Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire: Excavations at Brooklyn House 2015-16 1789698383, 9781789698381

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire: Excavations at Brooklyn House 2015-16 reports on exca

329 88 16MB

English Pages 296 Year 2021

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Prehistoric and Roman Archaeology of North-East Yorkshire 9780860541813, 9781407317120

225 46 123MB Read more

Excavations in the North Cemetery at Corinth in 1930 9781463221096

This is the site report from the excavation of the North cemetery in Corinth and represents a continuation of the report

192 89 6MB Read more

Excavations in the North Cemetery at Corinth in 1930 9781463221096

This is the site report from the excavation of the North cemetery in Corinth and represents a continuation of the report

199 66 6MB Read more

The Excavations of Roman Baths at Bath

284 43 1MB Read more

Life and Death in the Roman Suburb 0198852754, 9780198852759

Defined by borders both physical and conceptual, the Roman city stood apart as a concentration of life and activity that

969 181 9MB Read more

North Yorkshire Folk Tales 9780750955423

459 15 4MB Read more

Iron Age Cemeteries in East Yorkshire: Excavations at Burton Fleming, Rudston, Garton-on-the-Wolds, and Kirkburn 1850743517, 9781850743514, 9781848021662

With contributions by Janet Ambers, A. J. Clark, Elisabeth Crowfoot, Vanessa Fell, I. C. Freestone, J. R. A. Greig, Juli

586 84 22MB Read more

Iron Age, Roman and Saxon Occupation at Grange Park: Excavations at Courteenhall, Northamptonshire, 1999 9781841719979, 9781407320793

This report provides the results of archaeological investigations undertaken by Birmingham University Field Archaeology

217 40 67MB Read more

Cato the Younger: life and death at the end of the Roman republic 9780190869038, 0190869038

Family -- Early years -- Cato the tribune -- Cato and the formation of the triumvirate -- Cato's Cyprian mission an

1,529 162 25MB Read more

Excavations at 16-20 Church Street, Bawtry, South Yorkshire 9780860548348, 9781407318752

Report on urban excavations which revealed an occupation sequence from thirteenth century burgage plots into the moden p

184 25 136MB Read more

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire: Excavations at Brooklyn House 2015-16
1789698383, 9781789698381

Author / Uploaded
Janet Phillips
Pete Wilson

Categories
History
Archaeology

Table of contents :
Cover
Title Page
Copyright page
Contents Page
List of Figures
List of Tables
Acknowledgements
Chapter 1
Introduction
Pete Wilson
Figure 1: The Location of Malton and Norton
Figure 2: Roman sites in Malton and Norton
Figure 3: The Brooklyn House area – sites and find spots
Chapter 2
The Excavation
Janet Phillips and Pete Wilson
Figure 4: Overall Phase 1 plan with Trial Trenches
Figure 5: Plot of Geophysical Survey results
Figure 6: Phase 1 plan of Area 5
Figure 7: Area 5 – Section of Ditch 505
Figure 8: Area 5 – Section of Ditch 507
Figure 9: Area 6 – Section of Ditch 603
Figure 10: Area 2 – Phase 1 plan
Figure 11: Area 2 – Feature 287
Figure 12: Area 2 – Feature 287 (view to west)
Figure 13: Area 2 – Feature 2003
Figure 14: Area 2 – Ditch 298
Figure 15: Area 2 – Ditch 2005
Figure 16: Area 3 Phase 1/2 plan
Figure 17: Area 3 – East-facing section
Figure 18: Area 3 – Pit 3032
Figure 19: Area 3 – Gully 3028 and Pit 3032
Figure 20: Area 3 – Posthole 3034
Figure 21: Area 7 – South-facing section
Figure 22: Area 7 – West-facing section
Figure 23: Area 7 – North facing section
Figure 24: Areas 4 and 9 – South-facing section
Figure 25: Area 4 – Phase 2 Roman road and bustum plan
Figure 26: Area 4 – Access road north verge southern service – south-facing section
Figure 27: Area 4 – Section through the Roman road (view to north east)
Figure 28: Area 4 – Road metalling 418 (view to west)
Figure 29: Area 4 – Access road north verge northern service trench – south-facing section
Figure 30: Area 3 – Roadside ditch 397 (view to north)
Figure 31: Area 3 – Ditches 397, 3024 and 3036 (view to north)
Figure 32: Area 3 – Section of ditches 397, 3024 and 3036
Figure 33: Area 3 – Section of ditch 307 and gully 3097
Figure 34: Area 3 – Partial section of ditch 3024
Figure 35: Phase 2 – Location of burials
Figure 36: Area 2 – Grave 284, skeleton 286. 36a: Capstones partially covering SK286, 36b: SK286
Figure 37: Area 2 – Grave 284, skeleton 286 showing capstones
Figure 38: Area 2 – Grave 284, skeleton 286
Figure 39: Area 7 – Grave 716, skeleton 718
Figure 40: Area 7 – Grave 716 profile
Figure 41: Area 7 – Grave 284, skeleton 286
Figure 42: Area 4 – Section through bustum burial 475
Figure 43: Area 4 – Profile, north east to south west, through bustum pit 475
Figure 44: Area 4 – Profile, north west to south east, through bustum pit 475
Figure 45: Area 4 – Bustum during excavation – complete local grey ware beaker (No. 41, RF 417)
Figure 46: Area 4 – Bustum during excavation. Main bustum vessel, handmade calcite-gritted jar (No. 108, RF 416), under local grey ware beaker (No. 41, RF 417)
Figure 47: Area 4 – Bustum during excavation. Other pottery fragment overlying the main bustum vessel
Figure 48: Area 4 – Bustum during excavation. Main bustum vessel fully exposed
Figure 49: Area 4 – Bustum pit fully excavated
Figure 50: Area 2 – Plan of possible mausoleum (feature 270)
Figure 51: Area 2 – Section through robber trench (266) and construction cut (270) of possible mausoleum, showing overlying wall 205 of Structure A
Figure 52: Area 2 – Robber trench (266) and construction cut (270) of possible mausoleum, with overlying wall 205 of Structure A
Figure 53: Site plan showing all structures
Figure 54: Area 2 – Plan of Structure A
Figure 55: Area 2 – Structure A (view to north west)
Figure 56: Area 2 – Section through east-west line of post-holes within Structure A
Figure 57: Area 2 – with Structure A in the foreground (view to south east)
Figure 58: Area 2 Structure A – profile of wall 217
Fig 59: Area 2 Structure A – profile of wall 205/217
Figure 60: Area 2 Structure A – wall 205 (view to south)
Figure 61: Area 2 Structure A – wall 205/217 looking towards the possibly entrance in south-east corner of the building (view to south east)
Figure 62: Area 2 Structure A – wall 205/217 showing extent of robbing of parts of the wall (view to north west)
Figure 63: Area 2 Structure A – possible remnant of a flagged floor, context 225 (view to south)
Figure 64: Area 2 – Section through possible yard surface 236
Figure 65: Area 2 – Possible yard surface 236 (view to east)
Figure 66: Area 2 – Profile of wall 254
Figure 67: Area 2 – Wall 254 (view to north)
Figure 68: Area 2 – Plan of Structures B and C
Figure 69: Area 2 Structure B – profile of possible wall core 252
Figure 70: Area 2 Structure B – possible wall cores 252 (foreground) and 253 (rear) (view to south)
Figure 71: Area 2 – Section of possible surface 251, ash pit 275 and wall 246 of Structure C
Figure 72: Area 2 Structure C – profile of the inner face of wall 249
Figure 73: Area 2 Structure C – view along flue showing in situ capstone (view to west)
Figure 74: Area 2 – Overall view of Structure C (view to north west)
Figure 75: Areas 3 and 8 – Overall plan of Structures D, E and I
Figure 76: Areas 3 and 8 – Plan of Structure D and area to the north
Figure 77 Area 3 – East-facing trench edge section
Figure 78: Area 3 – The east-west walls of Structures D, E and I. Wall 314 (Structure E ) the top of the image, with wall 353 (Structure D) in the middle and wall 355 (Structure I) in front. In the left foreground are walls 315 (Structure D) and to the ri
Figure 79: Area 3 – The east-west walls of Structures D, E and I. Wall 314 (Structure E ) largely robbed out to the right, wall 353 (Structure D) in the centre and wall 355 (Structure I) to the left (view to west)
Figure 80: Area 8 – East-facing section
Figure 81: Area 3 – Section through walls 355, 353 and 314 (Structures I, D and E respectively)
Figure 82: Area 3 – West-facing trench edge section
Figure 83: Area 3 Wall 308 – West-facing profile
Figure 84: Area 3 – Section through Wall 308/Construction cut 307
Figure 85: Area 3 Structure E – Profile of Wall 313/327
Figure 86: Area 3 Structure E – Wall 313/327 (view to north)
Figure 87: Area 3 Structure E – Wall 313/327 (view to north)
Figure 88: Area 3 – Bird burial 316
Figure 89: Area 3 – Bird burial 316 – geese skulls
Figure 90: Area 3 – Structure F
Figure 91: Area 3 – Structure F (view to east)
Figure 92: Area 3 Structure F – possible internal wall 3092 (view to east)
Figure 93: Area 7 extension – West-facing trench edge section
Figure 94: Area 4 and 9 – Structure G
Figure 95: Area 4 – Northern verge trenches western end
Figure 96: Area 4 – Section through possible wall 4034
Figure 97: Area 3 Structure I – Walls 354 and 355
Figure 98: Area 3 – Section through walls 354 (Structure I) and 353 (Structure D). Wall 314 of Structure E removed. Wall 355 (Structure I) is under the ranging pole on the left of the frame.
Figure 99: Area 3 – Profile of wall 355 (Structure I)
Figure 100: Area 3 – Hearth within Structure I
Figure 101: Area 3 – Rubble spread 344 (view to east)
Figure 102: Area 7 – Service trench section
Figure 103: Plan locating grid squares in Area 2
Chapter 3
Prehistoric pottery
Blaise Vyner
Chapter 4
The Roman Pottery
I.M. Rowlandson and H.G. Fiske
Table 1: Quantification of samian by vessel class in Rim EVE and Maximum Vessel Number (MVN)
Table 2: Quantification of samian by Fabric
Table 3: Quantity of samian by stratigraphic phase
Table 4: Vessel Form and Function by Fabric Group (Production Area). Maximum Vessel Numbers (rim EVE in brackets) Vessel forms are Dragendorff (Drag) forms unless specified. (O and P = Oswald and Price 1920)
Figure 104: The decorated samian, scale 1:1
Figure 105: The stamped samian, scale 1:1
Figure 106: Illustrated plain samian, scale 1:2
Figure 107: Stamped Dressel 20 amphora: 107a: Rubbing of stamp (scale 1:1), 107b: photo of stamp, 107c: and microscope image of vessel fabric
Figure 108: The illustrated mortaria
Figure 109: Illustrated mortarium No. 2 of the potter GENIALIS. 109a: Rubbing (Scale 1:1), 109b: photograph of stamp, 109c: microscope fabric image
Figure 110: Grey ware mortarium with slag trituration grits (No. 6). 110a: general image and 110b: microscope fabric image
Table 5: MOCR Crambeck white ware mortaria by form type
Figure 111: A burnt Crambeck mortarium, interior and exterior, illustrated vessel No. 9. 111a: internal surface, 111b: external surface
Figure 112: A sooted Crambeck mortarium, exterior and interior, illustrated vessel 10
Table 6: MOCRF Crambeck Parchment ware mortaria by form type
Figure 113: Paint-decorated Crambeck Type 7 mortarium, illustrated vessel 13
Figure 114: Paint-decorated Crambeck type 7 mortarium, illustrated vessel 14
Table 7: CRPA Crambeck Parchment ware other forms by type
Figure 116: Crambeck parchment ware bowl, illustrated vessel 17
Figure 117: Crambeck parchment ware bowl type 5b, illustrated vessel 20
Figure 118: Crambeck parchment ware bowl type 5b, with painted crux gammata illustrated vessel 22
Figure 119: Crambeck parchment ware bowl type 5b with paint decoration, illustrated vessel 23
Figure 120: Crambeck parchment ware face pot and smith god pot fragments, left to right No. 25-27
Figure 122: Images showing the Crambeck parchment ware chicken/cockerel figurine (No. 28)
Figure 123: Stamped Parisian ware sherd (No. 29) and rubbing (rubbing scale 1:1)
Figure 124: Possible Triple vase single element in white ware fabric (No. 30)
Figure 125: Oxidised beaker base showing signs of being re-ground to form a small container (No. 33)
Table 8: NORGW1 forms as proportion of fabric (Norton types follow Hayes and Whitley 1950)
Figure 126: Norton reduced wares NORBB1, NORGW1 and NORCG
Figure 127: Folded beaker in Norton grey ware (No. 41)
Table 9: CRGR forms as proportion of fabric (Crambeck types follow Corder 1937)
Figure 128: Crambeck reduced wares CRGR, CRGR? and CRGRC
Figure 129: CRGR Basal sherd showing evidence of re-use (No. 58)
Figure 130: Mask-mouthed flagon in Crambeck grey ware (No. 60)
Figure 131: Left to Right- No. 61-3 Sherds from face-necked flasks in Crambeck grey ware
Figure 132: Face jar in Crambeck grey ware (No. 64)
Figure 133: Roundel-stamped face jar in Crambeck grey ware (No. 65)
Figure 134: Face pot fragments in Crambeck grey ware (left to right, 66-69)
Figure 135: Smith god pot sherds in Crambeck grey ware with applied tools and other motifs (Top row left to right 70-72, Bottom row left to right 73-76).
Figure 136: Applied wheels from Smith god and/or wheel god pots (Top row left to right 77-79; Bottom row left to right 80-81)
Figure 137: Crambeck grey ware sherd with applied snake (No. 82)
Figure 138: Various stamped and applied partial motifs. (Top row left to right 48 and 83. Bottom row left to right 84-86). Vessel 48 is in a Norton grey ware fabric NORGW1, the remainder in Crambeck grey ware CRGR
Figure 139: Holme-on-Spalding-Moor grey ware spindle whorl (No. 90)
Figure 140: Burnished sherd with applied pellet decoration (No. 94)
Table 10: Relative proportion of Norton, Crambeck grey wares and other reduced wares as proportion of whole assemblage by phase
Table 12: Calcite-gritted jar forms by % of Rim equivalents by phase
Table 13: Relative proportion of calcite-gritted wares and reduced wares as proportion of assemblage by phase
Figure 141: Illustrated calcite-gritted wares
Figure 142: Calcite-gritted jar from bustum cremation pit 475 (No. 108)
Figure 144: Proto-Huntcliff jar showing method of construction. Left- external, Right- internal surface showing hand building and wheel-finished rim (No. 126)
Figure 145: Illustrated calcite-gritted jars
Figure 146: Calcite-gritted other forms
Figure 147: Calcite-gritted ware miniature jar (No. 136)
Figure 148: Calcite-gritted ware body sherd cut down to a spindle whorl (No. 137)
Figure 149: Calcite-gritted baking plate (No. 138)
Figure 150: Fragment of portable oven in calcite-gritted ware (No. 140)
Figure 152: Fragment of portable oven No. 143 in calcite-gritted ware showing handmade manufacture technique (rim to the top)
Figure 154: Calcite-gritted ovens, large jars and a tray
Figure 155: Large calcite-gritted jar with internal finger pressed decoration jar (No. 147)
Figure 156: Vessels from Bustum burial 475
Table 14: Summary of selected excavated assemblages, the relative numbers of decorated and stamped vessels illustrating the potential size of those excavated in the twentieth century
Table 15: Quantity (by sherd count) of samian from recent excavations in Norton
Table 16: The Relative Frequency of Samian Form/Functional Categories at different site types, Brooklyn House, Norton and a comparative site from Worcester (Mills J. 2018)
Table 17: Crambeck Parchment wares as proportion of the assemblage by phase
Chapter 5
The post-Roman Pottery
C.G. Cumberpatch
Chapter 6
Figure 158: Counterfeit denarii of Trajan, Elagabalus and Orbiana (obverses only).
The coins
Richard Brickstock
Figure 159: Graphical summary of Shiptonthorpe (1985-91) coins
Figure 160: Graphical summary of Malton vicus (1968-70) coins
Figure 161: Graphical summary of Malton vicus (1949-52) coins (regular and copies largely undifferentiated)
Table 18: Summary of coin finds from Brooklyn House, Norton (2015-16)
Table 19: Summary of coin finds from Brooklyn House and Malton Vicus (1968-70)
7a – The small finds and vessel glass
H.E.M. Cool and S.J. Greep
Chapter 7
Table 20: The catalogued small finds by site phase and functional category. (Note: see text for exclusions. The column labelled ‘Later’ combines the Phase 7, unphased and unstratified material)
Figure 162: Personal ornaments. Copper alloy (Nos 1-3) and iron (No. 4) brooches and bone hairpins (Nos 6-7). Scale 1:1
Figure 163: Personal ornaments Bone hairpins. Scale 1:1
Figure 164: Personal ornaments. Bone hairpin( No. 17), glass beads (Nos 19-20), copper alloy bracelets (Nos 21, 23) and shale bangles (Nos 24-27, 29)
Figure 165: Personal ornaments and toilet equipment. Shale (No. 30), jet (No. 31) and bone (No. 32) bangles and bracelets and copper alloy tweezers (No. 34). Scale 1:1
Figure 166: Textile equipment. Iron needles (Nos 36-37), bone needle case (No. 38) and shale spindle whorl (No. 39). Scale 1:1
Figure 167: Household equipment. Copper alloy spoon (No. 40). Scale 1:1. Glass vessels (Nos 41-43) and iron ladle (No. 45). Scale 1:2
Figure 168: Household equipment. Iron candlestick (No. 46), upholstery nail (No. 47). Scale 1:1. Iron trivet (No. 49). Scale 1:2
Figure 169: Recreation and writing equipment. Glass counter (No. 50) and iron styli (Nos 51-52). Scale 1:1
Figure 170: Writing equipment. Iron styli. Scale 1:1
Figure 171: Writing equipment. Iron styli. Scale 1:1
Figure 172: Transport equipment. Iron bits. Scale 1:1
Table 21. Types of iron nails present according to Roman phases (quantified by heads and according to Manning (1985) types)
Figure 173: Structural ironwork. Joiners dog (No. 60), angle bracket (No. 62) and nails (Nos 64-70). Scale 1:1
Figure 174: Iron knives (Nos 71-76), bladed tool (No. 78) and punch (No. 79). Scale 1:2
Figure 175: Iron tongs. Scale (drawing) 1:3
Figure 176: Iron chisels (Nos 82-84) saw blade (No. 85), awl (No. 86) and handle socket (No. 87) and antler handle (No. 88). Scale 1:2 (Nos 82, 84, 86, 88), 1:1 (Nos 83, 85, 87)
Figure 177: Fittings. Iron key (Nos 90-95), box fitting (No. 97), ferrules (Nos 101-102), nailed socket (No. 103), loop-headed spike (No. 104) and copper alloy double-spiked loop (No. 96). Scale 1:2 (Nos 90, 91, 93-95, 97, 101-104), 1:1 (No. 96)
Figure 178: Agricultural and military equipment. Iron ox goad (No. 115), copper alloy lorica squamata scales (Nos 116-118), lorica segmentata fitting (No. 119), stud (No. 120) and phallic pendant (No. 121). Scale 1:2 (No. 115); 1:1 Nos 116-121; 2:1 detai
Figure 179: Weapons. Iron bolts (Nos 122-123) and arrowhead (No. 124). Scale 1:2
Figure 180: Miscellaneous items. Shale pendant (No. 125), iron rod (No. 132) and toothed strip (No. 136), copper alloy terminal (No. 136) and curved band (No. 137). Scale 1:1 Nos 125, 135; 1:2 Nos 132, 136, 137)
Table: 22: Distribution of the catalogued metal items through time
Table: 22: Distribution of the catalogued metal items through time (excluding nails and hobnails)
7b – The finds from the cremation burial
H.E.M. Cool and S.J. Greep1
Figure 181: A selection of the pyre goods. Scale 1:1
Figure 182: Reconstruction of the belt
Figure 183: Belt set. Copper alloy buckle and plates (Nos 1-2). Scale 1:1
Figure 184: Belt set and baldric fittings. Copper alloy plates with frogs (Nos 3-4) and baldric fittings (Nos 7-8). Scale 1:1
Figure 185: Other fittings from pyre. Iron and copper alloy blade (Nos 9-11), copper alloy brooch (No. 12), iron hobnail (No. 13), bone inlay (No. 15) and scroll holder fittings (Nos 16). Scale 1:1
Figure 186: Other fittings from pyre. Bone (No. 17), copper alloy (Nos 18-20) and copper alloy with iron (No. 21). Scale 1:1
Chapter 8
Table 23: Ceramic Building Material by form in relation to period
The ceramic building materials
J.M. McComish
Figure 187: Curving tiles with vents from context 758
Figure 188: Chimney sherd from context 206
Figure 189: Chimney or finial fragments from contexts 202 and 236
Figure 190: Combed box flues from contexts 253 and 309
Figure 191: Finger drawn keying on a box flue from context 4023
Figure 192: Parietalis from context 202
Figure 193: Tesserae from context 202
Table 24: Ceramic Building Material in relation to phase
Chapter 9
Table 25: Stone Building Material by form and stone type
The stone building materials
J.M. McComish
Figure 194: Two stone tesserae, RF 6 context 201 and RF 296 context 202
Figure 195: Stone roof flag of rectangular shape, context 816
Figure 196: Stone roof flag of elongated hexagonal shape, context 207
Figure 197: Stone roof flags with a curving upper edge from contexts 202, 220 and 830
Figure 198: Roof tiles re-used as lids, RF 194, context 235 and RF 308, context 202
Figure 199: Roof tile with scratched surface, possibly re-used as a hone, context 202
Figure 200: Limestone slab with bullnose profile, context 202
Figure 201: Carved stone RF 177, context 215
Figure 202: Carved block RF 96, context 202
Table 26: Stone Building Material from Phase 3 structures
Table 27: Tooling and burning on the Stone Building Material by structure
Chapter 10
Chipped stone lithics
Spencer Carter
Table 28: Lithics submitted finds composition summary
Table: 29: Lithics raw material types
Figure 203: Late Neolithic oblique arrowhead No. 8
Table 30: Lithic reduction technology
Table 31: Burnt lithics
Table 32: Chronological indicators
Table 33: Modified formal and non-formal tool forms
Chipped stone lithics Appendix
Chapter 11
Querns, millstones and associated material
John Cruse
Table 34: Saddle Querns, Rubbers and Polishers
Figure 204: Hand querns with a radial pattern of grooves on the rim. 204a: No. 11, RF 388, context 413; 204b: No. 13, RF 356, context 302
Table 35: Hand Querns
Table 36: Powered Millstones
Table 37: Phasing of the Querns
Table 38: Querns from excavations in Malton and Norton (from YQS Records)
Figure 205: Find spots of rotary querns excavated in Malton and Norton
Table 39: Roman sites in the vicinity of Malton and Norton
Figure 206: Roman sites with querns in the vicinity of Malton and Norton
Table 40: Comparison of Querns from the Study Area with YQS Quern Type Summaries
Figure 207: Fragment of a probable small millstone No. 15, RF 452, context 3117
Chapter 12
The whetstones
S. Tibbles
Table 41: Whetstones
Chapter 13
Human bone
Katie Keefe and Malin Holst
Table 42: Summary of osteological and palaeopathological results
Table 43: Summary of cremated bone assemblage
Table 44: Summary of cremated bone fragment size
Table 45: Summary of identifiable elements in the cremation burials
Chapter 14
Vertebrate remains
Alison Foster
Table 46: Hand-collected vertebrate remains (NISP – number of identified specimens) from Areas 3, 4, 5, 7, 8 and 9
Table 47: Hand-collected vertebrate remains (NISP – number of identified specimens) from Phases 1, 2, 3, 4 and 5
Table 48: Tooth-wear stages recorded for sheep and sheep/goat mandibles.
Table 49: Tooth-wear stages recorded for pig mandibles.
Table 50: Number of sheep and sheep/goat mandibles that could be assigned to the various age categories of Payne (1973; 1987), by phase
Table 51: Number of pig mandibles that could be assigned to the various simplified age categories of O’Connor (2003), by phase
Table 52: Metrical data (following von den Dreisch 1976) for domestic mammals, by element. All measurements are in mm
Figure 208: Frequency of cattle, caprine and pig remains (NISP) as a percentage of total fragments for Phases (2 -5)
Table 53: Metrical data (following von den Dreisch 1976) for birds, by element. All measurements are in mm
Table 54: Estimated withers heights in centimetres by context and phase
Chapter 15
Shell
John Carrott
Chapter 17
Discussion
Pete Wilson
Figure 209: Roman Norton showing proposed revised cemetery areas along Langton Road
Bibliography
Back cover

Citation preview

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire Excavations at Brooklyn House 2015-16 Janet Phillips and Pete Wilson

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire Excavations at Brooklyn House 2015-16 Janet Phillips and Pete Wilson with contributions by

Tony Benfield, Joanna Bird, Richard Brickstock, Spencer Carter, John Carrott, Hilary Cool, John Cruse, Chris Cumberpatch, David Dungworth, Hugh Fiske, Alison Foster, Stephen Greep, Kay Hartley, Malin Holst, Katie Keefe, Jane McComish, J.M. Mills, Ian Rowlandson, Sophie Tibbles, Blaise Vyner and David Williams illustrations by

Janet Phillips, Hilary Cool, Hugh Fiske and David Heslop photography for Chapters 7-9 and 11 by:

Dominic Powlesland

Archaeopress Roman Archaeology 77

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

ISBN 978-1-78969-838-1 ISBN 978-1-78969-839-8 (e-Pdf) © Archaeopress and the individual authors 2021

Cover Illustration: The bustum burial under excavation.

The site archive, including specialist data, will be deposited with Malton Museum, 36 Yorkersgate, Malton YO17 7AB. Accession Code: MALTM2020.1.

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 List of Figures�� ii List of Tables�� v Acknowledgements�� vii Chapter 1 Introduction�� 1 Pete Wilson Chapter 2 The excavation�� 8 Janet Phillips and Pete Wilson Chapter 3 Prehistoric pottery��77 Blaise Vyner Chapter 4 The Roman pottery��78 I.M. Rowlandson and H.G. Fiske Chapter 5 The post-Roman pottery��143 C.G. Cumberpatch Chapter 6 The coins ��148 Richard Brickstock Chapter 7 7a – The small finds and vessel glass��156 H.E.M. Cool and S.J. Greep 7b – The finds from the cremation burial��187 H.E.M. Cool and S.J. Greep Chapter 8 The ceramic building materials��200 J.M. McComish Chapter 9 The stone building materials��207 J.M. McComish Chapter 10 Chipped stone lithics��214 Spencer Carter Chipped stone lithics Appendix��219 Chapter 11 Querns, millstones and associated material��224 John Cruse Chapter 12 The whetstones��234 S. Tibbles Chapter 13 Human bone��237 Katie Keefe and Malin Holst Chapter 14 Vertebrate remains��248 Alison Foster Chapter 15 Shell��258 John Carrott Chapter 16 Discussion��260 Pete Wilson Bibliography�� 268 i

