The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703: Archaeological Investigations 1979-2009 9781407356877, 9781407356884

Table of contents :
Front Cover
Contents
Acknowledgements
List of Figures
1: Introduction
2: Vessel History: Building, Service and Loss
2.1 Background (1660–1674)
2.1.1 The Thirty Ships (1674–1688)
2.2 The Active Service of the Stirling Castle
2.2.1 1690–1699
2.2.2 Rebuild and Renewal
2.2.3 1699–1703
2.3 The Crew of the Stirling Castle: March 1701 to November 1703
2.4 The Loss of the Stirling Castle: 26–28 November 1703
2.5 Conclusion
3: The Archaeological History of the Stirling Castle
3.1 Archaeological Investigation
3.1.1 Discovery and Exploration: 1979–1991
3.1.2 1992–2002
3.1.3 2003–2009
3.1.4 Conclusion
3.2 Ownership, Archive Access and Dissemination
3.2.1 Ownership
3.2.2 Archive Access
3.2.3 Archive Dissemination
3.3 The Stirling Castle in the Context of Maritime Archaeology in England and the UK
3.3.1 Methodology and Management
3.3.2 Resources, Public Access and Dissemination
3.3.2.1 Resources
3.3.2.2 Access and Dissemination
4: Site Environment and Geophysical Survey
4.1 The Natural Environment
4.1.1 Geology
4.1.2 Hydrodynamic Regime and Sediment Transport
4.1.3 Goodwin Sands Historical (pre-1990s) Evolution
4.2 Geophysical Survey
4.2.1 Singlebeam Bathymetry
4.2.1.1 1997
4.2.1.2 1998
4.2.1.3 1999
4.2.2 Multibeam Bathymetry
4.2.2.1 July 2002
4.2.2.2 April 2005
4.2.2.3 September 2005
4.2.2.4 March 2006
4.2.2.5 August 2006
4.2.2.6 September 2009
4.2.3 Sidescan Sonar
4.2.3.1 2000
4.2.3.2 2002
4.2.3.3 2005
4.2.3.4 2008
4.2.3.5 2009
4.2.4 Acoustic Ground Discrimination
4.2.5 Magnetometer
4.2.6 Grain Size Analysis
4.2.7 Sub-bottom
4.3 Meso- and Wreck-Scale Site Evolution
4.3.1 Annual Meso-Scale Site Evolution
4.3.1.1 July 2002–April 2005
4.3.1.2 April 2005–August 2006
4.3.1.3 August 2006–September 2009
4.3.2 Annual/Monthly Wreck-Scale Site Evolution
4.3.2.1 July 2002–April 2005
4.3.2.2 April 2005–September 2005
4.3.2.3 September 2005–March 2006
4.3.2.4 March 2006–August 2006
4.3.2.5 August 2006–September 2009
4.3.3 Maximum and Minimum Depth Locations
4.4 Discussion
5: The In Situ Archaeological Remains of the Stirling Castle
5.1 Initial Site Formation Processes
5.2 Site Development: Methodology
5.2.1 Observations on Overall Correlation
5.3 Site Development: 1979–2009
5.3.1 Discovery: 1979
5.3.2 Re-Emergence: 1998–2000
5.3.2.1 Stern Area
5.3.2.2 Amidships
5.3.2.3 Bow Area
5.3.2.4 Summary
5.3.3 Reburial: 2002–2009
5.4 Discussion
6: Stirling Castle: Ship Structure and Shipbuilding
6.1 Overview
6.1.1 Construction Summary
6.2 Framing
6.2.1 Futtocks
6.2.2 Deck Beams
6.2.3 Stern Structure
6.3 Planking and Fastening
6.4 Rudder and Steering Mechanism
6.5 Discussion
7: Organic Material
7.1 Methodology
7.2 Recovered Artefacts by Material
7.2.1 Plant Remains
7.2.2 Bamboo
7.2.3 Bone and Ivory
7.2.4 Cork
7.2.5 Fibre
7.2.6 Gourd and Coconut
7.2.7 Leather
7.2.8 Textile
7.2.9 Tar
7.2.10 Wood
7.2.10.1 Containers
7.2.10.2 Dress and Personal Accessories
7.2.10.3 Entertainment/Leisure
7.2.10.4 Food Preparation and Consumption
7.2.10.5 Navigation and Measurement
7.2.10.6 Tools and Equipment
7.2.10.7 Ship’s Fixtures and Fittings
7.2.10.8 Weapons
7.2.10.9 Miscellaneous
7.3 Discussion
7.3.1 Galley Objects
7.3.2 Dress Accessories
7.3.3 Personal/Leisure Items
7.3.4 Tools and Equipment
7.4 Conclusion
8: Glass
8.1 Introduction and Methodology
8.1.1 Scientific Examination and Analysis
8.2 Recovered Artefacts by Type
8.2.1 Bottles
8.2.1.1 Onion Bottles
8.2.1.2 Square-Sided Bottles
8.2.1.3 Cylindrical Bottles
8.2.2 Drinking Glasses
8.2.3 Timing Glasses
8.2.4 Window Glass
8.2.5 Miscellaneous
8.3 Scientific Examination and Analysis
8.3.1 Methods of Analysis
8.3.2 Results
8.3.2.1 Tableware
8.3.2.2 The Bottles
8.3.2.2 The Plate Window Glass
8.4 Discussion
9: Metal
9.1 Introduction and Methodology
9.1.1 Scientific Examination and Analysis
9.2 Eating and Drinking
9.2.1 Plates
9.2.2 Bowls, Dishes, Lids and Porringers
9.2.3 Utensils
9.2.4 Jugs, Tankards and Drinking Vessels
9.2.5 Cooking Equipment
9.3 Personal Artefacts
9.3.1 Clothing and Attire
9.3.2 Lighting
9.3.3 Miscellaneous
9.4 Ship’s Equipment and Fittings
9.4.1 Navigation and Measurement
9.4.2 Scuppers
9.4.3 Miscellaneous
9.4.4 Copper/Copper-Alloy Fastenings
9.4.5 Unidentified Concretions
9.5 Arms and Armament
9.5.1 Swords
9.5.2 Small Arms
9.5.3 Ammunition
9.5.4 Cannon
9.5.5 Shot
9.6 Scientific Examination and Analysis
9.6.1 Methods of Analysis
9.6.2 Results
9.6.2.1 Navigational Dividers
9.6.2.2 Pewter
9.7 Discussion
10: Pottery
10.1 Introduction and Methodology
10.2 The Pottery
10.3 Discussion
11: Stone, Brick and Clay Pipes
11.1 Introduction and Methodology
11.2 Recovered Artefacts by Type
11.2.1 Navigational/Deck Slates
11.2.2 Whetstones
11.2.3 Bricks
11.2.4 Clay Pipes
12: The Stirling Castle: 1677–2009
12.1 Fighting the Ship
12.1.1 Personal Weapons and Small Arms
12.2 Navigation
12.2.1 Navigational Practice c.1703
12.2.2 The Navigational Equipment of the Stirling Castle
12.3 Artefact Distribution
12.4 Reflections on the Archaeology of the Stirling Castle
12.4.1 Restoration Warship or Vessel of 1699?
12.4.2 Shipwreck and Historical Archaeology of the Stirling Castle
12.5 Conclusion
References
Appendices
Index
Back Cover

Citation preview

THE MARITIME ARCHAEOLO GY TRUST MONOGRAPH, SERIES 4

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Archaeological Investigations 1979-2009 EDITED BY

JULIAN WHITEWRIGHT BAR BRITISH SERIES 656

2020

THE MARITIME ARCHAEOLO GY TRUST MONOGRAPH, SERIES 4

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Archaeological Investigations 1979-2009 EDITED BY

JULIAN WHITEWRIGHT BAR BRITISH SERIES 656

2020

Published in 2020 by BAR Publishing, Oxford BAR British Series 656 The Maritime Archaeology Trust Monograph Series, volume 4 The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 ISBN  978 1 4073 5687 7 paperback ISBN  978 1 4073 5688 4 e-format doi  https://doi.org/10.30861/9781407356877 A catalogue record for this book is available from the British Library © the editor and contributors severally 2020 Cover image  Timing glass (No 174] (Maritime Archaeology Trust) The Authors’ moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. Links to third party websites are provided by BAR Publishing in good faith and for information only. BAR Publishing disclaims any responsibility for the materials contained in any third-party website referenced in this work.

BAR titles are available from: Email Phone Fax

BAR Publishing 122 Banbury Rd, Oxford, ox2 7bp, uk [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

The Maritime Archaeology Trust The Maritime Archaeology Trust promotes interest, research and knowledge of maritime archaeology and heritage. Our objectives are to: • • • •

Investigate: maritime, coastal and underwater archaeology. Engage: people, communities and schools through involvement, enjoyment and education. Promote: hard to reach and inaccessible maritime heritage. Protect: through supporting heritage management for current and future generations.

Other Titles in the Series Analysing Maritime Archaeological Archives Collections, Access and Management Julie Satchell with contributions by Jesse Ransley and Julian Whitewright Oxford, BAR Publishing, 2016 Volume 3

BAR British Series 628

The Maritime Archaeology of Alum Bay Two shipwrecks on the north-west coast of the Isle of Wight, England Julie Satchell and Julian Whitewright with a foreword by Garry Momber and contributions by Nigel Nayling, Peter Northover, Shirley Northover, Nick Cokes, Philippa Naylor, Florencis Malamud, Joe Kelleher, Jon James and Paul Simpson Oxford, BAR Publishing, 2014 Volume 2

BAR British Series 608

The Archaeology and History of the Flower of Ugie, Wrecked 1852 in the Eastern Solent Edited by Julian Whitewright and Julie Satchell Oxford, BAR Publishing, 2011 Volume 1

BAR British Series 551

Of Related Interest

Stirling Castle Digital Archive The archive assessment for the Stirling Castle that was undertaken to develop this publication includes a database and artefact images which are available from the Archaeology Data Service: https://doi.org/10.5284/1058978 Stirling Castle Virtual Dive Trail Explore the site of the Stirling Castle through a new online resource available from https://historicengland.org.uk/get-involved/visit/protected-wrecks/virtual-dive-trails/ HM Submarine A7 An Archaeological Assessment: A report on the results of the A7 Project 2014 Peter Holt Oxford, BAR Publishing, 2015

BAR British Series 613

The Gresham Ship Project A 16th-Century Merchantman Wrecked in the Princes Channel, Thames Estuary Volume II: Contents and Context Edited by Gustav Milne and Dean Sully with contributions by Mark Beattie-Edwards, Lynn Biggs, Thomas Birch, Michael F. Charlton, Kelly Domoney, Clare Hunt, Phil Magrath, Marcos Martinón-Torres and Zofia Stos-Gale Oxford, BAR Publishing, 2014

BAR British Series 606

The Gresham Ship Project A 16th-Century Merchantman Wrecked in the Princes Channel, Thames Estuary Volume I: Excavation and Hull Studies Edited by Jens Auer and Thijs J. Maarleveld with contributions by Massimiliano Ditta, Antony Firth, Nigel Nayling, Delia Ní Chíobháin, Christian Thomsen, and Cate Wagstaffe Oxford, BAR Publishing, 2014

BAR British Series 602

The Hulks of Forton Lake, Gosport The Forton Lake Archaeology Project 2006–2009 Mark Beattie-Edwards and Julie Satchell with contributions by Paul Donohue, Mary Harvey, Alison James, Colin McKewan, Jane Maddocks, Daniel Pascoe, Philip Simons and Julian Whitewright Oxford, BAR Publishing, 2011

BAR British Series 536

For more information, or to purchase these titles, please visit www.barpublishing.com

Contents

Acknowledgements............................................................................................................................................................ ix Abbreviations...................................................................................................................................................................... x Series Editorial................................................................................................................................................................... xi Summary............................................................................................................................................................................ xii List of Figures................................................................................................................................................................... xvi List of Tables................................................................................................................................................................... xxiii 1. Introduction..................................................................................................................................................................... 1 Julian Whitewright 2. Vessel History: Building, Service and Loss................................................................................................................... 5 Julian Whitewright 2.1 Background (1660–1674).......................................................................................................................................... 5 2.1.1 The Thirty Ships (1674–1688)........................................................................................................................... 6 2.2 The Active Service of the Stirling Castle................................................................................................................... 7 2.2.1 1690–1699.......................................................................................................................................................... 7 2.2.2 Rebuild and Renewal......................................................................................................................................... 8 2.2.3 1699–1703.......................................................................................................................................................... 9 2.3 The Crew of the Stirling Castle: March 1701 to November 1703........................................................................... 10 2.4 The Loss of the Stirling Castle: 26–28 November 1703.......................................................................................... 11 2.5 Conclusion............................................................................................................................................................... 14 3. The Archaeological History of the Stirling Castle...................................................................................................... 15 Julian Whitewright 3.1 Archaeological Investigation.................................................................................................................................... 15 3.1.1 Discovery and Exploration: 1979–1991........................................................................................................... 15 3.1.2 1992–2002........................................................................................................................................................ 19 3.1.3 2003–2009........................................................................................................................................................ 22 3.1.4 Conclusion........................................................................................................................................................ 25 3.2 Ownership, Archive Access and Dissemination...................................................................................................... 26 3.2.1 Ownership........................................................................................................................................................ 26 3.2.2 Archive Access................................................................................................................................................. 26 3.2.3 Archive Dissemination..................................................................................................................................... 27 3.3 The Stirling Castle in the Context of Maritime Archaeology in England and the UK............................................ 28 3.3.1 Methodology and Management........................................................................................................................ 28 3.3.2 Resources, Public Access and Dissemination.................................................................................................. 29 4. Site Environment and Geophysical Survey................................................................................................................ 31 Amelia Couldrey 4.1 The Natural Environment......................................................................................................................................... 31 4.1.1 Geology ........................................................................................................................................................... 31 4.1.2 Hydrodynamic Regime and Sediment Transport............................................................................................. 32 4.1.3 Goodwin Sands Historical (pre-1990s) Evolution........................................................................................... 36 4.2 Geophysical Survey................................................................................................................................................. 38 4.2.1 Singlebeam Bathymetry................................................................................................................................... 38 4.2.2 Multibeam Bathymetry.................................................................................................................................... 40 4.2.3 Sidescan Sonar................................................................................................................................................. 43 4.2.4 Acoustic Ground Discrimination...................................................................................................................... 46 4.2.5 Magnetometer................................................................................................................................................... 46 4.2.6 Grain Size Analysis.......................................................................................................................................... 49 4.2.7 Sub-bottom ...................................................................................................................................................... 49

v

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 4.3 Meso- and Wreck-Scale Site Evolution................................................................................................................... 49 4.3.1 Annual Meso-Scale Site Evolution.................................................................................................................. 49 4.3.2 Annual/Monthly Wreck-Scale Site Evolution.................................................................................................. 50 4.3.3 Maximum and Minimum Depth Locations����������������������������������������������������������������������������������������������������� 53 4.4 Discussion................................................................................................................................................................ 53 5. The In Situ Archaeological Remains of the Stirling Castle........................................................................................ 55 Julian Whitewright 5.1 Initial Site Formation Processes............................................................................................................................... 55 5.2 Site Development: Methodology............................................................................................................................. 56 5.2.1 Observations on Overall Correlation................................................................................................................ 58 5.3 Site Development: 1979–2009................................................................................................................................. 59 5.3.1 Discovery: 1979............................................................................................................................................... 59 5.3.2 Re-Emergence: 1998–2000.............................................................................................................................. 61 5.3.3 Reburial: 2002–2009........................................................................................................................................ 64 5.4 Discussion................................................................................................................................................................ 65 6. Stirling Castle: Ship Structure and Shipbuilding....................................................................................................... 67 Julian Whitewright 6.1 Overview.................................................................................................................................................................. 67 6.1.1 Construction Summary..................................................................................................................................... 67 6.2 Framing.................................................................................................................................................................... 69 6.2.1 Futtocks............................................................................................................................................................ 69 6.2.2 Deck Beams...................................................................................................................................................... 71 6.2.3 Stern Structure.................................................................................................................................................. 72 6.3 Planking and Fastening............................................................................................................................................ 73 6.4 Rudder and Steering Mechanism............................................................................................................................. 74 6.5 Discussion................................................................................................................................................................ 75 7. Organic Material........................................................................................................................................................... 77 Kathryn Dagless, Angela Middleton, Gill Campbell, Polydora Baker, Simon Mays and Mariangela Vitolo 7.1 Methodology............................................................................................................................................................ 77 7.2 Recovered Artefacts by Material.............................................................................................................................. 77 7.2.1 Plant Remains................................................................................................................................................... 77 7.2.2 Bamboo............................................................................................................................................................ 78 7.2.3 Bone and Ivory................................................................................................................................................. 78 7.2.4 Cork.................................................................................................................................................................. 79 7.2.5 Fibre................................................................................................................................................................. 79 7.2.6 Gourd and Coconut.......................................................................................................................................... 80 7.2.7 Leather.............................................................................................................................................................. 80 7.2.8 Textile .............................................................................................................................................................. 86 7.2.9 Tar..................................................................................................................................................................... 86 7.2.10 Wood .............................................................................................................................................................. 86 7.3 Discussion ............................................................................................................................................................... 99 7.3.1 Galley Objects.................................................................................................................................................. 99 7.3.2 Dress Accessories........................................................................................................................................... 100 7.3.3 Personal/Leisure Items................................................................................................................................... 100 7.3.4 Tools and Equipment...................................................................................................................................... 101 7.4 Conclusion............................................................................................................................................................. 101 8. Glass ............................................................................................................................................................................ 103 David Dungworth and Julian Whitewright 8.1 Introduction and Methodology............................................................................................................................... 103 8.1.1 Scientific Examination and Analysis.............................................................................................................. 103 8.2 Recovered Artefacts by Type................................................................................................................................. 103 8.2.1 Bottles ............................................................................................................................................................ 103 8.2.2 Drinking Glasses............................................................................................................................................ 109 8.2.3 Timing Glasses................................................................................................................................................111 8.2.4 Window Glass................................................................................................................................................ 115 8.2.5 Miscellaneous................................................................................................................................................. 116 vi

Contents 8.3 Scientific Examination and Analysis...................................................................................................................... 116 8.3.1 Methods of Analysis....................................................................................................................................... 116 8.3.2 Results............................................................................................................................................................ 116 8.4 Discussion.............................................................................................................................................................. 122 9. Metal ............................................................................................................................................................................ 125 David Dungworth, Julian Whitewright and Angela Middleton 9.1 Introduction and Methodology............................................................................................................................... 125 9.1.1 Scientific Examination and Analysis.............................................................................................................. 125 9.2 Eating and Drinking............................................................................................................................................... 125 9.2.1 Plates.............................................................................................................................................................. 125 9.2.2 Bowls, Dishes, Lids and Porringers............................................................................................................... 127 9.2.3 Utensils........................................................................................................................................................... 129 9.2.4 Jugs, Tankards and Drinking Vessels............................................................................................................. 130 9.2.5 Cooking Equipment........................................................................................................................................ 130 9.3 Personal Artefacts.................................................................................................................................................. 133 9.3.1 Clothing and Attire......................................................................................................................................... 133 9.3.2 Lighting.......................................................................................................................................................... 134 9.3.3 Miscellaneous................................................................................................................................................. 135 9.4 Ship’s Equipment and Fittings������������������������������������������������������������������������������������������������������������������������������ 139 9.4.1 Navigation and Measurement......................................................................................................................... 139 9.4.2 Scuppers......................................................................................................................................................... 140 9.4.3 Miscellaneous................................................................................................................................................. 140 9.4.4 Copper/Copper-Alloy Fastenings................................................................................................................... 142 9.4.5 Unidentified Concretions................................................................................................................................ 142 9.5 Arms and Armament.............................................................................................................................................. 143 9.5.1 Swords............................................................................................................................................................ 143 9.5.2 Small Arms..................................................................................................................................................... 144 9.5.3 Ammunition.................................................................................................................................................... 145 9.5.4 Cannon........................................................................................................................................................... 145 9.5.5 Shot................................................................................................................................................................. 146 9.6 Scientific Examination and Analysis...................................................................................................................... 146 9.6.1 Methods of Analysis....................................................................................................................................... 146 9.6.2 Results............................................................................................................................................................ 147 9.7 Discussion.............................................................................................................................................................. 150 10. Pottery........................................................................................................................................................................ 153 Duncan H. Brown 10.1 Introduction and Methodology............................................................................................................................. 153 10.2 The Pottery........................................................................................................................................................... 155 10.3 Discussion............................................................................................................................................................ 159 11. Stone, Brick and Clay Pipes..................................................................................................................................... 165 Julian Whitewright and Kathryn Dagless 11.1 Introduction and Methodology............................................................................................................................. 165 11.2 Recovered Artefacts by Type............................................................................................................................... 165 11.2.1 Navigational/Deck Slates............................................................................................................................. 165 11.2.2 Whetstones................................................................................................................................................... 167 11.2.3 Bricks............................................................................................................................................................ 167 11.2.4 Clay Pipes .................................................................................................................................................... 168 12. The Stirling Castle: 1677–2009................................................................................................................................. 171 Julian Whitewright and Christin Heamagi 12.1 Fighting the Ship.................................................................................................................................................. 171 12.1.1 Personal Weapons and Small Arms.............................................................................................................. 174 12.2 Navigation............................................................................................................................................................ 174 12.2.1 Navigational Practice c.1703........................................................................................................................ 174 12.2.2 The Navigational Equipment of the Stirling Castle..................................................................................... 175 12.3 Artefact Distribution............................................................................................................................................ 177

vii

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 12.4 Reflections on the Archaeology of the Stirling Castle ........................................................................................ 179 12.4.1 Restoration Warship or Vessel of 1699?....................................................................................................... 180 12.4.2 Shipwreck and Historical Archaeology of the Stirling Castle..................................................................... 181 12.5 Conclusion........................................................................................................................................................... 183 References........................................................................................................................................................................ 187 Appendices....................................................................................................................................................................... 195 14.1 The Crew of the Stirling Castle: March 1701–November 1703.......................................................................... 195 14.2 The Stirling Castle in the Ordnance Survey of c.1696........................................................................................ 228 Index................................................................................................................................................................................. 231

viii

Acknowledgements This publication is the culmination of a number of stages of work on the archive from the Stirling Castle protected wreck site. The work has been funded by Historic England (formerly English Heritage). As outlined within the volume, site investigations began in 1979 and, in the intervening 40 years, many individuals, organisations and museums have been involved in archaeological work on the wreck site and/ or are now custodians of elements of the site archive. It is only with the generous help and assistance of those involved with the site that it has been possible to bring this material together.

366 (Fig 7.6), 395 (Fig 10.2), 423 (Fig 7.9), 437 (Fig 10.13), 450 (Fig 10.19), 462 (Fig 10.14), 459-61 and 464467 (Fig 10.1), 474 (Fig 11.10), 475 (Fig 11.11), 476 (Fig 11.12), 513 (Fig 9.5), 514 (Fig 7.10). Joshua Dack: 73 (Fig 8.11a), 170 (Fig 8.8a), 172 (Fig 8.9a), 173 (Fig 8.8b), 174 (Fig 8.9b), 179 (Fig 7.22), 180 (Fig 8.6a), 181 (Fig 8.11b), 182 (Fig 8.10), 185 (Fig 8.2a), 189 (Fig 8.2b), 192 (Fig 8.2c), 198 (Fig 8.7a), 227 (Fig 7.34), 250 (Fig 9.10), 251 (Fig 7.23), 276 (Fig 9.12), 291 (Fig 7.19), 313 (Fig 7.2), 329 (Fig 7.26), 334 (Fig 9.1a), 336 (Fig 9.4d), 337 (Fig 9.1b), 343 (Fig 9.1c), 344 (Fig 9.1d), 354 (Fig 7.28), 377 (Fig 7.29), 382 (Fig 7.30), 394 (Fig 9.2), 401 (Fig 9.8), 407 (Fig 8.7b), 413 (Fig 8.4b), 414 (Fig 8.6b), 415 (Fig 8.6c), 416 (Fig 8.4d), 417 (Fig 8.4a), 420 (Fig 8.4e), 421 (Fig 8.11c), 422 (Fig 8.6d), 489 (Fig 7.31), 494 (Fig 7.32), 525 (Fig 7.24), 535 (Fig 9.3), 593 (Fig 8.2e), 600 (Fig 8.4c), 607 (Fig 8.2d). Other illustrations kindly provided by individuals or organisations are acknowledged within the Figure captions.

First-hand accounts of the work undertaken on the site have proved invaluable for understanding the rationale and processes behind such work. Therefore, we would like to thank Tom Brown, John Chamberlain and Diane Chamberlain for providing insightful recollections of the 1979 seasons of work. We would also like to thank Bob Peacock, long-time site licensee, for his help with the project and for his extended efforts towards the stewardship of the site over many years.

We would like to thank David Dungworth, Angela Middleton, Duncan Brown, Gill Campbell, Polydora Baker, Simon Mays and Mariangela Vitolo of Historic England and Peter Northover of Oxford Materials for the investigation and analysis of various classes of material as part of this stage of the wider archive project. Likewise, we would like to thank Amelia Couldrey for her work on the archive of geophysical material.

A number of other people and organisations who have worked on the site or have been involved with archive material have also helped with the project, including the late Norman Temple, the late Bryan Smith, Ann Smith, Peter Holt, the late Martin Dean, Richard Bates, Ges Moody, Jenny Price, Richard Dunn, Historic England, the National Maritime Museum, Wessex Archaeology, the Mary Rose Trust and the Trust for Thanet Archaeology.

A number of staff, past and present, from the Maritime Archaeology Trust (previously the Hampshire and Wight Trust for Maritime Archaeology) have worked on the project since its inception. These include Julian Whitewright, Julie Satchell, Brandon Mason, Jasmine Noble-Shelley, Christin Heamagi, Kathryn Dagless, Virginia Dellino-Musgrave, Douglas McElvogue, Mark James, Alison James and Lauren Tidbury.

The archive of material from the site now resides in several locations and access to material has been made possible by a number of different people: Margaret Symonds (Isle of Thanet Archaeological Society), Michael Cates (Ramsgate Maritime Museum) and Jacqui Stanford (Hastings Shipwreck Museum). We would like to express our gratitude for their help. Illustrations of artefacts for this publication have been produced by Maggy Redmond and Joshua Dack, members and supporters of the Isle of Thanet Archaeological Society. They have produced the following illustrations of artefacts, listed by their ID numbers, that have formed or been incorporated into the figures indicated in brackets. Maggy Redmond: 106 (Fig 11.3), 233 (Fig 10.16), 235 (Fig 10.15), 240 (Fig 11.4), 241 (Fig 11.5), 242 (Fig 11.6), 243 (Fig 11.7), 252 (Fig 10.10), 255 (Fig 10.11), 267 (Fig 10.12), 268 (Fig 10.18), 273 (Fig 10.5), 274 (Fig 10.17), 290 (Fig 7.4), 304 (Fig 9.13), 305 (Fig 9.14), 328 (Fig 11.8), 331 (Fig 10.8), 332 (Fig 7.5), 335 (Fig 11.9), 341 (Fig 10.20), 348 (Fig 9.16), 355 (Fig 11.2), 365 (Fig 9.17),

ix

Abbreviations ACHWS

Advisory Committee on Historic Wreck Sites

ADU

Archaeological Diving Unit

AGDS

Acoustic ground discrimination systems

BSAC

British Sub-Aqua Club

CD

Chart Datum

EDXRF

Energy dispersive x-ray fluorescence

IHO

International Hydrographic Organisation

IOTAS

Isle of Thanet Archaeological Society

IOTAU

Isle of Thanet Archaeological Unit

MAS

Marine Archaeological Surveys

MCA

Maritime and Coastguard Agency

MOD

Ministry of Defence

NAS

Nautical Archaeology Society

NMM

National Maritime Museum

NRHE

National Record of the Historic Environment

ODN

Ordnance Datum Newlyn

PWA

Protection of Wrecks Act

RASSE

Rapid Archaeological Site Surveying and Evaluation

RMM

Ramsgate Maritime Museum

TAU

Thanet Archaeological Unit

TfTA

Trust for Thanet Archaeology

TNA

The National Archives

UCH

Underwater cultural heritage

UKHO

UK Hydrographic Office

VORF

Vertical Offshore Reference Framework

WA

Wessex Archaeology

x

Series Editorial The fourth volume in the MAT monograph series focuses on the wreck of the Stirling Castle, a highly significant vessel that provides evidence of Samuel Pepys’ Thirty Ships building programme and life onboard an early eighteenthcentury warship, and is representative of the historic event of the Great Storm in 1703. When the ship was discovered in 1979 the levels of preservation were stunning, with the two-decked wooden warship almost complete: it was a true ‘time capsule’ wreck. For 40 years the site has been recognised as one of Britain’s most significant, being designated under the Protection of Wrecks Act (PWA) in 1980. During this time the wreck has been subject to the dynamics of a constantly moving seabed which at times protects the remains, while at others exposing it to threat and ongoing degradation and loss, which continues today.

Finally, I would like to include a special mention of our colleague and dear friend Kathryn Dagless, who passed away in 2017 while work on this volume was ongoing. Kathryn was a finds specialist who analysed organic material and stone from the site and her work has enriched this volume. Her passion for artefacts is much missed. Julie Satchell Series editor

In chapter three Whitewright examines the history of the investigation of the wreck, stating that it can be seen as representative of the problems and challenges that have characterised maritime archaeology within England and the wider UK since the inception of the PWA in 1973. But, on a more positive note, it can also be viewed as symbolising the efforts of a generation of maritime archaeologists to address and solve some of the problems that they have inherited, while attempting to avoid the creation of similar circumstances in the future. There have, indeed, been enormous ‘efforts’ expended on the wreck, most of which have been undertaken by unfunded volunteer groups. The site demonstrates the value and challenges of long-running maritime investigations to which a wide range of individuals, groups, organisations, societies and companies contribute. These investigations, which take place on many Protected Wreck sites, often result in ‘open’ archives that accumulate over time and can be split over multiple locations and owners. While this has posed challenges for this project, the results demonstrate what it is possible to achieve through the assessment and analysis of these long-term, dispersed archives and how this work is vital for developing access to the data and promoting research, interpretation and reinterpretation as more comparative material becomes available in the future. The MAT is pleased to publish the results of this work within our monograph series. It is apt that the volume follows ‘Analysing Maritime Archaeological Archives’, which examined issues related to the lack of analysis and dissemination of shipwreck investigations. The culmination of the work on the Stirling Castle archive shows the rich potential within ‘back-log’ maritime archives and we hope it helps support the case for more such studies, resulting in future publications.

xi

Summary be the highest level of coherent archaeological remains, with debris from the upper deck, forecastle and quarter deck collapsed upon it. Beneath this, the orlop deck and hold are assumed to remain, while parts of the transom and stern quarters survive at the eastern end of the site. Within this volume the vessel’s framing, planking and fastening components are described with comparative material, both historic and archaeological, being used to contextualise and understand that of the Stirling Castle. Chapters considering artefacts by material include a catalogue and offer comparative analysis where possible and practical.

The Stirling Castle was a 70-gun third-rate ship-of-the-line, built in Deptford in 1679, rebuilt at Chatham in 1699 and lost on the Goodwin Sands during the Great Storm of 1703. The ship is highly significant from a historical perspective because it was launched as part of the Thirty Ships building programme overseen by Samuel Pepys and Charles II. The vessel formed part of the Royal Navy fleet in the early years of the War of Spanish Succession, with its operational life spanning a critical period in the development of the Royal Navy not just from the perspective of ship technology but also in terms of Britain’s becoming arguably the preeminent naval power in northern Europe. Added to this was the loss of the ship and crew in one of the most powerful and destructive storms to have been experienced and recorded in English history.

The large assemblage of organic remains demonstrates the preservation potential of the marine environment and includes plant remains, bone artefacts, rope, a range of leather objects including shoes and book covers, and textiles. Wooden objects range from containers, food preparation and consumption equipment and dress and personal accessories through to navigation and measurement equipment, tools and weapons. The organic assemblage is significant owing both to its variety and to the rarity of discovering such large numbers of organic artefacts on one site. It represents many aspects of life on board, from general daily tasks (e.g. scrubbing the desks) and specialist occupations through to leisure-time pursuits. The hierarchy of the crew is represented through the materials from which the artefacts are produced and their elaboration, from the basic, functional, undecorated objects used by the lower crew members through to fragments of expensive silk fabrics from the clothing of the higher ranks.

Following the discovery of the site in 1979 it has been subject to continued archaeological activity and has been designated under the 1973 Protection of Wrecks Act (PWA) since 1980. The wreck lies in a dynamic environment in which the movement of sand causes the site to be covered and uncovered, resulting in significant degradation; however, it is still an exceptional site, being one of the bestpreserved shipwrecks from the late seventeenth century in English waters. Numerous individuals, groups, Trusts and organisations have been involved in the investigation of the site over the past 40 years, which has resulted in the recovery of a significant artefact assemblage in addition to an extensive supporting archive recording the vessel structure and other in situ material. Work on the archive generated between 1979 and 2009, which is dispersed across a number of locations, has been funded by Historic England (previously English Heritage) and sought to summarise current understanding as an aid to long-term management and to encourage further research. This resulting publication is split into two parts. Part one considers the history of the Stirling Castle as a warship, covering its building, service and loss from both a broad contextual view and a more detailed view of the vessel and crew. A detailed account is then provided of the archaeological investigation of the site and the environmental factors impacting the seabed remains. The review of available geophysical survey data supports the presentation of the movement of sediments surrounding the wreck, analysis that provides the context for the periods of exposure and coverage.

Glass artefacts consist of bottles (onion, square-sided and cylindrical), drinking glasses, timing glasses and window glass. A number have been subject to scientific analysis, providing detail on their chemical composition. Over 230 metal artefacts range from personal possessions through fighting weaponry to parts of the ship structure. A number of metals are represented within the assemblage, with iron, pewter, copper, brass and lead being most notable. They are presented and analysed by functional category, including eating and drinking (plates, bowls, dishes, porringers, utensils, jugs, tankards, cooking equipment), personal artefacts (clothing, attire, lighting and miscellaneous), ship’s equipment and fittings (navigation and measurement, scuppers, fastenings) and arms and armament (swords, small arms, ammunition, cannon and shot). The collection of 57 pots from the site represents a wide variety of countries and production centres, reflecting the sailing history of the ship, the character of the crew or the pottery types available in the port of origin. They were used for a range of functions – storing foodstuffs, cooking,

Part Two presents the artefact archive by material and type in a series of chapters. Consideration of the ship’s construction uses records of structure remaining on the seabed as well as recovered material. The seabed remains consist of the vessel’s main deck, which is considered to

xii

Summary eating and drinking, and storing ointments and drugs. Stone and other ceramic artefacts include navigational or deck slates, whetstones, bricks and clay pipes.

De nombreux individus, groupes, trusts,organisations, se sont impliqués dans l’investigation du site au cours des 40 dernières années, ce qui a eu pour résultat la récupération d’un important assemblage d’artifacts, qui, en plus, s’appuie sur de nombreuses archives enregistrant la structure du vaisseau et d’autres matériaux in situ. Des travaux sur les archives générés entre 1979-2009 et qui sont dispersés entre divers endroits ont été financés par Historic England (anciennement English Heritage) et cherchaient à résumer la compréhension actuelle en tant qu’aide à sa gestion à long terme et à encourager de nouvelles recherches. Cette publication qui en est le résultat, est divisée en deux parties. La partie un examine l’histoire du Stirling Castle en tant que navire de guerre, couvrant sa construction, ses années de service et sa perte à la fois d’un large point de vue étendu au contexte et d’une vision plus détaillée du vaisseau et de son équipage. On offre ensuite un compterendu détaillé de l’investigation archéologique du site et de l’impact des facteurs environnementaux sur les vestiges restés sur le fond marin. Une revue des données des études géophysiques disponibles vient appuyer la présentation du mouvement des sédiments qui entourent l’épave, analyse qui fournit le contexte pour les périodes d’exposition et de couverture.

A review of the entire archive of material draws out wider trends to present an understanding of the ship, its crew and the archaeological record they have left behind. Interpretation considers two key questions related to the archaeological record: • How much of the original 1679 ship, in structural or equipment terms, survived its 1699 rebuild to become part of the archaeological remains of the vessel following its loss in 1703? • What was the impact of the wrecking event on the distribution of artefacts and structural materials on the seabed and how much did it reorganise the relationships of the people, their possessions and the equipment that they used on board the vessel? Where possible, these questions are addressed alongside further discussion of themes such as the fighting ship, navigation and artefact distribution. Conclusions reiterate how the archaeological remains of the Stirling Castle offer a compelling insight from a range of perspectives into the maritime world of the seventeenth and eighteenth centuries.

La partie deux présente les archives des artifacts par matière et type en une série de chapitres. L’étude de la construction du navire utilise les relevés de la structure du navire restée au fond de la mer ainsi que des matériaux recouvrés. Les vestiges au fond de la mer sont constitués du pont principal du vaisseau, considéré comme étant le niveau le plus élevé de vestiges archéologiques cohérents. Sur celui-ci, se sont affaissés des débris du pont supérieur, du pont de gaillard d’avant et du quart de pont. Là dessous le faux pont et la soute sont supposés avoir subsisté tandis que des parties de la traverse et des quartiers de poupe ont survécu à l’extrémité est du site. Dans le cadre de ce volume les parties constituantes de la carcasse, du bordage et de l’assemblage du vaisseau sont décrites, du matériel comparable, à la fois historique et archéologique, étant utilisé pour replacer dans son contexte et comprendre celui de Stirling Castle. Les chapitres qui examinent les artifacts par matière comprennent un catalogue et offrent une analyse comparative là où elle est possible et pratique.

Résumé Le Stirling Castle était un navire de combat en ligne de file de 70 canons troisième rang, construit à Deptford en 1679, reconstruit à Chatham en 1699 et perdu sur les sables de Goodwind Sands au cours de la grande tempête de 1703. Le navire est extrèmement important d’un point de vue historique parce qu’ il fut mis à l’eau faisant partie du programme de constructions navales des Trente Bateaux sous la supervision de Samuel Pepys et Charles II. Le vaisseau faisait partie de la flotte de la Marine Royale dans les premières années de la Guerre de Succession d’Espagne, sa durée de service actif couvrant une période critique du développement de la Marine Royale. Non seulement du point de vue de la technologie maritime mais aussi durant la période pendant laquelle la Grande-Bretagne est sans doute devenue la puissance navale proéminente en Europe du nord. A cela vint s’ajouter la perte du navire et de son équipage dans une des tempêtes les plus puissantes et les plus destructrices jamais vécues et relatées de ‘histoire de l’Angleterre.

Le vaste assemblage de vestiges organiques démontre le potentiel de préservation de l’environnement marin et comprend des restes de plantes, des artifacts en os, de la corde, une gamme d’objets en cuir, y compris des chaussures et couvertures de livres et des textiles. Les objets en bois vont des récipients, équipement pour la préparation et consommation de nourriture, habillement et accessoires personnels jusqu’aux instruments de navigation et de mesure, outils et équipement, et armes. L’assemblage organique est important en raison de la rareté de la découverte d’un si grand nombre d’artifacts organiques sur un seul site et est du à sa variété. L’assemblage réprésentée de nombreux aspects de la vie à bord du navire depuis les taches ordinaires quotidiennes, (par exemple frotter les ponts) et des occupations spécialisées en allant jusqu’aux

Suite à sa découverte en 1979, le site a fait l’objet d’activités archéologiques continues et est désigné sous l’acte de protection des épaves de 1973 (PWA) depuis1980. L’épave git dans un environnement dynamique avec des mouvements de sables qui font que le site est tantôt couvert, tantôt découvert, ce qui a eu comme résultat une importante dégradation, toutefois, il n’en reste pas moins un site exceptionnel, étant l’une des épaves les mieux conservées de la fin du 17ième siècle dans les eaux anglaises. xiii

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Zusammenfassung

loisirs. La hiérarchie de l’équipage est representée à travers les matériaux dont sont fabriqués les artifacts par exemple les objets ordinaires; utiles non décorés utilisés par les membres inférieurs de l’équipage jusqu’aux fragments de riches tissus de soie qui seraient venus des vêtements des rangs les plus élevés.

Die Stirling Castle war ein 70-Kanonen Schiff der 3. Klasse, wurde in 1679 in Deptford gebaut, in 1699 in Chatham renoviert und ist während des Großen Sturms von 1703 auf den Goodwin Sands gesunken. Das Schiff ist aus historischer Sicht von großer Bedeutung, da es im Rahmen eines Bauprogramms von dreissig Schiffen unter der Leitung von Samuel Pepys und Charles II. gebaut wurde. Das Schiff gehört zur Royal Navy Flotte in den ersten Jahren des Spanischen Erbfolgekrieges, dessen Betriebsdauer sich über eine kritische Phase in der Entwicklung der Royal Navy erstreckte, nicht nur aus der Perspektive der Schiffstechnik, sondern auch weil in dieser Zeit Großbritannien wohl zur herausragenden Seemacht in Nordeuropa wurde. Hinzu kam der Verlust des Schiffes und der Besatzung in einem der mächtigsten und zerstörerischsten Stürme, die in der englischen Geschichte erlebt und aufgezeichnet wurden.

Les objets en verre consistent en flacons (oignon, à cotés carrés et cylindriques).verres à boire, verres à messurer le temps, verre à vitre dont un certain nombre a fait l’objet d’une analyse scientifique qui a contribué des détails sur leur composition chimique. Plus de 230 objets en métal incluent des classes de matériaux allant de possessions personnelles en passant par des armes de combat à des parties de la structure du navire. Un certain nombre de métaux sont réprésentées à l’intérieur d’un assemblage; fer, étain, cuivre, laiton et plomb, étaient les plus remarquables. Ils sont présentés et analysés par fonction y compris manger et boire (assiettes, bols, plats, écuelles, ustensiles, pichets, chopes et batterie de cuisine), artifacts personnels, (vêtements, costumes, éclairages et divers), equipment et appareillages de navires (navigation et mesures, dalots) attaches et armes et armement (épées, petites armes, munitions, canons, et boulets).

Nach der Entdeckung des Fundortes im Jahr 1979 ist es Gegenstand anhaltender archäologischer Aktivitäten und wird seit 1980 unter dem Schutzgesetz für Wracks (Protection of Wrecks Act PWA) von 1973 anerkannt. Das Wrack liegt in einer dynamischen Umgebung und wird wechselseitig vom Sand abgedeck oder nicht abgedeckt, was zu einer erheblichen Erosion geführt hat. Dennoch ist es immer noch ein außergewöhnlicher Fund, der zu den am besten erhaltenen Schiffswracks aus dem späten 17. Jahrhundert in englischen Gewässern zählt.

La collection de 57 pots provenant du site représente une grande variété de pays et de centres de production reflétant soit l’histoire maritime du navire, le caractère de l’équipage, les types de poteries disponibles dans le port d’origine. Ils étaient utilisés pour une gamme de fonctions, conserver des produits comestibles, cuisiner, manger et boire, pots à onguent, flacons à remèdes. Les objets en pierre comprennent des ardoises de navigation ou de pont, des pierres àguiser, des briques et des pipes en argile.

Zahlreiche Einzelpersonen, Gruppen und Organisationen waren in den letzten 40 Jahren an der Untersuchung des Fundes beteiligt, was in einer signifikanten Ansammlung von Artefakten resultierte und einem umfangreichen unterstützenden Archiv, das den Schiffbau und sonstige Standortmaterialien enhält. Die Arbeiten an dem Archiv, das zwischen 1979 und 2009 entstanden ist und über eine Reihe von Standorten verteilt ist, wurden von Historic England (früher English Heritage) finanziert und war bestrebt das derzeitige Verständnis zusammen zufassen, um eine langfristige Verwaltung und weitere Forschung zu ermöglichen.

Une revue de l’ensemble du matériel archivé fait ressortir des tendances plus larges pour présenter une compréhension du navire et de son équipage et des témoignages archéologiques qu’ils ont laissés derrière eux. L’interprération examine deux questions clefs en rapport avec les dossiers archéologiques: • Combien du navire original de 1679, en terme de stucture ou d’équipements, a survécu à sa reconstruction en 1699 pour devenir une partie des vestiges archéologiques du vaisseau à la suite de sa perte en 1703; et • Quel fut l’ impact du naufrage sur la répartition des artifacts et des matériaux de la structure sur les fonds marins et dans quelle mesure il a réorganisé les relations entre les personnes et leurs biens et l’équipement qu’ils utilisaient àbord du vaisseau.

Diese daraus resultierende Veröffentlichung gliedert sich in zwei Teile: Teil 1 betrachtet die Geschichte des Stirling Castle als Kriegsschiff, den Schiffsbau, Schiffs-Dienst und den Verlust des Schiffes sowohl aus der Sicht des weiteren Zusammenhanges als auch aus einer detaillierteren Sicht auf das Schiff und seine Besatzung. Anschließend werden die archäologischen Untersuchungen des Geländes und die Umweltfaktoren, die sich auf den verbleibenden Meeresbodenreste auswirken, ausführlich beschrieben. Die Analyse der verfügbaren geophysikalischen Erhebungsdaten unterstützt die Darstellung, dass sich Sedimente rund um das Wrack bewegen, deren Analyse den Kontext für Zeiten der Expositions- und Abdeckung des Wracks liefern.

Là où c’est possible, ces questions sont abordées à-côté d’une prolongation de la discussion de thèmes tels que le navire au combat, la navigation et la répartition des artifacts. Les conclusions reitèrent comment les vestiges archéologiques du Sterling Castle offrent une vision fascinante du monde maritime au 17ième et 18ième siècles d’une gamme de points de vue.

Teil 2 unterteilt das Artefaktarchiv in Kapiteln nach Material und Typ. Für die Beschreibung des Schiffsbau xiv

Summary werden Aufzeichnungen über die auf dem Meeresboden verbleibende Schiffsstruktur sowie gefundenes Material verwendet. Die Überreste am Meeresbodens bestehen aus dem Hauptdeck des Schiffes, welches die höchste Stufe zusammenhängender archäologischer Überreste bietet, mit Trümmern des Oberdecks, des Vorderdecks und des Achterdecks.

Die Sammlung von 57 Töpfen aus dem Gelände repräsentiert eine Vielzahl von Ländern und Produktionszentren, die entweder die Segelgeschichte des Schiffes, den Charakter der Besatzung oder die Keramikarten, die im Ursprungshafen erhältlich sind, widerspiegeln. Diese wurden für eine Reihe von Funktionen verwendet – Lagerung von Lebensmitteln, Kochen, Essen und Trinken, Salbentöpfe und Gläser für Medizin. Steinartefakte umfassen Navigations- oder Deckschiefer, Wetzsteine, Ziegel und Tonrohre.

Unter diesem sind vermutlich das Orlopdeck und der Laderaum verblieben, während Teile des Heckbalken und Hecks am östlichen Ende des Fundortes erhalten sind. Hier werden auch Rahmen, Schiffsblanken und Verbindungsteile mit Hilfe von historischen und archäologischen Vergleichsmaterialien beschrieben, die zur Kontextualisierung und zum Verständnis des Stirling Castle beitragen. Kapitel, die Artefakte nach Materialtyp betrachten, enthalten einen Katalog und bieten soweit möglich und praktisch, vergleichende Analysen an.

Die Analyse des gesamte Materialarchiv zeigt breite Trends auf, die ein Verständnis des Schiffes, seiner Besatzung und der archäologischen Aufzeichnungen, die sie hinterlassen haben, vermitteln. Die Interpretation berücksichtigt zwei Schlüsselfragen im Zusammenhang mit der archäologischen Aufzeichnung: • Wie viel von dem ursprünglichen Schiff von 1679, in Fragen der Struktur oder der Ausrüstung, überlebte den Wiederaufbau von 1699, wurde Teil der archäologischen Überreste des Schiffes nach seinem Verlust im Jahre 1703; und • Welche Auswirkungen hatte die Senkung auf die Verteilung von Artefakten und Baumaterialien auf dem Meeresboden und wie sehr hat es die Beziehungen von Menschen und ihrem Besitz und die Ausrüstung, die sie an Bord des Schiffes verwendet haben, verschoben

Die große Ansammlung von organischen Überresten zeigt das Erhaltungspotenzial der Meeresumwelt und umfasst Pflanzenreste, Knochenartefakte, Seile, eine Reihe von Ledergegenständen, darunter Schuhe und Buchbezüge sowie Textilien. Holzobjekte reichen von Behältern, Lebensmittelzubereitungs- und Lebensmittelverzehr, Kleidung und persönlichen Gegenständen bis hin zu Navigations- und Messgeräten, Werkzeugen und Geräten und Waffen. Die organische Ansammlung ist bedeutend aufgrund der Seltenheit, eine so große Anzahl und Vilefalt organischer Artefakte auf einem Gelände zu entdecken. Die Sammlung stellt viele Aspekte des Lebens an Bord des Schiffes dar, von allgemeinen täglichen Aufgaben (z.B. Schrubben des Decks) und Facharbeiten bis hin zur Freizeit. Die Hierarchie der Besatzung wird durch die Materialien dargestellt, aus denen die Artefakte hergestellt wurden, zum Beispiel die grundlegenden, funktionalen und undekorierten Objekte, die von den unteren Besatzungsmitgliedern verwendet wurden, bis hin zu Fragmenten teurer Seidenstoffe, die von Kleidungsstücken der höheren Ränge stammen.

Soweit wie möglich, werden diese Fragen neben der weiteren Diskussion von Themen wie das Kampfschiff, Navigation und Artefaktverteilung behandelt. Schlussfolgerungen bekräftigen, wie die archäologischen Überreste des Stirling Castle einen überzeugenden Einblick in die maritime Welt des 17. und 18. Jahrhunderts aus einer Reihe von Perspektiven bieten.

