and charcoal which included a small amount of cremated bone. When it was refilled, its position was marked by a circular setting of rounded cobbles. They were between 7 cm and 12 cm in maximum dimensions. Many of them were quartzite.

Figure 3.17. Grave 2 showing the setting of boulders towards the left flanking the remains of the wooden coffin (Aaron Watson).

Figure 3.18. The first indication of a coffin stain within Grave 2 (Aaron Watson).

42

Richard Bradley, Amanda Clarke and Fraser Hunter Methodology The methodology of analysis used was the same as stated in Chapter 2 (p. 22). Cremated bone from Context 1031 The total weight of bone fragments recovered from this burial was 2063.8 g, which is consistent with the cremation of a single adult (McKinley 1993, 285) but would indicate unusually good post-cremation recovery and virtually complete burial of the remains. No clearly non-human bone fragments were identified, although the majority of fragments were too small for definite identification.

Figure 3.19. Two spiral rings and the stain left by the animal skull in Grave 2 (Aaron Watson). They had been placed at the exact centre of the older monument. Again the sample of burnt bone produced a radiocarbon date in the Roman Iron Age. Radiocarbon dates Samples from both the cremation burials were dated, but the bos teeth were too heavily mineralised to be studied by the same method. An isolated pit outside the stone circle (Context 1031): SUERC-44435 Cremated bone. 1802±29 BP/ad 130– 323 at 95.4% probability The recut pit in the centre of the stone circle (Context 1057): SUERC-44434 Cremated bone. 1673±29 BP/ad 259– 428 at 95.4% probability Cremated bone Fiona Shapland Cremated human bone was recovered from two contexts. Context 1031 was a shallow pit which contained a Roman Iron Age cremation burial. Context 1037 was a shallow feature capped by a setting of pebbles dug into the filling of an earlier pit in the centre of the ring cairn.

Bone colour and condition: the cremation process The cremated human bone from this burial was consistently white in colour, with very little variation between elements. Most recovered fragments measured less than 10 × 10 mm, although a significant proportion of fragments measured over 20 × 10 mm. Fracture lines were curved and irregular, with considerable warping. This indicates that the body was fleshed at the time of cremation, and that a temperature of over 900°C was maintained (Walker et al. 2008, 133). Identified fragments The majority (over 80%) of bone fragments from this burial were unidentifiable, but 391 fragments were identified as belonging to an area of skeleton or, in a small proportion of cases, a specific bone (Table 3.1). These identified fragments provide a minimum number of individuals of one, as there were no repeated elements identified, although given the large weight of bone recovered the presence of a second individual is a possibility. Evidence for age-at-death and sex Estimation of age-at-death was impeded by the fragmented and incomplete nature of the skeleton. No complete teeth were recovered, the state of the pubic symphysis and auricular surface of the pelvis were not recordable, and sternal rib ends were not observable. Where observable, cranial sutures were partially fused. All observable long bone epiphyses were completely fused, as was a single upper lumbar vertebral ring. In addition, one upper third molar root was observed to be complete. All these factors indicate that this individual was an adult over 19 years of age. One site of degenerative joint disease (see below) may suggest that this is an older adult (over 30 years).

3.  Excavations at Waulkmill, Tarland, Aberdeenshire Sex assessment was particularly problematic due to the poor survival of the pelvis, from which no pelvic criteria were recordable. Two cranial features, the orbital margin and occipital process, were observable, and both of these were possibly female. This does not provide strong enough evidence to assign a sex to this individual. Pathology Only one pathological feature was identified on this skeleton: slight marginal osteophyte development on the articular facet of the tenth or eleventh left rib. This is evidence of joint deterioration which would presumably have affected the corresponding vertebral facet as well. Degenerative joint disease of the costovertebral joints is common in individuals over the age of 40 years, and is usually associated with other degenerative changes in the spine (Ortner 2003, 549, 555). Cremated bone from Context 1037 This deposit is also dated to the Roman Iron Age. The surviving cremated bone from this deposit consisted of nine small unidentifiable fragments, only two of which were over 5 × 5 mm in size. It is impossible to be certain that these fragments represent human bone, although given the context of discovery this seems most likely. These fragments are predominantly white in colour indicating a high cremation temperature. No information on sex or age-at-death could be ascertained. Roman Iron Age artefacts Catalogue of the finds from the 2012 excavation Fraser Hunter and Penelope Walton Rogers, with surface X-ray fluorescence analysis by Maureen Young Grave 1: a penannular brooch Fraser Hunter Description: small penannular brooch of Fowler (1960) type Aa, with a circular-sectioned hoop, slightly oval in plan, expanding gradually to squared-off terminals (Figs 3.20–3.21). Humped pin, circular in section, the tip flattened into a narrow leaf shape (its extremity lost). Filemarks from finishing on the upper surface; very slight circumferential facets on the hoop from hammering into shape. Hoop: external diam. 21.8 × 24.7 mm; rod: diam. 2.7 mm (terminals: diam. 3.8 mm). Pin: length 38.4 mm, diam. 2.2 mm. Alloy: both hoop and pin are leaded bronze with trace levels (probably