List of Figures Figure 1: The Location of Malton and Norton�� 2 Figure 2: Roman sites in Malton and Norton �� 3 Figure 3: The Brooklyn House area – sites and find spots�� 5 Figure 4: Overall Phase 1 plan with Trial Trenches�� 9 Figure 6: Phase 1 plan of Area 5�� 10 Figure 5: Plot of Geophysical Survey results�� 10 Figure 7: Area 5 – Section of Ditch 505 �� 11 Figure 8: Area 5 – Section of Ditch 507 �� 12 Figure 9: Area 6 – Section of Ditch 603 �� 12 Figure 10: Area 2 – Phase 1 plan�� 13 Figure 11: Area 2 – Feature 287�� 14 Figure 12: Area 2 – Feature 287 (view to west)�� 14 Figure 13: Area 2 – Feature 2003 �� 15 Figure 14: Area 2 – Ditch 298 �� 15 Figure 15: Area 2 – Ditch 2005�� 15 Figure 16: Area 3 Phase 1/2 plan�� 17 Figure 17: Area 3 – East-facing section�� 18 Figure 20: Area 3 – Posthole 3034�� 19 Figure 18: Area 3 – Pit 3032�� 19 Figure 19: Area 3 – Gully 3028 and Pit 3032�� 19 Figure 21: Area 7 – South-facing section�� 20 Figure 22: Area 7 – West-facing section�� 20 Figure 23: Area 7 – North facing section�� 20 Figure 24: Areas 4 and 9 – South-facing section�� 21 Figure 25: Area 4 – Phase 2 Roman road and bustum plan�� 22 Figure 26: Area 4 – Access road north verge southern service – south-facing section�� 23 Figure 27: Area 4 – Section through the Roman road (view to north east)�� 24 Figure 28: Area 4 – Road metalling 418 (view to west)�� 25 Figure 29: Area 4 – Access road north verge northern service trench – south-facing section�� 26 Figure 30: Area 3 – Roadside ditch 397 (view to north)�� 27 Figure 31: Area 3 – Ditches 397, 3024 and 3036 (view to north)�� 27 Figure 33: Area 3 – Section of ditch 307 and gully 3097�� 28 Figure 34: Area 3 – Partial section of ditch 3024 �� 28 Figure 32: Area 3 – Section of ditches 397, 3024 and 3036�� 28 Figure 35: Phase 2 – Location of burials�� 30 Figure 36: Area 2 – Grave 284, skeleton 286. 36a: Capstones partially covering SK286, 36b: SK286�� 31 Figure 37: Area 2 – Grave 284, skeleton 286 showing capstones�� 31 Figure 38: Area 2 – Grave 284, skeleton 286�� 31 Figure 39: Area 7 – Grave 716, skeleton 718�� 32 Figure 40: Area 7 – Grave 716 profile�� 32 Figure 41: Area 7 – Grave 284, skeleton 286�� 32 Figure 42: Area 4 – Section through bustum burial 475�� 32 Figure 43: Area 4 – Profile, north east to south west, through bustum pit 475�� 33 Figure 44: Area 4 – Profile, north west to south east, through bustum pit 475�� 33 Figure 45: Area 4 – Bustum during excavation – complete local grey ware beaker (No. 41, RF 417)�� 34 Figure 49: Area 4 – Bustum pit fully excavated�� 34 Figure 47: Area 4 – Bustum during excavation. Other pottery fragment overlying the main bustum vessel�� 34 Figure 46: Area 4 – Bustum during excavation�� 34 Figure 48: Area 4 – Bustum during excavation. Main bustum vessel fully exposed�� 34 Figure 50: Area 2 – Plan of possible mausoleum (feature 270)�� 35 Figure 51: Area 2 – Section through robber trench�� 35 Figure 52: Area 2 – Robber trench (266) and construction cut (270) of possible mausoleum, with overlying wall 205 of Structure A��36 Figure 53: Site plan showing all structures�� 37 Figure 54: Area 2 – Plan of Structure A�� 38 Figure 55: Area 2 – Structure A (view to north west)�� 39 Figure 56: Area 2 – Section through east-west line of post-holes within Structure A �� 39 Figure 58: Area 2 Structure A – profile of wall 217�� 40 Figure 57: Area 2 – with Structure A in the foreground (view to south east)�� 40 Figure 60: Area 2 Structure A – wall 205 (view to south)�� 41 Fig 59: Area 2 Structure A – profile of wall 205/217�� 41 Figure 61: Area 2 Structure A – wall 205/217�� 41 Figure 62: Area 2 Structure A – wall 205/217 showing extent of robbing of parts of the wall (view to north west)�� 42 Figure 64: Area 2 – Section through possible yard surface 236�� 42

ii

Figure 63: Area 2 Structure A – possible remnant of a flagged floor, context 225 (view to south)��42 Figure 65: Area 2 – Possible yard surface 236 (view to east)��43 Figure 67: Area 2 – Wall 254 (view to north)��43 Figure 66: Area 2 – Profile of wall 254��43 Figure 68: Area 2 – Plan of Structures B and C��44 Figure 69: Area 2 Structure B – profile of possible wall core 252��45 Figure 70: Area 2 Structure B – possible wall cores 252 (foreground) and 253 (rear) (view to south)��45 Figure 71: Area 2 – Section of possible surface 251, ash pit 275 and wall 246 of Structure C��46 Figure 73: Area 2 Structure C – view along flue showing in situ capstone (view to west)��47 Figure 72: Area 2 Structure C – profile of the inner face of wall 249��47 Figure 74: Area 2 – Overall view of Structure C (view to north west)��48 Figure 75: Areas 3 and 8 – Overall plan of Structures D, E and I��49 Figure 76: Areas 3 and 8 – Plan of Structure D and area to the north��50 Figure 77 Area 3 – East-facing trench edge section��51 Figure 78: Area 3 – The east-west walls of Structures D, E and I��52 Figure 79: Area 3 – The east-west walls of Structures D, E and I. Wall 314 (Structure E )��52 Figure 80: Area 8 – East-facing section��53 Figure 81: Area 3 – Section through walls 355, 353 and 314 (Structures I, D and E respectively)��54 Figure 82: Area 3 – West-facing trench edge section ��55 Figure 84: Area 3 – Section through Wall 308/Construction cut 307 ��56 Figure 83: Area 3 Wall 308 – West-facing profile��56 Figure 85: Area 3 Structure E – Profile of Wall 313/327 ��56 Figure 86: Area 3 Structure E – Wall 313/327 (view to north)��57 Figure 87: Area 3 Structure E – Wall 313/327 (view to north)��57 Figure 88: Area 3 – Bird burial 316��58 Figure 89: Area 3 – Bird burial 316 – geese skulls��59 Figure 90: Area 3 – Structure F��60 Figure 91: Area 3 – Structure F (view to east)��61 Figure 93: Area 7 extension – West-facing trench edge section��61 Figure 92: Area 3 Structure F – possible internal wall 3092 (view to east)��61 Figure 94: Area 4 and 9 – Structure G��62 Figure 95: Area 4 – Northern verge trenches western end��64 Figure 96: Area 4 – Section through possible wall 4034��65 Figure 97: Area 3 Structure I – Walls 354 and 355��67 Figure 98: Area 3 – Section through walls 354 (Structure I) and 353 (Structure D).��67 Figure 99: Area 3 – Profile of wall 355 (Structure I)��67 Figure 100: Area 3 – Hearth within Structure I ��68 Figure 101: Area 3 – Rubble spread 344 (view to east)��70 Figure 102: Area 7 – Service trench section��71 Figure 103: Plan locating grid squares in Area 2 ��74 Figure 104: The decorated samian, ��82 Figure 105: The stamped samian, ��85 Figure 106: Illustrated plain samian, scale 1:2��85 Figure 107: Stamped Dressel 20 amphora ��86 Figure 108: The illustrated mortaria��88 Figure 109: Illustrated mortarium No. 2 of the potter GENIALIS ��89 Figure 110: Grey ware mortarium with slag trituration grits (No. 6)��91 Figure 112: A sooted Crambeck mortarium, exterior and interior, illustrated vessel 10��91 Figure 113: Paint-decorated Crambeck Type 7 mortarium, illustrated vessel 13��93 Figure 114: Paint-decorated Crambeck type 7 mortarium, illustrated vessel 14��93 Figure 115: Illustrated fine wares��96 Figure 116: Crambeck parchment ware bowl, illustrated vessel 17��96 Figure 117: Crambeck parchment ware bowl type 5b, illustrated vessel 20 ��96 Figure 118: Crambeck parchment ware bowl type 5b, with painted crux gammata illustrated vessel 22��97 Figure 119: Crambeck parchment ware bowl type 5b with paint decoration, illustrated vessel 23��97 Figure 120: Crambeck parchment ware face pot and smith god pot fragments, left to right No. 25-27��98 Figure 121: Crambeck parchment ware face necked flagon with missing plaque, illustrated vessel 24��98 Figure 122: Images showing the Crambeck parchment ware chicken/cockerel figurine (No. 28)��98 Figure 123: Stamped Parisian ware sherd (No. 29) and rubbing (rubbing )��99 Figure 124: Possible Triple vase single element in white ware fabric (No. 30)��100 Figure 125: Oxidised beaker base showing signs of being re-ground to form a small container (No. 33)��101 Figure 126: Norton reduced wares NORBB1, NORGW1 and NORCG��104 Figure 127: Folded beaker in Norton grey ware (No. 41)��105 Figure 128: Crambeck reduced wares CRGR, CRGR? and CRGRC��107 Figure 129: CRGR Basal sherd showing evidence of re-use (No. 58)��108 Figure 130: Mask-mouthed flagon in Crambeck grey ware (No. 60)��108 Figure 131: Left to Right- No. 61-3 Sherds from face-necked flasks in Crambeck grey ware��109 Figure 132: Face jar in Crambeck grey ware (No. 64)��110 Figure 133: Roundel-stamped face jar in Crambeck grey ware (No. 65)��110

iii

Figure 134: Face pot fragments in Crambeck grey ware (left to right, 66-69)��111 Figure 135: Smith god pot sherds in Crambeck grey ware with applied tools and other motifs��112 Figure 136: Applied wheels from Smith god and/or wheel god pots��113 Figure 137: Crambeck grey ware sherd with applied snake (No. 82)��114 Figure 138: Various stamped and applied partial motifsR��115 Figure 139: Holme-on-Spalding-Moor grey ware spindle whorl (No. 90)��116 Figure 140: Burnished sherd with applied pellet decoration (No. 94)��116 Figure 141: Illustrated calcite-gritted wares ��124 Figure 142: Calcite-gritted jar from bustum cremation pit 475 (No. 108)��125 Figure 143: Interior of calcite-gritted jar showing thick white deposit (No. 120)��125 Figure 144: Proto-Huntcliff jar showing method of construction��125 Figure 145: Illustrated calcite-gritted jars��126 Figure 146: Calcite-gritted other forms��127 Figure 147: Calcite-gritted ware miniature jar (No. 136)��127 Figure 148: Calcite-gritted ware body sherd cut down to a spindle whorl (No. 137)��127 Figure 149: Calcite-gritted baking plate (No. 138)��128 Figure 150: Fragment of portable oven in calcite-gritted ware (No. 140)��129 Figure 151: Fragment of portable oven in calcite-gritted ware (No. 141)��129 Figure 152: Fragment of portable oven No. 143 in calcite-gritted ware showing handmade manufacture technique��130 Figure 153: Fragment of portable oven in calcite-gritted ware No. 144��130 Figure 154: Calcite-gritted ovens, large jars and a tray��131 Figure 155: Large calcite-gritted jar with internal finger pressed decoration jar (No. 147)��131 Figure 156: Vessels from Bustum burial 475��134 Figure 157: Graphical summary of Brooklyn House coins��148 Figure 158: Counterfeit denarii of Trajan, Elagabalus and Orbiana (obverses only). ��148 Figure 159: Graphical summary of Shiptonthorpe (1985-91) coins��149 Figure 160: Graphical summary of Malton vicus (1968-70) coins��150 Figure 161: Graphical summary of Malton vicus (1949-52) coins (regular and copies largely undifferentiated)��150 Figure 162: Personal ornaments. Copper alloy (Nos 1-3) and iron (No. 4) brooches and bone hairpins (Nos 6-7)��160 Figure 163: Personal ornaments Bone hairpins. ��161 Figure 164: Personal ornaments ��162 Figure 165: Personal ornaments and toilet equipment. ��163 Figure 166: Textile equipment��164 Figure 167: Household equipment��167 Figure 168: Household equipment��168 Figure 169: Recreation and writing equipment��169 Figure 170: Writing equipment. Iron styli. ��170 Figure 171: Writing equipment. Iron styli. ��171 Figure 172: Transport equipment. Iron bits. ��172 Figure 173: Structural ironwork��174 Figure 174: Iron knives ��176 Figure 175: Iron tongs��177 Figure 176: Iron chisels��178 Figure 177: Fittings��180 Figure 178: Agricultural and military equipment��182 Figure 179: Weapons. Iron bolts (Nos 122-123) and arrowhead (No. 124)��184 Figure 180: Miscellaneous items��185 Table: 22: Distribution of the catalogued metal items through time��186 Figure 181: A selection of the pyre goods. ��188 Figure 182: Reconstruction of the belt��191 Figure 183: Belt set. Copper alloy buckle and plates 192 Figure 184: Belt set and baldric fittings. Copper alloy plates with frogs (Nos 3-4) and baldric fittings (Nos 7-8). ��193 Figure 185: Other fittings from pyre��195 Figure 186: Other fittings from pyre. Bone (No. 17), copper alloy (Nos 18-20) and copper alloy with iron (No. 21)��197 Figure 187: Curving tiles with vents from context 758��201 Figure 189: Chimney or finial fragments from contexts 202 and 236��202 Figure 188: Chimney sherd from context 206��202 Figure 191: Finger drawn keying on a box flue from context 4023��203 Figure 190: Combed box flues from contexts 253 and 309��203 Figure 192: Parietalis from context 202��203 Figure 193: Tesserae from context 202��204 Figure 194: Two stone tesserae, RF 6 context 201 and RF 296 context 202��208 Figure 195: Stone roof flag of rectangular shape, context 816��208 Figure 196: Stone roof flag of elongated hexagonal shape, context 207��209 Figure 197: Stone roof flags with a curving upper edge from contexts 202, 220 and 830��209 Figure 198: Roof tiles re-used as lids, RF 194, context 235 and RF 308, context 202��210 Figure 199: Roof tile with scratched surface, possibly re-used as a hone, context 202��210 Figure 201: Carved stone RF 177, context 215��210 Figure 200: Limestone slab with bullnose profile, context 202��210

iv

Figure 202: Carved block RF 96, context 202�� 211 Figure 203: Late Neolithic oblique arrowhead No. 8�� 215 Figure 204: Hand querns with a radial pattern of grooves on the rim �� 225 Figure 205: Find spots of rotary querns excavated in Malton and Norton �� 228 Figure 206: Roman sites with querns in the vicinity of Malton and Norton �� 229 Figure 207: Fragment of a probable small millstone No. 15, RF 452, context 3117�� 233 Figure 208: Frequency of cattle, caprine and pig remains (NISP) as a percentage of total fragments for Phases (2 -5)�� 255 Figure 209: Roman Norton showing proposed revised cemetery areas along Langton Road�� 262

List of Tables Table 1: Quantification of samian by vessel class in Rim EVE and Maximum Vessel Number (MVN)�� 79 Table 2: Quantification of samian by Fabric�� 79 Table 3: Quantity of samian by stratigraphic phase�� 79 Table 4: Vessel Form and Function by Fabric Group�� 80 Table 5: MOCR Crambeck white ware mortaria by form type�� 90 Table 6: MOCRF Crambeck Parchment ware mortaria by form type�� 92 Table 7: CRPA Crambeck Parchment ware other forms by type�� 95 Table 8: NORGW1 forms as proportion of fabric (Norton types follow Hayes and Whitley 1950)�� 103 Table 9: CRGR forms as proportion of fabric (Crambeck types follow Corder 1937)�� 106 Table 10: Relative proportion of Norton, Crambeck grey wares and other reduced wares�� 118 Table 11: Mean sherd weight (g) of reduced wares and calcite-gritted wares by phase�� 118 Table 12: Calcite-gritted jar forms by % of Rim equivalents by phase�� 120 Table 13: Relative proportion of calcite-gritted wares and reduced wares as proportion of assemblage by phase�� 121 Table 14: Summary of selected excavated assemblages, the relative numbers of decorated and stamped vesselsy �� 139 Table 15: Quantity (by sherd count) of samian from recent excavations in Norton�� 139 Table 16: The Relative Frequency of Samian Form/Functional Categories at different site types,�� 140 Table 17: Crambeck Parchment wares as proportion of the assemblage by phase�� 141 Table 18: Summary of coin finds from Brooklyn House, Norton (2015-16)�� 151 Table 19: Summary of coin finds from Brooklyn House and Malton Vicus (1968-70)�� 151 Table 20: The catalogued small finds by site phase and functional category. �� 156 Table 21. Types of iron nails present according to Roman phases)�� 173 Table 22: Distribution of the catalogued metal items through time�� 186 Table 23: Ceramic Building Material by form in relation to period�� 200 Table 24: Ceramic Building Material in relation to phase�� 205 Table 25: Stone Building Material by form and stone type �� 207 Table 26: Stone Building Material from Phase 3 structures�� 212 Table 27: Tooling and burning on the Stone Building Material by structure�� 213 Table 28: Lithics submitted finds composition summary�� 214 Table 29: Lithics raw material types�� 214 Table 30: Lithic reduction technology �� 215 Table 31: Burnt lithics�� 216 Table 32: Chronological indicators�� 216 Table 33: Modified formal and non-formal tool forms�� 217 Table 34: Saddle Querns, Rubbers and Polishers�� 224 Table 35: Hand Querns �� 225 Table 36: Powered Millstones�� 226 Table 37: Phasing of the Querns�� 226 Table 38: Querns from excavations in Malton and Norton (from YQS Records)�� 227 Table 39: Roman sites in the vicinity of Malton and Norton�� 228 Table 40: Comparison of Querns from the Study Area with YQS Quern Type Summaries�� 230 Table 41: Whetstones�� 236 Table 42: Summary of osteological and palaeopathological results�� 237 Table 43: Summary of cremated bone assemblage�� 238 Table 44: Summary of cremated bone fragment size�� 239 Table 45: Summary of identifiable elements in the cremation burials�� 239 Table 46: Hand-collected vertebrate remains (NISP – number of identified specimens) from Areas 3, 4, 5, 7, 8 and 9�� 249 Table 47: Hand-collected vertebrate remains (NISP – number of identified specimens) from Phases 1, 2, 3, 4 and 5�� 250 Table 48: Tooth-wear stages recorded for sheep and sheep/goat mandibles. �� 251 Table 49: Tooth-wear stages recorded for pig mandibles. �� 251 Table 50: Number of sheep and sheep/goat mandibles�� 252 Table 51: Number of pig mandibles that could be assigned to the various simplified age categories�� 252 Table 52: Metrical data (following von den Dreisch 1976) for domestic mammals, by element�� 254 Table 53: Metrical data (following von den Dreisch 1976) for birds, by element�� 255 Table 54: Estimated withers heights in centimetres by context and phase �� 255

v

vi

Acknowledgements With the various investigations into the archaeological remains at Brooklyn House occurring over a period of six years there have been many groups and individuals who have contributed to the project and the production of the volume you now have in front of you. The editors/producers of this volume would like the express their heartfelt thanks to the following people and organisations who at one time or another have contributed to the overall success of the project. John Buglass would, in particular, like to give special thanks to Richard Shields and John Lee of North Yorkshire County Council for the overall funding and client liaison; to Andrew Stamper and Paul Cockerill of Jacobs North Yorkshire for initiating the original project back in 2014. The support and ‘critical friendship’ offered by successive North Yorkshire County Archaeological Officers, Lucie Hawkins and Peter Rowe, helped immensely in focussing the project as it developed. During the investigations the on-site team, whose sterling work was skilfully led by Janet Phillips, consisted of Nathan Berry; Richard Coates; Angie Fawcett; Emma Samuels (with a mention in dispatches for her year long watching brief); Chris Scurfield; Liv Sharrad; Angie Walker; Jessica Watters and Maddy Walsh, along with James Lyall of Geophiz Biz who undertook the geophysical survey. Particular thanks go to the volunteers: Pattie Birch; John Bugden; Jo Ferries; Anne Shirling and Brian Smith from the Conisbrough Research and Archaeology Group who gave up their time to come along and join in the digging. The site work would not have been possible without the freely given assistance from the various contractors: in particular Mark Creggan and his team from Simpsons of York and Dave McCorrie and his team from Hobson Porter. Sophie Tibbles of East Riding Archaeology was responsible for overseeing off-site finds processing and the production of specialist assessment reports to a uniformly high standard. In this she was assisted by Karen Adams, Lisa Johnson, Janet Phillips and Sue Wilson. Janet Phillips undertook the stratigraphic analysis and produced a draft site text. We are grateful to all of the Specialists who have contributed to this report and also those others who undertook elements of the Assessment: Tony Benfield, Joanna Bird, Richard Brickstock, Spencer Carter, John Carrott, Hilary Cool, John Cruse, Chris Cumberpatch, David Dungworth, Charlotte England, Hugh Fiske, Alison Foster, Stephen Greep, Kay Hartley, Dave Heslop, Malin Holst, Deborah Jaques, Katie Keefe, Jane McComish, Jo Mills, Ian Rowlandson, Jane Sheppard, Sophie Tibbles, Blaise Vyner, David Williams and Charlotte Wilkinson. Particular thanks to Ian Panter of YAT for the many and varied conservation and technical analysis of the more significant finds. The project also benefitted from the expertise of various colleagues who willingly gave of their knowledge, including Mike Haken of the Roman Roads Research Association and Professor Dominic Powlesland of the Landscape Research Centre. Dr Tim Horsley kindly provided a copy of his report on the geophysical surveys he undertook on the site of Malton fort in 2007. David Snowden, of Malton Museum, kindly solved the ‘Lal Qila 1960’ mystery. John Buglass, Janet Phillips, Sophie Tibbles and Pete Wilson

vii

viii

Chapter 1

Introduction Pete Wilson

The site consisted of the access road and external areas of the former Brooklyn House Youth Centre located on Langton Road, Norton-on-Derwent (hereafter Norton), including the associated playing fields (NGR SE 7930 7086). Redevelopment as a primary school necessitated a programme of non-invasive work, followed by excavations undertaken between 2015 and 2017. Norton is located on the south bank of the River Derwent opposite Malton and its Roman fort. Roman Malton was, or more probably, Roman Malton and Norton together were known as Delgovicia (Drake 1736: map following p. 36; Creighton 1988: 403; Wilson 2017; 2019), rather than Derventio which was the accepted identification for much of the 19th and 20th centuries (see for example: Rivet and Smith 1979: 333-334).

Topography and Land-use The site is generally level and is generally at a height of around c. 25 m AOD, although to the west it slopes down towards the course of the Mill Beck. Prior to the excavations the areas investigated were in use as a playing field, car-parking area and access road associated with the former Youth Club. Archaeological Context of Project The location of occupation in Malton/Norton in all periods is, in large part, dictated by the need to utilise one of the relatively few potential crossing points on this part of the River Derwent (Figure 1). The Derwent originates on Fylingdales Moor some 6 km from the coast, but flows south and then south-westwards away from the coast to join the River Ouse at Barmby on the Marsh. The eastern part of the Vale of Pickering offers few north-south routes east of Malton/Norton, which is located west of the confluence of the Derwent with its major tributary the River Rye and above Kirkham Gorge which obstructs crossings to the west. Although the exact location of the Roman period river crossing is not known it is assumed to lie close to the present river bridge linking the two towns. That said Robinson identifies two possible locations for fords upstream of the island (1978: fig 6c, 34 no. 231, 40 no. 386; see Figure 2), although the latter location, no. 386, is based on an analysis of the medieval and later street pattern, rather than any Roman-period associations. Despite the lack of physical evidence, and even allowing for changes to the river as a result of post-Roman management of the River Derwent and its catchment, the existence of a Roman-period bridge at Delgovicia would be a reasonable expectation.

The programme of archaeological works included trial trenching and excavation and was undertaken by JB Archaeological Services on behalf of the construction contractors, variously Jacobs and Simpsons of York, working on behalf of North Yorkshire County Council. Previous work at Brooklyn House had included an evaluation in 2002 on the line of a new access off Langton Road (MAP Archaeological Consultancy Ltd 2002) and an extensive geophysical survey of the playing fields occupying the southern part of the site (Lyall 2014). BACKGROUND Geology and Soils The underlying geology of the site and the surrounding area are the Ampthill and Corallian clays of the Upper Jurassic (British Geological Survey 2001). Overlying this, the quaternary geology is one of lacustrine clays, silts and sands (British Geological Survey 1977). The soils which have developed from these deposits have been classified as the Landbeach association which is a permeable, calcareous loamy soil (Soil Survey of England and Wales 1983). While, within the areas investigated, the geology does not seem to have influenced the development of the archaeology as recorded, it did provide reasonably favourable conditions for geophysical survey (Figure 5).

In the Roman period the river crossing(s), whatever their form, would have been dominated by the fort located in Orchard Field and extending to the west under what became the site of Malton Castle and a (now demolished) Jacobean mansion that replaced it (Hudleston 1962: 117). To the south of the river a number of Roman roads are known, or suggested to extend to the south and east 1

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

N

Binchester

Piercebridge

Greta Bridge

r ive

es

Te

R Scotch Corner Catterick Bainbridge

Scarborough Filey

Malton / Norton Boroughbridge

Bridlington

York

Elslack

Stamford Bridge Hayton

Ilkley

e

erwent

Air

River D

use

er O

er

Riv

Tadcaster Riv

Hull Brough

Castleford

ent

r Tr

Templeborough

Rive

Manchester

Chesterfield

Key

Roman Roads (known)

Roman Roads (possible)

0

20km

Figure 1: The Location of Malton and Norton

(Margary 1973: roads 29, 81a, 812, 813, 816 – reassessed in RRRA Gazetteer of Roman Roads). Langton Road, approximates to the course of Margary’s Road 81a, which originates at York. About 3.5 km south of Malton/ Norton, Road 81a is joined by Margary’s Road 29 which originates at Brough-on-Humber (Roman Petuaria), the crossing point of the Humber for traffic from Lincoln (Lindum). Thus Langton Road was the major approach to Roman Malton/Norton from the south.

substantial excavations in and around the Roman fort from 1927-1930 (Corder 1930). As part of the 1920s campaign Dr J.L. Kirk excavated trial pits in the civilian settlement in Orchard Field and the grounds of The Lodge west of the fort (Corder 1930: fig. 52), although the report as published was focussed on the fort and its defences. In 1949-1952 Rev. Derek Smith undertook excavations in advance of development within the civilian settlement outside the south-east gate of the fort (Mitchelson 1964). It was these excavations that revealed the building known as the ‘Town House’ (Mitchelson 1964: fig. 3, plates VI-VIII), with its mosaic floor (Mitchelson

Until the advent of developer-funded archaeology the bulk of major archaeological effort was concentrated on the Malton side of the river (Figure 2) starting with 2

Pete Wilson: Introduction

N

Jack Berry House

New Rugby Club Old Maltongate

University of York excavations

Fort

St Leonards

Wenham 1968-70

Orchard Field 1949-52 Malt

went

e

ridg

on B

River Der

Church Street

ercial

Comm

Street

27 Wood Street

St

eet

Ni

d Str

as

Mill Str

ol

ch

Woo

Bev

erle

t

tree er S

et

St P

Welham Road

Grove Bungalow

yR

oad

Howe Road

Langton

Brooklyn House

Eastfields

eet

et

re

St

Norton Community School 2012

St Peter's St. St Nicholas St.

Road

Roman roads

The Ridings

500m

0

Lal Quila / 98 Langton Road

1:10,000

Figure 2: Roman sites in Malton and Norton

1964: plate IX; Neal and Cosh 2002, 344-346), as well as other buildings. From 1968-1970 further excavations in Orchard Field investigated a number of other buildings within the civilian settlement which it is suggested may have been furnished with defences and, possibly included a mansio in the 4th century (Wenham and Heywood 1997: 36-38). Dr Tim Horsley undertook earth resistance and magnetometer surveys on the eastern part of the fort and northernmost part of the civilian settlement within Orchard Field and within the adjacent walled garden. This work revealed considerable evidence relating to the stone buildings belonging to the latest phases of the fort’s occupation,

including part of the principia and other buildings in the central range, as well as barracks and other structures in both the praetentura (front) and retentura (rear) of the fort (Horsley unpublished). The barracks, as plotted (Horsley unpublished, figures 5. 13, 13, 28) appeared to include conventional cavalry barracks similar to those at Wallsend (Hodgson 2003, 37-121) in, at least, the praetentura, along with possible ‘chalet style’ barrack in both parts of the fort. See Daniels (1980) and Rushworth (2009, 114-133, 149-71; 2016, 468-520) for discussions of ‘chalet’ barracks on Hadrian’s Wall, which at Wallsend Rushworth (2016, 516-520) unequivocally interprets as cavalry barracks. 3

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire Earlier discoveries in Malton, and also Norton, have been brought together in a number of publications. These include: A Gazetteer of Roman Remain in East Yorkshire (Kitson Clark 1932: 99-108 – Malton, 113-118 – Norton) and The Archaeology of Malton and Norton (Robinson 1978: 5-12, 24-29 – Malton, 34-40 – Norton), referred to above. More recent small-scale developer-funded projects have examined a number of sites including within the civilian settlement in Orchard Field where 1st-century post-holes and stone buildings of 2nd- to 4th-century date have been found (MAP 1993: 32-34). South-west of Orchard Field test-pits on Sheepfoot Hill, close to the location of the 1949-52 excavations, revealed evidence of 3rd-century timber buildings and 4th-century stone structures (Finney 1990-91). However, the major fact that emerged from a review of the data available in 2006 is that there is an intensity of occupation apparent in areas other than those recognised at the time of Robinson’s survey (Wilson 2006). By that time there was evidence of occupation in Old Maltongate outside the north-west gate (or porta decumana) of the fort (MAP 1998a); St Leonard’s Churchyard west of the fort (YAT 1992); Sheepfoot Hill/ King’s Mill (Finney 1990-91; MAP undated) and the area of Castle Howard Road where Roman period boundary features have been noted (Stephens 1992a; 1992b; MAP 1996). What is not clear from most of these small-scale interventions is the character and extent of the deposits encountered, although St Leonard’s Churchyard has produced Roman-period stone-built structures of 3rdand 4th-century date with evidence of painted plaster amongst the demolition debris (YAT 1992). This site extends the known area of reasonably well-appointed civilian occupation into what was previously a ‘blank area’.

processes and burial activity (Wilson 2006: 43-45, fig. 3). Most unusual is the evidence for a goldsmith’s shop in the form of an inscription referring to its operation by a slave (RIB 712). The most substantial body of evidence relating to Roman-period industry in Norton is that for pottery production, with evidence for up to at least 25 kilns known along with additional material such as kiln fire bars (Swan 1984: 109-11, Bidwell and Croom 1997: 101103; Wilson 2006: 43-44; Stephens and Ware 2012: fig. 2). Many of the discoveries are the result of ‘casual observations’ but rescue excavations in 1948-1949 focussed on Howe Road on the Model Farm Estate, some 500 m east of Brooklyn House, revealed a group of 8 kilns, including two that intercut (nos 5 and 5a) and a T-shaped structure, similar to a ‘corn-drier’ or malting oven, which was suggested as a possible ‘drying floor’ for unfired pots (Hayes and Whitley 1950). Subsequent to the main excavations a further well-preserved kiln was located at Grove Bungalow, Langton Road (Hayes 1988: 72-76), some 250 m east of Brooklyn House. Citing information from Philip Corder, Kitson Clark (1932: 115) records a kiln ‘discovered in Wall Street, in 1862, in the foundations of Mr Jackson’s House on the road to Langton’. Wall Street, may have been a mistake for Wold Street, although Robinson (1978: 37) notes that most construction activity in 1862 was in Langton Road, not Wold Street, and places his gazetteer number (311) close to Brooklyn House, but there is no evidence to support that location as the find spot. Although the possibility of late 2nd-century (Swan 1984: 109), or even 1st- to 2nd-century (Bidwell and Croom 1997: 10-101, 110), pottery production at Norton had been suggested, no kilns pre-dating the 3rd century were known prior to 2012. The discovery of a kiln at Norton Community Primary School, c. 400 m northeast of Brooklyn House in 2012 pushed the evidence for pottery production back into the late 2nd to early 3rd century (Stephens and Ware 2012). In 1953 Hayes’ Grove Bungalow kiln was reopened and archaeomagnetic dating demonstrated that it was use in late 3rd to early 4th century (Cook and Belshé 1958; Hayes 1988: 77). The Norton Community Primary School and Grove Bungalow kilns demonstrate that the Norton Industry was active for in excess of a century, and possibly for at least 150 years, given that Evans (1988: 324) proposes the cessation of production at the Norton kilns ‘by the mid-4th century’.

Other development-led projects to the north-east of the fort have added evidence of substantial 2nd- to 3rdcentury buildings, apparently lining the road towards Old Malton at Malton New Rugby Club (Stephens and Ware 1995: 12-14), as well as early military occupation and later agricultural use at Jack Berry House (Burn et al. 2017). While outside the north-east gate of the fort excavations by the University of York have demonstrated the presence of major civilian stone buildings and occupation extending, probably, into the 5th century, as well as craft or industrial activity (Collins 2019: 424). Roman Norton (Figure 3), prior to the advent of developer funded archaeology, received much less formally organised archaeological attention than Malton. Despite that Robinson was able to list 147 discoveries and find spots (Robinson 1978: 34-40, nos 231-377). As was the case in Malton, many of the early discoveries were only recorded in the pages of the Malton Messenger, a source used by both Kitson Clark (1932) and Robinson. Of particular note is the cumulative evidence relating to both industrial/craft

Other industries also existed in Norton, including ironworking evidenced by a furnace, possibly for ironsmithing/smelting, found inserted into the mortar floor of a stone-built building (Robinson 1978: 35, no. 264; Hayes 1988: 86-88 – for the building see below). Other crafts industries were no doubt present in Norton (and Malton), with the discovery of ‘several fragments 4

lton

line

Ma

Railway

River D erwent

Bri

dge

ol

ch

Ni

Springfield Garth 1991

St tre et

as S

Street

R386

Brooklyn House

St

et

Welham Road

5 Langton 98 Langton Road

Road Lal Qila

Bright Steels 1994

Norton Community School Kiln 2012

Grove Bungalow 1949-50

St Peter's Vicarage Garden 1950

Sutton Street 2014

eet

d Str

Woo

27 Wood St.

treet

ial S

merc

Com

Mill S treet

Howe Road Kilns

oad

1998-9

yR

Bev erle

Figure 3: The Brooklyn House area – sites and find spots

Stre Peter

St Peter's Street St Nicholas Street 1951

Church

R231

0

R386

R231

1:5000

Roman roads

250m

Robinson's possible river crossings

Langton Road Cemetary

Wood Street / Langton Road Cemetary

Beverley Road / Mill Street Cemetary

Eastfield 1946-7

Other craft/industry

Kiln

Other Roman period find spot

Inhumation / cremation

Cremation

Inhumation

Structural evidence / paving

N

Pete Wilson: Introduction

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire of tesserae’ ‘near the Wesleyan Chapel, Commercial Street’ (Robinson 1978: 35, no. 247) providing further evidence of mosaic or tessellated pavements in Delgovicia and, perhaps, the existence of a locally-based mosaic workshop.

west side of Langton Road (David Snowden, Malton Museum, pers. comm.). 3.

Burials are the second category of data well-represented in Norton, but poorly understood due to many of the discoveries being Antiquarian finds or records, or more recent observations during building works and service trench excavations (Wilson 2006: 44-45). Again we are indebted to Kitson Clark (1932: 113-118) and Robinson (1978: 34-40) for their collation of antiquarian and other finds. In Howe Road, 500 m east of Brooklyn House, two inhumations were cut into the area of the kilns excavated by Hayes and Whitley (Hayes and Whitley 1950: 11), with a further inhumation recorded some 75 m south-west of the excavated area (Robinson 1978: 39, no. 353) and a skull reported from The Grove (Robinson 1978: 39, no. 352), some 50 m further west towards Brooklyn House.