Glasartefakte bestehen aus Flaschen (in verschiedenen Formen: Zwiebel, quadratisch und zylindrisch), Trinkgläsern, Zeitgläsern und Fensterglas, von denen einige einer wissenschaftlichen Analyse unterzogen wurden, die Details über ihre chemische Zusammensetzung lieferten. Über 230 Metallartefakte umfassen Beispiele aus persönlichem Besitz, Kampfwaffen und Teile der Schiffsstruktur. Eine Reihe von Metallen sind innerhalb der Assemblage vertreten: Eisen, Zinn, Kupfer, Messing und Blei sind am bemerkenswertesten. Sie werden vorgestellt und analysiert nach deren Funktionen, wie Essen und Trinken (Teller, Schüsseln, Geschirr, Besteck, Krüge, Drinkgefässe, Kochutensilien), persönliche Artefakte (Kleidung, Uniformen, Lampen und anderes), Schiffsausrüstung und Armaturen (Navigation und Messung, Speigatt, Befestigungen) und Waffen Ausrüstung (Schwerter, Kleinwaffen, Munition, Kanonen und Kanonenmunition). xv

List of Figures Figure 1.1. Location of the Goodwin Sands and the site of the Stirling Castle.................................................................... 1 Figure 2.1. The Resolution in a gale..................................................................................................................................... 6 Figure 2.2. The launch of the Stirling Castle at Deptford in 1679 as drawn by Willem van de Velde................................. 7 Figure 2.3. The Battle of La Hogue, 23 May 1692, as drawn by Willem van de Velde the younger................................... 8 Figure 2.4. The Great Storm of 26 November 1703........................................................................................................... 12 Figure 3.1. The 1979 team on board Tom Brown’s fishing vessel Shelandra.................................................................... 15 Figure 3.2. Artistic impression by David Perkins of the extent of the Stirling Castle in 1979.......................................... 16 Figure 3.3. The bell of the Stirling Castle being brought on board Shelandra in 1979..................................................... 16 Figure 3.4. The dive team during the 1999 season of work................................................................................................ 20 Figure 3.5. Licensee Bob Peacock inspecting the hull of the Stirling Castle in 1999........................................................ 20 Figure 3.6. The wooden log reel exposed during 1999....................................................................................................... 21 Figure 3.7. Simon Adey-Davies oversees the recovery of a cannon from the Stirling Castle............................................ 22 Figure 3.8. The in situ remains of a box of glass jars......................................................................................................... 22 Figure 4.1. Location map for Stirling Castle and Goodwin Sands bank features and location of metocean data recording sites............................................................................................................................................................. 31 Figure 4.2. Velocity direction and magnitude for neap and spring tidal diamond H (051°16’.3N, 001°27’.6E, 3.4km west of Stirling Castle site)........................................................................................... 32 Figure 4.3. Wave directional roses for Goodwin Sands wave directional waverider buoy and Hastings wavenet buoy...................................................................................................................................................................... 33 Figure 4.4. Time series of significant wave height for Goodwin Sands waverider buoy and Hastings waverider buoy.................................................................................................................................................................... 34 Figure 4.5. Transect location overlaid on multibeam bathymetry hillshade from March 2006.......................................... 34 Figure 4.6. Transect through bedforms 30m to east of wreck............................................................................................ 35 Figure 4.7. Location of sandwave transect (B to B’) overlaid on 2009 bathymetry........................................................... 35 Figure 4.8. Transect of sandwaves 100m west of the wreck site of B to B’....................................................................... 35 Figure 4.9. Bed-level change from April 2005 to August 2006 along transect B............................................................... 36 Figure 4.10. Historical chart contours for the Goodwin Sands........................................................................................... 37 Figure 4.11. Goodwin Sand bank bathymetry from 2009 MCA survey............................................................................. 38 Figure 4.12. Goodwin Sand bank dimensions and angles for 1795 to 2009...................................................................... 39

xvi

List of Figures Figure 4.13. Bed-level change of Goodwin Sands bank from 1997 to 2009...................................................................... 40 Figure 4.14. Singlebeam bathymetry survey of wreck site, 1997....................................................................................... 40 Figure 4.15. Singlebeam bathymetry survey of wreck site, 1998....................................................................................... 40 Figure 4.16. Singlebeam bathymetry survey of wreck site, 1999....................................................................................... 41 Figure 4.17. Multibeam bathymetry survey of wreck site, July 2002................................................................................ 41 Figure 4.18. Multibeam bathymetry survey of wreck site, April 2005............................................................................... 42 Figure 4.19. Multibeam bathymetry survey of wreck site, September 2005...................................................................... 42 Figure 4.20. Multibeam bathymetry survey of wreck site, March 2006............................................................................. 42 Figure 4.21. Multibeam bathymetry survey of wreck site, August 2006............................................................................ 42 Figure 4.22. Multibeam bathymetry survey of wreck site, September 2009...................................................................... 43 Figure 4.23. Sidescan sonar image from 2000 survey........................................................................................................ 43 Figure 4.24. Sidescan sonar image from 2000 survey........................................................................................................ 44 Figure 4.25. Sidescan from 2002 survey............................................................................................................................ 44 Figure 4.26. Sidescan from 2002 survey............................................................................................................................ 44 Figure 4.27. Sidescan sonar survey from 2002................................................................................................................... 44 Figure 4.28. Mosaic of sidescan sonar images from 2005 RASSE survey......................................................................... 45 Figure 4.29. Sidescan sonar image from 2005 survey........................................................................................................ 45 Figure 4.30. Sidescan sonar survey from 2008................................................................................................................... 45 Figure 4.31. Sidescan sonar, for the wreck area only, from 2008....................................................................................... 46 Figure 4.32. Sidescan sonar survey from 2009................................................................................................................... 46 Figure 4.33. Sidescan sonar, for the wreck area only, from 2009....................................................................................... 46 Figure 4.34. Line track data collected with Echoplus system, underlay of sidescan sonar................................................ 47 Figure 4.35. Magnetometer survey from 2002................................................................................................................... 48 Figure 4.36. Magnetometer survey image from 2008......................................................................................................... 48 Figure 4.37. Magnetometer survey from 2009................................................................................................................... 48 Figure 4.38. Mean grain size from 2005 analysis, overlain on April 2005 multibeam bathymetry................................... 49 Figure 4.39. Mean grain size distribution from the 2006 survey for a 200m × 200m area around the wreck................... 49 Figure 4.40. Meso-scale bed-level change for the period of July 2002 to April 2005........................................................ 50 Figure 4.41. Meso-scale bed-level change for the period of April 2005 to August 2006................................................... 50 Figure 4.42. Meso-scale bed-level change for the period of August 2006 to September 2009.......................................... 50

xvii

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Figure 4.43. Bed-level change from July 2002 to April 2005............................................................................................ 51 Figure 4.44. Bed-level change from April 2005 to September 2005.................................................................................. 51 Figure 4.45. Bed-level change from September 2005 to March 2006................................................................................ 52 Figure 4.46. Bed-level change from March 2006 to August 2006...................................................................................... 52 Figure 4.47. Bed-level change from August 2006 to September 2009............................................................................... 52 Figure 4.48. Positions of maximum depths and minimum depths for each multibeam bathymetry survey....................... 53 Figure 5.1. Representative site-plans based on archive material from 1979, 1999/2000 and 2006/9................................ 57 Figure 5.2. Schematic diagram showing the interpreted surviving extent of vessel structure............................................ 60 Figure 5.3. Schematic diagram showing the likely extent of outward collapse of the Stirling Castle’s hull..................... 60 Figure 5.4. Exposed planking collapsed outboard of the port quarter................................................................................ 61 Figure 5.5. Recently exposed deck planking in the midships area of the site in June 1999............................................... 62 Figure 5.6. Organic remains of a pulley block and rope, exposed and raised in 2000....................................................... 63 Figure 6.1. Overall site-plan of the Stirling Castle............................................................................................................. 68 Figure 6.2. Simplified diagram of the component parts of an assembled main-frame....................................................... 68 Figure 6.3. Simplified diagram of the main component parts of a large carvel-built ship................................................. 69 Figure 6.4. Section of exposed, collapsing hull at the stern port-quarter........................................................................... 69 Figure 6.5. Newly exposed deck structure recorded by SeaDive in 1999.......................................................................... 71 Figure 6.6. Outer planking secured with wooden treenails................................................................................................ 73 Figure 6.7. Deck planking secured with wooden treenails................................................................................................. 74 Figure 6.8. Artistic impression of the surviving rudder when still preserved to its full height in 2000............................. 74 Figure 6.9. The rudder assemblage of the Stirling Castle................................................................................................... 74 Figure 6.10. Fixed double block related to the steering mechanism.................................................................................. 75 Figure 6.11. Interpretation of the steering system of the Stirling Castle............................................................................ 75 Figure 7.1. Wooden blocks and associated cordage, in situ in 2000................................................................................... 77 Figure 7.2. Single-sided comb, ID 313............................................................................................................................... 78 Figure 7.3. Toothbrush, ID 314........................................................................................................................................... 78 Figure 7.4. Cup formed from half a gourd, ID 290............................................................................................................. 80 Figure 7.5. Drinking cup made from a gourd, ID 332........................................................................................................ 80 Figure 7.6. Wide-brimmed leather hat with repair to crown and one patch, ID 366.......................................................... 81 Figure 7.7. Wide-brimmed leather hat, ID 613................................................................................................................... 81

xviii

List of Figures Figure 7.8. Inner sole formed from three layers of leather, ID 014.................................................................................... 82 Figure 7.9. Slipper type shoe, mostly complete, ID 423..................................................................................................... 82 Figure 7.10. Fragmented shoe remains with layered construction, ID 514........................................................................ 83 Figure 7.11. Leather book cover, ID 292............................................................................................................................ 84 Figure 7.12. Leather book cover with ribbed spine, ID 293............................................................................................... 84 Figure 7.13. Leather book cover with small brass clasp on front and back, ID 295.......................................................... 85 Figure 7.14. Leather book cover with ribbed spine, ID 296............................................................................................... 85 Figure 7.15. Leather book cover stamped with intertwining scrollwork pattern, ID 297................................................... 85 Figure 7.16. Three fragments of coarse woven fabric, ID 403........................................................................................... 86 Figure 7.17. Silk pouch, purse or pocket, ID 585............................................................................................................... 86 Figure 7.18. Silk damask square, ID 429............................................................................................................................ 86 Figure 7.19. Cylindrical wooden case of barrel construction, ID 291................................................................................ 87 Figure 7.20. Barrel head piece, ID 664............................................................................................................................... 88 Figure 7.21. Warped and worm-eaten fragment of a wooden board marked in squares, ID 279....................................... 89 Figure 7.22. Unpolished wooden bowl, in three pieces, stamped with letters ‘IO’, ID 179............................................... 90 Figure 7.23. Smooth, lathe-turned wooden bowl, ID 251.................................................................................................. 90 Figure 7.24. Wooden dish, ID 525...................................................................................................................................... 91 Figure 7.25. End discs and spindles from various sizes of sand glasses, ID 175............................................................... 91 Figure 7.26. Telescope part, ID 329.................................................................................................................................... 91 Figure 7.27. Scrubbing brush, minus bristles, ID 281........................................................................................................ 92 Figure 7.28. Wooden pulley block, ID 354......................................................................................................................... 93 Figure 7.29. Wooden pulley block, ID 377......................................................................................................................... 94 Figure 7.30. Double wooden block, ID 382........................................................................................................................ 94 Figure 7.31. Woodend pulley block, ID 489....................................................................................................................... 95 Figure 7.32. Sheave with recess for a triangular coak, ID 494........................................................................................... 95 Figure 7.33. Wooden double block, ID 658 ....................................................................................................................... 96 Figure 7.34. Wooden pistol stock with internal construction shown by X-ray, ID 227...................................................... 98 Figure 7.35. Irregular wooden disc or plate, ID 300........................................................................................................... 99 Figure 7.36. Irregular wooden disc or plate, dark in colour, ID 301.................................................................................. 99 Figure 8.1. Freshly raised onion bottles on the deck of the dive boat Shelandra in 1979................................................ 103

xix

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Figure 8.2. Onion bottles.................................................................................................................................................. 105 Figure 8.3. Square-sided bottle, ID 117............................................................................................................................ 107 Figure 8.4. Square-sided bottles........................................................................................................................................ 109 Figure 8.5. Cylindrical bottle, ID 086............................................................................................................................... 110 Figure 8.6. Cylindrical bottles.......................................................................................................................................... 110 Figure 8.7. Drinking glasses..............................................................................................................................................111 Figure 8.8. Timing glasses................................................................................................................................................ 112 Figure 8.9. Timing glasses................................................................................................................................................ 114 Figure 8.10. Timing glasses.............................................................................................................................................. 115 Figure 8.11. Miscellaneous glass...................................................................................................................................... 117 Figure 8.12. Sodium and potassium content of the bottles............................................................................................... 119 Figure 8.13. Magnesium and calcium content of the bottles............................................................................................ 119 Figure 8.14. Sodium and magnesium content of the window glass.................................................................................. 121 Figure 8.15. Lead and calcium content of the window glass............................................................................................ 121 Figure 9.1. Pewter plates................................................................................................................................................... 127 Figure 9.2. Pewter porringer, ID 394................................................................................................................................ 129 Figure 9.3. Pewter porringer, ID 535................................................................................................................................ 129 Figure 9.4. Pewter tankards.............................................................................................................................................. 131 Figure 9.5. Silver jug, ID 513........................................................................................................................................... 132 Figure 9.6. Copper-alloy cooking cauldron, ID 234......................................................................................................... 132 Figure 9.7. Copper-alloy cooking cauldron, ID 270......................................................................................................... 132 Figure 9.8. Sieve or strainer, ID 401................................................................................................................................. 133 Figure 9.9. Buckles, buttons and a brass key.................................................................................................................... 134 Figure 9.10. Snuffer, ID 250............................................................................................................................................. 134 Figure 9.11. Pewter syringe, ID 118................................................................................................................................. 135 Figure 9.12. Nest of brass measurement weights, ID 276................................................................................................ 136 Figure 9.13. Pounce pot, ID 304....................................................................................................................................... 137 Figure 9.14. Inkwell, ID 305............................................................................................................................................. 138 Figure 9.15. Navigational dividers.................................................................................................................................... 139 Figure 9.16. The bell of the Stirling Castle...................................................................................................................... 140

xx

List of Figures Figure 9.17. Sheave, ID 365............................................................................................................................................. 141 Figure 9.18. Sword hilts.................................................................................................................................................... 143 Figure 9.19. Sword hilts.................................................................................................................................................... 144 Figure 9.20. Bronze cannon, originally of Dutch manufacture........................................................................................ 146 Figure 9.21. Shot gauge, ID 204....................................................................................................................................... 146 Figure 9.22. Effect of alloy additions on the hardness of tin............................................................................................ 148 Figure 9.23. Effect of alloy additions on the liquidus temperature of tin......................................................................... 148 Figure 10.1. Two-handled jar, Verwood-type ware, ID 459–461, 464–467 (Vessel 9)..................................................... 155 Figure 10.2. Mug, Bristol or Staffordshire slipware, ID 395 (Vessel 11)......................................................................... 155 Figure 10.3. Ointment pot, English white tinglazed ware, ID 617 (Vessel 14)................................................................ 156 Figure 10.4. Two small ointment pots, English white tinglazed ware, ID 618 (Vessels 14 and 15)................................. 156 Figure 10.5. Bowl, English tinglazed ware, ID 273 (Vessel 23)....................................................................................... 156 Figure 10.6. Rounded mug, Frechen-type stoneware, ID 232 (Vessel 25)....................................................................... 157 Figure 10.7. Narrow-necked jug, Frechen-type stoneware, ID 254 (Vessel 26)............................................................... 157 Figure 10.8. Narrow-necked jug, Frechen-type stoneware, ID 331 (Vessel 27)............................................................... 157 Figure 10.9. Bartmann jug, Frechen stoneware, ID 253 (Vessel 31)................................................................................ 157 Figure 10.10. Bartmann jug, lion rampant medallion, Frechen stoneware, ID 252 (Vessel 32)....................................... 158 Figure 10.11. Mug with applied rosettes, Westerwald-type stoneware, ID 255 (Vessel 36)............................................ 158 Figure 10.12. Handled jar, North French gritty white ware, ID 267 (Vessel 37).............................................................. 159 Figure 10.13. Jar/cooking pot with pronounced flanged rim, French redware, ID 437 (Vessel 38)................................. 159 Figure 10.14. Jar, white earthenware, ID 462 (Vessel 40)................................................................................................ 159 Figure 10.15. Olive jar, Seville-type coarseware, ID 235 (Vessel 42).............................................................................. 160 Figure 10.16. Costrel, Iberian redware, external clear lead glaze on top half, ID 233 (Vessel 47).................................. 160 Figure 10.17. Bowl, Iberian tinglazed ware, ID 274 (Vessel 49)...................................................................................... 161 Figure 10.18. Oil jar, Montelupo coarseware, ID 268 (Vessel 52)................................................................................... 161 Figure 10.19. Jar, Martabani stoneware, ID 450 (Vessel 55)............................................................................................ 162 Figure 10.20. Tea bowl, white Oriental porcelain, moulded petal motif, ID 341 (Vessel 56).......................................... 162 Figure 11.1. Navigational deck slate, ID 178................................................................................................................... 166 Figure 11.2. Sharpening wheel, ID 355............................................................................................................................ 167 Figure 11.3. Clay pipe, ID 106.......................................................................................................................................... 168

xxi

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Figure 11.4. Clay pipe, ID 240.......................................................................................................................................... 168 Figure 11.5. Clay pipe, ID 241.......................................................................................................................................... 168 Figure 11.6. Clay pipe, ID 242.......................................................................................................................................... 168 Figure 11.7. Clay pipe, ID 243.......................................................................................................................................... 168 Figure 11.8. Clay pipe, ID 328.......................................................................................................................................... 169 Figure 11.9. Clay pipe, ID 335.......................................................................................................................................... 169 Figure 11.10. Clay pipe, ID 474........................................................................................................................................ 169 Figure 11.11. Clay pipe, ID 475........................................................................................................................................ 169 Figure 11.12. Clay pipe, ID 476........................................................................................................................................ 169 Figure 12.1. The arrangement of the guns listed in the 1696 ordnance survey for the Stirling Castle............................ 172 Figure 12.2. Basic gun nomenclature applicable to the guns used on the Stirling Castle................................................ 172 Figure 12.3. Artefact distribution between 1979 and 2009, overlain on the combined all-season site-plan.................... 178 Figure 12.4. Historical connections of the Stirling Castle, 1678–1703............................................................................ 182

xxii

List of Tables Table 4.1. Metadata for singlebeam bathymetry surveys................................................................................................... 40 Table 4.2. Metadata for multibeam bathymetry surveys..................................................................................................... 41 Table 4.3. Metadata for sidescan sonar surveys.................................................................................................................. 43 Table 4.4. Metadata for acoustic ground discrimination surveys....................................................................................... 47 Table 4.5. Metadata for magnetometer surveys.................................................................................................................. 47 Table 4.6. Results from 2005 grain size analysis................................................................................................................ 49 Table 4.7. Bed-level change between multibeam bathymetry surveys for the area around the wreck............................... 53 Table 5.1. Sources of Stirling Castle site plans 1979–2009............................................................................................... 58 Table 6.1. Collated measurements of deck beams and ledges............................................................................................ 71 Table 8.1. Summary of timing glass dimensions, ordered by diameter of vessel body.................................................... 115 Table 8.2. Detail of fragments of flat window glass recorded under the ID 503.............................................................. 116 Table 8.3a. Chemical composition of the tableware (major elements)............................................................................. 117 Table 8.3b. Chemical composition of the tableware (minor elements)............................................................................ 117 Table 8.4a. Chemical composition of the bottles (major elements).................................................................................. 118 Table 8.4b. Chemical composition of the bottles (minor elements)................................................................................. 118 Table 8.5a. Chemical composition of the flat glass (major elements).............................................................................. 120 Table 8.5b. Chemical composition of the flat glass (minor elements).............................................................................. 120 Table 8.6. Chemical composition of the modelled alkali glass/flint glass........................................................................ 122 Table 9.1. Composition of copper-alloy map dividers (balance = copper)....................................................................... 147 Table 9.2. Chemical composition of the Stirling Castle pewter samples (balance = tin)................................................. 149 Table 9.3. Average composition of sadware..................................................................................................................... 150 Table 10.1. List of the pottery vessels associated with the ship....................................................................................... 153 Table 10.2. Total numbers of vessels by vessel type........................................................................................................ 162 Table 12.1. Summary of gun concretion dimensions recorded in situ on the seabed....................................................... 173 Table 12.2. Navigational artefacts from the Stirling Castle.............................................................................................. 176 Table 12.3. Navigational artefacts from contemporary shipwrecks................................................................................. 176

xxiii

1 Introduction Julian Whitewright In 1979 a group of sports divers from British SubAqua Club 106, working with a local fisherman out of Ramsgate, began the underwater investigation of net fastening locations on the Goodwin Sands, off the coast of Kent (Figure 1.1). The second site that the team identified and investigated produced a remarkable discovery: the apparently almost complete remains of a two-decked wooden warship. The ship had seemingly appeared from the Goodwin Sands, where it had been wrecked and was now sitting upright on the seabed, with its bow facing to the west. A large sandbank obscured much of the starboard side, but the port side was clear of sand, the bow of the vessel was unobscured and the vessel’s rudder was still in place at its stern. Artefacts were visible scattered among the wreckage of the uppermost surviving deck and some of the stairways and hatches could be clearly seen, and entered. Only the masts and rigging were missing. Wreck II, as it was simply labelled at the time, was the type of site that most wreck divers and maritime archaeologists can only dream of finding and working on, but few ever do.

At the time of its discovery the Stirling Castle was dubbed a ‘Mary Rose’ for the seventeenth/eighteenth century because of the high levels of preservation of the ship itself and the artefacts still surviving on the site. With hindsight, such comparisons were certainly not unfounded, and, if anything, the Stirling Castle had a greater percentage of coherent surviving hull structure and associated artefacts. The ship is also highly significant from an historical perspective because it was launched in 1679 as part of the Thirty Ships building programme overseen by Samuel Pepys and Charles II. The vessel was rebuilt in 1699 and formed part of the Royal Navy fleet in the early years of the War of Spanish Succession. As a result, the operational life of the Stirling Castle spanned a period in the development of the Royal Navy that was critical not just from the perspective of ship technology but also because it was at this time that Britain became arguably the pre-eminent naval power in northern Europe. Added to this was the loss of the ship and crew in one of the most powerful and destructive storms to have been experienced and recorded in English history.

The finding of Wreck II came at a time when the profile and awareness of maritime archaeology within the UK was increasing with ongoing work on the Tudor warship the Mary Rose in the eastern Solent. The local archaeology unit in Thanet, Kent, was contacted, along with the then archaeological unit at the National Maritime Museum. Early recoveries from the site of Wreck II included the vessel’s bell, dated 1701, two large copper kettles and numerous smaller artefacts. The date of the bell and a set of recurring initials on a series of pewter plates allowed the identification of the wreck to be confirmed as the Stirling Castle, a third-rate 70-gun ship that was lost on the Goodwin Sands in the Great Storm of 1703 along with two other Royal Navy third-rates, the Northumberland and the Restoration, and the fourth-rate Mary. Subsequent survey work by the same team located the remains of the Northumberland and another vessel thought to be either the Restoration or the Mary. Further investigation of the site of Wreck II/Stirling Castle was halted in 1980, when the dive team returned to the wreck only to find it reburied beneath the sandbank from which it had first emerged. The focus of work turned to the other Great Storm wrecks, with work on the Stirling Castle sporadic and dependent on the level of the seabed and the resultant exposure of the ship. Then, in the late 1990s, the vessel became re-exposed and work on the site was resumed, with the objective of recording as much of the exposed structural remains as possible, rather than the mass recovery of artefacts that characterised the 1979 season.

Figure 1.1. Location of the Goodwin Sands and the site of the Stirling Castle (base map created using ArcGIS).

1

Whitewright Despite its excellent preservation and obvious wealth of material culture, the site, like most maritime archaeological sites in British waters, did not benefit from centrally allocated resources to facilitate archaeological investigation and dissemination. The protection afforded the site through designation in 1980 under the 1973 Protection of Wrecks Act (PWA) meant the involvement of the government-funded Archaeological Diving Unit (ADU) from its formation in 1986, but the ADU were able to spend only a few days a year at the site, every few years, and carried a remit only to monitor the overall condition of the site itself, rather than to undertake proactive investigation. Meanwhile, although many of the recovered artefacts were on public display in Ramsgate Maritime Museum, the dissemination of the archaeological work on the Stirling Castle lacked a cohesive strategy and overall aim to take it towards a more publicly accessible state. Although a fundamental criticism of the wider archaeological process, the nature of the situation with regard to the Stirling Castle was by no means unique within British maritime archaeology in the 1980s, 1990s and 2000s. By the mid-2000s - despite the ongoing investment of time, effort and financial resources in the monitoring of the in situ remains by Bob Peacock, the site licensee – potentially one of the best-preserved maritime archaeological sites in UK waters was apparently in a state of rapid decline, with minimal public dissemination of the extensive archaeological and historical archive relating to the site having occurred.

enable public access to an archive of material that had already been identified as being of local, regional, national and international importance. A wide range of public bodies and organisations (from local to national) and several private individuals were consulted during that process. Building upon that, Stage Two (HWTMA, 2012) involved a detailed assessment of all archive material and the creation of an updated archive database. Critically, the results of that stage of the project demonstrated the wide variety and dispersal of the Stirling Castle archive, which is composed of the documentary (paper and graphic records, and digital archive including survey data) and the material archives (artefact assemblage and samples) of the site. The archive itself, in its broadest sense, now rests in a range of locations with a number of different owners, including the Isle of Thanet Archaeological Society (IOTAS), Ramsgate Maritime Museum, Hastings Shipwreck Museum, the National Maritime Museum in Greenwich, the National Record of the Historic Environment (NRHE) in Swindon, Wessex Archaeology (WA) and Historic England (HE), as well as in private ownership. At that juncture, the wider project process was revised slightly, with stages Three and Four being combined into a single Stage Three to address the analysis and publication of the archaeological archive. Stage Three of the project has resulted in this monograph, which aims to present the breadth and wealth of the archaeological archive of the site. The structure, scope and inevitable limitations of such a process are returned to below, but it is clear that such a publication is ostensibly an academic one. In that sense, it is clearly targeted towards the maritime archaeological and historical community in the first instance and then more widely to scholars, students and enthusiasts of the period. As such, there is more to be done to develop the wider educational potential of the site and to further increase public access to the material from the site.

It was against this background that a project to address the archaeological archive of the Stirling Castle was developed by Historic England (formerly English Heritage) and the Maritime Archaeology Trust (formerly the Hampshire and Wight Trust for Maritime Archaeology). That proposal built upon a desk-based assessment of the site conducted in 2003 by Wessex Archaeology in their capacity as the diving contractor for the PWA and a site management plan drawn up by Historic England. These reports highlighted the critical need for an evaluation of the recovered artefacts and a formal process to audit the archaeological archive, in all its forms, with the ultimate aim of bringing it to publication. The initial project design submitted to English Heritage by the HWTMA in January 2008 was developed by Julie Satchell and Douglas McElvogue and proposed four stages of work: • • • •

Some explanation is needed regarding the scope of this monograph and what is included and omitted here. Likewise, an overview of the subsequent structure and general aims of each element of the monograph will be useful. Perhaps most importantly, it is necessary to reaffirm that the Stirling Castle is very much a live archaeological site. Work by the licensee and monitoring by the PWA contractor are therefore ongoing, with the potential to produce further material, artefactual or otherwise, in the future. It is clearly impossible to produce a volume of the type presented here, through the methodology just described, if additional material must be accounted; simply for practical reasons a stopping point has to be established somewhere. Therefore, this volume, and the project that it forms part of, has chosen to cover the period from 1979 to 2009. This embraces the initial work on the site, when the majority of the artefacts were recovered, the exposure of the site and related work in the late 1990s and the work involved in beginning to empirically monitor the site in the 2000s. The end date is a compromise between the start of the present project, the last significant piece of diver survey on the site by Wessex Archaeology for which a report was

Stage 1 – Archaeological archive audit and appraisal Stage 2 – Archival assessment and analysis Stage 3 – Targeted archaeological recording Stage 4 – Publication and dissemination

The intention of the initial project design was that successful completion of each stage would facilitate the subsequent undertaking of the next phase. Stage One (HWTMA, 2009a) involved locating, visiting and appraising site archive material, including site records, surveys, images, videos and artefacts. This work provided an archive audit and appraisal as an initial step to help 2

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 available and a geophysical survey of the site undertaken by the UK Hydrographic Office (UKHO) that provides the most recent coverage of the site morphology within that time period. There is, therefore, a clear responsibility on the part of Historic England, in conjunction with the site licensee, to ensure that work conducted from 2010 onwards, in whatever form, is eventually afforded the same level of dissemination as this project has given to the 1979-2009 material. This includes further geophysical survey work commissioned by Historic England and undertaken in July 2015, which, it is hoped, will be able to build upon the wider geophysical analysis presented here within chapter four.

archive outlined in chapter five. Subsequent chapters then move through organic remains, glass, metal, pottery and stone materials. In each of these - chapters seven to eleven respectively - a catalogue of the material is set out and comparative analysis offered where possible, practical and appropriate. Following this, chapter twelve attempts to look across the entire archive of material in order to draw out wider trends and concerns, and to present an understanding of the ship, its crew and the archaeological record they have left behind that cuts across the material classes imposed in the preceding chapters. Attention is also given to how we can view the historical archaeology of vessels of this type in a general sense and in a specific sense in relation to the Stirling Castle.

The structure of this monograph is relatively straightforward and traditional. It falls very much into two parts, the content of which is dictated to by the story of the ship and its crew, their loss, their archaeological discovery and the subsequent work carried out on the site. This account therefore begins in chapter two with the history of the Stirling Castle as a warship, covering its building, service and loss and taking both a broad contextual view and a more detailed view of the vessel and crew. In effect, this takes us from Deptford Dockyard in 1679 to the Goodwin Sands on 26–28 November 1703. Attention then shifts in chapter three to the provision of a detailed account of the archaeological investigation of the site from 1979 to 2009. This is based largely on the documents, in all their forms, included in the site archive as well as interviews with those who have worked on the site. On reading this chapter it should become readily apparent that the archaeological archive of material from the site cannot be understood without such a developed appreciation of the people and processes behind its creation across a 30year period. The history of the archaeology is critical to our interpretation of the archaeology itself.

Throughout, two key questions can be identified and borne in mind, if not necessarily answered. The first of these revolves around the extent to which the original 1679 ship, in structural or equipment terms, survived its 1699 rebuild to become part of the archaeological remains of the vessel following its loss in 1703. The second perhaps echoes this, but on a smaller scale, and requires us to continually bear in mind the contribution of the drawn-out, but certainly cataclysmic, wrecking event to the subsequent distribution of artefacts and structural materials on the seabed. It is clear that the vessel underwent a less damaging process than others within its class (Northumberland and Restoration) that were also lost on the Goodwin Sands in 1703, but we must still consider the extent to which that process reorganised the relationships of the people, their possessions and the equipment they used on board the vessel. On the face of it, such consideration assumes a perfectly preserved, recovered and archived archaeological record upon which to base such an interpretation. The very existence of this project, to address the archive of the Stirling Castle so long after the initial period of work on the site perhaps indicates some of the challenges to such a process.

Having reviewed the archaeological history of the site, chapter four then addresses the environmental context of the Stirling Castle and the Goodwin Sands, drawing upon the extensive study of the sands and the substantial archive of geophysical surveys that have been conducted on the site, including the 2009 UKHO survey. This is the first time that such a dedicated, comprehensive, geophysically based overview of the site, its location and its morphology has been undertaken and published. With the foundations in place, attention turns more directly to the archaeological material, with an overview in chapter five of the extent of material that has been recorded in situ, material that potentially remains in situ and material that is assumed to have been lost. In a sense this can be categorised as the seabed archive and the analysis of that element of the site completes part one of the volume.

Finally, from the perspective of introducing this volume and the three decades of archaeological work that it represents, a final caveat must be observed: namely, that only material that was identified, included and recorded within Stages One and Two of the project has been included within this volume. It is likely that material has been omitted, not through deliberate selection on the part of staff involved in this project but because such material is not currently part of the wider archaeological process. In this regard such omission simply highlights the wider failings of maritime archaeology as a broad discipline within the UK since 1979. This project makes no excuse for that, but it is hoped that bringing a corpus of material such as the 1979-2009 Stirling Castle archive to publication will highlight the potential wealth of material, information and enjoyment that may be found within the archives of material from other English maritime archaeological sites. Many such sites still await a similar process in order to begin to reach their undoubted potential through a full process of public dissemination.

Part two has the simple aim of presenting the artefact archive in the fullest possible manner. As such, a series of chapters addresses the recovered artefacts by material and type. Chapter six discusses the understanding of ship construction and remains that can be gained from the recorded and recovered material as well as the seabed 3

2 Vessel History: Building, Service and Loss Julian Whitewright The present account largely continues this approach, at least initially, in treating the Stirling Castle as a continuous entity, even if the material nature of the ship that sank in 1703 is probably not the same as that of the one launched in 1679. The archaeological evidence of the vessel’s bell, bearing the date 1701, when the ship was first commissioned after its rebuild, provides a useful example of this. At the time, it was clearly the commissioned status of the vessel, rather than its date of launch, that was used to bestow additional kudos to such artefacts. Despite this, the rebuild of the Stirling Castle in 1699 does mark a convenient break between the two halves of the vessel’s service history from the perspective of the re-telling set out below, although this is as much because it corresponds with a cessation in warfare between 1697 and 1701, allowing for such rebuilding programmes to occur.

The history of the service of the Stirling Castle between its launch in 1679 and its loss in 1703 is closely intertwined with the wider political events of the period, including the various conflicts involving the naval fleets of England, the Netherlands and France. The following history of the Stirling Castle aims to contextualise the use of the vessel, its crew and their material culture within naval shipbuilding and operations in the late seventeenth and early eighteenth centuries. In that sense, it is very much a historical narrative of the period seen through the life of the Stirling Castle, rather than one that is complete and exhaustive in its account of every ship and naval engagement or the related political cause and effect. For the latter the reader is referred to the work of Roger (2004) for an overview of naval affairs at that time. An understanding from a historical perspective of the construction of vessels such as the Stirling Castle within the Thirty Ships is well covered elsewhere. In particular, the work of Endsor (2009) provides a detailed account, based mainly on the surviving historical records, of the Lenox, the first of the 70-gun third-rate ships, built by John Shish at Deptford. Coverage of the archaeological remains of one of the Stirling Castle’s sister ships, the Northumberland, also lost in the Great Storm of 1703, is provided by Pascoe and Peacock (2015) and Pascoe et al. (2015). Meanwhile, the basic skeleton of the life of the ship in operational terms is provided in the exhaustive overall work of Winfield (2009: 66, 76–77), from which the following account is developed.

Finally, and quite naturally given the archaeological nature of this monograph, attention will be turned to the crew of the vessel. Before the surviving material culture of the vessel and crew is addressed in chapters six to eleven, an examination is made of the historical documentation, notably the surviving paybooks, from the re-commissioning of the vessel in March 1701 to its loss in November 1703. 2.1  Background (1660–1674) From a naval perspective, the accession of Charles II in 1660 created a complicated and changeable series of political alliances in western Europe that, in a general sense, set England, France and Portugal in opposition to Spain and the Netherlands, with particular English concerns over increasing Dutch monopolisation of maritime trade (see Roger, 2004: 65–67). A war between England and the Netherlands (the first Anglo-Dutch war, 1652–54) had already been fought for similar reasons, during which the concept of deploying warships in a ‘line of battle’ was increasingly fully conceived (Gardiner, 2004: 147; Roger, 2004: 217; Tracey, 2004: 182). English confidence following success at sea in that conflict (at, for example, Portland, the Gabbard and Scheveningen), alongside skirmishes in west Africa and the north Atlantic, resulted in the outbreak of the second Anglo-Dutch war (1665–67). The outcome of that conflict (for an account see Roger, 2004: 68–79) was altogether different, including a Dutch victory in the ‘Four Days Battle’ at Galloper Sand (1–4 June 1666) followed by an English one in the ‘St James’s Day Battle’ (4–5 August 1666). The dire financial situation in England meant that peace negotiations were

Clarification can also be given to the two distinct phases in the life of the Stirling Castle either side of the ship’s rebuilding in 1699. In many accounts (for example those found in Winfield, 2009), a rebuild marks a completely new ship, signalled by a restarting of the historical account of the vessel. This is likely to have been physically true for some ships (see discussion below) and in some cases vessels carried a different name on either side of such a rebuild, such as the Duke/Prince George. However, throughout the archaeological investigation of the Stirling Castle and indeed also the Restoration and Northumberland on the Goodwin Sands, they have been clearly defined as, and continue to be defined as, vessels of the Thirty Ships programme (e.g. Dunkley, 2008: 4; WA, 2003a: 22) at least partially to establish their overriding historical significance. It is their status as part of the Restoration Navy and Pepys’ shipbuilding programme in particular that is emphasised, rather than the later career of the ships and their crews and the archaeological traces of such later phases that survive today. 5

Whitewright opened at the end of 1666. But the defining historical event in the war was the Dutch raid on the Medway over four days of June 1667, targeting the English fleet that was laid up there because of lack of money to send it to sea (Roger, 2004: 101). This daring attack resulted in one first-rate, three second-rates and three third-rates being captured or destroyed, including the flagship Royal Charles (Roger, 2004: 77). The war was concluded with the Treaty of Breda on 31 July 1667 with overall terms that favoured the Dutch. The ever-shifting sands of seventeenth-century international politics meant that by 1668 England and the Netherlands were in alliance, along with Sweden, against France. At the same time, Charles II had concluded a secret treaty (the Treaty of Dover) in June 1670 with Louis XIV of France in which the latter paid Charles subsidies in return for English support for France in a war with the Netherlands. This duly happened following the French invasion of the Netherlands in May 1672, resulting, from an English perspective, in the third Anglo-Dutch war (1672–74) (see Roger, 2004: 80–85). The naval actions of this conflict saw a combined English and French fleet take to sea against their Dutch counterparts, the first meeting seeing the Dutch victory at the Battle of Solebay in May 1672 (see Roger, 2004: 81). Two further, inconclusive, battles took place at Schooneveld in May and June 1673 before the final major naval engagement of the war at the Battle of the Texel on 11 August 1673, when the Dutch Admiral de Ruyter again held off the large allied fleet (Roger, 2004: 83–85). The changing political climate within England resulted in the signing of a peace treaty in February 1674.

Figure 2.1. The Resolution in a gale. The third-rate 70-gun ship was one of the first of the class and was launched at Harwich in 1667. Shown close-hauled on the port tack in a gale, painted by Willem van de Velde the younger c.1678 (© National Maritime Museum, Greenwich, London, BHC3582).

Roger, 2004: 218–20). The group of third-rate vessels, in their construction, formed what would now be considered an individual class, with efforts made to standardise the dimensions of their rigging, while their hull-form was still being used as the basis for two-decker ships of the line until 1755 (Roger, 2004: 218). They were themselves based on the existing class of third-rate vessels, such as the famously depicted Resolution (Figure 2.1).

2.1.1 The Thirty Ships (1674–1688) Reforms of the administration of the Admiralty in 1673 meant that the Royal Navy was no longer under the direct control of the King and instead resided somewhere between the King and Parliament (Roger, 2004: 107). This change in emphasis, along with the role of Samuel Pepys as Secretary of the Admiralty and Member of Parliament, meant that Pepys was able to persuade parliament to vote through £600,000 for a peacetime building programme of thirty ships of the line (Roger, 2004: 108). This programme took place between 1677 and 1685 and has become known as the ‘Thirty Ships’ building programme. It was overseen by Pepys in his role as Secretary of the Navy with the aim of reinvigorating the Royal Navy after the losses of the mid-seventeenth century and attempting to match the rapid growth in the Dutch fleet and especially the French Navy under Louis XIV (Lavery, 2004: 15; Roger, 2004: 108).

Winfield (2009: 66) provides a concise overview of the construction of the Stirling Castle, which formed part of the second batch of vessels ordered in 1678. The ship was built at Deptford Royal Dockyard by master shipwright John Shish and was the third third-rate built by Shish as part of the programme, after the Lenox, launched in April 1678 (see Endsor, 2009), and the Hampton Court in October 1678. The vessel was ordered on 9 July 1678 and launched on 29 July 1679 (Figure 2.2), a building time of around 12 months. The dimensions of the ship, as recorded at the time of building were 151 feet 2 inches (46m) on the lower gun deck, 120 feet 11 inches (36.8m) on the keel for the purpose of measuring the vessel’s tonnage, and 40 feet 7 inches (12.4m) in breadth. The depth of the hold was 17 feet 3 inches (5.25m) and the listed draught was 18 feet (5.5m). The first group of vessels was constructed with 12 gun-ports on the upper deck and 13 on the lower, while the second group, including the Stirling Castle, had 13 ports on both gun decks. Like many of the other third-rate ships

The Thirty Ships entailed the construction of a single firstrate ship, nine second-rate and twenty third-rate ships, the building of which to a standardised classification and tonnage, including the ‘rating’ of the armament, helped to put in place the foundations of Royal Navy building practices for most of the rest of the age of the sailing warship (Fenwick and Gale, 1998: 95; Lavery, 2004: 16; 6

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 2.2. The launch of the Stirling Castle at Deptford in 1679, as drawn by Willem van de Velde. The flags, from forward, are a jack on the stem, the admiralty flag, a Union flag and the naval ensign. The bare flagpole in the centre would normally carry the royal standard at a launch (© National Maritime Museum, Greenwich, London, PAH 3920/PY3920).

2.2 The Active Service of the Stirling Castle

built as part of the Thirty Ships programme, the Stirling Castle did not see immediate active service owing to a shortage of money, political manoeuvring, and the peace-time situation (Roger, 2004: 109–10). It, along with many of the other Thirty Ships, was mothballed termed ‘in ordinary’ - at anchor on the River Medway throughout the 1680s. The scale of this reserve fleet is illustrated by the record of mooring positions from September–October 1683 (see Endsor, 2009: 143), when the Stirling Castle occupied the second mooring position downstream from Rochester Bridge, along with 41 other ships of the line. These vessels included 15 of the original 20 third-rate ships built in the Thirty Ships programme and a detailed and lively historical account of the period ‘in ordinary’ of one of them, the Lenox, is provided by Endsor (2009: 125–46).

2.2.1 1690–1699 Some 11 years after its launch, the Stirling Castle was finally commissioned in 1690, when Captain Anthony Hastings took command. Hastings, who had been promoted to captain in October 1688, had among his previous commands another third-rate 70-gun of the Thirty Ships programme, the Essex. The date of the Stirling Castle’s commissioning is unclear, but it must have taken place early in 1690, in order for the ship to be part of AngloDutch fleet in that year. French policy during 1690 centred on a decisive fleet victory and subsequent control of the Channel (Willis, 2008: 202). The resulting Battle of Beachy Head, which took place on 30 June 1690 between the Anglo-Dutch fleet and the large French fleet, occasioned a decisive French victory and the loss of over half the Dutch ships involved. During the battle the Stirling Castle was part of the centre (red) squadron under the command of Admiral Lord Torrington (Winfield, 2009: xl-xli). From the perspective of the history of the Stirling Castle the action is notable because the vessel served alongside other third-rates built at Deptford dockyard, the Lenox and the Hampton Court. Meanwhile another Thirty Ships thirdrate, the Anne, was lost during the battle and burned on the beach at Pett Level (subsequently being designated under the PWA in 1974).

The accession to the throne of William III (William of Orange) and Mary II late in 1688 brought England into the anti-French League of Augsburg and the Nine Years War between 1688 and 1697. One direct result of this was the formation of a combined Anglo-Dutch fleet to oppose the French and to ensure both the security of William’s claim to the English throne against James II and the territorial integrity of the Netherlands in the face of French aggression. It is against this background that vessels such as the Stirling Castle were finally taken out ‘of ordinary’ and sent to sea. 7

Whitewright

Figure 2.3. The Battle of La Hogue, 23 May 1692, as drawn by Willem van de Velde the younger some years afterwards. Although less dramatic than some contemporary paintings, the image conveys a good impression of ships such as the Stirling Castle at sea through the third-rate vessel illustrated in the centre of the scene, as well as the overall visual impact of a contemporary fleet at sea (© National Maritime Museum, Greenwich, London, PAG6273/ PX 6273).

A constant theme in the service history of the Stirling Castle during this period is the almost yearly turnover of captains. In 1691 Hastings moved to command the secondrate 90-gun Sandwich (another of the Thirty Ships), in what can presumably be considered a promotion. He was succeeded by Captain Benjamin Walters, who had formerly commanded the fourth-rate 48-gun Swallow, on which he had been present at the Battle of Beachy Head. Unlike in 1690, French objectives in 1692 were concerned with gaining control of the Channel in order to support the landing of troops in Dorset with the aim of re-establishing James II. The decisive naval engagement came at the Battle of Barfleur (19–22 May 1692), when the Anglo-Dutch fleet defeated a much smaller French fleet. Subsequent engagements at Cherbourg and La Hogue (Figure 2.3) between 22 and 24 May saw the burning of three large French ships, including the flagship Soleil Royal, and the destruction of 12 other ships of the line (Roger, 2004: 148–49). The Stirling Castle was part of the rear division of the rear (blue) squadron under Admiral Ashby at Barfleur and part of the fleet commanded by Vice Admiral George Rooke in the action at La Hogue (Winfield, 2009: xlii).

fourth-rate 42-gun Ruby, and it was stationed with Admiral Berkley’s main Channel fleet, which attacked various French coastal towns including St Malo, Dunkirk and Calais, during the summer of 1695. More widely, French naval policy had altered slightly as a result of the failure of its battle fleet to secure a decisive victory, with a growing, but not total, emphasis placed on naval war against commerce (Black, 2002: 82; Roger, 2004: 157–60 and 172): in other words, the merchant vessels of England and the Netherlands. A direct result of this in England was a parliamentary provision of 1694 that allocated 43 ships as ‘cruisers and convoys’ to protect incoming merchant ships (Roger, 2004: 159–60). These vessels were distributed across five squadrons: off the East Coast, in the Downs off Kent, off the Channel Islands, off the Scillies and in the Western Approaches off the southern coast of Ireland. It was to the last of these squadrons that the Stirling Castle was assigned, still under the command of Deane, in 1696. The command of the Stirling Castle, along with its area of operation changed again in 1697, when Deane was succeeded by Jedediah Barker, formerly captain of the third-rate 52-gun Lion, and the ship was assigned to the Dunkirk squadron. The purpose of that squadron was to blockade the port of Dunkirk, which itself was the base of a number of highly successful privateers that had caused significant disruption to North Sea trade in the 1690s (Roger, 2004: 158).

Captain Walters was superannuated on 18 April 1693 and was succeeded as captain of the Stirling Castle by Humphrey Sanders, previously the captain of the first-rate 100-gun Royal Sovereign as the flagship of Vice-Admiral Ralph Delavall (Winfield, 2009: 12), who took command in May 1693, when the vessel was assigned to the Channel Fleet. The exact activities of the Stirling Castle in 1693 and 1694 remain unclear, but it is likely to have remained part of the Channel Fleet that facilitated the failed landings at Brest on 8 June 1694. In 1695 command of the Stirling Castle was taken over by Robert Deane, previously of the

2.2.2 Rebuild and Renewal The end of the Nine Years War, brought about largely by financial exhaustion, was signalled by the peace of Ryswick in September 1697 (Roger, 2004: 163 and 199). It also allowed the instigation of a substantial rebuilding 8

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 programme within the Royal Navy that involved many of the ships built during the Thirty Ships building programme. Of the 20 third–rates originally approved in 1677 and subsequently built, 15 survived in 1697 and 14 of these were rebuilt between 1697 and 1704 (Winfield, 2009: 76). As was the practice at the time, the ‘rebuild’ was exactly as its name suggests, with the old ship being taken to pieces as required to allow any rotten timbers to be removed, before being repaired and re-assembled (for a full account see Lavery 1980a; 1980b; Lavery, 2004: 17; Roger, 2004: 220). The rebuilt dimensions of the surviving third-rate Thirty Ships are generally the same as their original ones (Winfield, 2009: 76), allowing Lavery (1980a: 11) to conclude that the rebuilding programme between 1697 and 1702 did not involve a total rebuild of the entire vessel. Roger notes (2004: 220), however, that it was not unusual in some circumstances for the rebuilt ship to exceed the dimensions of the original vessel. Taking all of the above into account, the variety of possibilities within the phrase ‘rebuilt’ is well summed up by Lavery (2004: 124) who notes that this was ‘an elusive concept that that encompassed everything from a major reconstruction without substantial change of design to a purely administrative fiction that allowed the construction of a totally new ship’.

death of Charles II of Spain in November 1700 (see Harding, 2002: 164). Spain at this time had a global empire, with territories in the Americas, the Philippines, Italy and the western Mediterranean. Charles’ designated heir was Philip of Anjou, grandson of Louis XIV, while the son of Leopold I, the Holy Roman Emperor, was put forward as a rival. The latter was supported by England, the Netherlands, Portugal and the Duchy of Savoy, who all feared the consequences of a unification of France and Spain under a single dynasty but had different motives for entering the war. In particular, Roger (2004: 164) has highlighted the purely mercantile, maritime reasons behind English and Dutch involvement: access to the markets of the Spanish Empire, rather than to Spain itself. English strategy was based around maritime power and the blockade of Atlantic routes, in concert with control of the western Mediterranean to support Savoy (Roger, 2004: 164–65). Such a policy obviously required ships, and many vessels that had been rebuilt since 1697 were brought back into commission during 1701 and 1702 (see Winfield, 2009: 76–79 for details). Against this background the Stirling Castle was commissioned in March 1701 under Captain Henry Martin. The vessel’s pay book (TNA,ADM 33/210) records that rigging of the ship began at Chatham on 4 March 1701, with provisions and ordnance following afterwards. The Stirling Castle was assigned to the main fleet of Admiral Rooke and the vessel’s initial voyage saw the ship reach Spithead in late April 1701. Here Captain Martin left the ship, possibly as a result of being promoted to Rear Admiral, and assumed command of the third-rate 64-gun Defiance. Martin was replaced by Captain John Johnson, originally promoted to captain in May 1690 and with experience of commanding a series of third- and fourthrate vessels, including the fourth-rate 46-gun Kingfisher (1690–93) and the third-rate 70-gun Suffolk (1696–97), another of the Thirty Ships. The pattern of Captain Johnson’s assignment to various vessels between 1690 and 1699 largely matches the trend seen in the first period of commission of the Stirling Castle, with commanding officers moving between ships on a regular basis every couple of years or even every year. The Stirling Castle had five different captains during that period, while Johnson had commanded five different ships. The reasons for this regular shift in command from ship to ship are not clear and do not seem to be considered a matter for comment in the main sources, although Roger notes (2004: 206) the detrimental effects, recorded by the admirals of the day, of moving able seamen between ships on a regular basis, a practice known as ‘turning over’.

The Stirling Castle was rebuilt at Chatham Royal Dockyard under master shipwright Daniel Furzer and relaunched in 1699 along with two other third-rates of the Thirty Ships, the Eagle and the Expedition. The rebuilt vessel was 3 feet 4 inches (1m) longer on the keel, as measured for tonnage, and 1 inch (25mm) narrower. This in turn resulted in an increased tonnage from 105929/94 to 108764/94 using the builder’s measure. Overall then, the dimensions of the rebuilt Stirling Castle were almost the same as those of the original vessel, a pattern repeated in the other Thirty Ships third-rates rebuilt at this time (for figures see Winfield, 2009: 76–79). As a class of vessel allowing for the various dockyards involved, the rebuilding process itself and the reporting of measurements - the 1677 third-rates still had a relatively close conformity. The only real difference, across the class from the initial build to the re-build, is a consistent lengthening of the keel by a few feet as measured for tonnage, while maintaining the original overall length. Lavery notes (1980a: 9) that the breadth was the most common change during a rebuild, the very measurement that seems to remain most similar on the Stirling Castle as well as on other vessels of the same class. This might suggest that the rebuilding of these vessels did not result in the construction of a significantly different group of vessels afterwards, at least in terms of their size and form.

Such a practice is evident in the records of the Stirling Castle’s crew in May 1701. The vessel’s initial voyage was to Spithead, where paybook records indicate that eight men were discharged on the 24 April because they were infirm. The following month, on the 23 May, Captain Martin left the vessel, accompanied by 42 of the crew. For the most part these men comprised able seaman, but they also included the third Lieutenant and ten men who are listed as the ‘Captain’s Servant’. A further 16 men departed to

2.2.3 1699–1703 The rebuilt Stirling Castle was re-launched in 1699 (Winfield, 2009: 77) and, as at its initial launch, was then placed ‘in ordinary’ until it was required. Although the Nine Years War had ended in 1697, further conflict in the form of the War of the Spanish Succession soon arose following the 9

Whitewright 2.3 The Crew of the Stirling Castle: March 1701 to November 1703

join Captain Martin over the next few days. For his part, Captain Johnson arrived on the 24 May, and was joined by 39 new crew over the following days. Whatever its effect on crew structure and stability, the practice of ‘turning over’ nevertheless seems to have been commonplace with the commanders of many of the same vessels, perhaps reflecting the seafaring practice of the main battle fleet returning to port for the winter and the subsequent redistribution of captains to ships the following spring. Johnson did, however, remain in command of the Stirling Castle, apart from a temporary replacement as captain by Josias Crow in July 1703.

The historical archive associated with the Stirling Castle provides a wealth of information to supplement the material remains of the vessel and crew. Perhaps most notable from the perspective of this volume are the paybooks for the ship for the period from its re-commission in March 1701 to its loss in November 1703 (TNA, ADM 33/210 and 33/230). These detail the comings and goings of the crew across this period and allow some of the individual artefacts recovered from the site to be reconciled with individual members of the crew, usually through their initials. A complete list of all those who served on the Stirling Castle between March 1701 and November 1703, based on the paybook records, makes up Appendix 14.1.

