Later Prehistoric Settlement in Cornwall and the Isles of Scilly: Evidence from Five Excavations 1789699576, 9781789699579, 9781789699586

Citation preview

Later Prehistoric Settlement in Cornwall and the Isles of Scilly: Evidence from Five Excavations Edited by

Andy M Jones and Graeme Kirkham

Later Prehistoric Settlement in Cornwall and the Isles of Scilly: Evidence from Five Excavations Edited by

Andy M Jones and Graeme Kirkham

Archaeopress Archaeology

Archaeopress Publishing Ltd Summertown Pavilion 18-24 Middle Way Summertown Oxford OX2 7LG www.archaeopress.com

ISBN 978-1-78969-957-9 ISBN 978-1-78969-958-6 (e-Pdf) © the individual authors and Archaeopress 2021 Cover: The Iron Age settlement at Higher Besore reconstructed. (Painting: George Scott.)

All rights reserved. No part of this book may be reproduced, or transmitted, in any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior written permission of the copyright owners. This book is available direct from Archaeopress or from our website www.archaeopress.com

Contents Acknowledgements��������������������������������������������������������������������������������������������������������������������������������������������������������������������� xi

Section 1: Background to the Project

Andy M Jones Chapter 1: Introduction to the Volume�������������������������������������������������������������������������������������������������������������������������������������3

Section 2: Archaeological Recording during the 1996 Coast Protection Scheme at Porth Killier, St Agnes, Isles of Scilly

Charles Johns, Jeanette Ratcliffe and Andrew Young, with contributions from David Dungworth, Janice Light, Alison Locker, Henrietta Quinnell, Vanessa Straker and Roger Taylor Chapter 2.1: Background to the Excavations����������������������������������������������������������������������������������������������������������������������������9 Chapter 2.2: The Excavation Results����������������������������������������������������������������������������������������������������������������������������������������15 Chapter 2.3: The Artefacts���������������������������������������������������������������������������������������������������������������������������������������������������������24 Chapter 2.4: Mammal, Bird and Fish Bones����������������������������������������������������������������������������������������������������������������������������41 Chapter 2.5: Shell������������������������������������������������������������������������������������������������������������������������������������������������������������������������45 Chapter 2.6: Plant Macrofossils ������������������������������������������������������������������������������������������������������������������������������������������������52 Chapter 2.7: Radiocarbon Dating����������������������������������������������������������������������������������������������������������������������������������������������56 Chapter 2.8: Discussion��������������������������������������������������������������������������������������������������������������������������������������������������������������58

Section 3: Excavations at Killigrew 1996: an Iron Age and Romano-British Industrial Site on the Trispen Bypass, Cornwall

Dick Cole and Jacqueline Nowakowski FSA with contributions from Rowena Gale, Sophie Lamb, Albertine Malham, Gerry McDonnell, Henrietta Quinnell, Laura Ratcliffe–Warren, Adam Sharpe, Vanessa Straker and Roger Taylor Chapter 3.1: Background to the Excavations��������������������������������������������������������������������������������������������������������������������������65 Chapter 3.2: The Excavation Results����������������������������������������������������������������������������������������������������������������������������������������70 Chapter 3.3: Radiocarbon Dating and Charcoal Identification��������������������������������������������������������������������������������������������81 Chapter 3.4: The Artefacts���������������������������������������������������������������������������������������������������������������������������������������������������������82 Chapter 3.5: Charred Plant Macrofossils���������������������������������������������������������������������������������������������������������������������������������95 Chapter 3.6: Discussion �������������������������������������������������������������������������������������������������������������������������������������������������������������98

Section 4: Archaeological Investigations at Nancemere, Truro, Cornwall 2002: a Prehistoric and Romano-British Landscape

James Gossip with contributions from Rowena Gale, Andy M Jones, Julie Jones, Anna Lawson-Jones, Henrietta Quinnell, Clare Randall and Roger Taylor Chapter 4.1: Location and Background����������������������������������������������������������������������������������������������������������������������������������107 Chapter 4.2: The Excavation Results��������������������������������������������������������������������������������������������������������������������������������������110 Chapter 4.3: The Artefacts�������������������������������������������������������������������������������������������������������������������������������������������������������127 Chapter 4.4: The Charcoal��������������������������������������������������������������������������������������������������������������������������������������������������������136 Chapter 4.5: Animal Bone��������������������������������������������������������������������������������������������������������������������������������������������������������140 Chapter 4.6: Radiocarbon Dating��������������������������������������������������������������������������������������������������������������������������������������������142 Chapter 4.7: Discussion������������������������������������������������������������������������������������������������������������������������������������������������������������143

Section 5: Life Outside the Round: Bronze Age and Iron Age Settlement at Higher Besore and Truro College, Threemilestone, Truro, 2004–5

James Gossip with contributions from Justine Bayley, Paul Bidwell, Sarnia Butcher, Wendy Carruthers, Rowena Gale, J D Hill, Andy M Jones, Julie Jones, Anna Lawson-Jones, Roger Mcbride, Stuart Needham, Peter Northover, Cynthia Poole, Henrietta Quinnell, Roger Taylor, Anna Tyacke and Tim Young Chapter 5.1: Background to the Excavations������������������������������������������������������������������������������������������������������������������������149 Chapter 5.2: The Excavation Results��������������������������������������������������������������������������������������������������������������������������������������155 i

Chapter 5.3: The Artefacts ������������������������������������������������������������������������������������������������������������������������������������������������������190 Chapter 5.4: The Plant Macrofossils ��������������������������������������������������������������������������������������������������������������������������������������251 Chapter 5.5: The Charcoal �������������������������������������������������������������������������������������������������������������������������������������������������������258 Chapter 5.6: Radiocarbon Dating��������������������������������������������������������������������������������������������������������������������������������������������263 Chapter 5.7: Discussion �����������������������������������������������������������������������������������������������������������������������������������������������������������265

Section 6: Excavation of an Iron Age Settlement and a Roman Period Enclosure at Porthleven, 2014

Andy M Jones, with contributions from Paul Bidwell, Dana Challinor, Anna Lawson-Jones, Henrietta Quinnell, Clare Randall, Ryan P Smith and Roger Taylor Chapter 6.1: Background to the Excavations������������������������������������������������������������������������������������������������������������������������283 Chapter 6.2: Results from the Excavations����������������������������������������������������������������������������������������������������������������������������286 Chapter 6.3: The Artefacts ������������������������������������������������������������������������������������������������������������������������������������������������������302 Chapter 6.4: The Charcoal��������������������������������������������������������������������������������������������������������������������������������������������������������314 Chapter 6.5: The Animal Bone������������������������������������������������������������������������������������������������������������������������������������������������317 Chapter 6.6: The Radiocarbon Dating������������������������������������������������������������������������������������������������������������������������������������319 Chapter 6.7: Discussion������������������������������������������������������������������������������������������������������������������������������������������������������������321

Section 7: Review and Overview

Andy M Jones Chapter 7.1: From Beaker Pits to Living in the Round: Some Themes�����������������������������������������������������������������������������333 Bibliography������������������������������������������������������������������������������������������������������������������������������������������������������������������������������346

ii

List of Figures Section 1: Background to the Project

Figure 1.1 Map of Cornwall showing sites covered by this monograph.����������������������������������������������������������������������������� 3 Figure 1.2 Killigrew 1996, general view of excavation looking north. �������������������������������������������������������������������������������� 4

Section 2: Archaeological Recording during the 1996 Coast Protection Scheme at Porth Killier, St Agnes, Isles of Scilly

Figure 2.1 Location map showing Isles of Scilly and area of excavation.���������������������������������������������������������������������������� 9 Figure 2.2 Modern Landline mapping (area of excavation shown in red box).���������������������������������������������������������������� 10 Figure 2.3 Detail from 1876 OS map (area of excavation shown in red box).������������������������������������������������������������������� 11 Figure 2.4 Plan of the cliff edge showing the location of Zones A, B, C and D.����������������������������������������������������������������� 15 Figure 2.5 Diagrammatic section drawing of Zone A, showing the main features.��������������������������������������������������������� 16 Figure 2.6 Pre-, during and post-excavation plans of roundhouse 21.������������������������������������������������������������������������������ 17 Figure 2.7 Diagrammatic section drawing of Zone B, showing the main features.��������������������������������������������������������� 19 Figure 2.8 Detailed section drawing of pit [50] complex.���������������������������������������������������������������������������������������������������� 20 Figure 2.9 Diagrammatic section drawing of Zone C, showing the main features before cutting back.���������������������� 21 Figure 2.10 Diagrammatic section drawing of Zone C, showing the main features after cutting back.���������������������� 21 Figure 2.11 Pre- and post-excavation plans of Zone C.�������������������������������������������������������������������������������������������������������� 22 Figure 2.12 Porth Killier. Bronze Age pottery from Zone A, roundhouse area. Note P3 shows interior of sherd Nos P6–8 are from the Coastal Erosion Survey. ��������������������������������������������������������������������������������������������� 27 Figure 2.13 Bronze Age pottery from Zone B, area with pits. Note P11 shows interior of sherd. �������������������������������� 29 Figure 2.14 Porth Killier. Neolithic bowl P13 and Bronze Age pottery from Zone C, area with cairn.������������������������� 31 Figure 2.15 Porth Killier. Cross section of pivot stone SF175 from floor [27] roundhouse 21 Zone.����������������������������� 34 Figure 2.16 Porth Killier. Broken saddle quern SF154A, [52] in fill of Pit [50] Zone B. Arrows indicate battering on side. ���������������������������������������������������������������������������������������������������������������������������������������������������������� 34 Figure 2.17 Porth Killier. ‘Bruising mullers’ SF72, from (9) midden predating roundhouse 21 Zone A and SF151, [52] in fill of pit [50] Zone B. Cupped pebble hammer SF94 from floor [27] roundhouse 21 Zone A. Arrows indicate battering on side.��������������������������������������������������������������������������������������������������������������� 35 Figure 2.18 Rubbing stone SF148, from (84) soil build up over Zone C. ���������������������������������������������������������������������������� 36 Figure 2.19 Hammerstone SF176, from (150), cist infill in cairn in Zone C. ��������������������������������������������������������������������� 36 Figure 2.20 Histograms showing size distribution of limpets in Porth Killier shell samples: contexts (9), (13), (14), [49]. Figures 2.20a–2.20c are for bulk samples collected during the 1996 CAU excavation and Figures 2.20d–2.20g for bulk samples collected by Vanessa Straker (1989). Note variation in y-axis to allow for larger number of shells in VS samples.������������������������������������������������������������������������������������� 49 Figure 2.21 Results from the radiocarbon dating at Porth Killier and Porth Coose.������������������������������������������������������� 57 Figure 2.22 Foundation trench for the concrete sea wall at Porth Killier showing the archaeological features exposed in the cliff face. (Photograph: Cornwall and Scilly HER at Kresen Kernow.)������������������������ 58 Figure 2.23 Modelled land and intertidal areas in Scilly at about 1500 cal BC, showing intertidal and coastal field systems (purple dots) recorded in the Cornwall and Isles of Scilly HER. (Source: Charman et al 2016.)������������������������������������������������������������������������������������������������������������������������������������������������������ 59

Section 3: Excavations at Killigrew 1996: an Iron Age and Romano-British Industrial Site on the Trispen Bypass, Cornwall

Figure 3.1 Location map of Killigrew enclosure.������������������������������������������������������������������������������������������������������������������� 66 Figure 3.2 Killigrew enclosure and wider landscape. Later prehistoric enclosures (upstanding earthworks, place-names and cropmark enclosures). (Sources: HER and Cornwall and Isles of Scilly National Mapping Programme.)����������������������������������������������������������������������������������������������������������������������� 67 iii

Figure 3.3 Killigrew enclosure. Area of geophysical survey and extent of excavation. (Source: GSB Geophysical Survey, Gater 1996.)���������������������������������������������������������������������������������������������������������������������������������69 Figure 3.4 Killigrew enclosure excavation area showing features of all phases.��������������������������������������������������������������71 Figure 3.5 Furnace [82] land related metalworking activities within the enclosure. Late Iron Age site reused in the Romano-British period. For location see Figure 3.4 and section B-A in Figure 3.6.��������������������73 Figure 3.6 East – facing section across furnace [82].�������������������������������������������������������������������������������������������������������������74 Figure 3.7 East-facing section of outer downslope ditch [20].���������������������������������������������������������������������������������������������75 Figure 3.8 West-facing section of outer upslope ditch [31].�������������������������������������������������������������������������������������������������75 Figure 3.9 West-Facing section of inner upslope ditch [27].������������������������������������������������������������������������������������������������75 Figure 3.10 East-facing section of inner downslope ditch [36].�������������������������������������������������������������������������������������������76 Figure 3.11 Section across pits [4] and [7] – location of the tin dish.���������������������������������������������������������������������������������79 Figure 3.12 East-facing section across palisade ditch [58].���������������������������������������������������������������������������������������������������79 Figure 3.13 Beaker vessel from pit [113].���������������������������������������������������������������������������������������������������������������������������������82 Figure 3.14 Roman period gabbroic pottery: P1 [8], P2 [6], P4–7 [37] and P8 [58].�����������������������������������������������������������84 Figure 3.15 Coarse gabbroic storage jar P3 [112].������������������������������������������������������������������������������������������������������������������85 Figure 3.16 Romano-British ceramics from hollow 17: P9 (13), P10 (12), P11 (12), P12 (12), P13 (10), P14 (2), and P15 (2).����������������������������������������������������������������������������������������������������������������������������������������������������������������86 Figure 3.17 Romano-British ceramics: P16 wall 111, P17 soil in central area [55], P18 [55] and P19 [55].��������������������87 Figure 3.18 Part of a tin dish found in pit [7].�������������������������������������������������������������������������������������������������������������������������88 Figure 3.19 Stone weights: S1 fill (3) of inner ditch on north side, S2 unstratified.���������������������������������������������������������93 Figure 3.20 Whetstone S5 wall 111.������������������������������������������������������������������������������������������������������������������������������������������94 Figure 3.21 Furnace [82] during excavation looking north-west.�������������������������������������������������������������������������������������101 Figure 3.22 Furnace [82] during excavation looking south.�����������������������������������������������������������������������������������������������102 Figure 3.23 Tin dish in situ in pit [7].��������������������������������������������������������������������������������������������������������������������������������������103

Section 4: Archaeological Investigations at Nancemere, Truro, Cornwall 2002: a Prehistoric and Romano-British Landscape

Figure 4.1 Site location.������������������������������������������������������������������������������������������������������������������������������������������������������������108 Figure 4.2 Extent of excavation, geophysical survey results and principal features.����������������������������������������������������109 Figure 4.3 The Early Bronze Age: Location of Beaker pits and possible contemporary features in Trench 2.�����������111 Figure 4.4 Curvilinear ditch [3010].����������������������������������������������������������������������������������������������������������������������������������������112 Figure 4.5 Bronze Age activity in Trench 2.��������������������������������������������������������������������������������������������������������������������������113 Figure 4.6 The Late Iron Age / Romano-British enclosure.������������������������������������������������������������������������������������������������113 Figure 4.7 The round – Phase I. ����������������������������������������������������������������������������������������������������������������������������������������������114 Figure 4.8 The round – Phase Ib.���������������������������������������������������������������������������������������������������������������������������������������������115 Figure 4.9 The round – Phase 2.����������������������������������������������������������������������������������������������������������������������������������������������117 Figure 4.10 The round – Phase 2 entrance features. �����������������������������������������������������������������������������������������������������������118 Figure 4.11 The round – Phase 2 stone capped drain in entrance.�����������������������������������������������������������������������������������118 Figure 4.12 The round – Phases 3 and 4.��������������������������������������������������������������������������������������������������������������������������������119 Figure 4.13 The round – structure 1281 and associated features.�������������������������������������������������������������������������������������121 Figure 4.14 The round – structure 1085, hearth complex 1108 and associated features.����������������������������������������������122 Figure 4.15 east end round CA layout CA Portrait (1).jpg���������������������������������������������������������������������������������������������������124 Figure 4.16 Field systems found at Nancemere.�������������������������������������������������������������������������������������������������������������������125 Figure 4.17 Pottery P1 (top), P2 (middle), P9 (bottom).������������������������������������������������������������������������������������������������������127

iv

Section 5: Life Outside the Round: Bronze Age and Iron Age Settlement at Higher Besore and Truro College, Threemilestone, Truro, 2004–5

Figure 5.1 Site location.������������������������������������������������������������������������������������������������������������������������������������������������������������150 Figure 5.2 Extent of excavated areas and geophysical anomalies.������������������������������������������������������������������������������������151 Figure 5.3 The post-medieval landscape around Higher Besore.��������������������������������������������������������������������������������������152 Figure 5.4 Aerial photograph of Truro College playing fields during excavation.���������������������������������������������������������153 Figure 5.5 Higher Besore Area F – Early – Late Bronze Age activity.��������������������������������������������������������������������������������156 Figure 5.6 Higher Besore Area F, structure 1 (Early Bronze Age).�������������������������������������������������������������������������������������157 Figure 5.7 Excavation of Late Bronze Age Plain Ware from Higher Besore pit [6500].��������������������������������������������������158 Figure 5.8 Bronze Age activity across the excavated areas.�����������������������������������������������������������������������������������������������159 Figure 5.9 Late Bronze Age double sided mullers and quern fragments excavated from pit [5027] at Truro College, Area D.��������������������������������������������������������������������������������������������������������������������������������������������������160 Figure 5.10 Truro College Area D, Late Bronze Age pits.����������������������������������������������������������������������������������������������������161 Figure 5.11 Truro College, the Late Bronze Age post-ring structures (red circles represent the likely position of walls).����������������������������������������������������������������������������������������������������������������������������������������������������������162 Figure 5.12 Truro College, Late Bronze Age structure 3.����������������������������������������������������������������������������������������������������163 Figure 5.13 Truro College, Late Bronze Age structures 2, 4 and 6.������������������������������������������������������������������������������������165 Figure 5.14 Truro College, structure 5.����������������������������������������������������������������������������������������������������������������������������������166 Figure 5.15 Higher Besore Area F, structure 2 (Late Bronze Age).�������������������������������������������������������������������������������������168 Figure 5.16 Truro College, section through Late Bronze Age pit [2045].��������������������������������������������������������������������������168 Figure 5.17 Late Bronze Age pit [2045].���������������������������������������������������������������������������������������������������������������������������������169 Figure 5.18 Late Iron Age features at Higher Besore.����������������������������������������������������������������������������������������������������������170 Figure 5.19 Late Iron Age features at Truro College.�����������������������������������������������������������������������������������������������������������171 Figure 5.20 Higher Besore House 1 and Truro College structure 1.����������������������������������������������������������������������������������172 Figure 5.21 Higher Besore Late Iron Age houses 2 (right) and 3 (left).�����������������������������������������������������������������������������174 Figure 5.22 Higher Besore Late Iron Age house 3 during excavation.������������������������������������������������������������������������������175 Figure 5.23 Dressel 1A amphora rim from ditch [2500], Higher Besore house 3.������������������������������������������������������������175 Figure 5.24 Higher Besore Late Iron Age house 4.���������������������������������������������������������������������������������������������������������������176 Figure 5.25 Higher Besore Late Iron Age house 5.���������������������������������������������������������������������������������������������������������������177 Figure 5.26 Excavation of the stone-filled ditch surrounding Higher Besore house 5.�������������������������������������������������178 Figure 5.27 Higher Besore Late Iron Age houses 6 and 7 and oval ancillary structure (top centre).���������������������������180 Figure 5.28 Higher Besore Late Iron Age houses 8 (bottom) and 9 (top left) and structure 6107.�������������������������������182 Figure 5.29 Higher Besore house 9 during excavation.�������������������������������������������������������������������������������������������������������184 Figure 5.30 Higher Besore Late Iron Age houses 10, 11 and 12.�����������������������������������������������������������������������������������������186 Figure 5.31 Sections through Truro College Late Iron Age ditches [2010] and [2022].��������������������������������������������������188 Figure 5.32 a (top) and b (bottom) 60 Two images of the La Téne influenced Late Iron Age brooch from Truro College structure 1.�������������������������������������������������������������������������������������������������������������������������������������������190 Figure 5.33 Ironwork fragment SF14.�������������������������������������������������������������������������������������������������������������������������������������192 Figure 5.34 Pottery: Higher Besore. P1–2 Beaker from structure 1; P3–4 Late Bronze Age from dyke pockets. �������194 Figure 5.35 Pottery: P5–8 Late Bronze Age Plain Ware. Higher Besore pit [6500]. ��������������������������������������������������������196 Figure 5.36 Pottery: Late Bronze Age Plain Ware. Truro College Area D, P41–2 hollow [5030], P43 pit [5015], P44–5 pit [5027], P46–7 hollow [5053]. ���������������������������������������������������������������������������������������������������������197 Figure 5.37 Pottery: P49–50 Late Bronze Age Plain Ware from Truro College pit [2009] in structure 3, upper fill (2000). �����������������������������������������������������������������������������������������������������������������������������������������������������������198 Figure 5.38 Pottery: Late Bronze Age Plain Ware from Truro College Pit [2009] in structure 3, upper fill (2000) P51–8; lower fill (2041) P59–60. ���������������������������������������������������������������������������������������������������������������������199 Figure 5.39 Pottery: Late Bronze Age Plain Ware from Truro College. P61 pit [2003] in structure 3; P62 pit [2067] structure 4; P63–65 pit [2045] Area B; P66–69 pit [5055] western group Area D; P70 pit v

[5059] northern group Area D. Middle to Late Iron Age from Truro College P71 structure 1, P72 field ditch [2022] Area B. ���������������������������������������������������������������������������������������������������������������������������������������������201 Figure 5.40 Pottery: P9–15 Middle to Late Iron Age from house 1, Higher Besore. �������������������������������������������������������207 Figure 5.41 Pottery: Middle to Late Iron Age from house 3, Higher Besore, P16–18 amphorae, P16–17 the same vessel. ������������������������������������������������������������������������������������������������������������������������������������������������������������������208 Figure 5.42 Pottery: P23–26 Middle to Late Iron Age from house 3, Higher Besore. �����������������������������������������������������209 Figure 5.43 Pottery: Middle to Late Iron Age from house 4, Higher Besore, P27–29, from House 6, Higher Besore, P30–32. �������������������������������������������������������������������������������������������������������������������������������������������������������������211 Figure 5.44 Pottery: Middle to Late Iron Age from house 8, Higher Besore, P33–35; P36 from field ditch [1000] in Area A. Scale �������������������������������������������������������������������������������������������������������������������������������������������������212 Figure 5.45 Late Bronze Age mullers from pit [6585] Higher Besore, structure 1. ��������������������������������������������������������220 Figure 5.46 Late Bronze Age mullers from pit [5030] Truro College Pits Area D East. ��������������������������������������������������220 Figure 5.47 Large saddle quern, S5 from [6585].������������������������������������������������������������������������������������������������������������������221 Figure 5.48 Double sided rubbing stone, S7 whetstone fragment, S8 small cobble with use as rubbing stone. �������221 Figure 5.49 S40, S41 [5030] Double-sided mullers. ��������������������������������������������������������������������������������������������������������������223 Figure 5.50 S40 [5030] Double-sided muller.�������������������������������������������������������������������������������������������������������������������������223 Figure 5.51 S42, S43, S44 Late Bronze Age mullers from pit [5027] Truro College pits Area D East. ���������������������������224 Figure 5.52 Late Bronze Age stonework. S45–46 mullers, S47 saddle quern fragment, from pit [5027] Truro College pits Area D East. S48 disc unstratified from Area D. ���������������������������������������������������������������������224 Figure 5.53 Late Bronze Age cushion stones S49–50 from pit [2009] Truro College structure 3. Iron Age muller S52 from Truro College field ditch [2010]. ��������������������������������������������������������������������������������������������������225 Figure 5.54 Stone S51 from posthole TC[2184].��������������������������������������������������������������������������������������������������������������������226 Figure 5.55 Iron Age stonework from Higher Besore house 1. S9 notched slate, S10–11 slickstones, S12– 13 whetstones. ��������������������������������������������������������������������������������������������������������������������������������������������������������������227 Figure 5.56 Iron Age stonework from Higher Besore. S14 rubbing stone [2569] house 3, S15 ?weight [3500] house 5, S16 muller [4500] house 6, S17-18 unfinished spindle whorls [5000] house 7. �����������������������228 Figure 5.57 Iron Age stonework from Higher Besore. House 8 [5500] S19 hammerstone, S20 slate knife, S21 ?weight, S22 ?unfinished spindle whorl, [5594] hammerstone. House 9 [6014] S25 grooved hammerstone. Field ditch [1014] S26 rubbing stone/pestle, field ditch [7011] S27 whetstone. ���������������������230 Figure 5.58 Grooved cobble from [5581] house 8. ���������������������������������������������������������������������������������������������������������������231 Figure 5.59 Photograph of grooved cobble from [5581] house 8. �������������������������������������������������������������������������������������231 Figure 5.60 The Late Bronze Age sword mould from pit [6500].����������������������������������������������������������������������������������������231 Figure 5.61 Reconstruction of the shape of the sword hilt based on external inspection, radiography and (beyond the limits of the mould) closest parallels.�������������������������������������������������������������������������������������������������232 Figure 5.62 a and b Radiographs of the sword mould fragment from pit [6500]; top) in plan view; bottom) side view. �������������������������������������������������������������������������������������������������������������������������������������������������������233 Figure 5.63 Fired clay lump from (2041).�������������������������������������������������������������������������������������������������������������������������������238 Figure 5.64 Analyses of REEs of residues from iron smelting. Data normalised against Upper Crust standard of Taylor & McLennan 1981.�����������������������������������������������������������������������������������������������������������������������242 Figure 5.65 Analyses of REEs of residues from iron smelting. Data normalised against average ore composition.�242 Figure 5.66 Analyses of REEs of residues from iron smithing and technical ceramics. Data normalised against Upper Crust standard of Taylor and McLennan 1981.������������������������������������������������������������������������������242 Figure 5.67 Backscattered electron photomicrographs of polished blocks from representative specimens.������������244 Figure 5.68 Comparison of Higher Besore and courtyard settlement house layout.�����������������������������������������������������276

Section 6: Excavation of an Iron Age Settlement and a Roman Period Enclosure at Porthleven, 2014

Figure 6.1 Site location, excavation area, outlined in red.�������������������������������������������������������������������������������������������������283 Figure 6.2 Plan showing enclosure 1 and the main features uncovered in the excavation area.���������������������������������284 vi

Figure 6.3 Aerial photograph showing the site prior to development (© Cornwall Council 2005).����������������������������284 Figure 6.4 Roman period brooch CORN-09540E of first century AD date located near to the site (Photograph: Anna Tyacke, © Royal Institution of Cornwall).������������������������������������������������������������������������������285 Figure 6.5 Plan of features within group D.��������������������������������������������������������������������������������������������������������������������������287 Figure 6.6 Plan of features forming structure A1 and structure A2.���������������������������������������������������������������������������������288 Figure 6.7 Structure A1showing large central pit, viewed from the south.���������������������������������������������������������������������289 Figure 6.8 Plan of features forming roundhouses, structure B1 and structure B2.��������������������������������������������������������290 Figure 6.9 Structure B1showing post-ring and central features, viewed from the south.��������������������������������������������290 Figure 6.10 Plan of features within group E.�������������������������������������������������������������������������������������������������������������������������292 Figure 6.11 Plan of enclosure 1 and features within it.�������������������������������������������������������������������������������������������������������294 Figure 6.12 Photograph of enclosure 1 ditch [311] looking north west prior to section being cut back.�������������������294 Figure 6.13 Photograph of enclosure 1 ditch [311], with the bottom of furnace 806, exposed within the fill of the enclosure ditch. The furnace had clearly been cut into the ditch after it had been infilled.���������295 Figure 6.14 Plan showing internal features forming clusters S1 and S2 inside Enclosure 1.����������������������������������������296 Figure 6.15 Photograph of decorated hearth 704 after exposure.�������������������������������������������������������������������������������������297 Figure 6.16 Photograph of the layer (701) sealing hearth 704.������������������������������������������������������������������������������������������297 Figure 6.17 Photograph of stone-capped gully [772].����������������������������������������������������������������������������������������������������������299 Figure 6.18 Plan of features within group C.������������������������������������������������������������������������������������������������������������������������300 Figure 6.19 Beaker rims, P1 centre and left, P2 right. (Photograph: Gary Young.)��������������������������������������������������������302 Figure 6.20 Small Earliest Iron Age vessel P3. (Drawing: Jane Read.)�������������������������������������������������������������������������������303 Figure 6.21 The decorated hearth 704 after conservation. (Photograph: Ryan Smith.)������������������������������������������������305 Figure 6.22 Close-up of one of the rings in decorated hearth 704. (Photograph: Ryan Smith.)�����������������������������������305 Figure 6.23 Mould S1 from [155] group D Iron Age. (Photograph: Gary Young, drawing Jane Read.)�������������������������307 Figure 6.24 Grooved hammerstone S13 from [210] Iron Age (Photograph: Gary Young, drawing Jane Read.)����������307 Figure 6.25 Quartzite beach boulder, with slightly worn surface S29, unstratified find located within enclosure 1 with slightly worn surface. 25cm scale. (Photograph: Andy M Jones.)������������������������������������������308 Figure 6.26 Fragment of unfinished mould S2 from [715]. (Photograph: Gary Young, drawing Jane Read.)�������������309 Figure 6.27 Pivot bearing S3 (701) from soil around hearth 704. (Photograph: Gary Young, drawing Jane Read.)���309 Figure 6.28 Upper rotary quern fragment S7 (701), from soil around hearth 704. (Photograph: Gary Young, drawing Jane Read.)����������������������������������������������������������������������������������������������������������������������������������������310 Figure 6.29 Upper rotary quern S8 (794) from fill of ‘flue’ [792]. (Photograph: Gary Young, drawing Jane Read.)���310 Figure 6.30 Grooved hammerstone S15 (701) from soil around hearth 704. (Photograph: Gary Young drawing Jane Read.)������������������������������������������������������������������������������������������������������������������������������������������������������310 Figure 6.31 All flintwork – L1, L2, L3, L4, L5, L6, L7, L8, L9 and L10 (from top left, clockwise). (Photograph: Anna Lawson-Jones.)����������������������������������������������������������������������������������������������������������������������������������������������������312 Figure 6.32 Grooved hammerstone found in a ditch near, Tregeseal, West Penwith.����������������������������������������������������321 Figure 6.33 Reconstructed four-post structure at Castell Henllys. (Photograph: Andy M Jones.)�������������������������������323 Figure 6.34 Late Iron Age terret CORN-F6FD0F. (Photograph: Anna Tyacke, © Royal Institution of Cornwall.)�������326 Figure 6.35 Photograph of white stones around decorated hearth 704.��������������������������������������������������������������������������329