Further cremation and inhumation burials are know from the Malton side of the river including discoveries resulting from the construction of, or located close to, the former Malton to Thirsk railway close to the fort (Robinson 1978: 26, nos 67, 69-75), including a gypsum burial and a stone sarcophagus. In addition there are various reports of the discovery of ‘urns’ and inhumation from various locations across Malton (Robinson 1978: 24-29), including at least 29 infant burials within the fort (Corder 1930: 32, 67). Although the sample is small, the gypsum burial and sarcophagus, along with the discovery of the tombstone of a former trooper of the Imperial Household Cavalry (RIB 714) at Pye Pits outside the north-west gate of the fort, could indicate a bias towards more prestigious burials being located on the Malton side of the river.

In the broader Wold Street/Langton Road area a number of finds of inhumations, cremations, a tombstone naming one Aurelius (RIB 715) and a lidless stone sarcophagus, indicate the existence of an extensive cemetery extending for over 500 m (Robinson 1978: 3537, nos. 270, 271, 306-10, 316). Similarly, in the Wood Street/Beverley Road area various finds of inhumations and cremations suggest another extensive cemetery (Robinson 1978: 36-39, nos 275, 276, 278, 279, 285-88). More recently sewer works in Sutton Street revealed a late 1st- to 2nd-century female cremation burial (NAA 2014: 7). To the north, inhumation and cremation burials have been found either side of Commercial Street over a distance of some 200 m (Robinson 1978: 35, nos 245, 248-251).

The third element of the known archaeology of Norton, the form of the settlement itself, is perhaps the most difficult to draw together as much of the evidence is derived from antiquarian records, casual observations and ‘watching briefs’ (Figure 3). Within Norton Robinson (1978: 354-40) records at least 22 discoveries of structures or floors and 19 where paving is recorded, but in many cases the information is limited. In 1946-47 housing development in the Eastfields area of Norton impacted on extensive areas of Roman-period occupation. By working around the builders’ activities, Hayes (1988: 66-72)was able to record late 4th-century paving and at least two walls were found in Areas A and B, with further areas of paving, quantities pottery and other finds from a further 11 areas. The paving in Area A and the western part of Area B was well-laid and suggestive of an interior surface, with a less well-laid area to the east representing a possible yard. Nowhere were the excavators able to examine underlying stratigraphy. Further recording during construction works at Eastfields in 1998-99 revealed further worn stone surfaces indicative of occupation and dated to the late Roman period by a coin of Valentinian I (A.D. 364375) sealed by the earlier of two surfaces. Underlying the earliest surface was a flue and other deposits that produced pottery dating probably dating to the later 3rd century suggesting that occupation in that part of Roman Norton originated around that time (Anon 1999; Halliday and Sweeny 1999a; 1999b).

Immediately south of the Brooklyn House playing field some 60 inhumation burials and an unknown number of cremations of 2nd- to mid-4th-century date are known from a cemetery west of Langton Road. The evidence for this cemetery largely derives from three discoveries: 1.

an excavation at Leefe’s Bungalow (now known as ‘Sancroft’, 98 Langton Road), located the to the south of The Ridings (contra Stephens and Ware 2012: fig. 2) in 1953 which produced 26 inhumations and four or five cremations (Robinson 1978: 39, no. 354; Hayes 1988: 77-80);

2.

two inhumations were discovered in the garden of Lal Qila adjacent to 98 Langton Road (Malton Gazette and Herald 29th April 1960; Hayes 1988: 7780). ‘Lal Qila’ no longer exists, but was probably located to the north of 98 Langton Road as for many years 98 No. 98 was the last house on the

an apparently enclosed cemetery partially excavated in 1966-67 at The Ridings, which produced some 33 burials, including at least seven cremations (Robinson 1978: 39, no. 355).

Similar paving to that seen at Eastfields was observed near the junction of Wood Street and Wold Street 6

Pete Wilson: Introduction

(Robinson 1978: 35, no. 269; Hayes 1988: 83) and close to the kilns in Howe Road (Hayes 1988: 85). In 1950 a late 4th- to possibly 5th-century paved surface was excavated in St Peter’s Vicarage garden, c. 150 m east of Brooklyn House (Robinson 1978: no. 320). The paving was shown to seal a stone-lined flue, which in combination with the discovery of Hayes and Whitley (1950) type 3 firebars could indicate pottery production on the site (Hayes 1988: 87-88)

1995: 14). More recently a site was excavated at 27 Wood Street, located c. 150 m south-east of the Roman-period river crossing, which produced a sequence of timber and stone buildings dating from the late 1st to early 4th centuries (Burn et al. 2018). Turning to the Brooklyn House site specifically the earliest known reference to archaeological material is on the 25" Ordnance Survey Map (1911). The map records the discovery of ‘Roman coins found previous to A.D. 1851’ in land parcel 301 – the eventual location of Brooklyn House. In 1967 excavations for the foundations of the Youth Club produced:

Between St Peter’s Street and St Nicholas Street wellbuilt masonry walls, consisting of dressed blocks of limestone forming the south-west corner of a building, were found in 1951 c. 200 m north of Brooklyn House. At least two phases of floor were recorded and ‘three or four periods of occupation … from c. 150 A.D. or earlier to c. A.D. 370’, with evidence for iron-working in the form of a heavily burnt flue, associated with iron slag, that had been cut through the buildings’ mortar floor. The building sealed further structures, the earliest incorporated a semi-circular apsidal element which was itself sealed by an ‘fine hard mortar (Robinson suggests opus signinum) floor’, which underlay ‘two feet of ash, refuse and burnt matter’ on which was laid the latest mortar floor (Malton Gazette and Herald – 2nd March 1951). In 1954 a further building was excavated on the site which contained a flue and, close to Langton Road, a kiln (Robinson 1978: 35, no. 264; Hayes 1988: 88, pl. 24).

‘scattered Romano-British and mediaeval potsherds. A sewer trench across the site of the car park on the eastern side of the building cut through a pitched floor covered with stone roofing slates at a depth of 3 ft (c. 0.9 m). An infant burial beneath the floor. Finds: - fragment of the lower stone of a quern, a bone pin, fragments of a knife handle, a grey ware counter, oyster, mussel and limpet shells, sherds of Norton, Crambeck and calcite gritted ware.’ (Radley 1968: 115; Robinson 1978: 37, no. 312). In 2002, in advance of an extension to the car park and the creation of a new access to the Youth Club an archaeological evaluation was undertaken (MAP 2002). The excavation tentatively identified a Roman road, elements of at least two Roman-period structures, two infant burials and noted the presence of large quantities of demolition material, including ceramic roof and box-flue tiles. The latter being taken to suggest the possibility of a high status building with a hypocaust in the area.

Two developer-funded sites in Norton have provided evidence of structures. Work at Bright Steels in 1994 produced evidence for surfaces, probably representing floors, similar to those recorded at Eastfields in 194647, but associated with 1st- to 3rd-century pottery, rather than late Roman material (Stephens and Ware

7

Chapter 2

The excavation Janet Phillips and Pete Wilson INTRODUCTION (Figure 4)

ran from north to south from Norton towards Langton. All of the main features found on the site appear to be orientated in relation to the Roman road found in Area 4. Although the road, as recorded, appears to date to the late 2nd/3rd century it seems probable that a road or trackway pre-dating the engineered Roman road influenced the earliest phases of the field system identified through geophysical survey (Figure 4) and excavation.

The excavations, as noted above, were undertaken over two seasons. In 2015 six trenches were excavated as part of the programme of evaluation. Where discussed the context numbers from this phase of work are distinguished by a ‘TT’ prefix. The main season of mitigation works in 2016 involved the excavation of nine areas, several of which were extended or expanded due to the requirements of the construction programme and, in the case of Area 4, consisted of a series of machine-dug service trenches, located to the north and south of the access road to the site, that were recorded as a single area. Depth limits were imposed on the excavations that reflected the requirements of the construction programme and, as a consequence, the full depth of deposits could not be investigated in some areas. Similarly, the requirements of the construction programme, and issues of access to/ availability of some of the smaller areas, have in places meant that linking stratigraphy and/or features between trenches was problematic in places. The following section presents the stratigraphic sequences as recorded in both the evaluation excavations (Site Code BHN15) and main phase of excavation (Site Code BHN16).

The field system was identified on the geophysical survey (Figure 5) and in trial trenching carried out in 2015 (Figure 4) (JB Archaeological Services 2016). A simple form of agriculture was suggested, possibly dating from the later Iron Age to early Romano-British period, aiding the accumulation of a mixed soil and natural horizon. Into this deposit were cut two phases of grid-shaped enclosures, with the first dated to the 2nd/early 3rd centuries AD, and the 2nd to the 3rd/4th centuries. Both phases appeared to have silted up quickly with little sign of maintenance, suggesting that the land use was quickly altered to accommodate the expanding nearby Roman ‘small town’. Field System During the main excavation phase evidence for the field system aligned on the Roman road was identified in Areas 5 (Figure 6) and 6 (Figure 4). Both areas had seen little disturbance, lying on the periphery of both modern development and the Romano-period site, with topsoil and subsoil levels significantly shallower than those identified in the other areas. Further possibly related features were found in Areas 2 and 3. Due to trench locations and depth restrictions only ditches from the first phase of the field system were identified in the 2016 excavations. Unfortunately, the level of modern disturbance in Area 1 to the north of Brooklyn House obscured any evidence of this phase, so it was impossible to determine if the field system had continued northwards beyond the boundary of the geophysical survey (Lyall 2014). Further depth restrictions meant no features for this phase were identified in Areas 4, 7, 8 and 9.

An extensive programme of environmental sampling was undertaken throughout the fieldwork project. Unfortunately assessment demonstrated that there was little or no potential for useful analyses and therefore, where relevant, reference is made to the environmental assessment results in the text that follows. The full environment assessment is incorporated in the site archive. PHASE 0 – PREHISTORIC The site produced small quantities of lithics dating from the Mesolithic to the Later Bronze Age (Chapter 10) and several sherds of earlier prehistoric pottery (Chapter 3), but no features pre-dating the later Iron Age were identified. PHASE 1 - IRON AGE/ROMANO BRITISH (?1st century AD to 3rd century AD)

Field System Phase 1a – 2nd/3rd century

The earliest activity recorded on site was a phase of agriculture, with some scattered associated features spreading along the western side of a Roman road which

A 10 m length of an east-west aligned ditch, 505 (Figure 7), was visible in the southern part of Area 5 (Figures 8

Janet Phillips and Pete Wilson: The excavation

Figure 4: Overall Phase 1 plan with Trial Trenches

9

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

N

0

scale 1:500 @ A4

25m

Sec. 55 north

501

south

502

510

503 506

505 0

scale 1:20 @ A4

Figure 6: Phase 1 plan of Area 5

1m

10

Figure 5: Plot of Geophysical Survey results. Survey by James Lyall of Geophyz.biz

Janet Phillips and Pete Wilson: The excavation

1:2000 @ A4 A4

N A9

508 A8 A3

A1

A7

507 A6 A2

508

A5

Fig. 8

509

Tree Bole

510

504

503

504

505 506

Fig. 6

Figure 7: Area 5 – Section of Ditch 505 Geophysical features

0

5m scale 1:100 @ A4

11

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire the trench edge, however a maximum width of 2 m was recorded in the northern part of the trench and the ditch had a similar depth and profile to ditch 505. Pottery of the late 2nd century, or later, was recovered from fill 508 along with a fragment of oyster shell. The upper fill, 509, contained pottery dating from the late 3rd to 4th century along with a fragment of burnt clay/ daub. The southern arm of this enclosure, previously identified in the geophysics and in Trial Trench 5, as TT510, was not located in Trench 5, possibly due to the later Victorian sand extraction activity that was noted throughout the site.

Sec. 57 west

501

east

507

0

508

The furthermost north-south aligned ditch identified on the geophysical survey was found in Area 6, ditch 603 (Figures 4, 9). The corner of ditch 603 was visible in three of the pits excavated for structural pads, however only the easternmost edge of the ditch was exposed. A 1.80 m width of the ditch was visible to a depth of 0.72 m, with concave sides and a rounded base.

1m

scale 1:20 @ A4

Figure 8: Area 5 – Section of Ditch 507

4-6). This ditch can be identified on the geophysical survey (Figure 5) as forming the northern arm of an enclosure within the larger field system. It was 0.81 m wide and 0.48 m deep, with a U-shaped base.

Area 2 (Figure 10) Within Area 2 the level of ground reduction was largely limited to 24.48 m OD, meaning the natural sand was not fully exposed across the site. Unfortunately, this meant the field system ditches were not encountered during the main excavation and therefore not investigated further. It has, however, been possible to tentatively associate some further features on the same alignment with this phase. The finds recovered from ditch/feature 287 suggest it belongs to Phase 1a.

A 9.50 m length of a curving north-south ditch, 507, appeared to form the northern and western sides of a smaller enclosure seen in the geophysical survey (Figures 4, 5, 8). This ditch is a continuation of features recorded in Trial Trench 4 (TT407), and Trial Trench 5 (TT512, TT516) in 2015. Ditch 507 lay partially beyond

Sec.94

east

north

south

601

604

603

602

0

scale 1:20 @ A4

1m

Figure 9: Area 6 – Section of Ditch 603

12

13

298

Fig. 14

2005

Trial Trench

Fig. 13

Brooklyn House

Fig. 15

A5

A1

A2

A7

A8 A3

A9

A4

0

Geophysical Feature

Figure 10: Area 2 – Phase 1 plan

Service Trench

A6

1:2000 @ A4

scale 1:200 @ A4

Fig. 11

287

10m

N

Janet Phillips and Pete Wilson: The excavation

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Sec. 21 24.20 m O.D. north

south

293 288

294

292 290

nat

291 289

Figure 11: Area 2 – Feature 287

287

0

1m

scale 1:20 @ A4

Figure 12: Area 2 – Feature 287 (view to west)

Feature 287 was an east-west aligned elongated pit or ditch terminal (Figures 4, 11, 12 – 279 and 294) which extended for 2 m into the excavated area. In profile 287 was steep-sided and roughly U-shaped, was 1.92 m wide at the top and 0.94 m deep. Due to the steep sides of the feature, and the soft nature of the underlying geology of sand and gravel (203) significant slumping deposits were identified. This included 290, the primary fill on the southern edge, which seems to have slumped soon after the feature was opened. Pottery dating to the late 2nd to 3rd century was recovered from 288 higher in the sequence of fills, along with two fragment of large

mammal bone. Environmental sampling of 288 produced bone fragments, charcoal, ash and mortar crumbs, and spheroids of vitreous slag (Sample 12). The snails recovered, Vallonia ?excentrica Sterki, Pupilla muscorum (L.), Helicella itala and Helicidae sp., suggest the area was open grassland consistent with agricultural usage. The upper fill, 293, was considerably less mixed than the earlier deposits and also contained pottery dating to the 2nd to 3rd centuries, an oyster shell fragment, ten fragments of animal bone including mammal and chicken, stone roof tile, building stone and a fragment of imbrex. 14

Janet Phillips and Pete Wilson: The excavation

Sec. 26

Sec. 27 west

west

east

201

24.25 m O.D. 2004 nat

east

201

nat 24.20 m O.D.

nat 298

nat

299

2003 201 0

2001

1m

scale 1:20 @ A4

Figure 13: Area 2 – Feature 2003

0

There was no clear evidence of re-cutting this feature suggesting it was not maintained for a long period, consistent with the 2015 trial trench findings for the excavated field system ditches.

scale 1:20 @ A4

Figure 14: Area 2 – Ditch 298

a range of finds including stone roof and floor tile and pottery dating from the 3rd to 5th centuries, although the latest material was presumably intrusive from disturbance associated with later, Phase 5, dumping activity. The horizon also produced Roman brick, box flue tiles, imbrices and tegulae, fired clay/daub fragments, a piece of quern stone and a relatively large amount of oyster shell and some scattered remains of limpets. Animal bone was also recovered in large quantities with a total of 1149 fragments recovered (73 from 218, 726 from 231, 78 from 247 and 41 from 262). A number of the bones displayed evidence of butchery, and the species represented included cattle, sheep, horse, pig, dog, bird/chicken, frog/toad and small vertebrate. However, it should be noted that this was recovered across a large area suggesting general refuse rather than a specific activity/event and some of the material, at least, must be intrusive from later disturbance as suggested above. An environmental sample, Sample 15, was taken from deposit 262 located beneath Structure A (Phase 3) and

It may also be possible to associate a number of other features with Phase 1a/b (Figures 4, 10). They were identified during work on an east-west aligned service trench within Area 2 along the southern edge of Brooklyn House. A ditch terminal or pit, 2003 (Figure 13), and two ditches, 298 and 2005 (Figures 14, 15), were exposed on the same north-south alignment as the field system ditches. Unfortunately, due to their proximity to the existing services the features were not identified on the geophysical survey and could not be traced beyond the service trench edge. Overlying the ditches and features discussed above, and largely masking them, was a homogenous orangebrown, silt-sand deposit 218 (also recorded as 231, 247, 255-257, 262 and 265). This appears to represent a short-lived Romano-British ground level prior to the next phase of activity on site. These deposits produced

Sec. 28 west

east

201

24.23 m O.D. nat

nat

2006

2005

0

1m

1m

scale 1:20 @ A4

15

Figure 15: Area 2 – Ditch 2005

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire was found to contain ash, charcoal, a charred grain, bone fragments and crumbs of mortar.

posthole 3030, was located at the edge of the trench and the second, posthole (3034), was located in the terminal end of the ditch. Further indistinct and faint patches were observed in the base of ditch (3028) which may suggest smaller stakes had been used between these two more substantial posts. (Figures 4, 18-20)

Two similar ground level deposits were identified towards the centre of Area 2, contexts 255 and 256, as well as one towards the south eastern corner (265). As is clear from the discussion above the date of much of the material is questionable. It is possible that some, or possibly many, of the finds relate to occupation associated with the Phase 1 field system which is located outside the areas available for investigation. However, the soft, homogenous nature of the ground, and the presence of unequivocally late Roman pottery, could equally mean the majority of these finds are intrusive having originated in the later Roman period (Phase 5) of the site when it was used as a rubbish dump. The presence of 12th to 14th century Brandsby ware and Coarse Whiteware pottery in layer 231 suggests medieval agriculture further impacted on the stratigraphy.

Cut into probable ground surface 311, 4.25 m to the north of gully 3028, was a similarly aligned east-west ditch (304) (Figures 4, 16, 17). Ditch 304 was U-shaped, 1.48 m wide and 0.62 m deep, and was recorded for 3.50 m before being obscured by a later wall (308). A cut obscured much of the upper fill (3103) making it uncertain if the ditch was maintained for any length of time, however it does appear to have silted up prior to the development of Phase 2 Roman-period ground surface 388. Although there is no direct relationship with the features discussed above, it seems likely that 319/323 represents a short-lived Romano-British ground level. The deposit and finds suggest it accumulated while the field system was in use but prior to any major Roman activity. 319 yielded a fragment of Roman imbrex, nine fragments of large/medium mammal bone, including two identified as cattle and two as sheep/goat, and some stone roof tile. (Figure 17)

Area 3 Several features underlying the main Roman deposits were identified within Area 3. Although dating evidence was minimal, and a direct link with the other field system ditches could not be made, it seems likely these features form a continuation of the Iron Age/ Early Romano-British enclosures.

PHASE 2 – ROMANO-BRITISH (mid-3rd century?) The ditches and features identified in Phase 1 seem to have been in use for a relatively short period of time, in particular those closest to the trackway/Roman road. There was little or no sign of cleaning and maintenance, despite the soft, erodible nature of the natural sand. It is therefore likely that the field system fell out of use or was incorporated into the expanding Roman settlement. During this phase it is probable that the pre-existing trackway became a more formalised route, in the form of an engineered and metalled Roman Road. Third-century pottery was recovered from the core of the road, context 424, suggesting it was a later element of the Roman road system in the area.

In the southern part of Area 3 during machining c. 3 m of an east-west aligned ditch was identified cutting into the natural sand (3000). The U-shaped ditch (3099) (Figures 4, 16, 17) with concave sides, was measured 1.42 m wide at the top and 0.54 m deep with a rounded base. A 0.54 m thick primary fill (3100), a loose, mid orange-brown, silt-sand with frequent natural gravel inclusions was identified in the base. Sealing this was 0.24 m of what appeared to be redeposited natural material (3101), perhaps suggesting a deliberate backfilling event as the ditch was already nearly silted up with fill 3100. Overlying the undisturbed natural (3000) in the northern half of Area 3 was the remnant of a possible early ground level (311) (Figure 17). Cutting this deposit were several features possibly relating to an early field system.

The development of the road, and the location of the site on the fringes of the growing Roman town, led to a phase of burials spreading along the roadside. Roman burials have been recorded on other sites directly to the south, north and east, and an infant burial was found during drainage installations in the 1960s, the approximate location of which lies to the west of Area 3 (Robinson 1978, 37, no. 3120). The 2015 trial trenching identified isolated disarticulated human bone, and during the 2016 programme of works human bone was recovered from Area 2, 4 and 7.

In the northern half of Area 3 was a roughly oval shallow pit, (3032), which was truncated by the terminal of an east-west V-shaped gully (3028). The gully continued west for 2.90 m increasing in size from 0.06 m deep and 0.40 m wide to 0.34 m deep and 1.10 m wide at the trench edge. Its fill (3029) contained some, probably intrusive, pottery dating from the late 3rd to 4th centuries. Within the base of gully 3028 were two postholes. One,

Within the northern extent of Area 3 was a 0.30 m thick deposit, 330/338, overlying the earlier ditches 16

Janet Phillips and Pete Wilson: The excavation

Figure 16: Area 3 Phase 1/2 plan

17

18

3007

340 356

3011

3127

3129

3128

3106

3107

311

3108

330=388

315

3006

353

3005

314

351

396

Services.....

350

3139 3085

0

312 3105

3090

311

330=388

301 302 3104 303

scale 1:50 @ A4

3028

3029

3099

3100

3130 3039 3093

2m

3063

3062

3064

3082

Figure 17: Area 3 – East-facing section

352

3023

3101

3086=3084

333

3109

301 302 3104 313

3081

324

325

3102

356

311

323=319

305

341

304

3103 310

342 367 3007

340 356

311

330=388

3011

330=388

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Janet Phillips and Pete Wilson: The excavation

The northern access road verge trenches in Area 4 were not excavated to as great a depth as those in the southern verge trench and consequently any RomanoBritish ground levels were harder to identify or not reached at all. However, to the east of the Roman road a significant ground level deposit, 477/4022, was identified. It was into 477 that a bustum cremation pit (4) was cut (see pp. 29-34). 477 produced, probably intrusive, pottery dating to the late 3rd to 4th century.

Sec. 45 24.32 m O.D. south west

3033

north east

3032 0

scale 1:20 @ A4

1m

Several possible Phase 2 ground level deposits were identified in Area 9 to the west of the road and to the

Figure 18: Area 3 – Pit 3032

Sec. 46 24.30 m O.D.

south east

3033 311

3032

0

3029

3032

north west

3028

1m

scale 1:20 @ A4

Figure 19: Area 3 – Gully 3028 and Pit 3032

and features associated with the Phase 1 field system. It seems probable that this deposit accumulated as the field system went out of use, and the area came to be used for burial. 388 produced pottery dating to 3rd to late 4th century and was found with Roman brick, imbrices and a small amount of oyster shell. 330 produced, presumably intrusive, Crambeck Parchment ware. Environmental sample (23), from 388, produced crumbs of mortar, abundant cinder, traces of coal, charcoal and vitreous slag, a sphere and fused flake of hammerscale, one indeterminate charred grain fragment and one small vertebrate long bone fragment. (Figure 17)

Sec. 47 24.20 m O.D.

south east

3035

north west

3034 0

scale 1:20 @ A4

1m

Figure 20: Area 3 – Posthole 3034

Although there was no evidence for the early field system in Area 7, there was a similar ground build-up deposit to layer 388, context 727. 727 was identified as disturbed ground, similar to the natural but mixed possibly as a result of agriculture dating to the early field system. (Figures 21-23)

north of Area 3 - contexts 931, 916 and 934. Given the proximity of Area 9 to Area 3 it seems likely these deposits were stratigraphically the same as 388 and formed part of the same ground surface. (Figure 24)

In Area 4, located on the northern and southern verges of the site access road, the Phase 2 activity to the east of the Roman road in the southern verge trench was minimal, appearing to be largely made up of later dumps. As a result of this lack of disturbance three of the lowest deposits have been identified as likely to represent a Romano-British ground level, 412, 442 and 4054, with 412 producing stone roof tile. (Figure 24)

Possible evidence for the north-south aligned Roman road was first uncovered during MAP’s evaluation excavation in 2002 (MAP 2002). However, following the 2016 excavations it would appear the 2002 work had located deposits relating to yard surfaces, or similar remains, rather than the Roman road which was shown to lie 5 m to the east of the location proposed by MAP. The remains of the road camber were clearly visible 19

N16

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Sec. 64 west

705

east

701

702 711

709

703

710

726 709

702

704

703 728

705

Figure 21: Area 7 – South-facing section

727 0

2m

scale 1:50 @ A4

Sec. 65 north

south

702

704 705

703

Modern services

705

713

714

727

727 0

Figure 22: Area 7 – West-facing section

2m

scale 1:50 @ A4

Sec. 66 701

east

west

702

703

Modern service cut

705 727 712

0

708

2m

scale 1:50 @ A4

Figure 23: Area 7 – North facing section

Area 7, south facing trench edge section

under the access road of the existing entrance to the site. Unfortunately, only a narrow area of the road was revealed within three service trenches located in the northern and southern verges of the access road. However, this did provide a full profile and revealed a metalled, compacted surface at least 7 m in width, with possible roadside ditches and evidence of a later ditch, 415, truncating the western side. (Figures 24, 25)

charcoal but nothing else, possibly suggesting the road may have been established on relatively untouched land. Overlying the eastern side of 425 was a firm, red-orange, sand-clay deposit, 443, 2.50 m wide and 0.16 m thick. 443 appears to expand the footprint of the road, and it has been suggested by Mike Haken (Roman Roads Research Association – pers. comm. to J. Buglass) that it may have formed a clay shoulder to allow easier movement of hoofed animals. However, no trace of this deposit was found in the northern verge trenches and it may simply be a further deposit in the overall construction of the road. (Figure 24)

The primary deposit relating to the road was an area of heavily compacted orange-yellow sand with occasional small stone inclusions, (425) (Figure 24). The deposit extended for 6.75 m under the road and had a maximum observed thickness of 0.14 m, although its full depth was not reached. This deposit appears to have been laid as the foundation for the later layers and demonstrates a clear camber. An environmental sample (56) taken from 425 contained a trace of indeterminate rectilinear

On the northern verge a possible small circular feature, 429, was noted within deposit 425. Its shallow depth and irregular nature, along with a lack of similar 20

445 418 424

425

409

401

934 932

901

936

933

919

902

444 443

939

921

937

935

441 440

925

920 926

924

442

919 = 427 927 = 451 930=432=446

901

430

439 408

0

439 427

21

433

434

426 431 446

scale 1:50 @ A4

406=407

423

439

448

427

412

2m

449 447 450

Figure 24: Areas 4 and 9 – South-facing section

4042

4049

401 410

402 439 4054

413 415

411 4041 414 419

409 425

4042

417 418

401

424

439

445

Janet Phillips and Pete Wilson: The excavation

4003

462

22

24

447

Fig.

Modern tarmac road

482 483

6 Fig. 2 484

Fig. 2

9

N

410

415

Nat

461

466

456

419

4015

454 4018

443

455

Figure 25: Area 4 – Phase 2 Roman road and bustum plan

417

camber visible beneath modern road

Projected route of Roman Road

4003

459

457

9

Fig. 2

Sout

Charcoal

e Verg hern

442

Fig. 42

Fig. 44

4022

0

Cu. A

1: 20

Fe.

4044

475

26

420

scale 1:100 @ A4

421

475

. Fig

412

5m

24 Fig.

Fig. 43

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

23

4003

4010

Machine damaged

4001 4002

4011

4004

West

4015

4005

4021

483

482

480 481

486

487 491

4016

492

4018

4003

494

493

495

482

488

4013

0

4019=4011

4020

489

496

4006

scale 1:50 @ A4

497

4053

4004

4001 484

2m

4007 499

498

4000

Figure 26: Area 4 – Access road north verge southern service – south-facing section

4012 4014

484

481

485

4020

4022

4008

4011

4003

4001 4002

4004

475

4009

East

4010

projected line of Bustum cremation pit

4022

4019=4011

4020

4008

Janet Phillips and Pete Wilson: The excavation

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire features suggest it may only represent the removal of a stone or root during clearance of the area prior to the construction of the road.

early phase of the road, also possibly evidenced by ditch 3036 (see p. 27). Layer 424, the 0.20 m thick concreted core, produced pottery dating to the 3rd century and overlay a 5.5 m wide compacted sand deposit, 425. Overlying (424) were two stone metalling deposits, 422 on the eastern side and 418 over the majority of the road. (Figures 24, 28)

The primary excavated deposit relating to the road in the northern access road verge was a 0.38 m thick firm, dark orange-brown, clay-sand, (4003), which extended for 12 m under the width of the road. The eastern part of 4003 was overlain by a 0.20 m deep light brown, claysand (484), while the remainder was overlain by a 7 m wide secondary levelling deposit, 4002, which was also 0.20 m deep. 484 and 4002 probably represent the same deposit separated by a later intrusion 499/4007. (Figure 26)

Stony deposit 422 was lower lying that the main road surface and was located on the eastern edge of the road overlying the clay shoulder, 425. The deposit comprised moderately sized stones of up to 0.25 m diameter in a 0.12 m thick matrix and extended for c. 1 m before encountering deposit 418 which appeared to rise from this point. It is possible the stones forming 422 are the remnant of an earlier phase, or they could have been laid as make-up for the main surface, 418. Unfortunately, the western side of the road was truncated by a later ditch so the construction method there could not be confirmed.

Overlying 4002, the main levelling under the road was a firm reddish-brown compacted sand (4001) 0.26 m deep that extended for 5.25 m beneath the width of the road. However, a deposit recorded under the eastern part of the road, 4005, may represent an additional foundation layer, functionally contemporary with 4001 and also 4000, which overlay the western edge of 4001. As in the southern verge trench these deposits form the base for the road and serve to create its camber. (Figures 26, 27)

The main stony deposit, 418, extended for 5.20 m and formed the majority of the road surface. It comprised of larger stones, up to 0.30 m diameter, in a 0.16 m thick concreted matrix largely consisting of smaller stones. The deposit was truncated to the west by a later ditch, 415, and two possible repair patches were noted. To the east the stonework was significantly different, with the last metre containing flatter and squarer stones appearing to form a type of verge or roadside path before dropping to deposit 422. (Figures 24, 28)

Interestingly at the lowest point of the trench, and underlying 4003, the earliest excavated road foundation deposit, was a firm, stony deposit recorded as a metalled surface (4015), a 0.25 m width of which was visible lying directly on the natural sand (Figure 26). This deposit was c. 0.85 m below the main surface of the road and may represent evidence of an otherwise unobserved

Figure 27: Area 4 – Section through the Roman road (view to north east)

24

Janet Phillips and Pete Wilson: The excavation

clay, and occupied 1.7 m in the central area of the road. The stones appeared to be building material which may suggest reuse from nearby buildings. 417 produced calcite-gritted pottery dated to the 3rd to 4th century, however intrusive medieval/post-medieval brick fragments were also present and were probably derived from the insertion of modern services in the area. The other possible repair deposit, 44, was up to 0.18 m thick and was located over the easternmost 2 m of the road and appeared to significantly raise or define that edge of the road. (Figure 24) The northern verge trenches also appear to have two possible repair patches, 458 and 459, suggesting some longevity of use for the surface. 458 was a 0.05 m thick and 0.75 m long, elongated strip of stonework overlying the main metalled surface 457 on the western edge of the road. 459 was located to the east of 458 and was a 0.50 m wide strip of limestone, which also overlay 457. A 0.06 m thick dark grey-brown, silt-clay-sand, 4009 (Figure 25) and 409 (Figure 24), northern and southern verges respectively, was identified overlying the road surface. It is possible these deposits represent buildup over the surface prior to the road going out of use. Deposit 409 produced pottery spot-dated to the late 4th to early 5th centuries, which suggests the road may have remained in use into, if not beyond, the late Roman period. Also present were oyster and mussel shell fragments, as well as Roman brick and imbrex fragments. However, intrusive Reduced Greenware pottery dating from early 14th to 15th centuries and Redware from the 16th to 17th centuries, were also recovered from 409. This later pottery is most likely to be intrusive and result from the significant disturbance from modern services in the area. However, it could possibly suggest continuing use of the road into the medieval period, although this is unproven.