In 1702 Rooke was tasked with the capture of Cadiz as a means to strangle Spanish Atlantic trade, but, owing to his ambivalence towards the original plan, this operation achieved little (Roger, 2004: 165–66, cf. Harding, 2002: 169). The Stirling Castle was part of Rooke’s fleet off Cadiz in August 1702, but was not part of the detachment that took part in the victory over a combined French and Spanish fleet at Vigo on 12 October (Winfield, 2009: 77). In December 1702 the Stirling Castle became the flagship of John Leake upon his promotion to Rear-Admiral of the Blue (Martin-Leake, 1920: 105), although Captain Johnson would have remained in day-to-day command of the ship. Leake’s time as Flag Officer on the Stirling Castle was short, as he shifted his flag to the Resolution and then to the Royal William (Martin-Leake, 1920: 107); upon promotion to Vice-Admiral in March 1703 he moved his flag again to the second-rate 96-gun Prince George. That vessel was originally launched as the second-rate 90-gun Duke at Woolwich in 1682 as part of the Thirty Ships programme, having been built by Thomas Shish, younger brother of John Shish. Leake himself went on to become Admiral of the Fleet in 1708 and First Lord of the Admiralty in 1710.

One of the most intriguing aspects to come out of the examinations of the paybooks is the absence of any officer by the name of James Beverly. The significance of this lies in the association of such a person with initials found on some of the pewter plates raised in 1979 (section 9.2.1). In several of the published accounts of that work (e.g. Lyon, 1980: 341) the initials IB (interpreted as James Beverly), along with those of Captain John Johnson (II), were used to distinguish the wreck site as the Stirling Castle, rather than the Northumberland or Restoration. Those same accounts refer to James Beverly as the first Lieutenant of the Stirling Castle. But the first Lieutenant in November 1703 was Benjamin Barnet, who joined the ship in early 1702 as a replacement for David Wavill (TNA, ADM 33/210 and 33/230). Barnett is also listed as the first Lieutenant in another Admiralty document (TNA, ADM 8/8) which lists the ‘Present Disposal of all Her Majesty’s Ships and Vessels in Sea Pay’ as of 1 November 1703. Wavill himself joined the ship in May 1701 as a replacement for William Hamilton, who in turn had been first Lieutenant since the recommissioning of the vessel in March 1701. James Beverly was not part of the crew of the Stirling Castle, as recorded in the paybooks, at any point between 1701 and 1703. So, while matching John Johnson with a selection of appropriately stamped pewter plates is relatively simple, the previous interpretation, linking an individual called James Beverly with other pieces, must be questioned.

The Battle of Vigo drew England and Portugal into a close alliance and also caused the Mediterranean to become an ever more critical area of naval operation aimed at supporting land operations in Spain, as well as cementing the alliance with Savoy (Harding, 2002: 170; Roger, 2004: 168–69). To this end a fleet of 32 ships, including the Stirling Castle (Endsor, 2004: 92), based at Lisbon were dispatched to the Mediterranean under the command of Sir Cloudesley Shovell in August 1703. The fleet operated off the south coast of France without challenge from the French fleet at Toulon and, although Roger (2004: 168) notes that the expedition achieved little, Harding (2002: 170) and Blackmore (2011: 110–11) do link its unchallenged progress to a formal declaration of alliance by the Duke of Savoy in October 1703. The fleet was at Livorno (Leghorn) in September 1703 (Martin-Leake, 1920: 123). At that time, the lack of a permanent Mediterranean base meant that it was normal for the fleet to return to England for the winter and it departed from Livorno in early October 1703 (MartinLeake, 1920: 125), returning to the Channel where it was at anchor in the Downs, off the east Kent coast, on 26 November 1703.

The paybooks were obviously an important mechanism in the successful operation of the vessel and in the accounting of the pay of those who served on board. From the perspective of understanding the ship and its crew, the paybooks are far more than a simple accounting tool, however. They allow a relatively clear view of the structure, numbers and turnover of the crew of an individual vessel. The paybooks serve to highlight the exact numbers involved in the process of ‘turning-over’ of crew when a captain switched command as well as illustrate the wider turnover of crew that took place on a day-to-day basis. For example, across the whole period of the ship’s commission following the 1699 rebuild, from March 1701 to November 1703, a total of 772 men 10

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 signed onto the crew, with a further 129 marines also embarked from January 1702. Of these – on the basis of their names being listed in the paybook as having signed onto the ship, and having not been subsequently discharged – 285 sailors and 33 marines were aboard at the time of the vessel’s loss. Forty-two marines had been discharged from the ship on 22 November 1703. Given this seemingly very high throughput of crew, it is perhaps remarkable that 103 of the 285 had joined the ship during its initial period of recruitment in March and April 1703. Out of the 40 crew that arrived with Captain Johnson in late May 1701, 21 of them were still serving the ship when it was lost.

(section 8.2.1) which raises questions about how we think about the time depth present on shipwreck assemblages – a theme returned to in chapter twelve. 2.4 The Loss of the Stirling Castle: 26–28 November 1703 The naval vessels at anchor in the Downs on the 26 November 1703 were accompanied by a large number of merchant vessels. Estimates in different sources range from about 100 ships, to over 160, of which only 70 were present on the morning of the 27 November (TNA, ADM 51/291; TNA, ADM 106/3250). The storm that they had been subjected to was soon referred to as ‘the Great Storm’ (Figure 2.4) and caused havoc at sea and ashore, along the length of the English Channel coast and across southern England (for an extended account see Brayne, 2003). The storm also inspired Daniel Defoe to compile eyewitness accounts from across the south of England into his first book, entitled The Storm, alongside two shorter works, An Essay on the Late Storm and The Lay-Man’s Sermon upon the Late Storm, in 1704. These works both serve to emphasise the cataclysmic effect of the storm on the physical fabric of southern England and contain numerous accounts providing useful context to the loss of the naval vessels on the Goodwin Sands. A further intriguing element of Defoe’s account is the extent to which the storm itself was seen as a reflection of the wider woes of the country in 1703 (Hamblyn, 2003: xxxxv). In the context of the loss of the Stirling Castle, Defoe’s work is perhaps most wellknown for his assertion that the boatmen of Deal were more interested in plunder than in rescuing the stricken sailors marooned on the Goodwin Sands - a claim strenuously denied by the townspeople following the Great Storm.

The paybooks also remind us of some of the aspects of the vessel’s crew that are likely to remain invisible in the archaeological record. Today, it is easy to take the great variation in the forenames of individuals for granted. On board the Stirling Castle in November 1703 53 of the crew would have answered to the name of John, 42 to William and 38 of them to Thomas. At the other end of this scale, it is unlikely that any trouble with duplicate names would have been experienced by midshipman Polycarpus Taylor. A noteworthy occurrence is the appearance between 19 and 23 March 1701 of a Bartholomew Roberts (Ordinary Seaman). A sailor with the same name was later to achieve fame and notoriety between 1719 and 1722 as the pirate ‘Black Bart’, but nothing is known about his early life other than his going to sea in about 1695, aged 13. The appearance of this name in the paybook of the Stirling Castle is almost certainly coincidence. But the chronology of Black Bart’s early career at sea does fit with the dates of the Stirling Castle, and so the unlikely possibility that they are one and the same person must remain.

The sequence of events surrounding the loss of the Stirling Castle is not completely clear, but an attempt is made here to bring the various available accounts together as a mean to further our understanding of the subsequent archaeological remains. The vessel itself seems to have been originally anchored by two of its anchors, probably the ‘best bower’ and ‘small bower’ (Endsor, 2004: 94), the later recovery of which was reported to the Navy Board by Thomas Warren, the Navy Board representative at Deal (TNA, ADM 106/3250). The cable recovered with one of these anchors, of a length of 95 fathoms (174m), indicates that the ship was anchored at a single cable’s length (Endsor, 2004: 94). Endsor (2004: 94) highlights the difference in anchoring practice with the second-rate 96-gun Prince George, which was anchored with two and two thirds cables spliced together (Martin-Leake, 1920: 128). This more secure arrangement, presumably adopted in anticipation of the coming storm, was vindicated by the Prince George’s survival at anchor in the Downs. The greatest moment of distress for the vessels anchored in the Downs seems to have been at around 5 a.m., when Thomas Warren saw the majority of flashes from guns fired by ships in distress. Warren wrote to the Admiralty on the 27 November to describe the scene that had unfolded in front of him overnight and that morning;

Finally, the pay books also serve to illustrate another, perhaps often overlooked, aspect of the routine service of a vessel like the Stirling Castle: namely, the use of the ship for transporting naval seamen between ships. On the 13 September 1701 139 sailors from the third-rate 70-gun Royal Oak and the third-rate 80-gun Ranelagh boarded the Stirling Castle at the Nore in order to be transferred to the Newcastle at Plymouth. They were joined on the 14 October by a further 137 sailors from the third-rate 80gun Torbay, who were transferring to the Kingfisher, also at Plymouth, and on the 16 October by 145 sailors from the Ranelagh and Torbay who were also destined for the fourth-rate 50-gun Newcastle, itself wrecked at Spithead in the Great Storm of 1703. This voyage of the Stirling Castle, with its additional 421 passengers, must have been a swift one, for the transfer was completed on the 21 and 22 October 1701 and the wages for the additional men were paid on the 6 December. This aspect of the Stirling Castle’s service history is significant for one other reason. Included in the list of sailors that boarded on the 13 September from the Ranelagh is an able seaman called William Stonas. Although on board ship for only a few weeks, he left behind a permanent record of his presence in the form of an onion bottle bearing his name 11

Whitewright

Figure 2.4. The Great Storm of 26 November 1703. Engraving titled ‘The Great Storm. November 26 1703. Wherin Rear Admiral Beaumont was lost on the Goodwin Sands’. A large number of ships are shown in stormy conditions in a confined seaway and, to the left of the image, three vessels have sunk. A two-decked warship to the right of the scene is shown hardpressed and is rendered in a similar manner to the third-rate 70-gun Resolution in Figure 2.1. In the centre foreground a naval vessel is shown aground with only the poop above the water. Men scramble from the ship to a sandbank, from where they are rescued by a rowing boat. The vessel on the right and the vessel aground share a number of similar features and the artist may therefore be showing the sequence of events of the ship foundering and then running aground, before some of the crew escape to the sandbank for rescue. The disposition of the scene, the similarities with Figure 2.2 and the fact that there are survivors indicates that the wrecked vessel in the foreground is probably the Stirling Castle (© National Maritime Museum, Greenwich, London, PAH7410/ PY7410).

any man here, for when the Ships broke away, as the greater part of them did about five in the morning, most flashes of Guns being then seen, and the wind being then highest at WSW they could not carry a knot of Sail to cling the shore that tis feared the Goodwin Sand took up most of them, for when it was light we could see five sail, two of them pretty big ships fast upon it; one I am afraid was a man of war by the number of men we could discern to shelter themselves upon the masts, who all perished as the waters flowed and the sea broke upon her; and there is little of them now to be seen. (TNA, ADM 106/3250)

Right Honourable, I am sorry I must acquaint your Honours of a miserable slaughter that has happened amongst the ships in the Downs; yesterday it blew very hard all day, but about midnight it came on so violently at SW and continued between that and the WSW that about two in the morning Guns were seen fired from Ships in distress, but there being such a hurricane twas impossible for a man to stand against it, much more to offer them any help from the shore; that out of about 100 sail of one sort or another there is not above 70 now to be seen and a great many of them floating only bottoms, having I presume cut their Masts by the Board the better to ride it out; of the Men of War Rear Admiral Beaumont on the Mary is missing, with the Northumberland, Restoration, Stirling Castle and Mortar Bomb… tis a miserable spectacle to behold the Ships in the Downs as they now ride torn almost to pieces by this storme and the like weather has not been known in the memory of

The eventual rescue of some of the crew of the Stirling Castle indicates that it was almost certainly among the five ships reported by Warren that were aground and visible on the morning of the 27 November. The Mary, Northumberland, Restoration and Stirling Castle were recorded by Admiral Leake, aboard the Royal George, as visible, but as having been smashed to pieces and 12

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 lost by 10 a.m. (Endsor, 2004: 93; Martin-Leake, 1920: 128). A number of observations from other vessels in the fleet (Eagle (TNA, ADM51/291), Shrewsbury (TNA, ADM51/833)) that survived the night at anchor note the rapid destruction of the Mary, but state that the Stirling Castle, Northumberland and Restoration appeared to be relatively intact, at least initially. In fact, the Stirling Castle survived a further 24 hours in a grounded state. Seemingly the crew managed to let go another anchor, the smallest ‘spare’ anchor, which, although it could not hold the ship fast, did enough to allow it to go aground with its bow towards the wind rather than being turned broadside on and totally lost. This anchor was found in 2002 160m to the south-west of the wreck, with its shank pointing towards the wreck (Endsor, 2004: 95).

whose story is recounted in Defoe’s Storm (1704: 172– 76). Adams recalled that By Two a-clock we could here Guns firing in several Parts of the Road, as signals of Distress; and tho’ the Noise was very great with the Sea and Wind, yet we could distinguish plainly, in some short Intervals, the Cries of poor Souls in Extremities. By Four-a-clock we miss’d the Mary and the Northumberland, who rid not far from us, and found they were driven from their anchors; but what became of them, God knows: and soon after a large Man of War came driving down upon us, all her Masts gone, and in a dreadful Condition. We were in the utmost Despair at this sight, for we saw no avoiding her coming thwart our hawser: she drove at last so near us, when it pleased God the Ship sheer’d contrary to our Expectation to Windward, and the Man of War, which we found to be the Sterling Castle, drove clear off us, not two Ships Lengths to Leeward.

Prior to that, a key moment in the overall wrecking process of the Stirling Castle seems to be the point at which the Northumberland dragged anchors and, when these parted, was driven downwind, out of control, to foul the anchor cables of the Stirling Castle. This observation is included (highlighted in bold) in Thomas Warren’s letter to the Navy Board on the 29 November, in which he also reports on the eventual rescue of the survivors from the Stirling Castle:

It was a Sight full of terrible Particulars, to see a Ship of Eighty Guns and about Six Hundred Men in that dismal Case; she had cut away all her Masts, the Men were all in Confusions of Death and Despair; she had neither Anchor, nor Cable, nor Boat to help her; the Sea breaking over her in a terrible Manner, that sometimes she see’d all under Water; and they knew, as well as we that saw her, that they drove by the Tempest directly for the Goodwin, where they could expect nothing but Destruction; The cries of the Men, and the firing of their Guns, One by One, every Half Minute for Help, terrified us in such a Manner, that I think we were half dead with the horror of it.

Right Honourable The Wreck I gave your honours an account of in mine of yesterday [28 November] that was seen to have a weft out at the Goodwin Sand, proved to be the Sterling Castle, the boats that went from hence [Deal], and one that went from Ramsgate having saved about 70 of her men in which Number is four Officers of the Marines, the third Lieut., Chaplain, and Cook of the Ship; the rest of the Officers and men being drowned; amongst this Number that was saved belonging to the Sterling Castle, was one of the Mary’s Company who was Coxswain of her Yawl; he gives an account that she parted with all her cables and drove upon the Goodwin where she broke to pieces, he saw Rear Admiral Beaumont upon a piece of her Quarter, with his Clerk and Lieut. Sampson, and this man got upon another piece of Wreckage, and by swimming and getting from one piece to another, he was drove to the Sterling Castle with two men more, who died as soon as they got on board, that this man is all that is saved of the Mary’s Company; these people that are brought ashore can give no certain account of the Northumberland and Restoration; some say the former drove foul of their ship Sterling Castle, and put them away and soon after she sunk down, and others that they were both lost upon the Sand; I am afraid tis too certain they are gone, and none of their Company’s saved to give any account of them;. The Mortar Bomb is also still missing and no manner of account to be had of her. (TNA, ADM 106/3250)

Although entirely compelling, some aspects of Adams’ account are hard to reconcile with the other eyewitness accounts. Most notably, he notes that his ship was close to the Northumberland, which was driven from its anchors at 4 a.m., following which the Stirling Castle appeared, being driven downwind towards him. If the other accounts of the Stirling Castle’s anchors having been fouled by the Northumberland are correct, then the situation described by Adams is impossible, because the Northumberland could only have fouled the Stirling Castle after being driven from its own anchors, which would have happened downwind of where Adams was moored because of the direction of drift of the Northumberland. As a result, it would have been impossible for the Stirling Castle to have drifted past him. Given what is known about the subsequent orientation of the Stirling Castle at the time of striking, and the fact that the ship was dragging an anchor when it went aground, Adams’ assertion that the ship had ‘neither anchor, nor cable’ also seems at odds with the other evidence.

An eyewitness account of the Stirling Castle in the midst of the storm is provided by J. Adams, the captain of a ship that was blown from the Goodwin Sands to Norway,

Although the historical accounts of what happened in the early hours of the 27 November are hard to entirely piece together, the broader situation at daybreak is much clearer, 13

Whitewright with five ships visibly aground on the Goodwin Sands. The Captain’s log from the third-rate 70-gun Eagle, which survived the night at anchor, recorded that

windows into the operation of the ship on a day-to-day basis, while others, such as the letters of Thomas Warren, provide a better understanding of the process of loss in November 1703.

at the same time [9 in the morning] had the melancholy prospect of seeing the Mary, Sterling Castle, Northumberland and Restoration ashore on the Goodwin Sand, the first beat to pieces in a little time, the others set whole, above 100 sail of merchant ships, etc are missing, some hundred at an anchor, some upon the Break and Querne, other some no mast standing as are several men of war that lie in the Downes, in the afternoon [27 November] Vice Admiral Leake made the signal for all the boats to go and save what men they could from the wrecks, but is blew so hard that they could not effect it. (TNA, ADM 51/291)

How we consider some of this material is not a simple matter. Some material may be regarded as relatively straightforward, ‘traditional’ historical documents – the letters of Thomas Warren to the Navy Board are perhaps a good example of this. Others, however, are far more complex. While the ship’s logs and paybooks may be viewed simply as pieces of historical paperwork, especially when they are consulted in their present form within the National Archives, bound together with other documents of the same type and of a similar date from ships with alphabetically similar names, in their original form those documents were constantly evolving and changing to reflect the ongoing life of the ship and the ship’s company – its geographical movement, the comings and goings of crew, promotions, demotions and so on are all contained within their pages. We have to remember, therefore, that such documentation was a critical and fundamental mechanism in the successful day-to-day operation of the vessel itself, and of the wider institution of which it was a part. The direct historical archive of the Stirling Castle is therefore as much a part of the vessel’s material culture as the more physically tangible remains, be they cannons or candlesticks, that make up the archaeological archive that this publication was originally conceived to address. In that respect, when we concern ourselves with the material remains of the ship, the crew and the institution of which they were part, there can be no real divide between the historical record and the archaeological record. There is simply a record. It may be comprising a myriad of component parts, often constituted for very different reasons, but all had the same overall purpose of facilitating the social, ideological and economic operation of a British warship in the first few years of the eighteenth century.

Based on Admiral Leake’s account (Martin-Leake, 1920: 128), the Northumberland and Restoration seem to have broken up soon after the Mary, by about 10 a.m. on the 27 November. The Stirling Castle, with its bow partially towards the weather, fared better, with a number of survivors visible on the vessel’s poop deck, which was still above the water (Martin-Leake, 1920: 129). Having survived the initial storm and wreck, the men on the Stirling Castle then had to endure a further 24 hours on board the stricken ship before the weather abated enough for boats from Deal to reach them, as reported by Thomas Warren to the Navy Board. Of the 318 crew, both sailors and marines, on board the vessel at dusk on 26 November, only 70 were rescued from the shattered ship on 28 November. 2.5 Conclusion The account of the service history of the vessel provided above is intended in the first instance to provide context for the archaeological material discussed in subsequent chapters including the physical structure of the vessel and the artefacts contained on board. Through that account it is possible to place the vessel within some of the wider social, political and economic trends of the day, from which in turn we can account for some of the varying rationale behind the construction and deployment of the ship in a general sense. Within this contextual overview, the biographical narrative of the vessel and crew raises some important questions relating to our conception of shipwreck assemblages, the time depth of surviving material on the Stirling Castle at the time of loss, and whether or not we should consider the vessel a ‘Restoration Warship’, as it is commonly referred to, or simply as a warship of 1699–1703. All of these themes are returned to in chapter twelve. More directly related to the ship itself is the historical documentation concerning the vessel, which forms part of the wider archive of the Stirling Castle. Although now housed within the National Archives, rather than being preserved on the site in an archaeological sense, documents such as the paybooks represent significant 14

3 The Archaeological History of the Stirling Castle Julian Whitewright The history of archaeological investigation of the Stirling Castle spans a long period from the discovery of the wreck site in 1979 to the present day. The work undertaken on the site during this time has encompassed a wide range of people, groups and institutions. Some of the individuals involved have since passed away, while some of the groups and institutions no longer exist or have been replaced by successors or, in some cases, completely new entities. As a result of this, understanding the archaeological remains of the ship and its associated artefacts is difficult without first establishing the various processes of discovery, salvage, archaeological intervention, site management, archiving and conservation that have taken place since 1979. Figure 3.1. The 1979 team on board Tom Brown’s fishing vessel Shelandra (image courtesy of SeaDive).

This chapter therefore attempts to set out the history of the archaeological investigation of the site in the fullest way possible. It is based primarily on the surviving site archive, which includes licensee and Protection of Wrecks Act 1973 (PWA) contractor reports, formal and informal publications, correspondence between individuals and face-to-face interviews. Section 3.1 takes a chronological approach in setting out the work conducted on-site between the ship’s discovery in 1979 and the initiation of the current project and formal collation of archive for this project in 2009. Subsequently section 3.2 details the management issues relating to the ownership of archaeological material and the associated supporting archaeological archive in order to address and explain the currently dispersed nature of the site archive. Finally, the entirety of work on the site is considered in section 3.3 against the wider trends observable within English maritime archaeology across the same time period.

site of one of the vessels sunk during the infamous Great Storm of 1703. The method of locating and ground-truthing the known net fastenings that represented potential wrecks was straightforward, as described by Tom Brown during an interview in 2010. A fishing trawl was dragged across the location of the net fastening with a chain across the trawl opening. This would catch on the upstanding net fastening, at which point the fishing boat would be winched backwards until it was over the location, allowing the divers to descend down to investigate the fastening. The site that would be identified as the Stirling Castle was numbered simply as ‘Wreck II’ during the early seasons of work. The first diver on the site in 1979 was John Chamberlain, described by Tom Brown as returning to the surface with an onion bottle in each hand, shouting that there were ‘cannons everywhere’. Following this initial discovery, the dive team informed Fred Wall, the secretary of IOTAU, who in turn contacted David Lyon at the National Maritime Museum (Greenwich) and Keith Muckleroy. Lyon subsequently reported (1980) the discovery of the vessel in the pages of the International Journal of Nautical Archaeology.

3.1 Archaeological Investigation 3.1.1 Discovery and Exploration: 1979–1991 The discovery of the Stirling Castle took place as a result of exploration work undertaken by divers from British Sub-Aqua Club (BSAC) 106. Five members of the club, John Chamberlain, John Cayzer, Paul Fletcher, Roy Kennet and Keith Young (Figure 3.1), joined forces with the Isle of Thanet Archaeological Unit (IOTAU) through Dave Perkins to survey shipwrecks located off Ramsgate (Perkins, 1980: 3). Another key member of the group was Tom Brown, a local fisherman who provided the use of his boat, Shelandra, and the knowledge of many of the net fastening locations on the Goodwin Sands held by the local fishermen. Speaking in 2014, Chamberlain recalled that their initial searches had been for a vessel lost in the Anglo-Dutch wars of the seventeenth century, but, as these searches were unsuccessful, attention turned to the possible

The identity of the wreck was not known at the time of its discovery, but was confirmed as the Stirling Castle as a result of the stamps of personal initials on pewter plates recovered during 1979. Many of these carried I.I. or I.B., corresponding with the initials of Captain John Johnson and first Lieutenant James Beverly (Lyon, 1980: 341; although see section 2.3), at a time when ‘J’ was stamped as ‘I’. In 1979 the vessel was observed and reported as standing 6-8m clear of the seabed, orientated with the bow to the west and 15

Whitewright

Figure 3.2. Artistic impression by Dr David Perkins of the extent of the Stirling Castle as discovered in 1979 (image courtesy of the Isle of Thanet Archaeological Society/Dr David Perkins, late President of IOTAS).

with much of the port side visible, while the starboard side remained buried by a sandbank (Lyon, 1980: 339; Fenwick and Gale, 1998: 96-97). This was largely confirmed by John Chamberlain, who recalled that the sides of the vessel stood some 25 feet clear of the seabed. It is also very clear that the site was in a remarkable state of preservation when exposed during 1979. Those who dived the site during that period consistently report the presence of a preserved lower gun deck and parts of the main deck, with debris from the forecastle and quarter deck/poop present at either end of the vessel (cf. Lyon, 1980: 339-40). The ability of the divers to enter the hull of the ship has also been consistently stated during subsequent interviews (Brown, pers. comm.;1 Chamberlain, pers. comm.2). Despite this, both Tom Brown and John Chamberlain are adamant that the artistic representations (e.g. Figure 3.2) of the site produced by Dave Perkins exaggerate the appearance of the site as seen in 1979. Nevertheless, such visualisations have become the accepted representation of the completeness of the vessel remains at that time, something that is explored further in chapter five. The main work on the site in 1979 focused on the recovery of artefacts, including the vessel’s bell (Figure 3.3), many

Figure 3.3. The bell of the Stirling Castle being brought on board Shelandra in 1979 (image courtesy of SeaDive).

of which were exposed on the seabed and subject to rapid degradation or potential salvage by other parties (Lyon, 1980: 340; Perkins, 1980: 4). Efforts were made to record

Tom Brown recalled the work on the site during an interview in October 2010. 2  John Chamberlain was interviewed in October 2010 along with Diane Chamberlain, and on his own in June 2014. 1 

16

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 under the PWA on 7 May 1980, with full designation coming into force on 6 June 1980. Subsequent amendments were made to the centre-point of the site in September 1980 and an expansion of the protected area from a radius of 50m to 300m was made in 2004. At the time of its discovery and designation, ownership of the Stirling Castle rested with the Ministry of Defence (MOD), who sold the remains of the ship and its equipment to IOTAU on the 11 November 1980 for the sum of £200. In 1981 a letter from the Department of Trade (who administered the PWA at that time) to the Receiver of Wreck in Ramsgate confirmed that IOTAU was the legal owner of the ship and any equipment from it, but that personal items were not included in the material sold by the MOD. These were subsequently purchased from the Receiver of Wreck by IOTAU in 1982. IOTAU subsequently became the Isle of Thanet Archaeological Society (IOTAS) in 1988 and the ownership of the wreck and material from it now rests with that organisation (section 3.2).

the relative positions of material across the wreck site with a centreline and offset survey (John Chamberlain, pers. comm.); however, gauging the accuracy of the work undertaken during the 1979 season is difficult. Tom Brown recalled that the divers ‘did the best that they could, for a load of amateurs’. John Chamberlain also noted that the primary activity was the recovery of artefacts, rather than the accurate recording of where they were located on the site. This is largely reflected in Lyon’s published account (1980: 340), which stated that it was not possible to achieve the desired accuracy of locational recording because of the difficulty of the underwater conditions. A return to the site was planned for 1980 in order to undertake more extensive and detailed survey, but the vessel was found to have totally disappeared beneath the shifting sand of the area (Perkins, 1980: 7). However, as a result of the 1979 season of work the Stirling Castle was temporarily designated as a protected historic shipwreck

Archaeological recovery strategy on the Stirling Castle It will be clear from this account of the work undertaken on the Stirling Castle, along with the discussion of the wider context of the site presented in chapter four that the environment of the site was an extremely difficult and changeable one within which to work. Almost every report on archaeological activity, by the original dive team, the PWA licensee and the ADU/PWA contractor, records significant amounts of time being lost to the weather, the underwater conditions, or both. The challenging nature of the site has to some extent, along with the available archaeological expertise, dictated the types of archaeological intervention that have occurred at different times and has certainly played a large part in the successful application of the chosen strategy.

proceeding very much in the mould of a ‘rescue’ project, in which it was perceived that there was an imminent threat of many of the exposed artefacts being destroyed through tidal action, or of the site being rapidly reburied. Reference to the artefact list from the 1979 season confirms the approach recalled by Chamberlain, with 28 per cent of the recovered artefacts being organic in nature and 66 per cent non-organic (a further 6 per cent were of unknown class). The artefacts of non-organic materials included the pewter plates that were central to identifying the ship, as well as the vessel’s bell. Organic artefacts were clearly not a target for recovery by the dive team on the seabed, despite the published opinion, and neither were the human remains that members of the dive team reported seeing on the site.

The 1979 season of work was undertaken with very little formal archaeological assistance and this, in conjunction with the obvious vulnerability of newly exposed artefacts, led to a clear emphasis on the recovery of artefacts, rather than the recording of the disposition of the underwater remains. Some efforts were made to assign recovered material to one of three general zones on the site; fore, mid-ships and aft (Lyon, 1980: 340). Despite this, it is clear that the range of artefacts recovered in 1979 does not reflect the breadth of material that was exposed on the site at the time. Conservation facilities were extremely limited and so a conscious effort was made to raise only artefacts that would not need extensive treatment once on dry land (Chamberlain, pers. comm.). In contrast to Chamberlain’s memory of the rationale for selecting artefacts for recovery is the published statement at the time that an emphasis was placed on the recovery of organic material (Lyon, 1980: 340). The same account also makes it clear that the work on the site was

The result of the focus on artefact recovery meant that it was not until the late 1990s that concerted efforts were made to produce a cohesive record of the seabed remains. Even then, such work still had to take into account the range of artefacts uncovered at each visit to the site and which were sometimes destroyed or lost through natural processes during a period of survey work. The most well-documented example of this is the log reel that was uncovered and lost in 1998, while awaiting a licence to recover it. A feature of the more recent work on the site has therefore been to record the extant structure while attempting to document any changes to the disposition of the site, while at the same time being prepared to raise artefacts in danger of loss following exposure. Material recovered in this way ranges from an iron cannon and wooden carriage in 1999 to a collection of barrel staves raised in 2006. 17

Whitewright In 1981 bad weather prevented any diving on the site during periods when boats and divers were mobilised. Information on the site was provided by a sidescan survey of the Stirling Castle, Northumberland and Restoration, which indicated that the latter two were still clear of sand and that a portion of the Stirling Castle had emerged from the seabed. The surviving archive correspondence between the Board of Trade and IOTAU indicates that no formal diving took place during 1982. Diving on the site in 1983 included guests from the Medway Marine Archaeological Society and in one instance Alexander McKee of Mary Rose fame. The reports submitted to the Department of Trade in fulfilment of PWA obligations state that bad weather again hampered attempts to conduct any fruitful diving work.

From 1986 onwards the work taking place on the site was monitored by the newly established Archaeological Diving Unit (ADU), which had been set up to assist in the management of sites designated under the PWA. The director of the ADU, Martin Dean, had visited and dived on the site in 1979; following his visit in 1986, he was therefore well placed to comment on the archaeological remains of the vessel. It is instructive to quote him in full (ADU, 1986: 2): 2. Archaeological Remains The Application for Designation, written in 1980, described the wreck as being almost complete from gundeck to keel, and the writer [Dean] when he visited the site in 1979, confirms that gunports were visible on the port side in the mid-ships area, together with other recognisable features such as sternpost and rudder and a mast stub. All that can be seen now is a massive jumbled pile of displaced ship’s timbers with no obvious or recognisable form, lying in a scatter with an appreciable ‘dogleg’ towards the bow. The prediction on the Application that ‘… now the wreck is exposed from the sands deterioration will be fairly rapid’ appears to have been accurate.

Some meaningful archaeological work was carried out in 1983 and 1984, when a photographic and geophysical survey was undertaken by the ‘Goodwins Archaeological Survey’ that included the site of the Stirling Castle within a wider investigation of the maritime archaeology of the Goodwin Sands (see Redknap and Fleming, 1985; Redknap, 1990). During the same period concerns were recorded in the meeting of the Advisory Committee on Historic Wreck Sites (ACHWS) regarding unsubstantiated rumours of the site being plundered. Rumours of the use of explosives on the site were confirmed by John Chamberlain during an interview in 2014, in which he recalled that the dive team found dynamite left by another group scattered around the edge of the site during the early years of work. Further concerns about unlicensed activity on the site were raised in 1989 in letters to the Department of Transport regarding the availability of dive trips, for a nominal charge, to the Stirling Castle, despite the fact that the site was designated and such activity was therefore illegal.

Martin Dean’s personal observation of the site indicates that significant deterioration took place in the seven years between the ship’s initial discovery in 1979 and the visit of the ADU in 1986. The ADU report also confirmed the general inaccuracy of the existing site plans and recorded the fact that they were based on approximate measurements, verbal descriptions and post-dive sketches (ADU, 1986: 5). On a more positive note, the ADU were complimentary about the post-fieldwork treatment of recovered artefacts and of the publication of material that had been undertaken up to that point, comprising two locally produced pamphlets (Perkins, 1979; Perkins, 1980) and a short article published in the International Journal of Nautical Archaeology (Perkins, 1983) addressing the lead musket shot recovered from the site.

Further letters between IOTAU and the Department of Transport (now in charge of the PWA) during 1984 indicate that some diving did take place on the site and that the bow of the Stirling Castle was visible above the sand to a height of about two metres across an area of 20m × 15m. Whereas the port side of the vessel had been exposed in 1979, reports on the 1984 work (Perkins, 1984) indicate that the starboard side of the wreck was now exposed, primarily the forward half of the vessel. The overall impression of the divers on the site in 1984 was that little structural damage had taken place to the ship itself and it was noted that the 1979 datum line remained in place. The nature of the work, amounting to some 20 dives, during 1984 was restricted to photographic survey as prescribed by the ACHWS to the Department of Transport, which in turn issued the licence for the 1984 season of work. Despite this, a small number of artefacts deemed to be ‘at risk’, mainly bricks from the hearth, were raised. Petrological analysis of these assigned their production area to a brick-field to the east of Chatham in Kent, which tallies with the location of the rebuild and relaunch of the Stirling Castle at Chatham dockyard in 1699 and 1701 (described in section 2.2).

The ADU visited the site again in 1987 and conducted a single dive, their report recording that slightly more of the site was exposed than during the 1986 visit. This was the last visit to the site by the ADU until 1992 and brings to a close the first period in the archaeological investigation of the site. Direct diving work done under the PWA licence since the initial 1979 season seems to have been limited owing to bad weather and internal discord within the dive team, although in some seasons, such as 1984, a significant number of dives took place on the site. Despite the restricted nature of the archaeological work, in both extent and rigour, some progress was made off the water, notably in the consolidation of the post-excavation work and the establishment of Ramsgate Maritime Museum to put the artefacts from the Stirling Castle on public display under one roof. Alongside this, no other concerted licensed diving seems to have taken place prior to 1992, although the site was geophysically surveyed as part of 18

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 the wider Marine Archaeological Surveys (MAS) project that operated around the south-east coast of England in the late 1980s. Mark Redknap (Director of MAS) became the licensee of the Stirling Castle in the late 1980s and, although diving was undertaken on the site of the Northumberland, licensee reports from this period indicate that no diving occurred on the Stirling Castle (although a letter from Redknap to the Department of Transport in 1989 notes that sports divers were being offered low-cost excursions to the Stirling Castle, despite its designated status).

The ADU visited the Stirling Castle again in 1995 in conjunction with Bob Peacock. The site was again considered to be in a stable condition, potentially augmented by the build-up of sediment under the fishing nets that had accumulated on the wreck (ADU, 1995: 1). The wreck mound at that time was recorded as being some 3m high, with a ‘considerable quantity’ of wreck material visible, including disarticulated timbers, iron guns and a large anchor (ADU, 1995: 1). The concluding observation of the ADU was that the overall stability of the site was uncertain, primarily because of the ongoing lack of any baseline data with which to make year-on-year comparisons. This view of the site was largely repeated two years later in the ADU’s 1997 report. Peacock’s licensee report for 1996 provides further detail on the extent of structural remains visible (Peacock, 1996), recording lower levels of sediment than in previous years; although the starboard side of the vessel was still totally covered in sand, the port side was again exposed. This exposure included timbers in pristine condition, suggesting exposure for the first time. Additionally, two cannon tampions, the remains of a brick hearth and two anchors were visible on the site.

3.1.2 1992–2002 During their 1992 visit to the Stirling Castle the ADU recorded that the site comprised a well-defined wreck mound upon which archaeological material not seen since 1987 was exposed as a result of some loss of sand cover (ADU, 1992: 1). The ADU also commented that the exposed structural material comprised a complete run of planking on both sides of the site, in addition to guns and anchors. The stern of the ship was observed as having fallen away, with other material exposed up to 1.5m proud of the seabed. Rumours of the use of explosives on the site, along with unlicensed diving, were also reported, but could not be confirmed. The disposition of the site reported in 1992 bore no resemblance to any of the existing surveys, because of either changes in the intervening period or the incompleteness of earlier work (ADU, 1992: 2). The ADU also stated (1992: 2) that the creation of a baseline survey was urgently required to allow future assessments of change to the site to be made. This latter observation, along with the previous comments on the comparative disposition of material, confirms the fact that the locational accuracy of surveys conducted in the initial seasons of work was extremely limited. Consequently, it is very difficult to extract any significant information from the apparent distribution of recovered artefacts beyond their general position in the bow, stern or mid-ships areas. This is discussed further in chapter five in relation to the development of an overall site-plan for the Stirling Castle, as derived from all the seasons of work.

The extent of sediment cover had always been recognised as a critical factor in the stability of the site, with decreases in the overall sediment levels leading to a decrease in the stability of the exposed remains. Attempts were made to gather an informed dataset on this through geophysical surveys undertaken by the ADU in 1997, which concluded that there had been a slight increase in sediment levels, which in turn was providing some stability (ADU, 1997). Meanwhile, the UKHO surveyed the site and also reported that sediment accretion had occurred around the wreck (UKHO, 1997). Despite this, a trend for the site to become increasingly exposed was noted by Peacock’s work through the SeaDive organisation. In 1998 it was reported that sudden significant sediment loss had left the site in a state of exposure similar to that seen in 1979, a conclusion confirmed by the ADU (below). Fortunately, by 1998 plans were being made by the licensee, Bob Peacock, for a concerted period of fieldwork to take place in 1999. Called Project Man O’War, it comprised members from SeaDive, the Nautical Archaeology Society (NAS), the ADU and a group of divers from the USA. The sudden exposure of the vessel in 1998 to a level not seen since 1979 emphasised the timeliness of that planning and led to the ADU assigning two weeks to support the planned 1999 survey project. In partial justification of this extended period of time on a single site, the ADU (1998: 1) stated that

In contrast, a subsequent visit by the ADU in 1993 observed that no archaeological material was exposed because of higher sand levels on the site than in previous years, indicating that sediment levels could change dramatically in a short period of time, in conjunction with poor underwater visibility (ADU, 1993: 1). The high levels of sediment on the site were also recorded in his report for that year by Bob Peacock, who had become the licensee of the Stirling Castle, Northumberland and Restoration in 1993 under the PWA. The 1993 visit, along with those undertaken in 1992, recorded significant quantities of fishing net that had become snagged on the site and covered much of the top of the wreck mound. These appeared to comprise both nets from fishing at the site and loose nets that had drifted onto the site. Despite the protected status of the site, the ADU concluded that little could realistically be done to protect the site from damage caused by fishing gear impacts (ADU, 1992: 2).

The wreck is almost as exposed as when first found in 1979. Its condition has deteriorated and it is under serious threat from natural processes while exposed. Resources are not likely to be available to physically protect the remains, and there are no tested methodologies which can guarantee to preserve a site in such a difficult environment. 19

Whitewright The collapse of the hull of the Stirling Castle The subject of the extent of the collapse of the vessel’s structure was revisited by the ADU following their 1998 assessment, when sand movement exposed the site to a similar level to that seen in 1979 (ADU, 1998: 2). Martin Dean, who had visited the site in 1979 and 1986, was present again in 1998 and observed that in that year the site was again exposed to the level of the gundeck, indicating that the collapse of the hull was not as total as had been suspected in the 1980s and early 1990s. Despite the apparent continued survival of the lower elements of the vessel’s hull Dean did note that the timber debris on the surface of the site was more jumbled than in 1979 (ADU, 1998: 2). It was also clear to Dean that the hull was not intact in the sense of being a coherent structure, but that it contained several discontinuities at various points. What he could not conclude was whether or not that damage had occurred since 1979 or had always been an aspect of the site during its periods of exposure. As with most other diving activity on the site over the years, the problems posed by poor underwater conditions have hindered the consistent recording of the nature of the exposed remains at different times, making reliable observations about changes to the site very difficult to undertake. All of these elements are returned to in more detail in chapter five.

A critical element to understanding the overall site formation processes acting upon the Stirling Castle concerns the date at which the upstanding elements of the hull (as visible in 1979) seemingly collapsed to form the debris mound that has characterised subsequent observations of the site. The 1992 ADU visit provides initial comment on this, clearly stating (ADU, 1992: 2) that the collapse of the hull occurred in 1981/82, as a result of the inherent weakness of the wooden structure once the supporting sand had moved. Such conclusions about the collapse of the site also appear to tally with Martin Dean’s observations of the extent of the site based on his visits in 1979 and 1986. Further refinement of the date at which the collapse of the hull occurred is potentially provided, however, by the observations (above) of the dive team in 1984 that little change had occurred by that year; indeed, the 1979 survey lines remained in situ. On the basis of the archived reports and letters that address the extent and survival of the site, it could be reasonably suggested that the collapse of the site in fact took place between 1984 and 1986, rather than earlier. The ADU report from 1992 observing the presence of survey lines from six years previously (1986) also suggests that little further disintegration occurred in the intervening period and that the site had a reached a short-term, but delicate, level of stability across the 1986-1992 period. The initiative of the Licensee to organise an international project to record the site in the summer of 1999 is timely. The ADU recommends that two weeks of its time in the summer be allocated to supporting this project. In that way this important wreck can be at least partially preserved by record. As a result of the increase in exposure of the site the Stirling Castle was the primary target for the licensee during the 1998 dive season, resulting in 64 dive hours between July and September that year (Peacock, 1998: 1). This work demonstrated that the sandbank covering the starboard side of the wreck was reduced in size, with a resultant increase in exposure of material in that area. Likewise, the port side of the wreck was also more clearly visible than in recent years, allowing the positive identification of the first gun-port on that side of the ship. A similar pattern was reported at the stern of the vessel (Peacock, 1998: 2), where sections of framing, hull planking and the top of the rudder were all exposed. As a result of these observations, work was undertaken to begin to create a coherent and accurate plan of the site, revising the previous plans from the 1979 season (see chapter five for further discussion of site-plan progression).

Figure 3.4. The dive team during the 1999 season of work (image courtesy of SeaDive).

The 1999 season of work saw Project Man O’War (Figures 3.4 and 3.5) active on the site between May and September, although weather conditions restricted the main period of survey work to June and July (Peacock, 1999: 2). This work was successful in collecting a large amount of

Figure 3.5. Licensee Bob Peacock inspecting the hull of the Stirling Castle in 1999 (image courtesy of SeaDive).

20

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 also undertook further bathymetric survey of the site in 1999 to add to the data previously collected as a means to begin to quantify the rate of exposure of material taking place on the site (see box text on the collapse of the hull and chapter four).

survey data in conjunction with detailed video footage of the site (Peacock, 2000b: 8). The latter approach was an important part of the attempts made to carry out the ADU’s preferred policy, in the absence of viable alternatives, of ‘preservation by record’ (above (ADU, 1998: 1)). Concern was raised by the licensee about the level of exposure of the wreck structure and the rapid loss of exposed artefacts (Peacock, 1999: 2). This was particularly noted at the stern of the vessel, where there was a section of hull 17m in length and 3.3m high, including three visible gun-ports and associated cannon. A number of exposed artefacts were raised by the ADU following a decision to recover any exposed artefacts that could not be protected in situ. These included a copper utensil, a brass candlestick and a probable traverse navigation board (Peacock, 2000b: 8). The difficulty of working on the site was illustrated by the overnight loss of a wooden log reel (Figure 3.6); this had been prepared for removal, but had been mostly lost by the following day as a result of a deterioration in on-site conditions (ADU, 1999: 3; Peacock, 1999: 4).

Fieldwork continued in 2000, with some indications that there had again been significant movement of sands surrounding the site. This reduction in sediment levels led to the complete uncovering of the port side of the vessel, including the garboard-strake and associated outer-hull planking (Peacock, 2000a: 2). This exposure was associated with the outward collapse of the portside stern quarter and the associated collapse of the main gun deck onto the orlop deck, the former assuming an angle of 45 degrees from its previous 20 degrees. It was also observed that large quantities of fishing nets were being snagged on the vessel, with the potential to cause significant damage, including the complete loss of the main cross-beam on the stern of the vessel (Peacock, 2000a: 3). Overall, the work undertaken in 1999 and 2000 served to complete the pre-disturbance plan of the site and to demonstrate that the vessel was still substantially intact, although at clear risk of rapid deterioration. At the same time, further geophysical survey was undertaken by the ADU in order to document the site at the point of seeming maximum exposure (ADU, 2000). The ADU also observed that the vessel’s rudder was exposed to its full height of around 5m and that much of the transom was standing clear of the surrounding sand. They also recorded that at least two of the transom beams had been displaced from their original position in the stern structure. The ADU’s wider comments on the current and future trajectories of the Stirling Castle in 2000 are worth repeating here:

As with the 1998 season of work, the ADU worked closely with the licenced team during the 1999 fieldwork season. The ADU report reiterates (1999: 2) the 1998 observation that the vessel remains were almost as exposed as when first found during 1979. This was thought to have been caused by the north-easterly movement, by some 200m, of the sandbank on the starboard side of the wreck (ADU, 1999: 3). The ADU report also confirmed many of the observations of the licensee (above) regarding the extent of exposed structural and artefactual material during 1999. Unauthorised diving was recorded as having taken place on the site in 1999 (ADU, 1999: 3). This was evidenced by a written sign left on the seabed by the illegal divers that was found by the licensee at the beginning of the diving season. Artefacts from the site were offered for sale in the USA, indicating that artefacts were looted from the site at the same time (ADU, 1999: 3). The ADU

It is difficult to know how best this site might be managed, as there are no proven methods for in situ physical, chemical and biological protection of wrecks. Research being undertaken may eventually help identify the most cost-effective strategies to be employed on this site. Until responsibility for the 1973 Act [the PWA] changes there appears to be little chance of resources being made available for resolving such problems on Designated Historic Wreck Sites. The licensee and his team deserve praise for their constant efforts in caring for this important archaeo­ logical site. Unfortunately they are fighting a losing battle against natural degradation as the ship slowly disintegrates. It is also unfortunate that there is insufficient support from Government-funded heritage bodies to allow even basic preservation by record at a rate faster than that at which information is being lost. (ADU, 2000: 3, 4)

Figure 3.6. The wooden log reel exposed during 1999 and subsequently lost while awaiting licence clearance for recovery. The right-hand end plate was recorded as 30mm thick (image video still from ADU archive, © Crown copyright. Historic England Archive).

The 2000 season was also notable for the exposure of two intact gun carriages, cannon, trucks and tackle, all of which were in imminent danger of being swept out of the wreck and onto the open seabed. Consequently a 21

Whitewright

Figure 3.8. The in situ remains of a box of glass jars, likely to have belonged to William Deas, the surgeon of the Stirling Castle (image courtesy of SeaDive).

dive on the site, although a further sidescan survey was undertaken (ADU, 2001: 1). In 2002 the ADU undertook a planned series of multibeam bathymetric, sidescan and magnetometer surveys (ADU, 2002: 1). One of these was responsible for identifying a large anchor around 160m to the south-west of the site with the shank pointing towards the bow of the ship (Peacock, 2002: 3; WA, 2006: 2). It seems likely that the anchor may have been put out by the crew of the vessel during the Great Storm, prior to the vessel striking the Goodwin Sands (section 2.3). The general trends of exposure of structural material at the stern of the vessel, in conjunction with burial in the bow, were observed again. However, the licensee’s report noted that the level of exposure of the site had begun to decrease overall, with the exception of the stern area (Peacock, 2002: 2). Diving fieldwork was supplemented by the presence of RDF Media filming for Channel 4’s series Wreck Detectives (ADU, 2002). A notched wooden block containing a pair of angled sheaves was recovered in 2002 and was conserved by English Heritage.

Figure 3.7. Archaeological advisor Simon Adey-Davies oversees the recovery of a cannon from the Stirling Castle to temporary storage at a stable location in Ramsgate harbour in 2000 (image courtesy of SeaDive).

decision was made by the licensed team to recover one of the guns (Peacock, 2000a: 3-4; 2000c). The recovery (Figure 3.7) took place in unusually clement conditions and was overseen, albeit with some methodological reservations, by the ADU (ADU, 2000: 2). After initial work undertaken in Ramsgate Maritime Museum, the conservation of the gun and carriage was contracted through funds granted by the Heritage Lottery Fund, to the Mary Rose Trust, ownership of the gun has now been passed to the National Museum of the Royal Navy.

3.1.3 2003–2009 The cessation of fieldwork at the end of the 2002 season represents a convenient cut-off point in the narrative of the archaeological work undertaken on the Stirling Castle. The National Heritage Act (2002) passed responsibility for managing the PWA to English Heritage, in line with the management of terrestrial archaeological remains. In 2003 Wessex Archaeology (WA) took over from the ADU as the archaeological contractor for the PWA and it is this change in the statutory monitoring of the site that is really represented by this section break, rather than any change to the site licensee or any environmental consideration.

In 2001 work on the site by the licensed team demonstrated the continued exposure the vessel from midships to stern. This was concluded to have been brought about by a change in seabed current direction, coming from the south-east rather than the south (Peacock, 2001: 2). By 2001 the exposure of the stern area of the site in particular was considered to be greater than when the site was first observed in 1979. In contrast, the bow of the vessel was becoming buried along the port side up to 18m along the previously exposed structure. An area in the midships section of the wreck was also exposed and revealed the remains of a sea chest along with artefacts that could be associated with the ship’s surgeon (Figure 3.8). Before these artefacts could be raised a large number of them were removed by unauthorised diving, although the remaining items were raised. ADU activity on the site in 2001 was severely curtailed by the weather and they were unable to

Work on the site by the licensee continued in 2003 across some 27 hours of underwater survey. This activity demonstrated a set of overall trends similar to those seen in the 2002 and 2001 seasons of work, with the ongoing exposure of the stern areas of the wreck noted as a 22

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 particular concern, against a more general background of a reduction in exposure elsewhere (Peacock, 2003: 2–3). The licensee’s work between 2003 and 2007 comprised continued diving activities and the ongoing observation of a cycle of natural degradation affecting the physical integrity of some areas of the Stirling Castle. A continuing increase in structural exposure along the port side from midships to stern was recorded. The sediment reduction resulted in the exposure of artefacts, which were at immediate risk of loss or damage from tidal movement (Peacock, 2003; 2004; 2005; 2006; 2007). While the annual licensee reports from this period are generally very similar in nature and relatively brief in terms of content related to detailed site work, a notable exception is the 2006 report, which details the extensive and specific loss of material from the rudder: around 2.5m from the top section of the rudder had broken off, along with two cross-beams that made up the stern of the ship (Peacock, 2006: 2–3). Peacock completed his 2006 report with the following conclusion:

The 2003 report also highlighted an inconsistency between the existing diver-recorded site plan and the site extent recorded during the 2002 ADU multibeam survey (WA, 2003a: 16). Addressing this inconsistency was noted by the licensee as one of the key objectives for the licensed team and WA as the archaeological contractor (Peacock, 2003: 2). WA’s initial visit to the site, which took place in 2003, was intended as a means to guide subsequent fieldwork planning by identifying the extent of surviving material at that point in time, as well as to confirm the position of the site and to establish a series of monitoring points across the site to inform future understanding and recording of sediment levels and/or structural collapse (WA, 2003b: 1). The overall understanding of the extent of the site and the ongoing sediment regime was further informed in 2005 when the site was geophysically surveyed as part of year one of the Rapid Archaeological Site Surveying and Evaluation (RASSE) project (Bates et al., 2005: 23-28). This study suggested that the site had been subject to sediment accretion around the stern of the vessel and to the north-east of the vessel since 2002 (Bates et al., 2005: 27), when the ADU had conducted a previous multibeam bathymetry survey across the site.