Section 7: Review and Overview

Figure 7.1 Photograph of the quartz filled gully encircling house 9, Higher Besore.����������������������������������������������������335 Figure 7.2 Plan of the Late Iron Age round at Threemilestone. (Source: Schwieso 1976.)��������������������������������������������337 Figure 7.3 Plan of the Middle Iron Age enclosure at Boden. (Source: Gossip 2013.)������������������������������������������������������338 Figure 7.4 Map showing the distribution of enclosed later prehistoric and Roman period settlements in lowland Cornwall.���������������������������������������������������������������������������������������������������������������������������������������������������������339 Figure 7.5 Stone bowl, fragment from Trethurgy Road.�����������������������������������������������������������������������������������������������������339 Figure 7.6 Cornish type brooch CORN-DEC722. (Photograph: Anna Tyacke, © Royal Institution of Cornwall.)��������340 vii

Figure 7.7 Plan of Trethurgy Round, showing the organized layout of the enclosed settlement. (Source: Quinnell 2004.)��������������������������������������������������������������������������������������������������������������������������������������������������������������341 Figure 7.8 Pit [124], Tremough, half excavated. Note burnt stones in the fill.����������������������������������������������������������������343 Figure 7.9 Worked stone objects found within Higher Besore pit [5027].������������������������������������������������������������������������344 Figure 7.10 The Iron Age settlement at Higher Besore reconstructed. (Painting: George Scott.)�������������������������������345

List of Tables Section 2: Archaeological Recording during the 1996 Coast Protection Scheme at Porth Killier, St Agnes, Isles of Scilly

Table 2.1: Zones – key features and layers.�����������������������������������������������������������������������������������������������������������������������������15 Table 2.2: Summary pottery quantifications by Zone (P5–P8 allocated to illustrated vessels from the Coastal Erosion Project).������������������������������������������������������������������������������������������������������������������������������������������������24 Table 2.3: Summary of pottery from Zone A; contexts arranged so that earliest are at top, gradually becoming later down the table.������������������������������������������������������������������������������������������������������������������������������������25 Table 2.4: Material from the Coastal Erosion Project, arranged in sequence of 1996 stratigraphy. For details see Table 2.5. For contexts with no 1996 correlation, see Quinnell 1994; their approximate stratigraphic positions can be deduced from position in the table.����������������������������������������������������������������������26 Table 2.5: Summary of pottery from Zone B.��������������������������������������������������������������������������������������������������������������������������28 Table 2.6: Summary of pottery from Zone C.��������������������������������������������������������������������������������������������������������������������������29 Table 2.7: Stone artefacts.����������������������������������������������������������������������������������������������������������������������������������������������������������38 Table 2.8: Flint.����������������������������������������������������������������������������������������������������������������������������������������������������������������������������39 Table 2.9: Porth Killier, number of bones identified to species in all Zones.���������������������������������������������������������������������43 Table 2.10: Porth Killier, Zone A, bone from Old Land Surface contexts.���������������������������������������������������������������������������43 Table 2.11: Porth Killier, Zone A, bone from post-settlement soil.�������������������������������������������������������������������������������������43 Table 2.12: Porth Killier, Zone A, bone from the pre-roundhouse midden.�����������������������������������������������������������������������43 Table 2.13: Porth Killier, Zone A, bone from the roundhouse midden.������������������������������������������������������������������������������43 Table 2.14: Porth Killier, Zone A, bone from roundhouse 21.�����������������������������������������������������������������������������������������������44 Table 2.15: Porth Killier, Zone A, soils for debate?�����������������������������������������������������������������������������������������������������������������44 Table 2.16: Porth Killier, bones from Zone B.��������������������������������������������������������������������������������������������������������������������������44 Table 2.17: Porth Killier, bones from Zone C.��������������������������������������������������������������������������������������������������������������������������44 Table 2.18: Porth Killier, analysis of molluscs from the seven most shell-rich samples to show, 1: size range of limpets; 2: numbers and weights of limpet shell components; 3: presence of other mollusc species; 4: density of limpet component in relation to weight of bulk soil samples collected.����������50 Table 2.19: Porth Killier, analysis of molluscs from 22 samples of shell obtained from bulk soil samples excavated by CAU, May/June 1996.������������������������������������������������������������������������������������������������������������������������������51 Table 2.20: Porth Killier 1996 plant macrofossil assessment.����������������������������������������������������������������������������������������������55 Table 2.21: Radiocarbon determinations from Porth Killier.�����������������������������������������������������������������������������������������������56

Section 3: Excavations at Killigrew 1996: an Iron Age and Romano-British Industrial Site on the Trispen Bypass, Cornwall

Table 3.1: AMS dates from furnace [82]. ���������������������������������������������������������������������������������������������������������������������������������81 Table 3.2: Charred plant Macrofossils from Killigrew round.����������������������������������������������������������������������������������������������97 viii

Section 4: Archaeological Investigations at Nancemere, Truro, Cornwall 2002: a Prehistoric and Romano-British Landscape

Table 4.1 Details of Trevisker material from Nancemere.��������������������������������������������������������������������������������������������������128 Table 4.2 Standard gabbroic pottery from the entrance; includes a well-made gabbroic sherd from (1114); also mortarium from (1027) 23g, BB1, 3g, from (1112) and crucible, 3g, from (1018).�������������������������130 Table 4.3: All Roman period pottery except that in entrance way. ����������������������������������������������������������������������������������132 Table 4.4: Charcoal: identified taxa.���������������������������������������������������������������������������������������������������������������������������������������138 Table 4.5 NISP (number of identified species) and MNI (minimum number of individuals).���������������������������������������140 Table 4.6 Results from the radiocarbon dating.�������������������������������������������������������������������������������������������������������������������142

Section 5: Life Outside the Round: Bronze Age and Iron Age Settlement at Higher Besore and Truro College, Threemilestone, Truro, 2004–5

Table 5.1: Summary of Bronze Age pottery in broad chronological order (number of sherds/weight (g); HB=Higher Besore; TC=Truro College).���������������������������������������������������������������������������������������������������������������������192 Table 5.2: Summary of Middle to Late Iron Age pottery (number of sherds/weight (g)).��������������������������������������������202 Table 5.3: Internal neck diameters of the 11 illustrated BD6 vessels for which this dimension can be reliably obtained.����������������������������������������������������������������������������������������������������������������������������������������������������������203 Table 5.4: Dates and ceramic assemblages within houses; houses without Cordoned Ware are listed first.�������������203 Table 5.5: Distribution of sherds within structures (sherds /weight). Termini are defined as the 3m on either side of the entrance. ‘Other’ is sherds found over the structures or in features outside.��������������������206 Table 5.6: Sources of used and /or modified stonework: this does not include 40 Iron Age local notched slates, 14 Iron Age local quartz crystals and 12 Iron Age unused white quartz beach pebbles.����������������������216 Table 5.7: Stonework from Higher Besore Late Bronze Age contexts. In all tables f-g = fine-grained and m-g = medium grained.������������������������������������������������������������������������������������������������������������������������������������������������219 Table 5.8: Stonework from Truro College Area D east pits.������������������������������������������������������������������������������������������������222 Table 5.9: Stonework from Truro College: other Late Bronze Age contexts.�������������������������������������������������������������������225 Table 5.10: Stonework from Higher Besore house 1.�����������������������������������������������������������������������������������������������������������226 Table 5.11: Stonework from Higher Besore houses 3–7.�����������������������������������������������������������������������������������������������������227 Table 5.12: Stonework from Higher Besore houses 8–9.�����������������������������������������������������������������������������������������������������229 Table 5.13: Stonework from Higher Besore and Truro College Field Ditches.�����������������������������������������������������������������229 Table 5.14: Features of swords which best match the matrix from Higher Besore.�������������������������������������������������������236 Table 5.15: Sizes of wattle impressions in fired clay from [6505].�������������������������������������������������������������������������������������239 Table 5.16: Lithic assemblage.�������������������������������������������������������������������������������������������������������������������������������������������������241 Table 5.17: Truro College charred plant remains by area.��������������������������������������������������������������������������������������������������252 Table 5.18: Truro College playing fields: Charcoal from the Late Bronze Age and Iron Age settlements.������������������261 Table 5.19: Higher Besore calibrated date ranges.���������������������������������������������������������������������������������������������������������������263 Table 5.20: Truro College calibrated date ranges.����������������������������������������������������������������������������������������������������������������264 Table 5.21: Truro College Late Bronze Age post-rings – structural details.���������������������������������������������������������������������269 Table 5.22: Dimensions / area of internal living space of roofed domestic structures.�������������������������������������������������271 Table 5.23: Higher Besore and Truro College Late Iron Age structural attributes.��������������������������������������������������������274 Table 5.24: Characteristics of Iron Age and Romano-British structures at Chysauster, Carn Euny and Trethurgy.�274

Section 6: Excavation of an Iron Age Settlement and a Roman Period Enclosure at Porthleven, 2014

Table 6.1: Finds 1 Iron Age pottery by sherd number and weight in grams presented by Group.������������������������������302 Table 6.2: Roman period material; archive table in context order, by fabrics, sherd numbers and weight in grams.�������������������������������������������������������������������������������������������������������������������������������������������������������������������������304 Table 6.3: Quantification of non-local pottery.���������������������������������������������������������������������������������������������������������������������304 ix

Table 6.4: Analysis catalogue for Porthleven Shrubberies flintwork.������������������������������������������������������������������������������313 Table 6.5: Charcoal from Beaker pit [148] and Iron Age features.�������������������������������������������������������������������������������������315 Table 6.6: Charcoal from Romano-British features.�������������������������������������������������������������������������������������������������������������316 Table 6.7: Fragment inventory.�����������������������������������������������������������������������������������������������������������������������������������������������318 Table 6.8: Metrics.���������������������������������������������������������������������������������������������������������������������������������������������������������������������318 Table 6.9: Radiocarbon dating from Porthleven.�����������������������������������������������������������������������������������������������������������������319

x

Acknowledgements The editors would like to thank Henrietta Quinnell, Freya Lawson-Jones and Sara Homes for reading the finalized chapters. Porth Killier, St Agnes The programme of archaeological recording was commissioned by the Council of the Isles of Scilly. The Coastal Protection Scheme (CPS) was designed by consultant engineers Aspen Burrow Crocker, and constructed by T J Brent Ltd. Alison Locker would like to thank Dr Terry O’Connor (University of York) for identifying the grey seal immature pelvis and Dr Jacqui Mulville (University of Cardiff) for information on grey seal breeding habits. Jan Light is grateful to Steve Hawkins and Roger Burrows for allowing her access to their unpublished data on Patella spp. Within Cornwall Archaeological Unit, the 1996 fieldwork was undertaken by Andrew Young and Carl Thorpe, soil samples were sieved by Ann Reynolds, the archive report was compiled by Jeanette Ratcliffe and Andrew Young in 1997, and an unfinished site report (1999) begun by Jeanette Ratcliffe, who was the project manager. The report was edited and completed between 2006 and 2018 by Charles Johns and Andy M Jones, with illustrations by Sean Taylor and Carl Thorpe. Revisions to drawings were made by Connor Motley. Killigrew round The authors would like to thank Transportation and Estates division of Cornwall County Council for commissioning them to undertake the project. Expert advice was given by Vanessa Straker (Historic England), Henrietta Quinnell, Dr David Starley and Dr Justine Bayley of the Ancient Monuments Laboratory at the Centre for Archaeology, English Heritage, Margaret Brooks and Virginia Neal, Wiltshire Consultation Services. We are also grateful for the contributions of Albertine Malham and Gerry McDonnell (Ancient Metallurgy Research Group, University of Bradford), Dr R. Taylor and Sophie Lamb. Many thanks are also extended to Anna Tyacke, Jane Marley and Laura Ratcliffe-Warren of the Royal Cornwall Museum, Truro for advice and help with the storage and conservation of the Killigrew tin dish. Scientific dating was carried out by the University of Waikato Radiocarbon Dating Laboratory, New Zealand. Many thanks are also extended to Peter Rose and Henrietta Quinnell for their comments on earlier versions of the text. We are very grateful for the insights and valuable comments by colleagues Adam Sharpe and Andrew Young which have helped sharpen ideas on metalworking and enclosures in general. The Cornwall Archaeological Unit excavation team comprised Anna Lawson-Jones, Ann Reynolds and Carl Thorpe. Jane Powning drew the maps, Dick Cole the plans, and Carl Thorpe produced the artefact drawings. Revisions to the drawings were made by Connor Motley. A considerable number of unpaid volunteers worked on site. Many were Archaeology A-level students at St Austell and Truro Colleges and members of the public. The project was managed by Jacky Nowakowski. The authors would also like to thank the following people for their help with the excavation: Charlotte Andrew, Chris Barber, Phil Best, Andrew Cochrane, Mark Cole, Kevin Denton, Kirk Denton, Becky Edwards, Hannah Edwards, Steve Ellis, Ben Found, Polly Fryer, Murray Gould, Steven Grey, Richard Harwood, Val Hazel, David Henzell, Jo Higgins, James Keasley, Liz Keasley, Pam Lee, Darren Limbert, Jack Major, Calvin Malham, Mitch Mitchell, Mike and Tina O’Connor, Phil Oggleby, Andy Pearce, Isabel Popple, Stuart Pulley, Hannah Saint, Sophy Savage, Jago Titcomb, Andrew Venton, Imogen Wood, Pam Worthington and Duncan Yeates. Nancemere Thanks to Cofton Ltd for funding this project. Many thanks also to Sally Ealey, Pam Lee and Konstanze Rahn of Cornwall Archaeological Society for their assistance in processing environmental samples and finds. Thanks to all xi

those specialists and colleagues who have provided advice throughout the post-excavation stage and especially to the field team who were Dick Cole, Neil Craze, Peter Dudley (supervisor), Matt Mossop, Alex Osborne, Stuart Randall, Emma Ruddle, Sean Taylor, Anna Tyacke, Megan Val Baker, Dave Williams and Andrew Young. The project was managed by Andy M Jones. Many thanks also to Sally Ealey, Pam Lee and Konstanze Rahn of Cornwall Archaeological Society for their assistance in processing environmental samples and finds. Higher Besore The project at Higher Besore was funded by Interserve Project Services Ltd, CCC Planning Transportation and Estates Property Resources Group and Capita Symonds Ltd. Thanks to the many specialists involved in the project whose work has contributed to this report: Justine Bayley, Paul Bidwell, Sarnia Butcher, Wendy Carruthers, Rowena Gale, Alan Hogg, J D Hill, Julie Jones, Anna Lawson-Jones, Roger McBride, Stuart Needham, Peter Northover, Cynthia Poole, Henrietta Quinnell, Roger Taylor, Carl Thorpe, Anna Tyacke and Tim Young. Special thanks go to Henrietta Quinnell for the many useful comments and advice on the draft report. Henrietta Quinnell is grateful to E L Morris for discussion on the Bronze Age plain ware from Dorset. Within Cornwall Archaeological Unit, the project manager was Charles Johns and the fieldwork team comprised Pete Dudley, Megan Val Baker, Neil Craze, Carmello Grasso, Lynne Hendy, Anna Lawson-Jones, Hilary Orange, Stuart Randall, Theresa Rowell, Emma Ruddle, Jens Samuel, Marc Steinmetzer, Jo Sturgess, Helen Thomas, Anna Tyacke, Katie Watkins and Imogen Wood, site archaeologists. Carl Thorpe produced artefact illustrations and Graeme Kirkham and Peter Rose provided useful comments on the text. Thanks are also extended to Cornwall Archaeology Society volunteers, especially Michael Blake, Jean Hughes, Adrian Rodda, Trudy Staynings, Mick Triplett and Christine Wilson. Porthleven This Porthleven project was commissioned by 3MS Construction and carried out by Cornwall Archaeological Unit, Cornwall Council. The Project Manager was Andy M Jones and the Project Officer was Ryan Smith. The excavation team comprised Hayley Goacher, Paul Reddish, Ian Rose, Emma Ruddle and Megan Val Baker, with volunteers Martin Andrews, Bret Archer and Richard Mikulski. Maps and plans were produced by George Scott. Thanks to Anna Tyacke for information and photographs of the Roman period brooch and the Late Iron Age terret. Henrietta Quinnell thanks Cynthia Poole for constructive comment upon the hearth and Ruth Shaffrey for that on the pivot bearing.

xii

Section 1: Background to the Project Andy M Jones

1

2

Chapter 1: Introduction to the Volume

Despite the sometimes very long delays to publication, these programmes of archaeological investigation are very significant as they have revealed several settlement phases, which span the later prehistoric to Roman periods. They include Middle and Late Bronze Age roundhouses, field boundaries and unenclosed settlements of Iron Age date, and enclosures belonging to the Late Iron Age Roman period. Remarkably, despite the varying scale of the archaeological recording, when taken together the excavated evidence from all the investigated sites provide interesting and often broadly comparable sequences. For example, all four of the mainland sites reveal growing evidence for enclosure and occupation in the Late Iron Age and early Roman periods. Two of the sites also seem to have developed along specialist lines and were perhaps more closely associated with metalworking than inhabitation

Background In the period between 1996 and 2014 Cornwall Archaeological Unit, Cornwall Council, undertook archaeological investigations at five later prehistoric and Roman period settlements around Cornwall and the Isles of Scilly, which for a variety of reasons, have until now remained unpublished (Figure 1.1). Of these sites, three were small-scale and evaluative, as at Nancemere, targeted in response to coastal erosion as at Porth Killier, or carried out in challenging conditions as at Killigrew round (Figure 1.2). Higher Besore and Porthleven by contrast, were more substantive excavations, which resulted in larger areas being soil stripped.

Figure 1.1 Map of Cornwall showing sites covered by this monograph.

3

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 1.2 Killigrew 1996, general view of excavation looking north.

(Killigrew and Porthleven). Three of the sites in mid Cornwall are located either on the fringes of Truro (Higher Besore and Nancemere) or a few kilometres to the north of the city (Killigrew). This also allows for comparisons and contrasts to be made between them.

is varied in character but does reflect the then current context in which the work was carried out. For the purposes of this monograph we have therefore decided to leave these largely as completed by their authors, with a limited amount of updating, for example to reflect the publication of works which are no longer ‘forthcoming’. We have also updated radiocarbon dating citations, so that all quoted dates use a standard calibration curve.

In the light of the very significant cumulative results generated by these projects relating to the later prehistoric and Roman periods, the decision was made to draw them together into a single publication, which would provide an overview of settlement activity across Cornwall and the Isles of Scilly in later prehistory.

We have, however, not altered the content of specialist reports, which by and large were written close to the time of excavation. This means that, particularly the older sites do not contain references to more recent investigations, for example the TEDC site (Taylor, forthcoming), which is located near to both Killigrew and Nancemere is not referred to in either report. The arranging of chapters in the sequence in which they were excavated with the oldest first does nonetheless help reflect changing aims and objectives and perspectives as one progresses through the volume.

Report structure Following this introductory section, the resulting monograph is divided into five parts with a sixth discursive section at the end. Given the range of sites included and the detailed specialist work which has been done on many of them, it was decided that each section would be self-contained so that they can be read as separate, free-standing contributions. They have, however, been arranged in chronological order, by year of excavation, so that they can be read sequentially. A concise interpretative overview highlighting some of the themes to emerge is given at the end.

Any remaining potential inconsistencies between ideas expressed in the discursive chapters is remedied by the seventh and final section. This final synthetic chapter is concerned with the discussion of the more significant themes which have arisen from all five of the investigated sites, and also draws on other more recent work undertaken in Cornwall over the last decade.

Each section commences with the stratigraphical results from the project organized by chronological periods. This is followed by successive specialist chapters on the artefacts, the environmental analysis and radiocarbon dating. Each section concludes with the discussion. Given that each piece of work and its analysis was undertaken by a range of authors at different times, the style and focus of the discussions

Terminology used in this report Throughout this report structures are denoted by numbers without brackets; for example, House 4. Context numbers for cuts – ditches, pits, postholes and 4

Chapter 1: Introduction to the Volume similar features – are shown in square brackets: [126] and their fills, layers and other deposits are shown with round brackets: (126).

use the term ‘later prehistoric settlement’ as opposed to ‘later prehistoric and Roman settlement’ in the title of this monograph. This does not imply that the Roman Conquest and the subsequent colonial regime did not bring about disruption and very significant changes to the fabric of communities in Cornwall either in terms of administration or in social organization. Indeed, the potential implications of this process are touched upon in the synthesis at the end of the volume (Section 7). Instead, ‘later prehistoric’ is used in part to highlight continuity in some areas, such as the construction of settlement enclosures that had already started, on an albeit smaller scale, in the later Iron Age, as well as patterns of structured deposition which had considerably deeper roots. The title, however, also partly reflects the fact that Cornwall was distant from the nearest administrative centre in Exeter, and consequently little is directly known either about its governance or indeed of any of its inhabitants’ lives, whose names remain, like those of their Iron Age forebears, undocumented.

The term ring-gully is used throughout the report to denote ditching around the perimeter of both structures and hollows of circular or oval shape. Specialist reports are, for the most part, reproduced as accepted edited manuscripts and consequently terminology used within in them will reflect the time that they were written. The radiocarbon dating probability distributions (Sections 2, 3, 4, 5 and 6) were calculated using OxCal (v4.3). Unless stated otherwise, the 95 per cent level of probability has been used throughout this volume; calibrated determinations cited in the text may therefore differ from older published sources. Finally, despite the chronological range of the settlements under discussion, the decision was made to

5

6

Section 2: Archaeological Recording during the 1996 Coast Protection Scheme at Porth Killier, St Agnes, Isles of Scilly Charles Johns, Jeanette Ratcliffe and Andrew Young, with contributions from David Dungworth, Janice Light, Alison Locker, Henrietta Quinnell, Vanessa Straker and Roger Taylor In 1996 the Cornwall Archaeological Unit carried out a programme of archaeological recording associated with a Coast Protection Scheme on the north side of St Agnes in the Isles of Scilly. Work was mainly focused in Porth Killier, where building a new sea wall had an impact on nationally important Bronze Age remains exposed in the low cliff face. A small amount of recording also took place in Porth Coose, where a submerged peat deposit and stone walling are located towards the top of the beach. At Porth Killier the results fell into four zones. Working from east to west along the cliff face, these can be summarized as follows: Zone A, Bronze Age buildings, midden and a wall; Zone B, a series of Bronze Age pits, Zone C, an Early Bronze Age cairn or entrance grave and a prehistoric wall, Zone D, marine sand with no archaeological remains.

7

8

Chapter 2.1: Background to the Excavations

During May and June 1996, Cornwall Archaeological Unit (CAU) carried out a programme of archaeological recording associated with a Coast Protection Scheme (CPS) on the north side of the island of St Agnes on behalf of the Council of the Isles of Scilly (Figures 2.1, 2.2 and 2.22). The need for archaeological recording had been identified in the Environmental Statement prepared for the scheme (Nicholas Pearson Associates 1995). The archaeological work was mainly focused in Porth Killier, where building a new sea wall had an impact on nationally important Bronze Age remains exposed in the low cliff face (prehistoric settlement and field system close to Porth Killier; National Heritage List Entry 1014998). The overall aim of the 1996 recording work was to gain as much information as possible about the date, character and function of the site prior to the new sea wall being constructed in front of it.

Location and landscape setting The project area is located on the north side of the island of St Agnes in the Isles of Scilly (Figure 2.1). Here a flat, low-lying neck of land contains Big Pool, a Site of Special Scientific Interest (SSSI) and the only open fresh water on the island, apart from the much smaller and shallower Little Pool to the east (Figures 2.1, 2.2 and 2.3). Small pasture and bulb fields (enclosed by stone walls and hedging plants) occupy the headland to the north and the ground to the south, which slopes steadily up towards the island’s central, east-west ridge. The neck, however, is unenclosed and covered in rough grassland. It lies only just above high water (the surface of Big Pool being approximately 2.5m OD) and low dune banks separate it from Porth Killier to the east and the bays of Porth Coose and Periglis to the west. Prior to the

Figure 2.1 Location map showing Isles of Scilly and area of excavation.

9

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 2.2 Modern Landline mapping (area of excavation shown in red box).

Farm. During 1989, as an emergency coast protection measure, boulders were dumped at either end of this wall by the Royal Marines.

CPS, these banks rose only 1–2m above the level of Big Pool. They were originally formed by tidal and aeolian action but had been built up by islanders (using large stones and builders’ rubble) in an effort to counteract erosion by the sea. The water within the three bays is shallow and at low water the intertidal zone stretches out for some distance, exposing expanses of sand and seaweed-covered rock.

The other bays have sandy beaches, backed by dune banks. Periglis, which is sheltered on the north by Burnt Island, was once the main landing place for St Agnes and still serves as an anchorage for small fishing boats and sailing vessels. Porth Coose is a smaller, more exposed beach, facing north-west and bounded by rocky promontories – Porth Coose Carn to the north-east and Ginamoney Carn to the south-west. Prior to 1996, there was a dilapidated groyne (made up of stone and steel rails from the old lifeboat slipway) extending from the shore to Ginamoney Carn and separating Porth Coose from Periglis.

Facing north-north-east, Porth Killier is the largest and most rugged of the three bays, with a steep beach of sand, shingles, and cobbles, giving way to an expanse of bedrock and boulders. Prior to the 1996 scheme, the bay was enclosed by a dune bank (on its west and south side) and a low, crumbling, vertical cliff. The latter, comprised soil layers (containing archaeological features and artefacts) overlying ram (the natural granite subsoil) and raised beach deposits, extended along the eastern side of Porth Killier, out to the scrubcovered, rocky headland of Kallimay Point. In the southeastern corner of the bay the cliff was interrupted by a protuberance of outcropping granite. Along the bay’s south side, at the junction between dune and cliff, there were the 5m long remains of a 1m high granite and concrete sea wall, constructed in 1931 to protect a clifftop track and fields behind associated with Lower Town

Modern day settlement in the vicinity of the CPS consists of Lower Town Farm and a cluster of houses around St Agnes Church. The surrounding farmland is comprised of bulb strips hedged with tamarisk, pittosporum, and euonymus, which are used for flower production, and stone-walled fields which are mainly under pasture. The latter constitute Anciently Enclosed Land of prehistoric or medieval origin (Land Use Consultants 1996). 10

Chapter 2.1: Background to the Excavations

Figure 2.3 Detail from 1876 OS map (area of excavation shown in red box).

The geology of St Agnes is granite, with weathered periglacial head, known locally as ram, covering the lower hill slopes and valley floors; the geology supports soils suitable for cultivation and pasture (Geological Survey of Great Britain, 1975, Isles of Scilly, Sheets 357 and 358).

archaeologically-rich length of cliff face in Scilly. Documented for over 60 years, it is best known for its extensive limpet middens, which over this time have yielded a significant amount of artefactual and palaeoenvironmental material. Together with the walls of stone roundhouses, they represent the remains of a Bronze Age settlement whose inhabitants exploited marine resources (fish, seals, shellfish, and seabirds), as well as, to a lesser extent, practising farming and other land-based activities. Prior to 1996, there was nothing about the site to suggest that it was anything more than a domestic settlement, but the CPS work exposed a series of pits, which may have had a semi-industrial function, together with what appears to be the remains of a burial cairn (with a stone-lined cist or chamber).