Figure 28: Area 4 – Road metalling 418 (view to west)

Overlying levelling deposit 4001 in the northern verge was a 0.15 m thick spread of concreted cream stonysand measuring 4.6 m wide and forming a shallow domed agger, 4004. Sealing this was a 0.04 m thick lens of iron panning, 473, that appears to have built up beneath the upper stones, unable to permeate through 4004. (Figure 29) The road surface, 456/457/474, was up to 0.30 m thick and took the form of a concreted stony deposit including limestone pieces, fragments of stone roof tile and crushed limestone. Larger pieces of stone, up to 0.30 m x 0.14 m in size, were evident on the western side of the road, perhaps indicating repair work. However, no clear edge against the main road surface could be identified to confirm that suggestion. Late 3rd to 4th century pottery was recovered from 457 along with fragments of Roman brick. (Figure 29)

A possible roadside ditch, 420, unfortunately visible only in plan in the base of the trench, was identified 6.50 m to the east of the road in the southern access road verge. 420 appeared to run parallel with the road. However, it could not be further investigated as the trench was at its maximum permitted depth. 420 was 1.30 m wide and the surface of its fill, 421, produced pottery dating to the 3rd century. This feature was not identified in the northern verge due to depth differences and later deposits masking earlier features. (Figure 25)

Possible repair patches were identified in both exposed areas of the road, 417 and 445. In the southern verge trench two distinct patches of stonework overlay the main road deposit, 418. These could represent repair work or be the result of wheel ruts, however they may also represent an attempt to redefine the road at a later date following its truncation on the western side.

On the northern verge, overlying the initial levelling deposit, 484, was layer 482/ 4027/4040 which tipped westwards. It was 0.40 m deep, extended for 6.6 m and appeared to represent a gradual accumulation following the construction of the road and prior to any repairs or modifications. An intrusive sherd of Staxton/ Potter Brompton type pottery, dating to the early/mid-

Deposit 417 contained stones with a maximum diameter of 0.21 m, set into a 0.10 m thick grey-brown 25

West

26

415

4050

403

natural

466

4004

464

0

473

scale 1:50 @ A4

4004

2m

454

403 4051

455

4052

462

403

East

Figure 29: Area 4 – Access road north verge northern service trench – south-facing section

4001

402

474=456

463

402 415

4050

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Janet Phillips and Pete Wilson: The excavation

Figure 30: Area 3 – Roadside ditch 397 (view to north)

Figure 31: Area 3 – Ditches 397, 3024 and 3036 (view to north)

13th to early 14th century, was recovered from 4040. (Figure 26)

alignment it seems reasonable to assume the feature could represent an early phase of roadside ditch.

A large probable roadside ditch, 397, aligned north-east to south-west was identified crossing Area 3 (Figure 30), with maintenance evidenced by various re-cutting events. The ditch appears to cut through the ground surface deposit 388 suggesting it originated in the latter part of this phase. However, a small fragment of a possible earlier ditch, 3036, was recorded in section in the middle of Area 3 (Figure 31). Ditch 3036 could belong to Phase 1 however not enough of the feature was identified to confirm this and no finds were recovered (see metalled surface 4015 – p. 24). As the Roman road lies c. 12 m to the east of the ditch on a similar

Unfortunately, ditch 3036 extended outside Area 3 and was visible only in section (Figure 32). The majority of the upper part of 3036 was removed by ditch 397 consequently a full profile of it was not seen. As recorded, it was at least 0.46 m wide and 0.36 m deep with concave sides and a sharply rounded base. No finds were recovered. This suggests that ditch 397 was probably a recut of ditch 3036, however, not enough of the earlier ditch could be traced in the excavated area to confirm this. The full width of ditch 397 was not established as, like 3036, it extended out of the trench and later recuts disguised the original profile. At its 27

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Sec.48 399

24.51 m O.D.

388

West

3042

3043

397

398

East

3024

3022

3041 3037 3036

0

1m

scale 1:20 @ A4

Figure 32: Area 3 – Section of ditches 397, 3024 and 3036

Sec.50 24.60 m O.D.

East

3067 3072

3073

3024 3065

3060=3066 3070

397 3063 3023

3064

West

3071

0

1m

scale 1:20 @ A4

Figure 33: Area 3 – Section of ditch 307 and gully 3097

Sec.96 east

west 3042

3098

3022

3024=3095 3097 3096

3042

3023 397

3094

0

scale 1:20 @ A4

1m

Figure 34: Area 3 – Partial section of ditch 3024

28

Janet Phillips and Pete Wilson: The excavation

widest point 397 measured 1.95 m and was at least 1.2 m deep, its full depth extending below the agreed site depth limit. The primary fills yielded pottery dating to the late 3rd to early 4th century, contexts 3008 and 3094. Environmental sample 41, taken from 3094, produced traces of cinder, coal and indeterminate rectilinear charcoal and possible slag, as well as snail species that favour semi-heathland/meadows/woodland areas.

of Oxidised Sandy ware dating from late 13th–14th centuries. Infant burial SK718 (Figures 35, 39-41) Oval grave 716 was cut into Phase 2 ground surface layer 727. It contained an infant burial (SK718) that was possibly aligned north-south with the skull facing east. Unfortunately, the grave and surrounding area showed signs of animal disturbance, particularly on the southern and eastern sides. Grave cut 716 measured 0.40 m by 0.16 m and was 0.07 m deep with irregular sides and base due to the animal disturbance. There was no evidence of a coffin or shroud. Environmental samples 85 and 86 only produced tiny traces of coal and cinder.

Ditch 397 was recut by ditch 3024/3095. Restrictions on the excavation meant the full extent of 3024 was not exposed, with a maximum width of 1.40 m evident. Where visible the ditch had sloping irregular sides, with a slightly rounded bottom. 3024 was identified in several sections, having been initially noted below wall 353. Further sections through ditch 3024 demonstrated that its fills were loose sand-silts, with evidence of slumping events incorporating gravel from the natural subsoil. (Figures 31-34)

Bustum Cremation 475 (Figures 35, 42-49) A heavily burnt pit, 475, was identified within the northern verge trenches, cut into Phase 2 ground surface 477. Pit 475 was found to contain burnt material including bone, and several pottery vessels, and appeared to conform to a bustum-type cremation. Bustum cremation involves a pre-prepared pit with the funeral pyre constructed over the top, as a result of which the surrounding ground level, 477, was heavily burnt and included some of the burnt human bone remains. The majority of the burnt material had been scraped into the pre-prepared pit with a concentration of human bone being placed in a large calcite-gritted jar (Chapter 4 – No. 108, RF 416). Osteological analysis (Chapter 13) demonstrated that the deceased was an ‘old middle or mature adult’, but it was not possible to identify the sex of the burial from the surviving bone (p. 241). However, analysis of the finds allows Hilary Cool to argue persuasively that the deceased was male and ‘had undoubtedly been a soldier’ (Chapter 7).

On the eastern side of the ditch group was a further possible recut, ditch 3072 (Figure 34), which was 0.46 m wide and, while not bottomed, was at least 0.44 m deep. A further small recut, ditch 3097 (Figure 33), was visible for c. 2 m on the western side of the ditches in the central and southern area of the trench. Neither of ditches 3072 and 3097 produced any finds. This roadside ditch may also have been identified in the northern trench edge in Area 9. Deposit 936, (Figure 24), was identified as redeposited natural possibly forming a bank, which drops suddenly on its eastern extent as if cut by a feature. Unfortunately, this could not be investigated further as the full permitted depth of the service trench had been reached. Graves Infant burial SK286 (Figures 35-38)

The extent of heat damage caused by the pyre to the surrounding ground level has made the exact profile of the pit hard to discern, it does however appear to have been dug specifically to fit and support the pot. Large chunks of charcoal were identified around the base of the pot, suggesting branches were used to further support the pot within the pit. These branches appear to have been wedged into the base and sides of the pit and may account for its uneven profile.

Grave 284 was cut into the Phase 1 ground level 218. It contained a near complete infant burial (SK286) orientated east-west with the skull facing south. The U-shaped grave measured 0.75 m north-south by 0.65 m east-west and was 0.22 m deep. Due to the small size of the skeleton scattered bone from the burial was recovered from all the environmental samples taken from fill 285 (Samples 9, 10, 11 and 14). In addition, traces of cinder, coal, charcoal, vitreous/ vesicular slag, mortar crumbs, a single indeterminate charred grain fragment and a small piece of possible plaster were recovered. Similar material was recorded from beneath the burial along with a small sherd of glass. No evidence of a coffin or shroud was recovered, although the grave had been sealed with three large stones measuring approximately 0.20 m x 0.20 m x 0.20 m. The grave fill also contained an intrusive sherd

The roughly oval shaped cremation pit, 475, was 0.55 m long, 0.43 m wide and 0.40 m deep, with steep sides and an uneven base. A primary fill, 490, of heavily burnt sand was identified in the base of the pit containing pottery dated from AD 120 onwards. An environmental sample (80) taken from this deposit revealed rectilinear charcoal originating from oak trees and possibly heather, mortar crumb, a trace of vitreous slag, coal, elder fruits and a possible fragment of fish-scale. 29

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

N Bustum burial 475

A4

Projected route of Roman road

A9

A3

A8

1:2000 @ A4

SK 718

A4 A9

A8 A3

A1

A7

A7

A6 A2

SK 286 A5 0

14.5m 1:250 @ A4

Figure 35: Phase 2 – Location of burials

30

Janet Phillips and Pete Wilson: The excavation

N

284

SK 286

285

284

Skull

SK 286

Fig. a

Fig. b

0.50m

0 scale 1:10 @ A4

Figure 36: Area 2 – Grave 284, skeleton 286. 36a: Capstones partially covering SK286, 36b: SK286

Figure 37: Area 2 – Grave 284, skeleton 286 showing capstones

31

Figure 38: Area 2 – Grave 284, skeleton 286

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

N Fig. 40

Skull

716 SK 718

0.50m

0 scale 1:10 @ A4

Figure 39: Area 7 – Grave 716, skeleton 718

Sec.62 24.18 m O.D. north

south 716

Animal disturbance

Figure 41: Area 7 – Grave 284, skeleton 286

1m

0 scale 1:20 @ A4

Figure 40: Area 7 – Grave 716 profile

Sec.76 477

c. 1.10m below ground level

476 south west

north east 4044

c. 1.70m below ground level

475

1m

0

Figure 42: Area 4 – Section through bustum burial 475

scale 1:20 @ A4

Together the food stuffs could possibly be the remains of a funerary feast, or an offering intended to sustain the deceased on their journey to the afterlife.

appears to have been swept into the pit to surround the pot, with the main bulk of the burnt bone being deposited within it. Environmental sample 77, from fill 476, produced predominantly ash and burnt bone, rectilinear charcoal, with occasional round wood fragments, none of which retained bark or the waney edge to assist in determining wood growth. The species identified were largely oak, with some alder/birch/

The main fill of pit 475 was a loose, black, ashy-sand, 476, with frequent charcoal lumps and burnt bone remains (recorded as context 478). A single residual flint blade was also recovered from 478. This material 32

Janet Phillips and Pete Wilson: The excavation

Sec.79 north east

south west

475

1m

0 scale 1:20 @ A4

Figure 43: Area 4 – Profile, north east to south west, through bustum pit 475

Sec.80

north west

south east 475

1m

0 scale 1:20 @ A4

Figure 44: Area 4 – Profile, north west to south east, through bustum pit 475

hazel and possibly heather, a single small sherd of glass was also recovered.

These additional vessels presumably either related to a funerary feast, or had contained provisions to sustain the deceased on his journey to the afterlife. They were a complete local grey ware beaker (Chapter 4, No. 41, RF 417) with a curved rim and folded body, the lower part a large rouletted colour-coated flagon (Chapter 4, No. 16, RF 825) and a heavily burnt grey ware or Black Burnished ware 1 lipped dish (Chapter 4, No. 38, RF 826), fragments of which were incorporated in the surrounding surface deposit 477.

A further environmental sample (78) was taken from layer 477, where it displayed disturbed ground with apparently trampled elements of the bustum spreading out from the main pit. Sample 78 produced similar charcoal fragments and evidence of tree species sample 77. A further complete local grey ware beaker (No. 41, RF 417) and substantial parts of two other pots; one an unburnt colour-coated flagon (No. 16, RF 825), the other a heavily burnt grey ware or Black Burnished ware 1 lipped dish (No. 38, RF 826 were included in the pit next to the large jar containing the bulk of the bone.

The surviving profile of bustum pit fill 476 suggested that the backfill of the pit may have been left proud of the surrounding ground surface, perhaps as a visible marker of the location of the burial. The overlying deposit, 477, which as noted above incorporated 33

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 45: Area 4 – Bustum during excavation – complete local grey ware beaker (No. 41, RF 417)

Figure 46: Area 4 – Bustum during excavation. Main bustum vessel, handmade calcite-gritted jar (No. 108, RF 416), under local grey ware beaker (No. 41, RF 417)

Figure 47: Area 4 – Bustum during excavation. Other pottery fragment overlying the main bustum vessel

Figure 48: Area 4 – Bustum during excavation. Main bustum vessel fully exposed

cremation material and fragments of pottery vessels also found in the bustum pit, rises over the fill of the pit (Figure 42) suggesting that the grave location may have remained visible for some considerable time. However, the mound, as recorded, is only c. 0.5 m in diameter, in contrast with sites such as Petty Knowes, High Rochester (Charlton and Mitcheson 1984). At Petty Knowes the smallest recorded mound is 1.2 m in diameter (Burial 9) and most are in the range of 2.5 m to 4 m with in several cases a surrounding ditch, or ditch and bank (Charlton and Mitchelson 1984: 6). That said, given the evidence at Brooklyn House for post-Roman ploughing and other disturbance of overlying deposits, originally the mound could have been more substantial, albeit with no evidence of either a surrounding ditch or bank.

Figure 49: Area 4 – Bustum pit fully excavated

34

Janet Phillips and Pete Wilson: The excavation

Possible Mausoleum construction cut 270 (Figures 35, 50-52)

Approximately 0.2 m of construction cut 270 was visible in the bottom of Section 15 (Figure 51), where it was 0.59 m wide. A 2 m long section was excavated at the western terminal of the feature. Using robber cut 266 as a guide to the original shape of the feature, it would have measured approximately 4.65 m east-west before turning through 90o to the south. Unfortunately, the feature could not be traced further as it went below the agreed construction site formation height. All that remained of the profile of 270 was a flat base with a steep side to the north. The primary fill, 267, was 0.12 m deep, and comprised a semi-firm, yellow-brown, siltsand, that appears to have been a levelling/bedding layer in preparation for mortar deposit 269 which sealed it. 269 was a 0.06 m thick deposit of grey lime mortar which ran the length of the excavated slot.

The investigation of an area of slumped and subsiding wall in the main Roman structure in Area 2, Structure A, identified an underlying feature as having caused the stone walls to collapse. Further excavation identified a large feature, 270, cut into the Phase 1 ground level, 218, below Structure A. The feature appears to be a construction cut for an early structure, which has been tentatively described as monumental due to its depth and, given the later use of this area of Norton for burials, possibly a mausoleum. Only a small portion of the early feature remained, as the cut and any walling, or other structural elements had been removed at a later date by robber trench 266.

N

Fig. 51

1:250

ouse

lyn H

Brook

226

266

269

266

203 205 227

Structu

re A

226

Structure A

218 266

2m

0 scale 1:50 @ A4

Figure 50: Area 2 – Plan of possible mausoleum (feature 270)

Sec.15

24.53 m O.D.

205

north

218

south

226

227

266 273 272

271 268 269

274

267 270 0

266

scale 1:20 @ A4

1m

Figure 51: Area 2 – Section through robber trench (266) and construction cut (270) of possible mausoleum, showing overlying wall 205 of Structure A

35

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire trenches. 3060 produced pottery dating to the late 3rd to 4th century. A similar deposit, or deposits, were identified in Areas 7 and 8. In Area 7 705 was up to 0.35 m in depth, as was 736/757/760. In Area 8 823, 828 and 831 were up to 0.40 m thick. Pottery dated to AD 120 onwards was recovered from 828, and from the 3rd to 4th centuries from 831. Fragments of stone roof tile were found within deposits 823 and 831. The main Phase 2 structure that appears to have been removed at this time was the possible mausoleum (270). A large robber cut, 266/221, (Figures 35, 50-52) measuring 1.74 m wide at the top, with a stepped edge to the south, a steep convex side to the north and a stepped base, appears to have removed most of the earlier feature. The stepped profile appeared to have been created on the southern side of what would have been the mausoleum wall to allow easier access to the stonework. All large stonework barring one piece of masonry, which appears to have tumbled back into the cut and been left behind, was removed and may have been reused in the new buildings. With respect to the latter suggestion, the walls of Structure A (below) incorporated reused stone.

Figure 52: Area 2 – Robber trench (266) and construction cut (270) of possible mausoleum, with overlying wall 205 of Structure A

The masonry block and fragments of stone roof tile were found in primary fill 274, in the lower stepped area of the robber cut. Fill 274 underlay 268, which was sealed by a 0.05 m thick layer of redeposited natural sand (272), that had either been backfilled or had slumped in. Above 272, fill 271 contained building rubble, including fragmented stone roof tile and small pieces of masonry. The final fill, 273, was 0.84 m thick and included further building rubble. Finds recovered included pottery dating to the 3rd century from 274 and 46 fragments of bone from 273. The bone included both pig and cattle, including a cattle skull, but was generally poorly preserved and displayed evidence of abrasion suggesting that it was redeposited in the robber cut from elsewhere and could have originated off site.

PHASE 3 - ROMANO-BRITISH (late 3rd and 4th century) Following use of the area for burial the Norton settlement expanded to the south along the line of the Roman road. This resulted in the establishment of a series of stone structures to the west of the road, in the northern and eastern part of the development area. These were built over the former cemetery area and removed, or buried, any earlier structures. The most intensive development was in Area 3 where four structures and associated features were identified. The close proximity of these walls and the limited space within the trench made defining relationships difficult to determine, although a probable sequence is offered below.

Robber cut 266 was aligned east-west and was visible for 4.65 m before turning through a right angle to the south. As indicated above, the feature could not be traced further due to the agreed construction site formation height. A partial excavation on the corner was undertaken but the base was not reached, as the section clipped the corner of the feature only. Visible in the corner was a minimum of 0.25 m depth of a mixed slumped fill, 297, comprising a yellow-grey, silt-sand with natural stone inclusions. Overlying this was fill 222/296, a silt-sand, with small building stone inclusions and occasional flecks of charcoal. Seven fragments of cattle bone were recovered from 222 that displayed evidence of butchery and gnawing.

Post-burial Soil Development Some evidence of a build-up of deposits after the area ceased to be used for burials was identified in the southern part of Area 3. However, there was little indication of this over the rest of the area due to the significant amount of ground disturbance associated with the construction of the buildings. Deposit 3084/3086/3060 overlay probable Phase 2 roadside ditch 397 and extended across the full width of the trench. It was up to 0.32 m deep but was only visible for 2.50 m before being truncated by modern drainage 36

Janet Phillips and Pete Wilson: The excavation

Structure A (Area 2) – (Figures 53-57) Structure D Structure E Structure I

Structure H

Roman road

A4

1:200

Structure G

A9 A3 A8 A1 Brooklyn House

Structure F

A7

Structure A

Structure B

N

Structure C

A2

25m

0 scale 1:500 @ A4

Figure 53: Site plan showing all structures

37

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Modern Services

202

219 Fig. 56

210

212

240

242

237

244

282

206

280 229

Fig. 64

205

236

Fig. 59

N

232

Fig. 58

205

254 Fig. 66

5m

0 scale 1:100 @ A4

Figure 54: Area 2 – Plan of Structure A

A large stone-built structure (Structure A) was located in the north-eastern area of Area 2 (Figure 54). This structure was cut into the upper fills, 273 and 222, of the mausoleum robber cut (266) and was at least 15.50 m long and 5 m wide. The walls survived to a maximum of two courses and a height of 0.35 m. The building was aligned approximately east-west, perpendicular to the nearby Roman Road and, as recorded, the surviving evidence comprised of two well-defined walls to the south and east, and a possible third wall on its northern side. However, significant robbing had taken place and the whole structure was cut by a modern drain which together made the northern wall rather ephemeral.

Southern wall 205 was 3.50 m long and 0.55 m–0.70 m wide and, although robbed out in places, large stretches remained in good condition. It had been built in a construction cut, 226. Roman brick, box flue tile, imbrices and stone roof tile were recovered from directly on top of wall 205 along with pottery dating to the late 3rd to early 4th century. However, it should be noted that this material could have originated from the overlying rubbish deposit 206, rather than relating directly to the structure. The north-south aligned wall, 232, varied between 0.55-0.65 m in width, and ran from the eastern end of wall 205 for 5 m before being completely robbed out. No construction cut could be identified in plan (Figure 54). Three fragments of cattle/large mammal bone with butchery and gnawing marks, oyster shell and possible fragments of edible snail (Helix ?pomatia L.) were recovered from 233 overlying wall 232 and therefore are likely to have originated in the rubble dump overlying the walls, with 233 equating with layer 206 (above).

Within the structure a linear channel clear of all rubble and lying amongst the overlying debris may hint at a further division of the structure in the form of a robbed-out wall, however no firm evidence for this was recovered. Possible stone flagged floors were identified in places, along with a row of seven slightly off-centre postholes. A potential entrance may have been located in the south-eastern corner, and a possible yard or additional aspects of the building were visible in the eastern part of the area. Unfortunately, a modern service trench truncated the northern area of the structure, so further walls or features may have been lost.

It is possible that a 1.50 m wide gap between the eastern end of wall 205 and the southern end of wall 232 may represent the location of an entrance (Figure 61). However, no threshold could be identified and having a doorway so close to the corner would potentially have compromised the structural integrity of the building and it seems more likely that the gap is merely a product of the stone-robbing.

The two best preserved walls of Structure A (Figures 54, 58-62) were 205 (also recorded as 217 – see Figure 58) and 232, both of which were constructed from a double thickness of roughly squared oolithic limestone blocks c. 0.40 x 0.30 x 0.20 m in size with a rubble core and no apparent bonding. The blocks display occasional tooling and evidence of re-use.

The northern, possibly east–west aligned, wall 219 was very ephemeral in nature, having been largely robbed out with only the rubble core remaining along with 38

Janet Phillips and Pete Wilson: The excavation

an occasional facing stone. The wall was observed for c. 4.50 m before it could no longer be traced, and no intact areas were identified. It should also be noted that this feature may be a linear rubble spread, with the actual wall remains having been destroyed by the service trench directly to the north. This would account for an off-centre row of seven postholes which lie approximately 2.50 m north of wall 205, but only 1 m from possible wall 219.

probable Phase 1 ground surface 218 and extended for 5.75 m within the structure before terminating or being truncated by later activity. The postholes were evenly spaced with between 0.80 and 0.95 m from centre point to centre point. Each post-hole was packed with pieces of oolithic limestone ranging in size from 0.10 m x 0.08 m to 0.18 m x 0.29 m. No clear cuts were definable, as a result of the homogenous nature of the underlying material, and no clear post-pipes were identified.

Wall 219 yielded Roman brick, 12 fragments of animal bone including pig and large mammal with evidence of butchery and burning, stone roof tile and pottery dating to the 4th century. However, again it is possible the pottery derives from the overlying rubble deposit 206. An intrusive sherd of late 19th/20th-century bottle glass was recovered from 219, which presumably originated from the nearby modern drain. (Figure 54)

Due to the extent of later robbing a consistent floor level inside Structure A was not found, however several patches of a possible flagged floor, 229, 280 and 281 (Figures 54, 63), were identified overlying layer 218. The largest of these patches, 229, measured 2 m by 1 m, and was located 5 m east and 1.50 m north of wall 205. The stones were between 0.10 m - 0.15 m thick, 0.23 m - 0.40 m wide and 0.50 m - 0.30 m long. The stones forming the possible floor were roof tiles and therefore they could be no more than a remnant of the demolition/collapse of the structure. However, their consistent height of 24.48 m OD throughout Structure A is perhaps more indicative of them being reused as flooring. Pottery

Within the extant part of Structure A was an eastwest aligned row of seven postholes – 210, 212, 240, 242, 237, 244 and 282, aligned some obliquely to the building (Figures 13, 54, 55, 56). They were cut into the

Figure 55: Area 2 – Structure A (view to north west)

Sec.7 west 218

211

213

218

210

212

218

218

241

218 243

240

238 237

242

24.61m O.D. 220

218

283

218

east 0

scale 1:20 @ A4

1m

282

Figure 56: Area 2 – Section through east-west line of post-holes within Structure A

39

218

245 244

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 57: Area 2 – with Structure A in the foreground (view to south east)

24.82 m O.D.

205

west

227 226

218

205 227 218

226

205 205

227 226

218

205 227 218

east

0

scale 1:20 @ A4

1m

Figure 58: Area 2 Structure A – profile of wall 217

recovered from 229 included a sherd of Crambeck Ware dated to the late 3rd to 4th centuries.

fragments. Two flagstones were identified below rubble deposit 209 (Phase 4), all were similar in form to those forming layer 229 and were found at a similar height 24.46 m OD. The flagstones ranged in size from 0.30 m x 0.34 m to 0.28 m x 0.35 m and were 0.08 m - 0.10 m thick.

A further small remnant of flagged stone floor, 280 was found in the south-western part of the structure associated with 3rd-century pottery and stone roof tile 40

Janet Phillips and Pete Wilson: The excavation

It may also be possible to interpret a similar patch of stonework, 281, as a floor. The stones were identified to the north of the row of postholes, lying roughly centrally within Structure A. As well as the consistent height with (229) and (280), similar dimensions (0.31 m x 0.27 m x 0.12 m) and a similar wear pattern would strongly suggest this is a further remnant of a deliberately-laid floor.

Other areas of stone within Structure A, as seen in Figures 61 and 62 adjacent to wall 205/217, could be material deposited to support areas of flooring. However, they could simply represent collapse material from the building.

Sec.1 24.85 m O.D. 217

west

east

218

0

scale 1:20 @ A4

1m

Fig 59: Area 2 Structure A – profile of wall 205/217

Figure 60: Area 2 Structure A – wall 205 (view to south)

Figure 61: Area 2 Structure A – wall 205/217 looking towards the possibly entrance in south-east corner of the building (view to south east)

41

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 63: Area 2 Structure A – possible remnant of a flagged floor, context 225 (view to south)

with pottery dating from the late 3rd century to the later 4th century, including proto-Huntcliff type jars. At the western extent of the surface stones appeared to have been deliberately placed on end marking the edge of the yardage or potentially to provide drainage. The yard was made up of pieces of oolitic limestone ranging in size from 0.37 - 0.20 m to 0.30 - 0.30 m, with smaller stones placed on end, as well as some reused fragments of ceramic building material.

Figure 62: Area 2 Structure A – wall 205/217 showing extent of robbing of parts of the wall (view to north west)

Possible external yard surface 236 In the eastern part of Area 2, outside of Structure A, there was a further possible surface, 236 (Figures 54, 64, 65). This may have been a yard surface or hard standing, unfortunately due to the limited extent available for excavation its limits are unclear, however it seems reasonable to assume it may have extended to the nearby Roman road, to allow access to Structure A. There was approximately 1.50 m of surface 236 visible within the excavation area, starting 0.10 m from the eastern edge of wall 232. Surface 236 yielded 22 fragments of animal bone including cattle, pig and chicken, as well as frog/toad, with some bones displaying evidence of butchery marks. Other finds include oyster shell, mussel and limpet fragments, stone floor tile, possibly reused building stone, Roman brick, tegulae and imbrices

Located 1.70 m south of Structure A and wall 232 was a single thickness of roughly hewn oolitic limestone blocks, ranging from 0.30 m x 0.20 m to 0.20 m x 0.14 m in size. These blocks formed an east-west aligned wall (254) (Figures 54, 66, 67) located to the south of Structure A. Only 1.50 m of wall 254 was visible extending out from the trench edge before being truncated by geotechnical works carried out during the trial trenching phase. No continuation of the wall could be identified beyond the modern truncation. Wall 254 was cut into deposit 255, which has been previously suggested to be one of several deposits forming a ground surface that developed at the end of the Phase 1 activity on the site. As with many features on the site, the homogenous Sec.8

north

south

201 202 230

24.71 m O.D.

236

230

0

scale 1:20 @ A4

1m

Figure 64: Area 2 – Section through possible yard surface 236

42

Janet Phillips and Pete Wilson: The excavation

Structure(?) B (Area 2) (Figures 53, 68, 69)

Figure 65: Area 2 – Possible yard surface 236 (view to east)

Sec.9 24.52 m O.D.

west

2018 2017

0

east

254 255

scale 1:20 @ A4

1m

Figure 66: Area 2 – Profile of wall 254

Figure 67: Area 2 – Wall 254 (view to north)

nature of the deposits meant that a construction cut was not visible. Wall 254 was isolated from the nearby features, which it makes it difficult to assign it to a particular phase. However, the proximity of the wall to Structure A, and the similarity in alignment suggests it may have marked a plot boundary. 43

The possible remains of a second building, Structure B, were identified 10 m to the south of Structure A, 7.50 m south of possible plot division wall 254. Although no clearly defined walls remained two significant linear spreads of rubble were identified. The linear nature of the spreads along with the remains of mortar on some of the stone fragments strongly suggests they could be the rubble cores of walls from which the facing stones have been robbed, or possibly the remnants of floors or surfaces. Both spreads were aligned approximately eastwest, similar to the alignment of Structure A. The remains covered an area of 5 m northsouth by 2 m east-west and appeared to be cut into probable Phase 1 ground surface 256. The possible structural material survived as two separate elements, 252 and 239/253 (Figure 70). 252 was a possible wall or rubble spread measuring 1.50 m north-south and 5.20 m east-west. It incorporated a roughly straight edge on its northern side with a large patch of degraded limestone on the western side of the spread suggestive of the remains of further straight edge. On investigation, no coursing or in situ facing blocks were found, however the remaining stonework may represent the rubble core of a wall, with the considerable robbing on site accounting for the lack of any further remains and the diffuse nature of the material. Rubble spread 239/253 was located 1 m to the south of 252, with the main concentration of stone focussed 2.75 m south of 252). This spread contained a much more varied stonework than 252, with larger blocks 0.30 m x 0.20 m in size in a matrix of smaller rubble. Despite the initial linear appearance, the deposit could not be confirmed as part of a wall structure during excavation, a slot excavated through the eastern extent of the spread did not show any coursing, and the mixed size and quantity of stonework prevented any identification of possible rubble cores. Again, like 252, 239/253 could possibly represent remnants of floors or surfaces. Deposit 253 yielded Roman brick, imbrices, oyster and mussel shells and 30 fragments of animal bone, mostly from large mammals. Oyster shell and late 3rd- to 4th-century pottery was recovered from 239 along with five fragments of cattle/ large mammal bone. Although the section excavated through spread 252 did not find further structural evidence

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

N

Fig. 69

252 253

Structure B

259

258

251 249

Structure C

2014

Fig. 72

248

2013

2015 250

246 5m

0 scale 1:100 @ A4

Figure 68: Area 2 – Plan of Structures B and C

44

Fig. 71

Janet Phillips and Pete Wilson: The excavation

Sec.10

24.62 m O.D.

west

252

east

0

scale 1:20 @ A4

1m

Figure 69: Area 2 Structure B – profile of possible wall core 252

Figure 70: Area 2 Structure B – possible wall cores 252 (foreground) and 253 (rear) (view to south)

it did locate a cut underlying the rubble. The roughly U-shaped feature appeared to run east-west; however it was not identified on the preceding geophysical survey as joining with any of the earlier ditches. As the feature was cutting the probable Phase 1 ground level 265, it could be that the feature was a robbed-out construction cut for the remnants of a structure from which rubble spreads 252 and 253/239 originate. The cut, 263, was 0.82 m wide and 0.22 m deep and its fill, 264, produced three fragments of unidentified animal bone and pottery spot-dated from the late 3rd century onwards.