The continuing and rapid exposure of the Stirling Castle warrants a full review of the current site management strategies. Now that English Heritage regards the site as a tourist attraction, rather than a part of National Heritage, maybe it is time to de-designate the site, as the protection it has received, over the last 27 years has failed to protect the site from damage via fishing and the elements. It is vital that efforts are made to evolve both a long term and short term management plan if the full archaeological potential of this important wreck is to be fully realised. (Peacock, 2006: 3)

Further significant fieldwork was undertaken by WA during May and June 2006 in their role as the PWA contractor (see WA, 2006). Diving was conducted on site with the intention of combining the detailed multibeam survey undertaken in 2005 with direct diver observations to create a revised and updated site plan. As part of this process, many of the observable seabed features were rationalised with those historically reported and previously recorded on the site (see WA, 2006: 11–19). This work concluded that in the forward third of the site, where the most sediment accretion had occurred in recent years, the hull was likely to be coherently preserved from the keel up to the level of the gun deck (WA, 2006: 11). By contrast, the stern area of the site was considered to be highly unstable (WA, 2006: 11), concurring with the reports of the licensee in preceding years. In addition to the artefacts recorded in situ on the site – anchors, cannon and so on – a large number of barrel staves and components were also recovered during the course of the 2006 PWA fieldwork season (WA, 2006: 18) (see section 7.2.10 for further information on the barrel staves). Although the site had clearly been exposed to a significant degree in 1979 and 1998, WA were still able to note in their report (2006: 22) that ‘A very large late seventeenth century naval artefact assemblage is believed to exist in situ. Almost the entire content of the orlop deck and hold may be present. Preservation is likely to be excellent where not previously exposed’. The successful process of integrating multibeam survey with direct diver observation was continued during July and August 2007, with further work undertaken on the site by WA in relation to the PWA (see WA, 2007). During the course of this work it was also noted that sediment accretion was occurring on the site, particularly in the area of the stern, which had previously been subject to deep scour; this was observed to have been largely filled in by 2007 (WA, 2007: 5).

This matter was expanded upon further in his 2008 report, which recommended that the site could potentially be protected through covering in geo-textile, as he had suggested in 2001 (Peacock, 2008: 3). That year’s report also commented further on the continuing sediment accretion that was beginning to reduce the amount of exposed structure, with a transition towards sandy sediment with a lower shingle and mud content. However, as he also suggests (2008: 2), the apparent reduction in exposed structure could also be a reflection of the generally reduced height of the surviving hull structure as a result of the degradation observed since the re-emergence of the site in 1998. Throughout this period, running alongside the work of the site licensee, was the work of Wessex Archaeology as the contractor for the PWA. WA began their work with a full desk-based assessment of the Stirling Castle (WA, 2003a), with the intention of informing the development of a considered management plan for the site, including an on-going programme of monitoring and survey. The report’s recommendations were wide-ranging and expressed concern for the treatment of raised artefacts and the dispersal of the archive, as well as on-site, underwater considerations. With regard to the latter, the provision of a combined geophysical and diver-based investigation was recommended as a means to establish baseline data and to allow the ongoing site-formation processes to be fully understood. 23

Whitewright resulting in the loss of one of the transom timbers and probably the top of the rudder, confirming the reports of the licensee (WA, 2009a: 7-8). Shortly after this, a high resolution sidescan and magnetometer survey of the site was carried out in September and October 2008 by WA on behalf of English Heritage as part of the South East of England Designated Wrecks Project (WA, 2009b), adding to the geophysical survey record of the site.

It was also observed, as might be expected, that exposed timbers were subject to extensive damage from biological and physical mechanisms (WA, 2007: 5). The 2006/7 period of work undertaken by WA resulted in the development of a coherent and reliable site-plan, building on the 1999 survey project undertaken by the licensee but underpinned by repeated geophysical survey (see box text and chapter four). As a result, the extent of the site was fully documented, as was its relationship to surrounding wrecks. The nature of sediment movement, structural exposure and structural reburial also became better understood as a result of this period of work on the site.

Further work on the Stirling Castle was undertaken by WA through the PWA contract in June 2009 with the objective of continuing to inform and refine the overall site-plan (WA, 2010a). As in previous years, on-site conditions meant that several objectives, notably the use of sectorscanning sonar, had to be abandoned (WA, 2010a: 5). The overall condition survey undertaken in 2009 indicated that the stern-post had been subjected to further degradation and scour, but that the sides of the vessel towards the stern were now buried, as were the majority of features in the bow area. In general, exposed timbers were subject to extensive biological and mechanical erosion, but it was also observed that there was some indication of sedimentation occurring (WA, 2010a: 7). At this juncture in the history of the site it was noted by WA that the site itself is not stable, but that it is currently buried with little

Fieldwork in 2008 by WA as the PWA contractor aimed to undertake dendrochronological sampling in addition to further work on the integration of multibeam and diver observation across the site (WA, 2009a). The exposed timbers on the site were found to be too degraded to be usefully sampled for dendrochronological purposes and it was deemed undesirable to undertake intrusive excavation in order to obtain samples from buried timbers in a potentially better condition (WA, 2009a: 9). Ongoing condition monitoring of the site as part of this work observed that significant degradation had occurred at the stern of the vessel between August 2007 and August 2008,

Geophysical Survey undertaken by a variety of different organisations at different times. This has included the ADU and Wessex Archaeology (WA) as PWA contractors, as well as the site licensee, Marine Archaeological Surveys (MAS) in the 1980s (Redknap and Fleming, 1985; Redknap, 1990), the RASSE Project (Bates et al., 2011) and the UKHO, as well as surveys commissioned for the television series Wreck Detectives. As a result of all of this work, there is now a relatively well-established set of baseline data relating to the site that allow some of the longer-term processes that affect the archaeological remains to be better understood. The analysis of these combined surveys is presented and discussed in chapter four.

One of the recurring themes of the archaeological investigation of the Stirling Castle has been the continuing need to establish the baseline extent and condition of the site. This is a basic requirement of almost any maritime archaeological site and ideally such information can then be used to monitor the ongoing physical changes to the site on a year-by-year, or even month-by-month, basis. In the case of the Stirling Castle, such a survey was perhaps even more important because of the extremely dynamic environment that the wreck was located within and the sometimes extensive changes to its overall disposition that were being recorded with each season of work. The ADU had already noted the need for such a survey in 1992, but had also stated that a conventional survey was extremely difficult because of the on-site conditions.

The regular geophysical survey of the site has not meant an end to direct survey by archaeological divers. The multibeam bathymetry provides a coherent overview of the site, but the anomalies that such a record show still need to be investigated by archaeologists to confirm their identification. This two-fold approach formed the basis of the successful work undertaken by WA between 2006 and 2009, using acoustic systems to track and record both diver and artefact locations. WA noted at the end of the 2006 season that it was not possible to produce an improved site plan from multibeam data alone because the resolution was simply not sufficient to plot every object in detail (WA, 2006: 20). A combination of remote sensing and direct human observation therefore remains the most effective means to record and monitor the site.

One obvious solution to this problem lay in the application of marine geophysical survey techniques to the site, both as a means to gain an understanding of the broader cycles of exposure and burial on the evershifting Goodwin Sands and as a way to record the extent of the seabed structure itself. Such early surveys of the Stirling Castle were conducted with a singlebeam echo-sounder or sidescan sonar, before multibeam swath-bathymetry surveys were conducted from 2002 onwards. Alongside these have been regular sidescan sonar surveys as well as the use of magnetometers over the site. As with all of the other work described in this chapter, the geophysical surveys have been 24

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 potential for any non-intrusive fieldwork (WA, 2010a: 8). In 2009 the licensee reported that only around 30 per cent of the wreck structure was exposed and that the overall level of exposure had decreased (Peacock, 2009a: 3). Although the licensee’s official reports under the PWA are confidential, Peacock felt strongly enough about the management regime implemented on the Stirling Castle and on sites designated under the PWA in general to comment (Peacock, 2009b: 5) in public that, in regard to the Stirling Castle,

Secondly, and very much to the fore in any account of the site, is the undoubted and remarkable preservation of the ship itself and the people, along with their equipment and possessions, who once sailed aboard. Even now, several decades after the initial discovery of the vessel, and taking into account the documented degradation and destruction of ship and shipboard remains, it is worth reiterating how much of the archaeological record still remains in place. In that sense, a positive note can be taken from the assessment of WA’s dive team in 2006 (above) that a significant and well-preserved late seventeenth-century assemblage remained in situ.

the adoption of in situ preservation without protecting the site has meant it has been destroyed by the elements over the last 10 years.

The failure of such a significant assemblage to become fully accessible to the public is considered further in section 3.2. But the mere existence of the present project, undertaken as it is by a third party and tasked with the publication of the archive so many years after its creation, must be viewed as indicative of a chronic failing in the wider system. It is outside the scope of this publication to comment fully upon this, but, in writing with hindsight, it is clear that a systemic disregard for the realities of maritime heritage management and maritime archaeology in general accompanied the application of the PWA in the early 1980s (for contemporary comment see McGrail, 1989). While the formation of the ADU at least led to some continuity in the monitoring of sites it is equally clear that such sites as were designated, themselves representing only a small proportion of the total, were woefully catered for in terms of the overall resources allocated to them and the capacity of bodies such as the ADU to act in a proactive manner on a site-by-site basis.

If this is how we are to manage our protected sites then the wording should be changed from preservation (which it is not) to ‘staged and managed neglect’. By making a statement that we are not coping with maritime sites, a solution will have to be found. The remit of the present project was to undertake the assessment and analysis of the archive of the Stirling Castle up to and including the work undertaken in 2009. However, work on the site is ongoing, both by Bob Peacock as the site licensee and WA as the PWA contractor. Much of the work done by the latter has been undertaken through an overarching project addressing designated vessels in the wider south-east of England. The result has been further geophysical investigation and archaeological diving on a number of sites in the Goodwin Sands area, including the Stirling Castle. Information on this work is contained in a number of WA reports (2010b; 2010c; 2011; 2012). Work on the site has also been continued by the licensee and his team and in 2013 it was reported (Pascoe, 2013) that a number of framing scarph chocks had been identified and recovered from the sites of the Stirling Castle and the Northumberland (see also Pascoe and Peacock, 2015).

The extent to which the failings apparent at that time have been rectified since the 2002 National Heritage Act remain a matter for debate, one that is continued to an extent in section 3.3. Bob Peacock in his role as the site licensee has made clear his view (quoted above) that the notion of in situ preservation is not working as a means to preserve sites such as the Stirling Castle for the future. Set against this strongly held opinion must be the strides made in reaching an increasingly coherent understanding of the seabed remains since 2002, both at a site-specific scale and taking into account wider morphological changes.

3.1.4 Conclusion Offering a concise set of concluding thoughts to the work conducted on the Stirling Castle between 1979 and 2009 is difficult because of the sheer breadth of time during which activities, archaeological or otherwise, have taken place on the site. Despite this, a number of key factors that have influenced and shaped the rationale and outcome of work can be identified and are worth some further consideration here. Firstly, it is clear that, whatever the reasons for intervention on the site, which themselves have varied across the years, the defining factor has remained the environmental conditions in which on-site, underwater activity takes place. This has been uniformly hostile for all parties, whatever their motives. Strong currents, unreliable surface conditions, extremely mobile sediments and very poor visibility have all resulted in a loss of time on-site in every period of work and have certainly restricted the application of archaeological survey and recording methods, especially during the early seasons of work, when the archaeological experience of those involved was limited.

The remains of the Stirling Castle continue to be threatened by other factors, aside from the physical environment in which they are located and the problems of archaeological management tasked with overcoming that environment. Throughout the history of investigation of the site two further threats have been continually reported: fishing and unlicensed diving activity. While such factors have certainly had an effect on the ongoing formation processes of the site, their overall impact is extremely hard to quantify. The former has the destructive potential to displace large structural elements, while the latter has almost certainly led to the loss of archaeological information and physical artefacts at various points in the past. It is an unwelcome feature of the archaeological record that the loss of such information is permanent and, as such, it cannot play any 25

Whitewright part in our current or future interpretation of the ship, its crew or their loss in the Great Storm of 1703.

activity: dive logs, survey measurements and drawings, artefact records and photographs, geophysical data and the on-site photographic and video record. Ownership of that material rests with the organisation or individual responsible for creating or commissioning it. For example, material from the earliest phases of work is also held by and has indeed been published by some private individuals (e.g. Smith, 2010). Recent work has been undertaken by the SeaDive organisation through Bob Peacock as PWA licensee, who maintains a separate archive, in private ownership, of material relating to such work. Meanwhile, material produced by the PWA contractor since 2002 is owned by English Heritage, along with reports lodged by the site licensee, with more limited public access. In contrast, pre-2002 material that was produced by the ADU has been archived within the National Record of the Historic Environment at Swindon and is publicly accessible. In addition to this, a number of related historical sources such as Admiralty records or contemporary historical accounts of the Great Storm, can provide useful information on the ship and its loss and are retained in national institutions such as the National Maritime Museum or the National Archives.

The archaeological history of the Stirling Castle is almost as complex as the site itself. Yet, despite the challenges faced by the organisations that have worked on the site since 1979, the wreck has yielded a remarkable record of late seventeenth-century and early eighteenth-century material culture, both maritime and non-maritime. To understand that collection of material more fully in the context of the present project, our attention can now turn to the processes at play following the recovery of artefacts from the Goodwin Sands and in particular to the ownership and archiving of the archaeological record of the site. 3.2 Ownership, Archive Access and Dissemination The previous paragraphs set out some overall thoughts on the broader history of archaeological intervention that has taken place on the Stirling Castle from 1979 onwards. One obvious facet of that work not discussed above was the multiplicity of people and organisations that have worked on the site, with material from the site, or in the management of the site at different points in time. This has included archaeologists (professional and avocational), sports divers (legal and illegal), fishermen, historians, heritage managers, television crews and civil servants, all of whom represent a broad range of interest groups and an equally broad range of organisations. This section therefore address some critical issues arising relating to the ownership of the site and the site archive, including the artefacts present within it. It also sets out to explain the extent of subsequent archive dispersal that has in part led to the instigation of the present project and allows an insight into the subsequent limitations in the analysis of the Stirling Castle archive itself that inevitably stem from such dispersal.

The various archive-holders associated with the site have different responsibilities depending on their affiliation, mandate, stated aims and objectives and so on. These responsibilities range from a lack of an obligation to provide public access or even to publish material right through to a clear rationale to facilitate access and dissemination in the widest possible way because of the publicly funded status of the work. It is therefore unsurprising that over the course of 40 years the archive of material from the Stirling Castle, retained in such a broad cross section of ownership, has been curated in a variety of different ways and in different places. This has had an equally predictable outcome on the access to the archive that is currently enjoyed by the public, which can now be discussed.

3.2.1 Ownership

3.2.2 Archive Access

The ownership of the site, the artefacts raised from the site and the archaeological records resulting from work on the site is relatively straightforward to address. As noted above, the ship and its equipment is owned by the Isle of Thanet Archaeological Society (IOTAS) as a result of its purchase from the Ministry of Defence in 1980 for £200. The personal items of the crew that resided within the vessel were not included in that sale, but were purchased subsequently by IOTAS in 1982. All of the archaeological material that has been raised from the site or is still present on the seabed, including structural elements, ordnance and personal possessions, is therefore owned by IOTAS (see box text).

The varied ownership of the Stirling Castle archive has resulted in the archive itself becoming geographically dispersed over time. The majority of the artefacts raised from the site, especially from earlier periods of work, are held on loan by Ramsgate Maritime Museum, the Shipwreck Museum in Hastings and the National Maritime Museum, Greenwich. The Rupertino gun is held by the National Museum of the Royal Navy (Margaret Symonds, pers. Comm.3). A significant amount of archival material is held in the NRHE in Swindon, by SeaDive in Kent or by WA on behalf of English Heritage in Salisbury. Related historical documentation is primarily held in London, but across three different national institutions.

The sole ownership of the site and the material from it offers a marked contrast to the ownership of the archive produced as a result of the archaeological work undertaken on the site. That archive does not consist of physical artefacts, but comprises the material created during archaeological

Margaret Symonds is the Chair of the Isle of Thanet Archaeological Society and reported on the location of IOTAS-owned material in July 2014. 3 

26

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Clearly, this situation is not conducive to allowing effective and full public access to the archive in a manner that might be considered appropriate for an archaeological site of such demonstrably national importance. A powerful element of any future management of the site and its resulting archive must therefore be the identification of a way to allow public access to the entire archive, either in a physical or digital (virtual) form. Such access should allow members of the public to experience the material from the site for the purpose of simple interest, material from the site to be effectively used for outreach and education purposes (see also dissemination below) and finally elements of the archive to be used in future academic research. Thus, the issues of archive access and dissemination are closely bound together and it is to the latter that our attention can now turn.

This project has involved an audit of archive material, resulting in a database of information on the site archive. This includes record or object type, its location, and information on form and/or content. The artefacts from the site have been provided with a unique identifier that is used as the primary reference number in this volume (while any other reference numbers used previously are also recorded). This information has been deposited with the Archaeology Data Service and can be used to determine where individual records or artefacts are held by those wishing to consult the archive directly. In addition, it should be remembered that a significant corpus of material remains in situ on the seabed. As long ago as 1989 the dispersal and splitting up of site archives was noted as being a problem for maritime archaeology within Britain (McGrail, 1989: 10), a situation that has been reiterated recently (HWTMA, 2009b; 2009c; 2009d; Satchell 2016). The Stirling Castle archive represents a clear example of this from the date of its discovery to the present day.

3.2.3 Archive Dissemination The challenge surrounding the dissemination of the Stirling Castle archive stems to a large extent from the challenges just discussed relating to ownership, dispersal and access. In particular, dissemination of material for public engagement and for the purposes of education and outreach is extremely difficult without full access to the material. At least part of the value of the site and the material from it lies in the high levels of preservation associated with it and the range of well-preserved material that has been raised over the years. Existing maritime museums, such as that provided for the remains of the Mary Rose, have illustrated the high level of public interest in archaeological material comprised of artefacts representing the everyday life and work of people on board such sailing vessels. While material is on display at museums in Ramsgate and Hastings, there has been little or nothing in the way of current initiatives for wider public engagement with the site through digital sources4 or the provision of educational resources as has been achieved on other sites off southern England. Although the forthcoming online diver trail for the site will help to address this situation.

The Stirling Castle and archaeology in the Isle of Thanet At the time of its discovery in 1979, work on the Stirling Castle was undertaken in association with the marine archaeology section of the local Thanet Archaeological Unit (TAU). In 1988 TAU was renamed the Isle of Thanet Archaeological Society (IOTAS) to cater for the needs of enthusiastic amateur archaeologists. This resulted from the establishment of the professional and commercial organisation the Trust for Thanet Archaeology (TfTA). The first director of the TfTA was Dave Perkins, who had been closely involved in the early seasons of work on the Stirling Castle and who continued to deal with the site on behalf of IOTAS. The TfTA has never expressed an interest in any title to the Stirling Castle, which has remained entirely with IOTAS following its purchase of the site in December 1980 (Margaret Symonds, pers. comm.).

Dissemination to the academic community, a much smaller audience than the general public, has thus far been somewhat limited and piecemeal in nature. Published work has reflected the research interests of individuals, addressing subjects such as the statistical distribution of musket shot size (Perkins, 1983), the potential steering mechanism of the vessel (Endsor, 2004), elements of the vessel’s navigational equipment (Smith, 2010) and the application of geophysical methods to the recording of the site (e.g. Bates et al., 2011). In addition to this, material from the site has been included within wider historical work (Endsor, 2009) covering one of the Stirling Castle’s sisterships, the Lenox, as a means of providing artefact-based context to the historical narrative of the latter vessel. Finally, summary accounts of the site have appeared in general works (e.g. Fenwick and Gale, 1998), while some aspects of the current work have been

In 1982 TAU ownership of the shipwreck was reallocated in two blocks of 32 shares to Len Jay and John Cayzer, possibly because it was believed that a charitable trust such as TAU could not own a shipwreck. These shares were eventually returned to IOTAS after 1999, following the deaths of Len Jay and John Cayzer. Many of the artefacts from the early seasons of work were loaned to Ramsgate Maritime Museum, where they remain on public display. A further 184 artefacts that had been held in storage by TfTA were transferred to IOTAS and then loaned for display to the Shipwreck Museum in Hastings. The display of artefact material in these two museums provides an important element of ongoing public access to the archaeological remains of the Stirling Castle.

For online access to the digitised archive of HMS Invincible (lost 1758) see: https://www.maritimearchaeologytrust.org/mapguide/maps/ invincible/main.php 4 

27

Whitewright published in the newsletter of the NAS (e.g. Peacock, 2000a; 2000b; Pascoe, 2013).

Although the manner of work undertaken in 1979 would be deemed largely unacceptable by today’s archaeological standards, it should be remembered, and indeed recognised for the purpose of understanding the site, that those undertaking the work were not professional archaeologists and had little or no archaeological training. Their fears about the degradation of artefacts were well founded and borne out through the experience of subsequent archaeological teams in the late 1990s and 2000s. Likewise, there appears to have been a very real danger of interference on-site from more mainstream salvors and regular sports divers. It should also be realised that the ACHWS were largely aware of the type of work being undertaken on the site, in terms of the day-to-day methodology of the operation, but continued to license work to be undertaken; presumably it was considered better to have someone ‘working’ the site with material going to a local museum than nobody at all, with the related increase in the likelihood of even purer salvage work taking place.

From this summary alone it is clear that the academic publication of material has not been developed in a coherent manner, following a wider research and dissemination plan. This problem is partially mitigated by the analysis stage of the present project, which has resulted, in turn, in the publication of this monograph, incorporating as it does a number of pieces of specialist research based on the artefact assemblage and the accompanying archive (chapters six– eleven), along with the accompanying discussion. In this sense, the challenge of dissemination is one that should really focus more on the wider public, rather than the academic community. The site of the Stirling Castle and artefacts from it, representing life at sea over 300 years ago, have obvious potential to engage and educate a wide cross section of current society about their forebears. It is crucial that the challenge of archive access highlighted above is overcome in order to ensure both the dissemination of the archive to the public and future research to continue the process of academic dissemination as new material is discovered, recorded and raised from the site, as it surely will be. Such academic investigation, dissemination and discussion should in turn serve to update and refresh dissemination of the site to the wider public.

Some context is provided here by the discussion surrounding the proposed designation of the Brighton Marina site in the minutes of the 1983 ACHWS meeting. One view put forward was that the Brighton Marina site could be designated if the dive club that had found the site and applied for designation would ‘be prepared to work the site if it was protected’. In the days before the ADU was established it was considered necessary only to offer the otherwise archaeologically untrained divers some ‘archaeological direction’. It is difficult to imagine a similar approach being accepted if members of the general public with no archaeological experience, training and support facilities had applied to excavate a scheduled ancient monument in a terrestrial management context in the early 1980s. Only from the mid-1980s were professional diving archaeologists appointed, in the form of the ADU, to offer a more consistent approach to the monitoring of the historic wrecks that had been designated under the PWA.

3.3 The Stirling Castle in the Context of Maritime Archaeology in England and the UK Having presented an account of the archaeological investigation of the Stirling Castle based on the historical records left by those involved it is useful to consider the work done on the site against wider developments within British and English maritime archaeology across the same period. This will both provide some further context to some of the methodological and management approaches taken with regard to the site over the years and allow the development of a more considered opinion of the nature of the archaeological record produced through that work.

Yet, at the same time as material was being raised from the Stirling Castle in a largely salvage-based fashion, further along the south coast of England the work continuing on the site of the Mary Rose (see Marsden, 2003) in the eastern Solent applied underwater archaeological methods still considered to be of the highest standard. Other work around the coasts of Britain had also demonstrated the possibility of applying fully archaeological methods to the investigation of shipwreck sites in a range of conditions. Work on the sites of the Dartmouth (Adnams, 1974; Martin, 1978) and Kennemerland (Muckelroy, 1978: 196-214) spring readily to mind, with the latter in particular demonstrating the importance of, and analytical possibilities to be gained from, accurate underwater spatial recording. Perhaps the very fact that the majority of the material raised from the Stirling Castle found its way to the local museum to be catalogued and recorded, rather than being sold at auction, is testament to the impact that other projects around the country were having at that time. In retrospect, coming only some 12 years after the

3.3.1 Methodology and Management It is clear from the narrative of the history of intervention on the Stirling Castle outlined in section 3.1 that some extremely contrasting methodologies have been applied at different times to the site. Today, much of the work conducted in the 1979 season would be classified simply as salvage activity, with little in the way of meaningful archaeological data generated beyond the artefact collection. The initial visits by the ADU in the mid/late 1980s rapidly classified the early site-plans as highly misleading in terms of the extent of locational accuracy that they imply for the recovered artefacts (ADU, 1986: 5), a fact largely acknowledged by Lyon in his published report (1980: 340) on the work undertaken in 1979. This has obvious implications for the interpretation of any assumed spatial context or inter-relationship of those artefacts, and is addressed in chapter five. 28

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 systematic looting of the site of HMS Association, a direct contemporary of the Stirling Castle, in the Isles of Scilly, the 1979 season of work on the Stirling Castle represented progress in the treatment of underwater cultural heritage once it had left the seabed. In that sense, the archaeological archive from the 1979 season is very much a product ‘of its time’, and by extension is as much a record of maritime archaeology in Britain c.1979 as it is a record of a British warship c.1703.

financed project, were again evident in the raising of a World War II Dornier 17 aircraft in June 2013. The story of the archaeological work on the Stirling Castle remains a strong reminder of the sheer uniqueness of the Mary Rose project within British maritime archaeology and the difficulty encountered since in replicating the process for other sites, even those in relatively benign terrestrial contexts, such as the Newport Medieval Ship which at the time of writing has been conserved and researched (e.g. Nayling and Jones, 2014) but still lacks a publicly accessible destination. The success of the Mary Rose project, based on a system of mixed funding, allows attention to be focused on the comparable resources, or lack of them, allocated to sites such as the Stirling Castle, along with the subsequent public access and dissemination resulting from such resources.

Outwardly, little has changed in the overall approach taken to the management of sites designated under the PWA since the mid-1980s when the ADU was established. Licences can still be applied for by members of the public and, while many PWA site licensees are professionally trained archaeologists, many are not. As such, the system can be said to readily engage with the significant and important avocational archaeological sector within the country. The range of current licensees therefore represents the full sweep of opinion on how underwater cultural heritage should be managed, including the view that work on archaeological sites should be funded through the sale of underwater cultural heritage raised from such sites (Oldham, 2013:4). It is a viewpoint seemingly at odds with the modern heritage management objectives that are applied by bodies such as EH, but which to some extent harks back to the 1983 minutes of the ACHWS, noted above. A vivid contrast to such a view is provided by other current licensees who fully subscribe to the view set out in the UNESCO convention that ‘the commercial sale of archaeological material is fundamentally incompatible with the protection and proper management of underwater cultural heritage’ (Parham, 2010: 3). At the very least, it can be confidently stated that the range of opinion offered by PWA licensees is extremely broad and encompasses the full sweep of the archaeological spectrum. Meanwhile, monitoring of sites for management purposes is still carried out by a professional archaeological contractor and this is now, quite appropriately since the National Heritage Act 2002, overseen by national heritage agencies such as English Heritage, rather than by the Department for Transport.

3.3.2 Resources, Public Access and Dissemination 3.3.2.1 Resources The fact that the same overall system, with the addition of site monitoring by the PWA contractor (formerly the ADU), is still in place today says much about the potential inadequacies of the current system for managing marine archaeological sites in England. This was highlighted by Firth in 1999 in a review of the overall impact of the PWA on the creation of a maritime archaeological resource within the UK which noted (1999: 22) that the PWA is ‘a basically sound and beneficial statute which could provide a useful service for many years to come’. It is the application of the PWA to a small number of sites of limited temporal and spatial range that are the real shortcomings of the Act, rather than the protection that is offered – shortcomings that stemmed from the political climate in which the PWA was originally initiated in the 1970s (Firth, 1999: 22). While the PWA’s mechanics are relatively clear cut, it is equally apparent that the allocation of resources to support its actual application have grown in an organic way. It took over a decade for a professional archaeological diving unit to be established to independently monitor the sites listed through the PWA, and there is still no nationally funded facility to allow the conservation of all material raised from historic wreck sites. If such a facility had existed in 1979 and had been accessible to teams such as those working on the Stirling Castle, then perhaps the dive team would have been less reticent about recovering organic artefacts.

Seen from the distance of over 30 years, it initially seems strange that the archaeological remains of the Stirling Castle did not become a Mary Rose for the late sixteenth/early seventeenth century. Certainly, the level of archaeological preservation apparent on the site would have lent itself to a similar approach to the vessel’s crew and their equipment, even if raising the hull of the ship itself remained unfeasible. Despite the limitations of some areas of UK maritime archaeology at that time, outlined above, the work on the Mary Rose and other sites certainly demonstrates that such an outcome was possible. That the Stirling Castle was not subject to systematic archaeological recovery at the time, to be housed in a purpose-built museum and visited by hundreds of thousands of people, is perhaps simply a result of the extreme logistical difficulties of working on the site. The problems surrounding work on the Goodwin Sands, even with a well-equipped and

In more general terms, the absence of resources – financial, technical, physical or otherwise – allocated to the Stirling Castle was a constant source of comment in the reports on the site lodged by the ADU (e.g. 1998: 1; 2000: 3). The report lodged in 2000 in particular looked forward to the assumption of responsibility for the site by English Heritage and the assumed increase in resources that would be directed towards it as a result. It is clear from the history of the site set out above, that since 2002 there has 29

Whitewright been a clear upturn in the overall management of the site and in the extent of the archaeological work that has been conducted on the Stirling Castle in particular, as well as across the wider area of the Goodwin Sands.

lack of resources has already been discussed, to which can be added a widespread failure to fully facilitate public access and to disseminate the results of archaeological work, on those sites deemed worthy of designation under the PWA. It is no coincidence that the desk-based assessment of the Stirling Castle completed by WA in their newly appointed role as PWA contractor recommended that site licensees should have an obligation to publish the results of their work and that public understanding and appreciation of the site should be enhanced (WA, 2003a: i-ii). As before, the Stirling Castle can be taken as a proxy for wider issues within the discipline.

But maritime archaeology within England is clearly under-resourced from central and local government, given the relative numbers of sites protected and the obvious importance of those sites (given the apparent and often stated centrality of maritime affairs to our island nation). The allocation of resources for the survey, recording, monitoring and management of historic wreck sites, designated or not, around the coasts of England (and the UK) is still some way short of what might be considered reasonable and work is still heavily reliant on the time given freely by the avocational sector. The effective engagement of that sector is critical if the discipline is to remain inclusive in how it manages its designated sites, but it is not unreasonable for those volunteers to be fully resourced, trained and compensated for their time in a proactive, rather than a reactive, way. The strides made in the provision of central funding to directly support licensees on sites such as the Swash Channel, HMS Invincible and London are to be roundly welcomed and clearly demonstrate the quality of archaeological work that can be carried out if the situation allows. In that regard the situation is arguably much better than it has been in the past, although such funding is still largely limited to licensees from the professional side of the archaeological spectrum.

Tackling such problems has not, and cannot, happen overnight, a fact well illustrated by the time taken by the present volume to reach publication after its initial project commissioning in 2009. PWA licensees still have no statutory obligation to publish the results of their work, which, it should be remembered, takes place on sites protected in the national interest. But a greater number of England’s maritime archaeological sites, both designated and undesignated, are now becoming increasingly disseminated, including as a direct result of funding received from English Heritage and other public and private grant-making sources (e.g. Parham et al., 2013; Satchell and Whitewright, 2014). Publication of work that seeks to address maritime archaeology in a broad sense and to provide an agenda for future research has also taken place (see Ransley et al, 2013). Alongside this, the importance of public engagement and education as a means of dissemination is also becoming increasingly acknowledged and reflected. The situation is therefore not perfect, but it is immeasurably better than it has been in the past. The Stirling Castle can therefore still be seen as representative of the problems and challenges that have characterised maritime archaeology within England and the wider UK since the inception of the PWA in 1973, but on a more positive note it can also be viewed as symbolising the efforts of a generation of maritime archaeologists to address and solve some of the problems that they have inherited, while attempting to avoid the creation of similar circumstances in the future.

3.3.2.2 Access and Dissemination With regard to access and dissemination, the Stirling Castle can be seen to symbolise the wider challenges faced within the field. These range through the dispersal of archives, multiple ownership built up over long periods of time and even the sale/loss of artefacts outside the archaeological system. Some of these challenges – for example, that of ownership – are inevitable results of the wide range of people, organisations and so on, that become involved in such sites, especially over longer time periods. At a more practical level, such diversity of ownership and the related dispersal of material have inevitably resulted in considerable variation in how material is managed by its various owners and/or holders and the extent to which archive material has been formally deposited. This problem was noted in the Stage 2 Assessment report (HWTMA, 2012: 80) in relation to the currently accepted best practice for archaeological archives (see Brown, 2007). In this regard the status of the Stirling Castle archive perhaps epitomises the general crisis currently facing maritime archaeological archives within England (for further discussion see Ransley and Satchell, 2014) – a crisis that has been addressed by separate English Heritage initiatives for both designated and undesignated sites. The last point is worthy of further discussion with regard to maritime archaeology in England and relates to the situation that English Heritage inherited when they assumed responsibility for England’s territorial waters in 2002. The 30

4 Site Environment and Geophysical Survey Amelia Couldrey It is clear from the account of the archaeological work on the Stirling Castle provided in chapter three that the long-term management and wider understanding of the site is highly dependent on a coherent appreciation of wider environmental processes. In particular, phases of sediment loss and/or accretion have impacted dramatically on the exposure and/or burial of archaeological remains. Such processes have a direct relationship to the loss of archaeological material from the site, ease of access to the site itself, the fragility of archaeological material and the extent to which such material is considered vulnerable or under threat. It will be equally clear from chapter three, that for much of the period during which the site has been under archaeological investigation such an understanding has been quite lacking – a point continually raised by the archaeological contractors who have reported on the site and commented on the need for a baseline survey to be completed (e.g. ADU, 1992: 2; 1995: 1; WA, 2003a: 31). Only recently, have concerted attempts through geophysical survey been able to begin the process of recording comparable information across a range of complementary datasets from which an informed position regarding the geoarchaeological nature of the site can be established.

Figure 4.1. Location map for Stirling Castle and Goodwin Sands bank features and location of metocean data recording sites. Overlain on 2009 MCA contracted bathymetry of the Goodwin Sands bank (contains public sector information, licensed under the Open Government Licence v3.0, from the Maritime and Coastguard Agency).

With this in mind, attention in this chapter turns to the archive of material generated from work of a geoarchaeological nature on the Stirling Castle. Here are brought together a range of datasets, including single and multibeam bathymetry, sidescan sonar, magnetometer survey and sediment analysis. These are in turn informed and supported by wider existing studies of the Goodwin Sands and surveys undertaken for non-archaeological purposes, such as by the Maritime and Coastguard Agency (MCA). The results of such work are first contextualised with regard to the geology, the underlying hydrodynamic and sediment regimes, and the historical morphological evolution. Following this, the results of the identified, collated surveys are presented before analysis is offered regarding the recent evolution of the site itself and the implications of this for understanding its possible future pathway.

coast (Figure 4.1), 15km from the town of Deal, within an unnamed flow dominant channel that partially incises the North Sand. Presently (based on 2009 multibeam data) the wreck sits at a depth of 15m below Chart Datum (CD) 17.9m below Ordnance Datum Newlyn (ODN), using datum separation values taken from the Vertical Offshore Reference Framework (VORF). The Goodwin Sands is a complex sandbank system, incised by both flood and ebb tidal channels the long-term (decadal and greater) evolution of which is not fully understood, although many have laboured to produce forecasts (e.g. Cloet, 1954; Elderfield, 2001). 4.1.1 Geology The Goodwin Sands, which formed during the Flandrian Transgression (D’Olier, 1981: 222), consist of approximately 20–25m of unconsolidated sandy sediments resting on an Upper Chalk platform (British Geological Survey, Thames Estuary Sheet 51°N-00°, 1:250 000 Series). The bottom types are fine sand with gravelly sand forming the base of the gully in which the Stirling Castle rests.

4.1 The Natural Environment The modern Goodwin Sands bank can be divided into several named areas (displayed in Figure 4.1). The bank is presently split into two main sections, the North Sand, or Goodwin Knoll, and the South Sand, or South Calliper; these two areas are separated by the Kellett Gut. The wreck of the Stirling Castle is located 13km offshore of the Kent

Sediment samples from a 200m radius around the wreck were taken in 2006 (Bates et al., 2007: 78, 81–82). Median 31

Couldrey

Figure 4.2. Velocity direction and magnitude for a) neap and b) spring tidal diamond H (051°16’.3N, 001°27’.6E, 3.4km west of Stirling Castle site). Note there is a difference in the velocity scale between the two plots.

significant wave heights of 0.6m and average winter heights of 0.8m.

grain size was found to vary across the site from 1794μm (very coarse sand) in the west to 263μm (medium sand) to the east of the wreck, with a mean grain size of 1265 μm.

Storm events are defined using the Peaks-over-Threshold method. For the Goodwin Sands wave buoy (51° 14.996’ N, 01° 28.994’ E, 3km south-west of the site) 0.5 per cent of values annually exceed 2.2m. During the winter storms of 2013/14 peak significant wave heights of 3.69m were observed at the Goodwin Sands buoy, equivalent to a >1 in 30-year storm (Channel Coastal Observatory, 2014). From October 2013 to February 2014 seven storms of greater than 1 in 1 year return period (2.8m) were observed.

Dredging processes in the immediate vicinity of the wreck, within the North Sand Head zone, are thought to be likely to affect local sediment distribution (WA, 2003a: 27). Five licences were issued from 1976 and 1998 for the extraction of aggregate from the North Goodwin and South Goodwin areas, permitting over 9.5 million tonnes (6.3 million m3) of material to be extracted (Natural England, 2012).

The maximum tidal range is 5.3m at Deal. Spring flood currents reach maximum velocities of approximately 1.6m/s in a north-north-easterly direction and 1.4m/s during the ebb tide, in a south-south-westerly direction (Figure 4.2b) (Admiralty chart 1828, tidal diamond H, location shown in Figure 4.1). Neap tidal flows are on average half the strength of the equivalent spring tidal flow (Figure 4.2a). Unfortunately the nearest current measurement point to the site, for the purpose of recent in situ measurement, is located at a position more than 10km south of the site. Owing to the spatial heterogeneity of the flows around the bank these data are unlikely to be representative of the flow expected at the wreck site and are therefore not included here.

Waves with significant wave heights of approximately a tenth of the water depth produce an oscillatory velocity at the seabed (Soulsby, 1997: 72). For the Stirling Castle site, at low tide, this equates to waves of 1.5m. On average, between 2008 and 2014, records from the Goodwin Sands wave buoy indicated that significant wave heights exceeding that 1.5m threshold equated to 5 per cent. Put another way, it can be predicted that for 5 per cent of the time wave action is likely to have an impact on sediment transport at the site. Although severe individual storm events may impact the site, the 5 per cent incidence is relatively low. This is broadly supported by Manders (2009: 53), who observed that during model runs of the effects of a storm, storm waves did not significantly affect the pattern of residual sediment transport below the 10m depth contour.

Waves approach predominantly from the south and to a much lesser extent the north-east (Figure 4.3a). A maximum significant wave height (for the period of 2008 to 2014) of 3.69m and a mean significant wave height of 0.68m were recorded (Figure 4.4a). As expected, there is a large seasonality in the wave record, with average summer

As the Goodwin Sands wave buoy was deployed only in mid-2008 and we have bathymetric surveys dating back to 2002, data from the Hastings wave buoy (located 80km southwest of the site (50°44’.79N, 0°45’.30E)) are also presented here (Figures 4.3b–g and 4.4b). Owing to its offshore position significant wave heights at this buoy are

4.1.2 Hydrodynamic Regime and Sediment Transport

32

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 4.3. Wave directional roses for a) Goodwin Sands wave directional waverider buoy for the period of 2008–2014 and b–g) for Hastings wavenet buoy for the periods of b) 2002–2012, c) November 2002–April 2005, d) April 2005–September 2005, e) September 2005–March 2006, f) March 2006–August 2006 and g) August 2006–September 2009 (contains public sector information, licensed under the Open Government Licence v3.0 from the Coastal Channel Observatory).

33

Couldrey

Figure 4.4. Time series of significant wave height for a) Goodwin Sands waverider buoy and b) Hastings waverider buoy. Values over 0.5% exceedance in red and over 0.05% in black. (Contains public sector information, licensed under the Open Government Licence v3.0 from the Coastal Channel Observatory).

on average 1.5 times larger than observed at the Goodwin Sands site. Waves at the Hastings buoy also approach from a more south-westerly direction than at the Goodwin Sands site. Throughout the multibeam survey period (2002–09) wave directionality remained constant; however, from April 2005 to September 2005 and to a lesser degree from March 2006 to August 2006 the significant wave height was reduced because both surveys were performed during the spring/summer months. The presence of bedforms in the multibeam bathymetry data provides evidence of the local bedload transport and can be used to give an indication of the local seabed conditions. Two main types of bedform are observed in the region of the Stirling Castle: very large subaqueous dunes and medium subaqueous dunes, which surmount the former.

Figure 4.5. Transect location overlaid on multibeam bathymetry hillshade from March 2006 (based on data collected by the RASSE project).

By extracting depths along a transect 30m east of the wreck (Transect A, location shown in Figure 4.5 and transects in Figure 4.6) characteristic properties of the bedforms can be quantified. Wavelengths of 4–12m and wave heights of 0.2–0.8m are observed, giving a ripple index (wavelength/ height) of between 15 and 20. These bedforms are often strongly asymmetrical, with symmetry indices (stoss length/lee length) of up to 3.5. Bedforms are proximally flow aligned, with steeper faces angled towards the northnorth-east. Under Ashley’s (1990) classification system these bedforms fit predominantly within the medium

subaqueous dune category and are predicted to migrate at a rate of 100s of metres per year. While there are localised areas of medium subaqueous dunes, the tidal sand ridges that make up the Goodwin Sands are predominantly surmounted by very large subaqueous dunes (Ashley, 1990: 169) with heights in excess of 4m and wavelengths of approximately 250m (Figures 4.7 and 4.8). These bedforms are also asymmetric, with their orientation indicating a proximally north-easterly 34

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 4.6. Transect through bedforms 30m to the east of the wreck. Slope removed via polynomial regression fit. Bedform peaks marked with red asterisks and troughs with green asterisks.

Figure 4.7. Location of sandwave transect (B to B’) overlaid on 2009 bathymetry (contains public sector information, licensed under the Open Government Licence v3.0, from the Maritime and Coastguard Agency).

Figure 4.8. Transect of sandwaves 100m west of the wreck site of B to B’.

35

Couldrey Sands follows a seven-year pendulum swing from east to west and back again (Bathurst, 2005: 31). The instability of the bank has been described from the eighteenth century through to the twentieth century (see Cloet, 1954) and for the period from 1887 to 2000 (see Elderfield, 2001). These two accounts are brought together in the following section, with the addition of MCA data from 1997 and 2009. This, in conjunction with reference to historical charts, allows the multidecadal and even centurial evolution of the Goodwin Sands bank to be observed (Figure 4.10) and the history of exposure/burial of the site of the Stirling Castle to be followed. The presence of the Kellett Gut, swatchway or some more easily navigable channel through the bank has been documented since 1584 (Cloet, 1954: 208). In more recent histories of the Goodwin Sands the Kellett Gut was well established in 1795 and continued to exist until 1850 (though it was not formally named until 1926); by 1865 there was no trace of the channel and charts show that it was not present in 1887 either (Figure 4.10, first panel). The earliest chart presented here dates from 1887 and shows the Goodwin Sands as a single cohesive bank system with a single flow channel incised into the bank from the south-west to the north-east, creating a ‘calliper’ morphology. In such a system as recorded in 1887, the wreck of the Stirling Castle would have been fully buried, as water depths at the site were 0 to -4m CD (compared to -15m CD in 2009). Taking the bedrock layer to be -20m CD, this equates to between 16 and 20m of sediment.

Figure 4.9. Bed-level change from April 2005 to August 2006 along transect B.

transport direction. Using Knaapen’s (2005) migration rate equation and the bedform parameters extracted from the transects gives an estimated migration rate of 18m/yr. However, this value is likely to be a large underestimation of the true migration rate. By observing the crest and trough movement between April 2005 and August 2006 (16-months interval) (Figure 4.9) a rate closer to 60m/yr is found. Were the bedforms profiles symmetrical and the bedforms to pass directly over the wreck it is predicted that the wreck site would be buried under the bedform peak for a period of approximately two years.

The bank split into two more discrete sections at some point between 1887 and 1965, with the northern section remaining fairly stable in position and the southern section migrating south-east by approximately 10km. The tip of the bank, as defined by the 10m depth contour, extended a further 6km south. This is the first historical chart in which we see the presence of the channel in which the wreck is now located; however, its penetration into the bank was probably not sufficient to have exposed the wreck at this time. The water depth in 1965 was between -4 and -8m below CD, equating to between 12 and 16m of unconsolidated sediment.

4.1.3 Goodwin Sands Historical (pre-1990s) Evolution Having considered the overall location and hydrodynamic/ sediment regime, attention can now turn to assessing the morphology and historical evolution of the Goodwin Sands. At present the Goodwin Sands bank is approximately 24km in length and 10km in width and has a total area of approximately 220km2 (using the 2009 10m CD depth contour). At its deepest the bank reaches 40m below CD, although the average depth of the chalk bedrock is around 20m below CD (Carrizales, 2010: 11) and at its shallowest the bank sits 1m above CD: that is it is surface piercing at certain states of the tide. Above the 20m depth surface the bank has a volume of approximately 1km3 (estimated using 2009 Maritime and Coastguard Agency (MCA) data).

By 1986 the depth at the site had dropped to between -8 and -12m CD (12–8m of unconsolidated sediment coverage). From 1986 until 2000 the southern section appears to have remained in a much more stable position, while the northern of the two sections underwent a partial split to form a near-discrete central bank in addition to those to north and south. The depth of the channel in which the wreck is situated has also remained fairly constant over this period of time. Another way to observe the evolution of the bank is through the change in dimensions – that is axis length and orientation of the bank (Figure 4.11). These variables highlight the consistent anticlockwise rotation of the Goodwin Sands bank (Figure 4.12). The tides operate in a clockwise rotation, yet the Goodwins have been rotating

The sinuous shape of the Goodwin Sands bank is consistent with a site of bedload convergence: a location where there is no net bedload transport (Kenyon and Cooper, 2005: 22). Convergent-type sand banks are inherently unstable. Anecdotal evidence suggests the evolution of the Goodwin 36

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 4.10. Historical chart contours for the Goodwin Sands. ‘X’ shows the approximate location of the wreck of the Stirling Castle (image courtesy of Nick Elderfield (Elderfield 2001: 42)).

37

Couldrey 4.2  Geophysical Survey Sources of geophysical data include the MCA, ADU, the RASSE Project and WA. Where possible the original data were collated. Failing this, the processed gridded data were obtained. In a few cases, where even the preprocessed data were not obtainable, images of data were used and where possible these images were geo-rectified so that they were comparable to the other data. Included in this section are all known geophysical surveys of the wreck and surrounding area and interpretations of these data group by survey type. 4.2.1 Singlebeam Bathymetry Singlebeam bathymetry surveys were undertaken on the site between 1997 and 2000 (Table 4.1). Some datasets from these surveys were available for consultation. On the assumption that the position of the vessel superstructure has not changed from 1997 to 2002, the 1997 surface was shifted 30m east and 10m north to align with the 2002 surface. The 1998 surface was also shifted 23m east and 30m north. No shift was required for the 1999 data. The vertical datum for the 1997, 1998 and 1999 surveys are unknown, and without this information only a qualitative description can be extracted from these data. Singlebeam bathymetry data for the entire Goodwin Sands bank collected by the MCA were available via the UKHO Inspire portal and were collated for this analysis. Although these data were used within section 4.1.3, their resolution is too low to observe the wreck site morphology and so this survey’s interpretation is excluded from this section.

Figure 4.11. Goodwin Sand bank bathymetry from the 2009 MCA survey with annotations of bank morphological dimensions, as defined by Cloet (1954: fig. 5) (contains public sector information, licensed under the Open Government Licence v3.0, from the Maritime and Coastguard Agency).

4.2.1.1 1997

in an anticlockwise direction (Cloet, 1954: 204) with the main axis migrating from 38 degrees in the late 1700s towards 10 degrees in the early 2000s (UKHO, 2010: 13, fig. 10.4). Up until 2009 the bank system has continued the trend towards a north-south alignment. However, as discussed by the UKHO (2010: 13), this is largely due to the eastward migration of South Sand Head, while the rest of the bank system appears to be rotating in a clockwise direction away from the north-south alignment. Consequently the western edge of the general wreck site survey area has undergone a loss of sediment on a scale of 10–20m from 1997 to 2009, whereas the eastern edge has undergone a period of accumulation, with a net gain of between 3m and 10m (Figure 4.13).

In this survey (Figure 4.14) the wreck appears as a singular cohesive mound, elongated along the east-west axis. The mound is approximately 50m wide (north-south) and 70m long (east-west), with 10m of this length attributable to narrow protrusions of the mound at the western and eastern ends. The wreck mound stands 3m proud of the surrounding seabed. There are no obvious downstream or upstream scour pits, although this could be a result of the poor survey resolution. The channel in which the wreck is positioned slopes up towards the east by 4m over a distance of 150m. 4.2.1.2 1998 The topography of the wreck as surveyed in 1998 (Figure 4.15) is far more complex than that seen in 1997. The wreck mound has one main peak (4m proud of the seabed); then, disconnected from this, there are two more peaks to the north of the wreck site, both also standing 4m proud of the seabed. The primary mound is 37m wide and 58m long. There is prominent downstream (to the north of the wreck) scouring, observed as far as 240m downstream of the site and up to 2m deeper than ambient bed-level. The site slopes upwards towards the north-west and to the east of the wreck.

Owing to the complex and seemingly unpredictable nature of the evolution of the bank, caution must be applied when making predictions as to the future morphological trend. Were the trend from 1997 to 2009 to continue we would expect to see further burial of the wreck as a result of the encroachment of the bank previously found to the east of the channel and if the rate of migration of the 10m contour were to remain constant the site could be buried under 5m of sediment within the next couple of decades. However, these values are highly speculative and rely on the bank evolution remaining constant, which historical chart analysis has demonstrated is not always the case. 38

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 4.12. Goodwin Sand bank dimensions and angles for 1795 to 2009. Values for 1795 to 1947 approximated from Cloet (1954) and for 2009 extracted from MCA multibeam bathymetry.

accumulation upstream (to the south) of the site extends as far as 66m away from the wreck. There is a singular downstream scour pit, 2m deeper than the ambient seabed and extending at least 147m away from the wreck (the full extent of this is unknown, as it lies beyond the limits of survey). The site slopes to both the east and the west of the wreck, so that the wreck is positioned centrally within the channel.

4.2.1.3 1999 The 1999 survey once again illustrates a large change in the morphology of the wreck site over just a single year. The wreck mound (Figure 4.16) has two topographic highs along the length of the wreck (one to the east and one to the west). The wreck mound is 76m long and on average 30m wide; however, joined to this mound, 39

Couldrey Table 4.1. Metadata for singlebeam bathymetry surveys. Year 1997

Contractor MCA

Datum CD

Geographic coordinate system WGS84 31N

1997 1998 1999 2000

ADU ADU ADU Peacock

? ? ?

OSGB OSGB WGS 84 30N

Survey res. Line spacing 62.5m, vertical res. ±0.36m, horizontal accuracy 13m ? ? Line spacing ~5m

Survey area (m2) 150,000,000

Sourced? ü

25,000 100,000 34,000

ü (image) ü (image) ü (image) û

Figure 4.14. Singlebeam bathymetry survey of the wreck site performed in 1997. Vertical datum unknown (based on data collected by the ADU, ©Crown copyright. Historic England Archive).

Figure 4.13. Bed-level change of Goodwin Sands bank from 1997 to 2009. Areas of blue represent a decrease in bed-level with time (erosion) and areas of red an increase (accumulation) (contains public sector information, licensed under the Open Government Licence v3.0, from the Maritime and Coastguard Agency).