Coastal change in the last 100 years By comparing the 1876 25-inch map (Figure 2.3) with the modern (1981) 1:2500 OS map, it is possible to gain some idea of the extent to which the cliff in the southeast corner of Porth Killier has eroded during the last 100 years. In 1876 the drystone granite wall forming the north-west side of the small wedge-shaped field lying immediately behind the cliff still existed in its entirety (albeit in places right on the cliff edge). By 1981 this wall only survived at its north-east end, where the projecting rock outcrop had prevented erosion. Elsewhere wave action had cut back the cliff edge by at least 1.5 m, the field wall having presumably collapsed onto the shore below (though stones from it may have been reused in the 1931 sea wall).

The site was first recorded by amateur archaeologist Alec Gray in 1936, at which time the only recognisable feature was a midden yielding Bronze Age pottery (Gray 1972, 43). The only bones found were seal, leading Gray to suggest this was a seasonal habitation site where seal and fish meat was preserved. After 1936 more remains were exposed and destroyed by coastal erosion, and quantities of pottery, flint, animal bone, charcoal and unfired clay, and a granite quern rubber were collected from the cliff face, together with what was thought to be a fragment of a Bronze knife (now lost). The animal

Archaeological and historical background The low crumbling cliff in the south-east corner of Porth Killier, constitutes arguably the most 11

Later Prehistoric Settlement in Cornwall and the Isles of Scilly species identified were sheep, ox and cetacean, the last probably a porpoise (Turk 1968, 78).

recorded to the north-east. Numerous artefacts were retrieved, including a Romano-British brooch found by amateur archaeologist, Michael Tangye.

During October 1985 archaeologists from the Institute of Cornish Studies (ICS) made a sketch section drawing of the (three?) roundhouses and four limpet middens then visible (Ratcliffe and Thorpe 1991, 12). Numerous sherds of Bronze Age pottery (and several Iron Age/ Romano-British and early medieval) were retrieved from the cliff face, together with a couple of flint flakes and animal bone representing the remains of a variety of mammals (ox, sheep, red deer, horse and dolphin), birds (domestic fowl, razorbill, stone curlew, coot, and corncrake), fish (red gurnard, coalfish, grey mullet and pollack) (ibid, 71, 82, 105).

Many of the features and layers exposed in the cliff face in 1989 (and subsequent years) were still recognisable in 1996. In order to make it possible to tie in the results of previous work with those from the 1996 recording, a table correlating the two sets of site context numbers is included in the project archive. Project background St Agnes Coastal Protection Scheme The need for archaeological recording was identified in the Environmental Statement prepared for the scheme (Nicholas Pearson Associates 1995). The St Agnes CPS was designed by Aspen Burrow Crocker, consultant engineers working for the Council of the Isles of Scilly, and T J Brent were contracted to carry out the construction work which consisted of three different types of work:

CAU started monitoring the site in 1988 and in March 1989 English Heritage’s Ancient Monuments Laboratory carried out resistivity and magnetometer surveys in the field behind the cliff exposure to detect the presence of buried archaeological remains. The survey results suggested that the settlement extended inland for a distance of up to 12m (Jordan 1989; Ratcliffe and Parkes 1989, 9–11).

Construction of a granite-faced mass concrete sea wall – this extended westward for 80m from the rock outcrop in the south-eastern corner of Porth Killier.

During September 1989 a detailed record was made of a 28m length of cliff face (Ratcliffe and Parkes 1990, 27–32), which identified: the eroding remnant of a substantial circular stone building, a second structure comprising two pieces of walling, nine limpet middens, two layers of dark humic material apparently representing a limpet-free midden from which (unusually) limpets were absent; a buried land surface(s) predating the use of the settlement and post-occupation layers that had accumulated since its abandonment.

Reinforcement of the dune – this took place in Porth Coose and along the southern side of Porth Killier, and involved the use of Armorflex concrete block revetment, Enkamat erosion control matting, and imported Cornish granite. Raising the dune height – this was carried out in Periglis, using Enkamat and imported fill.

Numerous sherds of second millennium BC pottery were collected and bulk samples were taken for environmental analysis. The bones of a wide range of domestic and wild animals were identified, including seven fish and 10 bird species not previously recorded from the site. Cultivated plants (barley, emmer wheat and celtic bean) were identified amongst the plant macrofossils, which suggested that the settlement was set in a coastal environment which included blown sand, grassland and boggy areas, as well as arable fields and heathland. Radiocarbon dating revealed that the most extensive limpet midden had formed during the Middle Bronze Age, while the non-limpet midden was of Late Bronze Age date (Ratcliffe and Straker 1996, 62–73).

Archaeological impact In Periglis the nature of the CPS work mitigated against any disturbance of archaeological remains. In the other two bays, however, the scheme had an archaeological impact. Preparation work for the construction of the sea wall in Porth Killier involved removal of the short length of 1930s sea wall and boulders dumped by the Royal Marines during 1989, trimming back the cliff edge at one location (between 50m and 60m from the eastern end) and the excavation of a 1m deep foundation trench into the shelf of subsoil (ram) in front of the cliff face. The sea wall was constructed of successive layers of concrete behind a random granite face.

Over the following four years (1990–1993) the site was revisited as part of CAU’s annual coastal monitoring programme (Ratcliffe and Sharpe 1991, 48–50; Ratcliffe 1993, 33, 40; Ratcliffe 1994, 11–12). Several newly exposed features were recorded in the cliff section drawn in 1989 and settlement remains were also

This work had a direct impact on the nationally important Bronze Age remains exposed in the 1.5m high cliff face (part of prehistoric settlement and field system close to Porth Killier, National Heritage List 12

Chapter 2.1: Background to the Excavations Entry 1014998). Stone-built houses and limpet middens were already visible, and removal of the earlier sea defences revealed previously unrecorded remains – a series of pits and a burial cairn. Trimming back the cliff edge involved the partial removal of the cairn, and the foundation trench for the new sea wall cut through what was left of the interior of the main building. This loss was offset, however, by the fact that (except for a small part of the site that lies to the north-east of the rock outcrop) the new sea wall was constructed along the full length of the archaeological exposure, affording it protection from further coastal erosion. During the building of the wall, plastic sheeting was draped over the cliff section to prevent the wet concrete adhering to the archaeological remains. This measure should ensure that the latter remain undamaged if the wall is dismantled in the future. Anchoring the Armorflex concrete block revetment into the dune involved machine digging to a depth of 1–2m, but no archaeological remains were revealed by this process.

interesting from both a biological and archaeological point of view – the Scilly shrew and the toad. The latter is not present in Scilly today and, therefore, its discovery in a Bronze Age context at Porth Killier was very interesting. The shrew does populate the present islands and it was possible that the Porth Killier bones are later intrusions, but evidence from other early sites suggests this animal may have been present during prehistoric times. Retrieval and radiocarbon dating of the bones of both animals would greatly enhance understanding of their history in the Islands. Given the points mentioned above, the main objectives of the 1996 recording work were as follows: • Make a detailed record of the structures, features and layers exposed in the cliff face in order to gain a fuller understanding of the character of the settlement. • Date the exposed layers, particularly those relating to the use of the building(s). • Retrieve and date all pottery and other artefacts. • Obtain bulk samples of midden and other organic deposits in order to gain more information on the diet and economy of the settlement’s inhabitants and the surrounding environment. • Retrieve and analyse those types of environmental material not previously studied from this site (for example, marine molluscs, land snails, and pollen). • Retrieve and date the bones of the Scilly shrew and the toad to enhance understanding of these species in the Scilly.

At Porth Coose, machine excavations associated with anchoring the Armorflex into the dune revealed nothing of archaeological significance. Manoeuvring of heavy vehicles across the surface of the beach, however, disturbed a previously unrecorded intertidal peat exposure at approximately SV 87725 08610. The stone remains of a field wall and possible roundhouse were also observed nearby. Aims and objectives of the 1996 recording The overall aim of the 1996 recording work was to gain as much information as possible about the date, character and function of the site prior to the new sea wall being constructed in front of it.

Methodology The total length of cliff examined was 80m. Following removal of earlier sea defences, archaeology could be seen to survive in the eastern 60m of cliff face. Between 60m and 80m a series of machine-excavated test pits revealed only marine sand overlain by shingle.

During previous fieldwork (1985–1993) understanding of the exposed remains was hindered by the limited nature of the recording and the fact that the western part of the cliff face was obscured by the 1930s sea wall and the boulders dumped in 1989. In terms of the dating of the site, no radiocarbon results relating to the actual use of the main roundhouse had been obtained. In addition, the provenance of a handful of Iron Age/RomanoBritish and early medieval artefacts was unknown. The site had been shown to be of high palaeoenvironmental potential and much useful information had already been obtained about the diet and subsistence economy of the settlement’s inhabitants and the nature of the surrounding environment. There were, however, certain types of environmental material that had not yet been studied from the site – marine molluscs, land snails and pollen.

For the 60m of cliff in which archaeological remains survived the cliff edge was planned to scale and a detailed drawing was made of the cliff section. At two locations archaeological remains protruded beyond the general curve of the cliff and excavation by hand was required to accommodate the line of the new sea wall. This occurred between 20m and 30m from the eastern end of the section, where part of the interior of the main roundhouse survived on a shelf of ram, and at the western end of the site, where the remains of a burial cairn had previously been protected by the 1930s sea wall. Across the site, all identified layers and features were assigned context numbers and described on site context forms, and a full photographic record was made.

Amongst the animal bones already identified from the site were two species which were particularly 13

Later Prehistoric Settlement in Cornwall and the Isles of Scilly Artefacts were retrieved by two methods – by hand during the recording of the cliff section (with some artefacts being allocated small find numbers and their location plotted on the section drawings and plans), and from the bulk soil samples, during and after sieving (see below).

surface, context (40), in Zone A; a pollen sample from the pre-settlement old land surface, context (141), at the west end of Zone B; and a kubiena sample from the bottom of one of the pits, pit [50] in Zone B. All bulk samples were processed by wet sieving, the floats being collected on a 250 micron mesh and the residues on a 500 micron mesh. All the residues were completely sorted, and artefacts, animal bone, shell, charcoal and other material extracted from them.

The following environmental samples were taken: bulk samples of most of the primary layers and fills; a kubiena sample from the pre-settlement old land

14

Chapter 2.2: The Excavation Results This section provides a summary description of the features recorded at Porth Killier. Throughout this report, the context numbers for cut features such as ditches and pits are shown in square brackets – for instance [37] – and those for deposits, layers and fills in parentheses (47). Structures and groups of features are presented as unbracketed numbers.

Zone Location

For the purpose of describing the archaeological remains recorded in 1996 it is convenient to divide the cliff section into four zones (Table 2.1). Working from east to west along the cliff face, these can be summarized as follows: Zone A – buildings, middens and wall

Archaeology

A

0–29m, extending west from the Buildings, natural rock outcrop. Recorded in middens and 1989, apart from 28–29m obscured by wall the 1989 boulder sea defence

B

29–48m, previously obscured by 1930s Series of pits sea wall

C

Cairn or 48–60m, previously obscured by 1930s entrance sea wall and 1989 boulder sea defence grave, wall

D

60–80m, first 4m obscured by 1989 boulder sea defence

Marine sand – no archaeology

Table 2.1: Zones – key features and layers.

Zone A (Figures 2.4, 2.5 and 2.6) constituted the most easterly part of the cliff face recorded in 1996 (Figure 2.4). It was also that part of the site documented by Gray in 1936, monitored from 1985 by ICS and CAU and recorded and sampled in 1989 by CAU. Its eastern extent was defined by a natural granite promontory, to which the new sea wall is now anchored. Within this

stretch of cliff face were the remains of a settlement comprised of several distinct components relating to several phases of development. Table 2.2 and Figure 2.5 summarize the key structures and layers recorded and their stratigraphic relationship to each other.

Figure 2.4 Plan of the cliff edge showing the location of Zones A, B, C and D.

15

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 2.5 Diagrammatic section drawing of Zone A, showing the main features.

to trampling or abrasion). Fourteen sherds of pottery were retrieved from this layer in 1996; other artefacts retrieved in 1996 were parts of two stone mullers and a few residual pieces of flint (a core and two flakes). Animal remains identified in this layer in 1989 included pig, ox, grey seal, and deer, as well as a range of birds (cormorant, gannet, cf Godwit, cf Heron, puffin, Razorbill, redwing, cf Spotted flycatcher, and wren) and fish (bass, dab, common eel, conger eel, gurnard, mullet, plaice, pollack, saithe, whiting, and wrasse). Additional species recorded in 1996 were guillemot (bird), and gadid and sea bream (fish). Barley grains were also found in 1989, together with the remains of various arable weeds, grasses, and heathland and coastal plants.

Pre-settlement old lands surface (OLS) At various places within Zone A the natural ram was overlain by the remains of the original, pre-settlement old land surface (OLS), contexts (40) and (83). Where sealed below the walls of buildings this had remained completely undisturbed by the subsequent occupation of the site and these layers yielded no artefacts; and the same was true of contexts (5), (6), and (7) in the eastern half of the Zone which were probably also undisturbed remnants of the same old land surface. Pit [37] and limpet middens (9) and (38) Dug into the undisturbed OLS (halfway along Zone A) was a wide shallow cut or pit [37], 4.6m long by 0.5m deep, containing a rich limpet midden. The latter appeared in section as two limpet-rich layers (9) and (13) separated by a limpet-free lens (38), and presumably reflects two episodes of rubbish dumping. The time lapse between the two episodes is unknown, but the lower, more extensive layer (9), which fills the pit, was deposited before the construction of wall [39] and building [20], while the upper part of the midden, context (13), filled the gap between these two structures, though it is uncertain whether the rubbish was dumped before or after their construction. Layers (8), (10), (11), (12) and (14) are likely to have been naturally occurring soil build up which, being on the edge of the settlement, had become contaminated by occupation debris. However, layers (10), (11) and (14) may have been an eastern extension of (9) – the interface between these layers was obscured by overhanging vegetation.

Disturbed old land surface Towards the eastern end of Zone A the undisturbed OLS was overlain by context (4), a layer 7.5m long by up to 0.5m deep which yielded 31 sherds of pottery, part of a saddle quern (reused as a rubbing and hammer stone), and half of a bruising muller (reused as a small mortar). This layer appeared to be the result of the upper part of the OLS having become intermixed with settlement debris. This may have occurred as a result of ploughing after dressing with settlement rubbish. Alternatively, it may have been the result of the OLS becoming churned up during day to day occupation of the settlement, with midden material being dumped directly on this surface and spilling out from the adjacent settlement. The second scenario is probably more likely given that some of the 20 potsherds retrieved from layer (4) in 1989 had been broken by trampling. In addition to artefacts, two animal bones came from this layer in 1996, one sheep and one sea bream.

Lower midden layer (9) was radiocarbon dated in 1989; a date range of 3170 ± 65 BP, 1620–1280 cal BC (OxA3648) was obtained from its base and a range of 3220 ± 70 BP, 1670–1300 cal BC (OxA-3647) from its top. In 1989 it yielded 48 sherds, most of which had fairly fresh breaks, indicating that they had been deposited in the midden shortly after the vessels they came from had become broken (and were not subsequently subjected

In 1989 two dark humic layers, overlaying/lying within (4) towards its western end, yielded the bones of sheep, ox, and fish (grey mullet, pollack, whiting, and wrasse); 16

Chapter 2.2: The Excavation Results together with barley grains and seeds from various weeds. These layers were interpreted as limpet-free midden deposits. They had been eroded away by 1996 but appear to support the interpretation of (4) as a surface that was accumulating or in use during the life of the settlement. A combined sample from the two 1989 layers produced a radiocarbon date range of 2935 ± 55 BP, 1370–970 cal BC (OxA-4700). At its western end layer (4) overlay (11), which may have been the eastern extent of midden (9) (although overhanging vegetation obscured the join between the two). The 1989 radiocarbon dates, together with the stratigraphic evidence suggested that layer/surface (4) post-dated midden (9). Wall 39 and roundhouse 20 Wall 39 and roundhouse 20 were built into the top of midden (9). Though stratigraphically isolated from each other the two structures may have been contemporary owing to the fact that they occurred at the same level in the cliff section and both cut midden (9). Wall 39 was a drystone wall, one stone (0.54m) thick and four courses (0.85m) high. It appeared to be an enclosure wall running in on north to south direction on the eastern side of the settlement. Roundhouse 20 appeared to be circular or oval in plan, with its interior and most of its walling still buried beneath the cliff top. Visible in the cliff face was a section through the building’s double-faced wall (see Figure 2.5). Coastal erosion had removed part of the outer stone face, context 16, which comprised four courses of horizontally placed granite blocks 0.75m high. Between the two exposed stumps of 16 was the wall’s rubble and earth (ram) core, (18), 0.7m thick. Partial erosion of this core had revealed the back of the inner face of the wall 17. The wall core overlay undisturbed OLS (40), but the latter had been removed by the construction of the outer wall face; and on the east it had already been removed by the digging of pit [37]. Roundhouse 21 On the western side of roundhouse 20, tacked onto its outer wall face, was a second roundhouse 21. Although also circular or oval in plan, by contrast with 20, most of this structure had already been removed by coastal erosion. What remained was an arc of walling 0.8m high surviving in the cliff face, plus a small part of the building’s interior, lying on top of the shelf of ram that projected out from the base of the cliff at this point.

Figure 2.6 Pre-, during and post-excavation plans of roundhouse 21.

were horizontally set in three courses, with a saddle quern reused as one of the stones of the lowest course. On the east the basal course consisted of upright stones (with a couple of vertically set stones surviving above). It is at this point that wall face 22 could be clearly seen to butt onto the outer wall face, 16, of roundhouse

The wall was of similar construction to that of roundhouse 20 – an unmortared double-faced stone wall with and an earth and rubble core. The internal face of the wall, context 22, was 3.6m in length by 0.75m high, comprised of blocks of granite. On the west side these 17

Later Prehistoric Settlement in Cornwall and the Isles of Scilly 20. The coursed walling of the latter contrasted with the vertically set stones of 22 and followed a slightly different direction.

Layer (28) was in turn overlain by a deliberately laid area of flooring (27), consisting of yellow clay, redeposited ram 60mm thick, with cobbles set in it (Figure 2.6a). At a number of locations there were patches of dark loam (68), which were interpreted as being contemporary with floor (27). Also contemporary was a stone-lined and capped drain 32, 0.2m wide by 0.26m deep, the top of which was flush with floor (28), the capstones 33 being bonded together with similar clay. The construction trench for drain 32 had been dug down through the underlying occupation layer (28) and into the surface of the earlier ram floor (29). The drain extended from the eastern side of the building’s blocked doorway for a distance of 1.75m into the building’s interior, at which point it ended (although the stone lining was not intact as far as this point, the cut for the drain survived as a rounded off end). At its other end the drain could be clearly seen to continue under the blocking of the entrance. It is unclear whether its purpose was to direct water out through this entrance, to prevent waterlogging inside the building, or to bring in run-off water in for use by the building’s inhabitants.

In the centre of 22 was a 0.5m blocked doorway, 0.5m wide, flanked on either side by an orthostatic a jamb stone. The blocking consisted of two large blocks placed on top of each other, with smaller stones wedged into the gaps between these and the door jambs. Owing to its relationship with roundhouse 20, it was only on the western side of roundhouse 21 that the core and outer face of its walling was exposed in the cliff face. The core consisted of three spoil layers (80), (81) and (82) containing small stones and lenses of redeposited ram. On the west side of this (at the western extremity of Zone A) the outer face of the wall, exposed as three granite blocks placed one on top of the other, context 45. The area within roundhouse 21 measured 4.6m in diameter but it was impossible to be certain of the original dimensions of its interior (see Figure 2.6). The surviving interior consisted of a gently sloping shelf of ram, the northern, inland, half of which was cut into by features and partially overlain by layers associated with two or three phases of occupation and the postoccupation infilling of the building. On the southern, seaward side the sea had removed all evidence of human activity and eroded away the surface of the underlying ram.

The upper layers in roundhouse 21 related to its abandonment and subsequent use as a repository for domestic rubbish generated elsewhere in the settlement. Above floor (27) was a layer of midden material in a matrix of dark grey loamy sand (25) 0.14m deep, overlain by wall stones (25) that had either tumbled or been deliberately dislodged, into the interior. Overlying this was (23), a rich midden deposit, thickest (0.6m) on the eastern side of the interior where it clung to the inner wall face. Layers (25) and (23) probably represent the same general episode of dumping, comparable to midden layers recorded in 1989 (Ratcliffe and Parkes 1990).

The first phase of activity in 21 was represented by a series of features cut into the natural ram (29) which formed the floor of the building during this phase (Figure 2.6c). The floor was not level but sloped gently down from the back wall of the building. The features cut into its surface consisted of a series of postholes [43], [62], [70], [74], a stakehole [66] and a gully [41], together with two shallow depressions which were interpreted as worn areas of flooring or naturally occurring undulations in the ram [76] and [78]. The exact function of the postholes is uncertain, although the larger ones, such as [43] and [70] may have held roof supports and [74] is likely to have housed a doorpost, judging from its position on the western edge of the building’s entrance, where a post-stone was located during the second phase of occupation (see below). The gully may have been a drain designed to direct surface water away from the building – it had been truncated by coastal erosion but was clearly heading downhill towards the sea.

Soil containing settlement debris Overlying midden (9) in the eastern half of Zone A was layer (12). This consisted of soil containing significant quantities of settlement debris – 54 sherds of pottery, including sherds from the same vessel as some of those from midden (9), together with large stones, pockets of midden material, and blobs of clay, possibly used as mortar. This was interpreted as a natural build-up of soil, which accumulated during the life of the settlement or shortly after it had gone out of use and owing to its proximity to the eastern side of the settlement became contaminated with debris from the latter. Post-settlement layers

Overlying the floor (29) was a dark gritty loam (28)/ (59)/(61), some 80mm thick and containing patches of yellow clay (possibly redeposited ram) (Figure 2.6b). This layer was interpreted as an accumulation of occupation debris, with the clay possibly representing attempts to level out hollows that had developed.

Overlying the settlement remains were windblown sand and soils that had formed after it had been abandoned: contexts (1), (2), (3), (19) and (162).

18

Chapter 2.2: The Excavation Results

Figure 2.7 Diagrammatic section drawing of Zone B, showing the main features.

extent and, in places, overhangs had been scoured out by the sea (hence the dumping of the boulders). As a consequence, a considerable amount of trimming and cleaning was necessary before the section was in a recordable condition. Fragments of in situ archaeology were visible in plan, but these were not considered extensive enough for a plan to be produced.

Zone B – pits Zone B (Figures 2.4, 2.7 and 2.8) comprised an approximate 19m length of cliff face running west from wall 45 as far as the 48m mark (Figures 2.4 and 2.7). The natural soil profile throughout this Zone was best characterized around the 37–8m mark where the OLS was present as context (85), 0.28m deep and overlain by a dark sandy loam (84), 0.42m deep, which is in turn overlain by the sandy turf layer (1). Further to the west (for example, at the 45m mark) a pale sandy layer (112) appeared above (84) and below (1). The general assumption was that (85) equated to layers (5), (6), (7), (40) and (83) in Zone A, that (84) is equivalent to (3) and (112) is perhaps similar to (2). The important points are that layer (112) is recent in origin and that all of the archaeological features in Zone B are sealed by (84) – no features are cut into (84). All these naturally occurring soils form more or less horizontal layers.

Area B1 Figure 2.8 Detailed section drawing of pit [50] complex HERE. This was the most complicated area in Zone B (Figure 2.8). The earliest feature was a narrow gully [58], 220mm wide by 0.24m deep, visible in plan and running roughly north-west to south-east. It is possible that this feature was contemporary with wall 45 (which formed the western end of Zone C). If wall 45 represented the outer face of the wall of roundhouse 21 (in Zone A) then [58] might have been the drainage gully for it, in section it is clearly sloping down seawards. However, this interpretation assumes a chronology which was not demonstrable in the visible stratigraphy.

The archaeological features in Zone B took the form of a series of pits and comprised four discrete, multi-phase areas of activity: • • • •

The homogenous drain fill (56) was gritty and truncated by a large pit [50], 0.95m deep with a 2.5m length visible in section. In the west it steeply cut the ancient soil (85) and in the east appeared to use the wall 45 as its eastern edge. There was a slight cut through the ram immediately to the west of wall 45, suggesting that [50] post-dates 45.

Area B1) the pit [50] complex. Area B2) the pit [86] complex. Area B3) pits [97] and [101]. Area B4) pits [105] and [107].

All of these features cut the OLS [85] but it was not possible to establish any stratigraphic relationship between the different areas of activity.

A plausible interpretation is that 45 was the wall of roundhouse 21, gully [58] was drainage gully for the roundhouse and the use of pit [50] post-dated the use of the building. The floor of pit [50] in the west was lined with a thin layer of yellow clay [159], in the centre of the pit two later features were cut through this clay floor and into the ram.

Areas B2 – B4 were all concealed behind the 1927 sea wall which came away cleanly. Area B1, however, was protected by a pile of large boulders and quantities of modern material had been thrown against the cliff face. The boulders had also damaged the section to some 19

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 2.8 Detailed section drawing of pit [50] complex.

The earliest of these was pit [57], which also showed in plan. This was a shallow, ovoid pit, 0.72m wide by 0.2m deep, with two interesting fills. The lower fill (49) consisted of gravel with numerous shell fragments, quite different from the usual friable limpet midden material found in Zone A. This formed a horizontal layer sealed by (48) a deposit of yellow-reddish brown clay with extensive iron panning. This layer was in turn cut in the east by feature [161] comprising a shallow, rectangular cut [55], 0.46m wide by 0.1m deep, lined in the north and east by two flat stone slabs, suggesting that the whole feature was originally a stone box. It was filled with black humic soil (52). Its function is unknown, the stones showed no signs of burning hence it is unlikely to have been an oven or boiling pit and it is likely that (51) and (52) may date to the postabandonment of [161].

Area B2 The main feature here was a shallow pit [86], 1.56m wide by 0.14m deep with a gently sloping western side. The base of the cut consisted of a series of undulations, perhaps suggesting spade cutmarks. At either end of this pit were postholes [88] and [91], 0.24m wide by 0.28m deep, which were presumably contemporary with the pit and possibly for posts supporting a shelter over it. Pit [86] appears to have undergone two phases of infill; the primary fill [87] having the appearance of gradual silting and accumulation, but upper fill [89] being a deliberate backfilling using boulders. Immediately to the west was a second pit [93]/[95], 0.88m wide by 0.4m deep, which was initially considered to be two features. The primary fill [94] was probably silting but the upper fill [96] was a homogeneous dark humic soil containing four medium sized stones towards the top. The presence of these stones could be interpreted in a number of ways; was this feature a large post-pit with packing stones? Were they in situ stones over the pit fill which have subsided into the pit? Do they indicate deliberate backfilling of the pit? No stratigraphic relationship between pits [93]/[95] and [86] was demonstrable in the section.

Stones had slumped or been thrown against the western side of pit [50] prior to the accumulation of primary fill (51), a second phase of infill involved the deliberate dumping of large boulders contained in fills (47) and (48); the whole was sealed by (84). In general terms a common theme of the stratigraphy in Zone B was an episode, or episodes, of depositing stones and boulders into abandoned pits, particularly into their upper levels. This was suggestive of levelling off the area after the abandonment of the pits and prior to the build-up of soil layer (84). Pit [107] did not fall into this pattern but elsewhere there appeared to have occurred a conscious abandonment or change of use in Zone B. Certainly in the case of pit [50] the infill boulders appeared to have come from nearby roundhouse 21.

Area B3 This area also contained two pits. Pit [97] was 1.2m wide by 0.56m deep with gently sloping sides, containing three horizontal layers of fill (98), (999) and (100) and had possibly been recut. The infilled pit was subsequently cut on its western side by the deeper 20

Chapter 2.2: The Excavation Results

Figure 2.9 Diagrammatic section drawing of Zone C, showing the main features before cutting back.

Figure 2.10 Diagrammatic section drawing of Zone C, showing the main features after cutting back.

pit [101], 1.2m wide by 0.62m deep with a rectangular profile. Pit [101] also contained three distinct layers of deliberate backfill; a layer of stones (104) overlain by a clean sandy loam (102) and topped by a second stony fill (103).

approximately the 60m point (Figure 2.9) where the extremity of Zone C was defined by a substantial pile of large granite boulders dumped as sea defence by the Hicks family and Royal Marines. Some of these boulders originated from the upper courses of a post-medieval/ modern drystone wall running along the cliff top from the west and terminating more or less at the 56m mark along the cliff section. Tumbled boulders which had fallen randomly from this drystone wall were found in the upper levels of the cliff section at this point.

Area B4 This was perhaps the simplest area; a straight-sided pit [105], 0.78m wide by 0.52m deep, had been deliberately backfilled with stones (106) and the backfilled pit was later cut by a steep-side, flat-bottomed pit [107], 2.01m wide by 0.4m deep. A deposit of medium-sized boulders in a matrix of mottled light yellowish brown sandy clay (108) formed the west fill (deliberate infill) but the upper fill (109) was dark greyish brown sandy loam, resembling layer (84) and suggesting gradual accumulation of soil rather than deliberate infill.