Morphologically, Structure C was what is conventionally described as a ‘corn-drier’, but as a type is now believed to be multi-purpose, or at least suitable for a variety of uses including both corn-drying and malting grain for brewing (van der Veen 1989). As with the majority of features on the site no definitive construction cut could be identified, however due to the method of construction of the kiln the following is probable. Structure C appears to have been cut straight into the Phase 1 ground level (265), as a single construction event (2012), with minimal backfilling needed. A flagged stone floor, 2013, was then laid to form the base of the flue (Figure 72). Internally the flue was 0.25 m wide, with the flagstones continuing underneath the walls, and extended for approximately 0.95 m before appearing to slope downwards at the eastern end.

Structure C (Area 2) – (Figures 53, 68, 71-74) Part of a largely intact structure, Structure C, was found in the south-eastern corner of Area 2. As seen, it was 1.50 m wide north-south, and at least 1.75 m east-west with the remainder located outside the trench. Initial investigation revealed stonework, including three walls and an in situ capstone forming a flue with associated hard standing on its northern and southern sides (Figures 53, 71, 72). The flue was reasonably intact with ashy deposits remaining internally and two possible ash pits that extended into the eastern trench edge.

The northern flue wall, 249, which partially overlay floor 2013, was constructed of a double thickness of limestone blocks, with three courses internally surviving to a height of 0.47 m, but with only one 0.25 high external course remaining in situ (Figure 72). All of the blocks were roughly cut, measuring between 45

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 71: Area 2 – Section of possible surface 251, ash pit 275 and wall 246 of Structure C

0.20 m x 0.20 m x 0.13 m and 0.30 m x 0.25 m x 0.15 m There was no obvious bonding material present. The upper course was significantly heat-affected with crumbled limestone fragments clear throughout. Finds were recovered directly from wall 249 including late 3rd- to 4th-century pottery, ten fragments of animal bone including cattle and pig bones, fragments of box flue and stone roof tile, as well as oyster shell. The southern wall, 250 was also 0.35 m high internally and constructed in the same manner as 249, partially overlying floor 2013.

appeared to form an external surface or possible yard; however, they were not fully excavated, and it is possible that they masked a return of wall 250 that would have given the T-shape to the kiln/drier structure. Approximately 1 m north of Structure C was a moderately compacted rubble deposit, 251, measuring between 0.90 m - 1.20 m in width and extending from the trench edge westwards for up to 3.20 m. 251 produced 25 fragments of animal bone including cattle and large mammal with evidence of butchery and scavenging, oyster shell, Roman brick and imbrices, pottery dating to the late 3rd to late 4th/early 5th centuries, mixed limestone building material and stone roof tile fragments. Although not obviously structural when found, in section (Figures 68, 71), it was clear that 251 had been deliberately deposited, suggesting it could be a walkway or a surface associated with the kiln.

Both walls, 249 and 250, extended for c. 1.40 m before appearing to slope downwards to the east and stop, either by design or as a result of later robbing activity. Both walls terminated below the in situ capstone at the western end of the kiln. As a result, they were not fully excavated due to the restriction on depth of excavation. The rear wall of Structure C, 2014, was not constructed to the same standard as walls 249 and 250. It appeared as a roughly made join between the two walls with no coursing or bonding material visible. There were no apparent gaps for access to the flue, however this area was not fully explored, again due to constraints on the excavation area.

Located between the eastern end of the kiln and the trench edge was what may have been an ash or stoking pit, 275 (Figure 71). This 0.87 m wide feature only extended for 0.4 m into the trench. It is unclear whether it relates to a later truncation of Structure C or whether it was an integral part of its design. In section walls 249 and 250 appeared to stop as if truncated rather than having a clear edge, although the flagged flue floor (2013) did seem to be dipping towards the pit. The lower fill of 275, 278, was a loose, soft, grey-brown ash. Approximately 0.16 m of fill 278, was excavated as excavation stopped at the agreed site formation level. The presence of unconsolidated ash, traces of coal and cinder, vitreous slag, a single charred wheat grain and some possibly degraded burnt bone or mineral material was noted. The nature of the material perhaps supporting a direct association with the kiln. Fill 278 also produced a near complete straight sided bead and flange bowl in fine grey ware (Chapter No. 87) dating to the late 3rd century.

The roof of Structure C had taken the form of a series of capstones, one of which remained in situ (2015) at the western end of the structure. Several others were found slumped into the flue. The capstone measured 0.4 m wide and 0.08 m thick and appeared to have been reused, possibly originally having formed part of a window opening. Directly to the south of Structure C, and appearing to butt up to wall 250, there was an area of flat stones, 246, measuring 0.90 m by 0.60 m (Figures 68, 71) that produced pottery dating to the 4th century. The stones 46

Janet Phillips and Pete Wilson: The excavation

Sec.16

24.53 m O.D.

capstone

249

277

west

278

flagged base

0

east

scale 1:20 @ A4

1m

Figure 72: Area 2 Structure C – profile of the inner face of wall 249

3rd century onwards, four fragments of cattle/large mammal bone with butchery marks and inclusions of (probable) stone roof tile and stone building material. Located 6 m to the north-west of Structure C was a large pit containing burnt material, feature 259 (Figure 68). Due to its proximity to the kiln structure and the burnt nature of its fill it is a reasonable possibility that the two features may be related. Unfortunately, the pit was found at site formation level and was not excavated. It was roughly circular, measuring 2.90 m by 2.70 m and, as seen, contained a single soft, dark brown/black, siltsand fill, 258. An environmental sample (60) was taken from 278 contained indeterminate rectilinear charcoal, frequent coal, a little mortar crumb, a trace of glassy slag and eight bone fragments of which two were burnt. Further finds recovered from the surface of 258 included pottery dating to the late 3rd to 4th centuries, oyster and mussel shell, and Roman imbrex fragments. There were 40 fragments of animal bone recovered including cattle, sheep/goat, pig, horse and unidentified mammal bones. An intrusive sherd of Brandsby type pottery dating from the 13th to 14th century also derived from 258, presumably introduced by disturbance from medieval ploughing.

Figure 73: Area 2 Structure C – view along flue showing in situ capstone (view to west)

Structure D (Area 3) A large stone-built structure, Structure D, (Figures 53, 75-79) was located in the northern part of Area 3. Like Structures A-C it was aligned perpendicularly to the Roman road to the east. Approximately 5.50 m of an eastwest wall, 353, was located, with a small length, 1.25 m, of a north-south return wall, 315, visible in the western

Overlying 278 was a 0.12m thick compacted orangeyellow, ash and mortar fill, 277. Fill 277 was significantly different to the main ashy fills, suggesting a possible attempt to cap the earlier ash and form a new working surface. Sealing 277 was a 0.40 m thick, loose, grey ash fill, 276, which produced pottery dating from the 47

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 74: Area 2 – Overall view of Structure C (view to north west)

trench edge extending south. It is unclear whether any of the other east-west walls identified in Area 3 formed a third side of this structure. Although two courses of stone were visible in the western trench edge, only the lower course remained across the rest of the trench. A possible internal posthole (3016) was identified and an external yard or hardstanding, 352/3015, to the north of which was a possible boundary wall (308) (see pp. 5254).

Apparently bonded into wall 353 was a short, c. 1.25 m, length of a return wall, 315, aligned north–south. A construction cut could not be identified due to the proximity of the trench edge and later disturbance. The wall appeared to be similar in construction to wall 353 with roughly squared oolithic limestone blocks used as facing stones, however the core of the wall lay beyond the limit of excavation. Only a single course remained in situ and appeared to bond into the second course of wall 353. This shallower foundation perhaps suggesting an internal division rather than external wall. The blocks, bonded with lime mortar, measured a maximum of 0.34 m x 0.30 m x 0.18 m with selective robbing appearing to have taken place as the wall continued southwards. (Figures 75, 76)

A construction cut, 3005, for wall 353 was identified cutting ground surface 388 and fill 399 (ditch 397). The dimensions of the construction trench were not clear due to disturbance by later walls to the north and south; however, it appeared to be a minimum of 0.85 m wide and did not appear to extend below the base of the wall. Stone wall 353 measured between 0.65 m - 0.75 m in width and was constructed from roughly squared oolithic limestone blocks, with a rubble core. The blocks were a maximum 0.40 m x 0.30 m x 0.16 m with occasional tooling marks present and a lime mortar bonding. Two courses of stonework were visible in the western side of the trench, with only one course remaining throughout the rest of the wall with later selective robbing events removing some of the facing stones. The construction cut of the wall, feature 3005, fill 3006, contained pottery dating to the late 3rd century onwards.

Two mortar floors, with a later repair layer, were identified within Structure D to the south of wall 353 and were visible over an area that was 2.10 m wide and at least 1.20 m long. A 0.10 m thick moderately firm, yellow-grey, mortar deposit with frequent ashy patches, 3011/3078, appears to have been the primary internal floor deposit and 3011 produced pottery dated to the late 2nd to 3rd century (Figure 77). Environmental sample (20), taken from 3011, contained ash, crumbs of mortar, abundant vitreous slag, traces of coal and cinder and charcoal – possibly heather and a few flakes and spheroids of hammerscale.

A short length of the continuation of wall 353 was identified in Area 8 (Figures 75, 76, 80). Wall 815 was built of the same roughly squared limestone with lime mortar bonding, although it was not as well-preserved as 353. Pottery dating from the late 3rd century onwards was recovered from 815.

Overlying 3011 was another mortar deposit with fewer ash inclusions which may represent a resurfacing or repair of the original floor. Floor deposit 3007 was 0.10 m thick and was observed for 2.20 m in the trench edge and extended 2.60 m into Area 3. The deposit was made up of a compacted yellow-grey mortar, with lenses of ashy material and contained pottery dating to the late 48

Janet Phillips and Pete Wilson: The excavation

Fig. 77

Fig. 80

814

Fig. 82

312

373 352

316

3113 384

314

815

361

363 315

353

355 816

354 Fig.99

359 356

3114 3016

Fig.81

817

345

340

819

Burnt tile 3012

370

Fig.100

3010

349

3013 820

368

348

343 821

3115

330 341

347

342 822

330 313

823

327

324

Structure D

3116

349

Structure E

Fig. 85

326

Structure I

323=319

320

825 337

332

824 Fig. 80

?Structure E

3114

0

Fig. 77

Fig. 82

Figure 75: Areas 3 and 8 – Overall plan of Structures D, E and I

49

1:50

2m

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Fig. 77

Fig. 82

304

N

Fig. 84

308

3110

3102 Fig. 83

3111

3102

3112 311 Fig. 80

813

303

3113

Sk 812 Fig.81

312

814 352

815 3114

353

315

828

3016

356

3114

817

3115

340

Fig. 80

Fig. 77

0

1:50

349

2m

Figure 76: Areas 3 and 8 – Plan of Structure D and area to the north

50

Fig. 82

51

3007

340 356

South

3011

3128

3127

3129

3107

3106

311

3108

330=388

315

3006

353

3005

314

351

396

Services.....

0

352

3023

333

312 3105

311

330=388

301 302 3104 303

3090

3130 3039 3093

scale 1:50 @ A4

3028

3029

3099

3100

3101

3086=3084

2m

3063

3062

3064

3082

Figure 77 Area 3 – East-facing trench edge section

350

3139 3085

3109

301 302 3104 3081

313 324

325

3102

356

311

323=319

305

341

304

3103 310

342 367

3007

340 356

311

330=388

3011

North

330=388

Janet Phillips and Pete Wilson: The excavation

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 78: Area 3 – The east-west walls of Structures D, E and I. Wall 314 (Structure E ) the top of the image, with wall 353 (Structure D) in the middle and wall 355 (Structure I) in front. In the left foreground are walls 315 (Structure D) and to the right wall 355 (Structure I) (view to north east)

2nd to early 3rd century along with fragments of stone roof tile.

Figure 79: Area 3 – The east-west walls of Structures D, E and I. Wall 314 (Structure E ) largely robbed out to the right, wall 353 (Structure D) in the centre and wall 355 (Structure I) to the left (view to west)

A continuation of the mortar floor layers recorded in Area 3 was found 1 m to the west in the main service trench in Area 8. Located next to the east-west wall was a probable floor deposit, 827, a grey-yellow, 0.06 m thick lime mortar layer which extended for 0.45 m before being truncated by a later feature. (Figure 80)

also recovered, and environmental sample 22 contained abundant cinder, frequent coal, crumbs of mortar, indeterminate rectilinear charcoal and two pieces of vitreous slag. The deposit measured 1.50 m north south, extended 4.50 m into the trench and was 0.14 m thick. A further deposit with very frequent mixed size stone inclusions, 312, overlay the northern side of 352. 312 produced oyster shell, nine fragments of bone from medium-sized mammals and more pottery dating to the late 3rd to 4th centuries. This deposit was encountered at the same height as 352 and appears to form a level, compacted surface.

A small pit or posthole, 3016, which was largely destroyed by later construction, was identified lying in the central area of the trench, 0.10 m to the south of wall 353. It is possible it formed part of an internal wooden division within this structure however only one similar feature was found, possibly as a result of later disturbance. A further small roughly circular pit or possible posthole was also identified in Area 8 to the west of Area 3. Pit 833 cut through floor deposit 827 and was located 1.50 m from east-west wall 817, the continuation of 315. The pit was 0.46 m in diameter and 0.30 m deep It contained several large stone roof tile fragments, possibly used as post packing. (Figures 75, 76)

Within the small service trench on the eastern side of Area 3 (Figure 82), a similar deposit to 352 was recorded which seems likely to be a continuation of the hardstanding/surface. Deposit 3125 was a firm red-brown, clay-silt with rubble and gravel inclusions. It was a 0.20 m thick deposit and extended for 2.62 m within the trench before being truncated by modern disturbance.

Three compacted rubble deposits were located to the north of wall 353 which appeared to represent dumping events to create an external yard surface or hardstanding (Figures 74-76, 80). The lowest of these, 3105, was 0.06 m thick and was observed extending for 3.96 m in the trench edge. Overlying it was layer 352 which included frequent small pieces of stone (reused stone building material) and pottery dating to the late 3rd to 4th centuries. Fragments of Roman brick were

A further surface was identified in Area 8 – layer 826, a 0.06 m thick deposit of yellow-white mortar (Figure 80). Wall 308 has been tentatively assigned to this phase as a possible plot boundary (see p. 69), located 7 m to the north of wall 353 and beyond the associated yard 52

53

827

829

828

Area 7 - NE corner

south

830

824

nat

815

825

802

drain

cable

nat

828

802 803 814=813

801

828

822

scale 1:50 @ A4

826

north

803 814 820

2m

Figure 80: Area 8 – East-facing section

0

821

801

Sec. 85

817 828

818 nat

833

802

834

828

814

827

Janet Phillips and Pete Wilson: The excavation

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire surfaces. 308 overlay an earlier ditch, 304, and had been largely robbed out. A few stones remained in situ aligned east-west, with a c. 1 m length of the wall, up to three courses high, in the eastern trench edge (Figures 76, 77, 83, 84). Its construction cut, 307, was partially visible in section cut into deposit 388. Wall 308 was not as well built or wide as those seen in the main structure, being 0.35 m wide and 0.45 m high. Rough facing stones remained in situ on the northern side and a possible rubble core was seen. It was unclear if there was mortar bonding due to later robbing, which had also largely removed the southern face of the wall. 309, the backfill of construction cut 307, contained small stone roof tile inclusions, seven fragments of animal bone including cattle, a fragment of Roman box flue and pottery dating to the late 3rd to early 4th centuries.

and extended 5.6 m into the trench (Figures 76, 77, 81, 86, 87). The limestone blocks used in its construction were up to 0.26 m x 0.35 m x 0.16 m in size, with three courses and a rubble core bonded with a coarse lime mortar remaining in situ. Two of the lower courses were stepped out to form a foundation on the northern side, with the lowest built from very roughly shaped oolithic limestone. The two upper courses were completed to a higher standard with roughly finished squared blocks. Significant robbing had occurred leaving only the rougher material from the lowest course except for a 2.3 m length adjacent to the trench edge. Fill 351, of construction cut 350, contained pottery dating to the 3rd century or later. Some finds were in association with wall 314 including pottery dating from the late 3rd to 4th centuries, fragments of imbrex and five large/medium mammal bones with evidence of butchery. However, they may derive from the backfill material 351.

Structure E (Area 3) (Figures 75, 77, 81, 85-87) In a second iteration of Structure D a substantial stone wall, 314, was built against the northern side of wall 353. As this wall had a stepped foundation on one side only, away from 353, it is suggested this was a later structure, either a replacement wall or, possibly, an extension to the existing building. No return of wall 314 was located, although there are similarities with walls 313 and 327 located c. 7 m to the south, suggesting they may have formed a more substantial structure that replaced Structure D. An internal mortar floor was identified, 356/3077, and to the north it appears the pre-existing external surface or yard, 352/3015, remained in use and had three bird and animal burials cut into it (see below). A small area of yard was also identified to the south of the structure, 319/323, possibly enclosed by a boundary wall 3062. Finds from above 314 included: stone roof tiles, ceramic imbrices and ridge tiles, five animal bones and late 3rd- to 4th-century pottery.

Two phases of similar east-west wall were recorded c. 7 m to the south of wall 314 and it seems likely they are associated with it (Figures 75, 77, 85, 87). Wall 313 was earlier than wall 327, and may originally have crossed the full width of the trench. However, it was built across the top of the large Phase 2 north-south ditch, 397, and, as was seen in Area 2, the soft nature of the ditch fills meant that slumping into the underlying feature was a problem. It appears that significant repair and replacement work had to be undertaken on the structure with a later, and more substantial wall, 327, replacing the eastern part of wall 313. Finds recovered from directly on wall 313 included pottery dating from the late 3rd to 4th century. 324, the construction trench for wall 313, was cut into the top of ditch 397, with remnants of 313 surviving for a length of 1.90 m. 821, a possible continuation of wall 313, was identified in Area 8 (Figures 75, 80). 821 consisted of two courses of roughly cut limestone blocks, up to 0.50 m x 0.30 m x 0.28 m in size. It incorporated a stepped foundation and a rubble core with lime mortar

A construction cut, 350, for wall 314 was visible cut through the external yard surface, 352, associated with the north side of Structure D. However, the homogenous nature of the earlier deposits masked its southern extent. Wall 314 was 0.65 m - 0.75 m wide

Sec. 36

24.98 m O.D.

352 south

355

356 388

356 3001

3002

353

314 351

363

3006

350

388

3005

0

scale 1:20 @ A4

1m

Figure 81: Area 3 – Section through walls 355, 353 and 314 (Structures I, D and E respectively)

54

north

3125

3110

3122=3104 3126

55

3117

302 3122 3113

3133 3118 3134 3119

3115

3113

3120

302

3112

scale 1:50 @ A4

3116

2m

3113

Figure 82: Area 3 – West-facing trench edge section

0

3111

3120

302

3121

3122

3124

3123

Animal bone

3114

3125

3113

3122

Janet Phillips and Pete Wilson: The excavation

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Sec. 24 301 24.91 m O.D.

303 north

south 308

308

308

307 309

311

0

311

1m

scale 1:20 @ A4

Figure 83: Area 3 Wall 308 – West-facing profile

Sec. 25

east

24.91 m O.D.

Sec. 30 308

307

west

325 324

0

327

west

309

311

24.85 m O.D.

313

scale 1:20 @ A4

0

1m

Figure 84: Area 3 – Section through Wall 308/ Construction cut 307

east

326 328

scale 1:20 @ A4

1m

Figure 85: Area 3 Structure E – Profile of Wall 313/327

remaining. The foundation course was 1 m wide with the second course reducing to 0.65 m in width. Issues with linking the stratigraphy as observed in Area 8 with Area 3 to the east mean that, as planned, wall 821 does not align with wall 313. However, it appears likely that the two were part of the same structure, rather than 821 representing a surviving element of the wall removed by robber trench 341 to the east (p. 70).

wide, with the upper course reducing to 0.65 m. Smaller, roughly squared limestone blocks had been used in its construction; they were a maximum of 0.31 m x 0.27 m x 0.21 m in size. Third-century pottery and a fragment of sheep/goat bone were recovered from 327. A probable mortar floor level can be associated with Structure E extending from wall 314 to walls 313 and 327. Floor 356/3077 overlay the mortar floors associated with Structure D, 3011/3078 and 3007, and took the form of a compacted yellow, 0.10 m - 0.20 m thick coarse mortar (Figure 77). Pottery dating from the 3rd century onwards was recovered from 356.

Cutting wall 313 was the later repair or replacement wall, 327 (Figures 75, 77, 85-87). A partially visible construction cut, 326, was evident cutting through the earlier wall, very slightly off the exact alignment, 0.10 m - 0.15 m to the south. Wall 327 was more substantial than 313, with a deeper foundation trench and four courses of masonry remaining, including stepped foundations and a rubble core bonded with a coarse lime mortar. The stepped foundations were 0.90 m

Although Structure E was cut through the earlier northern external yard surfaces, 3105, 352 and 312, it appears likely that they remained in use at this time as there was no evidence of resurfacing. 56

Janet Phillips and Pete Wilson: The excavation

Figure 86: Area 3 Structure E – Wall 313/327 (view to north)

Figure 87: Area 3 Structure E – Wall 313/327 (view to north)

dating’ {Beta Analytic Laboratory number: Beta – 489837, Sample: 35N16C317B} ‘suggested they were of modern {20th century} date – probably waste from preparing the carcasses for cooking.’ (Carrott et al 2019: 31). However, while the radiocarbon date and size of the geese suggest that the burial could be a modern intrusion, this is contradicted by three pieces of evidence. Firstly, the fact that the matrix of the burial, 318, was largely cinder (environmental sample 19), rather than the overlying silt-sand soil matrix of rubble 303; secondly, the apparent care taken in the burial and thirdly the presence of another similar burial nearby (373 – below). A similar roughly oval cut, 373, measuring 0.20 m by 0.30 m and 0.04 m deep feature was also cut into surface 352 0.40 m to the north of pit 316 (Figure 88). Unfortunately, due to the shallow nature of the remains there was no clear articulation or definition of a skeleton. The bones, SK374, were fragmentary and were lifted as an unnumbered sample, analysis of which showed it to represent ‘at least two geese, or goose-sized birds’ (Carrott et al 2019: 31). The backfill, 375, was, again, shown to incorporate abundant cinder, with a little coal, some spheroids of hammerscale, five sherds of pottery, slag, and six small fragments of ceramic building material and no evidence of contamination from the overlying matrix of rubble 303.

Animal/Bird Burials

An indistinct patch of burnt material in a shallow scoop was located 0.75 m north-west of 374 which was also cut into the yard surface 352. The cut, 389, was roughly oval, measuring 0.19 m by 0.13 m and 0.25 m deep, with gradually sloping sides and a rounded base. No clear definition of an animal or bird skeleton was identified however, there was considerable burning and the contents were lifted as an environmental sample 24. No bone was recovered.

Butting against wall 314, and cutting into the yard surface 352, was the cut of a sub-circular pit (316), measuring 0.21 m by 0.26 m in area and 0.09 m deep. Within the cut were the skulls of two geese, SK317, one facing east, the other west, with their wings apparently laid over the top of both to cover them (Figures 88, 89). Post-excavation assessment suggested that there may have been elements of more than two geese present (Carrott et al 2019: 16). It also states that they were ‘large, domestic, birds …. and both their size and radiocarbon

A small length of a possible east-west plot division wall, 3062, was located 1 m to the south of wall 313 and seen in section (Figure 77). Wall 3062 appeared to cut through external surface 3082, although it was overlain by the same later rubble deposits that also sealed 3082. This suggests that it was built while Structure E was in use, perhaps to mark in a more permanent form the southern plot boundary. The lower portion of its construction cut, 3063, was partially exposed revealing three courses of stonework.

The heavily truncated remains of a further possible external surface to the south of wall 313 were only visible in section (Figure 77). Deposit 3082 was 0.28 m deep and formed from a compacted yellow-grey, claysand deposit with many mixed sizes of stones and heavy mortar inclusions forming what could have been a cobbled yard.

57

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

N

303 389 312

373 350

352

316

384

314

361 363

353

0

2m

scale 1:50 @ A4

352

352 316

316

350 351

350 351

Sk. 317 314

351

Sk. 317 314

0

351

0.5m scale 1:10 @ A4

Figure 88: Area 3 – Bird burial 316

There were no visible tool marks on the rough limestone used to build wall 3062. The wall, which was 0.38 m wide and 0.42 m high in the section, was only two stones thick with a limited amount of coarse mortar visible. The fill, 3064, of the construction cut for 3062 (3065) produced a fragment of imbrex.

On the same east-west alignment as wall 3062 was the robbed-out remains of a wall, 332. There was approximately 0.3 m between wall 3062 and the robber cut, 3138, which removed wall 332. In the absence of evidence linking either of them to possible structures, it is possible that walls 332 and 3062 represent different iterations of the same plot boundary. 58

Janet Phillips and Pete Wilson: The excavation

seems likely that this deposit may be up-cast relating to the construction of Structure F. Cut into the southern ground surface deposit, 3060, was 3075 a poorly defined construction cut for an east-west aligned wall, 3061. All that remained over most of the length of the wall was a single course of pitched rubble oolithic limestone foundation stones with no bonding material. The wall was 0.85 m wide, with c. 1.8 m of the wall visible in the trench. One squared facing stone survived from the second course; the rest having been robbed at a later date. At ninety degrees to wall 3061 was a short length of a north-south return wall, 3136. It was bonded into 3061 and had the same single rough foundation course remaining. 3136, built within construction cut 3135, was 1 m wide and the southernmost 2 m remained in situ, while the northern part had been extensively robbed.

Figure 89: Area 3 – Bird burial 316 – geese skulls

As the structure of wall 332 had been heavily robbed by undated later activity, 3138, the clearest surviving evidence for it was the surviving remnant of its construction cut, 3087. 3087 which was aligned eastwest, was recognisable for 2 m, and was shown to be 0.55 m wide and 0.20 m deep. 3082 had a steep southern edge suggesting that wall 332 was originally built up against it. A 0.20 m - 0.36 m thick mortar layer, 3083/3088, was present in the base with some evidence of degraded limestone surviving in situ. Pottery dating to the late 4th to early 5th century was recovered from 3083, presumably introduced during the robbing of the wall material.

The remains of a possible internal wall, 3092, were located after surface layer 331 was removed (Figure 92). Unfortunately, the level of robbing and modern disturbance in this area obscured the relationship between 3092 and the outer walls 3061 and 3136. A steeply sloping, 0.85 m wide, construction cut (3091) aligned east-west extended out from the trench edge for 0.75 m. A mortar bedding layer formed the primary fill, with the possible remnants of two courses of roughly hewn limestone walling left in situ. 3092 was 0.32 m wide and consisted of two facing stones with no core present. Fill 3089, of construction cut 3091, produced pottery dating to the late 4th to early 5th centuries, suggesting further late Roman robbing of structures on the site.

Structure D or E (Area 8) A short length of wall was identified within Area 8. Wall 821 clearly consisted of the same roughly hewn limestone with lime mortar bonding seen in both Structures D and Structure E. However, only one course was visible in the available area. It is not possible to definitely assign this wall to either structure (Figure 75).

A wall, 3118, was located within the small service trench on the eastern side of Area 3 and it is possible that it represents a return of one of the walls associated with Structure F (Figures 82, 90). 3118 was aligned east-west and comprised of two courses of roughly cut limestone blocks, up to 0.25 m x 0.18 m x 0.24 m in size, with no clear bonding material. No facing stones could be identified although there may have been a stepped foundation on the southern side.

Structure F This structure was located in the southern part of Area 3, c. 5.50 m from walls 313 and 327 of Structure E (Figures 90, 91). Unfortunately, only a small portion of its walls were visible, however they were substantial and appeared to be external, suggesting this structure was smaller than Structures D and E, perhaps forming an associated outbuilding. A possible internal wall was located, and the remains of a stone surface was identified, along with possibly associated rubbish pits and a sheep burial.

An area of mixed size worn stone, 331, was located within and to the north of Structure F and appeared to respect 3092, the possible internal wall (Figures 82, 90). It is possible that it represented remnants of contemporary internal and external surfaces, however, the extent of later robbing leaves some uncertainty. In addition, a later rubble deposit, 3044, overlying the southern part of the building masked the full extent of 331 in that direction.

Overlying the probable ground surface deposit, 3084, was an 0.14 m thick area of, possibly redeposited, natural, 3085. Although no direct link was established it

A layer of mortar, 3119, was identified within the service trench to the east of Area 3. It appeared to represent 59

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Fig. 82

N 3115 822 313 327

823 Area 8

3116

825 3114 332 339

Modern service trench

331

3115

3080

Fig. 21

3091 339=372

3092 331 3135

3136

3044

Fig. 22

Area 7

3117

Modern service trench

3061 3075

3118

3119

Fig. 82

Fig. 23

2m

0 scale 1:50 @ A4

Figure 90: Area 3 – Structure F

part of a mortar floor similar to those identified in Structure D. The floor, which was 0.02 m thick, butted wall 3118 and was observed for 0.35 m in the bottom of the trench (Figure 82).

Features possibly contemporary with Structure F The first of these, sheep burial 396, cannot be directly linked to Structure F due to the level of disturbance in the area but it has been tentatively assigned here as it is cut into deposit 3085 which is associated with the construction of Structure F. An oval cut (396) was seen 60

Janet Phillips and Pete Wilson: The excavation

Figure 91: Area 3 – Structure F (view to east)

in the western trench edge which measured 0.52 m by 0.24 m in area and was 0.18 m deep. The cut contained a young sheep skeleton (SK376) aged 10-12 months which was lifted for carbon dating and produced a date 284 -322 cal AD (@ 8.8% confidence) (128 -258 cal AD @ 86.6% confidence) {Beta Analytic Laboratory number: Beta – 489837, Sample: 35N16C376B}. It is possible reused stone roof tiles had been laid over the top of the skeleton, however the feature was overlain by rubble deposit 333 and the tiles could have been part of that layer (Figure 77).

Figure 92: Area 3 Structure F – possible internal wall 3092 (view to east)

steep-sided, 0.52 m deep with a rounded base and in plan was 0.58 m wide by 0.18+ m. Its lowest fills, 3093 and 3039 respectively contained ashy and burnt inclusions, with the latter also producing pottery dating to the late 3rd to 4th centuries.