4.2.2 Multibeam Bathymetry Several multibeam bathymetry surveys were undertaken on the site between 2002 and 2009 (Table 4.2), which were available for consultation. The surveys for 2002 to 2005 were vertically adjusted to August 2006 (adjustments given by Bates et al., 2007: table 8) and were given relative to ODN. This adjustment was based on the assumption that the wreck elevation has not changed and that changes in elevation of the surrounding seabed level is due to real effects. Figure 4.15. Singlebeam bathymetry survey of the wreck site performed in 1998. Vertical datum unknown (based on data collected by the ADU, ©Crown copyright. Historic England Archive).

The only multibeam dataset to be provided relative to CD was the 2009 MCA survey. Using datum separation values taken from the VORF CD values are expected to 40

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 4.17. Multibeam bathymetry survey of the wreck site performed in July 2002 (based on data collected by the ADU, ©Crown copyright. Historic England Archive).

The value of the shallowest point observed over the wreck (see section 4.3.3) is affected by the resolution of the survey; therefore, before taking this measurement all surveys were converted to the same point spacing as the survey with the lowest resolution (April 2005) before this value was extracted. These reduced resolution surfaces are also used to create bed-level change plots, presented in section 4.3.

Figure 4.16. Singlebeam bathymetry survey of the wreck site performed in 1999. Vertical datum unknown (based on data collected by the ADU, ©Crown copyright. Historic England Archive).

be 2.88m shallower than OD values for the 1km × 1km area surrounding the wreck. However, assuming that the elevation of the wreck has remained constant, a difference of +1.37m between the 2006 and 2009 surfaces is observed. Therefore, since the 2.88m VORF offset appears far too large and the more recent 2009 is assumed to be more accurate, an offset of 1.37m is applied to the 2002–2006 data to convert it to CD.

4.2.2.1 July 2002 In July 2002 the wreck mound measured 53m in length and 33m in width (Figure 4.17), though much of this is sediment accumulation and the exposed wreck is only approximately 12m in diameter. The shallowest part of the wreck stood at 12.5m below CD, so that parts of the wreck were 5.5m proud of the seabed. A dog-leg of material 4m in length and 0.6m in width was observed at the eastern edge of the wreck. Scouring was observed to the east of the wreck and downstream of the eastern end. The maximum scour depth, relative to the ambient seabed, was 3m.

Bathymetry data for 2002, 2005 and 2006 were provided at bank scale coverage (with resolution 1–1.5m) and for July 2002, April 2005, August 2006 and September 2009 at wreck scale coverage (with resolution 0.25–1m). The highest resolution survey data available for the wreck site has a point spacing of 0.25m and was recorded in September 2005. MCA data for the Goodwin Sands area are available from the UKHO Inspire portal at a point spacing of 2m. However, on request data can be acquired in a .GSF format at full survey resolution. Over the Stirling Castle this is approximately 0.5m.

4.2.2.2 April 2005 The much lower resolution of the April 2005 survey (Figure 4.18) means that individual artefacts cannot be picked out from the bathymetry surface. The wreck

Table 4.2. Metadata for multibeam bathymetry surveys. Area coverage (m2)

Area res. (m)

Wreck coverage (m2)

Wreck res. (m)

Year

Month

Contractor

Equipment

Datum

2002

July

ADU

Reson 8125

ODN

485,000

1.0

12,900

0.75

ü

2005

April

RASSE

Reson 8125

ODN

1,090,000

1.5

13,400

1.0

ü

September

ADUS

Reson 8125

ODN

-

-

12,200

0.25

ü

March

RASSE, WA Reson 8125

ODN

-

-

13,100

0.5

ü

August

RASSE

Reson 8125

ODN

677,000

1.3

12,900

0.75

ü

September

MCA

Kongsberg EM3002D

CD

128,500,000

2.0

13,400

0.5

2006 2009

41

Sourced?

ü

Couldrey

Figure 4.18. Multibeam bathymetry survey of the wreck site performed in April 2005 (based on data collected by the RASSE project).

Figure 4.20. Multibeam bathymetry survey of the wreck site performed in March 2006 (based on data collected by the RASSE project).

mound measured 50m long and 15m wide (not including the area of accumulation upstream, which when included extends the wreck mound width to 48m). The shallowest part of the wreck stood 12.6m below CD, 5.4m proud of the seabed. Scouring associated with the eastern tip was extensive and was at least 1m deeper than the ambient seabed. Scouring downstream of the main structure was more confined in width in comparison with the 2002 survey and reached a maximum depth, relative to the ambient bed, of 2m.

the western tip of the wreck and reached a maximum depth, relative to the ambient bed-level, of 2m. 4.2.2.4 March 2006 Once again the wreck dimensions in the March 2006 survey (Figure 4.20) remain unchanged (50 × 15m). There is an increase in accumulation downstream of the easterly end of the wreck in comparison with the September 2005 survey. Maximum scour depths remain fixed at 2m relative to the ambient level.

4.2.2.3 September 2005

4.2.2.5 August 2006

The September 2005 survey (Figure 4.19) provides the highest resolution image of the wreck. Again, the wreck mound measured 50m in length and 15m in width. The wreck stood 5.6m proud of the seabed. There appeared to be a spherical feature, 1.3m in diameter and 200mm in height within the upstream accumulation zone, 15m south of the main deposit. Downstream scouring emanated from

The most noticeable difference in the August 2006 survey (Figure 4.21), in comparison to the March 2006 survey, is the infilling of the bedforms to the east of the wreck. The wreck mound itself and the associated areas of scouring and accumulation remained largely unchanged over this period of time.

Figure 4.19. Multibeam bathymetry survey of the wreck site performed in September 2005 (based on data collected by ADUS).

Figure 4.21. Multibeam bathymetry survey of the wreck site performed in August 2006 (based on data collected by the RASSE project).

42

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 4.23. Sidescan sonar image from the 2000 survey. Darker areas indicate lower levels of return backscatter (based on data collected by the ADU, ©Crown copyright. Historic England Archive).

Figure 4.22. Multibeam bathymetry survey of the wreck site performed in September 2009 (contains public sector information, licensed under the Open Government Licence v3.0, from the Maritime and Coastguard Agency).

darker areas represent lower or higher return levels) are often excluded, limiting the usefulness of the images. On the whole, using sidescan sonar the Stirling Castle is visible as a large upstanding wreck with strong reflectors in comparison with the surrounding sediment.

4.2.2.6 September 2009 Following the RASSE project surveys from 2002 to 2006 the next survey over the wreck was conducted in September 2009 (Figure 4.22). This survey revealed a vastly different wreck site morphology. The upstream accumulation area was entirely gone and the wreck stood 3.1m proud of the seabed. The downstream scour still remained, but had a reduced maximum depth of 1.6m, relative to the ambient area. The wreck mound itself also appeared narrower by approximately 2.5m, which could have been as a result of increased burial. The site still slopes upward towards both the east and west; therefore during this period of time the wreck was still positioned within the channel.

4.2.3.1 2000 Within the 2000 sidescan sonar survey (Figure 4.23 and 4.24) the wreck is clearly visible as a large structure with scattered objects further afield of the main structure. Bedforms are observed within the survey of the wreck, although bedform dimensions cannot be extracted without further information. 4.2.3.2 2002 Within the 2002 sidescan sonar survey (Figure 4.25 and 4.26) the wreck appears as an upstanding structure with fewer observable individual objects than in the 2000 survey. Both upstream and downstream of the wreck bedforms are less common and the seabed is comparatively featureless. Downstream of the central section of the wreck there is a patch of lower backscatter, which probably represents finer sediments. In contrast, downstream of the western

4.2.3 Sidescan Sonar Sidescan sonar surveys of varying extent have been conducted across the site since 1998 (Table 4.3). Very little accompanying metadata were supplied alongside the sidescan sonar datasets; therefore, geographic position, insonifiction direction and backscatter polarity (whether Table 4.3. Metadata for sidescan sonar surveys. Year

Contractor

Equipment

Sourced?

1983

MAS

Waverley 3000

û

1998

ADU

Imaginex 858

û

GSE Rentals, ADU

Imaginex 858 & Klein 2000

ü (non-georeferred)

ADU

Imaginex 858

û

2000 2001

Date

2002

ADU

18/08/02–19/08/02

Imaginex 858

ü

2005

RASSE Project

15/04/05–16/04/05

Klein 3000

ü (non-georeferred)

2006

RASSE Project

2008

WA (PWA)

Klein 3900

ü (non-georeferred)

2009

WA (PWA)

Klein 3900

ü (non-georeferred)

û

43

Couldrey

Figure 4.24. Sidescan sonar image from the 2000 survey. Darker areas indicate lower levels of return backscatter. The wreck was insonified from the south? (Based on data collected by the ADU, ©Crown copyright. Historic England Archive).

Figure 4.26. Sidescan from the 2002 survey. Darker area represents finer sediments. Survey track aligned east–west. Overlain on July 2002 multibeam bathymetry (based on data collected by the ADU, ©Crown copyright. Historic England Archive).

Figure 4.25. Sidescan from the 2002 survey. Darker area represents finer sediments. Survey track aligned north– south. Overlain on July 2002 multibeam bathymetry (based on data collected by the ADU, ©Crown copyright. Historic England Archive).

end there is a patch of higher backscatter, interpreted as coarser-grained material. A target to the east of the wreck, previously identified within the 2002 bathymetry, is also observed within the sidescan sonar survey (Figure 4.27). Figure 4.27. Sidescan sonar survey from 2002. Target 75m south-east of the wreck. Overlain on July 2002 multibeam bathymetry (based on data collected by the ADU, ©Crown copyright. Historic England Archive).

4.2.3.3 2005 The wreck still remains as a large upstanding structure within the 2005 sidescan survey (Figure 4.28 and 4.29) and casts a large shadow. Bedforms are fairly ubiquitous across the sonar images; only a narrow corridor downstream of the wreck exhibits none.

strength is observed downstream of the eastern end of the wreck, possibly due to the deposition of finer-grained material in this area. Disruption in the bedform patterning is observed over 100m downstream of the wreck site.

4.2.3.4 2008

4.2.3.5 2009

The wreck structure in the 2008 sidescan sonar survey (Fig­ ure 4.30 and 4.31) appears less proud of the seabed than in the 2005 survey; however, this could be a result of the angle from which the wreck was insonified. An area of lower backscatter

Sandwaves observed within the area-wide sidescan surveys (Figure 4.30 and 4.32) migrated by 80m to 100m from 2008 to 2009, very close to the 60m/yr estimated 44

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 4.28. Mosaic of sidescan sonar images from the 2005 RASSE survey. Acquired in an east–west direction insonified from the north, casting shadows to the south of the wreck; therefore, darker areas represent weaker backscatter (reproduced from Bates et al., 2007: fig. 26).

Figure 4.29. Sidescan sonar image from the 2005 survey. Acquired in an east–west direction insonified from the north, casting shadows to the south of the wreck, therefore darker areas represent weaker backscatter (reproduced from Bates et al., 2011: fig. 4).

Figure 4.30. Sidescan sonar survey from 2008 (reproduced from WA, 2009b: fig. 22 © Crown copyright, data collected by Wessex Archaeology).

45

Couldrey

Figure 4.31. Sidescan sonar, for the wreck area only, from 2008 (reproduced from WA, 2009b: fig. 22 © Crown copyright, data collected by Wessex Archaeology).

Figure 4.33. Sidescan sonar, for the wreck area only, from 2009 (reproduced from WA, 2010b: fig. 4e © Crown copyright, data collected by Wessex Archaeology).

from the crest migration data presented in section 4.1.2. The wreck remained partially exposed in 2009, as evident from the sidescan sonar survey (Figure 4.33); however there appears to have been some burial of the outlying features, especially along the southern side of the wreck. Infilling of the downstream scour pit is also evident from the sidescan sonar surveys.

Generally speaking, values of roughness increase over the wreck (Figure 4.34), whereas values of hardness decrease over the wreck. Lawrence and Bates (2001) believe that this may be due to the exposed waterlogged wood affecting the return echo. AGDS does not give a difference in signal over the scour pit, and therefore is unable to pick up the subtle changes in sediment type over this area.

4.2.4 Acoustic Ground Discrimination

4.2.5 Magnetometer

Acoustic ground discrimination systems (AGDS) have been deployed during surveys of the Stirling Castle (Table 4.4). These provide return echo signals from which indexes of seabed ‘roughness’ and ‘hardness’ are created that have been used to classify seabed type (Lawrence and Bates, 2001).

A variety of magnetometer surveys have been carried out on the site (Table 4.5) from which it was possible to source a single geo-referenced survey dating from 2002. Images for the 2008 and 2009 surveys have been extracted from the relevant PWA contractor reports (WA, 2009b and 2010b).

Figure 4.32. Sidescan sonar survey from 2009 (reproduced from WA, 2010b: fig. 2 © Crown copyright, data collected by Wessex Archaeology).

46

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Table 4.4. Metadata for acoustic ground discrimination surveys. Year

Contractor

Equipment

Area

Sourced?

2000

RASSE

Echoplus

100m x 100m

ü (images)

2001

ADU/RASSE

Echoplus – 192KHz and 50KHz

2005

RASSE Project

û

2006

RASSE Project

û

û

Figure 4.34. Line track data collected with the Echoplus system, graduated for the roughness (E1) and hardness (E2) indexes from the 2000 survey, with underlay of sidescan sonar (reproduced from Lawrence and Bates, 2001: fig. 2). Table 4.5. Metadata for magnetometer surveys. Year 1983

Contractor MAS

1998

ADU

2000

ADU/Licensee

2001

ADU

2002

ADU

2008 2009

Date

Equipment ELSEC

Resolution/accuracy

Proton

Sourced? û û û

Geometrics Flux

û

Geometrics Caesium

ü

WA (PWA)

Marine Magnetics Explorer

ü (image)

WA (PWA)

Marine Magnetics Explorer

19/08/02

47

Sensitivity 0.02nT, absolute accuracy 0.2nT, res. 0.001nT

ü (image)

Couldrey anomaly of 169nT to the eastern end of the main wreck were identified. A further 23m to the east of the wreck an anomaly of 423nT was recorded. Outside the 200m × 200m survey area two further anomalies were identified: A, an anomaly measuring 283nT; and B, measuring just 125nT; both could represent buried wrecks or isolated debris and, as they could not be seen in the sidescan sonar survey images, required further surveying (carried out in 2009) to identify their source. In 2009 WA expanded their magnetometer survey of the site to cover a further 200m to the east of the site and a further 200m to the south-west of the site (Figure 4.37). The wreck itself gave a magnetic anomaly reading of 615nT. Slightly to the south-east of the site, but still within the survey area, two magnetic anomalies measuring 592nT and 221nT were identified. These are thought to be associated with a buried wreck and correspond to the same location as the previously identified anomaly within the 2002 sidescan sonar survey, the 2002 multibeam bathymetry survey and the 2008 magnetometer survey. Still further east of this wreck, outside the Stirling Castle survey area, but within the bounds of the eastern survey area, is another distinct anomaly with a value of 46nT. WA interpreted this as a deeply buried wreck site. Within the western survey area anomalies of 125nT and 319nT were observed that were thought to be associated with buried wrecks. Outside the designated survey areas, anomalies I– III are all thought to indicate the remains of further buried wrecks, though none were present within the sidescan sonar surveys.

Figure 4.35. Magnetometer survey from 2002. Yellows represent a low magnetic field strength and purples and blues a higher magnetic field strength. Overlain on April 2002 multibeam bathymetry data (based on data collected by the ADU, ©Crown copyright. Historic England Archive).

The 2002 magnetometer survey does not cover the Stirling Castle wreck area, instead passing 50m to the west of the site in a proximally south-westerly to north-easterly direction (Figure 4.35). A peak in magnetic field strength observed 100m south-west of the wreck site probably corresponds to another separate piece of isolated debris unrelated to the wreck, since it is unlikely that material will have been transported upstream away from the Stirling Castle wreck site. In the 2008 WA magnetometer survey (Figure 4.36) over the main wreck site an anomaly of 400nT and a smaller

Figure 4.37. Magnetometer survey from 2009 (reproduced from WA, 2010b: fig. 2c © Crown copyright, data collected by Wessex Archaeology).

Figure 4.36. Magnetometer survey image from 2008 (reproduced from WA 2009b: fig. 22b © Crown copyright, data collected by Wessex Archaeology).

48

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Table 4.6. Results from 2005 grain size analysis. Latitude and longitude in WGS84, Zone 31N. Latitude (m) 5682428 5681400 5681591 5681336 5681358 5681423 5681461 5681475 5681362

Longitude (m) 396216 396134 395773 395859 395900 395920 395890 395811 395781

Mean grain size (μm) 534 527 335 1835 345 2421 3868 780 742

Median grain size (μm) 372 315 303 263 276 353 1794 494 477

4.2.6 Grain Size Analysis In 2005 and 2006 grain size analysis was carried out by a group from the University of St Andrews. Nine samples are given for the 2005 survey (Table 4.6 and Figure 4.38). During the August 2006 survey 31 grab samples were taken (Figure 4.39), which were then analysed using a Sedigraph Laser Particle Analyser (Bates et al., 2007: 78); the raw data have not been found. Results from the 2005 survey (Table 4.6 and Figure 4.38) indicate that the coarsest sediments are found downstream of the wreck site. This suggests that surface fine-grained sediments have been eroded from the scour region. Sediments from the 2006 survey (Figure 4.39) show a bimodal distribution, with coarser-grained sediments found downstream of the wreck and finer-grained sediments to the east of the wreck, where deposition is occurring.

Figure 4.39. Mean grain size distribution from the 2006 survey for a 200m × 200m area around the wreck. Note: the alignment of the wreck and north arrow are not consistent.

4.2.7 Sub-bottom

below the gulley bottom and scouring down to bedrock had possibly occurred in places.

The only known sub-bottom profiler survey of the site was conducted in 1983 by MAS using an ORE sub-bottom profiler (Redknap, 1990). This survey covered a total of 7.5km2 over the Goodwin Sands area. While no copies of the data have been acquired Redknap (1990) observed that at the Stirling Castle site the bedrock layer was 0.5m

4.3 Meso- and Wreck-Scale Site Evolution The evolution of the Goodwin Sands bank has already been quantitatively described by comparing both singlebeam and multibeam bathymetry data sets (Figure 4.13). Such an approach can also be applied at a smaller scale to consider meso-scale bank change (coverage 500,000m2 and a resolution 1m-1.5m) and site-scale (coverage of 13,000m2 and resolution 0.25m-0.75m) changes for the period between 2002 and 2009. 4.3.1 Annual Meso-Scale Site Evolution 4.3.1.1 July 2002–April 2005 From July 2002 to April 2005 large parts of the channel underwent bed-level gain (Figure 4.40), notably to the south-west of the wreck, largely as a result of sandwave migration, and to the south-east of the, site where over 10m of sediment accumulated. In more scattered areas across the channel bed-level loss was observed at a rate of 1m/yr.

Figure 4.38. Mean grain size from 2005 analysis, overlain on April 2005 multibeam bathymetry (based on data collected by the RASSE project).

49

Couldrey

Figure 4.42. Meso-scale bed-level change for the period of August 2006 to September 2009. Scale as Figure 4.40. (Contains public sector information, licensed under the Open Government Licence v3.0, from the Maritime and Coastguard Agency).

Figure 4.40. Meso-scale bed-level change for the period July 2002 to April 2005. Cream represents areas that have not undergone significant changes, blues areas where there has been significant bed-level loss and reds areas where there has been significant bed-level gain. The location of the wreck is marked with black ellipse.

4.3.1.3 August 2006–September 2009 4.3.1.2 April 2005–August 2006

From 2006 to 2009 (Figure 4.42) the channel underwent a massive amount of bed-level accumulation, with values in the region of 4m, along the eastern flank engulfing the wreck site. The migration of sandwaves is still apparent to the north of the wreck site. The eastern edge underwent massive accumulation over the same period of time. Bed-level loss is observed to the south-west of the site at a rate of 2m/yr.

The most immediately obvious feature of the 2005 to 2006 bed-level change plot (Figure 4.41) is the striation of bedlevel loss and gain owing to the migration of sandwaves towards the north. The eastern half of the channel underwent bed-level gain over this period of time, with values as high as 5.5m found to the south-east of the site, where the bank had a slight protrusion.

4.3.2 Annual/Monthly Wreck-Scale Site Evolution 4.3.2.1 July 2002–April 2005 The most striking difference between the 2002 and 2005 bathymetry surveys (Figure 4.43) is the massive bed-level loss observed within the downstream scour pit, with values greater than 2m of loss. Half a metre of bed-level loss is also observed directly atop the wreck mound at the western end and up to 10m upstream of the wreck. At the eastern end a gain of 0.5m was observed on the wreck and downstream of the eastern tip of the wreck accumulation occurred, so that the bed-level was 2.5m higher in 2005 than in 2002. At a wreck-site scale the migration of medium subaqueous dunes is observed through the banding of bed-level change. 4.3.2.2 April 2005–September 2005 Over the period of five months from April 2005 to September 2005 most of the wreck site underwent no change (Figure 4.44), possibly as a result of the particularly calm wave environment over this period of time (Figure 4.3d). Bed-level change is restricted to the western and eastern ends of the wreck, and is opposite and equal to the change from 2002 to 2005: that is there was some

Figure 4.41. Meso-scale bed-level change for the period of April 2005 to August 2006. Scale as Figure 4.40.

50

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 4.43. Bed-level change from July 2002 to April 2005. Cream represents areas that have not undergone significant changes, blues areas where there has been significant bed-level loss and reds areas where there has been significant bed-level gain.

Figure 4.44. Bed-level change from April 2005 to September 2005. Scale as Figure 4.43.

restoration towards the level in 2002. Downstream of the western end of the wreck, slightly outside the scour region, bed-level loss is observed with values in the region of 0.3m-0.7m. Downstream of the eastern end, also slightly outside the main scour pit, there is an area of bed-level gain with values of approximately 0.5m.

stretch of bed-level loss emanating from the eastern end of the wreck and a small amount (0.3m) of sediment gain along the northern, downstream, side of the wreck. To the east of the wreck there is an increase in bed-level between the values of 0.3m and 1.3m, which is likely to have resulted from the westward encroachment of the channel edge over this period of time.

4.3.2.3 September 2005–March 2006 Bed-level change for the period of September 2005 to March 2006 (Figure 4.45), once again occurred in opposite directions at either end of the wreck. Bed-level loss was observed at the eastern end (0.3m–2.0m) and bedlevel gain was observed at the western end (0.3m–0.9m). Downstream of the wreck three strips of flow-aligned accumulation occurred.

Over the entire RASSE observation period, 2002–2006, there was a net deposition of sediment within the wreck area of 8677m3 (Bates et al., 2007: 95; 2011: 412). When observing each period of time separately (Figure 4.43 to 4.46) this change is not initially apparent, as there is a large spatial variability in this pattern that is often masked by bedform migration and localised change surrounding the wreck structure.

4.3.2.4 March 2006–August 2006

4.3.2.5 August 2006–September 2009

Very little change is observed directly over the wreck for the period between March 2006 and August 2006 (Figure 4.46). Elsewhere there is an isolated downstream

Over the final period for which multibeam data of the wreck site are available, from August 2006 to September 2009 (Figure 4.47), a massive increase in bed-level is 51

Couldrey

Figure 4.45. Bed-level change from September 2005 to March 2006. Scale as Figure 4.43.

Figure 4.46. Bed-level change from March 2006 to August 2006. Scale as Figure 4.43.

Figure 4.47. Bed-level change from August 2006 to September 2009. Scale as Figure 4.43. (Contains public sector information, licensed under the Open Government Licence v3.0, from the Maritime and Coastguard Agency).

52

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 observed across the wreck site. On average, bed-level gain across the site was 2m over the three-year period; however, large areas where more than 3m of bed-level gain occurred are observed to the east of the wreck and downstream of the wreck, where previously there had been a downstream scour pit. To the north-west of the site there is an isolated area of bed-level loss associated with the migration of a sandwave. Atop the wreck structure itself there is no observable change in bed-level; however, this could have been a result of the OD to CD conversion of the 2006 data. 4.3.3 Maximum and Minimum Depth Locations The evolution of the wreck site can be observed through bed-level change plots (sections 4.3.1 and 4.3.2 above) as well as through the presentation of the maximum and minimum depth locations, which highlight the peak of the wreck structure and the trough of the scour pit, respectively (Figure 4.48). While the position of the shallowest depth atop the wreck has remained fairly constant year-on-year, with the exception of April 2005, where the wreck mound peak is found 5m northwest of the other years, a much larger variability in the location of the maximum scour depth is observed. Initially the maximum depth was found 62m north-east of the wreck. This migrated south-west year-onyear at a rate of 5m/yr. A shift of 20m was observed from August 2006 to September 2009, bringing the deepest point of the scour pit to just 28m exactly north of the wreck. It has been observed that for structures orientated obliquely to the flow scouring is more extensive downstream for the end orientated into the flow (see discussion throughout Quinn, 2006). As the wreck structure has not changed in its orientation this then suggests that the flow has moved more towards the west-south-west. This is likely also to be connected to the rotation of the channel in which the wreck is positioned from 23°T in 1997 (79° to the orientation of the wreck) to 52°T in 2009 (48° to the orientation of the wreck), a rate of 2.4°/yr.

Figure 4.48. Positions of maximum depths within scour pit and minimum depths atop wreck for each multibeam bathymetry survey, overlaid on 2009 MCA bathymetry (contains public sector information, licensed under the Open Government Licence v3.0, from the Maritime and Coastguard Agency).

bathymetric survey, 5 × sidescan surveys, 3 × AGDS, 4 × magnetometer surveys and 1 × sub-bottom survey. Best efforts have been made to fully utilise the data available, even where positional and vertical reference systems are missing. Collation of this data, even allowing for the absent data, represents the most complete collation and analysis of such survey data in relation to the Stirling Castle site for the period 1979–2009. It should also serve to provide a set of comparative analysis to further inform and contextualise surveys undertaken since 2009. From the analysis of these data it is clear that the Goodwin Sands bank is a very complex system that operates on a full range of timescales, from medium subaqueous dunes that migrate at a rate of 100s of metres per year to the bank system as a whole, which rotates at a rate of less than a degree per decade. While the migration of the mediumsized subaqueous dunes is not important in terms of the burial and exposure cycle of the wreck, these bedforms are still critical to the overall transport of sediment within the Goodwin Sands bank system.

4.4 Discussion Through the collation of all known geophysical data associated with the Stirling Castle wreck site, 20 different geophysical surveys have been acquired (4 × singlebeam bathymetric surveys, 6 × multibeam bathymetric surveys, 4 × sidescan sonar surveys, 1 × AGDS, 2 × grain size analysis and 3 × magnetometer surveys) and presented within this section. Surveys that could not be acquired include: 1 × singlebeam bathymetric survey, 1 × multibeam

The sedimentary features that have the largest impact on the wreck site are the very large subaquous dunes (or sandwaves) and the migration of the bank itself. The former have wave heights in excess of 4m and wavelengths of approximately 250m and are capable of migrating up to 100m a year. Presently these bedforms traverse towards

Table 4.7. Bed-level change between multibeam bathymetry surveys for the 130m × 100m area surrounding the wreck and the equivalent rate of change per year. Time period July 2002–April 2005 April 2005–September 2005 September 2005–March 2006 March 2006–August 2006 August 2006–September 2009

Period in months 32 5 6 5 37

53

Change over wreck area (m) +0.4 0.0 +0.1 +0.1 +2.2

Rate of change (m/yr) +0.2 +0.1 +0.2 +0.3 +0.7

Couldrey the north-east and are largely located to the west of the wreck; therefore, they do not directly bury or expose the wreck, but their presence is still likely to influence the local hydrodynamics of the channel, which in turn will influence the movement of sediment around the wreck structure. Bed-level change across the site has largely remained within surveying errors (Table 4.7) (±0.3m assuming International Hydrographic Organisation (IHO) (2008) Special Order standards). The only period of time where we see a rate above 0.3m/yr is during the 2006–2009 period. This is surprising as it is generally assumed that there would be greater rates of change during those periods of time over which the site will have been exposed to storm waves (e.g. July 2002–April 2005, September 2005– March 2006 and August 2006–September 2009) compared with those which cover only summer months (April 2005–September 2005 and March 2006–August 2006). This indicates, as noted earlier in section 4.1.2, that the site processes are not strongly driven by storm conditions and that the continual sedimentary processes play more of a key role in determining the exposure or burial of the site. Over the observation period 1997–2009 the channel in which the wreck is positioned has been shifting towards the west. In the last three years of this observational period this has led to a massive increase in the bed-level (2–3m) around the wreck. It is predicted that this trend will continue for a decade from 2009, leading to the near total burial of the wreck. Predicting trends beyond this period of time becomes increasingly speculative owing to the complex and unpredictable nature of the evolution of Goodwin Sands bank system, involving sandwave migration, bank system rotation and so on. Previous episodic instances of exposure in the late 1970s and late 1990s strongly suggests that the site will not remain buried indefinitely, however.

54

5 The In Situ Archaeological Remains of the Stirling Castle Julian Whitewright investigation of medieval and post-medieval shipwrecks, understanding the wrecking event and subsequent site-formation processes is a critical part of our wider appreciation of such archaeological sites. The historically derived view of the loss of the Stirling Castle was set out in section 2.4 and is returned to here in section 5.1 as a means to elucidate the initial site-formation processes undergone by the vessel during, and in the immediate aftermath of, its loss in November 1703.

The artefact assemblage recovered from the Stirling Castle during the various periods of fieldwork on the site provides a highly informative record of the material culture associated with a late seventeenth-/early eighteenthcentury British warship. But it may be reasonably suggested that the extent of such recovered material is easily surpassed by the archaeological material that lies in situ on the seabed, especially when the surviving hull of the vessel is included. The biological, chemical and physical degradation of structural material, with resulting associated loss of artefactual material, has undoubtedly taken place during periods of greater exposure. But, although certainly not as complete as when discovered in 1979, the site of the Stirling Castle, in all probability, still comprises an extensive set of relatively complete, coherent remains.

Work on the Stirling Castle by WA in their role as archaeological diving contractor to English Heritage (Historic England from April 2015) for the PWA led to repeated monitoring of the site between 2003 and 2009. This work augmented that carried out during the 1990s and early 2000s by the previous contractor, the ADU, and Bob Peacock, the site licensee (details in chapter three). Work by all these parties has entailed the measured survey of the exposed elements of the site in conjunction with the creation of a photographic and video record of the seabed remains. In assessing that material, in conjunction with the more limited record from the 1979 season, the aim of this chapter is to provide an overview of the surviving seabed remains up to the point at which geophysical survey indicates significant sediment accretion and site reburial by 2009 (see chapter four). This aim is achieved by reviewing the known extent of material at three defined periods in the archaeological history of the site: 1979 (first discovery), 1998–2000 (re-exposure) and 2006–2009 (apparent reburial). The purpose of such an exercise is two-fold: first, it allows an assessment of the extent of material that has been lost since 1979; and, second, it provides an understanding of the extent of material still likely to survive in situ. In this regard the size and extent of the remaining ‘seabed archive’ is critical for allowing a comparison of the relative significance of the site when compared with remains of contemporary vessels, for allocating resources to manage those archaeological remains and for placing the recovered artefact assemblage in its more immediate archaeological context.

The ideas and processes that underpin the management of underwater cultural heritage (UCH), including shipwrecks, have developed beyond recognition since the Stirling Castle was discovered. Deliberate underwater archaeological work resulting in the large-scale recovery of material has diminished in the UK since the 1970s and 1980s, although examples do occur when deemed necessary for the purposes of preservation by record, academic research or as a result of marine development (e.g. Swash Channel, London (1665), Alum Bay 1, Gresham Ship). The focus has instead increasingly been on preservation in situ and the methodologies for the application of such policies and processes have become increasing complex (for discussion see Gregory and Manders, 2011). Such policies have not been without criticism; they risk, at a general level, leaving UCH in a situation of being ‘out of sight, out of mind’, and in the specific case of the Stirling Castle the policy was considered by the site licensee as little more than ‘staged and managed neglect’ (Peacock, 2009b: 5). It has certainly been recognised widely within the profession of maritime archaeology that the process of preservation in situ entails far more than simply leaving a site alone and prohibiting any form of interference. Active management, or in some cases intervention, is critical to preserve the environmental equilibrium that allows such a policy to be considered viable in the first place. Coupled with this must be an effective programme of site monitoring to record continuity and detect any changes. This in turn must be underpinned by clear knowledge of what the archaeological remains of the site actually comprise. To this end, as Adams (2013: 20–21) has pointed out in his wider consideration of the maritime archaeological

5.1 Initial Site Formation Processes The morphological processes acting on the site in the present day were covered in detail in chapter four, but it is useful here to address the initial stages of that process – the wreck and loss of the vessel itself. The narrative set out in section 2.4 makes it plain that this was a relatively prolonged process. The phrase ‘relatively prolonged’ is certainly important and worthy of further clarification. 55

Whitewright It is possible to reconstruct the human involvement in the wrecking process as lasting for around 36–42 hours, from the evening of Friday 26 November, which may be considered as ‘pre’ Great Storm, when the Stirling Castle was at anchor, to the rescue of the surviving crew sometime on Sunday 28 November, when the storm had abated sufficiently to allow boats from Deal to reach the vessel. The loss of the vessel was completed at some point after the crew had been rescued, but, as the subsequent archaeological remains so strikingly testify, the vessel was not subjected to a violent disintegration.

astern, more slowly, while dragging its anchor, before grounding and eventually being swallowed by the sands. Unlike the Stirling Castle, the site of the Northumberland is orientated with its bow to the south-east (Pascoe and Peacock, 2015: 134) indicating that the vessel was lying roughly broadside to the weather as it broke up. Despite this, it still comprises a coherent wreck mound with large areas of discernible structural remains and associated artefacts (see Pascoe and Peacock, 2015: 134–38). In highlighting the relationship between different timescales of loss and the resulting nature of the archaeological remains it is very easy to overlook the human contribution to that process. The relative preservation of the Stirling Castle would seem to be a clear product of the vessel orientation at the time of striking the Goodwin Sands. In turn, this is related to the ability of the crew to set a new anchor, at least partially, after the original ones had been fouled by the Northumberland. Finally, and perhaps most profoundly, it is almost impossible to imagine the two days spent by the surviving crew on board, being battered by a violent storm and washed over by huge waves while the majority of their fellow crew were lost. It would be reasonable to expect such an experience to have had at least some impact on the arrangement and location of the material culture, in all its forms, contained on board the ship as the surviving crew desperately attempted to remain alive.

When compared with other documented archaeological remains, the loss of the Stirling Castle was neither especially rapid nor particularly drawn out. In terms of the former, the well-documented losses of the Tudor Mary Rose (1545) and the Swedish Vasa (1628) perhaps provide the most graphic, and certainly well-known, examples of large warships that capsized and sank very rapidly while under sail. Neither of those vessels ran aground as part of the wrecking process, but one vessel that did was the nineteenth-century sailing barque Flower of Ugie, which struck the Horsetail Sand in the eastern Solent in December 1852. The Flower was recorded by observers as breaking up in a matter of hours, and archaeological testimony to this violent process is provided by the scattered nature of the seabed remains, including several large sections of wooden hull structure (for a full account see Whitewright and Satchell, 2011). At the other end of the timescale is the loss of HMS Invincible, also on the Horsetail Sand, in February 1758. After running aground and resisting attempts at refloating, Invincible remained sufficiently intact, for long enough, for a significant amount of the ordnance, stores and so on to be salvaged by the crew.

5.2 Site Development: Methodology The methodology for assessing the change to the site across the period 1979–2009 and for understanding the possible extent of surviving material was based initially around a rationalisation of the recorded site plans (Figure 5.1). This was founded upon work to produce an integrated site plan illustrating all seasons of work and through which all of the phases of survey and recording carried out since 1979 could be viewed at the same time. The annotation provided on the various available plans would also help to position artefacts raised at different times during the various phases of work on the site. Key to achieving this was moving conceptually from the systematic digitisation of individual site plans to a season-based series of overlaid plans.

The purpose of this brief review is to emphasise that the Stirling Castle, as far as the timescale is concerned, underwent a relatively drawn-out wrecking process. The grounding and loss of the ship was not a rapid event, in the manner of the Flower of Ugie. It was a protracted process that took place over a couple of days, covering the initial anchoring of the vessel, the development of the storm, grounding, loss and rescue. It is also possible to draw a useful comparison between the archaeological remains of the Northumberland and the Stirling Castle. The former is recorded by observers as having lost its anchors and being driven before the wind onto the Goodwin Sands, probably fouling the anchors of the Stirling Castle in the process. The re-deployed surviving anchor of the Stirling Castle, found on the seabed in 2002, seems to have kept the bow at least partly towards the storm, with the stern of the vessel grounded. Accounts that describe the surviving crew gathered on the poop deck fit the notion of the stern of the ship lying furthest from the wind and waves, while the security, albeit greatly reduced, afforded by the surviving anchor seems to have allowed the stricken vessel to remain at least partially stable. It is this that has resulted in the clear difference between the seabed disposition of the Stirling Castle and the Northumberland. The latter was driven before the storm onto the sandbank, while the former went

The site plans used were contained within the digital archive of material from the site and their origin, along with the principal survey method used to create them, and the information contained within each plan, is outlined in Table 5.1. Analysis of this dataset was done chronologically, beginning with the 1979 plan, which was deemed to be representative of the extant remains of the vessel at the point of discovery. This was overlain with site plans from 1999/2000, 2006 and 2009. In each case, the digitised versions of the original plans were scaled to a consistent scale of 1:100 and the main features present in each plan were plotted. For clarity, and because of the high level of variation in how such features had been drawn across the various seasons of work, a standard symbol was used to represent features such as anchors and cannons. The resulting combined plan should therefore be seen as a 56

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 5.1. Representative site-plans based on archive material from 1979, 1999/2000 and 2006/9. Features such as cannon and anchors have been given a standardised symbol and should not be used to differentiate actual types or sizes.

57

Whitewright Table 5.1. Sources of Stirling Castle site plans 1979–2009. Year

Type

Annotation

Survey Method

Source

1979

Published, processed overview plan, including main features

Artefact locations Areas of damage

Centreline and offset

Dave Perkins

1999/2000

Published, draft overview plan, including main features

Artefact locations Survey reference points

Direct survey measurement

SeaDive/Bob Peacock

2006

Published, processed overview plan of main features

None present

Interpreted multibeam

RASSE Project

2006–2009

Published, fully processed overview plan of main features

Feature IDs Material types Survey methods

Multibeam & sidescan sonar, Wessex Archaeology completed with diver survey located by acoustic positioning

representative, rather than an interpretative, tool. Classes of major seabed features, such as different types and sizes of cannon, cannot be distinguished from one another, but their relative visibility over time can be traced. Finally, a silhouette of the outline of the main gun deck for a third-rate of the Thirty Ships programme was then added to further the sense of scale, the dimensions were taken from Endsor’s reconstruction (2009: 14–15) of the Lenox, built by John Shish before the Stirling Castle. Although the Stirling Castle was refitted in 1699 the dimensions of the vessel, and those of the other refitted third-rates, stayed largely the same (see section 2.2). The gun-deck silhouette was aligned on the stern-post, as located in the 1999 and 2006/2009 site plans, because of its consistent location and status as a readily identifiable and integral element of the vessel’s structural remains.

are included in the site archive through the collection of licensee reports filed on a yearly basis. These two sets of documents retain individual views of the site’s formation processes across the period covered here that can supplement the wider surveys of the site. 5.2.1 Observations on Overall Correlation As might be expected from the account of work on the site provided in chapter three, the 1979 site plan shows the least correlation with the subsequent plans from 1999/2000 and 2006–09. This is perhaps to be expected given the 1979 team’s lack of archaeological expertise, the chronological separation between surveys and the reported degradation of the vessel in the 20 years between 1979 and 1999 and the subsequent 30 years to 2009. From 1979 to 1999 there is only a reasonable correlation between the positions of two of the vessel’s cannons located on the port quarter, and a third cannon lying fore-and-aft amidships near the aft end of the forecastle. From 1979 to 2009 there is almost no correlation in the main features, with the exception of one of the cannons just noted on the port quarter. By contrast, there is a strong correlation between some specific elements recorded on all of the plans from 1999 onwards – in particular, the large anchor in the central forward area of the site and the three cannons located on the forward starboard side. Other consistently positioned features are the line of guns along the port midships area and the stern-post/rudder/ transom. It can also be highlighted that WA were obviously confident enough in the accuracy of the 1999/2000 survey to use it as the basis for their own rendering of the structural remains present along the forward port side of the ship. As a result of this, further doubt must be cast on the detailed accuracy of artefact locations contained within the 1979 plan. Material rendered on that survey certainly existed on the site, but its location must be assumed to be accurate in only general terms within the three zones that the 1979 team identified; fore, mid-ships and aft. This problem was highlighted during the visit of the ADU to the site in 1992 during a period of slightly lower sediment levels. As noted in section 3.1.2 the ADU commented (1992: 2) that the site in front of them bore little resemblance to the published plan from 1979.

Following the integration of site plans from 1979, 1999, 2006 and 2009 it was possible to return to the video and photographic archive to try to identify areas where repeat recording had been undertaken. In particular, the stern and port quarter of the vessel have been the subject of prolonged work by SeaDive in 1999 and WA in 2006–9. The former of these is notable for having recorded sections and elevations, as well as plans, of the exposed structure. The measured surveys of the site impose a degree of (apparent) recorded certainty on the archaeological remains that they represent. This is perhaps most true of the surveys undertaken between 2006 and 2009 by WA that were underpinned by geophysical survey and located through acoustic positioning. But it is important not to lose sight of the observations of the changing on-site situation as recorded by those who visited it on a regular basis and were/are most familiar with it through the diver’seye view. In that regard, two sets of accounts contained within the site archive are of especial significance. The first is provided by Martin Dean, who visited the site during the 1979 work, returned in the late 1980s as director of the ADU, and continued to visit the site until the ADU contract ended in 2002. His observations on the changing nature of the site and in particular the levels of exposed material are recorded in the reports filed by the ADU following their visits to the site. The second set of eyewitness accounts come from Bob Peacock, who became site licensee in 1993 and whose observations

Initially there was a surprisingly poor match between the 2006 RASSE survey and the 2009 WA survey. It quickly 58

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 became apparent that the reason for this was that the drawn scale on the 2006 RASSE survey was wrong by a factor of 20 per cent; the 10 metres indicated on the scale bar actually represented 12 metres. The identification and rectification of this error produced a very close match between many of the features identified in the 2006 and 2009 surveys. Comparison of these two plans, produced separately within a relatively short space of time, also highlighted the value of direct diver observation of seabed remains. Many of the features highlighted, but not identified, from the multibeam survey undertaken for the RASSE survey were later identified as cannons during the WA survey. That work included direct observation by archaeological divers to supplement the geophysical survey and to clarify the identification of such features. In doing so it indicated the need for the physical ground-truthing of material, rather than relying solely on geophysical data.

port side of the ship. In addition, the positions of many, but certainly not all, of the artefacts recovered during the 1979 season are located on the plan. From the perspective of assessing the original extent of vertical exposure, the artistic impressions drawn by Dave Perkins, as well as some of the diver reports, claimed that both the main and upper gun decks were visible. Wreckage from the forecastle and the quarter/poop deck was located on the upper gun deck at either end of the site. Recent interviews undertaken during this project with Tom Brown and John Chamberlain suggest that Dave Perkins’ artistic impressions exaggerated the extent of the exposed remains. But much of the well-documented material raised in 1979 would have been situated on the upper deck – such as the copper cauldrons and exposed hearth – or would have fallen to the upper deck when displaced – such as the ship’s bell. In that regard, the claim that the upper deck, or more probably the main deck, of the ship, littered with the fragmented material from the forecastle and quarter/poop deck, was the uppermost exposed area of structure does not seem unreasonable, even if it cannot be confirmed. In the absence of other evidence, the interpretation of the extent of the original vessel exposed in 1979, shown in Figure 5.2, can only rely on the testimony of those present at the time.

Establishing the observed extent of the site and assessing the correlations between the three main periods allowed the subsequent interpretation to be further refined. There is an overall coherence to the recording of remains from 1999 onwards that allows the specific areas of the site, such as the stern-post and rudder, to be better understood. Critically, as noted in the introduction, this has been underpinned by archaeological monitoring of the site with an awareness of potential and ongoing areas of concern. In contrast, the material exposed in 1979 is not well located and the focus underwater was placed on the recovery of artefacts, rather than the recording of exposed material. Additionally, the 1979 work had no comprehension of the wider processes that the site was subject to, because there was no previous work to draw upon. Work on the site has developed in a largely organic manner (see chapter three) and each period of work was conducted by different groups with differing aims and objectives, often in response to a rapidly changing site morphology that dictated the type of work that needed to be carried out. Inevitably this has influenced the results achieved.

The coherence and structural integrity of the vessel’s remains in 1979 are much harder to assess because of the nature of the surviving record. It is clear that there was a considerable depth to the remains, as evidenced by the nature of the material on the surface of the site. But the size of the vessel illustrated in the 1979 plan is much larger, both in length and beam, than that of a ship of the Stirling Castle’s class at the level of the gun-deck. The difference amounts to c.4m at the bow, c.5m at the stern and c.2m along the port side. Even allowing for elements such as the beak-head not being fully recorded or accounted for at the bow of the vessel, or for the potential inaccuracies in the survey, such a difference is hard to reconcile. The most obvious explanation is that the hull of the vessel had already undergone some outward collapse at its time of discovery in 1979 (Figure 5.3). This is corroborated by the segment of hull and gun-ports collapsed outward and lying on the seabed at the port quarter, and to a lesser extent the alignment of the bowsprit and, it is assumed, the beak-head at the bow of the vessel. Given the nature of the wrecking process that the vessel underwent, and the general characteristics of the dynamic marine environment within which it was located for nearly 300 years, such outward collapse is certainly not surprising.

5.3 Site Development: 1979–2009 5.3.1 Discovery: 1979 The extent of the remains recorded in 1979 (Figure 5.1) and assumed to reflect the general nature of the seabed material are largely constrained by the apparent shape of the original ship and the encroaching sandbank along the northern edge of the site. The vessel’s total length is shown as 55m from stem to stern-post and a width from the stem/ stern centreline of 8.5m. Further material assumed to represent some of the beak-head and bowsprit remains are located to the west of the stem post and a piece of hull structure pierced by three gun-ports is shown as having collapsed outward from the port quarter. The plan is clearly demarcated into three areas that correspond roughly to the forecastle, midships area and quarter/poop deck. Some internal structural features are included such as a stairway, along with larger artefacts: three anchors and 19 cannons, five of which are located outside the demarcated line of the

The earliest video record within the archive is from an ADU dive in 1995 and contains coverage of the bow area. It is clear that sediment levels are high, but there was still some exposed structural material. The large anchor present on all the surveys is clear of the sediment, along with one or two other notable features, such as the large square beam shown to the south-west of the anchor on the 1999 site plan and some concreted guns. The edge of the 59

Whitewright

Figure 5.2. Schematic diagram to illustrate the interpreted surviving extent of vessel structure in the same manner as that produced following the 1979 season of work.

Figure 5.3. Schematic diagram illustrating the probable extent of outward collapse of the Stirling Castle’s hull. The main-deck is used as a reference point because it is the highest level of coherently preserved structure across all periods of investigation.

60

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 port side of the vessel is discernible, but it is certainly not exposed to a great extent or vertical height.

decay, caused by gribble and shipworm, is visible on most of the exposed wooden surfaces, indicating that the commonly observed cycle of exposure through loss of sediment cover, infestation by biological organisms, degradation and ultimately loss of material was underway. Such biological decay can result in the rapid destruction of even previously large and coherent structural remains and would certainly have served to accelerate the degradation of fully exposed vessel structure, accentuating the process of hull collapse (for discussion see Gjelstrup Bjӧrdal and Gregory, 2011: 78–80).

It is therefore extremely difficult to assess how much material from the site was exposed and/or lost following the discovery of the site in 1979. Its apparently rapid reburial – by 1980 – suggests, given our current understanding of the overall site morphology cycles (chapter four), that the site was exposed for perhaps a few years before the 1979 discovery, after which it was rapidly reburied through sediment movement. Martin Dean speculated on several occasions in the ADU reports of the 1980s and early 1990s about the point at which the hull of the ship collapsed. The hindsight provided by the re-emergence of the hull in the late 1990s, in conjunction with the comparisons undertaken here indicates that such collapse had probably already begun in 1979, but was much more gradual and less dramatically sudden than speculation at the time suggested.

5.3.2.1 Stern Area Video footage from 1998 onwards records increasing amounts of exposed structural material, especially in the stern area of the site, where the top of the rudder is clearly visible, standing about 1.5m clear of the seabed. It is notable that, as well as timber surfaces with extensive evidence of biological decay, there are also many surfaces visible with little or no decay, indicating their relatively recent exposure following a period of burial in a largely oxygen-free environment. A further feature of the site at this time is the large amount of fishing gear, in the form of nets, ropes and buoys, draped across, and entangled around, upstanding material. Non-structural material is also visible on the seabed, including highly perishable organic material such as cables and ropes and portable wooden material such as wooden pulley blocks and gun carriage trucks.

5.3.2 Re-Emergence: 1998–2000 By 1997 an ADU visit to the site recorded a coherent 5.5m length of 10 exposed frames and outer planking in the port quarter area that was around half a metre clear of the seabed (ADU video, 10 September 1997). This suggests that the exposure of the site reported from 1998 onwards was already beginning in 1997. The re-emergence of the site in 1998/1999 was described by Martin Dean as being comparable to the extent seen in 1979 (ADU, 1998: 1), which provides some indication of the potential one- to twoyear lead-in time to maximum exposure. It also indicates that the collapse of the site speculated upon during the 1980s was obviously not as dramatic as thought. It is likely that a process of sediment accretion and burial, on a site that was not properly understood in terms of its extent and character, gave the impression of collapse and loss of structural material. The extensive re-emergence of the site in the late 1990s and Dean’s comment upon this highlights that. The appearance of material across the site with its origins in the 1979 season, both from the Stirling Castle itself and detritus from the work undertaken on it, further serves to illustrate the ADU observation that sediment levels, and corresponding exposure, were similar in 1999 and 1979. Additionally, much of the fishing equipment tangled around the stern area of the ship was assessed by the ADU as having been in place since the late 1970s. A feature of the video record of the site from the 1999 work is the extensive remains of plastic pipe on the site, consistently described by the ADU as ‘old airlift pipes’ (ADU video, 23 June 1999), again presumably from the 1979 season of work.

By 1999 the considerable re-exposure of the wreck reported by the licensee was confirmed by the ADU. Their video record from June 1999 shows the extent of the exposure; it was possibile for a diver to swim a significant distance under the port quarter in the direction of the keel. At the same time, the material in this area shown on the 1979 plan as having collapsed outboard, probably from the poop or quarter deck, was rediscovered (Figure 5.4).

Work on the site in 1998–2000 was characterised by the measured recording of the site by the licensee in conjunction with an extensive video record taken by the licensee and the ADU. Consequently, analysis of the extent of material present when the site was re-exposed is much easier and founded on much greater certainty than the assessment of the 1979 exposure. The video archive also clearly indicates one major effect of exposure on the wooden structural elements of the vessel. Biological

Figure 5.4. Exposed planking that has collapsed outboard of the port quarter, probably from the poop deck or quarter deck. Observed in 1979 and relocated and recorded in 1999 (image courtesy of SeaDive).

61

Whitewright This indicated its survival within the sediment that had covered the wreck previously and the fact that sediment levels in that area had returned to those seen in 1979. In situ observation by the ADU (ADU video, 10 June 1999) of the visible gun-ports on the port quarter noted that the material was not freshly exposed and showed signs of weed growth and some degradation by marine organisms, further corroborating the view that the sediment levels and overall exposure were similar to those seen in 1979. The video record also indicates that there were no further gun-ports exposed below those visible on the port quarter, confirming that the latter are the lowest in the hull and are therefore representative of the main-deck ports. The absence of surviving structure forming the sides of the vessel above that row of ports illustrates that they form the highest coherent vessel structure surviving at the stern/ port quarter of the vessel in 1999. A caveat to this is a comment by Martin Dean (ADU video, 30 June 1999) that the stern/port quarter, although preserved, is far more complicated and jumbled than when he visited the site in 1979. During the same dive he also noted that the port quarter hull structure had undergone some outward collapse since 1979. Its position on the 1999 site plan, relative to the suspected centreline of the vessel, suggests that this outward collapse is c.1m on a horizontal plane.