Zone C was subject to three phases of recording: 1. 2.

Zone C – wall, cairn Zone C comprised a 12m length of cliff-face section running westwards from the 48m mark to

3. 21

A cursory record was made during the recording work on Zones A and B. A protruding length of cliff-face – from the 54m mark to the 60m mark – had to be trimmed back to allow practical siting of the sea defence wall. This was excavated by hand and recorded in plan and in section. At the same time the whole of Zone C was recorded in more detail. Too enable the most effective positioning of the sea defence wall; the whole of the cliff-face in

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 2.11 Pre- and post-excavation plans of Zone C.

Zone C had to be trimmed back by up to 1m. The first metre was cut back by machine, but the rest was excavated by hand and was recorded in plan and in section.

Wall 130 The earliest significant archaeological feature was wall 130, 1.2m wide by 0.4m high and visible for a length of 2m. Initially this was considered to be a field wall, however, excavation showed that it to be a doubleskinned structure, 153, with rubble core (158) – similar in construction to the walls of buildings 20 and 21 to the east: the shallow arc described in plan by wall 130 and the fact that it cuts into original ground surface also suggest that it is a surviving length of roundhouse wall. There were, however, three problems with this interpretation: 1) the rubble core (158) consisted entirely of mainly small stones in a friable sandy loam matrix with no trace of the mixed layers (with lenses of redeposited ram) characteristic of the walls to roundhouses 20 and 21; 2) whereas the interior of roundhouse 21 had been scraped down to the ram (ram formed the floor during the initial occupation phase), in the area bounded by wall 130 the original ground surface – though much mutilated – is intact to a depth of up to 0.3m; and 3) it was difficult to identify at what point in the eastern section wall 130 returns. One possibility was cut [117], but this cut was not visible once the section had been trimmed back (which was carried out by machine at this point).

The initial recording phase identified a cut [111] and boulder layer (110) [not recorded in section] and it was speculated that this might comprise a stone-filled pit similar to those recorded in Zone B. The later recording work, however, recognized these features as being part of a substantial stone-and-earth structure 154 which has been interpreted as a cairn (see below). Old land surface The OLS in Zone C, contexts (116) and (132), was broadly similar to that in the cliff section to the east and extant throughout virtually the whole 12m length of section. It was degraded and truncated by archaeological features. In places where no archaeological features occur the original ground surface was naturally overlain by soil build-up layers (84) and (112) which elsewhere seal the archaeological levels. At the 57–58m mark there occurred a narrow, curving gully [157] up to 0.3m deep (Figure 2.11). This is a somewhat peripheral feature and there are no clues as to its date, function or association with other features, except that it underlies (134) (not illustrated).

Despite these reservations the likeliest interpretation of wall 130 is that it was a roundhouse wall. Samples 22

Chapter 2.2: The Excavation Results were taken from the level at the bottom of wall 130 after removal of the rubble core (158) and also from layer (139), a thin patchy dark loam soil evident in places beneath the outer granite face.

post-dated the cairn and is interpreted as agricultural soil which has been pushed up against the western face of the cairn by successive cultivations. This soil was severely truncated in the west by a large deposit of marine sand (135?) which was probably modern in origin (though possibly post-medieval) and contains an animal burial (bovine?) [140].

Cairn 154 The dominant feature in Zone C was structure 154 which has been interpreted as a cairn (Figure 2.10). This structure was visible in section for roughly 9m, running westwards from 48m to 57m, and survived to a height of 0.5m. During recording stage 2 the cairn appeared to slope downwards towards the east, but after trimmingback 1m its profile was much flatter. The cairn was constructed of numerous granite stones and boulders of varying sizes and apparently piled randomly in a mixed sandy loam matrix. At its western end the cairn incorporated wall 130 into its construction. At more or less the central point along the visible length of the cairn was a probable cist or chamber [163], 0.8m wide by 0.44m high. This was of crude construction, consisting of three orthostats, a saddle quern stood on its side and large irregular-shaped boulder placed so as to form a roughly sub-rectangular chamber. There was no capstone and the chamber/cist appeared to continue back into the section.

The appearance of cairn 154 as a whole with very large boulders dumped, apparently at random, and with inclusions of yellow clay in the soil matrix, characteristic of floor level (127) in roundhouse 21 strongly suggests that it might be a roundhouse that has been demolished and converted into a cisted/chambered cairn, although the inconsistencies pointed out earlier should be given consideration. Zone D West of the 60m mark the whole nature of the cliff-face at Porth Killier changes dramatically (Figure 2.4). The spread of boulders dumped by the Hicks family obscure 4m of the section and it was not deemed feasible to remove them without undermining the cliff at this point. From the 64m mark (where the section was accessible again) onwards the soil profile characteristic of the preceding 60m disappears and is replaced by horizontal deposits of marine sand topped by shingle and there were no archaeological deposits.

Generally, the soil matrix of cairn 154 was homogenous brown/dark brown sandy loam with yellow clay inclusions and local differences. The fills of the chamber which were excavated in 50mm-100mm spits were consistent with this description (114), (150), (155), (156) and (157). The chamber contained large concreted lumps of yellow clay and numerous sherds of pottery from several different vessels. These sherds occurred over a depth of 0.3m and appeared to be complete vessels which had been broken in situ, layered one on top of another and mingled with the clay blobs. An upturned mortar or socket stone was also found in the chamber. No trace of cremation or inhumation was found but all the soil excavated from the chamber was retained for sampling.

Submerged peat and stone remains in Porth Coose The peat was exposed at more or less mid-tide in May 1996, at approximately SV 87725 08610. The exposure was some 30m by 3m in extent. A block 0.3m wide by 0.18m in depth was cut out, wrapped in plastic and placed in a plastic bread crate for transport. The top of the peat was levelled at +1.09m OD (as compared with the ground surface on the north-west side of the Big Pool which was +2.36m OD). Peat stratigraphy (laboratory description):

Beyond either end of cairn 154 there occurred soils which resemble the original ground surface but contain more stones than normal, for example (38) or are of a more mixed, gritty nature than normal, such as (115). These are best interpreted as original ground surface which has become contaminated by tumble and debris from the cairn.

0–0.14m undifferentiated sandy peat (lower boundary diffuse) circa 0.14–0.18m humic silty sand (upper boundary diffuse) 0.18m and below granite Samples for pollen analysis only were prepared from: 0–0.01m; 0.01–0.02m; 0.10–0.11m; and 0.17–0.18m. Full analysis was not done as part of this project.

To the west of cairn 154 occurred an extensive deposit of ploughsoil (132)/(134), 0.3m deep. Although a more precise date for this soil is not yet suggested, it clearly

23

Chapter 2.3: The Artefacts

associated with Wk-19902, 1420–1130 cal BC (Quinnell 2009–10). The other is Little Bay, St Martin’s (Neal 1983), where the later structural levels have associations with plain, sometimes biconical, wares and one radiocarbon date calibrated to 2780 ± 80 BP, 1190–800 cal BC (HAR1726) (ibid, 49). Much of the large Nornour assemblage is now recognized as comparable and of broad Middle Bronze Age date (Robinson 2007, 61–63).

Prehistoric pottery Henrietta Quinnell Some pottery eroding from the site was identified by Alec Gray in the 1930s (Gray 1972, 43 and fig 14b). Two groups of material eroding more recently from the site have also been studied. That from the Isles of Scilly Electrification Project (ISEP) was published (Quinnell 1991) but the second, from the Coastal Erosion Project (CEP), has remained as an archive document (Quinnell 1994). Relevant sections of the latter with illustrations where appropriate are included here. The range of material now extends from possibly Neolithic until into the first millennium BC (Table 2.2).

The assemblage from the nearby settlement at Higher Town, St Agnes, is mainly composed Late Bronze Age Plain Ware types with dating from the thirteenth to the ninth centuries BC but includes a few carinated vessels which may be eighth century or later (Quinnell, forthcoming a). The vessel types in Plain Ware are dominated by a range of necked jars which are not individually very distinctive, and it is possible that the presence of this material has not been recognized on some sites in the past.

The majority of 1996 finds were given small finds numbers, prefixed SF except those from the flotation of bulk soil samples, for which the soil sample SS numbers are used. Most sherds without SF numbers have been marked sequentially by context for example, 46/11, 46/12. A list of these is in the archive.

Vessels from the principal phases described above are all present at Porth Killier.

The report was first prepared in 1998 but is updated to include more recent significant work, especially that at Dolphin Town, Tresco (Quinnell 2009–10), Higher Town, St Agnes (Quinnell, forthcoming a) and at Old Quay, St Martins (Quinnell 2017a). The position of the site on an eroding coastline and the partial nature of the excavation restricted the search for conjoins and made the estimation of minimum vessel numbers very tentative.

Zone

Sherds Weight

Zone A 212

2366g

Zone B 119

1930g

Zone C 670

Totals 1001

A comment on the chronology of Bronze Age pottery on Scilly

5977g

Minimum vessel numbers and illustrations P1 – P4a, 6+ other vessels, minimum total 15+

P9 – P12, 6+ other vessels, minimum total 14+

P13 – P20, 7+, minimum total 15+

10273g Minimum vessels 44+

Table 2.2: Summary pottery quantifications by Zone (P5–P8 allocated to illustrated vessels from the Coastal Erosion Project).

The biconical, generally lugged, vessels with horizontally arranged decoration around their upper parts are now established as Early Bronze Age, following the publication of a series of radiocarbon dates from Knackyboy Cairn (Sawyer 2015, fig 8.2). Undecorated forms occur at this period. It is uncertain how far into the Middle Bronze Age decorated and undecorated lugged vessels continue to be made. There are two assemblages with single dates that suggest that by perhaps the fourteenth century BC simple undecorated biconical and slightly necked forms were current. One is at Dolphin Town, Tresco, where a spread of material was

Fabrics Pottery is described in the broad fabric groups previously used by the author (Quinnell 1991; 1994): granitic; coarse granitic with some inclusions up to 5mm or more across and 40–50 per cent inclusions; gabbroic; fine matrix granitic now recognized as Neolithic. Sherds are granitic unless otherwise stated; little coarse granitic and one gabbroic sherd were identified. Granitic pottery is generally thick (7mm-12mm being typical), with moderate (15 per cent) to very common (30 per cent) 24

Chapter 2.3: The Artefacts amounts of grit; most grit is 1mm-3mm but larger pieces occur. Reduction and oxidization varies even on the same vessel but there is a tendency for exteriors to be reduced. Hardness can also vary on the same vessel. The assemblage, as a whole verges towards hardness, most sherds can only just be scratched with a fingernail. Hardness, or its converse, is only commented upon in exceptional cases. Finger modelling marks are common. Any special characteristics are commented in the descriptions below. As far as possible the terminologies and procedures recommended by the Prehistoric Ceramics Research Group (1992) have been followed.

More recent work, at Dolphin Town, Tresco and at Higher Town, St Agnes has included petrographical analysis of granitic fabrics by Roger Taylor (in Quinnell 2009–10; forthcoming a). These showed some differences in that at the former granitic fabrics regularly contained crushed granite and the latter contained beach sand, differences which occurred over the long use of both sites. This suggests that methods of ceramic production may have differed slightly across the islands; Porth Killier might be expected to have fabrics similar to Higher Town but this needs confirmation through future work.

Context

Sherds

Weight

(4) OLS beneath roundhouses

31

256g

junction of (4) with (3)

3

37g

(3) soil over (4), East end of Zone

7

46g

(9), (13) midden preceding roundhouse 20

15

278g

(18) wall core of roundhouse 20

1

2g

(10) soil accumulation East of roundhouse 20

Details Includes small, flat-topped, everted rim and base angle

P1, P2, P3 and sherds from other vessels

8g

(12) soil accumulation East of roundhouse 20

54

28g

Includes sherd from P2

(8) soil accumulation East of roundhouse 20

1

2g

(80), (81), 82) wall core of roundhouse 21

4

32g

(78) depression in (290 basal floor roundhouse 21

1

40g

Eroded incised decoration

(63) fill of posthole [62] in (29)

5

55g

Two of same small flat-topped rim

(42) fill of gully [41] in (29)

1

64g

[70] post pit ? in [29]

7

96g

(28) build up over floor (29) in roundhouse 21

40

893g

One coarse granitic sherd with ? rusticated finish; four basal sherds from three vessels; many sherds are of size and freshness to indicate deliberate dumping

(27) material on/in upper floor in roundhouse 21

21

162g

P4, P4a and sherds of two other vessels of distinctive fabric

(68) part of floor (27)

2

5g

(31) fill of pit [30] in roundhouse 21

2

27g

(25) lower part of midden over roundhouse 21

7

99g

(23) upper part of midden over roundhouse 21

2

12g

(1) topsoil

7

224g

212

2366g

Totals

Includes small rounded rim and decorated sherd from vessel similar to P2 All eroded, including small rim

Table 2.3: Summary of pottery from Zone A; contexts arranged so that earliest are at top, gradually becoming later down the table.

25

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

1996 context

CEP context

Details of pottery

(83)

(19) OLS

one sherd

(4)

(6) midden in OLS

(9)

(14) midden

(82) wall core roundhouse 21 (18) No direct correlation, possibly part of (82)

(30)

(12) soil accumulation East of roundhouse 20

(12)

(8) soil accumulation East of (5) roundhouse 20

20 sherds from at least three separate vessels, some broken by trampling, flat-topped rim

48 sherds, most fairly fresh; sherds from base angle and shoulder of biconical jar with shallow finger impressions. P7 and sherds from P1, and P2 Seven sherds probably all from P6 and P7

P2, two sherds

Two sherds, abraded 11 sherds, mostly large and fresh Eight thick sherds ? from same vessel; + four others

No correlation

(4) dump on (5)

No correlation

(9) dump on (5)

Six sherds

(37)

Four body sherds, fresh breaks

(25) midden in roundhouse 21

(25)

Eleven sherds, mostly large and fresh, including part of a base P5

(3) (2)

(3) over (4), (5)

(1) topsoil

[2] soil build up

35 sherds

(1), (10), (17)

45 sherds

(27) on/in upper floor roundhouse 21

No equivalent material No equivalent material

No equivalent material

No equivalent material

one sherd

Eroded from 37 sherds, mostly small and eroded, could be second millennium BC. One stratified sequence grass-marked, early medieval period

About 56 sherds, probably eroded from (14), two rim sherds from ? P7, one Found unconnected sherd has a slight square lug with coastal 34 sherds from main part of site. Five join and form part of the base and monitoring lower wall of a vessel with heavy thumbing on the junction From ? structural remains to the North East of the site

17 sherds, second millennium BC type from possible structure NE of the recorded section (Ratcliffe 1994, 11)

Between structures and Kallimay Point 20 sherds, mostly small, all could be second millennium BC, including P8 over 300m

Table 2.4: Material from the Coastal Erosion Project, arranged in sequence of 1996 stratigraphy. For details see Table 2.5. For contexts with no 1996 correlation, see Quinnell 1994; their approximate stratigraphic positions can be deduced from position in the table.

Zone A Descriptions of enumerated vessels (Figure 2.12 and Tables 2.3 and 2.4)

unique with its two bands of slight and deeply scored lines.

P1 jar with rounded rim and slight internal bevel, hard; narrow zone of slight incisions below rim running horizontally around vessel but sufficiently irregularly to cross each other; below on girth, separated by plain zone, deep grooves roughly parallel and running horizontal around vessel, made while clay still very damp so that ridges between grooves have been flattened. Same vessel as APK 89 1/2 from then context (14). Grooves are a very common motif on second millennium biconical vessels but so far P1 appears

P2 three well-finished, hard, body sherds with irregular criss-cross decoration lightly incised during surface finishing, similar in technique to upper decorative zone in P1. Same vessel as APK 89 23/3, 23/5, 23/15 from then context (14) and almost certainly as APK 90 1/1, 1/2 from then context (30). The form of the vessel is very common in Scillonian second millennium BC assemblages, but lightly incised decoration is rare, occurring very occasionally for example, no 119, figure 33 from Nornour (Butcher 1978). 26

Chapter 2.3: The Artefacts P3 everted rim with pronounced internal bevel; on fabric this could just be the rim of P2.

such impressions are common in second millennium BC assemblages in Scilly, classically illustrated in plA XIV of the base of Urn X from Knackyboy Cairn (O’Neil 1952).

P4 simple rim with rounded top and slight interior and external bevels, oxidized fabric, rough finish, about six sherds. Selected to illustrate present of variety of simple rims in assemblage.

P6 APK.90.2/1 Flat-topped rim sherd from hard, straight-sided jar. Smooth finish, with mica showing on surface, on which there is a pattern of very fine incised lines.

P4a (not illus) 10 sherds, coarse granitic, circa 20mm thick, soft, oxidized; patches of roughly finished reduced surface survive on one side only of a few sherds; several perforations where organic material has been burnt out. The pattern of the faces on the sherds where surfaces do not survive suggest that quantities of organic material had been present, and that perhaps the vessel originally had been structured over a kind of basketry frame. If so, this is very unusual and has not been noted elsewhere in Scilly. The ‘rusticated sherds’ from context (28) below this (Table 2.3) may come from a similar vessel. The general nature of the material suggests that it is likely have been deposited in material over the floor after use; the fragile nature of the coarse granitic container would not have withstood any trampling (unless the pieces come from some inset feature such as a lined pit set within the floor).

P7 APK.89.20/1 Rim and upper part of biconical vessel, rim slightly everted with internal bevel, undecorated but exterior roughly smoothed; very reduced black. Irregular with marks of finger moulding clearly visible. Common second millennium BC type cf no 70 from Period 3, eastern side, Nornour (Butcher 1978, fig 29). P8 APK. 93.25/11 body sherd with a line of fingernail incisions, bordered on both sides by a double row of cord impressions, the direction of the cord twist in opposing directions in each row. A comparable sherd no 136, figure 34 comes from Nornour (Butcher 1978); twisted cord impressions are fairly frequent on Scilly but those of fingernails are not. The date and affinities of P8 are uncertain, the more so as it comes from outside the Zone A ‘hut circle’ or roundhouse settlement, from the area towards Kallimay Point for which possible

P5 (not illus) APK 89.25/1 Thick sherd from base at least 250mm across. Impression of matting on underside;

Figure 2.12 Porth Killier. Bronze Age pottery from Zone A, roundhouse area. Note P3 shows interior of sherd Nos P6–8 are from the Coastal Erosion Survey.

27

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Context

Sherds Weight Details

Area B1 [57] pit, earliest in pit [50] sequence

3

10g

[55] pit cut into (48), pit [50] sequence

2

18g

(490) fill of [57], pit [50] sequence

1

(52) fill of pit [55]

1

(51) primary fill of pit [50]

1

5g

As those in [57]

32g

Base angle with finger modelling, unusually hard

26g

(47) fill of pit [50] with boulders

55

753g

[50] not assigned to contexts in pit sequence

15

333g

(46) upper fill of pit [50]

12

Small body sherds with sparse fine grit < 2mm

307g

As those in [57] Body sherd

P9 and at least two other vessels on fabric differences

P10 and P11 (gabbroic) and sherds from ? other vessels P12 and sherds from other vessels

Area B2 (920 fill of posthole [91] in pit [86]

1

90g

Large body sherd

[100) fill of pit [97]

6

75g

Body sherds

(99) above (100) in pit [97]

5

185g

(1020 fill of pit [104] cutting pit [97]

8

22g

(84) sealing all prehistoric contexts

3

54g

(90) fill of posthole [91] in pit [86] Area B3

Area B general Totals

6

127

20g

1930g

Two tiny fragments from rim 7mm thick

Includes basal angle with outer side of base retaining slight impressions of coarse circular matting Small body sherds

Eroded body sherds

Table 2.5: Summary of pottery from Zone B.

settlement preceding the second millennium BC is suggested by lithic finds.

2 and probably in (25), the midden infilling the roundhouse. The finely finished and decorated rim P6 probably came from the wall of roundhouse 21 and is likely also to be redeposited, as is decorated sherd from (78) in the basal floor. Given the need to level out areas for roundhouse construction, with the consequent disturbance of earlier deposits, the gradual spread up through a stratigraphic sequence is understandable.

Discussion of Zone A pottery The distinctive vessels, P1, P2, P3 and P7, come from midden (9), CEP (14), from which radiocarbon dates are 3170 ± 65 BP, 1620–1280 cal BC (OxA-3648) from the bottom, and 3220 ± 70 BP, 1670–1300 cal BC (OxA-3647) from the top. The lack of close comparanda for both P1 and P2 points up how much we still have to learn about the detail of Scillonian forms at this period, their date range and area/mode of distribution within the islands. P7 is a simple biconical form which is likely to have been made over a long period. Midden (9) predates any detected structures on the site. Presumably the buildings related to it are either inland or have already eroded. Gray (1972, 43) comments on the fineness of the sherds he retrieved in the 1930s, comments appropriate for P2, and indeed P6, and suggest that his material may be connected either to this midden or to related buildings.

Two dates, 2533 ± 79 BP, 810–410 cal BC (Wk-5688) and 2794 ± 74 BP, 1190–800 cal BC (Wk-5689) come from (28), the build-up (probable dumping from the size of sherds) over basal floor (290) in roundhouse 21 and below upper floor (27). These dates suggest that roundhouse 21 belonged in the earlier first millennium BC. There is little pottery from contexts associated with roundhouse 21 and its infill that is distinctive. P4 is a basic form that had a long life. The mat-impressed base P5 may be either be redeposited or indicate a long use for these impressions. The roundhouse settlement in Zone A may have been contemporary with at least some of the pits in Zone B.

The distribution of sherds of P2 extends to contexts connected with the roundhouse 21: in soil east of the roundhouses, redeposited in the wall of roundhouse

A radiocarbon date from context (8) (CEP context (5)) was 2936 ± 55 BP, 1370–970 cal BC (OxA-4700). No distinctive material came from this context, which 28

Chapter 2.3: The Artefacts P10 shouldered jar, neck with fingertip impressions, rounded rim, roughly smoothed surface, oxidized but with heavy sooting on exterior of neck P11 rim with pronounced but irregular internal bevel from bowl; gabbroic, reduced exterior, roughly smoothed with sooting, oxidized interior more smoothly finished P12 flat-topped jar rim, slight out-turn forming neck upon which finger modelling visible; roughly finished; reduced interior, oxidized exterior. Discussion of Zone B material The value of the ceramics from the Zone B pits is that their fills were likely to have occurred quite rapidly. While redeposition cannot be entirely ruled out, there is no obvious evidence for intrusion. The pits come closest to single period sealed contexts of anything yet studied for the Scillonian prehistory. The finger tipped neck of P10 is paralleled by no 4 from Period 7 middens filling Building 3 at Nornour (Butcher 1978, fig 24), a Late Bronze Age Plain Ware type as now known from mainland Cornwall. There is a good mainland parallel in P3 from the structure at Callestick (Quinnell 1998– 9a) associated with NZA-7686, 1020–894 cal BC. The determination 2841 ± 68 BP, 1210–840 cal BC (Wk-5687) from pit [50] is entirely appropriate even if 2641 ± 71 BP, 980–540 cal BC (Wk-5686) is rather later and broader.

Figure 2.13 Bronze Age pottery from Zone B, area with pits. Note P11 shows interior of sherd.

is interpreted as late in the sequence of use of the roundhouses and associated soil deposits.

P11, the gabbroic vessel, has no close parallels on Scilly. It may be a lid, rather than a dish, and for that there is a good mainland parallel among the Late Bronze Age Plain Ware assemblage at Higher Besore (Quinnell Section 5, Figure 5.32, No 65). Bronze Age gabbroic vessels are rare on Scilly. Several gabbroic vessels of

Descriptions of enumerated vessels from Zone B (Figure 2.13 and Table 2.5) P9 flat-topped rim from necked jar with roughly smoothed surface (drawn from two non-joining sherds); several other sherds from neck of this or similar vessel. Context

Sherds Weight Details

OLS (138)

4

(134) sand over/in gully [137]

1

(1360 sand over/in gully [137]

47g

P13 and three sherds from other vessels

7g

Small sherd from slightly everted rim

2

11g

Cairn 154 make-up (121)

11

127g

Cairn 154 make-up (151)

6

67g

14g

Body sherds

Cist [163] infill (156)

54

517g

P15, eight small rim sherds ? from P16 or P17. All body sherds could be from P15, P16, or P17

Cist [163] infill (155)

136

1150g

Cairn 154 make-up (119)

Cist [163] infill (150)

3

210

2138g

Cist [163] infill (114)

242

1887g

Totals

666

5977g

(135) recent marine sand

1

12g

P14 and one other sherd

Body sherds

Body sherds

P16, five rim fragments, part of base and circa 20 body sherds from P17, other body sherds

P17, P18, 10 body sherds from P16, other body sherds

P19, P20, sherds from P16, P17 and ? P18b, other body sherds

Eroded body sherd

Table 2.6: Summary of pottery from Zone C.

29

Later Prehistoric Settlement in Cornwall and the Isles of Scilly broadly Plain Ware type come from the settlement at Higher Town, St Agnes (Quinnell, forthcoming a) together with two sherds from Middle Bronze Age Trevisker gabbroic vessels. (The present author regards the miniature gabbroic vessels from Nornour (Dudley 1968, 5 and fig 7, no 72; Butcher 1978, fig 36, no 156) as of probable Roman date).

sizable Old Quay Neolithic assemblage (Quinnell 2017a) and it is currently uncertain if the Neolithic carinated bowls were made on Scilly. If P19 is of Neolithic date the even size of the sherds and the fact that they only occur in the top layer of cist [163] suggest some form of deliberate redeposition. However, carinated bowls of rather different shape are a major component of Late Bronze Age plain assemblages of the early first millennium BC and have already been recognized at West Porth, Samson (Quinnell 1994; Ratcliffe and Straker 1996, 83–84).

Descriptions of enumerated vessels from Zone C (Figure 2.14 and Table 2.6) P13 rim sherd of bowl, granitic fabric with fine matrix and sparse grits, comparable in fabric and form to Neolithic material identified from the cliff at Old Quay, St Martins (Quinnell 1994) and more recently from excavations (Quinnell 2017a).

P20 rounded lug from side of biconical vessel; reduced brown/black fabric with external sooting; very irregular finish. Approximately 20 large body sherds could come from this vessel (sherds not noted in any other context).

P14 small sherd with row of impressed circles circa 4mm across; this form of decoration occurs in multiple lines on vessels of the second millennium BC for example, no 104 from Nornour (Butcher 1978, fig 32).

Discussion of Zone C material Porth Killier, with P13, and more tentatively P19, is now one of 12 sites on Scilly from which Neolithic ceramics have now been identified. Two others are on St Agnes, north-east of Periglis Cottage and near the Turk’s Head (Quinnell 1994). The publication of the Neolithic assemblage from Old Quay, St Martins, provides an update on Scillonian Neolithic ceramics (Quinnell 2017a).

P15 thin rim sherd narrowing to an almost pointed top, slight finger modelling traces, reduced grey; circa 10 body sherds in this distinctive grey fabric may belong to this vessel. P16 side of biconical vessel with slightly expanded, flattopped rim and squarish lug; reduced grey fabric with lighter exterior surface; fabric distinctive for abundant grit circa 40 per cent. As only one lug present, unclear whether vessel had two or four lugs.

P14 belongs to the decorated early second millennium BC Scillonian ceramics and may relate to early phases of activity in Zone A.

P17 rim sherds (four) from thick biconical vessel, with simple rounded top; very irregular lumpy fabric; reduced grey; at least 50 body sherds present; flat base probably that in (155).

Cairn 154 make-up layers contain a total of 20 sherds, none with any distinguishing features. The infill of cist [163] in the cairn contains 642 sherds, all of which, except possible Neolithic P19, could have come from the five vessels P15–P20. These five vessels are all plain biconical forms with lugs surviving on P16, P18 and P20. Most sherds are small, 30–35mm across or less. Where joins can be found, these are mostly within groups recorded as single small finds, suggesting sherds have fractured on deposition or through post-depositional processes. However, the small number of rims and virtual absence of bases (except for P17) indicate that only parts of vessels represented are present. It is of course uncertain how much of the deposit remains unexcavated and how much was lost before excavation. The general condition of the sherds suggests deliberate smashing into small pieces before deposition. Fractures are generally fresh.