Three other pits have a similarly uncertain relationship to Structure F. Pit 3090 butted wall 3062 (Figure 77), suggesting it could be contemporary with Structure F, but it continued beyond the western trench edge. It was

Similar ashy material was identified in pit 3080, located 2.25 m to the south of 3090, and cutting deposit 385 associated with the construction of Structure F. As

Sec. 74 701 702=730

702=730 north

738

735

737

735

732

729

south

734

736=705 733

739

0

740

1

Metres

scale 1:20 @ A4

Figure 93: Area 7 extension – West-facing trench edge section

61

739

740

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire recorded 3080 was shallow and 0.48 m by 0.29 m in area and truncated by modern service.

roof tiles laid flat with tiles on edge forming a border (431), all set into mortar bedding layer 434 (Figures 24, 94). The eastern part of the feature had been truncated at a later date, with the remaining portion forming a crescent shape which occupied the full 1 m width of the trench and was 0.60 m wide and 0.15 m thick. The tiles set on edge showed considerable wear suggesting that 431 had formed the threshold of a doorway giving access to the structure, or possibly between rooms within it, although initial investigation had suggested that it might represent part of a hearth.

It is possible that pit 713, identified in the base of a service trench that was a machine-excavated extension south from Area 7 (Figure 4 – inset), is also related to this phase of activity. 713, which was seen below the deposits recorded in Figure 93, was not fully exposed but appeared to be rounded with steep almost straight sides, 0.49 m wide and c. 0.45 m deep with a flat base. Structure G (Area 4 Southern verge)

One of the earliest deposits associated with this structure was 432/446/930, a possible compacted sand floor or bedding layer for a now removed floor surface. Its eastern edge was 9 m to the west of the Roman road, and it extended west for 6.75 m under walls 430 and 924. 432 did not appear to continue beyond a later robbing event, 4049, which probably had removed an external wall. Deposit 930 produced pottery dating to the 3rd century.

This structure comprised a stone wall 14.50 m to the west of the road, with a robbed-out possibly external wall 8.50 m west of the road (Figures 53, 94). Probable floor deposits, a hearth or more probably threshold and a possible internal division were all identified. Structure G was located towards the western end of the service trench at one of its shallowest points, as a result the full depth of archaeological deposits were not excavated. Additional deposits associated with this structure were identified in Area 9.

Overlying, and probably cutting, deposit 432 was a north-south aligned stone wall, 430, located 14.50 m to the west of the Roman road (Figures 24, 94). Stone roof tile fragments, oyster and mussel shell and pottery

Located on the eastern side of Structure G and 8.50 m west of the Roman road, was an area of reused stone

4024

N

Modern tarmac road

4032

Fig.

verge

24

426

924

427

431 939 432

. Fig

430

24

verge

925

923 0

Figure 94: Area 4 and 9 – Structure G

62

scale 1:50 @ A4

2m

Janet Phillips and Pete Wilson: The excavation

dating from the late 3rd to 4th century were recovered from the immediate area of the wall. The wall was constructed from a single thickness of roughly cut limestone blocks up to 0.28 m by 0.26 m by 0.20 m in size with evidence of lime mortar bonding. No construction cut could be identified and, in comparison to the other walls excavated, it appeared much less substantial, perhaps suggesting that it was an internal, rather than external, wall.

side and a roughly flat base. A considerable amount of stonework, 4053, was present within it in section, although damage during machining meant that it could not with certainty be identified as in situ walling. However, the stonework did not appear to be dumped and was most likely structural. Wall 4053 appeared to have up to four courses remaining in situ at the section, although the central coursing had been knocked out of place by the machine. The wall was roughly constructed against the stepped western edge of the construction cut, with no apparent bonding material and incorporating roughly hewn and irregular sizes of limestone. The backfill (4006) of the construction cut appeared to have been tipped in from the east. This irregularity again perhaps suggesting this structure could have belonged to a later sub-phase of construction along with Structure I in Area 3 (pp. 66-68).

A further possible wall aligned north-east to southwest was identified c. 1 m to the west of wall 430. Again, no clear construction cut could be identified. As recorded wall 924 was 0.30 m wide and appeared to be built on a rubble foundation (928). 928 consisted of large irregular pieces of limestone up to 0.75 m length and incorporated pottery dating to the 3rd to 4th centuries. The superstructure of wall 924 comprised two courses of roughly hewn limestone blocks with no apparent bonding material, and the lower course stepped out on the western side (Figure 94). However, as Areas 4 and 9 were excavated at different times, and under difficult circumstances, it is possible that the two features were part of a single wall, although this could not be demonstrated on site. If 924 was a separate feature its small size suggests that it could have, at best, supported a wooden superstructure. A 0.04 m thick firm, yellow-white mortar deposit, 925, representing a probable floor, was located to the west of wall 924. 925 appeared to drop westwards, possibly subsiding into a continuation of Phase 2 ditch 397.

Four probable post-holes, 486, 492, 494 and 496 were identified in section located within 3.2 m of wall 4053. The postholes were cut into the ground levelling deposits 480 and 481 and were irregularly spaced and cut to varying depths. The two outer postholes, 486 and 496 were similar in size with diameters of 0.10 m - 0.20 m and 0.45 m deep. The two central postholes, 492 and 494 were 0.10-0.12 m in diameter and 0.14-0.18 m deep. No clear postpipes were identified, and no post-packing was recorded (Figures 26, 53). A 1.85 m long and 0.14 m thick deposit of tabular stone, 485, was seen in section overlying postholes 492, 494 and 496 and possibly abutting post-hole 486 (Figures 26, 53). Although possibly the remnant of a more extensive surface, perhaps incorporating 4024 (see below), it appeared to be concentrated around the postholes, perhaps suggesting a small area of hardstanding was needed for an activity associated with them. A very large masonry block, 4028, was recovered from close to this deposit and could have represented a further structural element, possibly a post pad. 4028 was roughly cut and 0.50 m x 0.45 m x 0.13 m in size, with a flat base and slightly rounded top which showed signs of wear.

Structure H (Area 4 Northern verge) A series of features within the area of the northern verge of the access road may have formed elements of a further building, Structure H. Their proximity and lack of other activity nearby suggests they are at least related, whatever they represent. The location of the features shows evidence of possible ground levelling prior to four postholes and a possible wall being constructed (Figures 26, 53). A series of ground levelling deposits overlay 482, ground build-up after construction of the Roman road. The earliest of them, 488, contained frequent charcoal and burnt inclusions and was 0.45 m thick and, as seen, extended for 2.4 m. 488 could relate to industrial activity similar to that which produced ash dumps 402 and 440 recorded to the east of the Roman road. 488 was overlain by 481 which in turn was sealed by 480 which incorporated frequent burnt stone inclusions, possibly redeposited hearth material.

A narrow service trench was dug to the south of the above-mentioned features identifying several more tentative aspects of Structure H. Unfortunately, the depth and width constraints of the trench made it difficult to excavate fully and record, with a collapsing section forcing recording to stop. The earliest deposit identified was 4027 which may have been the same as, or equivalent to layer 482 (above). 4027 was overlain by 4025/4026, a 0.08 m thick compacted, pale-brown, stony deposit which appeared to form a bedding layer for a cobbled/metalled surface above 4024 (also recorded as 4030, 4031 and 4033). Surface 4024 was possibly a continuation of 485 and comprised moderately sized roughly cut limestone with the upper areas smoothed

A linear feature, 497, was identified c. 2 m to the west of the Roman road cutting through levelling deposit 488 (Figures 26, 53). 497 was 1.18 m wide and 0.70 m deep, with a stepped western side, slightly concave eastern 63

N

466

465

MAP Trench

9 Fig. 2

466

464

4033

64 0

scale 1:50 @ A4

4031 = 4033

4034

4023

Fig. 96

2m

Figure 95: Area 4 – Northern verge trenches western end

4024

4034

6

Fig. 2

463

462

Fig. 2

461 6

9 Fig. 2

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Janet Phillips and Pete Wilson: The excavation

keep traffic away from the nearby structures or as a result of repair work. A narrow, 0.50 m wide and 0.11 m deep, linear depression, 419, was visible in road surface 418, that was initially interpreted as a wheel rut. However, the feature appeared to be straight and regular, as if stonework had been removed (Figures 24, 29). Its fill, 414, produced stone roof tile fragments, oyster shells and pottery dating to the 4th century, along with intrusive post-medieval brick fragments. The possibility of 419 representing the location of a stone kerb has been suggested (M Haken, Roman Roads Research Association, pers. comm.), and it is possible that as part of the works on the road at this time the kerb was removed and a roadside ditch created to replace it.

Sec. 90 4034 4035

west

4033

4035

0

4036

east

1m scale 1:20 @ A4

Figure 96: Area 4 – Section through possible wall 4034

In Area 4, on the southern verge of the access road, a U-shaped ditch 415 was found cut into the western edge of the Roman road with the upcast, 413, appearing to form a bank on its western side (Figures 24, 29). The ditch was 1.66 m wide at the top and 0.46 m deep. Its upper fill, 411, produced stone roof tile fragments, oyster shell and pottery dated to the late 4th century, including a painted Crambeck red ware bowl. The same U-shaped ditch was identified as 498 in the service trenches to the north of the access road where it was recorded as being 1.04 m wide and 0.48 m deep. The ditch was cut through surface 463 which appeared to have remained in use.

with use, surrounded by crushed stone. The surface was covered by a layer of soft, mid grey-brown, sandy-silt, 4036, of 0.06 m thickness, which appeared to be a buildup deposit accumulated through usage. Layer 4036 was overlain by a 0.03 m thick mortar bedding layer, 4035, on which was set a roughly constructed possible wall, 4034 (Figure 95), suggesting the potential existence of a later, otherwise unknown, structure. A single course of 4034 remained in situ which was also seen in a later extension to the main trench in the verge (Figure 96). 4034 was damaged and/or robbed out in places but appeared to be aligned roughly northwest to south-east, that is obliquely to the line of the Roman road.

On the eastern side of the Roman road no structures or features were identified excepting later dumps of rubble and pit 404 mentioned above. There were however several large dumps of very ashy material in the trenches in the northern and southern verges of the access road. The extent of these deposits is perhaps suggestive of an industrial process taking place nearby with the waste material being dumped on open ground over the road, or they could possibly relate to the earlier phase of cremations (Figure 24).

The only definite feature identified to the east of the Roman road was a small pit, 404, which produced undiagnostic Roman-period pottery. Located to the west of the Roman road in the northern verge was a 6 m long area of apparently deliberately placed flat stones with crushed and smaller stones filling the gaps, 462/463/464, that possibly formed a surface. Pottery dating from the 3rd or 4th to the 5th centuries, oyster shell, stone roof tile and Roman brick and box flue tile was recovered from 462 and 463. However, a sherd of intrusive 19th-century plant pot was also recovered from 462 and therefore it possible that the later Roman pottery could also be intrusive. No indication of a structure was found, and it is unclear if the possible surface represented an internal feature or part of an area of yardage. The full depth of the surface could not be ascertained as it extended below the permitted depth of the trench.

Within the southern verge of the access road a series of ashy deposits overlay the clay shoulder, 443, of the road. Deposit 440 was a loose, mid-grey, ashy siltsand 0.28 m thick which extended for 5.80 m. This was overlain by a similar deposit, 408, of loose, light greybrown, ashy silt-sand of 0.66 m thick and recorded for 6.75 m which contained pottery dating to the late 3rd to 4th centuries, as well as oyster and limpet shell. The nature of these deposits suggests that they may have been related to industrial processes located nearby (Figure 24). In the trenches on the northern verge similar deposits, 454, 455, 4051 and 4052 were located over the eastern part of the compacted core of the road, 4004. All the layers contained ashy and burnt material and 454 produced a single sherd of a samian mortarium dating to AD 150+ (Figure 29).

Roman Road There appears to have been a redefinition of the western roadside ditch during this phase, perhaps to 65

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire This build-up of ashy material continued in the second of the northern verge trenches over metalling 4015: layers 4012, 4013, 4014, 4018 and 4021 (Figure 26).

running into the eastern trench edge. It was 0.40 m wide with only a single course remaining in situ. An isolated patch of mortar, 392, 0.13 m wide and 0.06 m thick, was located within the area defined by the south-eastern part of the wall. A similar area of, possibly degraded, mortar, 391, was also identified, however it is unclear if this was in situ or was material accumulated after the structure was abandoned.

A small extension to Area 7 revealed the remains of a stone surface, 732 (Figure 93). The stones had been laid on a bedding layer, 734, which produced pottery dating from the 3rd century onwards. Overlying 734 was the probable surface, 732, consisting of roughly shaped, worn flat-laid stones of up to 0.10 m x 0.04 m in size.

Turning through 90o, again closely following the earlier walls of Structure E, was a length of a northsouth return wall, 354. Due to the rough construction of the walls it is uncertain if 345 and 355 were bonded together. As with wall 355, 354 appeared to cut mortar floor 356 but was not well-defined. A previously laid 0.06 m thick deposit of coarse mortar, 386, acted as a bedding layer under the southern part of construction cut 3131 associated with wall 354. However, mortar 386 faded away before reaching return wall 355. Wall (354) was as roughly constructed as 355, but displayed more evidence of the reuse of larger limestone blocks up to 0.35 m by 0.24 m in size (Figures 75, 97-99).

PHASE 4 - ROMANO-BRITISH 4 (4th century) The continuing use of the area can be directly linked to the pattern of expansion and contraction of the Roman settlement, reflecting its roadside location. Following the previous phase of substantial activity, the area then seems to have gradually been abandoned, probably reflecting a contraction of the urban area. Whether by deliberate clearance or simple abandonment, Structures A to F appear to have gone out of use, with all of them showing signs of robbing and removal of stonework. However, the roadside location meant that some interest was maintained in the area with possible late or even post Roman activity continuing on the site in the form of Structure I, possibly Structure H (pp. 63-65) and perhaps a further structure incorporating wall 4053 (p. 63), constructed prior to or during the demolition or abandonment of the pre-existing structures. Interestingly there is no rubble build up over the road confirming it continued to be used after the disuse of the nearby structures. However, rubble deposits were present on the eastern side of the road and the origins of these are unclear.

Although it likely that such a structure would make use of existing floors where possible, a mortar layer, 377 was identified which may belong to Structure I, overlying floor 356 of Structure E. Unfortunately, there was not a clear relationship between 377 and the walls of Structure I and it could have been a final repair of the floor for Structure E. 377 was up to 0.07 m thick, and as seen consisted of patches of mortar which occupied an area 1.76 m by 3.10 m and areas of coarse sand and gravel which may represent degraded mortar, or bedding for no longer extant mortar.

Structure I (Area 3) (Figures 53, 75, 78, 97)

A rough hearth was partially exposed c. 1.30 m to the south of wall 355. The feature lay on the eastern limit of excavation and was truncated by a Victorian/modern pit on its western side which had removed the majority of its stratigraphic relationships. However, the feature was not seen until floor 377 was exposed and therefore it is likely that it belongs to this structure. The burnt area extended for 0.30 m into the trench and for 2 m north-south along the trench edge. It comprised of a series of burnt layers and one visible reused, heavily burnt, stone roof tile, 3012, with only 0.20 m x 0.35 m of it visible (Figure 100). The burnt material around the stone was sampled (environmental sample 31) and found to contain mostly cinder, frequent coal, vitreous and amorphous slag, traces of mortar, indeterminate bone fragments, decayed wood, ceramic building material, fragments of corroded iron nails and some spheroid hammerscale.

A roughly built structure was recorded within Structure E, having been constructed against its southern side wall, 314, and suggesting a possible further sub-phase of construction, possibly contemporary with Structure H and perhaps wall 4053. The two walls identified were poorly constructed with no clear coursing or finishing perhaps suggestive of occupation making use of the nearby abandoned resources. A possible surface and hearth may also be associated with this structure. An east-west aligned construction cut, 3001, was partially visible cutting into mortar floor 356 of Structure E. However, due to 3001 being inserted against existing wall 314 and impacted on by Victorian and modern disturbance it was not fully understood. Wall 355 appeared to have been constructed from a mix of reused roughly cut limestone blocks and smaller unshaped pieces and, in comparison to the other walls found during the excavation, it is poorly built. A single sherd of pottery dating to the 3rd century onwards was recovered from 355. 355 was recorded for c. 5 m before

Two heavily burnt deposits, 3013 and 3010, possibly the result of raking out the hearth, were located to the west of stone tile 3012. The earlier deposit, 3010, consisted of compacted, heavily burnt black and red sand, and was 66

Janet Phillips and Pete Wilson: The excavation

Figure 97: Area 3 Structure I – Walls 354 and 355

Figure 98: Area 3 – Section through walls 354 (Structure I) and 353 (Structure D). Wall 314 of Structure E removed. Wall 355 (Structure I) is under the ranging pole on the left of the frame.

Sec. 37 west

349 3012

3013 3014

3010 3024

24.82m O.D.

3008

371 east

3015

0

1 scale 1:20 @ A4 Figure 99: Area 3 – Profile of wall 355 (Structure I)

67

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Sec. 32 354

387 south west

24.84 m O.D.

386

388

386

387

north east

3131 0

1 scale 1:20 @ A4 Figure 100: Area 3 – Hearth within Structure I

0.29 m wide and 0.04 m thick. This deposit then appears to have been cut or scooped out close to the hearth stone, suggesting that the hearth continued in use, with 3013 representing a further phase of rake-out. Deposit 3013 was a linear spread of compact, burnt black and red sand, environmental sampling produced flecks of charcoal and vitreous slag. A further ashy deposit, 349, was located to the south-east of the hearth stone and also appeared to be rake-out material.

variety of sizes of rubble, with stone roof tile, oyster shells and pottery dating to the late 3rd to early 4th centuries. 209 contained Roman brick fragments and 52 fragments of animal bone including cow, pig and horse with evidence of gnawing, butchery and burning. A single residual flint bladelet (RF 320), possibly dating from the Later Mesolithic, was recovered from 214, along with 46 fragments of animal bone including cow, pig, chicken and small invertebrate bone, again with evidence of some butchery, burning and gnawing.

Demolition or abandonment deposits

A further two rubble spreads were located to the north of wall 219 in the central area of Structure A. Spreads 228 and 295 were made up of mixed building rubble, stone roof tiles, Roman brick fragments and oyster shell. Both produced pottery dated to the late 3rd to 4th centuries. Thirty-two fragments of animal bone, including cattle bones, were recovered, some with evidence of butchery and gnawing. Spread 228 appeared to end in a very straight north-south aligned edge. It would seem possible that this material was initially butting up to a wall that was later robbed out leaving a gap in the rubble, feature 234, which produced pottery dating to the 4th century.

Structure A (Area 2) Large rubble deposits were present throughout Structure A, overlying both the external walls and internal features. All of the rubble deposits contained a mix of crushed/degraded limestone, stone roof tile fragments and pieces of oolithic limestone of varying size. Very few larger stones were visible in the rubble deposits, with most having been robbed out and presumably removed from the site. Rubble spread 204 was located approximately 1.50 m to the west of Structure A and extended for 3.25 m eastwest and 5.50 m north-south. The spread contained frequent fragments of oolithic limestone building rubble along with fragments of stone roof and floor tiles. Roman pottery recovered from the deposit dated to the late 3rd to 4th century, although with some intrusive modern sherds. The area of the deposit had fairly distinct edges and did not merge into the nearby remains of Structure A suggesting it was not from collapse but may have been from sorting material as the structure was robbed out.

In the eastern part of Structure A two more rubble deposits were identified, 235 and 233. 235 was located approximately 0.70 m from rubble spread 228 and again had a very straight north-south aligned edge, suggesting this spread was contemporary with 228 and that it too was deposited prior to the wall being robbed out. Spread 233 was located inside the eastern wall of Structure A. Both 233 and 235 produced stone building material, stone roof tile fragments and, in the case of 233, Roman brick, imbrices and tegulae. Further finds include oyster shell, fragments of Helix sp (edible snail?) and substantial quantities of pottery, much of it residual, but with material dating to the 4th century and one sherd of probable late 4th- to early 5th-century

Rubble spread 209/214 was located in the western part of Structure A, to the north of wall 205. It contained a 68

Janet Phillips and Pete Wilson: The excavation

date from 233. A moderate quantity of animal bone was also recovered, 49 fragments from 233 and 95 from 235, including cattle, pig, horse, dog and large mammal some of which showed evidence of butchering, burning and scavenging.

bone came from 215 which, on balance, is perhaps most likely to be an intrusive from disturbance of the Phase 2 burials, perhaps by scavenging animals. Alternatively, given the presence of the near complete pot (Chapter 4, No. 42), the miniature jar (Chapter 4, No. 136) and a shale pendant (Chapter 7, No. 125) in 215, it is possible that the bone could have derived from a much later, otherwise unrecorded, burial.

A small area of a rubble deposit, 230, extended east from wall 232 into the trench edge. 230 consisted of small fragments of building material, stone roof tile, Roman brick and tegulae, and incorporated a large group of pottery, including Crambeck Parchment ware, oyster shell, china limpet shell, and large Helix sp (possibly edible snail). Fifty-one fragments of animal bone were recovered including cattle, pig and horse bone some of which demonstrated butchery marks and damage from gnawing.

Structure C (Area 2) (Figure 68) Although some material had to be cleared to reveal the kiln, it had been subject to less disturbance and robbing out than Structures A and B with most of the stones remaining in-situ. This may suggest the feature remained in use for a longer period than the buildings, or was a later feature, possibly even built after Structures A and B were robbed, and potentially reusing some of the robbed building material. However, there was a patch of rubble 1 m east of the kiln, 248, which extended out of the excavation area, but was at least 1.50 m east-west and 0.50 m north-south. Layer 248 produced 31 fragments of animal bone including cattle, sheep/goat, pig and dog, some of which showed evidence of butchery and gnawing, as well as pottery dating to the 4th century and oyster shell. Given its similarity to the rubble deposits found across the site it is reasonable to suggest that it is further evidence of demolition or abandonment, although it cannot be definitively linked with Structure C.

An isolated area of rubble, 208, was identified to the south of Structure A. 208 contained a mixed array of stone building material including stone roof tiles, however, unlike the other spreads, larger blocks were also present, measuring up to 0.45 m x 0.65 m in size. This may reflect the sorting and reuse of material following the abandonment of the buildings, possibly suggesting systematic dismantling of the structures when the area became a dump. The deposit yielded seven fragments of bone, including pig and large mammals, with some evidence of burning, as well as pottery dating to the late 4th to early 5th centuries. Structure B (Area 2)

Structure D (Area 3)

Rubble spread 215 lay between Structures A and B, however, if wall 254 does represent a plot boundary, 215 should be seen as associated with Structure B. Initially 215 was thought to be a wall; however, on investigation the quantity of finds recovered and mixed nature of the stone building material, suggested it was a dump or spread. In addition to very frequent stone inclusions, the deposit incorporated finds of oyster shell, stone roof and floor tile, imbrex fragments and a carved stone block (RF 177). The pottery from 215 dated to the late 4th to 5th centuries and included part of a head from a cockerel figurine in Crambeck Parchment ware (Chapter 4, No. 28), a miniature calcite-gritted jar (Chapter 4, No. 136), proto-Huntcliff and Huntcliff types jars. In addition there was a near complete large calcite-gritted lug-handled jar, lacking its rim, with a white internal deposit and a pierced lower wall possibly for a spigot (recorded as context 225). One hundred and thirtynine fragments of animal bone were also recovered including cattle, horse, cat and bird bone some of which demonstrated butchery, burning or gnawing. Environmental sample (3) from 225 revealed traces of coal, cinder, mortar crumbs, amorphous slag, a single spheroid of possible hammerscale, small vertebrate bone fragments, a blackberry/raspberry fruit stone and elder fruit fragments. A single fragment of human

Unfortunately few demolition or abandonment deposits could be specifically linked to Structure D. However, there was robbing of the associated walls, and some further deliberate robbing events were noted. The possible plot division wall, 308, to the north of Structure D, was robbed out across most of its visible length. A robber cut, 3102, 1.40 m wide and 0.54 m deep was present in section (Figures 17, 76, 83, 84). Its fill, 305, contained a large piece of limestone, 0.35 m x 0.28 m x 0.15 m in size, presumably from the robbedout wall, and pottery dating to the late 4th to early 5th centuries, as well as two fragments of sheep/goat and three fragments of large mammal bone, also mussel and oyster fragments. Located to the north of the robber cut was an area of mixed stone rubble, 306, which may have been the result of the robbing activity. It produced pottery of the late 3rd to 4th centuries, a fragment of large mammal bone and stone roof tile fragments. In Area 8, rubble deposit 818/819 was possibly equivalent to 306. Environmental sample 91, from 818, contained frequent fragments of heat-affected clay, traces of coal and cinder and a single charred barley grain. 69

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire In Area 3 a further probable robber trench, 341, was identified running east-west parallel to, and 0.56 m north of, walls 313 and 327 of Structure E. 341, which crossed the full 5.25 m width of Area 3, had straight steep sides, a flat bottom and was 0.86 m wide (Figure 77). The stone of the wall was entirely robbed-out, but it contained a primary fill of compacted, orange-grey sandy mortar, 367, 0.13 m thick that remained in situ. Over 367 backfill 342 contained pottery dating to the late 4th century or later. From its location it would seem likely that 341 had removed the southern wall of Structure D. However, this would mean that internally Structure D would have been slightly over 12 m wide and, presumably therefore would have incorporated internal supports, suggesting that it was either aisled, or had, in the absence of any evidence of footings for a spine wall, a line of central posts.

two mixed rubble deposits, 820 and 822. 820 contained pottery dating from the late 3rd to 4th century and stone roof tile fragments. Structure F Again robbing and later disturbance made the identification of demolition or abandonment deposits directly associated with the building difficult. A rubble spread, 3044/3045/3076, (Figure 101) partially obscured the walls and surviving floor deposits and sealed the majority of the structure. It overlay stone surface 331, as well as walls 3061 and 3136. The spread incorporated considerable quantities of mixed stone and also included roof tile, pottery dating to the late 3rd to 4th centuries and fragments of Roman brick. 3044 also produced an intrusive 13th to 16th century fragment of plain tile.

A small pit, 3127, was identified in the trench edge section cutting through floor deposit 3107 within Structure D (Figure 77). Although no function for the pit could be determined with certainty it is possible that it relates to the removal of an internal feature or post. The full extent of pit 3127 was unclear, but as recorded it was 0.66 m wide and 0.36 m deep with rounded sides and bottom.

Layer 3074, a further 0.17 m thick probable dumped deposit overlay 3044 and also contained pottery of the late 3rd to 4th centuries, as well as stone roof tile. North of wall 3118, the service trench on the eastern side of Area 3 was occupied by a rubble deposit, 3117, which dipped down to the north of the wall and contained pottery dating from the 3rd century onwards. In addition to the Roman material, there was some intrusive 19th-century clay pipe, 13th- to 16thcentury ridge tile fragments, as well as a sherd of Gritty ware dating to the 13th to 14th centuries.

Structure E (Area 3) As with Structures D and F, the extent of robbing activity, the proximity of the remains to the surface and the intensity of modern activity had largely removed any rubble layers relating to demolition or abandonment. However, overlying wall 327 was a mixed deposit of rubble and dumped material, 329. Within Area 8 wall 821, also recorded as 313 in Area 3, was overlain by

A further four deposits were identified in the service trench extending from Area 7 past the southern part of Area 3 (Figure 102 – exact location unrecorded on site). These deposits have been tentatively associated with Structure F due to their proximity to it. Overlying ground build-up layer 760 were layers 759 and 762, both of which contained large quantities of mixed limestone building material, oyster shell, stone roof tile and 3rd to 4th century pottery. Both deposits were relatively shallow, and they may represent the result of sorting of demolition material following the abandonment of the structure.

Figure 101: Area 3 – Rubble spread 344 (view to east)

70

Also overlying 760 was an area of mixed limestone and mortar, 756, which contained pottery dated to the 3rd to

Janet Phillips and Pete Wilson: The excavation

Sec. 87 west

763 761

701

730 759

764

Modern services

756

east

760

740

0

scale 1:50 @ A4

2m

Figure 102: Area 7 – Service trench section

4th centuries and fragments of Roman brick and stone floor tile. This may be a similar rubble deposit to 759 and 762, or possibly represent a surface made of reused material. Unfortunately, given the limitations of the service trench 756 could not be further investigated. A further small patch of dumped mortar, 758, containing one sherd of 3rd to 4th century pottery, oyster shell, stone roof tile and fragment of Roman imbrex with a knife cut hole possibly designed for a chimney or finial attachment, was also identified within these dumps of stone.

western part of a deposit that included layer 919 (see below), which as a group dipped away, again possibly into a continuation of ditch 397. 923 produced pottery dating to the late 3rd to 4th century. The threshold of Structure G, 431, (Figure 24) was sealed by a 0.22 m thick deposit incorporating much building rubble, 426. Layer 426 produced 4th-century pottery, stone roof tile fragments, oyster shell and china limpet shell. Overlying 426 was layer 427, also recorded as 919 in Area 9. This deposit, which was up to 0.40 m thick, included very frequent rubble inclusions and extended east from wall 430 for 8.75 m to seal wall 924. Both 427 and 919 contained 4th-century pottery along with oyster fragments, Roman imbrex fragments and stone roof tile.

Structure G (Figure 24) Two deposits were identified which both lacked rubble and were noticeably siltier in nature, perhaps suggesting a period of dereliction prior to demolition or abandonment, to the west of wall 924. Deposit 926 directly overlay the probable floor, 925, and was a soft reddish silty-sand of 0.15 m thickness stretching for c. 1.50 m from which fragments of Roman imbrices were recovered. Overlying 926 was layer 920 which was 0.31 m thick. Both layers dipped down to the west, possibly into a continuation of ditch 397.

A deposit of dark grey-black, silt-sand with possible charcoal inclusions, 448 (Figure 24), which extended for 2 m, was identified as one of the lowest deposits reached in the access road southern verge trench. Approximately 0.06 m of it depth was excavated before the working depth limit was reached. The deposit underlies the later Structure H, and no direct link can be made to any other features, it is possible the deposit represents the Phase 1 ground level as seen in other areas of the site. Environmental sample (64) found the presence of ash, rectilinear charcoal, identified as heather and alder/birch/hazel, a single charred grain possibly of wheat and four indeterminate bone fragments.

As 926 dipped westwards it merged with deposit 921, as recorded as 922. 921/922 overlay 937, and with 937 formed a 0.33 m thick area of concentrated stone rubble containing a sherd of burnished grey-ware dating to the 3rd to 4th centuries. Within the deposit there were intermittent patches of burnt stone and environmental sample 76, from 937, produced burnt material, including ash, rectilinear charcoal and some cinder.

Two small pits or postholes, 435 and 437, were identified cutting into 448, but due to depth restrictions were not investigated further. Both pits/postholes were roughly circular with a 0.40 m diameter and loose, mid-dark brown, silty-sand fills, 436 and 438 respectively. The pits lay within 0.25 m of each other, partially beneath 431, the possible threshold of Structure G. Their close

To the south of 921 was an area of compact material, including stonework with attached mortar, that may have been laid as a surface, 923, or possibly represent material displaced from mortar floor 925 associated with Structure G. This material may represent the 71

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire proximity suggests that they might have been related to Structure G, possibly in its construction phase.

identified in the central area of the trench, partially overlying mortar floor 356 of Structure E. 3009 incorporated frequent stone roof tile fragments, oyster shell and pottery dating to the late 3rd to 4th centuries. Also, partially overlying floor deposit 356 and feature 341, the robber cut for the south wall of Structure D, was layer 340 which consisted of mixed stone rubble and debris and extended for 4.30 m across the trench to a depth of 0.16 m.

Rubble deposits extended east from Structure G towards the Roman road and, although they are not directly linked to the structure, it is probable that they originated from it. A small patch of stone rubble with attached mortar, 450, 0.60 m thick and 0.90 m long, lay 3 m east of the threshold, or hearth 431. Although appearing structural there was no coursing and the material is likely to have been redeposited. Overlying rubble deposit 427 on its eastern side and partially overlying rubble 450 was a 0.26 m thick and 1.82 m long deposit containing frequent stone rubble, 447 (Figure 24).