Figure 5.5. Recently exposed deck planking in the midships area of the site in June 1999. The exposed timber at the centre-top of the image is a barrel stave (image ADU archive, © Crown copyright. Historic England Archive).

of material is not horizontally uniform, but has a peak towards the centreline of the ship. The presence of deck planks overlaying deck beams was observed by the ADU in June and July 1999 in several areas along the rough centreline of the ship (Figure 5.5). Although these decking features cannot be confirmed as the main or upper deck, the nature of the surviving outer hull structure at the bow and stern strongly suggests that they form structure from the main deck.

Depth observations at the rudder by the ADU in June 1999 recorded a vertical difference of 4m (compared to 1.5m a year earlier) from the tiller notch at the rudder-top to the seabed at the stern of the vessel (ADU video, 10 June 1999). This distance was confirmed as 3.2m of exposure during the measured recording of the rudder in early July. Taking Endsor’s (2009) reconstruction of the Lenox as a guide, with a rudder height of c.30 feet (9.1m), this would suggest that in 1999 c.5–6m of hull structure lay below the seabed at the stern of the vessel, placing the seabed at roughly the level of the orlop deck. A large amount of structural material was present forward of the transom, in the area of the vessel’s gunroom, although this was not fully recorded until later work by WA. Dive video from 1999 also highlights that on the starboard quarter, despite several investigations by the ADU, along with the licensee’s recording, there is no surviving/exposed material forward of the transom area and the displaced transom timber/ deck beam that lies to the north of the site. In contrast to the extensive remains of the vessel’s transom, stern-post and rudder, to the north-west of this along the line of the starboard side the site trends rapidly to featureless seafloor with no indication of any archaeological remains.

5.3.2.3 Bow Area The bow of the Stirling Castle was subjected to several dives by the ADU in July 1999. These recorded a significant area of upstanding, exposed hull structure with a curved north-west–south-east alignment along the forward port side of the vessel. This structure extends for about 15m and is accurately recorded on the siteplan produced by the licensee during that season of work. From the perspective of understanding the extent of exposed material, two critical features were recorded. The first was an in situ lead scupper and the second a complete gun-port and surrounding elements of hull structure standing c.2m clear of the seabed. The latter comprises the forward end of the chain-plate shelf that anchors the foremast shrouds on the port side. The gunport set directly underneath this is therefore probably port No. 2 on the main deck on the port side of the ship. On this basis the scupper, which is situated above and to the east of the gun-port, can be identified as one of those serving the upper deck in the same area of the vessel. The displacement of this structure from the centreline of the vessel, relative to the stern-post, indicates that in 1999 the side of the ship in this area had collapsed outward by c.4.5m on a horizontal plane from where it might have been expected to be.

5.3.2.2 Amidships Further forward the sides of the vessel are seemingly much less coherent, or at least less exposed. However, there is a rise in the seabed of around 4.5m from the stern to a position amidships and about one-third distance from the bow of the ship. This would place much of the main deck of the ship beneath the seabed, in the centre of the vessel. Outward collapse of the sides and the generally moundlike nature of the site indicates that the preservation

A similar pattern of outward collapse for the starboard bow of the vessel is confirmed by the structure recorded 62

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 by the licensee and his team, in conjunction with ADU records, in 1999. There, coherent hull structure runs for around 25m from the stem-post, around the bow and along the side of the vessel. This exposed structure was later recorded by WA in their work on the site (below). In some places only framing timbers were visible, but in others gun-ports and planking were also observed. As on the port side, no gun-ports are visible below the observed ones, indicating that the latter are main-deck gun-ports. Some evidence of the foremast chains located above these ports was also observed, matching the extent of exposed material seen on the port side. The ADU’s interpretation during diving was that the two sides of the vessel had split open at the bow, leaving the stem-post in situ, and had fallen away to either side. This is seemingly confirmed by the fact that the starboard structure had also collapsed outward by c.4.5m, mirroring the collapse of the port side. Finally, some indication of the potential amount of hull surviving beneath the seabed in the bow area is given by observations of the top of a wale, located below the gunports, at the level of the seabed. As with the stern area discussed above, comparison of this structural feature with Endsor’s (2009) reconstruction of the Lenox indicates that c.6m of material was contained beneath the seabed in the bow area in 1999. Finally, in 1999 (ADU video, 9 July 1999) a large amount of material was observed scattered on the seabed forward of the stem-post. This constitutes the remains of the beak-head structure, noted on the 1979 site-plan, and which appears to have survived exposure in 1979 to be buried and re-exposed in 1999.

Figure 5.6. Organic remains of a pulley block and rope from the main-deck cannon, carriage and related gun-tackle that was exposed and subsequently raised in 2000 (image courtesy of SeaDive).

observed gun-ports and associated hull structure, indicates that the main-deck ports were exposed in 1999, with limited surviving hull structure above them. The seabed itself is considerably lower and was in the general vicinity of the top of the wale, beneath the main-deck gun-ports. Amidships, the vessel appears to have remained largely buried in 1999, although with protruding elements of main-deck beams and deck planks in some places. The centre of the wreck mound was around 4.5m higher than the seabed at the base of the stern-post and rudder, the latter of which was still in an intact state, including the tiller notch at its upper end. Like the forward end of the ship, the port quarter was exposed at the level of the main-deck ports, although in a more disarticulated state than when exposed in 1979.

5.3.2.4 Summary It is clear from the personal accounts of those who worked on the site, and their record of that work, that the Stirling Castle underwent a significant period of exposure between 1998 and 2000. The nature of the material visible on the seabed indicates that this was similar to when the site was discovered in 1979. A highly visible record of that work, in the form of ID tags, airlift pipes and other material left on site at the end of the work, was re-exposed by 1999. Despite the exposure of the site, very few surface artefacts were present in 1999, perhaps indicating that seabed levels were not quite at the same depth as in 1979. By 2000, however, more well-preserved artefacts are present (Figure 5.6), indicating the ongoing nature of the exposure, as reported by the licensee and the ADU (section 3.1.2). However, the apparent collapse of the main deck in the stern of the ship onto the orlop deck in 2000 is difficult to confirm from the available video archive.

The extent to which the surviving hull of the ship retained its overall structural integrity is on the face of it difficult to assess. However, comparison between the hull structure recorded in 1999 and the original dimensions of the ship suggest that the hull had been subject to outward collapse of c.1m at the port quarter and c.4–5m on both the port and starboard bow, at the level of the main deck. This process of collapse seems to have entailed the sides of the vessel falling away from the centreline, leaving the stern-post and stem-post in largely their original positions. In 1999 a recorded angle of the latter of 20 degrees from the vertical to port gave a further indication of the effect of the outward collapse of the port quarter on the remaining material along the centreline of the ship. What is slightly unclear from the extent of material exposed so far is whether this collapse originated from the keel or lowest elements of the ship, or was facilitated by a break further up the hull, for instance at the level of the orlop deck. The eventual extent of exposure at the stern of the vessel (below) suggests that the latter scenario is unlikely and that the sides of the ship have collapsed outwards from the keel, aided by the fracturing of deck beams visible along the centreline of the main deck.

An assessment of the extent of the vertical exposure of the original vessel, on the basis of the observations taken between 1998 and 2000, is provided in Figure 5.2. Material visible in the forward part of the site in 1998 included bricks and a large copper cauldron, presumed to be from the hearth, and a wooden capstan bar lodged under a jumble of other wreckage. Such material would be expected to be present on the upper deck in this area of the ship. Its presence, along with the disposition of the 63

Whitewright 5.3.3 Reburial: 2002–2009

is potentially because WA were building on the 1999 site plan, but principally because from 2006 the survey method was founded upon ground-truthing anomalies identified through the high-resolution multibeam surveys. The process of locating and inter-locating features across the site was therefore dramatically speeded up and the accuracy greatly increased. The effectiveness of this approach is well illustrated in the bow area of the site, where, despite sediment accretion, the WA survey was still able to locate, position and record 16 iron guns, the brick hearth, two anchors and a selection of structural timbers (WA, 2006: 10). This methodology was also able to successfully identify and locate a line of guns in the midships area of the site, roughly aligned along the port side of the vessel. Comparison of the different site-plans suggest that some of these are the same as those annotated onto the 1979 site-plan but absent from the 1999/2000 plan. As with the bow area, sediment accretion should not be confused with the total reburial of all wreck material.

The ongoing reburial of the Stirling Castle by sediment accretion on the site by 2009 should not be seen as a uniform process. Instead, the records contained within the archive outline a process of ongoing exposure in different areas of the site at different times, coupled with reburial in other areas, before the more general reburial recorded by 2009. Overall, this process can be summarised within the 2002–2009 period as gradual sediment accretion in the bow and midships area of the site, observed and reported by the licensee from 2002 and WA from 2006, while at the same time, the stern area of the vessel continued to be exposed as a result of net sediment loss, before also becoming subject to prolonged sediment accretion. Fortunately, the contrasting processes occurring at either end of the site are well documented through the geophysical archive, and particularly the multibeam surveys (sections 4.2.4 and 4.4). Review of that material within chapter four has highlighted the cause of reburial as the westward migration of the channel in which the site is situated. It is possible to state that at the end of the period covered by this publication sediment had accumulated across the site by 2–3m compared with the point of maximum exposure and was predicted to continue to accumulate. But, as hinted above, the complexity of this process dictates that a review of the extent of physical degradation and loss of any exposed material from 2002 must be undertaken in order to complete the narrative of the surviving seabed archive and to provide an overview of the material still surviving in situ at the end of the period concerned.

In the midships area the same survey process allowed the identification and location of the ends of a number of exposed timbers and timber ends that follow the fore-and-aft centreline of the vessel and which have been interpreted as deck beams (WA, 2007: 7–9). These timbers are noted as either being at an angle of 45° (No. 2282) or as appearing to be at such an angle owing to the exposed length and vertical height recorded. The burial of these to the north, and their break line roughly along the centreline of the vessel, may therefore provide further evidence of the overall process of outward collapse of both sides of the ship to their current location and the associated downward collapse of the decks between them to the level of the main deck.

A video record of the site taken by the ADU in July and August 2002 confirmed the observation of Bob Peacock in his licensee report for that year: namely, the accumulation of sediment towards the bow and the loss of sediment towards the stern. The bow had much less exposed material and the seabed itself was comprised of much finer sediment than that recorded in previous visits. As a result of this, the preservation of material in the bow area of the site can probably be considered to be broadly similar to that observed and recorded in 1999/2000, before sediment accretion processes began to take place. The caveat to such a statement is that the loss of some artefactual material did occur, as a result of exposure and degradation, prior to reburial. The best example of that process is probably the log reel that was exposed in 1999 but subsequently lost before it could be recovered. WA note (2006: 6) that by 2003 reburial extended to around 70 per cent of the site, with only the stern area still subject to exposure and scouring. The seabed around the stern was recorded by WA in 2006 as being scoured to the full depth of the surviving hull and, as a result, such structural remains were deemed to be very unstable (WA, 2006: 11).

The overall integrity and coherence of the site across the period from 2002 to 2009 is mixed. As noted above, the accretion of sediment in the bow area suggests that the surviving archaeological material located in that area is broadly similar to that observed and recorded in 1999/2000. Correlation of the site plans from 1999 and 2009 suggest that there has been little further outward collapse in the sides of the vessel’s hull at the bow, which is to be expected given the increase in sediment surrounding and supporting it. In this regard, the bow area of the vessel is likely to continue to contain material from the upper deck that has collapsed downwards onto the main deck. The hearth area observed by WA in 2006 is the prime example of this. Beneath the surviving main deck, the orlop deck and forward lower areas of the vessel are likely to remain in a highly preserved state, with little potential for exposure since the loss of the vessel in 1703. This scenario is broadly similar in the midships area of the vessel, where the continued presence of a series of guns, assumed to be from the main deck, has been observed since 1979. As with the bow area, although there has been outward collapse of hull structure, material relating to the orlop deck and hold, including hull structure, seems highly likely to continue to be in a well-preserved state.

The final site plan (see Figure 5.1) from the 2002–2009 period is initially misleading. Despite relating to a period when sediment levels were increasing, the plan shows a greater amount of ship structure and associated artefacts present, especially within the outline of the vessel. This 64

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 As indicated above, the stern area and port quarter of the ship represents a very different scenario. Extensive sediment loss, including scouring around the sternpost continued the processes of exposure and resultant degradation of organic material witnessed from 1998. This is perhaps best indicated by the loss of the rudder section above the preserved transom timber after 2003. Comparison of the uppermost transom timber recorded in 1999 with that recorded by WA in 2006 indicates that the timber present in 1999 was also lost. The fact that the rudder was previously preserved to its full extent, up to the tiller notch, only serves to highlight the severity of this loss of material. Even with a truncated rudder, the distance from its top to the seabed was 6m between 2003 and 2006 (WA, 2006: 17). When compared with the 4.5m maximum height between the complete rudder top and the seabed recorded in 1999 the scour at the stern of the vessel is dramatically apparent. Likewise, the rapid sediment accretion in the same area is illustrated by a reduction in the same measurement to 1.4m by August 2007 (WA, 2007: 9). Allowing for the loss of the rudder top, this probably places the seabed in roughly the same place as in 1999/2000. Further damage to the stern took place between August 2007 and August 2008 (WA, 2009a: 7–8), during which time the surviving exposed part of the rudder was completely lost and the stern-post re-aligned to the north-east. Transom timbers associated with the sternpost and recorded during previous seasons of work were also deemed to be lost. The cause of such dramatic damage was not clear at the time of recording, but was considered (WA, 2009c: 8) to be the result of winter storms in late 2007 or impact from fishing equipment. The presence of large amounts of snagged fishing gear has been a feature of archaeological surveys on the site since at least the mid-1990s.

of the nature of remains recorded in 1979, that outward hull collapse in the bow (Area 1) had probably taken place prior to discovery of the site in that year (section 5.3.1). This may have been due either to the length of exposure prior to discovery or to previous exposure. The latter is considered the most likely, given the collapse of material from the decks above the main deck, onto the main deck, allowing the ‘as-found’ situation encountered in 1979. Only through reference to the ongoing work on the site, in particular during the period of significant re-exposure around 1999, is it possible to gain a longer view of these processes. In particular, comparison of all survey records strongly suggests that the bow area underwent further subsequent outward collapse between 1979 and reexposure in 1999 to reach the position recorded between 1999 and 2009, relative to the projected centreline of the vessel. Multibeam surveys indicate that no further similar collapse had occurred by the time the bow area was substantially reburied by 2009. Moving aft, the midships (Area 2) of the vessel has undergone generally less exposure as a result of sediment loss. However, as a result it is less well understood because it has been subject to a less intensive survey regime. Despite this, it seems reasonable to conclude that the outward collapse of the hull is broadly similar to that recorded in the bow area of the site. The stern (Area 3) of the vessel, including the transom, stern-post, rudder and port quarter, has, at various times, undergone extensive, prolonged and significant exposure. Although this area of the vessel appears to have been relatively intact upon discovery in 1979, it had undergone outward collapse by 1999. Continued sediment loss and scour around the stern area led to further outward collapse of the port quarter, in conjunction with the loss of the upper portion of the rudder and probably one of the surviving transom timbers, after 2003. This was followed by the total loss of the rudder and further damage to the stern-post by 2008. The stern area of the vessel can be considered to have been the most exposed area of the site for the longest period of time, as a result of which it has suffered the highest level of degradation.

In conjunction with the loss of material at the stern of the vessel, the port quarter underwent further outward collapse as a result of the loss of supporting sediment. This collapse is certainly over and above that commented upon by Martin Dean in 1999, itself in comparison with the situation that he observed in 1979. Although subject to sediment accretion and reburial by 2009, it is clear that the stern of the vessel had seen the permanent loss of a significant amount of structural material, and with it physical integrity, between 2002 and 2009. This is in contrast to the bow of the vessel and it is probable that such loss at the stern has in turn resulted in a reduction in the coherence of the artefact assemblage in that area during the same period.

On the basis of that summary of the exposure of the vessel’s hull, and taking into consideration the nature and extent of the material that has survived and been recorded at various times, an account of the probable extent, integrity and coherence of the surviving elements of the ship can be offered on a deck-by-deck basis. The highest levels of the ship – the poop, quarter deck and forecastle – seem likely to have been severely damaged during the initial loss of the vessel, following its grounding in 1703. However, the preservation of material from these locations through to recovery in 1979 indicates that further remains of similar material may still surround the site, or be preserved within lower deck levels. The upper deck of the Stirling Castle collapsed onto the main deck presumably at the juncture when the sides of the vessel began their outward collapse, concluded to be prior to 1979. This is evidenced by the

5.4 Discussion Having rationalised the various seasons of work undertaken on the site, a number of points can be summarised in relation to the probable formation processes that the site has undergone between 1979 and 2009. Firstly, consideration must be given to the question of the collapse of the hull, including the extent of such collapse and the chronological sequencing of it. It seems likely, on the basis 65

Whitewright presence of material from the upper deck, such as the hearth and related equipment, on the main deck. The highest level of structural material on the site, as currently understood, relates to the main deck of the vessel. Gunports, guns and other material relating to this structure have been observed and recorded since 1979. Material from this level of the ship is certainly not complete and has undergone degradation and loss since 1979, the stern area and port quarter suffering the greatest impact in this regard because of the lengthy exposure period. However, the ongoing recording of material relating to this deck of the vessel on the surface of the site serves to illustrate the ongoing preservation of the main deck of the vessel and material associated with its operation. Below that, the condition of the orlop deck and hold are unknown because they have never been exposed. The nature of the material observed since 1979, in conjunction with the understanding of seabed levels and ship structure reached since 2002, strongly suggest that both those levels of the ship have the potential to be significantly intact, in a highly coherent state and containing enormous quantities of in situ archaeological material. This situation was summarised by WA (2007: 11) as follows: ‘[a] ‘very large late seventeenth century naval artefact assemblage is believed to exist in situ. Almost the entire content of the orlop deck and hold may be present. Preservation is likely to be excellent where not previously exposed.’ It will be clear from the narrative presented above that the Stirling Castle is not a complete vessel, in the manner of the early seventeenth century Swedish Vasa. However, it is doubtful, given the trauma of the loss of the ship, that it was ever as complete as is generally recounted. Certainly, the available evidence points to hull collapse and material loss having taken place prior to the discovery of the site in 1979. Likewise, there can be no doubt that archaeological material has been lost from the site in the intervening years. Similarly, the understanding of the site morphology set out in chapter four highlights the likelihood that future instances will occur in which sediment cover is reduced, archaeological deposits become exposed, degradation processes are accelerated and structural and/or artefact material is lost. But, as this chapter has attempted to demonstrate, the site should not been seen as one that is archaeologically exhausted, with little new material to be discovered. Instead, the site of the Stirling Castle is probably still one of the largest, most coherent British naval assemblages within the archaeological record from this period. Significant, extensive parts of the vessel survive in situ with an equally large amount of well-preserved artefacts. To echo comments made elsewhere in relation to the contemporary remains of the Northumberland (Pascoe and Peacock, 2015: 142), it is critical that planning and preparation is undertaken to ensure that the full potential of the surviving material is reached, rather than lost, when the Stirling Castle is re-exposed at a future date.

66

6 Stirling Castle: Ship Structure and Shipbuilding Julian Whitewright In addition to the artefact assemblage (chapters seven to eleven), the in situ archaeological remains of the Stirling Castle represent a significant quantity of hull structure, as outlined in chapter five. That chapter highlighted the potential extent of material still surviving at the site, or which has been observed, since the initial discovery of the site in 1979. This chapter builds upon that account by aiming to describe the wooden structural remains of the ship itself from a maritime archaeological perspective. However, the nature of the work that has been undertaken on the site (chapter three) is not always easy to reconcile with the requirements of a written account detailing the technical construction features of the ship. As will be clear by now, the earliest season of work was primarily concerned with the recovery of artefacts, with subsequent periods of activity focusing largely on attempting to record and monitor the overall extent of the preserved material. Within that work, however, detailed recording of the surviving ship structure in discrete areas has been completed through SeaDive/Licensee work during 1999/2000 and more recently by WA between 2006 and 2009. Records from these phases of activity, together with more limited measured recording undertaken by the ADU, have been utilised to produce the following account.

remains, with debris from the upper deck, forecastle and quarter deck collapsed upon it. Beneath this, the orlop deck and hold are assumed to survive, while parts of the transom and stern quarters survive at the eastern end of the site. Since its discovery, the seabed remains of the vessel have been categorised into three general areas (see also section 5.3):

With that in mind, the following narrative is structured around the ordered description of the vessels’ framing, planking and fastening components as far as they can be gleaned from the information available within the 1979– 2009 archive. Comparative material, both historical and archaeological, is integrated at each stage of this process as a means to contextualise and understand the Stirling Castle material in the most direct manner. Foremost among these is the ongoing archaeological work on the Northumberland (Pascoe and Peacock, 2015) and the historical reconstruction of the Lenox (Endsor, 2009). As such, it should be remembered that the material presented here is certainly not complete or exhaustive, and simply serves to offer a snapshot of the exposed, accessible elements of the vessel and a glimpse of the potential of the material still preserved at the site. Future recording of the position and fastening of the structural hull elements is likely to enable the creation of a more detailed and specific account of the construction techniques.

Chapter five addressed these as discrete areas, as observed and recorded during the different seasons of excavation. However, it is also instructive to offer a combined site plan (Figure 6.1) that draws together all of the main seasons of work, and recorded features within each season, from 1979 to 2009. Once features that are repeatedly present across several seasons are combined, in conjunction with those from more recent seasons (1999 onwards) whose position is recorded with some confidence, a picture of the site emerges. This allows the locations of the various examples of structural recording discussed below to be positioned within the overall plan. Strikingly, it also highlights the fact that the spatial integrity of major features, such as the port-side alignment of guns that still roughly retain their original spacing, has remained relatively intact. This offers further confirmation of the main conclusion of chapter five that, beneath the exposed surface of the wreck-mound, the preservation of material is potentially very high, from both spatial and material perspectives.

6.1 Overview

6.1.1 Construction Summary

The hull remains of the Stirling Castle comprise a long low mound lying on a west-east axis with the bow to the west (WA, 2006: 10). The remains of the vessel’s main deck is considered to be the highest level of coherent archaeological

In keeping with the normal building practices for large carvel-built warships of the Royal Navy in the late seventeenth and early eighteenth centuries, the Stirling Castle was probably built around a series of established

• Area 1: Relating to the bow of the vessel, containing the remains of the forward part of the hull, including remnants of the upper, main and orlop decks, the hold, the stem-post and part of the forecastle. Elements of the beak-head and bowsprit may lie outboard of this area to the west of the main wreck mound. • Area 2: The midships area of the vessel, where exposed archaeological remains are primarily related to the upper and main deck. • Area 3: The stern of the vessel, including the surviving elements of the stern-post and transom timbers. Elements of the port quarter have collapsed outboard, including parts of the vessel’s side from the quarter deck or poop deck. The starboard quarter is still largely buried while the last vestiges of the rudder, intact in a complete state until 2003 but now thought to be lost, may be buried at the stern of the vessel.

67

Whitewright

Figure 6.1. Overall site-plan of the Stirling Castle based on combining all seasons of survey and recording. It must be emphasised that a large proportion of the material shown has since been lost, such as the rudder and wing transom. The plan shows that some key features, such as the two alignments along the fore-and-aft axis of the surviving guns, have been recorded, when viewed across multiple seasons of work. Features such as cannon and anchors have been given a standardised symbol and should not be used to differentiate actual types or sizes.

main framing pieces that were bolted to the keel of the vessel (for an extended account see Endsor, 2009: 31–70) after the stem-post and stern-post had also been set in place. These main-frames were not made from a single piece, but were comprised of timbers spanning the keel and defining the bottom of the ship, known as floors, to which were attached a series of increasingly vertical timbers, termed futtocks, that formed and defined the sides of the vessel (Figure 6.2). The lowest of these futtocks, and the first to be placed, were known as the first futtocks (or lower futtocks), followed by the second futtocks (or middle futtocks), third futtocks (or upper futtocks) and finally a series of top-timbers that demarcated the highest part of the ship’s hull. Where possible, framing timbers were created following the natural curve and grain of the wood, rather than simply cutting shapes from whatever material was available. Further framing timbers were placed within the gaps between the main-frames and attached to the outer planking. In keeping with their role, these additional pieces were known as filling-frames and could be fitted only after the outer planking (below) was in place.

Figure 6.2. Simplified diagram to highlight the component parts of an assembled main-frame for a ship such as the Stirling Castle.

Reinforcement to the ship’s hull was provided by a keelson, laid along the centreline and bolted through the floors into the keel, with further strength added by stringers, risers, and shelves and clamps laid fore-and-aft and bolted to the framing. Transversely, deck beams were secured to the shelves and clamps and spanned across the vessel, tying the two sides of each main-frame together and providing the structure upon which the decks could be laid. A critical element in this, and other parts of the vessel, were wooden

knees, often shaped in a right angle following the natural curve and grain of the wood; these were used to provide a horizontal connection (a lodging knee) between deck beam and beam-shelf and a vertical connection (a hanging knee) between the beam and the frame beneath it. To this integrated framing structure (Figure 6.3) two layers of planking were added: outer planking, around 68

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 wing-transom, into which vertical timbers called fashion pieces were rebated that supported and framed the stern gun-ports, ornate windows and decorative stern of the vessel. At the bow, a structure known as the beak-head was carried forwards from the stem-post and provided a means to secure the vessel’s bowsprit, a location for the figurehead and a suitable location for the heads (toilets) used by the common sailors when at sea. 6.2 Framing As noted above, the nature of the work done on the site, coupled with the extent of structural material that remains buried, dictates that provision of a full account of the framing timbers of the Stirling Castle is impossible. Instead, focus is given here to the more sporadic recording of framing components undertaken at different times in different parts of the site and which in a general sense all relate to the vessel’s futtock timbers. 6.2.1 Futtocks Exposed futtocks were recorded by WA in 2006 (WA, 2006: 16–17) on the starboard side of the ship’s bow (Area 1, WA context 2267). Two sample futtock heads (upper ends) measuring 170–200mm sided and 170–230mm moulded were recorded. A run of exposed frames, assumed to be futtocks, at an unconfirmed location on the port side of the vessel was recorded by the ADU in 1997 (ADU Video, 10 September 1997). The line of frames had almost no visible gap between the frames, had a sided dimension of c.300mm and was orientated along an east-west axis. This probably indicates a location amidships, at the very aft end of the observed bow timbers or the forward end of the port quarter structure. The latter of these locations seems most likely, based on the site-plan produced during the 1999/2000 work. Finally, as part of the 1999/2000 survey, a section of exposed framing with in situ outer and ceiling planking was recorded at the stern-port quarter of the ship, where it had fallen partially outboard (Figure 6.4). The work recorded two futtocks with a wooden chock

Figure 6.3. Simplified diagram to highlight the main component parts of a large carvel-built ship such as the Stirling Castle. For clarity only one deck is shown, while other elements are included only to indicate their location, rather than every instance of their occurrence.

the outside of the vessel’s frames to form a watertight skin; and ceiling planking, laid onto the interior face of the frames. Both of these layers of planks offered a further means to bind the hull structure together along a longitudinal axis. This was enhanced in some places by thicker bands of outer planking, known as wales, that ran around the sides of the ship and added further strength. Additional reinforcement timbers, known as riders, were then secured through the ceiling planking and into the floors and futtocks. The majority of the fastenings used in a wooden ship such as the Stirling Castle were also wooden and were called treenails. These were used throughout the vessel to secure planks to floors and futtocks and deck planking to deck beams and ledges, to provide a lateral fastening between the adjacent elements of the main-frames, and a myriad of other functions. After being driven into place, the exposed heads of the treenails were tightened to prevent them coming loose and to waterproof them. This was often done with a wooden wedge, or through one or more deep cuts filled with caulking, and called a caulking cut. Not all fastenings were of wooden treenails, and riveted iron bolts up to 1¼ inches in diameter were used in selected areas of the ship, to fasten the floor timbers to the keel, for example, or in the scarphs between keel sections. At the stern, large transverse elements called transom timbers spanned the gap between the two sides of the vessel and were rebated to be attached to the stern-post at their centre point. The widest of these was called the

Figure 6.4. Section of exposed, collapsing hull at the stern port-quarter, drawn up from the survey originally recorded in July 1999 by Alison Shanck and Bill Utley.

69

Whitewright between them. The maximum moulded dimension of the upper futtock was at its lower end (above the chock), and was 220mm, while at the upper, broken end the moulded dimension was 160mm. The lower futtocks had a maximum moulded dimension of 190mm. No sided dimensions were recorded for either timber. The chock between the futtocks measured 195mm moulded, 180mm along its inboard (widest) face and 100mm on its outboard (narrowest) face.

The sizes of such filling-frames are un-documented, but were probably of variable dimensions according to the space required to be filled. Another filling-frame head seems likely to have been visually recorded by the ADU in July 1999 in the vicinity of station S9 on the starboard bow of the vessel, adjacent (to the west) to context 2267 recorded by WA. The ADU record (ADU video, 9 July 1999) shows the head of a frame timber, assumed by the diver to be a futtock in the audio transcript, with a completely flat, sawn end, between two adjacent futtocks. The timber is well-squared with the tree-rings centred in the cross section. The significant analytical point of this timber is the absence of any form of scarph-joint between the in situ futtock and the futtock that would be expected to have continued the run of the frame in a vertical direction. Recent archaeological work on the site of the Northumberland (Pascoe and Peacock, 2015: 137–38) has indicated that by the time of the loss of that ship and the Stirling Castle in 1703 it was common for framing timbers to be joined together with a triangular wooden chock that was recessed into the head of the lower timber and the heel of the upper one (see also Endsor, 2009: 44–45). The absence of any such chock-scarph on the timber head observed by the ADU indicates that it may have been the head of one of the filling-frames, which were simply attached to the outer planking upon installation, and which had no requirement to be scarphed to adjacent filling-frames in the manner of the floors and futtocks.

The analysis outlined in chapter five indicates that all of the timbers just discussed are likely to be related to the side of the ship at the level of the main deck. As such, their most likely initial identification is as third futtocks or the upper portion of second futtocks. Records of the scantling lists for the Thirty Ships and later comparable vessels (reproduced by Endsor, 2009: 226–29) indicate that both second and third futtocks had the same sided dimension of 12¼ inches (311mm). The moulded dimension, in the language of the day termed ‘in and out at ye gundeck’ is not completely clear for each class of futtock, but is likely to be c.10 inches (254mm). The same figures are given in the 1691 building contract for the 70gun Yarmouth (launched 1694) and by master shipwright Daniel Furzer, later responsible for rebuilding the Stirling Castle at Deptford in 1699. The visible consistency across sources perhaps indicates the relative stability in scantling dimensions used by builders in the decades after the Thirty Ships were constructed. The dimensions of the futtocks recorded by the ADU offer a clear correlation, as might be expected, with the historically recorded dimensions and provides further confirmation of the analysis (chapter five) regarding the extent of the exposed structural material present at the site. However, the sizes of the futtock heads recorded by WA are much smaller than might have been predicted. Assuming that relatively well-preserved timbers were recorded, rather than heavily degraded ones, the timbers observed in 2006 may come from a higher point in the vessel, indicating the extent of the outward collapse and the exposure of material from that collapse. In this regard, the historically attested sided dimension at the head of the top-timbers of 8 inches (203mm) is instructive and offers a close match to those recorded in 2006 (170– 200mm). However, the moulded dimensions of the same top-timbers ‘at the gunwale’ is given in the historical records as 4½ inches (114mm) which presents a further large discrepancy with the seabed material recorded in 2006 (170–230mm). However, despite their location at opposite ends of the ship there is a reasonable correlation between the WA records and the 1999 section drawing, the latter of which offers a moulded dimension of 160– 220mm. The most likely rationalisation of these different measurements (ADU, WA, 1999 survey) at the same constructional level of the vessel is that the ADU had recorded parts of the main-frames of the ships – hence the close correlation with the historical documentation – while the other two surveys had recorded parts of the filling-frames that were laid between the main-frames.

Further evidence for the use of chocks does survive from the Stirling Castle and was described above in relation to the section of the stern-port quarter recorded in 1999 (Figure 6.4). As observed by Pascoe and Peacock (2015: 138), the nature of that chock, including its shape and lack of fastening to the adjacent frame timbers, suggests that it is unlikely to be a formal chock-scarph and may just be a temporary repair. Alternatively, if the adjacent recorded frame timbers are in fact filling-frames, as suggested above, then the loose chock may just serve as a means to continue the process of filling the space between the main-frames in a manner that imitates the wider practice of using chockscarphs to join more structurally important timbers. The recording of the floor–futtock arrangements, including whether they were joined laterally in overlapped pairs or relied on their fastenings to the outer and ceiling planking for structural integrity, should be a priority for future seabed recording. This information would enable a more detailed understanding of the specific hull construction and help determine whether there was freestanding construction of completed frames prior to adding the planking, or whether the planking was added to each run of frame elements in an alternating process as the structure progressed upwards. The former method requires a more mathematical approach, requiring calculations to complete the pre-shaped run of completed frames, whereas the application of planking at each stage of framing is based in traditions of shipbuilding by eye and experience. 70

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 6.2.2 Deck Beams Work by WA on the site in 2006 and 2007 observed and recorded a number of structural elements in the centre and east of Area 2 that are likely to represent deck beams (WA, 2006: 17, fig. 3; 2007: 6–8, fig. 3). These form a coherent pattern aligned roughly north-south and appearing out of the sand overburden lying along the northern part of the site. Their disposition fits a pattern that might be expected for the vessel’s primary deck beams and associated intervening ledges, the latter notched into carlings that ran fore-and-aft between deck beams. Meanwhile, their orientation is noted as either being at an angle of 45 degrees or as appearing to be at such an angle owing to the exposed length and vertical height recorded. The burial of these to the north, and their break line along the general centreline of the vessel, may therefore indicate that the starboard midships and stern quarter of the ship have collapsed outward. The measurement of all these timbers, as recorded on the 2006/2007 site plan and accompanying description, is summarised in Table 6.1, and an image of an exposed timber, probably a ledge, in the same area and taken during good conditions in 1999 is shown in Figure 6.5.

dimensions shown in the later site-plan, presumably reflecting the move from a sketched record of the area to a fully measured one.

Comparison of the measurements of these features, in particular the sided dimension, with the scantling lists for vessels of the period (see Endsor, 2009: 226–27) suggests that beams and the accompanying ledges from different decks can potentially be identified. Although some allowance has to be made for the clear variation in dimensions, this is to be expected given their sometimes degraded state and the context of breakage that has occurred across the vessel during its loss and subsequent site-formation processes. There are also differences in the illustrated dimensions of features 2031–2034 from 2006 to 2007 (WA, 2006: fig. 3; 2007: fig. 3), with smaller

The most westerly timber (2029) is much larger than all of the others and its sided dimension suggests it is the remains of a main-deck beam. To the east, the sided dimensions of features 2031 and 2032 are most compatible with upper-deck beams, although they are somewhat on the small side. Critically, the distance between these timbers is c.1.45m, although allowance must be made for the fact that the damage they have sustained means that they now lie slightly out of their original alignment. This distance fits very closely with that given in the historical sources for upper-deck beams, and is both smaller and larger than the value given for

Figure 6.5. Newly exposed deck structure recorded by SeaDive in 1999. The timber in the centre of the image is a ledge, with deck planking landing upon it. A treenail is preserved in the top of the ledge (image courtesy of SeaDive).

Table 6.1. Collated measurements of deck beams and ledges. Feature numbers are set out in order from bow to stern, as recorded between 2006 and 2009. Comparable dimensions are taken from those reproduced by Endsor (2009: 226–27). Feature No. 2029 2031 2032 (2288) 2308a 2308b 2033(2287) 2034 2283 Un-numbered Un-numbered 2307

Sided (mm) 400 260 240 190 170 160 270 340 270 320

–

Comparable dimensions Sided

Main-deck beam 412–420mm (16¼"–16½") 381–387mm (15"–15¼") 2.44m (8')

Main-deck ledge 127–152mm (5"–6") 101–127mm (4"–5") 229mm(9")

Moulded Space between

Moulded (mm) – – 280 – – 190 –

Exposed length (m) 1.80 1.32 1.48 0.80 0.72 0.77 2.72 2.56 1.30

–

1.44 Upper-deck beam 305–330mm (12"–13") 254–279mm (10"–11") 1.52m (5')

71

Interpretation Main-deck beam Upper-deck beam Upper-deck beam Upper-deck ledge Upper-deck ledge Upper-deck beam Upper-deck beam Upper-deck beam Upper-deck beam Upper-deck beam Upper-deck beam Upper-deck ledge 101–114mm (4"–4½") 89–101mm (3½"–4") 229mm (9")

Quarter-deck beam 203–244mm (8"–9½") 152mm (6") 0.6m (2')

Whitewright main-deck and quarter-deck beams respectively. The same sizing and distribution pattern can be observed for features 2034 and 2283, further to the east, and so the same interpretation as upper-deck beams can be applied. Between these pairs of beams lie features 2308a, 2308b and 2287, whose recorded location, dimensions and spacing allow an interpretation as upper-deck ledges, although their sided dimension and the spacing between them is larger than given for ledges from any area of the ship. Further refinement in the interpretation of these timbers can be reached through reference to the 2005 RASSE multibeam survey of the site (Bates et al., 2011: fig. 5c; Figure 4.19), which clearly shows that feature 2033/2287 is of the same size and orientation as features 2031, 2032, 2034 and 2283, indicating that its proper interpretation is also as part of the series of upper-deck beams. Further east again, the same patterning of beams as for 2034 and 2283 can be seen for two un-numbered features and feature 2307.

original location of this timber within the vessel is difficult to ascertain, but its sheer size, and in particular its length, indicates that it is not part of the stern framing assemblage. Its eventual location outboard of the stern of the vessel probably prevents an interpretation as a main-deck beam, given that the main deck in the part of the vessel over 9m in breadth is probably largely preserved in place. One remaining possibility, given the overall length, is that timber S15 is one of the aftermost vertical elements of the vessel’s rudder (discussed below), which would have been comprised of several vertically set pieces of timber that would have broken away during the grounding and loss of the ship. The stern-post itself was clearly visible in 1999 and formed part of the recording undertaken at the stern of the vessel in that season and the following year. At that time it leant to port (the south) at an angle of 20 degrees, perhaps reflecting the overall angle of heel of the ship, with dimensions of 400–480mm breadth (side to side) and 630–970mm fore and aft. Survey in July 1999 recorded that the aft-most face of the stern-post was 300mm clear of the stern planking, itself rebated into the stern-post (below). Historical sources (see Endsor, 2009: 226–29) provide a range of comparative dimensions for the head of the stern-post, within which the breadths span from 1 foot 10 inches (558mm) to 2 feet 2 inches (660mm). Meanwhile, the fore-and-aft measurement at the head of the stern-post is consistently given as 1 foot 10 inches (558mm). The clear variation in the breadth between the historically listed measurements and the seabed record is difficult to reconcile, but the problems are perhaps hinted at by the variation in the recorded measurements themselves, which in conducting the present analysis are assumed to be the result of degradation over time, different levels of marine growth, different survey techniques and so on.

The analysis of the recorded features in this area of the site, interpreted as deck beams and ledges primarily from the upper deck of the ship, serves to remind us of the process of collapse of material from the upper parts of the ship down to the level of the main deck. It also highlights three further important and inter-linked factors. The first of these is the usefulness of correlating the recorded archaeological features with available historical documentation as a means to further elucidate/ confirm the likely identification – if not immediately obvious – of the archaeological remains. Second is the reminder that this process provides that the historical documentation, where it may exist, should not necessarily be taken completely at face value given the ability of the archaeological record to highlight the subtle differences between how ships were actually built and with how they were supposed to have been built. The differences between the supposed, and actual, dimensions and spacing of the deck beams extant on the Stirling Castle, even allowing for the impact of the wrecking process and subsequent degradation, are likely to be an example of this. Third is the apparent differences between the archaeological and historical sources (although small) for this element of the vessel structure, which may be some of the first evidence that can be identified for the actual extent of the 1699 rebuild of the vessel and its resulting impact on the ship’s structure.

The in situ uppermost transom timber was recorded by the licensee’s team in 1999/2000 and was 600–650mm sided, 450mm moulded and c.7.8m in overall length. It had a rebate to receive the stern-post cut into the centre of the moulded face that measured c.750mm across and c.280mm deep. The length of this timber correlates very closely with the historically given breadth of the main deck at the transom of 7.77m (25½ feet). That, in conjunction with its position relative to the rudder and top of the stern-post, strongly suggests that it is the wing transom, the largest of the transom timbers fitted across the stern of the vessel. This interpretation is further strengthened by the presence of a series of rebates cut into the upper face of this timber and observed (although not measured) in ADU video records of the structure. These rebates would have taken the heels of the vertically set counter-timbers. While the location and width of the wing transom provide a close correlation with the historical material, its other dimensions do not. For example, the building contract for the Yarmouth specifies a thickness of 14 inches (355mm) for the wing transom, a much smaller dimension than the one used on the Stirling

6.2.3 Stern Structure The well-preserved stern of the ship has been a feature of the commentary of the site since its discovery in 1979. Although subject to significant loss of material after 2003, the main wooden elements of the stern structure have been relatively well recorded. The largest timber in this is located on the seabed to the northeast of the stern-post, orientated north-south and labelled S15 at its northern end when recorded in the 1999/2000 survey. This timber is c.9m in overall length and the face lying uppermost is 500mm across; no thickness has been recorded. The 72

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 above, assumed to be the port quarter. This planking was attached to the exposed frames by wooden treenails (not measured), and had a thickness of 70mm (ADU video, 10 September 1997). A section of the hull at the very stern of the port quarter, recorded in 1999 (Figure 6.4), gives a range of plank thicknesses between 24mm (for one plank) and 90–100mm (for the majority of the others). The outer planking on the transom of the ship was recorded in 1999/2000 in the area directly adjacent to the stem-post and rudder. There, the run of the planks around the hull of the ship brings them into a rebate in the stern-post at an angle of nearly 45 degrees. The average width of these planks was 300mm and the upper-most surviving plank had a thickness of 80–100mm. ADU and SeaDive records from 1999/2000 observed the use of wooden treenails in the outer planking, two per plank at each fastening location, the heads of which were tightened with triangular caulking-cuts (Figure 6.6).

Castle. The extent of the preserved remains mean that it is not possible to state whether or not the head of the rudder (see below) was carried through the counter, or remained outside it. However, it seems likely, given the conclusions below regarding the steering mechanism of the ship, that the rudder-head extended through the counter in the manner that Lavery (1987: 13) suggests was common during the final years of the seventeenth century. On the inside of the transom planking (below) recorded in 1999/2000 was a series of frames set on the same foreand-aft alignment as the stern-post, to which the transom planking was treenailed. The heads of these timbers finished below the wing transom and were not attached to it, although treenails surviving in their upper faces suggest that at least one of the transoms, or false transoms, originally secured them vertically. The dimensions of these timbers are recorded only on the large-scale site plan for 2000, which gives measurements of 300–330mm sided and 270–240mm moulded.

An area of exposed deck timber was recorded in Area 1 in conjunction with the deck beams noted above. Context 2308 is a group of timbers interpreted as planking that lies a short distance to the north of the two recorded deck beams (2287 and 2288). The thicknesses of these planks are not recorded, but their widths are given by WA as 320mm, 350mm and 380mm and they are considered likely to be the decking that overlaid feature 2033/2287 (WA, 2007: 8). The identity of this area of decking is considered uncertain by WA, but it probably formed part of the gun deck, or collapsed upper deck (WA, 2006: 8). The discussion above with regard to the deck-beam remains indicates that the latter interpretation is the most plausible. As with the fastening of outer planking to frames, the deck planks were fastened to the deck structure with wooden treenails, a group of which were observed by the ADU in Area 2 of the site. Here a single treenail was used per plank, per ledge, and each treenail was tightened with a cross-cut caulking-cut (Figure 6.7) rather than the triangular caulking cuts seen on the outer planks.

The damage to the rudder after 2003 (section 5.3.3) also resulted in the loss of the wing transom. However, survey by WA in 2007 recorded another transom timber (Feature 2268) measuring 540mm sided and 200mm moulded. It spanned the stern of the vessel to a maximum length of 4.9m and was centrally rebated to facilitate attachment to the stern-post. Licensee Bob Peacock considered that this timber represented one of the lower transom timbers, possibly the third transom (WA, 2007: 8). While the timber is certainly one of the lower transoms, confirmation of its exact identification as the third transom is difficult, in part because of further discrepancy between the in situ archaeological remains and the relevant historical material. These indicate an expected moulded dimension for all the lower transoms of between 10 inches and 12 inches (254mm–305mm), much larger than the actual dimensions in place on the vessel. The seabed material therefore provides a potential indication of over-sizing timbers for one element (the wing transom) and under-sized ones in an adjacent related piece (the third transom). As before, this alerts us to the potential variation between the historically stated building practices and the ship actually built. Alternatively, if the validity of the historical documentation is adhered to as a principle, then the discrepancy again serves to highlight a potential area where the impact of the 1699 rebuild is visible in the preserved seabed structure. In either case, the archaeological details of the transom timbers are significant for our wider understanding of the Stirling Castle and post-medieval wooden shipwrecks more generally. 6.3 Planking and Fastening Outer planking with a thickness of 90mm was observed on the starboard side of the bow (Area 1) in 2006 (WA, 2006: 17). Outer planking was also recorded by the ADU in 1997 in the same area of the port side as the run of frames noted

Figure 6.6. Outer planking was secured with wooden treenails. The example pictured here had its head tightened with a triangular caulking cut (image courtesy of SeaDive).

73

Whitewright 6.4 Rudder and Steering Mechanism As intimated above, the ship’s rudder was largely complete and in situ until 2003 (Figure 6.8), and as a result was recorded in some detail in early July 1999 (Figure 6.9) as part of the wider recording work in the stern area of the vessel. At that time the top 3.2m of the rudder, out of a probable total height of c.9m (section 5.3.2), stood clear of the seabed, compared with a maximum exposure of 6m between 2003 and 2006 (section 5.3.3). Two surviving but heavily concreted pintle and gudgeon fittings 2.1m apart

Figure 6.7. Deck planking was secured with wooden treenails. The example pictured here had its head tightened with a crosscut caulking cut (image ADU archive, © Crown copyright. Historic England Archive).

Comparison of the observations of plank dimensions with historical sources is difficult because of the variety of different plank thicknesses used, depending on whether the planking in question was regular outer planking (4 inches (102mm)), the lower or upper wale (10 inches (254mm)), the ‘thick stuff’ either side of the wales or the chain-wales (both 6 inches (152mm)), the plank between the chain-wales (3 inches (76mm)) or the outer-planking of the quarter deck (2 inches (51mm)). Assuming a situation of perfect preservation, outer planking could therefore vary between c.50mm and c.250mm, depending on its exact location on the side of the vessel. Rationalisation of this variety with the imperfect vertical location of the recorded archaeological material is therefore extremely difficult and again illustrates the problematic nature of combining the archaeological and historical material. But, for the most part, where dimensions have been recorded in the archive they seem to relate to the sizes that would be expected from regular outer planking. Turning to the deck planking, the use of one treenail per plank per ledge is directly at odds with the historical sources (Endsor, 2009: 227), which specify two treenails in each ledge for each deck plank. Finally, the caulking cuts used to tighten the treenails, in all their forms, are becoming increasingly well documented in archaeological sources relating to English/British shipbuilding and for later periods it is possible to relate potentially different building locations with the type of tightening method; caulking cuts in southern England, multiple pegs in north-east England or a single plug in Scotland (for discussion see Whitewright et al., 2014: 100–103). Although this is far less certain for the seventeenth and early eighteenth centuries, the tightening methods observed on the Stirling Castle are certainly consistent with those seen on the roughly contemporary Dartmouth (Martin, 1978: 47–48, fig. 21), which was also originally built and later refitted in the south and south-east of England, respectively.

Figure 6.8. Artistic impression by Ted Westhead of the surviving rudder when still preserved to its full height in 2000 (image courtesy of SeaDive).

Figure 6.9. The rudder assemblage of the Stirling Castle, drawn up from the survey originally recorded in July 1999 by Bill Utley and Tom Berkey.

74

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 and with an estimated centre-to-centre spacing of 2.5m were visible in 1999. The preserved extent of the main rudder timber, termed the ‘main-piece’ (see Lavery, 1987: 11; Divers, 2004: 65–71) was 500mm fore-and-aft and 400mm side-to-side, while the upper 830mm of the rudder body projected aft a further 320mm. Beneath this a narrow central ridge measuring 140–260mm fore-and-aft and 120–170mm side-to-side projected aft of the main-piece. The forward vertical edge of this, when recorded in 1999, retained what appeared to be a caulking seam. The head of the rudder is on top of the main-piece, and at its base it is 500mm square, projects 500mm above the top of the stern-post and has a semi-circular side elevation. A foreand-aft notch 270mm wide and 500mm deep is cut into the rudder-head to seat the tiller. Beneath this, a transverse hole with a diameter of 130mm is located in the upper part of the main rudder timber, 500mm below the top. It is clear from reference to Figure 6.9 that the shape of the upper portion of the rudder surviving and recorded in 1999 is different from the standard shape that might be expected (see Lavery, 1987: 10–11), especially in regard to the overhang of the rudder head on its aft side. The most obvious explanation for this is that some of the component pieces of the rudder that were attached to the main-piece were displaced during the wrecking process. The tapered aft edge of the main-piece and the presence of a caulkingseam is further evidence for this. Bearing that in mind, it may be suggested that timber S15 (above), recorded lying outboard of the vessel on the starboard side, is one such component.

Figure 6.11. Interpretation of the steering system of the Stirling Castle based on the arrangement described by William Sutherland (1717: 87) and the observed fixed blocks (image courtesy of Richard Endsor).

The published interpretation of these (Endsor, 2004) is that they form part of the blocks and tackles that would have led lines from the vessel’s steering wheel downward to the deck below, through the angled sheaves (fixed to a location on the vessel’s centreline), out to the side of the vessel, through the fixed single block and from there to the tiller (Figure 6.11). The critical importance of Endsor’s interpretation, which there is no reason to doubt, is that it represents the earliest evidence for the use of the steering wheel, replacing the previously used and less effective whipstaff.

Two further elements related to the steering mechanism of the vessel have been recorded and published (see Endsor, 2004). These are two examples of fixed sheave blocks found in the gunroom of the ship (the furthest aft space on the main deck). The first of these, comprising a single sheave, was observed in situ by the licensee and was still attached to the surviving ship’s structure. The second was found loose in the gunroom area and contained two individual angled sheaves (Figure 6.10).

In this regard, the direct archaeological evidence from the Stirling Castle supersedes the historical and ship model evidence, the latter of which is of uncertain date (Lavery, 1987: 19; Endsor, 2004: 95), and indicates the use of the steering wheel for helming the Stirling Castle at the time of its loss in 1703. It can be speculated that retrofitting such integral equipment as the steering mechanism between commissioning in 1701 and loss in 1703 is unlikely, especially given the wider absence of historical documents ordering such work (Lavery, 1987: 19; Endsor, 2004: 95), and a reasonable conclusion would be that the Stirling Castle was fitted with a steering wheel during its 1699 rebuild, the evidence for which remained preserved in the archaeological record following its loss in 1703. 6.5 Discussion While the bulk of the discussion that can be undertaken in regard to the structural remains of the Stirling Castle has been set out above, some further general observations can be made. These revolve largely around the correlation between the archaeological remains, as recorded during

Figure 6.10. Fixed double block recorded in 2002 and related to the steering mechanism (image courtesy of Richard Endsor).