P18 flat-topped rim from biconical vessel, reduced but with slightly oxidized patches on exterior. Oval lug drawn as part of this vessel because fabric is similar. A second lug, eroded, may come from (114), but unclear how many lugs vessel had. P19 (not illus) carinated vessel represented by about 15 sherds, including one rim, which do not make a continuous profile; distinctive fabric, fine matrix with sparse grit some of which is circa 5mm: reduced grey with darker external surface, well-made and wellfinished externally, rough on interior. Noticeable that smooth surface is continuous despite large grits. The only sherds of P19 not in (114) are two in (150) directly beneath it. The shape of the carinated form is not certain as the profile is not continuous. It is just possibly of Neolithic date; the fabric, with its large grits in a smooth matrix, could represent a fine-ware variant of the Neolithic granitic fabrics so far identified on Scilly. However, carinated bowls do not occur in the

In deposits from ritual/funerary contexts, certain constraints might have applied to their selection and the material may not have straight forward chronological implications. Plain lugged biconical vessels are found throughout the second millennium BC domestic sequences at Nornour (Butcher 1978) and Little Bay 30

Chapter 2.3: The Artefacts

Figure 2.14 Porth Killier. Neolithic bowl P13 and Bronze Age pottery from Zone C, area with cairn.

appears to be the only cairn in which plain biconical forms alone occur. The radiocarbon dates from the Porth Killier cist are not helpful 3512 ± 70 BP, 2030–1660 cal BC (Wk-5690) comes from (156) from the base of the infill and 2973 ± 73 BP, 1410–1000 cal BC (Wk-5691) from higher in the cist. The interpretation of these dates is complicated by the recognition of possible Neolithic material in the cist infill, introducing the possibility that earlier charcoal might be incorporated in the infill. Whatever its date, the assemblage was likely to have been carefully selected for deposition in the cist. A possible parallel for this is the assemblage from a shallow pit at Dolphin Town, Tresco, with parts of a plain lugged vessel (Quinnell 2009–10). The smashing of the vessels indicates certain symbolic factors were

(Wardle 1983). A clear case has been made at Little Bay for this plain material to have survived later than decorated biconical forms; a single radiocarbon date calibrated to 2780 ± 80 BP, 1190–800 cal BC (HAR-1726) comes from later contexts with undecorated forms. The classic assemblage from Obadiah’s Barrow on Gugh, very close to Porth Killier in terms of second millennium BC geography, contains a wide range of decorated biconical forms (Hencken 1933, fig 10), remains undated. Both the assemblages at Bant’s Carn (Ashbee 1976) and at Knackyboy Cairn (O’Neil 1952) contain decorated and plain biconical forms. For the latter the sequence with plain vessels deposited later than those with decoration now supported by modern dating (Sawyer 2015). Salakee Down on St Mary’s (Grimes 1960, 170–180) 31

Later Prehistoric Settlement in Cornwall and the Isles of Scilly involved in the infill, as does the probable redeposition of earlier material. This selection could have occurred at any time during the second millennium BC, and the question of the relationship of the cairn and cist to the nearby roundhouses must remain open.

Topsoil (1) SF 6, 14, 46, 53, 80 six small eroded sherds including a rim. Area B1 – sequence in fill of pit complex [50]. [57] pit SF147 three small body sherds with sparse fine grit < 2mm. (49) fill of [57] SS16 small sherd as those in [57]. (48) sealing [57] SS15 15 sherds are probably all daub. [55] cut into [48] SF141 two sherds as those in [50]. (52) fill of [55] one sherd. (51) primary fill of [50] SS17 body sherd. (47) fill of [50] with boulders, includes SF 137, 138, 139, 140, 144, 146 and SS14; total of 55 sherds from at least three vessels on fabric variations including P9 (47/2, 47/4). (46) upper fill of [50] 12 sherds including P10 (46/11, 46/11); P11 (46/8). [50] (not assignable to more detailed contexts) 15 sherds including P12.

Details of SF numbers for pottery by context Zone A Old land surface (4) beneath roundhouses, includes SF 4, 5, 8, 11, 12, 16, 17, 19, 20, 22, 25, 26, 37, 38; 31 sherds. Junction of (4) with soil (3) above it includes SF 21, 33, 34; three sherds. Soil (3) over (4) at east end of site, includes SF 2, 9, 10, 13, 18, 27, 44; seven sherds. Midden (9)/(13) preceding roundhouse 20, (9) includes SF 58, 63, 64, 67, 68, 69 and SS3, [13] SS4; 15 sherds including P1 (SF 68/1, 68/3 and non-joining 68/2),same vessel as APK 89 1 / 2 from then context (14); P2 (SF 64, 67, 68/4), same vessel as APK 89 23/3, 23/5, 23/15 from then context (14) and almost certainly as APK 90 1/1, 1 / 2 from then context (30); P3 (SS3). (18) wall-core of roundhouse 20 SF17 one sherd. Soil/midden accumulation (10) east of roundhouse 20 SF73 one sherd. Soil/midden accumulation (12) east of roundhouse 20 SF 49, 50, 54 four sherds including one from vessel in midden (9). Soil/midden accumulation (8) east of roundhouse 20 SF48 one sherd. (82) wall-core of roundhouse 21 SF78, 79 two sherds; (80) SF84 one sherd; (81) SF79 one sherd. (78) depression in basal floor (29) of roundhouse 21 SF136 one sherd from girth of vessel with eroded incised decoration. (63) fill of posthole [62] in (29) roundhouse 21 SS24 five sherds. (42) fill of gully [41] in (29) roundhouse 21 one sherd. [70] post pit ? relating to basal floor (29) roundhouse 21 seven sherds. (28) build up over basal floor [29] roundhouse 21 SF103,104, 113, 115, 118, 119, 120, 121, 122, 123, 124, 125, 127, 128, 130, SS9. (27) material in/on upper floor in roundhouse 21 SF 96, 97, 105, 108, 109, 111, SS8, 21 sherds which, unusually, group clearly into four vessels including P4 (SF96) and P4a (SF 97, 105, SS8). (68) ? part of floor (27) SF 100, 102 two sherds as SF 108, 111 in [27]. (31) fill of pit [30] SF 75, 76 two sherds. (25) lower part of midden over roundhouse seven sherds. (23) lower part of midden over roundhouse SS6 two sherds.

Zone B2 (92) fill of posthole [91] in pit [86] SF152 large body sherd. (90) fill of posthole [88] in pit [86] SS31 six sherds including two tiny rim fragments from vessel 7mm thick. Zone B3 (100) fill of pit [97] SS33 six body sherds. (99) above (100) in pit [97] SF156 five sherds including basal angle, the outer side of the base retaining slight impressions of course circular matting. (102) fill of pit [104] (cutting [97]) eight small body sherds. Zone B4 No ceramics. Zone B (84) sealing all prehistoric contexts SF 88, 89, 90 three eroded body sherds. Zone C with cairn 154 Old land surface (138) four sherds including P13. (136) sand over/in gully [137] two sherds including P14. (134) sand over/in gully [137] small rim sherd with slight out turn. Cairn 154 make-up layers contain a total of 20 sherds, none with any distinguishing features. Cairn 154 make-up context (121) SF162, SS39 eleven sherds. Cairn 154 make-up context (119) SS38 three sherds. Cairn 154 make-up context (151) SS2 six sherds. 32

Chapter 2.3: The Artefacts Cist [163] infill contains 642 sherds. Most sherds are small, 30-5mm across or less. The interior was excavated in 5mm–10mm spits. These are, upwards. Cist infill (156) SF 191, 200, 201, 202, 205, SS53, a total of 54 sherds. Eight small rim fragments could belong to P16 or P17 except P15. All body sherds could come from P15, P16 or P17. Cist infill (155) SF 189, 186, 193, 196, 197, 199, 200, 203, 204, SS52, a total of 136 sherds including five rims, part of base and circa 20 body sherds from P17; also P16 Cist infill (150) SF 179, 180, 181, 182, 183, 184, 185, 188, 190, 192, 194, SS50, SS51, a total of 210 sherds including circa 10 body sherds from P16 and P17; at least 50 body sherds present; base probably that in [155]; P18c. 20 body sherds could come from this vessel. Cist infill (114) SF 163, 164, 165, 167, 168, 177, 178, SS35 a total of 242 sherds, and three chunks of daub, includes sherds from P16 and P17, a squarish lug which may belong to P18; also P20 and P19.

coarsest grained to the finest present in the assemblage: coarse, medium coarse, medium fine, fine. The three non-granite pieces are [25]/2 – quartzite, SF148 – spotted slate, and SF176 – veined quartz porphyry. The latter two derive from rocks in the contact zone around the local granite. The quartzite is of an appropriate type to arrive, as an erratic, from the same source as the Cretaceous flint which is found in Scilly. Eighteen pieces come from Zone A, distributed through the contexts below, associated with and post-dating roundhouses 20 and 21. Four are from Zone B, three from fills in pit [50]. Three are from Zone C of which two pieces unique to the assemblage come from cairn 154. Stone artefacts from Scillonian settlements have received little detailed study. Because of the frequency of finds of pebbles, distinguishing those which show signs of use or working is not easy; until the last few decades all but the most obvious artefacts were probably not recognized. Only the reports on Nornour (Butcher 1978), Little Bay (Neal 1983) and Halangy Porth (Ashbee 1983) have sufficient detail for comparative discussion. Stone artefacts from the ISEP were all without contexts and are briefly summarized (Ratcliffe 1991, 67–71); the small number of artefacts from the CEP has not so far been discussed as a group (Quinnell 1994); the relevant extract from the archive report is included at the end of this section. Of the artefact groups so far represented in prehistoric contexts, troughs, whetstones and holed weights are the principal types not present at Porth Killier.

Possible crucible sherd David Dungworth The possible crucible sherd recovered from gully [157] at Porth Killier consists of a ceramic material with abundant quartz grain temper (fragments of mica are also visible). The outer surfaces are vitrified and dark in colour. The sherd is too small to allow the form of the vessel to be reconstructed. Energy dispersive x-ray fluorescence analysis of the sherd detected the following elements: Al, Si, Ca, Ti, Fe, Cu, Zn, Sr, and Zr. One small peak was seen in a position which could indicate either As or Pb but it was not possible to determine which element was present.

More recent studies, most notably of the artefacts from the later Bronze Age structure at Callestick (Quinnell and Taylor 1998–9), have highlighted the way in which suitable stones can be used and reused as artefacts in different ways. This is understandable when a site may be some distance away from sources of suitable stone. Table 2.7 shows that quern SF154A was also used as a rubbing stone and hammer and that muller SF28 was also used as a small mortar. As suitable stone is likely to have been easily obtainable this reuse presumably relates to the convenience of using stones which were on the site, as opposed to collecting from the beach which may then have been some distance away.

The abundant temper, the vitrified surfaces and the detection of non-ferrous metals all point to the sherd being a crucible. Stonework Henrietta Quinnell Twenty-five artefacts with indications of working or of usage were found (Table 2.7). All, except pivot stone SF175, were ‘pebbles’ and could have come from a local beach which would have been some distance to the North of the present St Agnes coastline. Geological terminology is used for size: pebble 4-64mm, cobble 64mm-256mm, boulder >256mm. Twenty-two pieces are of granite, for variations of which the geological terminologies are complex and varied. A subjective division has therefore been used here, grading from the

Pivot stone SF175 (Figure 2.15) was found set in (27), the upper floor in roundhouse 21. This stone was a flattish pentagonal block, the outline of which appeared to be natural but with the bottom made roughly level by hammering. The pivot socket was set eccentrically, a circle 90mm across and 60mm deep. The interior had worn very smooth, but with a series of scratches concentric to the circle, and the flattish base rough and 33

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 2.15 Porth Killier. Cross section of pivot stone SF175 from floor [27] roundhouse 21 Zone.

scarred. The edge of the socket was convex all round. Pivot stones are fairly frequent finds on Scilly (for example, Ratcliffe 1991, 67–68) and occurred at Little Bay (Neal 1983, 59), a site with a long occupation within the Bronze Age but with no subsequent use. While pivot stones are not uncommon, there appears to be no clear published account of how they actually functioned. The concentric scratches are most likely to be caused by sharp quartz grains. It may be suggested that the tip of a post rested in the pivot packed with sand, both to help stability and to ease friction. If a wooden collar was fitted around the post, overlapping the edge of the pivot, this would have had the effect of keeping the packing sand in the pivot and producing the kind of wear around its top, the convexity has been remarked upon above. The subject is one which would benefit from a more detailed examination of surviving pivot stones and from experimental work. (The author is much indebted here to Carl Thorpe for discussing possibilities).

Figure 2.16 Porth Killier. Broken saddle quern SF154A, [52] in fill of Pit [50] Zone B. Arrows indicate battering on side.

have been available. Small querns may relate to some special form of food processing, as may the mullers. Bruising mullers Twelve mullers, none complete, were retrieved. ‘Muller’ is the normal term describing the stones used for attrition of food, usually assumed to be cereal, on saddle querns. As saddle querns are chosen with an interior concavity to aid the collection of processed material, and as this concavity increases through wear, any ‘muller’ needs a (slightly) convex surface. This convexity of worn surfaces is a major feature in distinguishing mullers. Those at Porth Killier were all of elongated shape and smooth oval cross-section, of which SF72 (Figure 2.17) is typical; only [23]/2 comes from a tool of more rounded shape and thicker cross-section. Battering, on the ends or around parts of the edges, was noted on all the Porth Killier examples. This, on close examination, was not the result of subsequent use of mullers as hammers; traces of abrasion were intermixed with the smooth wear on the flat surface(s), sometimes partly smoothed away and sometimes intruding on it on the same artefacts. This interrelationship of two types of wear suggests that the ends or sides of tools had been used for bashing or crushing while the flat surface(s) were used for attrition. Such mixed traces of usage would result from the regular practice of partly flattening

A second pivot stone, in general similar to SF175, was found upside down in (114), the upper layer of fill cist [163]; this was removed by Francis Hicks to the Old Lighthouse. It may be noted that the cist also incorporated a saddle quern, and, with so much stone readily available, the incorporation of used objects could have had some deliberate, symbolic purpose. Saddle querns Three were retrieved; a fourth was left in position forming the side of cist [163] in cairn 154. They are made on boulders or cobbles probably less than 0.5m long; one, SF154A (Figure 2.16), has had its narrow edges removed to produce a more solid working surface. These querns are generally less than half the size of those recorded from Little Bay, Nornour and Halangy Porth and smaller than Bronze Age saddle querns from mainland Cornish sites of this date, for example those from Trethellan (Nowakowski 1991, 145–148). There seems to be no reason why larger blocks should not 34

Chapter 2.3: The Artefacts

Figure 2.17 Porth Killier. ‘Bruising mullers’ SF72, from (9) midden predating roundhouse 21 Zone A and SF151, [52] in fill of pit [50] Zone B. Cupped pebble hammer SF94 from floor [27] roundhouse 21 Zone A. Arrows indicate battering on side.

as a bruising muller. A number of the 16 artefacts from Halangy Porth described as ‘rubbing and polishing stones used as hammer stones’ (Ashbee 1983, 27) are probably mullers of this type. It may be expected that more will be subsequently identified as more is learnt about stone artefacts, a topic which has generally been neglected in the past. Further discussion regarding mullers is included in the report on the later Bronze Age structure at Callestick (Quinnell and Taylor 1998–9).

grain, or other material, before grinding. Flattened or bruised grain in small quantities might have been easier to control and work on the small saddle querns described above. Alternatively, there may have been some factor special to the cereal or other food product worked which benefited from the suggested process. It should be noted that these distinctive mullers are found throughout the stratified sequence in Zone A and in pit [50] in Zone B. It has been thought appropriate to distinguish these mullers as bruising mullers. Of the four mullers identified in the CEP, APK 93.19, from then context (32), can, on re-examination, be classed

Rubbing stones This is a general term used for stones with worn facets which do not fall into any distinctive 35

Later Prehistoric Settlement in Cornwall and the Isles of Scilly category, and generally have concave as opposed to convex wear. Of the six from Porth Killier, [123] may be part of an elongated ‘lapstone’, a tool traditionally used for the working of leather. SF148 (Figure 2.18), with SF40 and [25]/2, are similar in that their shape provides a good grip on an object used in a way which would produce concavity. This effect would have been produced by working leather, or perhaps cloth or fibres, which were backed by a curved convex surface such as the top of a boulder. The four pieces suggested for leather working are all of fine-grained rock.

had surviving maximum dimensions of 115mm by 105mm by 55mm. The pebble has a naturally rounded end; its two opposing long faces had been worn very flat. APK 93.19, probably from the same context, was similar in shape and the two flat surfaces, and of much the same dimensions. Both are appropriate in size and shape to have been used as riders for saddle querns. APK 93.17, context (24) or (26) in upper fill of the roundhouse, and APK 93.14, from material fallen from the roundhouse, were both too small for any suggestions to be made as to their use.

Figure 2.19 Hammerstone SF176, from (150), cist infill in cairn in Zone C.

Figure 2.18 Rubbing stone SF148, from (84) soil build up over Zone C.

Cupped pebble hammer SF94 (Figure 2.17) from the floor in roundhouse 21 is the only example. Pebble tools, often with opposed cups or depressions, have been reviewed by (Roe 1985) with no examples quoted from contexts after the Early Bronze Age. No Cornish site published since Roe’s review has produced these except Callestick (Quinnell and Taylor 1998–9, no 87); the context here is later Bronze Age with the possibility of the redeposition of earlier material. No example is so far known from Scilly. If the earlier Bronze Age date is upheld by future work, then the cupped pebble hammer may be redeposited from activity earlier than the roundhouses.

Extract from the 1994 CEP report Parts of four granite pebbles used as rubbing stones were found; they all showed signs of heavy burning before breakage. APK 93.20 from context (25), midden inside roundhouse 32, had surviving maximum dimensions of 115mm by 105mm by 55mm. The pebble has a naturally rounded end; its two opposing long faces had been worn very flat. APK 93.19, probably from the same context, was similar in shape with the two flat surfaces, and of much the same dimensions. Both are appropriate in size and shape to have been used as riders for saddle querns. APK 93.17, context (24) or (26) in the upper fill of the roundhouse, were both too small for any suggestion to be made as to their use. The use of local pebbles for tools is especially common on Scilly (for example, Butcher 1978, 95) but unfortunately details are too often not reported (see Neale 1983) (Quinnell 1994, 9).

Hammer stone SF176 (Figure 2.19) is of distinctive black and white veined quartz porphyry with a very smooth surface showing the colours of the rock. Its surviving end has been heavily shattered by hammering and it has broken across a line of weakness. Hammer stones are generally quite common finds, although many granite pebbles have only small areas affected by use as hammers. At Halangy Porth the use of quartz, as opposed to granite, is marked (Ashbee 1983, 27) and it is suggested that a non-granite rock was chosen for its durability. Function may have affected the special choice of rock for SF176. Its distinctive appearance, however, may relate to the specialized deposition of objects in cairn 154; its deposition might be considered as related to that of a layer of white pebbles in the cist in the barrow at Trelowthas (Jones et al 1997, 59). Parts of four granite pebbles used as rubbing stones were found; they all showed signs of heavy burning before breakage. APK 93.20 from context (250), midden inside roundhouse 32, 36

Chapter 2.3: The Artefacts

Artefact Pivot stone SF175

Saddle quern SF154A

Illus ? Rock type

Width

Amount Description

Yes

Coarse granite block

270mm complete See text – found upside down

Yes

Coarse granite boulder

Sides dressed by battering to consolidate edge; one face worn concave through long use

(add)

2/3rds

Saddle quern SF29

No

Coarse granite boulder

120mm

¼

Saddle quern ? [25]/3

No

Coarse granite cobble

130mm

½

No

Medium granite cobble

Yes

Medium fine granite cobble

No

Medium fine granite cobble

Bruising muller SF150

No

Medium granite cobble, hardly water worn

Bruising muller SF117

No

Bruising muller [53]

Bruising muller SF28

Bruising muller SF72

110mm

½

94mm

½

Oval, both surfaces convex, wear flattening some quartz crystals, battering on end; one (4) OLS Sector A face worn concave from use as small mortar Oval, both sides worn smooth, very slightly concave; some battering around much of surviving edge

95mm

2/3

Oval, with flatter side worn but slightly convex, end battered; wear on fracture suggests continued use after breakage

Medium granite cobble

110mm

1/3

Oval, both sides worn and slightly convex, battering on end

No

Medium granite cobble

97mm

1/3

Oval, one side worn, slightly convex, battering on end

Bruising muller SF95

No

Medium granite cobble

80mm

1/5

Segment of oval muller with worn convex surface and battered edge

Bruising muller [31]/1

No

Medium coarse granite cobble

57mm+

2mm

Modern pupae. O-M >2mm.

*

O roots/tubers. Modern fly pupae.

F>2mm incl. twigs. Bone. Modern fly pupae.

-

F>2mm. Bone. Modern fly pupae.

M>2mm. Bone.

O>2mm. Modern fly pupae.

M>2mm

Comments; Charcoal >2mm

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

O including. corn spurrey.

-

-

-

M-F. Ulex, Carex, Potentilla

O including. Danthonia

O

-

-

O>2mm

Modern pupae. M>2mm

O>2mm

O>2mm

-

Modern pupae. O>2mm

-

I topshell. O>2mm. -

O-M tubers. F>2mm - twigs.

O>2mm.

-

-

O barley M-F Chenopodium, Carex, O tubers. O>2mm rachis Raphanus, Danthonia

-

-

-

-

-

-

-

-

-

M wheat, barley. Celtic bean

O barley

-

-

Chaff

M-F barley

O wheat

Crops

Chapter 2.6: Plant Macrofossils

(90)

31

32

(99)

(119)

38

39

54

(121)

(138)

(139)

(143)

(150A)

(150B)

(155)

(156)

46

47

48

50

51

52

53

54

-

(158)

(144)

(136)

(134)

44

45

(133)

(132)

(131)

(123)

43

42

41

40

(118)

(116)

36

37

(114)

(102)

(100)

35

34

33

(87)

(77)

29

30

Context

Sample

5 / 100

3 / 20

9 / 100

14 / 20

12 /40

9 / 35

14 / 50

1/1

3/1

2 / 10

13 / 50

5 / 30

4/5

13 / 40

12 / 30

5 / 50

3/1

10 / 40

10 / 100

2/5

39 / 300

14 / 50

9 / 30

8 / 40

4/1

2/1

6

4

10

14

14

10

14

3

5

3

4

7

6

16

18

6

4

12

11

6

45

18

13

10

6

3

lower pit fill

make up for cairn 154

fill of gully [137]

O

O

O (fragments)

O

-

O (fragments)

-

-

-

O

-

F grain mainly naked barley; O celtic bean

O barley

-

O

O

-

Crops

soil + rubble fill of wall

fill of chamber or cist in 154

fill of chamber or cist in 154

bottom half fill cist / chamber in 154

top half fill cist / chamber in 154

marine sand

marine sand

disturbed old ground surface below wall

-

-

-

-

O naked barley

O barley inc. naked. O celtic bean

O naked barley, indet. barley

-

-

-

-

-

O

O

O

-

-

O

-

-

-

-

-

-

O-M

-

O

O

-

-

-

-

-

O hulled and naked barley. 1 oat

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

O

-

Weeds

-

O barley

-

-

-

Chaff

-

O

-

old ground surface and tumble from O grain (fragments); O cairn celtic beans

fill of gully [137]

build up against cairn

old ground surface

tumble from cairn

fill of natural hollow

make up for cairn 154

make up for cairn 154

old ground surface

make up for cairn / upper cist fill

backfill pit [101]

primary fill of pit [97]

secondary fill pit [98]

posthole fill

pit fill in roundhouse 21

Vol. sample/ Weight Type float (l/ml) (kg) M>2mm

-

O>2mm. Modern fly pupae.

M>2mm

F>2mm

O-M >2mm

O-M >2mm

-

-

-

O>2mm. Modern fly pupae.

O>2mm

O tuber fragments. Modern fly pupae.

M>2mm. O roots/tuber fragments. Modern fly pupae.

O>2mm. Modern fly pupae

Modern fly pupae

-

O>2mm

O>2mm

O>2mm

F>2mm, include. twigs

O>2mm

O>2mm

O>2mm

O>2mm

-

Comments; Charcoal >2mm

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

737kg

99

596 l

8 / 20

10

9 / 10

Oat: Avena sp. Hulled barley: Hordeum sp. Naked barley: Hordeum sativum var nudum Wheat: Triticum sp. Grain fragments: Triticum / Hordeum (wheat / barley) Celtic bean: Vicia faba var minor Wild / sea radish: Raphanus raphanistrum / maritimus Corn spurrey: Spergula arvensis Goosefoot: Chenopodium sp. Gorse (seeds): Ulex sp. Tormentil: Potentilla sp. Heath grass: Danthonia decumbens (L.) DC Sedge: Carex sp.

-

-

Crops -

-

Chaff -

O

Weeds

Table 2.20: Porth Killier 1996 plant macrofossil assessment.

old ground surface below wall 130

old ground surface below cist [163]

Vol. sample/ Weight Type float (l/ml) (kg)

Key: Rare: 0–10; Moderate: 11–50; Frequent: >50 items

(116B)

(116A)

55

56

Context

Sample -

O>2mm

Comments; Charcoal >2mm

Chapter 2.6: Plant Macrofossils

55

Chapter 2.7: Radiocarbon Dating Charles Johns contexts at Porth Killier and two samples from an intertidal peat deposit at Porth Coose. All the samples taken in 1996 were sent to the radiocarbon laboratory at the University of Waikato in New Zealand.

Three radiocarbon determinations were obtained from Porth Killier during the Coastal Erosion Project in 1989 (Figure 2.21 and Table 2.21). Six samples for radiocarbon determinations were taken from sealed

Feature

Lab. no

Age BP

Calendrical years 95.4%

Material

Samples collected from Porth Killier in 1996 Pit [50]

Wk-5686

2641±71

Pit [50]

WK-5687 2841±68

Roundhouse 21

Wk-5688

2533±79

Charred plant tuber circa 1gm. Not further identified

Charred barley grains (Hordeum sp.) weighing circa 1gm. In tube A Hordeum sativum var nudum (naked barley); in tube B Hordeum sp. indeterminate barley (could be hulled or naked)

980–540 BC 1210–840 BC

Circa 12gm comprising two fragments of charred Vicia faba var minor (celtic bean) and three fragments of charred plant tuber which was not further identified. It is assumed that all this was used

810–410 BC

Wk-5689

2794±74

Context (156) from cist / Wk-5690 chamber [163] – cairn 154

0.35gm charcoal fragments comprising Ericaceae (heather family) wood, a twig that was unidentified but probably from the Leguminosae 1190–800 BC (gorse/broom family) and two unidentified charred root fragments that were probably from the Ericaceae family

3512±70

0.13gm charred twig of Ulmus sp. (elm) of 10-20 years’ growth

2030–1660 BC

Context (150A) from – cist / chamber [163]

Wk-5691

2973±73

0.12gm grains of charred Hordeum sativum var nudum (naked barley)

1410–1000 BC

AGNPK 1 (layer (14), top of midden)

OxA-3647 2794 ± 74 Hordeum sp. (barley) grain, charred

1670–1300 BC

OxA-3648 3512 ± 70 Hordeum sp. (barley) grain, charred

1620–1280 BC

Roundhouse 21

Samples collected from Porth Killier in 1989

AGNPK 2 (layer (14), base of midden)

AGNPK 3 0.3gm [circa 16 fragments] Calluna vulgaris (heather) charcoal. Age of (context (5), base OxA-4700 2973 ± 73 wood circa 4 years of Bronze Age sequence

1370–970 BC

Intertidal peat deposit APC7

Samples collected from Porth Coose in 1996 Intertidal peat deposit APC8

Wk-5693

1650 ± 60 Peat deposit

AD 250–550

Wk-5694

1510 ± 50 Peat deposit

AD 420–640

Table 2.21: Radiocarbon determinations from Porth Killier.

56

Chapter 2.7: Radiocarbon Dating

Figure 2.21 Results from the radiocarbon dating at Porth Killier and Porth Coose.

57

Chapter 2.8: Discussion ‘Tapestry’ excavations

Coastal change

Parker Pearson has described the excavation of thin trenches through eroding mounds at Dun Vulan on South Uist as ‘tapestry’ excavations. He considered these limited excavations were considered doomed to fail in understanding the spatial ramifications of the complex machair settlement sites that were investigated, (Parker Pearson et al, forthcoming). This epithet could be applied to the work undertaken at Porth Killier during the 1996 CPS scheme, and previously, and by extension the term can be used to describe the general recording of eroding cliff-face sections which for the last 20 years has been one of the main forms of archaeological investigation undertaken in Scilly (Figure 2.22). Whilst the recording of such sections, collecting of finds and occasional sampling has added greatly our knowledge of prehistoric Scilly while extracting the maximum amount of information from the minimum amount of intervention, the interpretation and understanding can only be partial, often raising as many questions as those that are answered.

Palaeogeographic reconstructions by the Lyonesse Project, a recent study of the coastal and marine historic environment of Scilly, indicate a rapid rise in sea-level during the late Mesolithic (Charman et al 2016). The single land mass of Scilly that had existed at 9000–7000 cal BC, began to break up and St Agnes and the other western islands became separated from the northern islands by around 5000 cal BC, although part of the area between St Mary’s and St Agnes would have been intertidal flats. By the beginning of the Middle Bronze Age, circa 1500 cal BC, the configuration of the islands was approaching that of the present day, with the most dramatic difference being that there was a vast intertidal area stretching right across St Mary’s Road and covering almost all of the region between the main islands of St Martin’s, the Eastern Isles, Tresco, Bryher and Samson. At this time the inhabitants of Porth Killier would have looked out on an intertidal area of about 300m in extent (Figure 2.23).