Above 340 there was another large rubble spread, 3081, which was up to 0.21 m, deep, that occupied the full width of the middle and southern part of Area 3 and extended for over 3.35 m north-south. 3081 contained large pieces of stone building material and stone tiles. Towards the southern end of the trench 3081 was sealed by a similar rubble deposit, 333/3040, that incorporated frequent mixed-size stone.

A further rubble-like deposit, 451/927/4032, was located between walls 924 and 3083 and produced pottery dating to the late 3rd to 4th centuries from 927 and 4032. It is possible that this material represented a remnant of a damaged internal or external surface as it was level. However, it was not possible to confirm this given the limited extent of the trench (Figure 24).

In Area 8, overlying the east-west wall 817, which belonged to either Structure D or E, was a mixed rubble and mortar deposit, 818/819, which extended for 2.60 m and was 0.28 m thick. Further rubble/demolition deposits were noted in the Area 8 service trenches – 804, 805, 806 and 809 – all of which contained mixed limestone and mortar.

Cutting through rubble deposit 427 was a probable stone-robbing trench, 4049, possibly dug to remove an external wall of Structure H (Figure 24). As it was cut through the rubble deposits, it indicates that the robbing of the wall would have been completed some time after the abandonment of the building, but prior to the later Phase 5 soil build-up and dumping. The U-shaped cut was 1.36 m wide and at least 0.58 m deep, its fill, 423, contained mortar and stone inclusions, oyster shell and pottery dating to the late 3rd to 4th century.

In Area 3 two rubble deposits, the earlier one, 348, perhaps representing robbing and sorting of materials, was located in the central area of the eastern trench edge. 348 which overlay the possible mortar floor 377 of Structure G, was a loose, roughly piled, 0.10 m thick, collection of stone roof tiles covering an area c. 1 m square. It incorporated pottery dating from the 3rd century onwards. It was overlain by 347 which contained pottery dating to the late 3rd to early 4th centuries and more stone roof tile. A similar rubble deposit, 3116, located close to 348 and 347 was recorded in the small service trench to the east of Area 3.

Possible rubble relating to Structure H A mixed rubble deposit, 461 overlay probable surface 462/463/464 in the northern roadside verge. 461 itself was sealed by a dumped deposit, 466, that incorporated larger limestone fragments and late 3rd to 4th century pottery. Separated from 466 by a thin silt deposit, 465, was a pile of stone floor tiles, 467. Unfortunately, only a small portion, 0.5 m wide, of 467 was within the trench but it is a further indication that building materials were being sorted on site, presumably for re-use (Figure 29).

Rubble deposits in Area 7 A 0.16 m thick dump of degraded mortar, 711, was identified in the north-western corner of Area 7. Sealing it was a further dump of material, 709, containing frequent stone rubble including damaged roof tiles and imbrex fragments.

Wall 4034, tentatively assigned to Structure H (above) and cobbled surface 4033 were overlain by a 0.41 m thick sandy-silt 4037 which produced pottery dating to the late 4th to early 5th centuries, as well as oyster, Roman brick and stone roof tile fragments.

In the north-eastern corner a similar patch of mortar was found, 728, that incorporated considerable quantities of ashy material and evidence of burning. 728 was overlain by 704, a 0.12 m thick deposit of mixed rubble, mortar and ceramic building material. 704 appeared to be at its thickest in the south and west facing trench edge, suggesting the material may be in part derive from the mortar floors identified in Area 3 for Structures D, E and G. This deposit also overlay pit (713) (p. 62) in the west facing trench edge.

General demolition layers not associated with individual structures In Area 3 a rubble or demolition deposit, 3009, presumably associated with Structures D, E and G, was 72

Janet Phillips and Pete Wilson: The excavation

A further spread of rubble, 708, was located in the southern part of Area 7 and could have derived from any of the structures in the area. In the central part of the extension to Area 7 a rubble deposit, 729, had been truncated by modern services, but as seen was 0.15 m deep with 0.60 m visible in plan.

however there was an increased frequency of it noted within the refuse deposits, consistent with dumping rubbish in open ground. Cattle, sheep and pig bones were recovered, largely characterised by butchery marks. Grid Square 9 (Figure 103) showed the largest percentage (85%) of these, with split cattle long bones predominantly making up this figure. This may suggest a specific deposition spot for a trade, or a single deposition event by a tradesman. Evidence of craft working was present in the form of a cattle metatarsal having been worked or worn to a point, two horn cores having been chopped from a cranium, and two phalanges having knife cuts associated with skinning. Younger animals were also found including lambs and calves, with a part skeleton of a very young lamb recovered from over Structure A. These younger animals appear to have died soon after birth rather than be evidence of food waste, suggesting regular breeding of livestock.

Although no structures were identified to the east of the Roman road some rubble deposits do seem to have accumulated in this area, overlying the probable industrial deposits 440 and 408. Deposit 444, which overlay 440 and the eastern side of the road, was 0.32 m thick and at least 1.98 m long. 444 incorporated moderate quantities of stone and pottery dating to the late 3rd to 4th centuries. The easternmost of these deposits, 406/407, which overlay industrial deposit 408, contained oyster and mussel shell, Roman box flue, pottery dating from the 3rd to 4th centuries and stone roof tile. PHASE 5 - ROMANO-BRITISH (late 4th to early 5th century)

A variety of other species were also represented, including horse, hare and some probably intrusive rabbit bones. Bird bone was represented by chicken, a goose radius, crow or rook and raven bones. The dog bones recovered, mainly from Grid Square 6, appear to largely represent three individuals including a young puppy and a medium sized adult. A further few isolated and scattered dog remains were identified along with some cat bones. Evidence of deer hunting was also found with two mandibles and a metacarpal originating from roe deer. A single antler tine from a red deer was found perhaps indicative of craft activities.

When occupation on the site ceased and the buildings were demolished or left to decay, the area then appears to have become the focus of rubbish disposal from the nearby settlement. A thick spread of homogenous material was identified across Areas 2, 3, 7 and 8 overlying the earlier features, and it was from this deposit that the majority of finds were recovered. Area 2 Deposit 202 was an orange-brown, silt-sand with frequent inclusions of natural gravel and stone building material. It was present across the whole of Area 2 becoming noticeably thicker from west to east, increasing from 0.10 m to 0.42 m. The majority of finds recovered from the above Structure A were differentiated as context 206, although some were recorded as 216 which was initially thought to represent a second structure, although that proved not to be the case. The material from 202, 206 and 216 will be considered as one here.

Human Bone (see Chapter 13) A single, probably intrusive, human bone was also found, the proximal half of an adult tibia. It is perhaps most likely to have originated from disturbance of the Phase 2 burials. However, as discussed with respect to the disarticulated bone from layer 215 (p. 69), given the extent of post-Roman disturbance on the site, late burials may have existed, but did not survive intact enough to be recognised in the ground, although Trial Trench 4 did produce human leg bones during machine stripping (context TT429).

Pottery (see Chapter 4) The majority of the assemblage belongs to the late 4th to early 5th centuries, including painted Crambeck Parchment ware and Huntcliff types, with some residual late 3rd century material.

Ceramic Building Material (see Chapter 8) A range of Roman ceramic building material was recovered from this rubbish dumping event including bricks, three fragments from chimneys, including one from 206 with horizontal bands of pie-crust decoration, ridge tiles, tegulae, a single fragment of parietalis tile used to line the walls of rooms, two tesserae probably originating in a coarse tessellated pavement, fragments of box flue tiles and imbrices.

Vertebrate Remains (see Chapter 14) As expected with refuse deposits a large number of animal bones (6508) were found, with many showing evidence of butchery and scavenging. Evidence of dog gnawing was seen throughout the remains and phases, 73

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

N

GS13 GS3

GS2

GS1

GS6

GS5

GS4

GS7 GS10

0

GS14

GS9

GS8

GS12

GS11

GS15 GS16

A2

scale 1:500 @ A4

25m

Figure 103: Plan locating grid squares in Area 2

Fired clay/daub (Archive)

at Brooklyn House, or indeed any of the cemetery areas know from the Langton Road area.

Three fragments of daub/fired clay with roughly smoothed convex or flat surfaces were recovered from the deposit. A single corner fragment was recovered with rod/sail impressions indicative of wattle and daub, and all pieces demonstrated direct heat exposure. Although the fragments were not directly associated with an onsite structure, they are likely to have had a structural function.

Shell (Chapter 15) Significant quantities of oyster, mussel, limpet, common whelk and venus shell were found across the area, scattered in a consistent manner with the deposition of domestic refuse. Lithics (see Chapter 10)

Stone Building Material (see Chapter 9)

Some residual prehistoric flint tools and debitage were recovered from 202 and 206, indicative of the use of the area from the late Mesolithic onwards, probably due to its vicinity to the nearby river crossing.

A large quantity of stone roof tile, building stone and floor tiles was recovered consistent with the general debris that must have been present following the nearby structures going out of use. A micaceous sandstone tessera (RF 296) was also recovered. A possible fragment of an oolitic limestone coffin (McComish, Chapter 9, RF 96, Figure 203), given its location in context 202, could potentially derive from anywhere in Roman Malton/ Norton and have no association with the burial activity

Intrusive Material Intrusive medieval and modern pottery (see Chapter 5) and other finds were recovered from throughout this deposit (see below), emphasising the impact of medieval 74

Janet Phillips and Pete Wilson: The excavation

ploughing and the extent of modern disturbance. Material included 19th-century clay pipe, 16th- to 17th-century Frechen-Kohn Stoneware, 17th-century Hollow Ware and a variety of 19th-century coarsewares. In addition various sherds of late 19th/20th-century vessel glass were identified.

in the small extension to Area 7, and as 764 within the service trench. Pottery dating from the late 3rd to 4th centuries was recovered from 703/715 and 735. Presumably as a result of disturbance from nearby modern services 715 also produced a fragment of late 19th/20th-century green glass.

Areas 3, 7 and 8

Within Area 8 a similar deposit was recorded overlying the Romano-British features. Deposit 814/813/816/832 produced stone roof tile and pottery dating from the 3rd to 4th centuries. An isolated dump, 835, produced pottery dating from the late 4th to early 5th century, as well as stone roof tile, Roman box flue and brick fragments, which may also form part of the main dumping event. As in the dumped material in Area 2, some intrusive fragments of 19th-century clay pipe were found in deposit 813 along with a single sherd of Splash Glazed Sandy Ware dating from the early 13th to 14th century; however given the considerable modern disruption in the vicinity this is not surprising.

Within the northern part of Area 3 a deposit was identified which appeared very similar to the general dumping layer identified in Area 2. Deposit 303 was a dark, grey-brown, silt-sand with stone building material and roof tile inclusions. The deposit was present in the northern 9.60 m of the trench and was up to 0.57 m thick. A variety of finds were recovered including 137 fragments of animal bone from cattle, pigs, horses and dogs, some with butchery marks and evidence of burning or gnawing. Pottery from 303 dated from the 4th century onwards. Ceramic building material included Roman brick, ridge tiles, imbrices and tegulae. Some intrusive post-medieval and modern brick was also recovered, as well as pottery sherds from the late 18th to 19th centuries, including Creamware and Brown Salt Glazed Stoneware, reflecting the significant Victorian disturbance of the area.

Areas 4 and 9 Although there was no evidence for an equivalent deposit to 303, the main dumping layer found in Area 3, there was a build-up of material over the rubble and remnants of Structure G. Deposit 449/416, a silt-clay that incorporated moderate quantities of rubble, was a 0.28 m thick and extended for over 10 m. 416 produced 4th-century pottery. Further west another silty deposit 939) overlay rubble deposit 919.

Two deposits, 3115 and 3121 identified in the small service trench on the eastern side of Area 3 may represent disturbance from robbing out of wall 327 and Structure E (Phase 3). It is possible that both deposits represented the preliminary stages of dumping in the area, as they were sealed by dumped deposit 3110. However, 3121 was a pale yellow, firm, clay-sand deposit which appeared similar to the mortar floors identified in Structures D and E, but was stratigraphically too high to be contemporary with them. It might be taken as a further hint of the loss of evidence of later 4th- and 5th-century occupation in the area as a consequence of medieval and later activity.

West of Structure G 934, a 0.9 m thick deposit was probably part of the same dumping horizon as 303 in Area 3. PHASE 6 – MEDIEVAL/POST MEDIEVAL There is little evidence from the site for the period between the Romano-British and the medieval/post medieval periods, other than a single sherd of Late Saxon type ware from 730 (Chapter 5). Some medieval agriculture and manuring must have taken place to account for the disturbance of the latest Roman-period deposits and the intrusive pottery. Some possible plough scarring was recorded in the Trial Trenches and in the areas of site where natural sand was reached.

Deposit 3110, also recorded as 3113, was a possible continuation of layer 303, however this was not confirmed. The deposit comprised a firm, red-brown, sandy-silt with frequent gravel inclusions and pottery dating to the late 3rd to 4th centuries. Sealing 3110 was a firm, red-brown, clay-silt with frequent rubble inclusions 3126.

A range of medieval pottery was recovered, largely from the main rubbish dumping deposit in Area 2, 202, with the earliest dating from the 11th to 12th century including Brown Sandy ware and Gritty ware. Brandsby ware, and Reduced Sandy wares dating to the 13th and 14th centuries were found, along with Humberware, Cistercian ware and Late Medieval gritty and sandy wares. Given the limited quantities of medieval and later pottery published from Malton/Norton the site assemblage from Brooklyn House is of local if not regional interest and provides a firm basis for future work.

A similar deposit may have overlain Structure F to the south. Deposit 339/372 was up to 0.45 m deep and was exposed for 2 m in the southern part of the area before being cut away by the modern service trenches. It too produced pottery dating to the late 3rd to 4th centuries. Layer 703/715 was a 0.35 m thick deposit of similar material in Area 7, which may also represent the same rubbish dumping event. It was also recorded as 735 75

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire The 2015 trial trenches located a series of sand extraction pits, presumably excavated during the construction of the Victorian houses and villas lining Langton Road to the east of the site. More evidence of sand extraction was found in Area 5, where a large area of disturbed ground, 503, was identified. Due to constraints on the depth of excavation in Area 6 further, possibly similar, features were identified but could not be recorded.

PHASE 7 – VICTORIAN/MODERN The majority of disturbance to the Romano-British remains in the northern part of the site came from the construction of a Victorian/modern lean-to type structure in Area 3. In addition a number of refuse pits belonging to this phase were recorded – 335, 359, 3111, 763 and 829, along with four animal burials – 322, 3124, 812 and 905. In Area 9 a large pit, 912/932, which extended into Area 3 as 3026, was recorded cutting Roman-period rubble deposits including layers 934 and 939.

76

Chapter 3

Prehistoric pottery Blaise Vyner

A small quantity of prehistoric ceramic was included among finds.

247 GS13 Salt container Sherds (2+) from a vessel of uncertain shape, light terracotta exterior surface and fabric, the interior surface missing, a few small and occasional mediumsized angular limestone grits, occasional voids from which calcitic grits have leached, weight 13 gm.

Catalogue BHN16 279 GS4 Peterborough Ware Single sherd of a thick-walled vessel, buff-brown exterior surface, dark grey interior surface, the fabric contains occasional angular medium-sized stone grits but its main component is chunks of ‘grog’, giving it a soapy feel, wall thickness 15 mm, weight 21 g. Decorated with short comb impressions. Peterborough Ware is widely distributed across East Yorkshire, usually from non-funerary contexts and mostly from antiquarian excavations. A recently excavated assemblage from Sewerby Cottage Farm, Bridlington has produced an assemblage of Peterborough Ware with reliable radiocarbon dates indicating a currency of 3636-3101 cal BC at 2 sigma (Manby 2009, 182).

Following the recognition of fragments from salt containers in ceramic assemblages of latest Iron Age date in North Yorkshire (Willis 1999, 23-26), the material has increasingly been found in assemblages from the lower Tees valley area (Willis 2016b, 259-261 and figure 12.3). Evidence for salt manufacture in the form of ovens, kiln supports and shallow pans has been excavated at Loftus, on the coast of north-east Yorkshire (Sherlock and Vyner 2013), with fragments of salt container present in the ceramic assemblage from Kilton Thorpe Iron Age settlement, no more than 3 km distant. Salt could fairly readily have been distributed by boat along the coast: to the salt manufacturing evidence at Street House, Loftus, should be added the limited manufacturing evidence from Easington, Spurn Head, which is more likely to be latest Iron Age rather than the late Bronze Age as suggested (Richardson 2011, 63), although detail is sparse. The coastal manufacturing locations are unsurprising, but complexity is introduced with the manufacturing evidence from Newbridge Quarry, Pickering (Vyner 2010), which includes several pieces of fired clay as well as a rod (cf. Sherlock and Vyner 2013, figure 9.6), now seen to derive from salt-working. Its presence at Pickering is perhaps best explained by inland refining of a coarse primary product, which raises interesting questions concerning the control of manufacturing and distribution.

430 T4 Beaker Sherds (3) from a well-made vessel, surfaces and fabric dark terracotta with few obvious grits, although there are voids from which small calcitic grits have leached, which are numerous on the interior surface, but less common on the exterior, wall thickness 7 mm, weight 25 g. Decorated with concentric lines of impressed fine cord. This is an All Over Corded Beaker, a style which is distributed across the Wolds, although most examples are from antiquarian excavations (Manby 1988, 71 and figure 4.16). 247 GS13 Single sherd from a small thin-walled vessel, surfaces orange-brown, fabric core light grey, numerous small limestone grits in a smooth well-made fabric, wall thickness 5 mm, weight 1.5 g. Iron Age or native style Romano-British.

Acknowledgement Blaise Vyner would like to thank Jane Richardson of ASWYAS for provision of information on the Newbridge Quarry assemblage.

77

Chapter 4

The Roman pottery I.M. Rowlandson and H.G. Fiske with K.F. Hartley, J.M. Mills, J. Bird and D. Williams

Malton settlement. The other vessels of note were the grave goods and funerary urn that were used as part of the bustum rite, a type of burial that appears to have been used in the vicinity of 3rd-century military sites but that was uncommon in rural eastern Yorkshire. The picture from this pottery assemblage is one that suggests that the inhabitants were more closely aligned, in both pottery use and cultural beliefs, to the inhabitants of the Malton fort and other military sites in the north of England than those living in more basic rural settlements in eastern Yorkshire.

INTRODUCTION Twenty one thousand and sixty Roman sherds (403.637 kg, 304.56 RE - Rim equivalents) were recovered from the excavation and evaluation trenching on this site. The Roman pottery from Brooklyn House is the largest quantified assemblage from Norton. Although a small proportion of the pottery was produced before AD 150 the majority of the pottery from the site appears to date to the 3rd and 4th centuries AD. Good smaller assemblages were retrieved from Phases 1-3 but the majority of the pottery from the site was retrieved from Phases 4 and 5 and contained Crambeck Parchment wares and Huntcliff jars that dated the group to the later 4th century AD. Good groups from phases of construction were recorded but the majority of the pottery was retrieved from Phase 5 soil layers. Typically many of the largest groups of later 4th-century AD pottery excavated from Lincoln and York (Darling and Precious 2014: Monaghan 1997: 866867) have been retrieved from similar soil layers when there was probably a breakdown in the typical patterns of rubbish disposal in the settlement at the end of the Roman period and vacated plots became the focus for dumping domestic refuse. Similar patterns have been observed by the author at roadside settlements such as Navenby, Lincolnshire (Rowlandson et al. 2011; 2015). Therefore it is possible that the pottery from the final Phase 5 soil layers may represent material that was in use on the plot or dumped on the site from adjacent plots, perhaps as the area of settlement contracted (Dicus 2014; Bryant 2011; see Gerrard 2013 for extended bibliography). What can be said is that the inhabitants of the settlement appeared to have access to a broad range of pottery at the end of the Roman period probably until the end of the 4th century AD, broadly Monaghan’s Ceramic Period 4b at York (1997: 866-867) or perhaps the early 5th century AD. No Early Anglian/Anglo-Saxon style pottery was retrieved and little that might suggest that the site had an extended phase of activity into the 5th century AD.

THE SAMIAN POTTERY J.M. Mills Much of the samian from Brooklyn House was residual in levels deposited in the 4th century AD and later (Phases 3+). The vast majority came from the kilns of Lezoux in Central Gaul, Rheinzabern and Trier in East Gaul. A few vessels from La Graufesenque (South Gaul) and Lavoye in Argonne (East Gaul) were also identified. The bulk of the samian dates from the mid-2nd to the mid-3rd centuries AD. The East Gaulish vessels dating from the mid-3rd century were some of the latest to reach Britain; samian imports having ceased around AD 260. The vessel function profile for the assemblage is unusual, having a high proportion of plain bowls and mortaria and a correspondingly small quantity of cups. The samian accounts for less than 2% of the ceramic assemblage, a figure likely to be influenced by the late date of the site and proximity to a local pottery industry. Methodology and Quantification Each sherd was weighed and a fresh fracture, where necessary, was examined with a X10 hand lens in order to identify the fabric and hence the kiln site or production centre. The data was recorded on an Excel spread-sheet using standard codes. Moulded decoration and potters’ stamps were recorded using graphite rubbings (Figures 104, 105). The database records context number, fabric code, presence of cross-context joins, vessel form, sherd type, sherd count and weight, condition, rim diameter and EVE (Estimated Vessel Equivalent), presence of decoration and potters’ stamps, and date ranges (early– late date). The presence of use-wear, repair and graffiti was also systematically recorded where present.

The assemblage was significant as a broad range of face pots, Smith god pots and other specialist vessels produced at Norton and Crambeck were recovered that give an insight into the religious observances of and cultural beliefs of the inhabitants of the Norton/ 78

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Vessel Function CLOSED FORM DECORATED BOWL PLAIN BOWL

DISH OR BOWL CUP

DISH

MORTARIA

Rim EVE

MVN

No

% of total

No

(form present)

0

2

1.6

0.11

3.6

21

17

0.73

23.8

39

0.69

22.5

0.18

5.9

0.05 1.31

1.6

42.7

% of total

31

3

2.5

40

32.3

8

11

6.5 8.9

Table 1: Quantification of samian by vessel class in Rim EVE and Maximum Vessel Number (MVN) Fabric

Count

SAMSG SAMCG SAMLA SAMRZ SAMTR ?SAMTR SAMEG SAM

7 153 2 88 34 3 2 2 291

Phase

Count

Wt. (g)

1 2 3 4 5 7 Unphased

20 6 39 18 149 11 48 291

272 30 734 168 1580 68 587 3439

Mean wt. (g) 13.6 5 18.8 9.3 10.6 6.2 12.2 11.8 g

Rim EVE 0.25 0 0.55 0.21 1.84 0 0.22 3.07

Table 3: Quantity of samian by stratigraphic phase

% of Total Weight (g) % of Total Rim EVE % of Total Mean sherd MVN No Weight EVE weight (g) 2.4 58 1.7 0.08 2.6 8 6 52.6 1067 31 1.65 53.7 7 53 0.7 26 0.8 0 0 13 2 30.2 1197 34.8 1.03 33.6 14 47 11.7 1068 31 0.18 5.9 31 15 1 10 0.3 0 0 3 0 0.7 7 0.2 0 0 4 1 0.7 6 0.2 0.13 4.2 3 2 3439 3.07 11.8 126

The samian assemblage (Table 1) from the excavation (291 sherds, 3439 g) includes vessels from kiln sites in the three main production centres: La Graufesenque in South Gaul (SAMSG), Lezoux in Central Gaul (SAMCG), Argonne (SAMLA), Rheinzabern (SAMRZ), and Trier (SAMTR) in East Gaul. A few sherds were too small, or so heavily burnt that the fabric could not be identified confidently, codes include SAM, ?SAMTR, and SAMEG. The assemblage is summarized by fabric in Table 2 and by stratigraphic period in Table 3. A further 20 sherds (192 g) were recovered from the assessment phase (Monteil 2016) 13 from Lezoux, seven from East Gaul. These are not included in the totals here, but are referred to within the text. Distinct from the quantification by Estimated Vessel Equivalent (EVE) derived from rim measurement, MVN estimates the Maximum Vessel Numbers identified from the sherd count after taking sherd joins, decoration, and finish into consideration. This is only possible because of the standardization in samian production and provides quantification that is useful to enable comparison with assemblages which were recorded without the use of rim EVE. Table 4 gives quantification by vessel form and fabric group in both MVN and rim EVE.

Table 2: Quantification of samian by Fabric

and Drag 31R/Lud Sb are used for East Gaulish vessels where they are more variable than the Central Gaulish forms. Note: Totals do not include the two Wa 79/80 vessels which do not fit into a functional category The data are summarized in Table 1 to illustrate how the two quantification methods can vary, MVN exaggerates the quantity of decorated ware present compared with the EVE value because a decorated form can be identified from a very small body sherd. In contrast the MVN for cups appears to be much lower than the EVE value suggests; small vessels are strong and resist breakage, consequently four cup sherds (40 g) from vessels with small rim diameters give an inflated rim EVE value. The larger the assemblage the less obvious these inconsistencies tend to be. Condition The mean sherd weight for the site is pretty standard at just under 12 g. The mean weight of the East Gaulish sherds, however, especially that from Trier, is above average as might be expected of the latest material in the collection subjected to less attrition and redeposition than earlier material.

The standard vessel form type series for samian are used throughout: Dragendorff (Drag), Walters (Wa), Curle (Cu) and Ludovici (Lud). The term Drag 31/Lud Sa 79

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Function

Form

CLOSED FORM

Closed

SAMSG SAMCG SAMLA

8 (0.11)

2

9

1

1

31R/ Lud Sb

10 (0.12)

Bowl

1 ((0.06)

CUP

O&P LV 13

1 (0.12)

18-18/31

2 (0.08)

18 or 18/31

PLAIN BOWL

DISH

38 33

Cup

15/17R or 18R 18/31

18/31(R)

18/31R or 31R

?32

31/ Lud Sa 32 36

32 or 36

SAMEG

2

1

37

SAMTR

29

DECORATED BOWL

SAMRZ

5 (0.49)

2 (0.09)

1

4 (0.1)

4 (0.22)

4 (0.11)

7 (0.04)

4 (0.24)

4 (0.07)

6 (0.16)

19 (0.51)

1

1

1 (0.05)

1 (0.07)

1

1 (0.08)1

1 (0.05)1

Walters 79

Lud Tb/ Cu 23 var

1 (0.07)

Dish

2

1

45

8 (0.18)

1

FORM UNCERTAIN INDETERMINATE Count/Wt (g)

Cu 21 Wa 79/80

Lud Sa/Sb

18/31R or 31R

2

2 (0.05)

69/211 g

1

1

1

22/158 g2

10/53 g

MORTARIA

1

SAM & ?SAMTR

1/5 g

3/10 g

Table 4: Vessel Form and Function by Fabric Group (Production Area). Maximum Vessel Numbers (rim EVE in brackets) Vessel forms are Dragendorff (Drag) forms unless specified. (O and P = Oswald and Price 1920) 1 Heavily burnt rendering fabric identification impossible 2 Includes one roughly shaped disc and one possible disc

The general condition of the samian is good, with some large fresh sherds; little of the material is battered or abraded with no sign of alteration to the appearance by aggressive soil conditions. There is a high proportion of small and shattered sherds, especially within the Central Gaulish material, this explains the high number sherds of indeterminate form (69, or 45% from that kiln site) (Table 4). Use-wear was noted on just a few sherds, only three vessels, all bowl forms, had interiors that were worn smooth with an absence of slip from Lezoux a Drag 37 (202 (GS02)) and a Drag 31R (202 (GS05)), and a Trier Cu 21 (207 (GS06 and GS09)) all had been heavily used. A couple of foot-rings were also noticeably worn,

both Trier vessels: a Drag 31/Lud Sa (207 (GS12)) and a Drag 31R/Lud Sb (200). No drilled holes or cut slots indicative of leaded repairs were observed, nor were any alterations such as trimmed bases recorded. A single Drag 31 from assessment Trench 5 was worn on the underside of the base and is the only example of this potential type of vessel re-use (Monteil 2016). This low incidence of use-wear, repair and re-use perhaps suggests the inhabitants had access to a good supply of samian and there was little necessity to extend the useful life of the vessels. The only exceptions are two possible discs: both sherds from Rheinzabern vessels, one more obviously worked (202 (GS08)) weighs 13 80

I.M. Rowlandson and H.G. Fiske: The Roman pottery

g and is quite carefully chipped around most of the circumference, the other is smaller, weighing only 7 g, and may have just broken this way, but seems to be a rough-out for a smaller disc (202 (GS07).

date. No potters’ stamps were found, and the decorated sherds are generally small and with the exception of a sherd from the assessment attributed to Cinnamus and a larger sherd from the Phase 5 dumping in the style of Banuus (Figure 104. DS3) are of little help in refining the site dating. The overall evidence suggests that the bulk of the Central Gaulish samian dates from AD 160 or perhaps even AD 170.

South Gaul (SAMSG) South Gaulish products (six sherds, 58 g) make up a very small percentage of the samian from the site; none was identified from the assessment trenches. Only dish forms Drag 15/17R or 18R and Drag 18-18/31, and a flake from a Drag 29 decorated bowl were identified. A foot-ring sherd was burnt as was a further sherd, classified as SAM, which was too burnt to confidently identify, but probably of South Gaulish origin. The fabrics indicate all the vessels came from the main kiln site at La Graufesenque, probably Flavian in date.

East Gaul Overall the quantity of East Gaulish samian is only slightly less than that from Central Gaul by sherd count (132 sherds, 45%), but 67% of the assemblage by weight (2314 g). In Britain East Gaulish samian usually forms around 10% of a samian assemblage where the site was in continuous occupation from the 1st century AD. This high proportion of samian from East Gaul emphasises the late date of this group and the late 2nd-century date for the start of activity on this site.

Central Gaul No sherds from the Trajanic/Hadrianic kilns of Les Martres-de-Veyre were identified. All of the Central Gaulish samian came from Lezoux, and forms the largest part of the assemblage, about 53% by sherd count and rim EVE, although the small sherd size means that Lezoux samian formed only 31% by weight. The mean sherd weight (7 g) is low, suggesting that this material had perhaps been subject to repeated re-deposition; however, none was noticeably battered or abraded and just eight had been subject to post-depositional burning.

Although minor kilns exported small quantities to Britain from the early 2nd century the main period of export from the major centres of Rheinzabern and Trier began around AD 160. The usual ratio between the two is 70% Rheinzabern : 30% Trier (Bird 1993: 2) and that is reflected here. The earliest East Gaulish wares identified are two Drag 37 sherds from Argonne, probably Lavoye (Figure 104. DS5 and DS6) dating from c. AD 150-180. Another Argonne sherd was recognised from the recent Wood Street excavation (Monteil 2018) but few sherds from this source have been identified from Malton or Norton to date.

The main period of exportation from Lezoux to Britain was from about AD 120/125. Little material was identified that dates from before AD 160 or even 170; although the relatively high proportion of indeterminate sherds (Table 4) may mask the presence of some earlier forms the quantity is not thought to be significant. Early plain forms are dishes Drag 18/31 and 18/31R, a single Drag 18/31 was also recovered from the assessment trench 3 subsoil (Monteil 2016); these are dated AD 120-160. A single decorated sherd (Figure 104. DS2, 3007) has been tentatively identified as being from an under-fired, early 2nd-century Lezoux vessel and maybe as early as AD 100-120. There is no contemporary samian on this site suggesting that it, like the vessels from South Gaul may derive from earlier occupation in the area.