75

Whitewright seabed survey, and the historical documentation relating to Royal Navy building from the period. At a wider scale the seabed remains offer a reminder of the primacy of the archaeological record as the physical remains of what was actually built, rather than what may have been conceived, planned or thought to have been built (see also Muckelroy, 1978: 215). In this regard, the capacity of archaeological sites such as the Stirling Castle to offer reference points for technological developments, such as the introduction of the steering wheel or the application of chock-scarphs is clear. The value of such archaeological reference points lies in the fact that they are less muddied by dating issues reliant on stylistic interpretation or inclusion at a modelbuilders scale. In a related way, it is possible to affirm the use of such archaeological/historical comparison through the process of being able to identify previously unclear structural elements on the basis of their recorded dimensions. But, in conjunction with that same process, we are again reminded of the pitfalls of such comparison where the as-built material obviously differs from the historical sources. Such differences certainly complicate the process at a micro-level, while serving to highlight the capacity for individual variation in the components between vessels of the same class and overall dimensions. In some ways, such variation should not be a surprise, given the stated differences in preferred component dimensions from builder to builder and, as a result from shipyard to shipyard, in the surviving historical documents. The result of this is really a reminder of the dangers of generalising a wider building tradition on the basis of one archaeological example or a single historical document. In reality, both sources show similar small nuances in the outward results of converting timber resources into highly complex ship structures within the wider context of late seventeenth/ early eighteenth century England. At the heart of these differences, it must be assumed, lie the master shipwrights who oversaw the building of each individual vessel and who had the capacity to stamp their own opinion as to ‘best practice’ onto the ships that they were producing.

76

7 Organic Material Kathryn Dagless, Angela Middleton, Gill Campbell, Polydora Baker, Simon Mays and Mariangela Vitolo The conditions for the preservation of archaeological material on the site afforded by rapid deposition and burial in deep sediment, as noted in chapter three, have meant that a large quantity of organic material has been observed in situ (Figure 7.1) and in some cases recovered during the various seasons of work. This includes the early years of work when the preference (section 3.1.1) was the recovery of non-organic material because of the costs and perceived complications of dealing with organic material. Despite the preference for non-organic material, a large number of organic artefacts have been recovered from the site over the course of the work undertaken on it. This chapter considers the organic artefact assemblage recovered from the Stirling Castle present in the 1979–2009 archive, which forms an important element of the record of the site.

discussion of the assemblage as a whole and comparisons with artefact assemblages from contemporary wrecks and terrestrial sites of the period (section 7.3). 7.2 Recovered Artefacts by Material Each class of material is described in general terms, after which follows a summary artefact catalogue, including the condition and dimensions: Length (L), Width (W), Height (H), Diameter (D), Thickness (T). In all cases the ID number is that given to the artefact during the assessment stage of the current project; any external ID number is provided in parenthesis. Artefacts that were selected for and underwent investigative conservation are indicated by (IC). 7.2.1 Plant Remains

7.1 Methodology

041 [397] Grape seeds. Vitis vinifera L. One organic sample recovered from the site contained 369 fully mature grape seeds, 154 immature or undeveloped grape seeds and 72 fragments of grape skin. These could be the remains of dried grapes or raisins. However, the recovery of vine stalks from the same sample suggests the presence of fresh grapes. Other plant remains (probably straw) recovered in the sample might belong to the container in which the grapes were found, although no information concerning the container survives. It is also possible that this material represents the remains of straw used as packing material to stop the grapes from spoiling (cf. Stainforth, 1979: 150).

The further study, analysis and review of the artefacts included here was undertaken using the existing detailed catalogue and photographs. Closer study of the photographic catalogue enabled some of the unassigned objects to be identified and classified. In addition to this, a programme of investigative conservation was undertaken by Angela Middleton of Historic England using X-radiography and microscopy to provide further information on the selected artefacts. Analysis of the plant remains was provided by Gill Campbell and Mariangela Vitolo, also of Historic England. A catalogue of organic artefacts by their material grouping is presented below (section 7.2), sub-divided where appropriate according to function. This is followed by a

312 [218.19] Nutmeg (Myristica cf. fragrans Houtt.). Whole inner seed, partially decayed. Originally identified as a hazelnut with a pierced shell, but later confirmed by G Campbell to be nutmeg. 320 [225] Pine Cone. Pinus pinaster Ait (maritime pine). This pine is native to the coasts of the central and western Mediterranean (Mitchell, 1978: 169) and was first grown in Britain by John Gerard in 1596 (Aiton, 1789: 367). Maritime pine was one of the conifers planted during the great period of British landscape gardening (Hadfield, 1976) and was among the seed supplied by Alexander Eddie in 1765–67 as part of the landscaping of Oatlands Park, Surrey (Harvey in Symes, 1981). One of the crew of the Stirling Castle may have either picked up the cone for decoration – they are widely used as Christmas ornaments today – or collected the cone for its seed either for their own or an associate’s use. Although such acquisition could have taken place in the UK, a Mediterranean origin is more likely owing to the very limited distribution of the tree in the UK at the time.

Figure 7.1. Organic material, in this case wooden blocks and associated cordage, in situ in 2000 (image courtesy of SeaDive).

77

Dagless, Middleton, Campbell, Baker, Mays and Vitolo 237 [151.1] (IC) Handle from a cut-throat razor, wider end broken off obliquely across one of two rivet holes containing rusted remains of iron fasteners. Surface rounded on one side. Good condition. L 155mm; W 12–15mm.

7.2.2 Bamboo A single artefact was identified through microscopic study as being made from bamboo. Although originally identified as an apple corer, its size and the presence of a similar object from HMS Invincible (Bingeman, 2010) suggests a more probable identification as a barrel spout.

238 [151.2] (IC) Handle from a cut-throat razor, wider end suddenly narrows to a point; narrow end bevelled. Two rivet holes with rusted remains of iron fasteners. Another attempted hole at the front. Surface rounded on one side. Good condition. L 146mm; W 12–16mm.

280 [199] (IC) Bamboo object, one end cut at an acute angle, the other tapered, pared and smoothed with some brown deposits suggesting the depth of insertion into a barrel. Good condition. L 152mm.

239 [151.3] (IC) Handle from a cut-throat razor, both ends broken; evidence of two rivet holes. Surface rounded on one side and flat on the other. Good condition. L 125mm; W 11–17mm.

7.2.3 Bone and Ivory Worked Bone and Ivory

285 [201.4] (IC) Socketed ivory handle, slight crack at socket end. Presence of Schreger lines indicates that the material is ivory (Penniman, 1952: pl. 1; Krzyszkowska, 1990: 35). L 88mm; W 25mm.

The worked bone and ivory artefacts comprised personal items including a comb (ID 313), toothbrush (ID 314), razor handles and handles for small knives or other implements. Both the comb (Figure 7.2) and toothbrush (Figure 7.3) were recovered in a good condition, with many of the teeth still intact in the comb and the copper wire that would have held the bristles remaining in the head of the toothbrush.

286 [201.5] Socketed bone handle, slight crack at socket end. L 76mm; W 23mm. 287 [201.6] (IC) Socketed handle, cracked at both ends. Made from ivory. Fair condition. L 81mm; W 20mm.

Four riveted bone handles were identified, into which an iron pin would have been inserted to provide a pivot for the blade. Although no blades survive within them, comparison with other riveted handles of the period suggest that they are probably razor handles. Four other whittle-tanged handles were identified as having been carved from sheep or goat metapodials (identified by K Dagless). The pricker (ID 326) was originally described as a ‘small wooden spike’ when first recovered.

288 [201.7] (IC) Socketed handle. Schreger lines indicate manufacture from ivory. Good condition. L 198mm; W 56mm. 313 [219] (IC) Comb, possibly a nit comb. Schreger lines indicate manufacture from ivory. Dark in colour, with a lighter patch showing grain pattern; teeth intact but some show signs of discolouration. The comb has been broken and repaired, with teeth missing in the area of the break. Good condition. L 102mm (Figure 7.2). 314 [220] (IC) Toothbrush, made from bone and copperalloy wire. The bone has been scored on one side (the back) to create channels for the wire to run in. The wire is in one piece, looping in and out of the holes, while it never emerges on the front, where the bristles would have been. The bristles have not survived. However, most of the holes have remains of soil and possibly fibre remains inside. Handle tapers to a point. Good condition. L 89mm (Figure 7.3). 326 [228] (IC) Pricker, possibly made from ivory, but certainly an osseous material. Heavily worked and polished. The wider end has a cavity, which contains lead at the mouth, as confirmed through X-radiography and XRF (X-ray Fluorescence). Good condition. L 95mm.

Figure 7.2. Single-sided comb, ID 313.

327 [229] (IC) Handle, made of osseous material, possibly ivory, and with a flattened hexagonal cross section. It is heavily worked and highly polished, which makes identification difficult. There is a cavity at the slimmer end, that still seems to contain remains of a high-density metal, possibly remains of the tang, as indicated on the

Figure 7.3. Toothbrush, ID 314.

78

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 X-ray. The other end possibly broke off. A hole was made and a pointed tip inserted. The tip is made of the same material as the handle itself. Good condition. L 90mm.

with transversal butchery marks, probably chops, in the middle of the anterior shaft, and three slightly diagonal finer cuts/chops on the proximal anterior shaft. The bone is adult in size but the proximal end is unfused, indicating that the bone is from a subadult animal younger than 2.5– 3.5 years at death (Silver 1969; Popkin et al 2012).

477 [369] (IC) Handle, made of osseous material, probably ivory. Eight-sided faceted handle with a circular void through its length. Good condition. L 85mm; W 11– 19mm; H 13–24mm; D(void) 5mm.

257 [177] Complete left fallow deer (Dama dama) antler, with a first (brow tine), a second tine and palm with seven preserved points (including spellers). Length of the main beam is approximately 450–500mm (after Haltenorth and Trense 1956). There is a square shaped hole (c. 10mm), near the base of the palm but the origin is uncertain, whether pre- or post-depositional. The coronet of the antler appears to be fully formed, and it is after this stage that the antler is shed.

478 [370] (IC) Handle, ivory. Conical form. Fair condition. L 82mm; D 13–23mm. Animal Bones The small assemblage of five animal bones includes a complete fallow deer antler, a medium size mammal femur, two large mammal vertebrae and one large mammal rib. The bones were not examined directly, but from photographs and finds sheets by Polydora Baker. All but the fallow deer show evidence of butchery, suggesting that they derive from food rather than from drowned livestock, although it not possible to say if they formed part of the ship’s stores. The origin of the antler is uncertain; it may be an intrusive find or perhaps carried onboard as part of cabin furnishings, for example as a curiosity or decorative item. A sixth specimen is from an indeterminate large mammal (animal/human).

534 [1480] Large mammal (non-human/human) partial cranium, with fused frontals. Human Bones As four fifths of the ship’s crew are reported to have perished it is likely these may be some of the victims of the sinking. The human bones were not examined directly, but from photographs and finds sheets by Simon Mays. The assemblage includes the following specimens: 1090 [071] Left humerus, almost complete, missing proximal end with distal end fused. Adult size.

The butchered cattle and pig or sheep/goat bones are typical of victuals carried onboard late medieval and postmedieval vessels, including casked pork and beef, and fresh mutton from live animal cargo (Armitage 2004; Coy and Hamilton-Dyer 2005; Wijngaarden-Bakker 1987). Split vertebrae and chopped ribs were carried onboard the Mary Rose, including portions of cattle backbone (Coy and Hamilton-Dyer 2005). Consumption of fresh mutton is more typical of later vessels, for example on the Earl of Abergavenney, and possibly reserved for the captain’s table (Armitage 2004).

533 [1480] Left femur, almost complete, broken at proximal and distal ends. Adult size. 7.2.4 Cork The two cork objects recovered were identified as bottle or jar stoppers. One (ID 500) had a maximum diameter of 20mm, while the other (ID 601) was larger, with a diameter of 49mm, indicating the varying sizes of container that the ship would have been carrying.

087 [074] Fragment of large mammal rib, possibly cattle (Bos taurus). Includes the articulation (partly broken) and proximal third of the blade. The medial side of the blade was chopped through/broken and shows four additional transverse (slightly diagonal) cutmarks, though in a slightly different direction to the chop/break.

500 [322] Vessel cork. Good condition. H 27mm; D 18–20mm. 601 [326] Fragment of vessel cork. Poor condition. H 24mm; D 49mm.

088 [073] Cervical vertebra of a large mammal, possibly cattle (Bos taurus). The bone may have been butchered, cleaved through on a slight medio-lateral diagonal, removing most of the centrum and part of the right caudal articular process and the right transverse process.

7.2.5 Fibre Eight fragments of rope were recovered. Two were in a fragile condition but could be identified as three-stranded ropes with strand diameters c.20mm. These three strands were then wrapped together in narrow ropes of c.25mm (ID 068) and c.35mm (ID 067) diameter. Another four are in much better condition. Three are again three-stranded, of similar thickness to the above but without the smaller rope encircling them; the fourth is single-stranded (all ID 521). There is also a fragment of rope with part of a nail attached (ID 264).

089 [072] Thoracic vertebra of a large mammal, possibly cattle (Bos taurus), cleaved through (probably butchered) just to the right of mid-line, removing part of right side of centrum and vertebral arch and right transverse process. 145 [016] Femur of a medium size mammal (cf pig, Sus scrofa, or possibly sheep/goat, Ovis aries/Capra hircus), 79

Dagless, Middleton, Campbell, Baker, Mays and Vitolo 067 [94] Fragile rope fragment. Three-strand rope each strand diameter c.20mm. Right hand lay. L 155mm; D 41mm. Wrapped with twine/ smaller rope of diameter approx 35mm.

484 [382] Coconut shell, possibly used as a drinking or other vessel. Base and ‘lid’ present. Good condition. L 120mm. 7.2.7 Leather

068 [93] Fragile rope fragment. Three-strand rope, right hand lay. L 137mm; D 41mm. Wrapped with narrow twine/ rope, diameter approx 25mm.

Forty-four fragments of leather were recorded, including items of clothing (hats, shoes and part of a waistcoat), a belt/strap and six book covers. A number of smaller fragments were recovered that could not be identified with any certainty. Three large fragments of what is believed to be a waistcoat were identified (ID 584). No fastenings were present but button holes were evident.

101 [60] Fragment of rope, three strands. May be associated with block (ID 59). L 150mm; D 23.5mm. 264 [183] Fragment of rope with part of a nail attached. L 127mm.

Positively identified fragments of clothing include two leather hats (ID 366 and 613). These both carry a flat crown that is straight on one side and slants at an angle on the other, presumably denoting the front/back. The crown is made up of five parts: two side panels, a strip of leather, a leather thong and the top. The side panels overlap at the front (?), and are secured with a straight stitch. The side panels do not join at the back (?) and the gap is backed with a strip of leather from the inside. Leather thong crosslacing attaches the strip to either side of the side panels (four holes on ID 366 and three holes on ID 613). The side panels are attached to the brim using straight stitch. The top piece of the crown is attached to the side panels in ‘closed seam with flesh/ edge stitching’ (Goubitz, 2001: 184, 188). Both examples have a wide brim made of one piece, with perforated holes around the outer edge (c.4mm diameter). A leather thong is wound around the perforated holes in a whip stitch technique. The inner edge is slashed to allow it to be folded up for attachment to the crown. Lately the provenance of these hats has been called into question (Middleton et al, forthcoming).

521 [323] Four rope fragments. Three with three strands with left hand lay, laid up right handed. One with single strand and left hand lay. L 165–480mm; D 24–38mm. 7.2.6 Gourd and Coconut Four cups/vessels were identified in the catalogue as being made of gourd or coconut, and ranged from 56 to 102mm in diameter. These had originally been identified as coconut halves, and similar items in the HMS Invincible assemblage were also identified as such. 289 [202.1] Dark brown spherical gourd with hole cut in top, possibly for use as a drinking cup. Some small barnacles adhere. Good condition. D 102mm. 290 [202.2] Cup formed from half a gourd (Figure 7.4). Surface black and smooth. Good condition. D 76mm. 332 [234] Drinking cup made from a gourd, some cracks on rim (Figure 7.5). Dark inside, paler exterior. H 56mm; D 70mm.

366 [439] (IC) Wide-brimmed leather hat with repair to crown and one patch (Figure 7.6). Stitching evident between crown and base and evenly spaced holes around the brim. Diameter of brim c.250mm. Condition assessment: despite its overall good condition the hat is mouldy and has a greasy appearance. The top piece of the crown is probably a replacement, as it is not as greasy as the rest of the hat but shows signs of migrating grease on the outer edges, where it joins the side panels. All stitching has failed and been replaced with white or orange thread in running stitch. Thread impressions of the straight stitch can be observed, however, especially at the back, where the side panels overlap. The thread was Z-twisted. Some lacing around the outer edge of the brim is frayed or missing. The cross-lacing at the back is not complete any more. One of the side panels is rather thin and crushes easily. The other side panel is firm. There are areas of loss on the softer side panel, which have been backed and glued in place with replacement leather. There is a circular impression running around the lower edge of the crown, probably stemming from an attachment, such as textile binding. A similar impression can be observed at the front (?) next to the side panel seam. Both impressions appear ribbed.

Figure 7.4. Cup formed from half a gourd, ID 290.

Figure 7.5. Drinking cup made from a gourd, ID 332.

80

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 7.6. Wide-brimmed leather hat with repair to crown and one patch, ID 366.

613 [1110] (IC) Wide-brimmed leather hat (Figure 7.7). Good condition. Stitching evident between crown and base and evenly spaced holes around the brim. Diameter of brim c.250mm. Condition assessment: this leather hat is in good condition. Its surface is greasy and sticky and has attracted quite a bit of dust. It is also mouldy. The hat overall contained its shape very well and is very firm. All stitching has failed and been replaced with white thread in running stitch. The original stitch was probably straight stitch, as can be seen by very faint thread impressions. A lot of the lacing around the outer edge of the brim is frayed or missing. Some of it has been replaced with a new lace.

There are quite a few areas of surface loss where the grain has come away. Identifiable footwear included slipper-type shoes (and fragments) and what appears to be the sole, heel and upper of a boot. Many of these were in good condition. Other fragments included soles and heels together, as well as separate soles and heels. Many of these were in good to fair condition. The heel fragment (ID 017) was formed from at least 12 layers of leather, which were held together by a central peg of wood. Seven larger peg holes surrounded this. The sole and heel found together (ID 003)

Figure 7.7. Wide-brimmed leather hat, ID 613 (©Historic England, drawn by Judith Dobie).

81

Dagless, Middleton, Campbell, Baker, Mays and Vitolo 024 [413] Inner sole from a shoe, formed from two layers of leather, evidence of stitch holes around the margin and nail holes in the heel. Good condition. L 260mm; W 93mm; T 5mm; Heel width 70mm.

were joined together with a single central iron peg and nine wooden ones. The heel was formed from many layers of leather, stitch holes were visible on the upper surface and a groove was cut on the lower surface. The other shoe sole fragments all showed evidence of stitch holes around the margins holding together the layers of leather that formed the sole and fixing holes were evident at the heel end. One sole had traces of tar on its heel end. Stitch holes were noted on the other fragment. The footwear collection would benefit from further analysis/research on shoe types.

025 [412] Inner sole from a shoe, two layers of leather, wear on upper surface. Lower surface scored to take stitching, four nails in the heel and cracked across the heel. Good condition. L 259mm; W 85mm; T 5mm, heel width 70mm. 026 [411] Inner sole from left shoe or boot. Evidence of nail and stitch holes. Evidence of wear, a second layer of leather on lower surface and traces of tar on the heel. Good condition. L 275mm; W 96mm; T 2.4–3mm.

003 [425] Sole and stacked heel. Stitch holes on upper surface and cut groove on lower. Sole and heel riveted with central iron rivet and nine wooden pegs. Good condition. L 275mm; W 92.6mm; Heel length 72.8mm, width 77.8mm, height 39.5mm.

028 [409] Leather sole fragment, stitch holes along worked margins and evidence of square nail holes. L 242mm; W 93mm; T 4.5–5mm.

013 [424] Leather from the sides of two shoe soles with stitch marks. Fair condition. L 550mm; W 3.5mm and L 380mm; W 3–4mm.

423 [1243] Slipper type shoe, mostly complete. Sole, heel and upper all present with stitch marks. Good condition. L 260mm (Figure 7.9).

014 [423] Inner sole formed from three layers of leather (Figure  7.8). Stitch holes around the margin and fixing holes in heel. Fair condition. L 60mm; W 90mm; T 10mm.

514 [G:08] (IC) Shoe fragments, sole and upper no longer attached. Sole cracked and the upper filled with a large concretion, identified by X-radiography to be nails. Poor condition. L 220mm (Figure 7.10).

015 [422] Leather inner sole. Evidence of iron staining around rivet hole in heel. Small fragments of wood and some stitching. Good condition. L 248mm; W 71mm; T 3mm.

524 [331] Three fragments of shoe leather. One upper made from two layers of fine leather with stitching around edge, L 157mm; W 122mm; one layer of leather rolled with stitching holes, L 170mm; W 21mm; one layer of thick leather with single stitching hole L 61.5mm; W 34.5mm. Good condition.

017 [420] Shoe heel formed from at least 12 layers of leather held together by central rivet of wood. Evidence of seven large rivets around the margin. Good condition. L 68.5mm; W 73.5mm; T 26mm. 020 [417] Inner sole from a shoe, two layers of leather with stitch marks around the margin and nail holes in the heel. Good condition. L 253mm; W 70mm; T 3–4mm. 021 [416] Leather sole/inner sole with stitch marks around the margin and nail holes in the heel. Good condition. L 253mm; W 75mm; T 4.5–6mm. 023 [414] Shoe fragment, possibly slipper shaped. Leather has been shaped and gathered with stitch holes around the margin and traces of tar. Fair condition. L 250mm; W 95mm; T 1.5–2.5mm.

Figure 7.8. Inner sole formed from three layers of leather, ID 014.

Figure 7.9. Slipper type shoe, mostly complete, ID 423.

82

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 7.10. Fragmented shoe remains with layered construction; detail revealed through X-ray, ID 514 (X-ray not to scale ©Historic England).

614 [No ID] Four shoe fragments. A) possibly part of a boot, substantial heel (30–40mm high), sole and upper all present. L 250mm. B) heel with nail marks and part of upper attached. C) sole and upper with stitch marks. L200mm. D) sole and upper with stitch marks. L 200mm. All in good condition.

other end of the long piece has a large elongated hole at the slightly pointed end. The shorter piece of leather was inserted into the belt buckle and sat with a hole over the area for the pin. One end of this short piece of leather has been folded over and was originally secured with a couple of stitches. The other end of the short piece is also pointed and has three small holes. The grain of the leather is worn down or obscured by a conservation treatment. The flesh side is laminating off the leather surface. This is probably a musket sling or similar, rather than a conventional belt. See also ID 205, 207 and 208, section 9.3.1. Good condition.

A complete belt with rectangular buckle (ID 424) and a long strap were recovered. Both were in excellent condition. Two other fragments of leather strap were also recovered; these were slightly curved and formed from two layers of leather with evidence of stitching.

481 [384] Long strap. No buckle but folded back on itself at one end and loop present. Excellent condition.

424 [1244] (IC) Complete belt with D-shaped buckle with central bar. The leather itself consists of two long pieces. The longer one is wound around the central bar and secured with two rivets, which have not survived. The

030 [407] Two fragments of leather strap, slightly curved. Stitching present, including fine thread on margins. 83

Dagless, Middleton, Campbell, Baker, Mays and Vitolo

Figure 7.11. Leather book cover, ID 292.

Figure 7.12. Leather book cover with ribbed spine, ID 293.

Formed from two layers of leather. L 365mm; W 25– 29mm; T 3mm.

stamped front and back with rectangle around edges and flower motifs in corners. Good condition. L 127mm; W 102mm.

584 [No ID] Three large fragments, identified as parts of a waistcoat. Front identifiable with button holes. Tattered and torn but still in good condition.

297 [209] Dark leather book cover stamped with intertwining scrollwork pattern and two small brass clasps front and back (Figure 7.15). Good condition. L 102mm; W 76mm.

The six dark leather book covers survived in good condition. The contents are unknown but the books were probably almanacs, prayer books or bibles which were common personal possessions during this period. Some showed faint evidence of decoration in places and centrally positioned clasps were still present on a few.

A leather pump washer was formed from two layers of leather. There was a single central hole with a diameter of 26mm. 019 [418] Pump washer formed from two layers of leather. Fair condition. D 112mm; T 10mm. D (central hole) 26mm.

292 [204] Dark leather book cover, stamped faintly with a circle front and back, plus double line around the edge (Figure 7.11). Good condition. L 127mm; W 178mm.

Thirteen pieces of unassigned leather fragments were recovered from the site. Some had been worked and shaped and also demonstrated evidence of sewing/stitching. It is possible that some of these are also fragments of clothing, but their exact form when complete could not be identified.

293 [205] Dark leather book cover, stamped faintly with double line decoration and leaf motif in corners (Figure 7.12). Four ribs evident on the spine. Good condition. L 152mm; W 222mm. 294 [206] Dark leather book cover. Faint evidence of double line decoration on front near spine and of three spinal ribs. Torn and with holes. Fair condition.

016 [421] Three pieces of worked leather, shaped and cut with stitch holes. Good condition. A) L 61mm; W 109.5mm; T 1.5–2mm. B) L 48.7mm; W 83.4mm; T 4mm. C) L 85.7mm; W 93.5mm; T 21mm.

295 [207] Dark leather book cover with small brass clasp on front and back (Figure 7.13). Vertical creases noted on the spine. Good condition. L 102mm; W 152mm.

018 [419] Four pieces of worked leather, three attached together, largest piece made from two layers, L 245mm; W 114mm; T 1.5mm. Several fragments of strap and other worked leather. Strap dimensions L 126mm; W 17mm; T 3mm.

296 [208] Dark leather book cover, six ribs noted on the spine, which is torn (Figure 7.14). The cover is scored/ 84

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 7.13. Leather book cover with small brass clasp on front and back, ID 295.

Figure 7.14. Leather book cover with ribbed spine, ID 296.

022 [415] Worked leather, cut and pierced with sewing holes. Possible shoe upper. Good condition. L 250mm; W 205mm; T 2.5–3mm.

031 [406] Fragment of worked and shaped leather, evidence of sewing holes. Fair condition. L 156mm; W 137mm; T 1.5–2.5mm.

027 [410] Fragment of worked leather. Good condition. L 185mm; W 137mm; T 4.5–6mm.

032 [405] Large fragment of worked and shaped leather. Evidence of fine sewing holes along all margins, formed from two pieces of leather. Good condition. L 380mm; W 155mm; T 6mm.

029 [408] Worked and shaped leather, evidence of stitch holes on three worked margins and remains of a fourth. Largest fragment L 277mm; W 180mm; T 3mm.

A further interesting item, which was unassigned, and unfortunately had no photograph, was noted in the

Figure 7.15. Leather book cover stamped with intertwining scrollwork pattern, ID 297.

85

Dagless, Middleton, Campbell, Baker, Mays and Vitolo catalogue (ID 078). It is described as a ‘concretion overlaid with decorated leather on one surface and three leather straps’. The decoration is visible and shows a lion rampant, a round seal with shield and the remains of gold lettering. 078 [83] Concretion overlaid with decorated leather on one surface and three leather straps. Decoration visible showing lion rampant, round seal with shield and remains of lettering all in gold. L 310mm; W 140mm; H 160mm. 7.2.8 Textile Seven textile fragments were recovered, comprising a damask square, a pouch and two small fragments of silk, and three fragments of a woven fabric. A round textile button was also found. 121 [40] Finely woven fabric, possibly silk, dark grey, fine, too fragile to record before conservation. No photograph available.

Figure 7.18. Silk damask square, ID 429.

480 [377] Textile button. D c.20mm.

403 [1131] Three fragments of coarse woven fabric, one of which has decoration woven into it that appears to be an anchor (Figure 7.16). Good condition.

585 [No ID] Square silk fragment, ripped. L c.300mm; W 300mm.

404 [No ID] Silk pouch, purse or pocket (Figure 7.17). Drawstring still present.

7.2.9 Tar

429 [1330] Silk damask square, comprising four sewn strips with design of trees and foliage in Chinese manner (Figure 7.18). Holes and stains. L 480mm; W 450mm. Individual strips L 480mm; W c.120mm.

Two lumps of tar were recovered that would have been used for various purposes, including caulking and preserving rope. 167 [10] Two lumps of tar. L 102–114mm; W c. 89mm. 7.2.10 Wood A total of 161 wooden artefacts was recovered, including armaments such as cannon parts (wooden trucks, carriage fragments), musket fragments and sabots; personal and dress items such as a ‘pill box’, buttons and a comb; navigational equipment; utilitarian items such as bowls and a platter, implement handles, a brush and cylindrical case; and various other miscellaneous artefacts that would all have had specific uses on board the ship.

Figure 7.16. Three fragments of coarse woven fabric, ID 403.

7.2.10.1 Containers Two small wooden containers were recovered. One was a round pill box with a screw top, the other was a cylindrical wooden case of barrel construction (five thin wooden staves held by two hoops of withies), together with related circular wooden end pieces (ID 291). 291 [203] (IC) Cylindrical wooden case of barrel construction. Six individual wooden staves bound by two withies. One circular wooden end piece was also preserved. Was found to contain rolled fabric. Wood species analysis indicates that the main body of the case and the bindings are made from Abies or Cedrus

Figure 7.17. Silk pouch, purse or pocket, ID 404.

86

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 Head Pieces

sp. (fir or cedar) (identification by G Campbell), with the former being more likely owing to the colour. L 381mm; D 70mm (Figure 7.19). Objects of a similar size and construction were found on the Invincible wreck, these have been interpreted as writing kits (Bingeman 2010).

659 [WA2242] One of the central parts of the head piece, or lid, of the barrel. Almost complete. Two dowel holes on either side with dowel remains. 660 [WA2246] One of the central planks of a head piece. Slightly damaged; two dowel holes and dowel remains visible on one side.

319 [224] Round wooden pill box with screw top. Over 50 per cent of the rim of the lid is broken off. Lid has screw threads and corresponding threads on the inside of the box. D 35mm.

661 [WA2251] One of the central elements of a head piece. Complete. Two dowel holes with dowel remains on either side. Four inscribed lines on one side, which could be the mark of the maker or owner.

Barrel remains recovered from the site in 2006 included 13 staves and six head or lid pieces, all believed to be made of oak (WA, 2006: 18). They were recovered from a single find spot, but it is not known whether they represent parts of the same barrel. Stave length varies from 720mm to 950mm, with at least three different lengths apparent. The staves are slightly tapered along their length. Chimes (the tapered edge at the end of the stave) and crozes (grooves to receive the barrel head) are apparent. A pronounced curve is apparent on several staves. One of the head pieces, with a length of 0.61m, appears to be full or almost full barrel diameter. The ends of the head pieces are tapered to fit the croze of the staves (Middleton, 2016). Although the parts of more than one barrel appear to be present, all of the sizes suggest small– medium barrels.

662 [WA2252] Almost complete head piece, one end slightly damaged. One of the central planks of a head piece. Two plugged holes on the surfaces and three dowel holes on the sides. 663 [WA2253] One of the central planks of a head piece. Complete with one iron nail in one corner. Bevelled ends. Four dowel holes, two along either side. Two holes have dowel fragments. 664 [WA2255] Complete. One of the outer planks of a head piece. Bevelled edges, two holes with dowel remains on the side (Figure 7.20).

Figure 7.19. Cylindrical wooden case of barrel construction, ID 291.

87

Dagless, Middleton, Campbell, Baker, Mays and Vitolo

Figure 7.20. Barrel head piece, ID 664 (©Historic England, drawn by Chris Evans).

Staves

7.2.10.2 Dress and Personal Accessories

665 [WA2243] Almost complete, one end broken off, two iron nails in complete end.

A small number of wooden dress and personal accessories were recovered including 19 round buttons, each of which was domed on one side and had a single central hole. A wooden comb (ID 063) was also identified. This had broken into two fragments but still had 33 of its teeth remaining (no photograph was available).

666 [WA2244] Slightly damaged on either end; one end has two grooves, other end has a large crack running from the end towards the centre. 667 [WA2245] Slightly damaged at one end; the undamaged end has two grooves.

063 [98] Wooden comb in two fragments, with 33 teeth remaining. Good condition. A) L 72.5mm; W 60mm; T 6mm. B) L 29.5mm; W 55mm; T 5.5–6mm.

668 [WA2247] Slightly damaged at one end; the other end is grooved and bevelled. Two iron nails at the damaged end, and possibly two more.

306 [218.1–9] Nine circular wooden buttons with one central hole, domed on one side. Good condition. D 28mm; T at centre 10mm.

669 [WA2248] Slightly damaged on the ends, one groove in each end.

307 [218.10–11] Two circular wooden buttons, domed on one side, with one central hole, flat surface in friable condition. D 28mm; T at centre 10mm.

670 [WA2249] Possible barrel stave; it is knotty and bent. Damaged on both ends, remains of one iron nail. 671 [WA2250] Slightly damaged on both ends. Two grooves at both ends. A wooden peg at one end.

308 [218.12–13] Two circular wooden buttons, domed on one side, with one central hole. Good condition. D 31mm; T at centre 10mm.

672 [WA2254] Broken at one end; the other end has two grooves.

309 [218.14–16] Three circular wooden buttons domed on one side, with one central hole. Good condition. D 26mm; T at centre 10mm.

673 [WA2256] Damaged at both ends. 674 [WA2257] Damage at both ends and by marine wood borers.

310 [218.17] Circular wooden button, domed on one side, with one central hole. Good condition. D 27mm; T at centre 16mm.

675 [WA2258] Severely damaged. One iron nail remains in the wider end.

311 [218.18] Circular round wooden button, domed on one side, with one central hole. Good condition. D 17mm; T at centre 6mm.

676 [WA2259] Almost complete. One groove at either end. One end is bevelled. The remains of one iron nail near the bevelled end.

630 [120] Domed wooden button; single central fixing hole c.4mm diameter. Good condition. H 10.2mm; D 29mm.

677 [WA2260] Severely damaged. One groove is visible at one end. 88

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 061 [100] Wooden handle with remains of ‘blade’; much detail obscured by concretion, but metal (iron?) blade appears to be folded. Handle turned from close grained wood. Fair condition. L 170mm; D 20–43mm; Handle length 122.5mm, blade width c.34.5mm, thickness 1.5–7mm. 081 [80] Wooden tool handle, one end broken; appears to be a bit and brace handle, eight faces on main part of handle narrow to rough octangle. Straight-grained timber. Fair condition. L 135mm; W 46.5mm; T 25.3mm.

Figure 7.21. Warped and worm-eaten fragment of a wooden board marked in squares, ID 279.

093 [68] Turned wooden handle, circular in cross section. Good condition. H 55mm; void depth 14.5mm in depth. ‘Knob’ D 37.5mm.

7.2.10.3 Entertainment/Leisure A few leisure items were recovered including a fragment of a game board. It had warped and was described as ‘worm-eaten’, but had clearly been marked into squares (ID 279). A single drumstick was also identified (ID 398).

099 [62] Large fragment of turned fine-grained wooden platter. Radius c.175mm. Incised lines. Good condition. L 340mm; H 78mm; T 8.5mm.

279 [198] Warped and worm-eaten fragment of a wooden board marked in squares (Figure 7.21). Probably part of a chess board. Fair condition. L 241mm; W 76mm.

142 [19] Two fragments of fine-grained wooden platter. Incised lines on lower and upper surfaces. Good condition. L 340mm; H 78mm; T 8.5mm; Radius 205mm.

398 [1242] Drumstick. Good condition. L 420mm.

147 [14] Wooden handle, circular in cross section. Wound with wire. Evidence of leather cover. Void for blade rectangular to oval. Located by raised cannon. Good condition. L 101.5mm; D 30–31.5mm; Wire diameter 0.4mm; Void dimensions 9×6.8–16.3×11.5mm.

7.2.10.4 Food Preparation and Consumption Galley equipment relating to food preparation and consumption included complete and fragmented bowls/ dishes and platters. Three were body and rim fragments from a single artefact (ID 525) with carved decoration on the upper surface of two of the pieces that looks like a shield. An unpolished bowl, 355mm across when the fragments were assembled, was stamped with the letters IO (ID 179). A complete smooth, lathe-turned wooden bowl (ID 251) was recovered. Its two raised hoops are separated by a slightly waisted band running mid–high around the outside. Two platters were described in the catalogue, but no photographs were present to study them further. One was a large fragment of a fine-grained turned wooden platter with incised lines (ID 099). Two fragments of another platter (ID 142), also made of a fine-grained wood, were recovered and had been decorated with incised lines on both upper and lower surfaces and an additional incised mark on the lower surface.

149 [12] Wooden handle. Wound with wire. Circular void filled with corrosion products. Much detail obscured by corrosion. Stained black. L 112.4mm; D 27–46.5mm; D of void 2.5mm. D of wire 0.5mm. 179 [97] Unpolished wooden bowl, in three pieces, 355mm across when assembled (Figure 7.22). Stamped with letters I.O. Bowl has a smooth edge. Fair condition. W 356mm. 251 [159] Smooth, lathe-turned wooden bowl (Figure 7.23). Interior dark, exterior lighter with dark patches. Two raised hoops separated by slightly waisted band runs mid– high around outside. Very good condition. H 152mm; D 127mm. 282 [201.1] Socketed wooden handle. Hexagonal cross section. Bulbous end, narrowing towards socket. Good condition. L 106mm; T 31mm.

A number of wooden handles were recovered; a few had small fragments of blades still embedded within them but the blades were missing from the majority. It cannot be said for certain whether these would have been used just as pieces of cutlery or as multi-purpose tools. A turned wooden tool handle with a void to take a square sided tang was decorated with three incised lines at the widest point of the handle and at the tang end (ID 105). Three other wooden handles for whittle-tanged implements (ID 282, 283, 284) were identified, as well as a wooden grip with two brass bands 110mm long and a central hole to take a whittle tang (ID 501). The spigots would have been used to plug holes in barrels or other vessels.

283 [201.2] Socketed wooden handle. Hexagonal cross section. Bulbous end, narrowing towards socket. Good condition. L 106mm; T 30mm. 284 [201.3] Socketed wooden handle. Hexagonal cross section. Bulbous end, narrowing towards socket. Good condition. L 101mm; T 29mm. 490 [304] (IC) Two wooden spigots. Tapered cylinders. One with hole, diameter 6mm, cut through upper part

89

Dagless, Middleton, Campbell, Baker, Mays and Vitolo

Figure 7.22. Unpolished wooden bowl, in three pieces, stamped with letters ‘IO’, ID 179.

artefacts presented here (see section 8.2.3 (glass), 9.4.1 (metal) and 11.2.1 (stone)). The wooden material comes primarily in the form of a number of devices for recording navigational observations (fragmentary remains of a backstaff and cross-staff) or for making related calculations (gunter rules). A number of the items are now housed at the National Maritime Museum, Greenwich (and viewable through their online catalogue), where they are indicated by the secondary catalogue prefix NAV. Further work has been undertaken on the back-staff recovered in 2001 (ID 127), which was published by Bryan Smith (2010). In addition to the back-staff and cross-staff that are confirmed as being present on board, Smith also postulated (2010: 179) the presence of a Davis Quadrant on the vessel, on the basis of some of the smaller more fragmentary pieces recovered, such as ID 654.

Figure 7.23. Smooth, lathe-turned wooden bowl, ID 251.

and a circling band of zig-zag metal. Good condition. A) L 69mm; D 15–20mm. B) L 65mm; D 15.5–20mm.

7.2.10.5 Navigation and Measurement

127 [34] Back-staff rod. Broken, square in cross section. Scale on all faces: face 1, 0/90 to 30/60; face 2, 30/60 to 50/40; face 3, 60/30 to 70/20; face 4, included 80/10. One end face shows bevelling and marks. Other end face broken and worn. Location near S17, 2001 season. A full description, analysis and illustration is given by Smith (2010). Good condition. L 266mm; W(faces 1 and 3) 16.5mm; W(faces 2 and 4) 15.8mm.

Artefacts relating to navigation and measurement occur in a variety of material groups other than the wooden

146 [15] Fine-grained turned wooden eye piece and cover. Eye piece with glass secured by circlip. External thread

525 [330] Three fragments from wooden dish: two rim one body. Carved decoration on upper surface of two pieces. Fair condition. A) L 148mm; W 69.5mm; T 6.5–9.5mm. B) L 105mm; W 34mm; T 6.5–10.5mm. C) L 106mm; W 26mm; T 7.5–10mm (Figure 7.24).

90

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 7.24. Wooden dish, ID 525.

on eye piece fits internal thread on cover. Glass stained dark brown. Located by raised cannon. Good condition. H 15.2mm; D 40.1mm; Cover outer D 34.8–41.5mm; Cover inner D 33.5mm; H 10.5mm. 175 [72] Sandglass parts (Figure 7.25): Four end discs from sandglasses of various sizes, the largest being 95mm in diameter, the smallest 50mm in diameter. Also two complete and three broken spindles from sandglasses. Good condition. 329 [231] Telescope part. Outer casing (of hardwood, poss. Greenheart) unscrews. Glass disk or lens at either end. Very good condition. Screwed together L 147.5mm; D 26.5mm (max). Unscrewed larger part L 80mm; smaller part L 73.5mm (Figure 7.26).

Figure 7.25. End discs and spindles from various sizes of sand glasses, ID 175.

Figure 7.26. Telescope part, ID 329.

91

Dagless, Middleton, Campbell, Baker, Mays and Vitolo 360 [388] Strip of wood rebated at each end, possibly part of a back-staff or cross-staff. Fair condition. L 910mm; W 21mm.

are both inscribed on the rule. L 610mm; W 44mm; T 7mm (see Smith, 2010: fig. 8). 656 [NAV1977] Gunter rule, broken, with brass studs for locating dividers. Back - slightly altered order of scales: meridian and eq parts are at the top (normally at the bottom). L 175mm; W 42mm; T 6mm.

400 [294] Glass lens in broken circular casing. Good condition. Diameter c.50mm. 485 [No ID] Hourglass end disc. Good condition. Diameter c.70mm.

7.2.10.6 Tools and Equipment

650 [NAV0510] Cross-staff and four vanes, all incomplete. The vanes are in two sections held together by brass pins. The three longest vanes are stamped ’90’, ‘50’ and ‘30’, while the fourth has no marking. A full description, analysis and illustration is given by Smith (2010). Location: according to the National Maritime Museum catalogue, in a chest in the after part of the orlop deck, and, according to Smith (2010: 173), in a chest in the forward part of the gun deck. L 837mm; W 16mm.

A small number of wooden tools and equipment were recovered and identified. These included a turned handle and a scrubbing brush pierced with 32 symmetrically arranged bristle holes, but minus the bristles. It tapers to a point at either end and one of the piercings has been cut through, either accidentally or to take a hanging loop or lanyard. 105 [56] Turned wooden tool handle. Void to take squaresided tang (void D 7.5mm). Decorated with three incised lines at widest point of handle and at tang end. Good condition. L 97.5mm; D 16–31.3mm.

652 [NAV1974] Back-staff sighting vane, with brass plate inserted for the pinhole sight and ‘4’ stamped into the wood. Wooden sight vane with an inlaid brass pinhole sight and a brass pressure plate for holding it tightly in place on the 30 degree arc. L 118mm; W 37mm; H 43mm.

281 [200] Scrubbing brush, minus bristles (Figure 7.27). Tapers to a point at either end. Pierced 32 symmetrically arranged bristle holes, plus one that was possibly made to take a hanging loop or lanyard. Good condition. L 175mm; W 64mm.

653 [NAV1973] Cross-staff vane (part). Trapezoidal in section. L 62mm; W 24mm; H 24mm; internal measurement 16mm × 17mm (14mm to metal).

501 [311] (IC) Wooden grip with two brass bands of length 11mm. Central void, rectangular at one end (11mm × 7.5mm), which contains the remains of a metal rod. Some wire remains attached. Fragments of leather/fibrous material that had been trapped under bands, was recovered as a sample. Good condition. L 112.5mm; D 25mm.

654 [NAV1975] Back-staff vane (part), with ‘R I R’ stamped into it. A broken part of a wooden sight vane or shadow vane from a Davis Quadrant. L 37mm; W 41mm. 655 [NAV1976] Back-staff or cross-staff vane (part). Central trapezoidal part of a sighting vane with square hole. The arms have both been broken off. L 57mm; W 53mm; H 20mm.

7.2.10.7 Ship’s Fixtures and Fittings A variety of wooden objects identified as parts of the ship’s fixtures and fittings have been identified and are included within the 1979–2009 archive.

651 [NAV0593] Folding rule in two fragments with a brass hinge in each. Inscribed on two faces only, one marked in inches and the other non-linear. A) 201 × 21 × 9mm. One side shows a scale of inches from 11 to 14, divided into tenths, the other shows an unidentified scale 110 to 130. B) 124 x 20 x 9mm. One side shows a scale of inches from 20 to 30, divided into tenths, the other shows an unidentified scale 20 to 40, and 30 to 40. It has four brass studs for locating dividers and is marked 12 at the hinge. 649 [NAV0119] Gunter rule, inscribed with scales to aid navigational calculations by using a pair of dividers to measure off lengths against the different scales marked on the rule. The scales on the front include a linear scale in inches, a diagonal scale to set exact lengths, scales of leagues, chords, rhumbs, sines, tangents, semi-tangents and miles of longitude. Those on the back include sines of rhumbs, tangents of rhumbs, a line of numbers, logsines, logversines, logtangents, meridional lines and equal parts. The maker’s name, Thomas Arnold, and the date, 1700,

Figure 7.27. Scrubbing brush, minus bristles, ID 281.

92

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 043 [395] Timber fragment with evidence of two squaresided nails. Good condition. L 104mm; W 125mm; T 32mm; D 40mm.

091 [70] Shaped timber in two pieces, one end face shaped, evidence of metal fastening and one edge bevelled, straight grain, pale brown. Good condition. L 275mm; W 22mm.

045 [393] Timber fragment. Three drilled voids of 19.5mm. Diameter with flat base depth of 235mm. Surface recessed D 4.7mm.

102 [59] Block and sheave. Damage includes gribbling on one surface. Close-grained. Good condition. L 250mm; W 218mm; T 110mm. Sheave diameter 150mm. Sheave axle diameter 37mm.

046 [392] Timber fragment, no evidence of fixing. L 150mm; W 70–103mm; T 26.5mm.

130 [31] Block and sheave. Good condition. L 255mm; W 225mm; T 124mm. Sheave diameter 155mm. Sheave axle diameter 32.5mm.

047 [391] Fragment of curved worked timber, no fixing holes evident. Good condition. L 234mm; W 30–37mm; T 14.5–32.5mm.

131 [30] Block and sheave. Strop grooves retain fragments of rope. Close-grained, dark brown with orange iron stains. Good condition. L 250mm; W 216mm; T 120mm. Sheave diameter 151mm. Sheave axle diameter 35mm.

048 [390] Fragment of carved timber with one treenail (D 15.5mm) and one nail hole. One face plain with slight curve. Good condition. L 138mm; W 18.2–24.5mm; T 5– 33mm.

133 [28] Block and sheave. Strop grooves retain fragments of rope. Good condition. L 195mm; W 170mm; T 90mm. Sheave axle diameter 27.8mm. Located under raised cannon.

054 [107] Worked timber. One face flattened and other slightly rounded. Ends curved and bevelled on both faces. Incised marks on flattened face. Good condition. L 625mm; W 118mm; T 25.5mm.

134 [27] Block and sheave. Good condition, some damage to one cheek of block. L 157mm; W 120mm; T 58.6mm. Sheave diameter 85.5mm, sheave axle diameter 18–16mm. Located under raised cannon.

055 [106] Worked timber. One face flattened, other slightly rounded. Curved edge bevelled on both faces. Marks on flattened face. Good condition. L 535mm; W 160mm; T 28mm.

354 [266] Large cat or main-yard lift block with lignum vitae sheave. Shell L 1100mm; W 500mm; T 350mm. Sheave diameter 500mm (Figure 7.28).

056 [105] Worked timber. One face flat, other slightly rounded. Ends curved and bevelled. Indented circle and v on flat face. Good condition. L 510–600mm; W 125mm; T 16–25mm.

356 [268] Block and sheave. L 533mm; W 356mm; T 203mm. Sheave diameter 305mm.

Figure 7.28. Wooden pulley block, ID 354.

93

Dagless, Middleton, Campbell, Baker, Mays and Vitolo

Figure 7.29. Wooden pulley block, ID 377.

377 [485] Wooden pulley block with parts broken off and missing (Figure 7.29). Fair condition. L 242mm; W 185mm; T 119mm; swallow/breach height (total) 196mm; swallow/ breach width c.50mm; axle hole diameter 30–32mm.

386 [648.1] Wedge for tightening a treenail end. Good condition. L c.60mm. 486 [379] Sheave. Good condition. No measurement given.

382 [538] Double block with lignum vitae sheaves. One side of shell much degraded. L 497mm; W 330mm; T 330mm. Sheaves 290mm Diameter (Figure 7.30).

489 [380] Block shell, sheave missing (Figure 7.31). Fair condition. L 253mm; W 184mm; T 130mm; swallow/ breach height 205mm; swallow/breach width 35mm; axle hole diameter 35–37mm.

385 [648] Five treenails. Fair condition. L 203mm.

Figure 7.30. Double wooden block, ID 382.

94

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 7.31. Wooden pulley block, ID 489.

494 [302] Sheave (Lignum vitae) sheave with coak missing (Figure 7.32). Good condition. D 140mm; T 22mm. Aperture for coak, triangular sides of 69mm, 70mm, 69.5mm.

section 9.5.4) and a straight-grained wooden truck held on axle 154, with concretion on both flat surfaces and a broad arrow mark cut into its outer face (ID 637), was catalogued as being ‘from cannon 154’. An axle in place through truck 153 had one flat surface but otherwise was circular in cross section. It had evidence of vertical fixings (ID 638).

658 [WA1050] (IC) Fixed double block made of elm containing two angled sheaves thought to be made of lignum vitae, in position on their original pegs. The remains of three iron bolts can been seen in the block (Middleton, 2016). L 1220mm; W 366mm; D 178mm (Figure 7.33). The block is also shown in Figure 6.10.

A number of fragments of muskets were identified but photographs were available for only some of these for further study. One of the musket butts had remains of a metal plate on its upper surface and two sets of initials carved or stamped into it (not described and no photograph). The wood is stained dark brown and also has some red iron staining (ID 132). In addition, five other musket fragments were excavated: two butts and three fragments that were identified as the wooden casing of the barrel. A single pistol butt was recovered. No photograph was available, but it was described as ‘fragment of pistol butt in concretion, wood capped with non-ferrous metal’. Green corrosion products could be seen on the metal (ID 110). Three pistol

7.2.10.8 Weapons Wooden fragments of weaponry included fragments of cannon carriage. Although descriptions were available there were no photographs to enable the pieces to be studied further and their location is currently unknown. A carriage (ID 640) and wooden truck (ID 634, 635) were identified as having been parts of ‘cannon 155’ (ID 639,

Figure 7.32. Sheave with recess for a triangular coak, ID 494.

95

Dagless, Middleton, Campbell, Baker, Mays and Vitolo

Figure 7.33. Wooden double block, ID 658 (©Historic England, drawn by Chris Evans).

stocks were identified, two of which had remains of the ram-rod holds and traces of the trigger guard (IDs 224, 225). Two similar items (ID 482, 492) are thought to be parts of cartridge formers. Finally, four whole sabots and

11 sabot fragments were recovered (all listed in the site archive under ID 517). These were devices used in firearms (including cannon) to fire a projectile smaller than the bore diameter or that needs to be held in a precise position. 96

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 110 [51] Fragment of pistol butt in concretion, wood capped with non-ferrous metal. Green corrosion products on metal. L of butt remaining 108.7mm, D of capping 47.3–56.3mm.

634 [150] Wooden truck from cannon 155, axle void of 130–155mm. Showing wear at one surface, concentric marks on one flat surface. Good condition. D 460mm; T 130mm.

132 [29] Wooden musket butt. Two sets of initials carved into one face, and some carved or stamped into upper surface. Remains of metal plate on upper surface. Butt end, evidence of fixings. Wood is stained dark brown with some red iron staining. Good condition. L 320mm; W 45– 135mm; T 50–53mm.

635 [151] Wooden truck from cannon 155, axle void, no surface marks. Good condition. D 460mm; T 130mm. 637 [153] Straight-grained wooden truck held on axle 154, axle void diameter 130mm, concretion on both flat surfaces. Broad arrow mark cut into outer face. Good condition. D 460mm; T 125–130mm.