Figure 2.22 Foundation trench for the concrete sea wall at Porth Killier showing the archaeological features exposed in the cliff face. (Photograph: Cornwall and Scilly HER at Kresen Kernow.)

58

Chapter 2.8: Discussion

Figure 2.23 Modelled land and intertidal areas in Scilly at about 1500 cal BC, showing intertidal and coastal field systems (purple dots) recorded in the Cornwall and Isles of Scilly HER. (Source: Charman et al 2016.)

The sediments outcropping on the beach at Porth Coose were re-sampled by the Lyonesse Project and are dated to the first millennium AD, between 1835 ± 35 BP, cal AD 80–320 (SUERC-38089) and 1507 ± 29 BP, cal AD 430–640 (OxA-23860) overlapping with those from Old Town, Lower and Higher Moors, St Mary’s, but representing a different part of Scilly. The pollen spectra suggest a similarly open landscape with minor woodland elements but with more elm (Ulmus) in the woodland cover, as well as birch, oak and hazel. The pollen and algae show that the sediment accumulated under aquatic conditions and some elements such as Cyperaceae (sedges) are probably of mostly local origin. The grassland flora has very high abundances of disturbed ground plants with almost 60 per cent plantain species (Plantago) in the lower part of the core.

Until the excavations at Old Quay, St Martin’s in 2013/14 (Quinnell 2017a and above), two pits from East Porth, Samson, had produced the largest assemblage of Neolithic pottery found in Scilly (Neal and Johns, forthcoming). Indeed, until 1989 the only other identified Neolithic artefacts from Scilly were a few stone axes, mace-heads, adzes, flint arrowheads. There was also some misidentified pottery from entrance graves, but recent reappraisal has shown that no Neolithic ceramics can now be attributed to these monuments (Jones, Breen et al 2012, 59; Sawyer 2015, 117). The 1989–93 Coastal Erosion Project recovered definite Early Neolithic pottery from Old Quay, St Martin’s, and also at cliff-edge site below Bonfire Carn, Bryher (Quinnell 1994), leading to the identification of more Neolithic sherds in the Isles of Scilly Electrification Project assemblage from two sites on St Agnes, Veronica Farm on Bryher, Dolphin Town on Tresco and Lower Town, St Martin’s (Quinnell 1991).

Neolithic pottery The 15 sherds from a probable carinated Neolithic vessel recovered from the cist during the 1996 CPS recording at Porth Killier, St Agnes bring the total number of sites in Scilly associated with Neolithic pottery to 9 (when the 3 entrance-grave finds are discounted as of Bronze Age date); the various findspots contained several different variants of a fabric made on the islands (Quinnell 1994).

The occurrence of pits containing Early Neolithic pottery and other artefacts in Cornwall and elsewhere has been discussed by Cole and Jones (2002–3). The East Porth pits are similar to the features examined 59

Later Prehistoric Settlement in Cornwall and the Isles of Scilly by them at Tregarrick Farm, Roche, and the Neolithic pottery in the East Porth pits is almost certainly the result of structured deposition, as is the inclusion of the two small granite and pumice stones in pit A. The provenance of the probable Neolithic pottery from Porth Killier may also suggest curation and structured deposition.

destroyed, there are 384 recorded cairns, the majority of which occur in cairnfields, and date from the first half of the second millennium BC, the Early to Middle Bronze Age. Most are small, between 4m and 7m in diameter and perhaps 0.5m high although there are some larger ones, up to 22m in diameter and 2.3m in height. They vary in form, some being a simple mound of stones, others having a stone kerb. A proportion contain stone-lined cists or cover shallow burial pits.

Numerous Middle Neolithic features were excavated at Old Quay, St Martin’s, in 2013/14 including at least 17 pits and 37 postholes (Garrow and Sturt 2017, table 3.16). An assemblage of 27 pieces of worked quartz and 10 fragments of pottery from a curved-sided Neolithic bowl were recovered from one pit (f5). Another pit nearby (f9) contained three further worked quartz fragments and another pit a very large sherd of pottery (f12) (ibid, 85). Three pieces of pumice, 738 sherds of pottery, 308 pieces of flint/quartz and four worked stones were recovered from another large pit (f33) (ibid, table 3.16). The nature of these artefacts certainly suggests an element of structured deposition.

Both entrance graves and cairns are found predominantly on high ground around the Atlantic coastline. The majority of cairns occur within four large cairnfields; Shipman Head Down on Bryher, Castle Down on Tresco, Chapel Down on St Martin’s and Wingletang Down on St Agnes, although there are a number of cairns nearer to the present coastline, such as the example at Pendrathen, St Mary’s which was finally destroyed by the severe storms in the winter of 2013/2014 (Johns and Mulville 2008). Cairn 154 could arguably represent either a cairn with a cist or a ruined entrance grave; although the former is probably more likely. It is also, however, a possibility that the site started out as a roundhouse and was later converted into a cairn.

It had generally been thought that Scilly was not permanently settled until the Bronze Age and that the few Neolithic artefacts found so far represent seasonal visits from West Penwith (Thomas 1985, 101; Ratcliffe and Johns 2003, 5). However, the evidence from Old Quay, St Martin’s, seems to indicate a Neolithic presence in Scilly that was fairly permanent (Garrow and Sturt 2017, 132). The earliest unequivocal land clearance for agricultural use occurs in 4377 ± 30 BP, 3090–2910 cal BC (OxA-23825) at Par Beach, St Martin’s, where the pollen data collected during the Lyonesse Project shows that birch-dominated woodland was replaced by grassland unconnected to sea-level change (Charman et al 2016, 203). The pollen sequence from Higher Moors shows evidence of initial forest clearance and cultivation of cereal crops in the Neolithic (Scaife 2005, 76). The majority of surviving prehistoric houses and entrance graves are unexcavated and some of these may have their origins in the Neolithic, although the recently obtained sequence of radiocarbon measurements from Knackyboy Cairn, St Martin’s, indicates an Early to Middle Bronze Age date for the time of the entrance grave’s principal use (Sawyer 2015).

The pits The digging of pits and burying of pottery and other artefacts and deposits is a well-attested phenomenon in the British Isles during the later prehistoric period (cf Richards and Thomas 1984; Bradley 1990; 2007; Pollard 1992; Cole and Jones 2002–3; Jones and Reed 2006) and extensive archaeological evidence has been discovered in Cornwall for the curation and ‘structured deposition’ of broken potsherds. It has been suggested that the act of pit digging and deposition may have been intended to render activity memorable and fix a connection between people and place. In Cornwall the shape of pits and repertoire of materials placed in them seems to have changed little from the beginning of the Neolithic period into the Bronze Age, other than the changing ceramic types placed in them, although it has been argued that the character of such pits generally developed with time, with more care taken over the objects selected and the pits themselves being better crafted (Cole and Jones 2002–3, 134; Jones and Reed 2006).

It is quite possible that the extent of settlement and activity in the Middle Neolithic on Scilly has previously been underestimated. As mentioned above the use of very well sorted clays in the East Porth pottery indicates that those potting the vessels were very familiar with locally available resources and are likely to have been on the islands for more than a short time

A growing number of prehistoric pits are being identified in Scilly and their significance being reassessed. Two pits excavated at East Porth, Samson in 1971, contained the largest assemblage of Neolithic pottery so far found in Scilly (Neal and Johns, forthcoming). During the investigation of prehistoric structures at Bar Point, St Mary’s in 1972 a layer of occupation soil was found to contain several small pits, one of them filled by a

Bronze Age cairns In addition to 83 recorded entrance graves in Scilly, of which nine are only ‘alleged’ sites and three have been 60

complete pot standing upright (rim missing) whilst there were sherds in several of the others (Butcher and Johns, forthcoming). At Pendrathen, also on St Mary’s, the lower part of a flat-based coarse vessel was found in a clay-lined pit dug into the ram (Samuels 1975, 117), this was similar to a vessel found in a clay-lined pit, at Halangy Porth, St Mary’s in 1936 containing about a dozen sherds from two or three different pots, some calcined bone and charcoal (Gray 1972, 34–35). In 2007 the bases of two Bronze Age urns, containing the nested fragments of smaller vessels, were recovered in 2007 during a watching brief only 250m north of Porth Killier (Johns and Quinnell 2014).

only 20 can be securely dated to the Bronze Age. The excavated sites are: the earliest phase of Nornour (Dudley 1968; Butcher 1978), Little Bay, St Martin’s (Neal 1983), Halangy Porth/Down, St Mary’s (Ashbee 1996) and Dolphin Town, Tresco (Taylor and Johns 2009–10; and Higher Town, St Agnes (Taylor and Johns, forthcoming). Such settlements invariably lie within or adjacent to the remains of broadly contemporary field systems, small rectilinear fields defined by boulder walls, stony banks and lynchets. There is a strong coastal distribution of known settlement sites, possibly near safe landing places. Most occur on lower hill slopes adjacent to the coastline. The preference for the leeward side of hills is a pragmatic choice providing shelter and protection from the prevailing south-west wind (Robinson 2007, 174–180). Those on the inner edges of the archipelago may have overlooked a fertile plain, occasionally inundated saltmarsh or shallow sea depending on which model of sea level rise is adopted.

The pits recorded during the 1996 CPS at Porth Killier only contained a small assemblage of pottery sherds. The types present are simple forms which occur all through the long sequence at Nornour. The function of the pits is unknown. They do not seem to have been ovens or boiling pits and the stones and boulders deposited in them suggest of some sort of levelling off of the area after they went out of use possibly marking a conscious abandonment or change of use in the area.

Radiocarbon dates from a sequence of hearths in House B at Little Bay demonstrate occupation between 2124–1525 cal BC and 1206–800 cal BC with episodic abandonment and modification evidenced by layers of soil, rubble and midden material (Neal 1983, 58). House 6 at Nornour has produced a radiocarbon date of 1430– 1020 cal BC.

Bronze Age settlement Porth Killier has for many years been recognized as the site of a major settlement during the second millennium BC (Quinnell 1991, 77). The geophysical survey suggests that the settlement lies in a depression in the bedrock and extends inland for a distance of up to 12m, although no archaeological stratigraphy or building remains were clearly defined (Jordan 1989; Ratcliffe and Parkes 1989, 9–11).

There is strong evidence that the houses on settlement sites such as Nornour, Little Bay and Halangy were not necessarily occupied simultaneously but constructed, modified, abandoned, reoccupied and rebuilt implying a continuity of settlement pattern (Robinson 2007; 170, 174).

Prehistoric settlements on Scilly consist of small groups of stone-walled circular or oval houses, ranging from 2m to 12m in diameter. Door posts are usually large granite orthostats or substantial coursed pillars, and pivot stones are often a feature suggesting the hanging of hinged wooden doors. Postholes are largely absent from house interiors and it is not certain how they were roofed. Typical internal features are hearths, usually horseshoe, box or circular-shaped; partition walls and benches (often around the central hearth).

The domestic economy of the Bronze Age settlement at Porth Killier seems to have been based on mainly cattle and sheep although animals caught by hunting, trapping and fishing and shellfish seem more important as sources of protein, perhaps reflecting an economy where few domestic animals were kept and not primarily for their meat, although larger bone debris may have been disposed of elsewhere reflecting some differential disposal practices.

Approximately 130 such houses, forming some 57 settlements have been identified on Scilly, of these

61

62

Section 3: Excavations at Killigrew 1996: an Iron Age and Romano-British Industrial Site on the Trispen Bypass, Cornwall Dick Cole and Jacqueline Nowakowski FSA with contributions from Rowena Gale, Sophie Lamb, Albertine Malham, Gerry McDonnell, Henrietta Quinnell, Laura Ratcliffe–Warren, Adam Sharpe, Vanessa Straker and Roger Taylor The partial excavation of a double-ditched sub-circular enclosure discovered by geophysical survey near Killigrew Farm, Trispen, in 1996, has produced new data on enclosures generally classified as ‘rounds’. The site contained a furnace dated to the Late Iron Age, which set the scene for a marked phase of industrial ironworking during the Romano-British period from the second to the fourth centuries AD. The specialized character of the site sets it apart from apparently ‛domestic’ settlement enclosures known as rounds recorded across Cornwall, and highlights complex functional and social relationships between rounds throughout the Romano-British period. An incomplete but unique tin dish was found buried in a pit within the enclosure and this is a new addition to a wide range of tin alloy objects from Roman Cornwall. Evidence of earlier activity at the site included the discovery of sherds of a Beaker vessel in a pit.

63

64

Chapter 3.1: Background to the Excavations In 1996, Cornwall Archaeological Unit carried out an archaeological assessment in advance of the upgrade of the A39 between Carland Cross and Trispen in mid Cornwall (Figure 3.1). The assessment comprised a desktop study, field reconnaissance and a geophysical survey (Gater 1996). This took place in August and September 1996 and was followed by excavation later on that year (Cole 1996).

present A39 have Cornish place-names, a number of which contain the prefix tre denoting a ‘farm estate’, and these would have been in existence by circa AD 1000 (Padel 1985). Local soils are brown earths of the Denbigh Series, a type characteristic of Cornish lowlands. Such soils comprise well-drained fine loams over slate or slate rubble. High levels of acidity did not favour the preservation of bone or pollen. The underlying geology is Ladock Beds or Grampound Grit (Soil Survey of England and Wales 1983).

At two locations within the proposed road corridor, the buried remains of sites of major archaeological significance were detected by magnetometer survey (Gater 1996).

At least 19 possible enclosures or rounds of assumed late prehistoric or Roman date lie within an approximate 3 kilometre radius of Killigrew (approximately one for every 1.5 square kilometres) (Figure 3.2). The majority are cropmarks which have been plotted from aerial photographs by the National Mapping Programme. Only two survive as substantial upstanding earthworks: Tregear, immediately east of Trispen (PRN22364), and the hillfort at Bishop’s Wood, St Allen (PRN25196) to the south of Trispen and 3–4 kilometres south of Killigrew. A handful of cropmark enclosures lie within the immediate vicinity of Killigrew, all in St Erme parish: two to the west on Killigrew Farm (PRN32027) and Ennis Farm (PRN55289) and another to the north at Chyton (PRN55292) (Figure 3.2).

The first was the remains of a sub-circular, doubleditched enclosure which lay to the east of Killigrew Farm (NGR SW 8466 5234, Figure 3.2). At approximately 50m by 40m in extent, this had the form, size and character of a later prehistoric or Romano-British ‘round’: a classic site type for settlement in the far south west. A number of linear anomalies were also found, likely to represent the remains of a contemporary ditched field system (Figure 3.3). The second site (NGR SW 8465 5271) comprised a pair of perfectly circular ditches, interpreted as the probable remains of two Early Bronze Age round barrows. Following negotiations between the County Archaeologist and colleagues in Transportation and Estates, Cornwall County Council, the road scheme was redesigned so that the barrows, located within a valley bottom, were buried beneath an embanked roadway to ensure their long-term preservation in situ. The ditched enclosure at Killigrew, however, lay on a hillslope and here the design of the road necessitated its partial destruction. Available resources were therefore focused on the excavation of the enclosure within the footprint of the road corridor. This excavation took place over three weeks in December 1996.

Within the wider area cropmark enclosures have been recorded at Boswiddle (PRN55198), Landrine (PRN55226), and several in St Allen parish: St Allen (PRN55277), Tolcarne (PRN55279), Lanner (PRN25194) and Trevalso (PRN32057) (Figure 3.2). Three possibly associated enclosures have been recorded at Trehane Vean (PRN55152; PRN55153; PRN55154) (Figure 3.2). A possible site has also been identified from fieldname evidence at Trelassick (PRN22466), Ladock, an allegedly ‘defensive’ bank was recorded near Carland Cross (PRN32049), and it is also possible that St Erme churchyard (PRN25252) is located on the site of a late prehistoric enclosure. The place-names of Tregear (PRN22364) and Hay (PRN22403) (both in Ladock parish) suggest the existence of enclosures, although it is possible that these refer to nearby known rounds.

Location and setting The site at Killigrew (Historic Environment Record Primary Record Number 55297, henceforth PRN) is located on a north-facing slope, 8 kilometres north east of Truro, St Erme at 115m OD (Figures 3.1 and 3.2). It is situated within a tract of Anciently Enclosed Land, in which field patterns are characterized as medieval or earlier in origin (Cornwall County Council 1996). Farms in the surrounding landscape date to at least the Norman period. All the farms within the vicinity of the

All of these sites vary in form and size, none have been excavated so we do not know how many of these sites, which may be assumed to be settlements, were contemporary with Killigrew: the evidence does, however, suggest a well-populated landscape during the later prehistoric and Romano-British periods. This 65

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 3.1 Location map of Killigrew enclosure.

66

Chapter 3.1: Background to the Excavations

Figure 3.2 Killigrew enclosure and wider landscape. Later prehistoric enclosures (upstanding earthworks, place-names and cropmark enclosures). (Sources: HER and Cornwall and Isles of Scilly National Mapping Programme.)

67

Later Prehistoric Settlement in Cornwall and the Isles of Scilly busy picture has relevance for our understanding of the potentially varied characters and functions of enclosed settlement during this period and how a site like Killigrew may have developed through time within such a landscape.

Principal aims and excavation strategies Prior to its discovery the enclosure at Killigrew was unknown as no earthworks were obvious but the excellent snapshot provided by the geophysical survey (Figure 3.3) indicated considerable potential for wellpreserved sub-surface remains. Rounds typically have deep ditches likely to contain valuable dating information as well as providing insights into the overall character of settlement for the Romano-British period (for example, Penhale Round: Nowakowski 1993; 1998 and Nowakowski and Johns 2015; Johnston et al 1998–9). Given the fact that nearly half of the enclosure was to be destroyed by the road, the excavation presented a good opportunity to examine the enclosure ditches as well as interior features. Resources were, however, limited and by necessity, overall excavation strategies had to be both pragmatic and opportunistic. The excavation focused on a number of key objectives (Nowakowski 1996):

A large univallate hillfort lies some 4 kilometres to the south in Bishop’s Wood (PRN25196) and may have been a major central settlement at this period (Henderson 1914). Other earlier prehistoric monuments in the form of Bronze Age barrows dot the skyline to the north of Killigrew, on higher ground at Newlyn Downs and at Penglaze, and around Carland Cross to Hendra Wood in the east (Figure 3.2), giving the discovery of a Beaker vessel found during the excavations at Killigrew (see below) a broader landscape context. The geophysical survey In advance of road construction and as part of an environmental impact assessment, Geophysical Surveys of Bradford undertook magnetometer surveys of 15 areas along the length of the road (about 40 per cent of the entire route). The survey of Area F, which discovered the enclosure at Killigrew, was widened by an additional 20m beyond the confines of the road corridor, so that a more complete picture of the entire circuit could be produced (Figure 3.3).

• Recovery of evidence for the date, character, social and economic status of the site. • Recovery of diagnostic artefacts to date chronological and structural phases. • Recovery of palaeoenvironmental data from well-sealed contexts for analysis and scientific (radiocarbon) dating.

Results were promising Gater (1996) commented:

The topsoil overlying the enclosure was removed by machine under archaeological supervision and cleared down to the level of the subsoil. Subsequent excavation was carried out by hand.

‘a partial ring feature, approximately 30m [sic] in diameter [. . .] the very strong anomaly on the eastern side[,] may indicate a stone wall or a former boundary bank. Inside the ring, there are some very strong readings that suggest the presence of burnt material. The results suggest this feature is a round. Just outside the ring feature, there are two linear features which intersect at its north-east corner. On the southern extremity of the survey, there is another small length of ditch that appears to be curving outside the limits of the survey’.

The corridor for the proposed new road cut through the heart of the enclosure, leaving undisturbed portions of the site to both the east and west (Figure 3.3), with unexcavated portions of the site extending under the original A39 road and into farmland to the west. The area of excavation measured just over 60m north-south and varied between 20m and 28m in width. Approximately 40 per cent of the enclosure lay within the road corridor and around 75 per cent of the archaeological features found during the excavation were investigated. Most major interior features were investigated through sampling and the enclosure ditches were sectioned in a number of places, although this amounted to less than ten per cent of the entire circuits.

The survey also showed that the enclosure had two ditches. A series of very strong parallel linear anomalies were also detected within the northern sector of the survey area, approximately 40–60m north of the enclosure (Figure 3.3). These were not investigated; one was the ditchline of a former field boundary recorded on the 1st edition Ordnance Survey 25 in: 1 mile map (circa 1880).

68

Chapter 3.1: Background to the Excavations

Figure 3.3 Killigrew enclosure. Area of geophysical survey and extent of excavation. (Source: GSB Geophysical Survey, Gater 1996.)

69

Chapter 3.2: The Excavation Results

Principal discoveries and main phases

period, represents its final phase. In the medieval period small farming hamlets grew up in the vicinity, and the remains of now vanished ditched field hedges dating to the historic periods were detected by the geophysical survey within the road corridor. None of these features were investigated.

The excavation at Killigrew round discovered features dating to a number of different periods, which revealed a previously unsuspected time-depth. The main features of each chronological phase are summarized below.

The site

Phase 1 Beaker activity, third millennium BC

Stratigraphic summaries

A small sealed pit contained a deposit of Beaker pottery, indicating that there had been prehistoric activities at this location approximately 4000 years ago. The discoveries by geophysical survey of two potentially Early Bronze Age sites (ring-ditched barrows) to the north (Gater 1996 and see above) suggest that the Beaker pit was not an isolated discovery.

The following stratigraphic summaries are presented in the main chronological phases outlined above. Principal features are described and discussed in terms of their significance for each major phase. Overall interpretations are presented in the concluding overall discussion. Phase 1 Beaker activity

Phase 2 Late Iron Age activity, first to second centuries BC

A very shallow and irregularly cut circular pit [113], 0.5m in diameter and 0.1–0.12m deep, was discovered just outside the enclosure to the south (Figure 3.4). This was filled by a single deposit of friable light brown silty clay intermixed with shillet (derived from the parent bedrock). It contained 46 sherds of a single distinctive Beaker vessel. The (incomplete) broken pot had been deposited in one event and then the pit was sealed.

The next major phase is represented by a metal-working furnace, which has been scientifically dated to the first to second centuries BC. As found the furnace lay within the enclosure, but there was no conclusive evidence that the furnace and the enclosure were contemporary (below). Phase 3 Romano-British activity, second to fourth centuries AD

This is the only feature which revealed earlier prehistoric activity at Killigrew, although a number of potentially related irregular hollows and pits ([120], [122], [124], [126], [128] and [130]) were located to the west nearby (Figure 3.4). None, however, produced any datable material. All these features were relatively shallow and filled with homogenous light brown silty clays with shillet inclusions. There were no stratigraphic or clear structural associations and their relative shallowness may have been due to truncation through ploughing. If they were related and all date to end of the third millennium BC, then their survival is remarkable, given that the surrounding landscape has probably been regularly ploughed since at least the medieval period. It is therefore possible that these shallow features represent ritualized activities spread across a much larger area but about which we know nothing. Very little flint was found during the excavation.

The ceramic assemblage shows that by the RomanoBritish period, all the principal structural features at Killigrew had been established. These comprised double ditches (once accompanied by banks) which made up the enclosure and within lay walled hollows, stone-lined drains and evidence for industrial activity in the form of working hollows and burnt pits. No domestic buildings were found, which is unusual for an enclosure of this period. This suggests that Killigrew was not a settlement but a place in the landscape with a specialized role. Phase 4 Post-Roman demise and landscape By the fourth century AD, the enclosure’s demise and general abandonment towards the end of the Roman 70

Chapter 3.2: The Excavation Results

Figure 3.4 Killigrew enclosure excavation area showing features of all phases.

71

Later Prehistoric Settlement in Cornwall and the Isles of Scilly (93), 0.3m deep, which contained substantial lumps of burnt wood, many greater than 100mm in size. Probable remnant spent fuel from the last firing. This deposit was bulk-sampled and when analysed was found also to contain mixed plant macrofossils and included oats, barley and wheat as well as a few weeds and tubers (Straker and Lamb, below). A charcoal sample identified as Salicaceae Salix/populus from this context produced an AMS determination of 2140 ± 60 BP (Wk-8333), which calibrates to 366–41 cal BC (Table 3.1). In the northern pit a correspondingly thick compact deposit (85) overlay [115] across the floor. This had less dense charcoal and contained 40 per cent dark / yellowish brown silty clay mixed with shillet. Macrofossil plant remains identified as Ulex/Cytisus (Table 3.1) from this layer gave a radiocarbon (AMS) measurement of 2040 ± 70 BP (Wk-8334), which calibrates to 208 cal BC to cal AD 123 (Table 3.1). The only finds within the furnace were burnt granite ‘pounders’ and vitrified clay lining from contexts (80), (83), (84) and (93).

Phase 2 Late Iron Age metalworking Furnace area [82] Lying within the southern (upslope) part of the enclosure was the reasonably well-preserved remains of a purpose-built furnace [82]: this produced direct evidence for metalworking (Figure 3.4). It had been cut into the hillslope in an area marked by amorphous spreads of redeposited clay and shillet, (135), (162) and (166) (Figures 3.5 and 3.6), which contained much charcoal and burnt stone. It is possible that these spreads were derived from earlier industrial activities which predated the furnace structure. The furnace [82] comprised two large linked oval pits which formed an hour-glass (figure of eight) shaped structure which was aligned north-south (Figure 3.5). The furnace had been dug into the sloping hillside and the two pits were set on different levels: the northern one was 0.3m lower than its southern counterpart (Figures 3.6, 3.21 and 3.22). The southern pit was 2m long and the northern one 1.7m and both, approximately just over 1m wide. The upslope (southern) pit was 0.60m deep and its downslope (northern) counterpart was approximately 0.5m deep. Both pits were interdependant and it may be assumed that the upper pit (to the south) was the principal firing area, and the lower pit (to the north) at the front, was the cleaning out area. Pressure of time meant that only 50 per cent of the structure was fully excavated. Our interpretation of this key feature has been based on analysis of a relatively small sample of the furnace deposits (15 tenlitre samples of 78 samples (780 litres) were recovered). Samples for future detailed work are retained in the archive. There was clear evidence for several repeated episodes of industrial metalworking in this area which resulted in a complex stratigraphy. The furnace appears to have been located in an area with a long sequence of metalworking and this was confirmed by the scientific dating (below).

Exactly how these two features operated is uncertain. The working or firing end may have been upslope (to the south) with the lower pit, to the north and downslope, a receptacle for material raked out of the actual furnace. An alternative interpretation is that the pits represent two different features which are adjacent and were used in a similar fashion but perhaps not at the same time. Although the radiocarbon dates from the two areas overlap, suggest intermixing and likely contemporaneity. The detritus of many firings taking place immediately upslope would have resulted in accumulated mixed material. (We are very grateful to Adam Sharpe for his insight and comments here). After the furnace had ceased to function it was infilled by dumps (83) and (84), which contained shillet and burnt stone, charcoal and vitrified clay; this was all secondary, redeposited, material. A number of spreads associated with the furnace were recorded, and these extended to the west. Directly beneath the overlying topsoil and over the immediate area of the furnace was a stony layer (67) (Figure 3.5) which contained charcoal and other metalworking waste, very similar to the top fill of the furnace (83). A small sondage was excavated through the approximate centre of this deposit and three sealing layers, (116), (117) and (118), were recorded. Deposits (116) and (118) were lenses of charcoal, while (117) was a dump of redeposited shillet. Although not fully excavated, (116) contained 1.51kg of metal slag with fragments of furnace lining. This material overlay the furnace, and therefore post-dated it: it may well have been redeposited material dumped from perhaps another unidentified furnace area close by.

Along the entire length of the southern (upslope) pit was a narrow gully, 0.2m wide by 0.2m deep and 1.8m long (Figure 3.5). This was possibly an under-flue. A series of profiles cut across the southern, central and northern parts of the furnace revealed its well-cut vertical (eastern) edges (Figure 3.22). Exposed bedrock immediately around the furnace had been scorched a deep red but it was only at the top end of the southern section of the feature that fragments of furnace lining were recorded in situ. The two related parts of the furnace contained very distinctive deposits. At the base of the northern pit there was a dense but thin charcoal layer containing much burnt stone [115]. Filling the base of the southern (upper) pit was an extremely dense charcoal deposit

Evidence of metalworking from within and around the furnace included a number of fragments of slag in 72

Chapter 3.2: The Excavation Results

Figure 3.5 Furnace [82] land related metalworking activities within the enclosure. Late Iron Age site reused in the Romano-British period. For location see Figure 3.4 and section B-A in Figure 3.6.

contexts (93) and (116), and possible metal objects and fragments of clay lining of the furnace in (67) and (116). Clay lining was not, however, extensively recorded in situ. X-Ray Fluorescence (XRF) analysis of a sample of the lining carried out by Albertine Malham and Gerry McDonnell of Bradford University, demonstrated iron smelting had taken place within the clay-lined furnace, an interpretation reinforced by the recovery of a small amount of iron ore on the site (below).