224 [138] (IC) Broken flintlock stock, with remains of ram-rod and traces of trigger guard. The wood of the musket stock is dry and flaky and the metal components are now absent. There is a white residue at the butt end that could possibly be salt efflorescence. L 775mm.

638 [154] Axle in place through truck 153, one flat face, otherwise circular in cross section, probably broken close to missing truck. The flat face, with width 135mm, has evidence of fixings vertical to it. Good condition. L 960mm, W 135mm (flat face) D 110mm (min.).

225 [139] (IC) Broken flintlock stock, with remains of ram-rod holder and traces of trigger guard. The wood of the musket stock is dry and flaky. Fair condition. L 902mm.

640 [156] Carriage from cannon 155, complete with stepped cheeks. Good condition. L 1510mm; H 540mm; T 137mm; T (of base) 124mm.

227 [141] (IC) Wooden pistol stock with inner construction features visible in X-radiograph. Fair condition. L 229mm (Figure 7.34).

7.2.10.9 Miscellaneous A number of miscellaneous wooden objects were identified that could not be assigned to specific categories. These included four disks of unknown function. One may originally have been a plate, but is now irregularly shaped and slightly warped. It was carved with the initials ID, IP or JP and demonstrated some repair work (ID 300). Another irregular disk, possibly a plate, was carved on one side with the initial ‘R’ (ID 301). Two wooden disks were catalogued, each of which has a central hole for a rope to be passed through. The rope knot remains in one of them (ID 495). The final disk (ID 094) had a central ‘pin’ hole, a slight ridge around its upper rim and a scratched line on its lower surface.

375 [483] Wooden musket stock, iron has completely eroded. Fair condition. L 864mm. 492 [305] Part of cartridge former. Turned with head, grain running along length. Good condition. L 49mm; D 14.7mm. 508 [47] (IC) Pistol body. Fragmentary remains of a pistol, including the pointed-ovoid brass butt cap and wooden butt. A number of iron concretions also preserved that represent the remains of a blade and hilt, but are heavily corroded. 517 [No ID] Sabots. Eleven fragments, four whole, four halves all found together.

044 [394] Wooden plug. Tapering cylinder with evidence of friction wear 18.3mm from ‘top’ surface. Good condition. L 40mm; D 16.8–25mm.

527 [No ID] Wooden musket fragment. No further details or measurements given.

064 [97] Table leg. No further details or photograph available.

528 [I: 05] (IC) Wooden musket fragment. Single copper-alloy ram-rod holder remains attached, with a piece of concretion attached. A fragment of ramrod appears from X-ray to be left in the concretion. There are some iron corrosion products on the inside and outside of the gun barrel part. The wood has several longitudinal splits. L 540mm.

094 [67] Wooden disk, ‘upper’ surface slight ridge around rim, some fine-grained, cream substance. ‘Lower’ surface central pin hole and scratched line. Straight grain. Good condition. T 61–64.5mm; D 3–5.5mm. 168 [18] Three wooden nails, one with end painted white. Worn, cracked and splintered. Fair condition. L 191mm.

529 [No ID] Wooden musket fragment. L 450mm. 530 [No ID] Wooden musket fragment. Fragile condition. L 300mm.

300 [212] Irregular wooden disc or plate. On one side are carved the initials ‘IP’ (Figure 7.35). Slightly warped with some repair work but in good condition. D 152–171mm.

590 [B: 05] (IC) Wooden musket fragment. Attached ramrod and two copper-alloy ramrod holders. No further details or measurements given.

301 [213] Irregular wooden disc or plate, dark in colour. Carved on one side with the initial ‘R’ (Figure 7.36). Good condition. D 127mm. 97

Dagless, Middleton, Campbell, Baker, Mays and Vitolo

Figure 7.34. Wooden pistol stock with internal construction shown by X-ray, ID 227 (X-ray not to scale ©Historic England).

369 [442] Crudely carved wooden object of unknown origin of cube and rod form. Each of the three cubic sections is 28mm wide, and two of them have patterns incised into them. The object is pierced longitudinally at one end. L 120mm.

488 [378] Unidentified wood fragment. L 95mm. 491 [300] Wooden oval, red lacquer on one surface. Good condition. H 66.7mm; W 40mm; T 5.5mm. 495 [301] (IC) Two wooden disks. One of which is turned and has a knot attached to it. The other one is damaged on the side and underside and has a hole. A) Good condition, slice across branch, rope knot remains, D 98–106.5mm;

399 [295–298] Wood and brass unidentified fragment. 482 [383] Peg. Smooth with spherical finial. L 80mm. 98

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 657 [NAV1978] Nine wooden fragments of unknown origin and use. No photographs available. 7.3 Discussion Artefacts made of organic materials would have been everyday common objects in the past but are infrequently found on archaeological sites because of the specific burial conditions needed for their survival. The anaerobic conditions of wreck sites are more conducive to the preservation of these materials and their instances on such sites are higher than on terrestrial ones. However, they remain fragile and their survival rate is still lower than artefacts constructed of more robust materials. It is likely that this was compounded in the case of the Stirling Castle by the deliberate selection of non-organic artefacts for recovery during the early seasons of work on the site (section 3.1.1).

Figure 7.35. Irregular wooden disc or plate. Initials ‘IP’ carved on one side, ID 300.

The scarcity of such finds means that there are few comparative sites for the organic assemblage as a whole. However, individual types of artefact can be reviewed with other wrecks of the period and also a small number of waterlogged terrestrial sites. Of the excavated wrecks of the same period some have produced only a few organic artefacts and others none at all. Those reviewed include the Warship Hazardous (1706), the Dartmouth (1690), the Anne (1690) (Marsden and Lyon, 1977), the Sapphire (1696), the Coronation (1699) and the Northumberland (1703). Other wrecks that have significant assemblages are of slightly earlier and slightly later periods, but have been also reviewed to understand the assemblage from a maritime perspective; these include HMS Invincible (1758) (Bingeman, 2010) and HMS Swift (1770) (Elkin et al., 2007). A significant assemblage was recovered from the Mary Rose excavations and, although this is much earlier in date, the functional, personal and galley objects were reviewed as their practical uses remained broadly similar in later periods.

Figure 7.36. Irregular wooden disc or plate, dark in colour. Initial ‘R’ carved on one side, ID 301.

Land-based site assemblages of the same period have been reviewed for evidence of the material culture of the time, as many of the same artefactual categories – personal and dress accessories and food preparation equipment, particularly those items more functional than decorative were used in everyday activities both in houses and on ships. Sites reviewed include the dockyard remains at Deptford (Divers, 2004), where the Stirling Castle was built as well as other waterfront sites in London, such as Adlands Wharf (Divers, 2002), Victoria Wharf (Tyler, 2001) and Narrow Street (Killock et al, 2005), where one of the local producers supplied rope to the naval dockyard at Deptford. The large and comprehensive assemblage of post-medieval artefacts recovered from excavations in Norwich (Margeson, 1993) was also consulted.

T 35mm. B) Incomplete, wood species ID of Pinus sp. (pine) (identification by G Campbell), D 111–120mm; D(hole) 7mm; T 21mm. 502 [303] (IC) Two pieces of wood, shaped and rebated. Evidence of two iron fixings on both pieces. Orange iron staining. Good condition. A) L 160mm; W 70mm; T 40mm. B) L 145mm; W 65mm; T 35mm. 504 [50] Fragment of wood in poor condition. L 96mm. 516 [392] Unidentified wood object. L 120mm.

7.3.1 Galley Objects

612 [1470] (IC) Wooden plank or board, possibly from the base of a chest. X-ray reveals 14 nail cavities around the outside edge: three at one end, five along each side, one at the other end. L 480mm; W 170mm.

Wooden vessels and utensils must have been common in every kitchen as they would have been cheap and in 99

Dagless, Middleton, Campbell, Baker, Mays and Vitolo of the lower status) probably wore similar garments to each other. The surviving fragments were plain with little or no decoration and would have been purely functional as would be expected for lower-ranking crew members.

plentiful supply (Margeson, 1993: 73). Wooden items would also have been useful on board ships as they were less likely to break in unstable conditions. Wooden bowls were made in many shapes and sizes and in large quantities in the post-medieval period (Morris, 1993: 95). Of those found in the significant assemblages recovered from Norwich, 80 per cent measured between 90mm and 260mm in diameter, and Morris states that this size would have been used for eating and drinking, with the larger platters used for food preparation and presentation. Wooden bowls were also found on some of the waterfront sites reviewed, as at Victoria Wharf (Tyler, 2001: 83). Dark colouring on the inside of these vessels is common and is an accumulation of grease and stains. Serving vessels recovered from the Stirling Castle include complete and fragmented wooden bowls and platters, some of which had engraved decoration. One of the bowls was decorated with what appears to be a shield carved into its upper surface, and another was stamped with the letters I.O. Unlike many of the initials seen on the metal artefacts described in chapter nine, ‘I.O.’ cannot be associated with any of the ship’s crew or marine complement between 1701 and 1703. Similar lathe-turned bowls were recovered from site of HMS Invincible (Bingeman, 2010).

Woven textiles were used for many personal garments during this period, such as shirts and caps (Crowfoot, 1993: 45), but their survival from the wreck was poor, with only a few fragments recovered. The leather objects are well preserved and are present in the Stirling Castle assemblage in the form of hats, a waistcoat and shoe fragments. One of the hats was patched, indicating that it had either been torn or worn through by constant use. The leather waistcoat fragment appears to have been a simple utilitarian garment with no decoration. Similar waistcoats (or jerkins) recovered from the Mary Rose were thought to have belonged to the gun crew (Gardiner, 2005: 18–19), with cloth jerkins being more expensive and belonging to officers rather than members of the lower ranks. The fragments of footwear require further study to try to determine the exact shape of the shoe/boot, but other post-medieval maritime assemblages (e.g. Invincible and Mary Rose, although respectively earlier and later in date) indicate that shoes worn by the crew were often simple utilitarian slip-on types and were official issue (Gardiner, 2005: 20).

Other galley artefacts recovered were objects that would have been in daily use such as the cups, bottle/jar stoppers, barrel spout and various wooden and bone handles. It is possible that these handles may have been parts of knives, the precise use of which is uncertain in assemblages of this period, although many must have been in everyday use for a variety of functions on board the ship (as discussed in the section on tools and equipment below) as well as for preparing and eating food. In the seventeeth century personal knives were still carried and would have been numerous on ships; both bone and wooden knife handles were also identified among the assemblages of the Dartmouth and Sapphire (Holman, 1975: 260). Wooden spoons must have been the chief eating and cooking utensil after the knife (Margeson, 1993: 73), but, as with those that must have been present on the Stirling Castle, these rarely survive. Forks did not become part of table cutlery until later in that century (Moore, 2006: 19), and to begin with were probably confined to the higher levels of society.

Small fragments of silk and damask (a reversible single colour silk fabric) would probably have belonged to clothing of the higher-status crew members, as they were both expensive and decorative. The fragments were small and could not be identified as specific items of clothing or accessories (apart from the pouch), but could have been used as the linings of hats, cloaks or other garments. Such a use is attested through examples recovered from sites in post-medieval Norwich (Crowfoot, 1993: 47). A use as ribbons is seen in the assemblage from the Mary Rose (Gardiner, 2005: 19–20). These items were often imported from Europe (Crowfoot, 1993: 49) and further afield overseas. 7.3.3 Personal/Leisure Items Personal and leisure items recovered include a toothbrush, combs, book covers and a fragment of a games board – reminders of the everyday lives and activities of the crew. The maintenance of health and hygiene on board ship was largely the responsibility of the barber-surgeon (Richards, 2005: 153), and it is likely that many of the personal grooming items particularly the razors, were his property. The fine-toothed nature of the bone and wooden combs recovered indicates that they would probably have been used not only for grooming but also for removing fleas and lice from hair. Although washing and grooming were not every day practices, the need to control pests and parasites was understood by crews (Richards, 2005: 157) and they would have combed their hair and been shaved regularly.

7.3.2 Dress Accessories It is the sturdier of dress accessories that survive in the archaeological record, as textiles are extremely fragile and among the most perishable of archaeological materials (Margeson, 1993: 45). Leather fares better but it is unusual to find complete items. Fragments of textiles that survive in the Stirling Castle assemblage potentially represent all social levels and ranks within the crew, with items of a coarse woven fabric, leather and fine silk all present. There was no official naval uniform at the time, and owing to the utilitarian nature of the clothing, many of the crew members (particularly those

Two small wooden containers were recovered. Their contents is unknown, as no residues were clearly seen 100

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 within them. Future analysis may be able to be able to determine this. One was a round pill box (ID 319), the other a cylindrical wooden case found to contain rolled fabric.

The leather belts and straps that were identified would not necessarily have been intended to hold up breeches, but were more likely to have been used to suspend hanging objects such as knives, pouches, swords and daggers, as suggested for those recovered from the Mary Rose (Gardiner, 2005: 20).

The book covers that were raised from the Stirling Castle were embossed or etched with decoration, but it was not possible to identify their subject matter. They could have been bibles, or prayer books which would probably have been common personal items on board a ship of this period, or possibly almanacs, and their presence indicates ownership by an educated person. As they were valued possessions they were often kept in chests (Richards and Gardiner, 2005: 127), as the ones from the Stirling Castle were. Copper-alloy clasps were common on such books, and were found on those recovered from the Norwich excavations (Margeson, 1993, 74) as well as at Victoria Wharf (Tyler, 2001: 89); they would have been used to keep the book closed as well as protect the pages.

Another reminder of the crew’s daily tasks is a scrubbing brush pierced with 32 symmetrically arranged bristle holes, but minus the bristles. It tapers to a point at either end and one of the piercings has been cut through, either accidentally or to take a hanging loop or lanyard. 7.4 Conclusion The significance of the organic assemblage from the Stirling Castle must be emphasised not only because of the rarity of discovering such large numbers of organic artefacts on one site but also because of the extensive variety of the finds. The assemblage represents many aspects of life on board the ship, from general daily tasks (e.g. scrubbing the decks) and specialist occupations through to leisure time. The hierarchy of the crew is represented through the materials present and their use, ranging from the basic, functional and undecorated objects used by the lower crew members through to fragments of expensive silk fabrics that would have come from the clothing of the higher ranks.

The recovered gaming board had a chequered pattern and would have been used for chess or draughts when the crew had some leisure time. Similar boards have been found on both terrestrial and maritime sites. A similar board was recovered from the Mary Rose (Redknap, 2005: 139), and a gaming board made of tile was found at Deptford (Divers, 2004: 62). 7.3.4 Tools and Equipment

Overall, this significant assemblage can add to our knowledge of shipboard culture of the time and could be enhanced by further work on certain categories of artefact, such as the leather clothing (hats, waistcoat, footwear), which could be explored in more detail to supplement the understanding of related items of the period recovered from both maritime and terrestrial sites. Equally, the textiles (wool and silk) would benefit from more detailed research to identify their exact fabric and possibly understand where they may have been traded from.

Wood, bone and horn, as well as being a common materials for sturdy containers and plates for preparing and serving meals, would also have been used to make many small implements. As discussed above, small multi-purpose knives were common objects in the post-medieval period and would have served a variety of functions on board the ship. Whittle-tanged utensils such as knives or other blades (where the blade narrows to a tang that is inserted into the handle) are generally accepted to have been utilitarian objects, usually lacked any decoration (Goodall, 1993: 125). Examples of multi-purpose knives recovered from other wrecks include those from the Mary Rose (although earlier in date). The rope fragments are representative of cordage that would have had multiple uses throughout the ship. The excavations at Deptford (see Divers, 2004) recovered treenails, hemp rope (cordage) and tarred animal fibre (caulkage), all of which were present on the Stirling Castle. The specialist artisans who worked with these materials – ropemakers and caulkers – were often based in dockyards, and finds of this sort are typical of dockside and quayside sites and are associated with the building, refitting or dismantling of ships. Rope was supplied to the Deptford dockyard from the shipbuilding industries in 43–53 Narrow Street, Ratcliffe, following a visit by Pepys to the site in 1664 (Killock et al., 2005: 3). The bone (or possibly ivory) pricker found on board the Stirling Castle would probably have been used for splicing lighter lengths of cordage. 101

8 Glass David Dungworth and Julian Whitewright 8.1 Introduction and Methodology

Height (H), Diameter (D), Thickness (T). In all cases, the ID number is that given to the artefact during the assessment stage of the current project; any external ID number is provided in parenthesis. The following classes of artefact were identified: bottles and flasks, drinking glasses, timing glasses and miscellaneous glass. The latter includes fragments of window glass recovered from the site.

Eighty-seven glass artefacts were contained within the artefact archive for the Stirling Castle for the period 1979–2009. The majority were raised during the initial seasons of investigative activity on the wreck site in 1979 and are currently dispersed between displays at Ramsgate Maritime Museum and Hastings Shipwreck Museum.

8.2.1 Bottles

A catalogue of glass artefacts, grouped by type, is presented below (section 8.2), and is followed in section 8.3 by the scientific analysis of a sample of the assemblage. Discussion of the assemblage as a whole and comparisons with artefact assemblages from contemporary wrecks and terrestrial sites of the period, where applicable, are contained in section 8.4. Glass and glass fragments are common on post-medieval sites and as such the descriptive nomenclature itself is relatively consistent. Some differences do remain, which were apparent when drawing together a catalogue compiled over a 30–year period. Therefore, the terms set out by Jones et al. (1989) have been adopted and applied in a standard way across the existing catalogue, including, where appropriate, the revision of existing artefact descriptions.

Although a variety of bottle types was recovered from the site, three forms dominate the assemblage: dark green ‘wine’ bottles, commonly termed ‘onion bottles’, two groups of square-sided bottles and a smaller number of cylindrical bottles. The presence of a large number of onion bottles tallies closely with the accounts given by those who dived the site in 1979 and reported seeing them scattered everywhere on the seabed. A large number were subsequently raised (Figure 8.1). In addition to this, a number of smaller bottles, both round and square, were recovered from an exposed case, thought to be a medicine case, during the 2000 season of work. 8.2.1.1 Onion Bottles

8.1.1 Scientific Examination and Analysis

The commonest type of bottle within the archive are the globular onion bottles that were a particular feature of recoveries during the early phases of work on the site. Thirty-seven such bottles are recorded in the archive, the majority of which are complete. Onion bottles such as those listed below are commonly found on British archaeological sites dating to the late seventeenth and early eighteenth centuries (e.g. Lucas et al., 2003: 183;

In addition to the presentation of the artefact catalogue undertaken here, a small selection of glass artefacts were sampled in order to examine their chemical composition. This work is presented in section 8.3 and all artefacts sampled and investigated with this purpose are marked (S) after their ID number. Precise determination of chemical composition has the potential to provide detailed information, and chemical analysis can shed light on the nature of the raw materials used to produce the artefacts as well as the technologies employed in their fabrication. The assemblage of artefacts from the wreck of the Stirling Castle is particularly valuable as the date of the wreck is known and the ship was thoroughly refitted just a few years before its loss. While a few objects may have been curated by owners for decades, most are likely to have been manufactured immediately prior to 1703. This well-dated glass assemblage, containing a variety of bottles, flat glass and tableware, can thus be compared with contemporary material from terrestrial contexts (Dungworth, 2011; 2012; Dungworth and Brain, 2009; 2013). 8.2 Recovered Artefacts by Type Each class of material is described in general terms, after which comes a summary artefact catalogue including the condition and dimensions: Length (L), Width (W),

Figure 8.1. Freshly raised onion bottles on the deck of the dive boat Shelandra in 1979 (Image courtesy of SeaDive).

103

Dungworth and Whitewright 184 [102] Onion bottle. Dark green. No other information.

Moore, 2000: 56–57; Watson et al., 2010: 187). Although generally classed as wine bottles, some examples were used for other purposes, such as for storing medicine. The Stirling Castle assemblage is easily dated, in simple terms, through its presence on the vessel at the time of its loss in 1703. The general form of bottles present in the assemblage match those set out by Noël Hume (2001: figs 8–13) dating between 1688 and 1704. No examples of the more angular forms from before 1688 were recovered. The replacement of the ‘onion bottle’ in the late 1730s by a fully cyclindrical body (Jones, 1986: 73) means that the material from the Stirling Castle is a useful record of vessel form around 1700. In addition, it is clear from the comparative examples noted above that the survival of complete vessels, including those with seals, is relatively rare.

185 [103] Onion bottle. Dark green, dimpled base and moulded seal on shoulder reading ‘CB 1703’. Lip chipped and barnacle-spotted but good condition. H 170mm; D 150mm, 151mm over stamp (Figure 8.2A). 186 [104] Onion bottle. Dark green, steep slope to shoulders. Lip chipped but good condition. H 159mm. 187 [105 ] Onion bottle. Dark green. No other information. 188 [106] Onion bottle. Dark green. No other information. 189 [107] Onion bottle. Dark green, dimpled base and moulded seal on shoulder reading ‘CB 1703’. Lip and seal chipped but good condition. H 170mm; D 140mm (Figure 8.2B).

A number of the bottles from the Stirling Castle do carry seals on the shoulders, including dates of 1700 (ID 192) and 1703 (IDs 189, 190, 591 and 621), while two have initials scratched into the surface of the body (IDs 593 and 607). The use of seals at this time was generally to denote the ownership of the bottle, and the initials of the owner were usually incorporated into the seal (Noël Hume, 2001: 61). The significance of this is discussed further below.

190 [108] Onion bottle. Dark green, with patches of paler green. Moulded seal at shoulder reading ‘CB 1703’. Chipped lip but good condition. H 178mm. 191 [109] Onion bottle. Dark green, part of cork still in neck. H 140mm. 192 [110] Onion bottle. Dark green, moulded seal at shoulder reading ‘William Stonas 1700’. Lip and seal chipped, some devitrification on body but good condition. H 156.5mm; D 140.5mm (Figure 8.2C).

Although the onion bottles from the shipwreck have generally been classified as wine bottles there are two apparent exceptions to this, IDs 138 and 140, the contents of which were fat or oil. These two bottles may represent the use of wine bottle forms for the storage of oils or vinegar, noted by Noël Hume (2001: 75) as being common practice in the late seventeenth and very early eighteenth centuries. Meanwhile, onion bottle forms associated with pharmaceutical use from an assemblage excavated in London are described and analysed by Gibson and Evans (1985: 151). The examples from the Stirling Castle are slightly bigger than that illustrated by Noël Hume dating to 1710 (2001: fig. 17) but comparable to those illustrated by Gibson and Evans (1985: fig. 1) dating to 1700–1730. It is also possible to include ID 139 within this interpretation on the basis of its similar dimensions and recovery from the same location, along with IDs 138 and 140, recovered during the 1999/2000 fieldwork.

193 [111] Onion bottle. Dark green with paler patches. Lip chipped but good condition. H 152mm. 194 [112] Onion bottle. Dark green. Barnacle spotted and with chipped lip but good condition. H 203mm. 199 [115.1] Onion bottle, neck fragment. Green glass with some concretion. Top of neck has thick ridge. Fair condition. L 102mm. 333 [234.1] Onion bottle. Green glass with rounded shoulders. Neck badly chipped, fair condition. H 155mm. 357 [286] Onion bottle. Dark green. Good condition. H 178mm. 379 [487] Onion bottle. Dark green. Good condition. H 178mm.

138 [23] Onion bottle. Stained blue, brown and black. Appears to be filled with fat or oil, sealed at neck with concreted matter. Location: S16. Complete but glass is defoliating, fair condition. H 150mm; D 150mm.

384 [638] Onion bottle. Broken neck and concretion over 50 per cent of remaining surface, fair condition. No other information.

139 [22] Onion bottle. Complete with remains of stopper, possibly cork. Location: S16. Some defoliation of glass but good condition. H 142mm; D(general) 150mm; D(neck) 26mm; D(shoulder) 53mm.

427 [1328] Onion bottle. Dark green. Lip chipped and some delamination of glass. Good condition. D(body) 155–159mm; D(collar) 41–43mm.

140 [21] Onion bottle. Clear-green, black and red/brown stain. Contents appear to be fat or oil. Location: S16. Delamination of glass and staining but fair condition. H 143mm; D(general) 150mm; D(neck) 26mm; D(lip) 23mm.

526 [316] (S) Onion bottle. Base fragment, dark green glass with pontil and occasional large air bubbles. Fair condition. D(base) 150mm; T 4–15mm. 104

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 8.2. Onion bottles. A, with seal, ‘CB 1703’ [ID 185]. B, with seal, ‘CB 1703’ [ID 189]. C, with seal, ‘William Stonas 1700’ [ID 192]. D, with engraved initials ‘WH’ [ID 607]. E, with ‘IR’ etched into body [ID 593].

105

Dungworth and Whitewright

Figure 8.2. (Continued).

591 [381] (S) Onion bottle. Dark yellowish-green glass with occasional air bubbles. Moulded circular seal at shoulder (D 37–41mm) reading ‘CB (or possibly GB) 1703’ with a bird emerging between the letters. Lip broken and overall surface in poor condition. H 175mm; H(shoulder) 100mm; H(neck) >69mm; D(base) 103mm; D(max) 150mm; D(shoulder/neck join) 62mm; D(neck) 25–51mm. Estimated capacity 1195ml or a little over a quart.

condition. H 141mm; D(base) 135mm; D(body) 151mm; D(neck) 25–51mm (Figure 8.2E).

592 [318] (S) Onion bottle. Dark yellowish-green glass with occasional air bubbles. Lip, some neck and part of the body missing, but good condition. H 110mm; H(shoulder) 52mm; H(neck) >52mm; D(base) 90mm; D(body) 120mm; D(neck/shoulder join) 58mm; D(neck) 25–46mm; T(wall) 2.5–4.5mm. Estimated capacity 310ml or a little over half a pint.

597 [324] (S) Onion bottle. Neck and part of the shoulder of a very large onion bottle. Green glass with rare small air bubbles. Applied collar. Good condition. H(neck) 111mm; D(shoulder/neck join) 100mm; D(neck) 44–79mm; D(collar) 57mm; D(lip) 44mm; T(lip wall) 6–9mm. Estimated capacity around 4–5 litres or around 1 gallon.

593 [317] Onion bottle. Dark green. ‘IR’ etched into body. Lip broken, glass pocked and delaminating but good

598 [327] (S) Onion bottle. Neck and part of the shoulder of a very large onion bottle. Green glass with rare small

594 [201] Onion bottle. Dark green, broken at base of neck, body lost. No other information. 595 [328] Onion bottle. Dark green. Two incised marks around neck. Poor condition. H 79mm; D(neck) 25–53mm; D(lip) 25mm.

106

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 they could be fitted efficiently into wooden cases (Killock et al., 2005: 57). Such bottles are frequently described in the literature as ‘Dutch gin bottles’, but were often of English origin and represented a large part of English bottle-making in the early seventeenth century (Noël Hume, 2001: 62). Many of those listed below were found within the group of bottles from the ‘surgeon’s store’, and very few of them are of a size comparable with the illustrated case bottles. Based on their size, perhaps only IDs 487 and 633 can be classified as true case bottles, the remaining examples simply being a convenient form for storage for a variety of liquids for use on board the ship.

air bubbles. Applied collar. Fair condition. H(neck) 118mm; D(shoulder/neck join) 100mm; D(neck) 48– 77mm; D(collar) 62mm; D(lip) 53mm; T(lip wall) 6–9mm. Estimated capacity around 4–6 litres or around 1 gallon. 599 [325] (S) Onion bottle. Neck and part of the shoulder of a very large onion bottle. Green glass with rare small air bubbles. Applied collar. Good condition. H(neck) 112mm; D (shoulder/neck join) 98mm; D(neck) 45–83mm; D(collar) 61mm; D(lip) 45mm. Estimated capacity around 4–5 litres or around 1 gallon. 602 [1054] Onion bottle. Dark green. Lip broken and some delamination of glass, fair condition.

060 [101] Bottle. Small square-sided bottle, pontil scar and flared lip. Fair condition. H 35.8mm; W 25mm; D(neck) 10mm; D(lip) 17mm.

603 [1053] Onion bottle. Dark green. Lip chipped and some delamination of glass, fair condition.

108 [53] Bottle. Small square-sided clear bottle, stained brown, with cork. Narrow neck, flared lip and pontil scar. Good condition. H 76mm; D(shoulder) 46mm; D(neck) 14mm; D(lip) 25mm.

604 [1055] Onion bottle. Dark green. Lip chipped and some delamination of glass, fair condition. 605 [1052] Onion bottle. Dark green. Lip damaged and body pock-marked with concretions and a hole, fair condition.

117 [44] Bottle. Square-sided clear glass stained red/black. Short narrow neck with rounded shoulders, remains of cork in place, one face broken away but present. Pontil mark. Fatty deposit on interior. Some devitrification of vessel body, but fair condition. H 138mm; W(max) 81mm; D(neck) 19mm; D(lip) 28–29mm; T(body wall) 3–5mm (Figure 8.3).

606 [1051] Onion bottle. Dark green. Lip damaged and body cracked in several places, fair condition. 607 [1046] Onion bottle. Dark green, ‘WH’ etched into body. Lip damaged, Good condition. H 165mm; D 140.5mm (Figure 8.2D).

120 [41] Bottle. Small square-sided clear/light green glass bottle, flared lip, square shoulder. Appears to have been blown against flat surface to form four faces. Broken pontil. Location: surgeon’s store. Good condition. H 72mm; D(neck) 16–19mm; D(lip) 26mm; T(lip) 2.7mm.

608 [1047] Onion bottle. Dark green. Lip damaged and body pock-marked, fair condition. 621 [487] Onion bottle. Dark green. Moulded seal at shoulder reading ‘CB 1703’. Some devitrification of glass, fair condition. 8.2.1.2 Square-Sided Bottles Nineteen bottles with square-sided bodies made up the other main component of the glass assemblage. Seven were relatively small bottles less than 80mm in height and the remainder – including three fragmentary pieces – were much larger bottles. Of these two groups, the former have been classified within the archive as medicine bottles, while the latter have been broadly classified as ‘case’ bottles, although this is discussed further below. Bottles from both groups, as well as some cylindrical bottles, were part of a group recovered from a wooden case interpreted as being part of the store of the Stirling Castle’s surgeon, William Deas. The larger square-sided bottles have been classified within the archive variously as ‘Dutch gin’ bottles or ‘case’ bottles. Both terms are somewhat generic and in many cases interchangeable. ‘Case bottles’ were commonly used to store and transport a range of liquids in a practical manner because

Figure 8.3. Square-sided bottle, ID 117, line drawing by Ann Smith (courtesy of Ann Smith and Isle of Thanet Archaeological Society).

107

Dungworth and Whitewright 123 [38] Bottle. Square-sided bottle, rounded shoulder, narrow neck and flared lip. Cork in place. Faint pontil scar. Odour of fish oil/medicinal. Location: surgeon’s store. Glass delaminating but good condition. H 139mm; D(neck) 18mm; D(lip) 25.5mm.

irregular flared lip, horizontal shoulder. Good condition. H 64mm; W (body) 27mm; D (body) 27mm (Figure 8.4E). 487 [1491] (S) Fragment of base and body from dark green square-sided gin bottle. The base is slightly pushed in and has a pontil scar. Little or no corrosion, good condition. H(max surviving) 56mm; W(base) 106mm ; T 3–5mm.

124 [37] Bottle. Square-sided bottle, clear glass stained dark-brown, rounded shoulder, narrow neck and flared lip. Cork in place. Pontil scar. Distinct odour: oil/medicinal. Signs of flaking but good condition. W(base) 73mm; W(shoulder) 81mm; D(neck) 20mm; D(lip) 27–28mm.

596 [No ID] Bottle. Square-sided body, green/brown bottle, rounded shoulder, neck missing. Body cracked and some devitrification, fair condition.

125 [36] Bottle. Square-sided bottle, clear glass stained dark-grey, short narrow neck and flared lip. Cork in place. Some silt and dark residue remains. Location: surgeon’s store. Good condition. H 136mm; W(base) 73mm; W(shoulder) 76mm; D(neck) 20mm; D(lip) 25mm.

600 [No ID] Bottle. Square-sided green/brown bottle, rounded shoulder, narrow neck and flared lip. Cork stopper still in place. Body cracked and some devitrification, fair condition. H 143mm with cork; H 136mm without cork; W/D 80mm (Figure 8.4C).

126 [35] Bottle. Small square-sided clear glass bottle with flared lip. Broken pontil. Cork in place. Location: surgeon’s store. Large fragment broken out of one face but generally good condition. H 76mm; W(base) 45mm; W(shoulder) 45mm; D(neck) 21mm; D(lip) 27mm.

633 [123] Bottle. Square-sided. Well-defined shoulder and narrow neck, broken pontil in base. Glass delaminating but good condition. H 275mm; W 54mm; D(neck) 38–54mm. 8.2.1.3 Cylindrical Bottles

195 [113] (S) Fragment of base from a square-sided bottle. A green glass with pitted and iridescent corroded surface. The walls are 3–7mm thick and the glass is very seedy (lots of small bubbles). The base is slightly pushed in (the push in is circular and is not central and so was probably not a feature of the mould. The overall form is square (the walls are 87mm and 89mm apart). The highest wall survives to a height of 105mm but the original bottle could have been twice this (Van den Bossche, 2001: plate 84.1).

In addition to the onion and square-sided bottles just described, the Stirling Castle archive contains a further seven bottles with a cylindrical body form. Two of these (ID 097 and 104) come from the ‘surgeon’s store’. As a result of this, and comparable examples (e.g. Gibson and Evans, 1985), all the cylindrical bottles within the assemblage can perhaps be interpreted as fulfilling a pharmaceutical role on board the ship, and as such be classed as vials.

196 [114] Fragment of neck. Green glass bottle with lead top or seal. No marks. Possibly the neck and seal from a broken square gin carafe. Broken but good condition. D 51mm.

086 [75] Bottle. Small clear/pale green glass, part of cork possibly in place. Narrow neck, wide flanged lip. Prominent pontil scar. Good condition. H(overall) 89mm; H(base to shoulder) 64.5mm; D(body) 42–44mm; D(neck) 13.5mm; D(lip) 27.5mm; W(lip) 8mm (Figure 8.5).

413 [1342] Bottle. Small square-sided clear/light green glass bottle with cork stopper in place. Narrow neck, wide irregular flared lip, horizontal shoulder. Good condition. H 77mm; W(base) 47mm; D(neck) 17mm; D(lip) 27mm (Figure 8.4B).

097 [64] Bottle. Clear to blue-green glass with many small air bubbles, cork in place. Narrow neck, wide flanged lip. Pontil scar obscured by corrosion products. Location: surgeon’s store. Good condition. H 83mm; D(body) 38– 40mm; D(neck) 14–15mm; D(lip) 30–32mm.

416 [1345] Bottle. Small square-sided clear/light green glass bottle, wide irregular flared lip, horizontal shoulder. Good condition. H 71mm; W(base) 45mm; D(neck) 16mm; D(lip) 27mm (Figure 8.4D).

104 [57] Bottle. Pale green glass. Short narrow neck, flared lip, deep pontil scar. Corrosion stains. Location: surgeon’s store. Fair condition. H 115mm; D 48–54mm; D(neck) 15–16mm; D(lip) 28mm.

417 [1346] Bottle. Square-sided green/brown bottle, rounded shoulder, narrow neck and flared lip that is damaged. Glass delaminating slightly, fair condition. H 148mm; W(base) 74mm; D(neck) 20mm (Figure 8.4A).

180 [98] Bottle. Clear, blue–grey glass. Rounded shoulders and conical base. Cylindrical neck finishes in wide flanged lip, partially broken off. Fair condition. H 97mm; D(body) 51mm; D(neck) 13mm; D(lip) 28mm (Figure 8.6A).

418 [1347] Bottle. Square-sided blue/brown bottle, rounded shoulder, narrow neck and flared, rounded lip. Fair condition. H 151mm; W(base) 75mm; D(neck) 20mm.

414 [1343] Bottle. Blue–green glass. Rounded shoulders and conical base. Cylindrical neck finishes in wide flanged lip, partially broken off. Fair condition. H 87mm; D(body) 41mm; D(neck) 13mm; D(lip) 27mm (Figure 8.6B).

420 [1490] Bottle. Small square-sided clear/light green glass bottle with cork stopper in place. Narrow neck, 108

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 8.4. Square-sided bottles. A, ID 417. B, ID 413. C, ID 600. D, ID 416. E, ID 420.

415 [1344] Bottle. Blue–grey/green glass with cork stopper in place. Rounded shoulders and conical base. Cylindrical neck finishes in very wide flanged lip. Good condition. H 85mm; D(body) 36–38mm; D(neck) 14mm; D(lip) 31mm (Figure 8.6C).

8.2.2 Drinking Glasses The remains of a number of drinking glasses were recovered from the Stirling Castle. These largely take the form of stemmed vessels, usually with one or two knops, and are presumably associated with the numerous wine bottles described earlier.

422 [1492] Bottle. Blue–grey glass. Rounded shoulders and conical base. Cylindrical neck finishes in wide flanged lip. Good condition. H c.125mm; D(body) 50–51mm; D(neck) 15mm; D(lip) 27mm (Figure 8.6D).

096 [65] Vessel base. Clear/pale amber poor-quality glass with many air bubbles. Appears to be base of stemmed 109

Dungworth and Whitewright glass, rolled lip, evidence of pontil scar. Good condition. H 20mm; D 58–60mm; T 2–3.5mm. 141 [20] Drinking glass. Stem and lower part of bowl of clear, heavy drinking glass. Two ball knops and a tapering stem broken below lower knop. Fair condition. H 91mm; D(bowl) 48mm; D(stem) 12–17mm; D(knop) 33mm. 148 [13] Drinking glass. Small-stemmed clear glass, base missing. Lower part of bowl decorated with fluting. Bowl joined to stem by sharp-edged button. Black deposit on interior. Stem broken off below ball knop but good condition. H 74mm; D(lip) 40mm; T(bowl) 1.5mm. 197 [115.1] (S) Pushed in base of beaker or tumbler. Pale green glass with very little sign of corrosion. The glass is extremely thin (~0.5mm). The walls have sheared close to the base, the underside of which has a pontil scar (~20mm diameter). The form is typical of the seventeenth century (Willmott, 2002). Poor condition. D 63mm. 198 [115.2] (S) Drinking glass. Lower part of simple cone, goblet bowl, tapering stem with a single ball knop and upper part of a plain conical foot. The glass is colourless glass with little sign of surface corrosion, rather thick and dense (i.e. lead glass). The form is typical of the period 1695–1715 (Elville, 1960). Fair condition. H 85mm; D(stem) 13–27mm; D(knop, max) 31mm (Figure 8.7A).

Figure 8.5. Cylindrical bottle, ID 086, drawing by Ann Smith (courtesy of Ann Smith and Isle of Thanet Archaeological Society).

Figure 8.6. Cylindrical bottles. A, ID 180. B, ID 414. C, ID 415. D, ID 422.

110

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 8.7. Drinking glasses. A, ID 198. B, ID 407.

407 [1340] (S) Drinking glass. Clear glass. Lower part of simple cone or hemisphere goblet bowl, and tapering stem with double ball knop. The top knop is larger and separated from the smaller, lower one. The glass is colourless glass with little sign of surface corrosion, rather thick and dense (i.e. lead glass). The form is typical of the period 1700–1720 (Elville, 1960). Stem broken below lower knop at upper edge of foot. Fair condition. H 91mm; D(stem) 11–17mm; D(lower knop, max) 19mm; D(upper knop, max) 33mm (Figure 8.7B).

of sand. This joint was then bound around with cloth or twine to seal and secure it. This two-piece manufacture of the glass body contrasts with later examples, some of which (though not all) were made from a single piece of glass (e.g. Maciaszek, 2008: 238–9) by the second half of the eighteenth century. The whole glass was then housed in a wooden frame comprised of two end pieces and a number of wooden spindles to connect the two ends. The paired nature of timing glass elements is well illustrated by ID 182, comprising five spare glasses recovered as a group. Within these five, there are two clear pairs, based on dimensions and form (182A/B and 182C/D), while the fifth glass (182E) has slightly different dimensions and is missing its partner glass.

408 [1349] (S) Vessel foot, stemmed goblet. A yellowish green glass with very little sign of corrosion. The glass surface is a little rough and contains a small amount of seed (small bubbles). The foot has a slightly uneven folded rim and is distinctly oval in plan (57–60mm across). The overall quality is rather poor. The folded foot was common through the sixteenth and seventeenth centuries and began to disappear only from the middle of the eighteenth century (Elville, 1960). Good condition. D 57–60mm.

170 [67] Half sandglass, tear-drop shape, clear/pale blue glass. Narrow neck and wide flanged lip. Good condition. H 229mm; D(body) 87mm; D(lip) 30mm; D(bore) 12.5mm. Found on the upper deck in association with a frame of five wooden spindles and two circular wooden end pieces. Spindle ends carry tenons to meet corresponding mortices in end pieces. Interior faces of end pieces are decorated with a double concentric incised line, exterior face is decorated with triple concentric incised lines (Figure 8.8A).

615 [1341] Drinking glass. Clear glass bowl of straightsided bucket form with diagonal fluting around the lower part. Stem broken at a ball knop. 8.2.3 Timing Glasses

171 [68] Complete sandglass, tear-drop shape, clear/pale blue glass. Narrow neck and wide flanged lip. The lip of one half-glass is broken. Good condition. D(body) 60mm; D(lip) 23mm; D(bore) 8mm. Found on the upper deck in association with a frame of five wooden spindles and two wooden circular end pieces. Spindle ends carry tenons to meet corresponding mortices in end pieces. Interior faces of end pieces are decorated with a single concentric incised line, exterior face is decorated with double concentric incised lines. H 178mm; D(end piece, max) 102mm.

A selection of timing glasses, which would be described today as ‘hour glasses’, were recovered. Such items had a two-fold purpose: to keep track of the time of day on board the ship, and in so doing to indicate the changing of the watch, mealtimes and so on; and to provide a chronological reference point for the purpose of navigation, through either keeping track of the time of the day or providing a measurement of a fixed period of time for the purposes of calculating the speed of the vessel. The latter was a critical element of the wider navigational positioning of the ship through dead reckoning, especially when estimating the longitude of the vessel.

172 [69] Complete sandglass, tear-drop shape, clear/pale blue glass. Narrow neck and wide flanged lip. Chip to one lip, good condition. H 186mm; D(body) 65/70mm; D(neck) 15mm; D(lip) 23mm. Found on the upper deck in association with a frame of five wooden spindles and two

Timing glasses during the period of the Stirling Castle comprised two glass vessels, conical or tear-drop in shape, with a diaphragm between their necks to regulate the flow 111

Dungworth and Whitewright

Figure 8.8. Timing glasses. A, ID 170. B, ID 173.

112

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703 circular wooden end pieces. Spindle ends carry tenons to meet corresponding mortices in end pieces. Interior faces of end-pieces are decorated with a single concentric incised line, exterior face is decorated with double concentric incised lines (Figure 8.9A).

from shipwrecks are extremely rare, although a roughly contemporary pair, one large and one small, was found on the Swedish warship Kronan, lost in 1676.1 The Kronan examples have octagonal end pieces, with five frame spindles carved into decorative geometric shapes. Meanwhile, a pair of octagonal end pieces, pierced for five spindles, have been identified within the collection of material from the Swedish warship Vasa, lost in 1628.2 The same combination of octagonal end pieces and decorative carving is seen in an example dating to c.1700 held by the National Maritime Museum, Greenwich (AST0110).3 Two further examples held at Greenwich both carry similar carved spindles; one has round end pieces and dates to c.1630 (AST0080),4 while the other has hexagonal end pieces and is dated broadly to the sixteenth century (AST0074).5

173 [70] Half sandglass, tear-drop shape, clear/pale-blue glass. Narrow neck and wide flanged lip. Good condition. H 310mm; D(body) 117mm; D(lip) 33mm; D(bore) 12mm. Found on the upper deck in association with five wooden spindles and one circular wooden end piece. Spindle ends carry tenons to meet corresponding mortices in end piece. The end piece is pierced to accommodate six spindles in total. The end piece is decorated with a triple concentric incised line (Figure 8.8B). 174 [71] Complete sandglass, conical shape, clear/pale blue glass. Narrow neck and wide flanged lip. Good condition. H 146mm; D(base) 37mm; D(neck) 15mm; D(lip) 28mm. Found in association with three wooden spindles and two circular wooden end pieces. Spindle ends carry tenons to meet corresponding mortices in end pieces. These are pierced to accommodate four spindles in total. Interior faces of end pieces are decorated with a single concentric incised line, exterior faces are decorated with double concentric incised line (Figure 8.9B).

Further comment can also be passed on the relative size and form of the timing glasses from the Stirling Castle, which can in turn provide some indication of their function within the navigation system of the ship. The dimensions of the half-glasses are summarised in Table 8.1, ordered according to body diameter. A group of the smallest half-glasses is readily apparent, with a body diameter ranging between 32mm and 39mm. This group is further characterised by their conical, rather than tear-drop, body form. The overall frame dimensions and body form of ID 174 correspond closely with the example held in Greenwich (AST0074) that represents a 15–minute timing glass. It therefore seems plausible to suggest that the group of conical timing glasses from the Stirling Castle were those used for calculating the speed of the vessel, using relatively short interval, possibly of about 15 minutes.

182 [100] Five spare half-sandglasses, conical shape, clear/pale blue glass. Narrow neck and wide flanged lip, conical push-up base (Figure 8.10). 182A: D(base) 37mm; D(neck) 17mm; D(lip) 29mm; D(bore) 13mm; H 66mm.

The remaining timing glasses all have a tear-drop body form. Of these, ID 173 is distinctly larger than all the others and nearly twice the diameter of ID 171 and ID 644. The diameter of ID 172 is comparable to those latter examples and so those three timing glasses can probably be grouped together. They also share identical lip diameters and similar bore diameters. Lying midway between the group of three (ID 171, ID 172 and ID 644) and the larger ID 173, is ID 170, which is appreciably smaller than ID 173 in terms of body diameter but very similar from the perspective of lip and bore diameter. On the basis of the latter measurements, these two glasses can probably be grouped together.

182B: D(base) 38mm; D(neck) 16mm; D(lip) 28mm; D(bore) 12mm; H 65mm. 182C: D(base) 32mm; D(neck) 13mm; D(lip) 24mm; D(bore) 9mm; H 62mm. 182D: D(base) 34mm; D(neck) 14mm; D(lip) 23mm; D(bore) 9mm; H 63mm. 182E: D(base) 39mm; D(neck) 14mm; D(lip) 24mm; D(bore) 10mm; H 60mm. 644 [ASTO111] Complete sandglass, tear-drop shape, clear/pale blue glass. Narrow neck and flared, flattened lip. Push-up base. Good condition. D(body) 61mm; D(neck) 15mm; D(lip) 23mm. Found in association with a frame of five wooden spindles and two circular wooden end pieces. Spindle ends carry tenons to meet corresponding mortices in end pieces. Interior faces of end pieces are decorated with a single concentric incised line, exterior face is decorated with double concentric incised lines. H 184mm; D(end piece, max) 84mm.

It is very difficult to make further inference about the time-scales represented by the two larger groups of glasses because of the loss of the diaphragms that would have been secured between the half-glasses to regulate the flow of sand. Depending on the diaphragm, the same size glass could be set to measure longer or shorter time-scales. At https://www.regalskeppetkronan.se/en/the-finds/two-sandglasses/. https://www.vasamuseet.se/en/collections/search-the-collectiondigitaltmuseum. 3 http://collections.rmg.co.uk/collections/objects/10272.html. 4 http://collections.rmg.co.uk/collections/objects/10242.html. 5 http://collections.rmg.co.uk/collections/objects/10236.html. 1

2

When taken as an entire object, glass plus frame, the timing glasses within the Stirling Castle archive are relatively plain in nature. Comparative complete examples 113

Dungworth and Whitewright

Figure 8.9. Timing glasses. A, ID 172. B, ID 174.

114

The Stirling Castle, a 70-gun Ship Lost in the Great Storm of 1703

Figure 8.10. Timing glasses, ID 182A–E.

503, 1–19 [307, 308] (S) Window glass. Nineteen fragments of flat glass of variable thickness (see below). The glass has a slightly iridescent and pitted surface. The iridescence is present on all surfaces, but the pitting is largely absent from the cut and broken edges. The thickness of the glass (typically 5mm) indicates that this is plate glass. The presence of a rounded edge (SC503.08) suggests that this was formed using the cylinder or broad glass process and then flattened (i.e. not cast plate glass). The flat surfaces of the glass are very close to plane surfaces and were probably ground and polished. The variation in thickness of individual fragments suggests that the two flat surfaces are not perfectly parallel. A patent for the manufacture of plate glass (for mirrors, coach windows and some domestic windows) was issued in 1662 and manufacture took place in Vauxhall, London. At this time plate manufacture would have used the cylinder or broad process, as the cast process was not invented – in France – until 1688. While an attempt was made to introduce the cast process into England in the 1690s, this did not happen until the 1770s.

Table 8.1. Summary of timing glass dimensions, ordered by diameter of vessel body. ID 173 170 172 644 171 174 182A 182B 182C 182D 182E

Body Ø (mm) 117 87 65/70 61 60 37 37 38 32 34 39

Lip Ø (mm)

Bore Ø (mm)

33 30 23 23 23 28 29 28 24 23 24

12 13 c.9 c.11 8 c.11 13 12 9 9 10

Form Tear-drop Tear-drop Tear-drop Tear-drop Tear-drop Conical Conical Conical Conical Conical Conical

best, it can be suggested that the largest timing glasses (ID 173 and possibly ID 170) could have been used for setting the length of an entire watch of four hours, while the others (ID 171, 172 and 644) were used to set intervals within each watch.

It is not clear when or why the pitting in the glass occurred. It is unlikely to have been an artefact of the polishing process in order to deliberately reduce transparency (privacy glass) as the glass generally has a good transparency (e.g. it is quite easy to read printed text when the glass is laid on printed paper). In addition, the density of pitting is rather variable between fragments and even within single fragments. It is likely that the pitting occurred after production but before the ship was wrecked (otherwise the pitting would be expected on the edges), however, the aim (if any) or the cause(s) remain uncertain. These fragments have been given additional numbers (ID503.01–503.19) to allow individual identification (Table 8.2).

8.2.4 Window Glass 479 [376] (S) A fragment of plate glass similar in most respects to ID503.1–19 (below). The glass has flat plane surfaces and a thickness of 4.2–4.7mm. There is an apparent arrow impressed on one surface. This has been produced by the removal of some material and the impression comprises numerous very small fracture surfaces. This could have been produced by pressure flaking with a very small but hard tool. There is no sign of any pitting or iridescence on the surface of the apparent arrow, which suggests that this occurred after the recovery of the glass from the shipwreck.

522 [315] Glass fragment. Probably window glass. Good condition. L 33mm; W 24mm; T 3.4mm. 115

Dungworth and Whitewright Table 8.2. Detail of fragments of flat window glass recorded under the ID 503. No. 503.01 503.02 503.03 503.04 503.05 503.06 503.07

Thickness 3.3–3.7mm 5.3–5.5mm 4.2–4.3mm 4.2–4.3mm 3.3–3.9mm 4.4–4.5mm 5.5–5.9mm

No. 503.08 503.09 503.10 503.11 503.12 503.13 503.14

Thickness 3.6–4.0mm 4.8–5.6mm 5.8–6.8mm 3.9–4.5mm 6.4–7.1mm 4.7–5.0mm 4.9–5.1mm

8.2.5 Miscellaneous

No. 503.15 503.16 503.17 503.18 503.19

Thickness 4.0–4.7mm 5.1–6.3mm 6.0–7.2mm 5.7–6.5mm 4.9–5.5mm

fluorescence (EDXRF) was used to determine the presence of a range of elements (EDAX Eagle II, 30µm monocollimator, 40kV accelerating voltage, 1mA current, 100sec livetime duration). EDXRF can provide such data without damaging artefacts, but the data may be affected by surface corrosion. Use of EDXRF in this way was especially useful where sampling would significantly detract from the aesthetic value of an artefact (e.g. ID329, the possible microscope, and ID554, a pewter spoon). Some objects (e.g. navigational dividers) could be analysed using EDXRF by carefully removing corrosion products from a small area (