Phase 3 Romano-British period Killigrew was defined by a double-ditched circuit which is a relatively rare feature of excavated enclosures dating to the Romano-British period (Figure 3.4) (cf A Young 2012). It is not known if the ditches were constructed during the later Iron Age when the ironworking furnace was operating (above) or during the Romano-British period. Their appearance does, however, show how some specialized places emerged in the landscape in later prehistory and were made distinctive through acts of enclosure (cf Nowakowski and Quinnell 2011). Whether furnace [82] was originally set within an enclosure cannot be resolved.

It is clear that activity continued around the furnace area after the structure itself had fallen out of use. A circular pit [161] was dug into the upper dump layer (83) (Figure 3.5). This contained a single fill of burnt stone, charcoal flecks and pieces of slag; and was overlain by a revetted stone edge 164 which abutted a spread of charcoal (163). Two other pits, [136] and [160], were located immediately to the west of the main furnace area. These pits, circular and with flat bases, may be contemporary and could represent reuse of the area during the Roman period but were not fully excavated.

The outer ditch circuit The outer ditch circuit was sectioned and fully excavated in three places: one segment to the north, [20], and two in the south, [29] and [31] (Figures 3.7 and 3.8). The profile recorded in the north [20] was over 2.5m wide, although later ploughing had reduced the depth of the feature to only 0.6m. At this downslope location, ditch [20] had a stepped base (a feature not recorded elsewhere) and this may have represented a minor

Two sherds of Romano-British pottery were recovered from the surface of layer (67) but a general absence of pottery of Romano-British date from all these features was noted. 73

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 3.6 East – facing section across furnace [82].

recut (Figure 3.7). It contained four fills, each with considerable amounts of shillet. Ditch fills (22) and (23) were slight deposits at the base of the ditch and both were sealed by the main fill, a brown-yellow clay loam (24) which also contained a small amount of charcoal. This was sealed under a red-brown silty clay (21).

and 1.1m deep with a generally wide mouth and uneven profile (Figure 3.10). It contained three deposits, (37), (38) and (39). The top fill (37) was a dark brown silty clay while the middle fill (38) was a light yellow-brown clay loam; both deposits contained a considerable amount of charcoal which, in the case of (38), clustered at the base. The basal fill, (39), was similar in make-up to (38) immediately above it but contained a significant amount of shillet. Upper fill (37) produced 39 sherds of pottery, largely of the third century AD, and a greisen steelyard balance weight (Figure 3.19; S1).

Following the removal of topsoil a large quantity of pottery was visible in the top fill (112) of the outer ditch to the east of [20]. Here 163 sherds were recovered and these represent part of a thick-walled and decorated storage jar (P3) dated to the third century AD (Figure 3.15).

Two further segments of the inner ditch circuit were excavated, [25] (not illustrated) and [27] (Figure 3.9). These measured 1.9m and 2.6m wide respectively. Both were up to 0.9m deep. Ditch [25] contained one fill of a yellow-brown clay loam, with some dense pockets of shillet, and two sherds of pottery. The northern face of ditch [27] was almost vertical and its southern face a gentle slope. Here the ditch was filled by two yellow-grey brown clay loam layers, (41) and (28), with a compact orange brown silty clay (42) at its base. No finds came from this part of the inner ditch circuit.

Upslope to the south, two segments of the outer ditch, [29] (not illustrated) and [31], were excavated and here both ditch profiles were U-shaped. Their widths varied between 1.7m and 2m and they were 0.85m and 0.8m deep respectively. Each segment contained single fills of yellow-brown clay loams, with some evidence (tip lines) for dumping (Figure 3.8). The upper fill (32) of ditch [31] contained four sherds of Romano-British pottery which are not closely datable. Although this general area was clearly ploughed in the medieval period and buried remains were partially truncated, it is likely that the depths of the outer ditches were not originally that much deeper as excavated.

Summary of ditches The inner and outer ditches at Killigrew contained relatively homogenous dumped soils. No datable finds were found in the basal fills. It is likely that during the main phases of activity within the enclosure, the ditches were maintained and cleaned out on a regular basis. If so, where the ditch upcast was removed to is unknown. Single-episode and dumped fills recorded in sections of both the inner and outer ditches does suggest that some deliberate backfilling took place, an idea reinforced by the amount of dumped pottery recovered from the upper fills of the ditches. This activity can be broadly dated by pottery to the third century AD and provides a

The inner ditch circuit The inner ditch circuit was also investigated in three places: two segments, [25] and [36], were excavated downslope to the north, with a further segment, [27], investigated upslope to the south (Figures 3.9 and 3.10). Again, what appears to have been the more substantial ditch was to the north. Here ditch [36] was 2.4m wide 74

Chapter 3.2: The Excavation Results

Figure 3.7 East-facing section of outer downslope ditch [20].

Figure 3.8 West-facing section of outer upslope ditch [31].

Figure 3.9 West-Facing section of inner upslope ditch [27].

likely date for the final infilling episodes and end-use of the site. No slag or metallurgical debris was recovered from the ditch sections investigated.

the inner, averaged just over 1m deep and the Killigrew ditches are comparatively shallow when compared to the more substantial ditches associated with other excavated rounds, where ditches are generally 2m (or more) wide, and more than 1.5m deep; examples include Penhale Round (Nowakowski and Johns 2015),

The varied profiles of the ditches at Killigrew appear to reflect the natural slope. The deepest ditch circuit, 75

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 3.10 East-facing section of inner downslope ditch [36].

Trevisker 2 (ApSimon and Greenfield 1972), Trethurgy (Quinnell 2004) and Castle Gotha (Saunders and Harris 1982).

Hollow 17 was broad and sub-circular, measuring at least 16m (east-west) and 9m (north-south) in extent. It had been dug as a sunken feature into the shillet bedrock. Along the base of its southern side was a slight depression which was 1m wide and 0.1m deep. Here a section of a stone revetment, which formed a partly tumbled wall face, 9, was exposed. A fragment from broken rotary quern S3 was recovered from wall fabric. Wall 9 was part of a longer curvilinear wall line other segments of which were found to the south (see below).

There was also no definitive evidence of accompanying banks or ramparts at Killigrew, or indeed evidence of where ditch upcast had been deposited (see above). A space up to 3m wide separated the inner and outer ditches (both upslope and downslope). This feature was also found at Threemilestone (Schwieso 1976). Evidence for a wide inner bank was found alongside the edge of the deeper inner ditch at Threemilestone but there were no traces of a bank between the inner and outer ditches (Schwieso 1976, 52–54, fig 20). At Threemilestone it was suggested that the outer ditch acted as a stock barrier to prevent cattle falling into the open deeper inner ditch (Schwieso 1976, 65).

The earliest layers lining the base of the hollow comprised a gritty brown silt (14) and a localized dump of redeposited natural shillet (43) (within the depression along its southern edge). These deposits were overlain in the southern part of the hollow by a brown-yellow clay and shillet spread (13).

The absence of surviving traces of banks or ramparts at Killigrew may be largely due to the siting of the enclosure on a pronounced hillslope and subsequent gradual attrition through agricultural cultivation in the medieval and post-medieval periods.

In the centre of the hollow was a sub-rectangular pit [33], which had been dug through the lower deposits (13) and (14) and into the natural bedrock. Measuring 1.3m by 0.9m, and aligned east-west, the pit was 0.3m deep with sheer sides and a flat base. There was considerable discolouration (reddening) of the natural stone in and around the pit, indicating that it had been the site of open fires. The pit contained three fills. The bottom fill, (70), was a thin lens of charcoal and the overlying fills (34) and (35) also contained some charcoal. The top fill (34) contained six sherds of Romano-British pottery. This feature had the characteristics of a typical Romano-British hearth, similar to those found at Trethurgy, St Austell (Quinnell 2004).

Working areas Hollow 17 Within the north-west zone of the excavation and contained by the northern inner ditch of the larger enclosure was a hollow 17 (Figure 3.4). This presented some of the most complex stratigraphy on the site and was identified as a major anomaly by the geophysical survey (Figure 3.3). During excavation the extent of the hollow was defined but it was only partly excavated, although sufficient information was obtained to determine its shape, form and date. A trench was cut across the hollow by machine so that a section could be recorded.

In the hollow there were also three postholes (not illustrated). Posthole [15] was positioned to the west of pit [33] and also cut through clay layer (13). Postholes [16] and [69] were recorded at the lowest point of the hollow dug straight into the natural bedrock, although it was not clear whether these features were earlier in 76

Chapter 3.2: The Excavation Results date as their relationship with the overlying fills within the hollow was difficult to ascertain. These can only be interpreted as random discoveries and it is not possible to determine whether they had once formed a wooden structure set within the hollow (see below).

features recorded to the north west. But like hollow 17 it too was edged by a stone revetment 103 on the south and contained a small number of features which suggests that it too had functioned as a working area. The southern edge of the hollow was revetted, with a section of stone walling 103 exposed and built up and set against a cut in the natural slope. Here the revetment was 0.8m wide and 0.7m high. 103 formed part of an irregular wall line, sections of which were exposed as stone revetted faces recorded in various states of preservation as 73, 110 and 45 to the west of the hollow (Figure 3.4 and see below). This wall line was moulded into and against the natural slope but appeared to contain the area upslope where the furnace [82] was located.

Over clay shillet layer (13) was an extensive grey-brown ashy silt clay deposit (12), which appeared to infill much of the hollow. Along the southern edge this layer was in part covered by a quite extensive dark grey brown clay loam which contained charcoal (10). Pottery was found in contexts (revetment) 9, (10), (12) and (13). The earliest layer, (13), contained P9 (Figure 3.16), a Type 21 bowl, probably of third century AD date, while layers (10) and (12) both contained pottery of probable third to fourth century AD date. Sherds recorded from within and around the stone revetment 9 were not datable.

This much larger hollow contained a stone spread in its centre (52) (Figure 3.4), which covered an area 2m by 3m, and three further features were recorded to the south east. The largest of these was the remains of a possible open hearth [104] made of two large flat stones surrounded by an area of burnt clay (0.7m by 0.4m). The other two features in this area were an isolated small shallow posthole [105] (0.1m in diameter and 0.1m deep), and a triangular arrangement of flat shillet stones (107) which measured 0.65m by 0.8m.

Overlying the top of revetment 9 were traces of another gritty brown silt clay loam (68). Only a small amount of this deposit was visible because hollow 17, once filled, had been cut into by a later hollow [1]. This later feature was filled by two deposits, a grey-brown silt clay loam (2) and a brown silt clay loam (3). Layer (3) contained only five sherds of pottery but deposit (2) contained 85 sherds, all broadly dated to the third to fourth centuries AD.

The most recent deposit within the central hollow was a dark topsoil (55). To the west this extensive layer sealed the ‘orangey’ layers (53), (54) and (108), which partially overlay another revetted segment of stone walling 73, part of the same irregular wall line as 103 to the south and 45 and [110] further to the north west (Figure 3.4). A number of pot sherds were recovered from this general area: 13 sherds from (53), 11 from (108) and 32 from layer (55). Dating of these, where possible, indicates a date in the third to fourth centuries AD, again apparently contemporary with the infilling of the enclosure ditches.

This evidence suggests that the demise of hollow 17 and the later cut [1] was broadly contemporary with the backfilling of the ditches, also dated to the third century from pottery in the upper fills. It is difficult to ascertain with certainty exactly what went on in hollow 17. The structural evidence is patchy and confusing and its overall functions difficult to pin down. It is, however, possible that it had served, as suggested by copper residue on a sherd (see Roman period pottery [34]), an unroofed working hollow, and was perhaps the focus for a whole range of minor workshop activities somehow related to industrial processes carried out within the wider enclosure. Alternatively, given the dominance of sherds of jars and bowls from this area of the site, it could be that food was being warmed here (rather than prepared or cooked) as most of the pieces from 17 were fragments of jars and bowls all used for serving food (see Quinnell, below).

Hollow 17 and this other major hollow appear to be demarcated zones, operating as open settings for activities related to the main business within the enclosure. Clearly contemporary they were fashioned into the natural slope and the network of revetted walls appeared to separate activities from the upper slope within the wider enclosure from those lower down. A network of stone revetments

Romano-British Hollow in the centre of the enclosure

Segments of low stone walls in various condition were recorded across the centre of the enclosure. These were all sections of stone revetments (of varied build but principally shillet and other stone) which had been built up and against the natural slope. Some segments were more intact than others where stone tumble showed collapse and neglect. Their overall arrangement which

Another large hollow (labelled ‘Romano-British Hollow’ on Figure 3.4) lay within the central area of the excavation and this had also been detected as a large amorphous anomaly by the geophysical survey (Figure 3.3). It too had been dug into the natural slope and was more amorphous in definition and blended into 77

Later Prehistoric Settlement in Cornwall and the Isles of Scilly divided up areas of the round interior conveys the intention to contain and demarcate activities as well as separating the upper and lower parts of the round interior. Their overall arrangement formed at least three curvilinear stretches of walling which resulted in a twisted network across the central part of the round interior and which converged as a mass of tumbled stone in the centre of the excavation at 45 and 110 (Figure 3.4).

The stronger feature was ditch [58], aligned north east – south west. Up to 5m of this ditch was visible and its western terminal was excavated and recorded as 1m deep. The ditch had steep, near-vertical sides which narrowed to a rounded base where it was 0.35m wide (Figure 3.12). It contained four fills: the upper deposit (59) contained granite and shillet pieces deliberately laid. This upper fill with packed stone was embedded into compact lower fills – (80), (60) and (81) – which produced three sherds of pottery dating to the second century AD. Stones in the upper fill looked like stone packing in situ and it is possible that this deep feature was a ‘foundation trench’ for a wooden fence or palisade. If so, it is likely to have been a major structural fixture in this part of the site. Iron slag was also recovered from (59).

In the north west one wall line comprised wall 9 and wall 57 which shared a southern curvilinear trajectory with a segment recorded as 51. In the north, wall 57 was built up in a cut which measured 0.4m wide and 0.1m deep and contained two sherds of undiagnostic Romano-British pottery. Walling 9 as discussed above was recorded as the southern edge of hollow 17. Wall 51 in the south was made up of shillet pieces which had been laid in foundation cut [50], 0.5m wide and 0.25m deep. Wall 51 converged with segments of walling recorded as 45 and 110 to the south (see below).

Immediately to the north of ditch [58], was a slighter parallel ditch [62]. It was 0.6m wide and only 0.1m deep; it was irregularly cut but contained a number of medium-sized shillet pieces. Also located in this area were two probable post-pads, [94] and [96], and a posthole or pit [98] (not illustrated). Post-pad [94] lay to the south of ditch [58] while post-pad [96] lay between the two ditches. Both post-pads were cut into the natural bedrock but no more than 0.12m deep. Both contained packing stones. Positioned to the immediate west of ditch [62] was isolated pit [98], 0.6m by 0.5m and very shallow at only 0.1m deep. Evidence of post-pads could even suggest the former presence of wooden buildings in this area of the interior.

So, this long stretch of a curvilinear wall encircled a raised area upslope to the west and extended from the southern edge of hollow 17 (see above). Another related but separate wall line was defined by stone wall 111 which lay to the south west of the junction of walls 45 and 110. This marked out the southern edge of the raised area to the west and would appear to have contained a zone characterized by clusters of pits where fires had once burnt (see below).

Many of the upper spreads and layers from this central area contained ceramics dating to the third-fourth centuries AD. A stratified second century AD pottery sherd was recovered from ditch [58], suggesting earlier activities within this general area.

A considerable jumble of stone was visible in the general area where the four walls 51, 45, 110 and 111 converged. Leading into this was a stone-lined drain [149] on an east-west alignment. This feature was constructed of granite pieces, with vertical sides and a single capping stone. The inner channel of the drain was 0.4m wide and 0.3m deep. This well-made feature was unusual in that it was made of granite, a non-local stone, and appeared to be sited in an (unroofed) open space. The construction of a well-made drain hints at the need for water management in activities related to industrial working. Stone-lined drains in open (and roofed) spaces have been recorded at the enclosed settlement at Goldherring, Sancreed, alongside outdoor hearths which had clear evidence for iron smelting in the late Roman and post-Roman periods (Guthrie 1969, 13–18, 24–25).

Romano-British industrial activities Industrial pits The ground surface to the south and west of the central hollow was raised, perhaps from upcast dug out of the hollow (Figure 3.4). In the west and contained within the curve of stone revetments (see above), there were clusters of pits with evidence of industrial activities involving the use of fire. Fragments of slag and pieces of iron were also found displaced into contexts 45, ditch fill (59) and the dark clay loam layer (63), and all point to evidence of small-scale smelting or secondary working such as iron smithing.

Ditches [58] and [62] – a possible interior palisade Beyond the Romano-British hollow 5m to the north the traces of two narrow ditches [58] and [62] were recorded (Figure 3.4).

Within this area, a group of three related features were the earliest. These were a large sub-oval pit [137] which was 1.6m long and 0.22m deep, a sub-rectangular pit [139] with vertical sides and a flat bottom filled with charcoal, and a sub-oval pit [146] (0.8m long, 0.4m wide 78

Chapter 3.2: The Excavation Results

Figure 3.11 Section across pits [4] and [7] – location of the tin dish.

Figure 3.12 East-facing section across palisade ditch [58].

and 0.1m deep). All three pits contained fragments of wood charcoal and their sides were heavily scorched, which suggests that they had all once contained open fires. They were also cut into by later pits. Pits [137] and [139] were later replaced by a sub-oval pit [141] measuring 0.7m by 0.5m which contained wood charcoal. Pit [141] and earlier pit [146] were also cut by later pit [143], which was oval, 1.8m by 1m, and 0.25m deep. It was U-shaped in profile and was cut into bedrock which had reddened due to intense heat. It contained two fills with numerous fragments of wood charcoal.

pits or postholes – [46], [48], [154], [156] and [157] – and a further squarish area of burning (158).

Pottery from these related industrial features date the pits to the Romano-British sequence on the site. Pit [137] contained 18 pottery sherds, and pits [143], [149], and [154] also contained pottery; much of the pottery is undiagnostic although storage jar sherds were identified and date to the third century AD. To the east of this busy complex there were a number of very small

Both pits are likely to be related in function and may represent further evidence of small-scale industrial activities centred on metalworking, although the overall quantity of slag from the site is fairly small. Unfortunately, the samples were not scanned for hammerscale, (a sure indication of secondary metalworking) and close dating of these features is

A pair of small, shallow pits [74] and [77] lay to the north east of hollow 17 (Figure 3.4). Both contained evidence for burning. Pit [74] measured 0.70m by 0.45m and was 0.05m deep. It contained two distinct fills; the lower (76) was a charcoal layer which lined the pit base. Its companion, pit [77], 1m to the south, was larger at 1m in diameter. This also contained a reddish clay deposit (78), further evidence of burning. There was no direct stratigraphic relationship between the two pits.

79

Later Prehistoric Settlement in Cornwall and the Isles of Scilly impossible. They are, however, a further indication of the general poorly defined spread of activities which characterized the overall interior of Killigrew enclosure.

0.1m deep. Near these linear cuts was a shallow flatbottomed pit [88]: 0.85m by 0.60m and 0.1m deep. All of these features contained only single fills and there is no direct stratigraphic relationship between any of these features and an association is assumed. In form and shape they share some similarities to postRoman ironworking pits excavated at Gwithian in the 1950s (Nowakowski et al 2007, 40–44; fig 12).

A tin dish deposit – pits [4] and [7] Immediately to the south of hollow 17, a solitary pit [4] was identified (Figures 3.4, 3.11 and 3.23). Subcircular and flat-bottomed, 1.4m in diameter and 0.22m deep, it contained two deposits of brown silty clay with charcoal fragments. Cut into its base was a smaller pit [7], and this contained half a tin dish (Figure 3.18).

Phase 4 The post-Roman and medieval landscape There is no evidence for any later occupation of the site following its demise around the third to fourth centuries AD. During the subsequent post-Roman and medieval periods upstanding earthworks and remains appear to have been left to gradual attrition and decay. The excellent preservation and survival of many of the interior features as well as the ditch circuits defining the enclosure suggests that, to some degree, subsequent immediate land use in this area did not heavily impact on the abandoned site. It is clear, however, that eventually parts of the enclosure were eroded with the possible levelling of banks and truncation of the upper mouths of the ditches as a result of arable farming. This may have intensified during the later and postmedieval periods with no above-ground remains of the enclosure surviving into the modern period.

These pits had no direct stratigraphic relationship with the main hollow 17 but pottery found in the lower pit with the tin dish dates to the second to early third centuries AD. A handful of sherds from the fill of the upper pit [4] are earlier dating to the late first or second centuries AD. These may be residual. We can only speculate as to the reason why the dish was deposited in the pit and that only part of a dish was placed within the feature is of great interest and may indicate some form of structured deposition (see below). Other features within the enclosure To the north and west of the two main hollows, just inside the inner ditch of the enclosure, there was a group of three very slightly cut features (Figure 3.4). Ditch [86] was shallow, elongated and aligned northsouth. It was 3.5m long and 0.6m wide, but only 0.15m deep. Ditch [90] was aligned east-west, and measured 5m long and 1m wide. It too was shallow at less than

A road was subsequently constructed through the enclosure and the two segments of the enclosure were split between two separate landholdings with origins in the medieval period; Chyton, first recorded in 1513, to the east, and Killigrew, documented in 1284, to the west.

80

Chapter 3.3: Radiocarbon Dating and Charcoal Identification

One of the principal objectives of the excavation was to obtain material to date the Killigrew enclosure and its history of use. Two Accelerator Mass Spectrometry (AMS) dates were obtained. Given the uniqueness and rarity of the furnace structure [82], it was decided that available resources would be spent on obtaining a date for this major feature. Suitable wood charcoal samples from the furnace were selected and prepared by Rowena Gale.

• 4 fragments gorse/broom (Ulex/Cytisus)*, diameter 5mm. • 6 fragments oak (Quercus) rounded and sapwood. • 3 fragments hazel (Corylus), roundwood. Radiocarbon dating: results of the dating programme Two samples for AMS dating were submitted to the University of Waikato Radiocarbon Dating Laboratory, New Zealand, in 2000. Both were from furnace [82] and the dates are shown in Table 3.1.

Charcoal identification Rowena Gale

Lab no

Two charcoal samples were examined and identified to genus level to provide short-lived material for radiocarbon dating. Both samples were from fills in furnace [82].

Context

WK8333 Layer (93)

The material was prepared for examination by fracturing to expose fresh transverse, tangential and radial surfaces. These fragments were supported in sand and examined using a Nikon Labophot incident light microscope at magnifications of up to x400. The anatomical structure was matched to reference material. The results are shown below. Samples indicated with * were selected and submitted for dating.

WK8334 (85)

material Willow/ popular (Salix/ Populus) Gorse/ broom (Ulex/ Cytisus)

Age BP years Ỏ 13C 2140 ± 60 -26.0 ± 0.2

2040 ± 70 -24.5 ± 0.2

Calibrated at 95%

366 BC–41 BC

208 BC–123 AD

Table 3.1: AMS dates from furnace [82].

Context (93)

Dating the furnace

A large sample, predominantly composed of narrow roundwood. A small proportion only of the sample was examined to allow sufficient material for dating using a single genus.

The two radiocarbon dates from furnace [82] were earlier than the diagnostic pottery for the main phase of the enclosure. The determinations overlap, indicating probable later Iron Age activity, thereby suggesting an early date for metalworking on the site, although whether this took place within an enclosure is unknown. While the dates suggest pre-Roman metalworking activities no finds dating to the later first millennium BC were found in the areas excavated. This could indicate something about the way the site evolved – perhaps it was unenclosed initially and metalworking was a sporadic and infrequent activity. The reuse of the location during Romano-British times may have effectively obliterated much earlier traces. The question of whether the furnace was located within an enclosed space therefore remains unresolved as the only available dates for the enclosure ditches are from pottery found within upper fills.

• 8 fragments of willow/poplar (Salix/Populus)*, diameters approximately 5mm, 4-5 growth rings. • 7 fragments oak (Quercus), roundwood, diameter 5mm-6mm, 4-8 growth rings, plus a fragment of wider sapwood. • 3 fragments gorse/broom (Ulex/Cytisus), roundwood, diameter 5mm-15mm. • 2 fragments hazel (Corylus), diameter 8mm. Context (85) • A relatively small sample, from which the following were identified. 81

Chapter 3.4: The Artefacts A total of 708 finds was recovered during the excavation. The majority were ceramics (645 sherds) and of Romano-British date. Three pieces of flint, 13 pieces of stonework, 30 pieces of slag, some fragments of furnace clay lining and eight unidentified iron pieces made up the rest of the finds. Ceramics Henrietta Quinnell with petrology by Roger Taylor Beaker pottery The pottery fabric was visually inspected using a hand lens. Forty-six sherds weighing 145g from a single Beaker came from (114), the fill of pit [113] just south of the enclosure (Figure 3.4). Some of the breaks were fresh but less than 10 per cent of the vessel is represented. Sherds present include rim, neck, base angle and base and allow a possible vessel shape to be reconstructed (Figure 3.13). The fabric is hard, with very common inclusions up to 1.5mm; the exterior has been very well smoothed, almost burnished. Both surfaces have been oxidized reddish yellow 5YR 6/6. The reconstructed shape corresponds broadly to Clarke’s (1970) S2 type. The decoration consists of small, neat, paired fingernail impressions forming vertical lines up the vessel. Its capacity is somewhere between 1700 and 1900cc, falling within Case’s group of medium size (1995, 56).

Figure 3.13 Beaker vessel from pit [113].

Discussion Radiocarbon dating programmes have shown that Clarke’s typological sequence cannot be used as a guide to chronology (Kinnes et al 1991), although stratigraphic data suggests that vessels of S2 or longnecked (Needham 2005) shape are unlikely to belong to the beginning of the Beaker period. Work on Cornish Beakers by Jones (2005) and Jones and Quinnell (2006) indicates that most Beakers in Cornwall are likely to be late, perhaps occurring in the centuries either side of 2000 cal BC.

Beaker petrography Inclusions: quartz, abundant angular grains up to 1.5mm, 75-80 per cent; feldspar, white and translucent angular and cleaved grains up to 1mm, about 15 per cent; mica, muscovite flakes up to 0.5mm, less than 1 per cent, biotite, a few flakes seen; tourmaline, angular black grains and striated prismatic crystals, less than 1 per cent; rock fragments, a few granitic fragments and possible slate or micaceous hornfels fragments.

The comparatively local sourcing of the fabric is typical of Beakers in Cornwall (Parker Pearson 1990, 11), as opposed to the extensive use of gabbroic clays both in the Early Neolithic and in the Bronze Age with Trevisker material.

Comment: a granitic derived temper. The granitic mineral content is not in the proportions expected in a granite source rock. The presence of a non-granitic rock fragment indicates a source of temper, probably stream sand, outside the margin of the granite. The nearest granite is the St Austell outcrop, 10 kilometres to the north east, so the source of temper could have been rather closer to the site.

No precise comparanda for the Killigrew Beaker has been published from Cornwall, although the shape is present in vessels such as that from Trevedra (Clarke 1970, 477). Fingernail decoration of this type has been recognized among the group associated with a burnt mound at Boscaswell in West Penwith (Jones and Quinnell 2006). More irregular and coarse fingernail82

Chapter 3.4: The Artefacts impressed sherds are among the surface collection from field 24, Polcoverack, on the Lizard (Smith 1987, fig 20), and recent work on Beakers in Devon (Quinnell 2003) has demonstrated that this decorative mode was used quite widely in the county. It was noted by Clarke (1970, 214) that fingernail decoration was seldom used on vessels chosen for deposition in barrows.

is included in the report on Trethurgy (Quinnell 2004, 108). This report contains a discussion of the gabbroic fabrics used and a type series for Roman period gabbroic pottery with comments on dating (ibid, 5.6.3) which is used below. The primary intention of this report is to provide chronology for the activities at Killigrew. Fill (8) of pit [7] in association with tin dish

There has been a tendency in the past to term Beaker pottery not deposited in barrows as ‘domestic’ (Gibson 1982), and to regard it as of lesser importance than that which was. It is more appropriate to regard Beaker pottery as a broad group from which certain vessels, very much the minority of those made, were selected for barrow deposition. For Cornwall, figures compiled in 2007 (Jones and Quinnell 2006) indicated some 11 vessels deposited in barrows out of 60 known Beakers, and for Devon seven out of 55 (Quinnell 2003).

Seven sherds, including P1 (Figure 3.14), jar with grooved rim and tooled cordons beneath; Trethurgy Type 11, second and early third centuries; fragments from another similar jar. Fill (6) of cut [4] over pit [7] Three sherds, including P2 (Figure 3.14), bowl with grooved rim and tooled cordons beneath; Trethurgy Type 19, St Mawgan-in-Pydar Type R (Threipland 1956), late first and second centuries.

During the currency of Beaker pottery it is still unclear how domestic sites are represented in the archaeological record (Thomas 1999, passim) and therefore how ‘domestic’ Beakers may be recognized. Thomas (1999) has highlighted a tradition of pit deposition for the Neolithic and Beaker periods in which the pits are of distinctive shape. The pits are open-topped, shallow and rapidly infilled, and appear to have been dug for the burial of selected deposits which can include pottery, animal bones and burnt material. Some may have been situated on sites which were lived on for a while, others could have had their own significance as locales for meetings or ceremonies.

Context (32) in fill of outer ditch on south side Four sherds, including girth with incised line from large Type 4 jar; not closely dateable. Context (112) top of outer ditch on north side

The Killigrew pit fits well into this tradition as do the other pits in Cornwall with Beaker pottery, including those at Bosmaugan, near Lostwithiel (Cole 1999), Porthleven and Nancemere (Sections 4 and 6). In Devon there are now five sites with pits (Quinnell 2003) of which the best published example is the group at Mare Lane, Beer (Tingle 1998, 67–80). The increasing recognition of sub-surface features such as pits, and appropriate study of their contents, will enable a pattern of activity during the period of Beaker use to be developed into which finds from barrows can be fitted, rather than one in which barrows are perceived to dominate. In the present state of studies, good publication of pit groups and other non-barrow contexts may be considered at least as important as material from barrows.

P3 (Figure 3.15), 163 sherds from a thick-walled coarse gabbroic storage jar with frilled ‘pie-crust’ rim and criss-cross incised pattern on girth. Trethurgy Type 16. Storage jars form a continuous tradition in Cornwall from large versions of Iron Age Cordoned Ware through to the fifth – sixth centuries. Dating of vessels such as P3 is heavily reliant on the currency of storage jar forms in Devon and Somerset; South-Western Greyware storage jars were in use in Exeter from the later second century, although the main period of their currency is the third and fourth centuries (Holbrook and Bidwell 1991, 175). Independent dating for the Cornish series is scant. The earliest occurrence at Trethurgy is in period 3, tentatively dated to the mid-third century (Quinnell 2004, 120), a date with which the South-Western Grey storage jar in the ditch deposit at Kilhallon is in accord (Carlyon 1982, 161). Currency in Cornwall of decorated examples such as P3 is probably mostly safely accepted as third and fourth centuries with the possibility that the date may extend back into the later second century.

Roman period pottery

Context (37), upper fill of inner ditch on north side

Some 600 sherds of Roman date were recovered. Restricted resources limit the present report to descriptions of distinctive pieces and to brief summary statements of the material from each context. All are of standard gabbroic fabric unless otherwise stated. A petrological description of this fabric by Dr D F Williams

Thirty-seven sherds including P4–P7 (Figure 3.14). P4 two sherds from Type 16 coarse gabbroic thick storage jar, rim top with slashes (P4a), girth with incised grooves (P4b); comments as for P3.

83

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Figure 3.14 Roman period gabbroic pottery: P1 [8], P2 [6], P4–7 [37] and P8 [58].

84

Chapter 3.4: The Artefacts

Figure 3.15 Coarse gabbroic storage jar P3 [112].

P5 dish with curved wall and simple, out-turned, flattopped rim, Type 20, late second to third century date.

Context (34) above (13) in sequence infilling hollow 17 Five sherds, including a chunk of coarse gabbroic storage jar and a piece of South East Dorset blackburnished base. Third century or later.

P6 part of small necked jar with globular body and girth groove, Trethurgy Type 3 continuing from St Mawganin-Pydar Type N (Threipland 1956), where they are common; these belong to the earlier Roman centuries with a date after the second century unlikely.

Significant levels of copper were detected on the blackened interior surface of a sherd from this context by XRF analysis in 1997 (David Starley, pers comm).

P7 (not illus) scrap of samian, very worn, almost certainly from the ovolo of Dr 37 bowl, Central Gaulish, with one edge worn smooth as though to produce powder; it is generally accepted that samian had a long currency in the West Country and small sherds may have provided medicines or colourants (for example, Cadbury Congresbury: Rahtz et al 1992, 150 and references). The chronology of context (37) is difficult to establish in view of jar P6, but a date in the third century with P7 regarded as a survival would accommodate the problems.

Context (12) above (34) in sequence infilling hollow 17 Thirty-five sherds, including 20 of coarse gabbroic storage jar P10 (Figure 3.16), of which the rim has semicircular stamped decoration; see comments on P3, no parallels known for form of stamp. P11 (Figure 3.16) everted rim from simple slackprofiled Type 4 jar, long date range from second to fifth centuries. P12 (Figure 3.16) bowl with out-turned rim and deep groove along top, Type 21 third or fourth centuries and another similar vessel.

Contexts (59), (60) and (80) in the fill of cut [58] Three sherds, including small jar P8 (Figure 3.14; comments as P6) and probably no later than second century.

A third-century date would accommodate the material from context (12)

Context (13), low in infill sequence of hollow 17

Context (10) above (12) in sequence infilling hollow 17

P9 (Figure 3.16), bowl or dish with flat-topped grooved rim, Type 21, third or fourth centuries.

Twenty-six sherds, including 20 from coarse gabbroic Type 16 storage jars, a rim as P11 and P13 (Figure 3.16), 85

Later Prehistoric Settlement in Cornwall and the Isles of Scilly P15 (Figure 3.16), girth sherd from Type 4 jar with wavy line incised on wet clay, probably third or fourth centuries. Contexts associated with wall cut [56] and wall 110, including 51, 45, 73 and (107) Forty-five sherds, none diagnostic except some dozen coarse gabbroic storage jar body sherds. Contexts [154], [137], [149], and [143], and associated with pits with charcoal and burning Twenty-nine sherds, none diagnostic except some 15 coarse gabbroic storage jar body sherds. Wall 111, post-dating pit (55) Forty-five sherds, including coarse gabbroic storage jar and Type 4 jar rim. Also P16 (Figure 3.17), Type 11 bowl or jar with small pointed everted rim, second or early third centuries. Soil spread (108) and related layers (53) and (52), late in sequence in central area Twenty-five sherds, including edge of lid and Type 13 storage jar. Soil accumulation (55) over central area Thirty-two sherds, including parts of Type 20 rim. Also P17 (Figure 3.17), rounded rim jar with tooled cordon beneath, simple version of Type 13; this Cordoned Ware-derived type, while commoner in the first to second centuries, probably continues through the third to fourth centuries (Quinnell 2004, 118).

Figure 3.16 Romano-British ceramics from hollow 17: P9 (13), P10 (12), P11 (12), P12 (12), P13 (10), P14 (2), and P15 (2).

P18 (Figure 3.17) Type 13 jar with rolled rim, a heavier version of P17.

dish with simple curved side, Type 23, third to fourth centuries.

P19 (Figure 3.17) Type 16 coarse gabbroic storage jar rim with simple out-turn.

Context 9 above (12)

The chronology and function of the ceramic assemblage

Contained three Type 4 jars and context (3) above five sherds, including three of coarse gabbroic storage jar.

Certain features such as pit [7] and ditch [58] should be no later than the second century. Type 19 bowls as P2 and Type 11 jars such as P1 and small Type 3 jars as P6 and P8 are unlikely to be third century. There is a lack of diagnostic pottery from the lower levels of the enclosure ditch fills, but upper fills of both ditches contain material best accommodated within the third century. A third-century date is the most appropriate that can be suggested on current data for the general range of forms from the interior, especially the Type 16 storage jars and the Type 20 bowls with flat, out-

Context (2) at top of sequence infilling hollow 17 Eighty-five sherds, mostly coarse gabbroic storage jar. P14 (Figure 3.16), Type 21 bowl, and two others, third to fourth centuries.

86

Chapter 3.4: The Artefacts

Figure 3.17 Romano-British ceramics: P16 wall 111, P17 soil in central area [55], P18 [55] and P19 [55].

turned, rims. If activity had continued into the late third or fourth centuries, bowls with conical flanged rims would be expected to occur (cf no 61 from Reawla: Quinnell in Appleton-Fox 1992, fig 17). The introduction of this type, Type 22, the one apparently distinctive later Roman form current in Cornwall, is discussed by Quinnell (2004, 124) with reference to Trethurgy. The enclosure at Killigrew may either have started as an open site in the second century, or the ditches could belong to this period but not have accumulated rubbish until their original purpose was forgotten and they ceased to be cleaned out.

site, with the vessels present relating mainly to storage and to eating. This pattern fits well with the suggested industrial function of the site as a whole. Metalwork finds The tin dish Henrietta Quinnell A cast metal dish with a footring (Figures 3.18 and 3.23) was found in pit [7]. It survived in fragmentary form, its metal content being almost completely oxidized (cf Tylecote 1979). The dish was about 225mm across, with a slightly pointed thickened rim; it was shallow, some 15mm deep, and had a footring 90mm across but only 4mm high.

The assemblage, although small, is distinguished by a range of storage jars and of bowls. Cooking pots, however, are poorly represented, compared to the numbers found on sites normally regarded as ‘domestic’ such as, for example, Castle Gotha (Saunders and Harris 1982). In the large assemblage at Trethurgy cooking pots formed 68 per cent, compared to bowls 19 per cent and storage jars 13 per cent (Quinnell 2004, chapter 5.6.3). Storage jars may have been used to bring to the site commodities such as prepared food, water or, indeed, materials not connected with nourishment; bowls are presumed to have been used as eating dishes. That the latter, at Killigrew, were frequently sooted suggests that food was being kept warm or heated up. No detailed study of the function of different ceramic types, especially in peripheral areas of the province, has yet been undertaken. The Killigrew assemblage at present suggests a pattern of food preparation off the

The metal was qualitatively analysed by X-ray fluorescence in 1996 by Dr David Starley of English Heritage Ancient Monuments Laboratory. It had been almost entirely tin, 95 per cent, with a small proportion of lead, traces of copper and of iron, the latter almost certainly deriving from burial contamination. The pottery found in pit [7] strongly suggests a second century date for its infill and therefore for the plate. Almost all cast plates and dishes of non-precious metal from Roman Britain are of late third or fourth century date. These are generally described as pewter, an alloy 87

Later Prehistoric Settlement in Cornwall and the Isles of Scilly pewter tableware industry does indeed belong to the late third and fourth centuries. The Killigrew dish, like those from Bucklersbury, does not conform to Peal’s typology for pewter ware (1967, figs 3–4), and its shape appears amongst the simplest on record. The best published comparanda are two rather smaller dishes, described as ‘in association’ with pottery of around AD 300 from Camerton (Wedlake 1958, 84 and fig 57). It should, however, be stressed that very few tin and tin alloy vessels survive from the first and second centuries AD.

Figure 3.18 Part of a tin dish found in pit [7].

Beagrie (1989) has summarized the data available for the Romano-British pewter industry and published analyses which show that vessels described as pewter fall into three broad groupings, with around 50 per cent, 70 per cent and 95 per cent of tin respectively. Beagrie makes the important points that comparatively few vessels have been analysed, that the broad groupings are not distinct, and that there were no standards for tin alloys in Roman Britain. He also shows that vessels with 95 per cent tin could occur in ‘pewter’ hoards dated to the fourth century, and that while pewter became generally established in the third and fourth centuries, tin with small amounts of additive or of impurity could still be used. A variety of tin alloy vessels survive from the Roman period in Cornwall. (It may be noted, presumably because of the high profile of tin production in Cornwall, that such vessels, when unanalysed, are always referred to as being of tin, whereas elsewhere in the country they tend to be referred to as pewter.) These are illustrated and discussed by Penhallurick (1986). The Treloy bowl and lid were possibly associated with a brooch of second or early third century date (ibid, 202 and fig 107a). The Hallivick or Halviggan cup is of a form replicated in late third- and fourth-century British pewter hoards (ibid, 187 and fig 93). A simple bowl with a bead rim from Carnon had no real associations (ibid, 195 and fig 100), nor did a similar bowl from Parson’s Park, St Neot (ibid, 195 and fig 100), which analysis has shown to be 99.1 per cent tin. The Caerhays jug held a coin hoard with the range AD 253–282 (ibid, 187 and fig 124) and was 96 per cent tin. The Bosence jug (ibid, 215 and fig 124) has been analysed and is 85 per cent tin, 10 per cent copper, 3 per cent lead and 2 per cent iron (Brown

of tin and lead, of which the proportions vary but may be of the order of 60–80 per cent tin and 40–20 per cent lead. Only a minority of vessels have, however, so far been analysed for metal content. Wedlake (1958, 82–93) provided the first substantive discussion of the subject following his excavations at Camerton in Somerset; his conclusion of a late Roman date for the pewter industry has been supported by later research such as that of Peal (1967). Peal published a typology of rim types for pewter plates and dishes and noted that there were two examples which did not fit into this, those of probable second-century date from Bucklersbury House in the Walbrook area of London. The Bucklersbury plates, together with the other Walbrook material, were subsequently studied by Jones (1983), who published analyses showing that these plates were of 98 per cent and of 96 per cent tin (ibid, fig 3). Jones’ discussion of the Walbrook material demonstrates that occasional manufacture of tin ware took place throughout the Roman period, although the full development of a 88

Chapter 3.4: The Artefacts 1970, 111); this was found in association with a saucer or patera with an inscription to Mars, which analysis showed was 85 per cent tin, 15 per cent lead and 2mm in radial cross-section were considered for species identification.

Pyrus sp., pear; Sorbus spp., rowan, service tree and whitebeam. These taxa are anatomically similar. Prunoideae. Prunus spinosa L., blackthorn. Salicaceae. Salix sp., willow, and Populus sp., poplar. In most respects these taxa are anatomically similar.

The charcoal varied from firm and well-preserved to poor and friable. Samples 611 and 628 contained insufficient charcoal for identification. The samples were prepared using standard methods (Gale and Cutler 2000). Anatomical structures were examined using incident light on a Nikon Labophot-2 compound microscope at magnifications up to x400 and matched to prepared reference slides of modern wood. When possible, the maturity of the wood was assessed (namely, heartwood/sapwood).

Features from which charcoal was examined included pits [2012] (trench 2) and [5006] (trench 5). The fill of [2012], described as a hearth pit, included oak (Quercus sp.) and cf hazel (Corylus avellana). The [5006] sample, was considerably larger and consisted mainly of oak (Quercus sp.) with moderate growth rates (probably from largewood) but also included birch (Betula sp.), hazel (Corylus avellana) and the hawthorn/Sorbus group (Pomoideae). The origin of the charcoal is unknown.

Early Bronze Age

Romano-British

Results

Charcoal was relatively abundant in the fill of the Phase 1 ditch [1026]; associated samples were similar in species content and comprised mostly oak (Quercus sp.) but also included blackthorn (Prunus spinosa) and gorse (Ulex sp.) and / or broom (Cytisus scoparius). From the north side of the entrance where remnants of the lower portions of the banks forming the ditch / ramparts of the round (1280) were found, a small deposit of oak (Quercus sp.) charcoal was obtained from the fill of the revetment.

The taxa identified are presented in Table 4.4. Classification follows that of Flora Europaea (Tutin, Heywood et al 1964–80). Group names are given when anatomical differences between related genera are too slight to allow secure identification to genus level. These include members of the Pomoideae (Crataegus, Malus, Pyrus and Sorbus), Leguminosae (Ulex and Cytisus) and Salicaceae (Salix and Populus). When a genus is represented by a single species in the British flora, it is named as the most likely origin of the wood, given the provenance and period, but it should be noted that it is rarely possible to name individual species from wood features, and exotic species of trees and shrubs were introduced to Britain from an early period (Godwin 1956; Mitchell 1974). The anatomical structure of the charcoal was consistent with the following taxa or groups of taxa:

Deposit (1114), a build-up of accumulated silts over the surface of the enclosure entrance, appears to have included general household waste, for example, pottery, animal bones and charcoal. The latter included oak (Quercus sp.), hazel (Corylus avellana) and gorse (Ulex sp.) and/or broom (Cytisus scoparius). Continued use of the round saw the closure of the original entrance by the construction of ditch [1087] aligned across the causeway. The origin of samples 551 and 552, from the primary fill of the ditch [1087], is uncertain, although domestic waste is possible, especially as the deposit also contained burnt animal bone. This fill also appeared to contain the decayed remains of timber. The charcoal was named as oak (Quercus sp.), gorse (Ulex sp.)/broom (Cytisus scoparius) and the hawthorn/Sorbus group (Pomoideae).

Betulaceae. Betula sp., birch Corylaceae. Corylus avellana L., hazel Fagaceae. Quercus sp., oak Leguminosae. Cytisus scoparius (L.) Link, broom and Ulex sp., gorse. These taxa are anatomically similar. Rosaceae. Subfamilies: Pomoideae, which includes Crataegus sp., hawthorn; Malus sp., apple; 136

Chapter 4.4: The Charcoal Context (1135) related to a fill of rubble below (1112) close to the entrance: the associated sample included oak (Quercus sp.), gorse (Ulex sp.)/broom (Cytisus scoparius) and hazel (Corylus avellana).

(1034), the fill of ditch [1026]; (1197), a fill behind the revetment; (1199), the flagstone floor to the entrance; (1044), the fill of ditch [1087]; (1250), the fill of gully [1249]; (1135), the fill below (1112); and (1177), fill of ditch [1009]. The inclusion of charcoal in these deposits probably came about as a result of the deliberate dumping of hearth waste into these contexts with other waste material. Deposit (1114) over the flagstones at the enclosure entrance, an accumulation of silt, appeared to have built up over some period of time and thus probably included the scattered fragments of discarded fuel debris from various hearths or fires. The source of the charcoal in these samples is unknown but is assumed to have originated from domestic hearths. Evidence from the charcoal indicates the frequent use of oak, with species such as gorse/broom, hazel, the hawthorn group and blackthorn occurring less often (Table 4.4).

Charcoal was also examined from hearth samples located within the ‘round’, probably dating from earlier phases of occupation. The elaborately constructed hearth feature 1108 implies a more specialized use than a domestic hearth, perhaps for industrial / craft activities. Evidence of in situ burning was recorded on the basal deposit and on the shillet stone-work. The remains of fuel debris (charcoal) was recovered from several related contexts and identified as mostly oak (Quercus sp.) and gorse (Ulex sp.)/broom (Cytisus scoparius) but also birch (Betula sp.), hazel (Corylus avellana), the hawthorn/Sorbus group (Pomoideae) and willow (Salix sp.) and / or poplar (Populus sp.) (Table 4.4). Context (1147) also included cinder-like material and possibly coal, perhaps implicating non-domestic uses. A large deposit of narrow roundwood from context (1219) underlying (1214), the quartz cobbling in 1108, consisted entirely of gorse (Ulex sp.)/broom (Cytisus scoparius).

Charcoal was collected from a number of associated contexts in and around the hearth 1108 including (1146), (1147), (1148), (1150), (1193), (1194), (1195) and (1216). These consistently indicated the use of firewood obtained mainly from oak and gorse/broom; including a high ratio of narrow roundwood, with infrequent use of birch, blackthorn willow/poplar and probably hazel (Table 4.4). Context (1219), a large deposit under (1214), the quartz cobbling in 1108, consisted entirely of gorse/ broom.

The fill of hearth / cooking pit [1069], (1070) contained heat-cracked pebbles, burnt bone, charcoal and a quantity of porous black ‘cokey’ material. The charcoal consisted of fast-grown oak (Quercus sp.) sapwood, probably from fairly wide roundwood.

A small quantity of oak charcoal from hearth feature [1069] indicated the use of moderately fast-grown oak.

Features relating to post Romano-British use of the enclosure suggest the area was given over to agriculture. A small quantity of oak (Quercus sp.) charcoal was recovered from the fill of the east-west ditch [1009], context (1177), which flanked the entrance to the round. The origin of the charcoal is unknown.

Environmental evidence and fuel supplies The rather sparse data available from Beaker deposits suggests the selective use of oak for wood fuel, thereby verifying the availability and, probably, the dominance of oak in the area at this time. More certain evidence of woodland composition is illustrated by the frequency of oak in the Romano-British deposits, although it is clear that gorse (Ulex sp.) – the more likely taxon than broom, Cytisus scoparius, in this landscape, was also common. Gorse forms tough spiny thickets, capable of withstanding extreme exposure in harsh conditions. The combination of oak and gorse in the fuel deposits, both of which provide excellent firewood, suggests that gorse was prevalent on the higher, more exposed hillside and probably gathered from farther away, whereas oak woodland was probably confined to the lower, more sheltered sides of the valley. The growth rates recorded in the oak charcoal were typically those of trees growing in reasonable, although not optimal, conditions. Had the firewood been gathered from trees growing in stressed conditions, for example, on exposed scarps, this would have been reflected in the wood structure. Birch (Betula sp.), hazel (Corylus avellana) and hawthorn (Crateagus sp.) are typically

Discussion The assemblage of charcoal included deposits from both Early Bronze Age and Romano-British features. Samples were obtained from Early Bronze Age pits [2012] and [5006] of which the former is a possible hearth pit. Thus, while both deposits almost certainly represent fuel debris, it is difficult to attribute these to a specific activity but in the absence of industrial waste, hearth debris from cooking or heating seems the most likely source. It is probable that oak provided the bulk of the fuel, with some use of birch, hazel and the hawthorn group. The majority of the samples examined were obtained from Romano-British features in trench 1, mainly from contexts associated with the entrance to the ‘round’ but also from the hearth complex 1108 located within the enclosure. Samples / contexts associated with sequential modifications of the entrance included 137

Later Prehistoric Settlement in Cornwall and the Isles of Scilly

Context

Sample

2000

504

5007 1034 1197

1199 1044 1250 1135

1177

1070

1146

1147

1148

1150

Betula

Corylus

-

cf. 1

-

-

529

1

619

-

564

631

-

551

-

624

552

628

-

-

1

-

-

-

-

-

-

3

-

-

1

-

-

-

594

-

3

-

560

-

-

-

590

577

578

579

-

-

-

-

-

-

-

-

Insufficient for id

611

-

-

-

-

11h, 1s 5h

3h,2r 5h

-

-

-

8s

-

8r

-

1h, 1r, 4s

-

1s

5r

-

3

3

-

1h, 1r, 2s

5r

3r

-

3h, 5r, 3s

-

-

-

-

-

1216

614

-

-

cf. 2

-

2s

-

-

cf. 1

-

-

-

-

-

-

-

602

-

-

-

6h, 11r, 4s

1

-

-

1

-

-

2h

Salix/ Populus Ulex/ Cytisus

25h, 5s

2

601

598

-

Quercus

-

-

1194

1219

-

2

1

1195

Prunus

4

580

1193

Pomoideae

-

1r

-

-

-

10

-

3

-

-

-

26r

3s

3

18r

-

1r, 1s

-

3r

-

-

-

Insufficient for id

2r

-

-

10

38r

Key. h = heartwood; r = roundwood (diameter 2mm in radial cross-section were considered for species identification. One large sample from pit TC[5030] was 50 per cent subsampled.

Area B2 Structure 2 A small sample from the lower fill TC(2107) of the bowlshaped pit TC[2067], located close to the northern perimeter of the structure, indicated the use of oak (Quercus sp.), the hawthorn/Sorbus group (Pomoideae) and gorse (Ulex sp.)/broom (Cytisus scoparius). Oak fragments included both slow and moderately fastgrown wood. The pit did not appear to have served as a hearth.

The charcoal examined from Area B2 was mostly firm and well-preserved whereas samples from Area D were more degraded and friable. The samples were prepared using standard methods (Gale and Cutler 2000). The anatomical structures were examined using incident light on a Nikon Labophot-2 compound microscope at magnifications up to x400 and matched to prepared reference slides of modern wood. When possible, the maturity of the wood was assessed (that is, heartwood or sapwood) and stem diameters and the number of growth rings recorded. It should be noted that charred stems may be reduced in volume by up to 40 per cent.

Structure 3 The burnt lining in pit TC[2009] was consistent with its use as a hearth. Charcoal from the basal layer probably represents fuel residues remaining in situ from its use. Deposit TC(2041) from hearth pit TC[2009]contained large fragments, almost entirely oak (Quercus sp.) although a small amount of alder (Aldus glutenous) was also named. The oak consisted mainly of moderately fast-grown roundwood (including a stem fragment: diameter of 15mm and six growth rings). The origin of the sample from the uppermost fill TC(2000) is less certain but may also consist of in situ hearth debris. The charcoal was mainly oak (Quercus sp.) but also included birch (Betula sp.), holly (Ilex aquifolium) and gorse (Ulex sp.)/broom (Cytisus scoparius). This context also included a large quantity of pottery sherds and firecracked stones.

Results The taxa identified are presented in Table 5.18. Classification follows that of Flora Europaea (Tutin, Heywood et al 1964–80). Group names are given when anatomical differences between related genera are too slight to allow secure identification to genus level. These include members of the Pomoideae (Crataegus, Malus, Pyrus and Sorbus), Leguminosae (Ulex and Cytisus) and Salicaceae (Salix and Populus). When a genus is represented by a single species in the British flora, it is named as the most likely origin of the wood, given the provenance and period, but it should be noted that it is rarely possible to name individual species from wood features and exotic species of trees and shrubs were introduced to Britain from an early period (Godwin 1975; Mitchell 1974). The anatomical structure of the charcoal was consistent with the following taxa or groups of taxa:

Hearth pit TC[2003] was shallow and irregular in shape. Associated charcoal from fill TC(2004) consisted entirely of oak (Quercus sp.). Structure 4 The fill TC(2153) of bowl-shaped feature TC[2152] contained pottery fragments and charcoal. Although there was no evidence to identify this as a hearth pit it is probable that the charcoal represents domestic fuel debris; the taxa identified included narrow roundwood from oak (Quercus sp.), hazel (Corylus avellana) and the hawthorn/Sorbus group (Pomoideae).

Aquifoliaceae. Ilex aquifolium L., holly Betulaceae. Alnus glutinosa (L.) Gaertner, European alder; Betula sp., birch Corylaceae. Corylus avellana L., hazel Fagaceae. Quercus sp., oak Leguminosae. Cytisus scoparius (L.) Link, broom and Ulex sp., gorse. These taxa are anatomically similar. Rosaceae. Subfamily: Pomoideae, which includes Crataegus sp., hawthorn; Malus sp., apple; Pyrus sp., pear; Sorbus spp., rowan, service tree and whitebeam. These taxa are anatomically similar. Salicaceae. Salix sp., willow, and Populus sp., poplar. In most respects these taxa are anatomically similar.

Structure 5 This L-shaped structure consisted of a double row of postholes. Its function is unknown and dating is problematic, although its proximity to the Late Bronze Age structures suggests it is contemporaneous with these. Charcoal samples from the lower fills of postholes TC[2160], TC[2223] and TC[2230], contained multiple species with a high ratio of narrow roundwood and clearly originated from the remains of firewood, not from burnt posts. The taxa named included predominantly oak (Quercus sp.) but also hazel (Corylus 260

Chapter 5.5: The Charcoal

Sample Context

Salix / Populus

Ulex / Cytisus

Alnus Betula Corylus

Corylus/ Ilex Alnus

Pomoideae Quercus

-

-

-

-

-

6

11h, 10r, 4s

-

1

Fill of hearth pit / posthole [2003]

-

-

-

-

-

12h, 5r, 3s -

-

-

2

-

-

4r

-

3h, 18r, 6s -

6r

1

-

-

-

-

-

4h, 9r, 5s

-

-

2

-

-

-

-

-

7h, 23r, 8s -

-

Fill of pit TC[2152]

-

-

3

-

-

2

6r, 1s

-

-

Fill of posthole TC[2160]

Description

Area B2 Structure 2 238

TC(2107) Fill of pit [2067]

Structure 3 200

TC 2004)

211

Upper fill of TC(2000) hearth pit TC[2009]

213 232

TC(2041) Basel fill of hearth pit TC[2009]

Structure 4 246

TC(2153)

Structure 5 249

TC(2161)

-

-

13

-

-

-

1h, 18r, 1h -

-

256

257

TC(2224) Charcoal from 1 posthole TC[2223] TC(2229) -

-

-

-

-

-

2h, 6r, 6s

-

-

258

Fill of posthole TC(2231) TC[2230]

1

-

9r

-

-

-

1h, 1r

-

-

TC(2018)

Fill of posthole TC[2017]

-

-

-

-

-

-

5h, 4r, 37s -

-

TC(5014)

Upper fill of posthole TC[5012]

-

-

-

-

-

2h, 1r, 23s -

-

TC(5000)

Fill of hollow / pit TC[5030]

-

-

-

-

-

-

64h, 13s

-

-

1

-

-

1

-

-

6r

1

4

-

-

-

-

-

-

33h, 1r, 11s

-

-

-

-

6

-

4

2h, 5s

-

-

Other features 207 Area D 501 502

503 505

Charcoal within TC(5004) pit TC[5039]

506

TC(5058)

Charcoal within pit TC[5057]

-

-

-

-

-

-

32h, 2s

-

-

Table 5.18: Truro College playing fields: Charcoal from the Late Bronze Age and Iron Age settlements. Key. h = heartwood; r = roundwood (diameter