The range of vessel forms (Table 4) is limited to fairly standard dish and bowl forms. A single gritted body section from Trier mortarium, and two Cu 21 ungritted mortaria, one from each centre were recorded; the Trier example probably 3rd-century in date. Bowl Drag 31R/Lud Sb was the most common type, 19 from Rheinzabern and seven from Trier; dish Drag 31/Lud Sa was evenly represented with four examples from each centre. At least five vessels of the Trier Drag 31 series and two or three from Rheinzabern are of 3rdcentury date as are the Drag 36 dishes. Other dishes came from Rheinzabern, seven Drag 32 of which two were stamped, and a barbotine rim from a Curle 23/ Lud Tb var dish (Figure 105. PS3). None of the 3rdcentury jar and flagon forms such as the ones found at Piercebridge (Ward 2008: 178) were identified, and although uncommon such forms might be expected here given the date of this group. Other forms include nine Drag 37 bowls from Rheinzabern, but none from Trier. These were not closely-dated spanning the late 2nd to 3rd centuries with the exception of one vessel in the style of Helenius (Figure 104. DS8) that appears to be an example of re-use of a mould in the second quarter of the 3rd century.

Many of the Lezoux forms were introduced after AD 160, these include Drag 31R, Drag 45, Wa 79 and 80; others earlier in the 2nd century like Drag 31, Drag 38 and closed forms of which two very small body sherds were found were introduced c. AD 150. Cup form Drag 33 was produced throughout the life of the samian industry, but the small cups, 100-120 mm in diameter, found here are probably Antonine, if not mid-late Antonine in date. The barbotine decorated dish Drag 36 was also produced over a long period, from c. AD 70 to AD 260; two of the Lezoux examples are large and mid-late Antonine in 81

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 104: The decorated samian, scale 1:1

82

I.M. Rowlandson and H.G. Fiske: The Roman pottery

No cups in East Gaulish fabrics were identified; no Drag 40 cups to complement the Drag 32 dishes, nor any of the more common straight-sided Drag 33/46 types. This is discussed more fully below (pp. 138-140).

Bird pers. comm., October 2019). AD 100-120. Area 3, mortar floor 3007, Structure D, Phase 3 DS3 Drag. 37, SAMCG (Lezoux). Body sherd from upper part of decoration with ovolo Roger B157 and beads below. The design is divided into panels with a bead row surmounted by large rosette Rogers C165. To one side is a medallion, to the other a figure or motif which is too blurred to identify. The impression of the medallion appears triple-bordered, but this seems to be a result of careless mould-making. The ovolo divider with large rosette and the use of medallions suggest the work of Banuus. The rosette, beads and medallion are on a bowl stamped in the mould by Bannus in the Birley collection but with a different ovolo (Stanfield and Simpson 1990: pl. 140, 10). AD 180210. Area 2, subsoil/Roman dump deposit 202 (GS08), Phase 5

The only stamped vessels recovered were from Rheinzabern and Trier (see Catalogue of potters’ stamps); two late 2nd- early 3rd-century examples from Rheinzabern (Figure 105. SS2 and SS3) from the Phase 5 soil dumps and a large portion of a 3rd-century Trier bowl (Figure 105. SS1 and Figure 106. PS5) from robbed wall 205 (GS03) assigned to Phase 3. Although largely residual, the East Gaulish samian is an interesting group with several 3rd-century vessels, a selection of which are illustrated (Figure 106. PS1-8). The Decorated Samian – J.M. Mills and Joanna Bird (Figure 104)

DS4 Drag. 37, SAMCG (Lezoux). Small body sherd with only part of a lion, or possibly lion attacking boar O.1491, this figure type was used by many potters including some of those working in the later 2nd century. Mid-late Antonine. Area 2, topsoil 201 (GS01), Phase 7

The following catalogue lists, dates, and where possible identifies the decorated pieces to individual potters or groups of potters. The catalogue is organised by production area then by phase and context within each. The entries give vessel form, fabric/kiln site, description, date range and excavation references. The catalogue entries for the Argonne vessels (DS5 and DS6) have been written by Joanna Bird.

East Gaul DS5 Drag. 37, SAMLA (Argonne). Burnt body sherd, probably from the Argonne. Similar bands of spirals between plain lines are on bowls attributed to Germanus (Oswald 1945: fig. 7, 12) and Tribunus (fig. 8, 29). The main leaf is generally similar to one used by Cesatus (e.g. Muller 1968: taf 17, 441), but both larger and more complex. The fragmentary motif to its right is probably a narrow pinnate leaf (cf. taf 17, 441, but that leaf has a short straight stem); the motif to its left may be a repeat of the pinnate leaf or perhaps a narrow bunch of grapes (cf. taf 17, 443, another Cesatus bowl), but the stem, though the same shape, is longer). AD 150-180. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5

Abbreviations: Drag Rogers RF O.

Form as defined by Dragendorff Motif in Rogers 1974 Motifs in Ricken and Fischer 1963 Figure type in Oswald 1936-37

The Inventory Numbers (Inv. No) quoted are taken from European intake of Roman Samian ceramics. http:// www.rgzm.de/samian/home/frames.htm South Gaul DS1 Drag. 29, SAMSG (La Graufesenque). Small body sherd with a band of reflexed trifid motifs below the central cordon. c. AD 60-85. Area 3, mortar floor 3007, Structure D, Phase 3

DS6 Drag. 37, SAMLA (Argonne). Probably Argonne, on fabric and slip. Only part of the ovolo band with bead row below remains. Bead rows below the ovolo are not common on Argonne ware, suggesting that this may be an early bowl from the workshops. A similar ovolo with beads occurs on an Argonne bowl from Shiptonthorpe (King and Millet 2006: fig. 7.2, 59), and it is also known from Corbridge (Hofmann 1990: fig. 1, nos 4 and 6), attributed there to Germanus of Lavoye. AD 150-180. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5

Central Gaul DS2 Drag. 37, SAMCG (Lezoux). Small body sherd with a bright orange slip and a paler, micaceous, fabric with occasional large calcareous inclusions. Slightly underfired? The only decoration is a large, horizontally placed S or scroll motif (Rogers S68) within a panel with zig-zag borders. The S motif is recorded by Rogers (1974; 1999) for Igocatus and the anonymous potter Me… . The fabric is not right for Les Martres and Igocatus although an Igocatus mould could have been traded and reused at Lezoux. The fabric is, however, similar to one identified as Me… from Hooper Street, London (original context reference: 1469, SF 656, Joanna

DS7 Drag. 37, EG (Rheinzabern). Body sherd with only small boar RF T68 surviving. This figure was used by Comitialis on style IV and VI bowls (Ricken and 83

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire Fischer 1963: 130). AD 170-240. Area 4, cobbled surface 4031, Phase 3

of Names on Terra Sigillata (NoTS). The potters’ name and the die have however been added to the digital record at https://www1.rgzm.de/samian/home/ frames.htm

DS8 Drag. 37, EG (Rheinzabern). Body sherd from lower edge of the decoration, with trophy motif (RF O160b) upside down within a double bordered medallion (RF K19a), simple pinnate leaves (RF P79) and an eagle (RF T207) at the base of the decoration. The lower part of the eagle seems to have been trimmed off in the finishing of the bowl. The inverted motif in the medallion, leaves and eagle are all on a bowl attributed to Helenius (Ricken and Thomas 2005: taf. 174, 19), a potter dated to the period AD 180-200. This sherd, however, is poorly finished, is nearly 10mm thick, and of a paler, coarser fabric than would usually be expected for an Antonine potter. It is possible that this bowl was made later, the mould re-used. Reuse of moulds is an activity that is now accepted to have happened within the Trier industry and is increasingly suspected to have happened at Rheinzabern too (Bird 2002: 33-34). A date in the first half of the 3rd century, perhaps AD 220-250, is therefore proposed for this piece. Area 2, rubble spread 233 (GS13), Phase 4

Catalogue of Potters’ Stamps J.M. Mills The following catalogue lists the potters identified in alphabetical order. Each entry gives: potters name (i, ii etc., where homonyms are involved), die number, production centre (fabric codes), form (form codes). Reading. Comment. Date range. Excavation references. SS1 Atilido,1a, Trier, Drag 31R/Lud Sb. TIIDOF(I) The initial blind A has a long serif at the bottom right side and in some impressions a central dot shows, but not in this example. The F is incomplete, or has very short cross strokes, and there is no sign of the last letter. Very coarse and pale fabric as per description in NoTS (Hartley and Dickinson 2008: 294). The rouletted band is narrow. Late 2nd – first half of the 3rd century, but the pallor of the fabric suggests second quarter of the 3rd century. Area 2, Structure A, robbed wall 205 (GS03) Phase 3; Area 2, subsoil/Roman dump deposit 202 (GS13), Phase 5. (Figure 105. SS1 and Figure 106. PS5)

DS9 Drag. 37, EG (Rheinzabern). Small body sherd with two impressions of ovolo RF E42. This ovolo was used by Julius I, (220-255) Lupus (190-225) and Perpetus (230-260?). AD 190-260. Area 2, subsoil/ Roman dump deposit 202 (GS08), Phase 5

SS2 Dignus, iii, Φ1, Rheinzabern, Drag 31R/Lud Sb. ) INVF This is the first example of a potter of this name working at Rheinzabern. This stamp almost certainly reads Dignus, the end of the stamp was not cleanly impressed and the beginning of the D is missing. The positioning on the base suggests that no letters or symbols precede the name, but until a complete example is recorded this is not a certainty. The finish and fabric suggest that this vessel was made c. AD 170220/230. Area 2, subsoil/Roman dump deposit 202 (GS02), Phase 5. (Figure 105. SS2)

DS10 Drag. 37, EG (Rheinzabern). Body sherd/scrap with a fragment of ovolo RF E25 or E26. Late 2nd – early 3rd century. Area 2, 207 (GS09), Phase 7 DS11 Drag. 37, EG (Rheinzabern). Body sherd from lower edge of the decoration, two double-bordered medallions RF K20 with a vertical bead row between and an 8-petalled rosette with a central hollow RF O48 are the only surviving elements of the decoration. Late 2nd – early 3rd century. Area 2, rubbish deposit 216 (GS06), Phase 5

SS3 Lupus iv, 2a, Rheinzabern, dish, probably Drag 32. V(PVFE) This die was used mainly on form Drag 37, but there are occasional records on dishes and bowls (Hartley and Dickinson 2009: 144-145). AD 190-225. Area 2, subsoil/Roman dump deposit 202 (GS15), Phase 5. (Figure 105. SS3)

The Samian Potters’ Stamps J.M. Mills Very few potters’ stamps were recorded amongst this collection of samian. Of the five stamped dishes recorded two were from Trier and three from Rheinzabern. It is unusual for no stamped vessels from Central Gaul to be recovered where reasonable numbers of sherds are recovered; many have been published from the sites north of the river, the reduced nature of much of the Lezoux samian from this excavation undoubtedly affected the survival of the stamps. Just three of the five could be identified; one was of just a single letter, and the fifth, a Trier Lud Sb bowl (207 (GS12)), retained only a scrap of the frame edge and is not listed below. Stamp SS2 is from a die not previously recorded and consequently it will not be found in the relevant volume

SS4 Not identifiable, Rheinzabern, Drag 32. )( Late 2nd – first half 3rd century. Area 2, Roman ground surface 231 (GS16), Phase 1. (Figure 106. PS8) Catalogue of Illustrated Plain Samian J.M. Mills PS1 Drag 36, Rheinzabern. Coarse fabric and thin orange slip. c. AD 220-250/260. Area 4, rubble 410, Phase 4 PS2 Drag 36 Trier. Large vessel with heavy beaded rim. c. AD 220-250. Area 2, rubble 230 (GS13), Phase 4 84

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 105: The stamped samian, scale 1:1

Figure 106: Illustrated plain samian, scale 1:2

PS3 Barbotine rim sherd from a dish, Rheinzabern. The thick rim has more in common with the rims of straight-sided cups which were occasionally decorated ‘en barbotine’ (Oswald and Price 1920: pl. LV, 24) than the more common Drag 36 dish. This vessel must then be a variant of the Curle 23 (or Lud Tb) dish which was the companion form to the cup (Oswald and Price 1920. pl. LIX). Not a common variant, Stanfield illustrated two similar dishes from London (1929: fig. 10, 48 and 50), no. 50 is the most similar to the Norton vessel. Late 2nd – early 3rd century. Area 2, subsoil/Roman dump deposit 202 (GS03), Phase 5

Phase 3; Area 2, subsoil/Roman dump deposit 202 (GS13), Phase 3 PS6 Drag 31R/Lud Sb, Rheinzabern. With a internal groove not a step. c. AD 170-230. Area 2, Roman ground level 231 (GS16), Phase 1 PS7 Drag 31R/Lud Sb, Trier. Grooved interior and flattened bead rim. c. AD 180-240. Area 2, subsoil/ Roman dump deposit 206 (GS02), Phase 5 + a joining unstratified sherd. PS8 Drag 32, Rheinzabern. Incomplete stamp (SS4) c. AD 180-230. Area 2, Roman ground level 231 (GS16), Phase 1 and rubble 251 (GS15), Phase 3

PS4 Dish, probably Drag 36, Lezoux. A large unstamped dish with a beaded foot of the type usually associated with Drag 37 and Drag 38 bowls. Occasional examples on Drag 36 are known (Oswald and Price 1920: pl. 13, 13). c. AD 160-200. Area 2, subsoil/Roman dump deposit 202 (GS03), Phase 5

THE OTHER WARES I.M. Rowlandson and H.G. Fiske Methodology The pottery has been archived using count and weight as measures according to the guidelines laid down for the minimum archive by The Study Group for Roman Pottery (Darling 2004) using the codes and database format developed by the City of Lincoln Archaeological Unit (CLAU) (see Darling and Precious 2014). The fabric

PS5 Drag 31R/Lud Sb, Trier. Stamped by Atilido of Trier (SS1). Late 2nd – first half of the 3rd century, Area 2, Structure A, robbed wall, 205 (GS03), 85

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire series developed by Bidwell and Croom (1997; 2012) and the National Roman Fabric Reference Collection (Tomber and Dore 1998) have, where possible been concorded with this report. Rim equivalents (RE) have been recorded and an attempt at a ‘maximum’ vessel estimate has been made following Pollard (1990). Vessels selected as suitable for illustration (D*, see catalogue) and fabric samples (FTS) have been bagged separately for ease of future reference. A preliminary pottery ‘spotdate’ was produced presented prior to the development of the phasing scheme for the site. The pottery has been discussed by ware type and according to the site structure presented in this report. The illustrated vessels have been referred to in the text as No*. Catalogue entries follow the format: illustration number, fabric, discussion, area, feature/layer, fill, Phase and Drawing reference number. The terms ‘conical flanged bowl’ and ‘straight-sided bead and flanged bowl’ have been used interchangeably to refer to vessels within the range of Gillam’s Types 228-231 (1970: fig. 24). The archive record contains the following tabulated data that could not be included within this report. • • • •

Quantified dating summaries by context Full quantified archive Full fabric summary Full form summary

This data forms part of the site archive and will also be curated in an Access database, available from the author in a digital format. Amphora I.M. Rowlandson with D. Williams One hundred and eighteen sherds of amphora (11.464 kg, 0.49 RE) were recorded from this assemblage. In contrast to rural assemblages from this area a good range of amphorae were recorded. The proximity of the site to the fort at Malton and the integration with military supply lines account for the quantity of amphora sherds from this site. A far broader range of amphora types were recorded from the Malton vicus excavation (Williams 1997) although no sherds of North African amphora were identified amongst that assemblage. It was also noteworthy that a considerable proportion of the amphora from Brooklyn House might be dated to the 3rd century AD. The date range represented probably accounts for the absence of many of the Early Roman types. The amphora that was recovered were mostly found in large fragments and retrieved from Phase 5 deposits with the remaining material found within compacted surfaces or rubble surfaces from Phase 1-4 groups. Rural assemblages from this region have small quantities of amphorae but sherds often show signs of having been reused as tools (Darling and Precious forthcoming; Rowlandson and Fiske 2019a; Rowlandson et al. 2017). Only one example of possible reuse was

Figure 107: Stamped Dressel 20 amphora: 107a: Rubbing of stamp (scale 1:1), 107b: photo of stamp, 107c: and microscope image of vessel fabric

noted amongst the assemblage from Brooklyn House: an area of abrasion on the stamped handle fragment (Figure 107) possibly suggests it was reused as a rubber. The remaining material from this assemblage was in fairly fresh condition. DR20 Dressel 20 amphora Unsurprisingly the majority of the amphora were globular Dressel 20 olive oil type amphora from Southern Spain. It was noticeable that the majority of the sherds in this group were either the finer ‘saltsurfaced’ or the pale grey reduced fabric variants more commonly seen amongst later assemblages from the later 2nd to earlier 3rd century AD. Sherds in the grittier fabric variant, more commonly seen amongst 86

I.M. Rowlandson and H.G. Fiske: The Roman pottery

assemblages dating to the Early Roman period, were also present in small quantities. There were few feature sherds present but a rim fragment retrieved from Phase 4 would typically date to the 3rd century AD (MartinKilcher 1987: beilage 2, Class G AD 210-280) and a stamped handle is discussed by David Williams below (Figure 107).

variation across the range of amphorae produced (Laubenheimer 1985; Laubenheimer and Schmit 2009). NAAM North African amphora A single fragment from a North African amphora (Tomber and Dore 1998: NAF AM 2) was recorded from Phase 5. These vessels were often used to transport olive oil products and were most commonly found in Britain in the 3rd and 4th centuries AD (Tyers 1996: 104).

DR20: David Williams writes: Part of a much damaged Dressel 20 handle which contains a worn partial stamp in ansa. A reading of the existing end letters of the bottom line that can be made out show … E(?) LISSE. It is clear that there is also a top line above this, where the last three letters appear to be SSI. Given this information, it seems quite likely that a full reading of the complete stamp may be II IVNI MELISSI/ET MELISSE. If this reading is correct, it refers to the estate of the two Melissi et Meliss(a)e, which Callender suggests may have been a husband and wife or brother and sister partnership (1965: no. 879b). The stamps belonging to this family have a very wide distribution, including many found in Roman Britain, and date to the 3rd century AD (Callender 1965; Carreras and Funari: 1998: no. 138, 1-5; Berni Millet: 2008: no. 849). Kilns associated with the Norton stamp have been found on the banks of River Genil, a tributary of the River Guadalquivir, at Las Delicias, north of Astigi (Berni Millet 2008: 430). Area 2, subsoil/Roman dump deposit 202 (GS04), Phase 5, D140

AMPH Lusitanian amphora? A further unusual amphora fragment was presented to David Williams who writes: Top section of an ovoidshaped handle, with multiple grooves on the upper part, in a hard coarse fabric, reddish-brown on the surfaces and with a reduced grey core (Phase 5, 202 (GS14)). It is difficult to be sure but it is possible that this handle may belong to a Lusitanian amphora, perhaps one of the Almagro 51 variants that were current during the 3rd and 4th centuries AD and carried fish-based produce (Pinto et al. 2016: passim). Mortaria I.M. Rowlandson with K.F. Hartley (Figure 108) Two hundred and ninety-three sherds of mortaria (14.546 kg, 13.27 RE) were recorded. Mortaria made up 1.39% of the assemblage by sherd count and 4.36% by RE. The proportion of mortaria sherds from this assemblage was broadly similar to the Wood Street Norton assemblage (Mills, P. 2018: 1.4% by sherd count and 3.8% by RE) and greater than that recorded at the Malton Jack Berry House assemblage (Evans and Mills 2017). In contrast the majority of the Brooklyn House assemblage was of Late Roman date with a bias towards the products of the local Crambeck industry; it was notable that the proportion of York or Malton mortaria dating to the 2nd century AD was much lower. With the exception of a small number of Mancetter-Hartshill mortaria that may have dated to the later 2nd century AD the assemblage could be dated to the 3rd or 4th century AD. A range of slag-gritted vessels in oxidised fabrics, many probably locally produced and probably dating to the 3rd century AD, were also recorded.

GAU Gaulish wine amphora The next most numerous amphora type were Gaulish wine amphora, mostly in the typical GAL AM1 fabric variety (Tomber and Dore 1998) with two fragments similar to the NOM AM fabric, perhaps from a Gauloise 12 type vessel, and two sherds from a vessel in the GAL AM 2 fabric variant. Feature sherds were rare with one rim from a Gauloise 4 vessel in the GAL AM1 fabric and another in GAL AM 2. David Williams was presented with sherds from two vessels for comment: A thin-walled ribbed bodysherd from a Gauloise 4 flatbottomed wine amphora (Phase 5, 202 (GS6)). This form was exported to Roman Britain from the 2nd half of the 1st century AD till at least the end of the 3rd century (Laubenheimer in Williams and Keay: 2006).

MOMY Malton/York area mortaria (Tomber and Dore 1998: EBO WS; Mills, P. 2018: M01)

Part of a broad strap handle and thin-walled bodysherd in a fairly micaceous fabric (Phase 5, 202 (GS08)). This is almost certainly from a Gaulish flat-bottomed wine amphora, probably Gauloise 4, as this is the most common Gaulish amphora form imported to Roman Britain, though in a different fabric from the other vessel (above). This particular fabric is not the most common of Gaulish 4 fabrics but these vessels were made at a large number of kiln sites, mostly but by no means exclusively in the Narbonnaise area of France, and there are accordingly some nuances of fabric

Fourteen sherds of York or Malton area mortaria in oxidised fabrics with ‘mixed trituration grits’ were recorded (MOMY) all of the rims present appeared to be hooked types and many of the vessels had a white slip (Nos 2 and 3) including a stamped vessel (see below). The majority of these sherds were retrieved from Phases 4 and 5. A single vessel (No. 3) was retrieved from a Phase 2 ditch dated to the 3rd century AD. It is 87

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 108: The illustrated mortaria

likely that all of these sherds were produced some time in the later 1st to 2nd century AD. Examples of these types of mortaria were more common at Norton Wood Street and the Malton vicus (Mills, P. 2018; Bidwell and Croom 1997).

320: fig. 129, no. 95) and two from Hayton (Hartley 2015, illus. 9.17, nos 3 and 4, (no. 3 is mounted upside down)). Mortaria stamped with the same die have now been recorded from: Brough-on-Humber (2); Castleford; Hanging Grimston, North Yorkshire; Warren Hill Spring in the parish of Hotham, East Yorkshire (Didsbury forthcoming, Plot 123 NGR: SE 8876 3450, Context 22001 in the ‘Ganstead to Asselby Gas pipeline, East Yorkshire, 2005-2006); Hayton (2, see above); Lease Rigg (Frere and Fitts 2009: 248-249); Norton on Derwent (3); Malton (1949.52, unpublished); Slack (2); York (see above); Scarborough Museum, provenance unknown; Yorkshire Museum, provenance unknown; Cambridge Museum (the Bateman Collection, provenance unknown, but probably from York). The last three may all be assumed to be from York and elsewhere in Yorkshire).

2 MOMY: A white-slipped hooked rim mortarium featuring a stamp of the potter GENIALIS (see Hartley below). Area 2, subsoil/Roman dump deposit 202 (GS05), Phase 5, D126 3 MOMY: A white-slipped hook-rimmed mortarium with internal wear. Possibly a York product (Dickinson and Hartley 1971: fig. 19.7-8; Bidwell and Croom 1997: no. 145). Area 3, ditch 397, fill 3042, D120 The stamped mortarium – K F Hartley Figure 109, illustrated vessel No. 2. Wt 70 g diameter c. 270 mms 10% A flange fragment from a mortarium in hard, red-brown fabric with blackish-grey core and cream slip. The inclusions are fairly frequent, moderatesized quartz with rare black and red-brown material. No trituration survives. Area 2, subsoil/Roman dump deposit 202, Phase 5

The distribution of these eighteen mortaria is rather strikingly limited to Yorkshire and Humberside. They are in generally similar fabric and it seems very likely that they were made within this area. A possible source on present evidence is the Norton/ Malton area itself, but it must be remembered that this is speculation unless or until a kiln or other confirmatory evidence is found.

This stamp, )NIA(, is from a die which reads GENIALI(ligatured)S when complete (only one example of unknown provenance in the Yorkshire Museum, preserves the LIS clearly). It can be confidently identified as a stamp of Genialis from the same die as a stamp from York Minster (Phillips and Heywood 1995:

The range of rim-profiles used is quite wide, but all fit within the period AD 90-140 and the optimum date for this example is c. AD 90-110/120 Six other dies giving GENIALIS are known; some or all of these could belong to the same ‘firm’, with productions 88

I.M. Rowlandson and H.G. Fiske: The Roman pottery

MOG Reduced mortarium (local, slag-grits) A range of mortaria with slag trituration grits and iron-rich fabrics fired to orange, orange with whiteslip, or grey reduced colours were recorded from the Brooklyn House assemblage. Vessels of this type are known to have been produced in the vicinity of Lincoln in the Swanpool tradition, at Cantley near Doncaster and in the Catterick area (Tomber and Dore 1998: Fabrics SWN WS, CAN WS and CTR WSA). Three body sherds appeared to be a good match with Swanpool vessels from Lincolnshire (MOSPT, Phase 5), and three sherds from reeded rim and bead and flanged mortaria similar to Cantley products were recorded from Phase 5 (MOCA). In addition to these were a group of reduced mortaria, two with reeded rims (No. 6 and No. 8) and a further example with a hammer-head rim. Although grey ware mortaria like these are known to have been manufactured in East Anglia (Darling 1993: 198) the vessels from this site would appear to have a similar fabric to the grey ware vessels produced at the Norton kilns. Examples of sandy mortaria in both grey and oxidised orange fabrics have been recorded from the Norton ‘Grove Bungalow’ kiln (Hayes 1988: fig. 55.4345) and the Norton ‘Model Farm Estate’ site (Hayes and Whitley 1950: 15-18). Sherds from a further four oxidised mortaria from this site (recorded as MORT, Nos 4-5) also appear likely to be local products. Although a few sherds appear to be local products dating to the 3rd century AD it appears likely that only a limited number of such vessels were produced at the Norton kilns prior to the development of the Crambeck industry. 4 MORT: A reeded-rim mortarium in a pale orange oxidised fabric with slag trituration grits and internal wear. Area 9, rubble deposit 926, Phase 4, D69 5 MORT: An oxidised mortarium with slag trituration grits, possibly a local product. Area 2, subsoil/Roman dump deposit 202 (GS09), Phase 5, D20

Figure 109: Illustrated mortarium No. 2 of the potter GENIALIS. 109a: Rubbing (Scale 1:1), 109b: photograph of stamp, 109c: microscope fabric image

6 MOG: A reduced ‘grey ware’ mortarium, probably a local product. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D022 (Figure 110)

probably in Caistor St. Edmunds and at least one production in the Midlands. The die used for the above mortaria from Yorkshire and Humberside produced a closely similar stamp to one of the two probably used in Caistor St. Edmunds.

MOCR Crambeck white ware mortarium (Tomber and Dore 1998: CRA WH) The vast majority of the mortaria from the site could be attributed to a Crambeck production source. Two hundred and five sherds from a maximum of 187 vessels were in the coarser MOCR fabric. Table 5 provides a breakdown of this group by mortarium type. The flanged Crambeck Type 6 variants (plain MBF and reeded MRR; Corder 1937) appear to be the most common; Evans (1989: 79) has suggested that this type may have remained in production from the late 3rd

The other mortaria – I M Rowlandson MOSPT Swanpool type mortarium (Tomber and Dore 1998: SWN WS) MOCA Cantley type mortarium (Tomber and Dore 1998: CAN WS) MORT Local oxidised mortarium (slag-grits) 89

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 110: Grey ware mortarium with slag trituration grits (No. 6). 110a: general image and 110b: microscope fabric image Form

Description

M

Unknown rim type

MCR5

Sherds Sherd % Weight (g) Weight % Total RE RE % 99

48.29

3978

39.62

0.27

3.26

Corder type 5 hemispherical flanged type

3

1.46

140

1.39

0.23

2.78

Type 6 plain bead and MBF/ MFL Corder flange

49

23.9

2974

29.62

4.05 48.97 2.98 36.03

MRR

Corder type 6 reeded rim

41

20

2285

22.76

MCR7/ MHH

Corder type 7 hammer head types

10

4.88

560

5.58

0.72

8.71

MCR8

Crambeck Type 8 double flanged types

0

0

0

0

0

0

MWS

Corder 1928, Pl. V. 134-5 wall sided

3

1.46

104

1.04

0.02

0.24

TOTAL

205

100

10041

100

8.27

100

Table 5: MOCR Crambeck white ware mortaria by form type

century AD until the late 4th century AD. As there were limited numbers of the finer Parchment ware mortaria from this assemblage it appears possible that the coarser Crambeck Type 6 vessels continued to be used to fulfil the needs of the local inhabitants.

small sample of mortaria sherds from these phases the MOCR fabric was the most common type present. MOCR was also the most common mortarium type from the remaining phases with the vast majority of sherds unsurprisingly being retrieved from the large assemblage attributed to Phase 5.

A small proportion of this material was retrieved from Phase 1 and this made up the majority of the mortaria from Phase 1, these vessels were of Crambeck Type 6 and may represent material deposited towards the end of the 3rd century AD or were perhaps intrusive. Small numbers of sherds were recorded from Phases 2 and 3 including illustrated vessel number 8. From the

7 MOCR: A Crambeck mortarium with a bead and flange rim and internal wear, Unstratified, D108 8 MOCR: A Crambeck reeded rim mortarium with scored wavy line decoration. Area 3, construction cut 3091, fill 3089, Phase 3, D116 90

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 112: A sooted Crambeck mortarium, exterior and interior, illustrated vessel 10

MOCRF, gritted with finer slag trituration grits (Tomber and Dore 1998: CRA PA). This fabric type has been seen as a key dating indicator for the latest phases of pottery production in the north and production is considered to have begun in the late 4th century AD (Bidwell 2005; Bidwell and Croom 2010). Intrusive Parchment ware mortaria were recorded from Phase 1 and 2 deposits, a small quantity of sherds was retrieved from the late 4th-century Phase 4 deposits but the majority of the sherds in this fabric were recovered from Phase 5.

Figure 111: A burnt Crambeck mortarium, interior and exterior, illustrated vessel No. 9. 111a: internal surface, 111b: external surface

9 MOCR: A Crambeck Parchment ware mortarium, heavily burnt and possibly cut down for re-use as a lid. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D11 (Figure 111) 10

Hemispherical flanged Crambeck Type 7 mortaria (Nos 11-14) were the most numerous types along with a small number of bead and flanged Type 6, and wall-sided types. A notable absence amongst this assemblage was the double flanged Crambeck Type 8 mortarium form. From a sample of less than 40 sherds and 2.93 RE it is difficult to be certain if this is due to a chronological bias or merely by chance as there was a reasonable quantity of Parchment ware in other tableware forms (193 sherds, 0.91% of the assemblage or 5.77 RE 1.89% of the assemblage). In their study of

MOCR: A Crambeck Parchment ware mortarium, the sooting pattern suggests re-use as a lid or cover once broken. Area 4, topsoil 4023, Phase 7, D97 (Figure 112)

MOCRF Crambeck white ware mortarium (Tomber and Dore 1998: CRA PA) Thirty-nine sherds from 34 vessels could be attributed to the Crambeck Parchment ware mortarium fabric 91

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Form

Description

M

Unknown rim type

6

15.38

394

21.67

0

0

MCR5

Corder type 5 hemispherical flanged type

0

0

0

0

0

0

10

25.64

274

15.07

0

0

0

0

21

53.85

1049

57.7

0

0

0

0

0

0

2

5.13

101

5.56

0.2

6.83

39

100

1818

100

2.93

100

MBF/ MFL

Corder Type 6 plain bead and flange

MRR

Corder type 6 reeded rim

MCR7/ MHH

Corder type 7 hammer head types

MCR8

Sherds Sherd % Weight (g) Weight % Total RE RE %

Crambeck Type 8 Double flanged types

MWS

Corder 1928, Pl. V. 134-5 wall sided

TOTAL

0.66 22.53 0

0

2.07 70.65

Table 6: MOCRF Crambeck Parchment ware mortaria by form type

the pottery from Carlisle Millennium Project Swan, McBride and Hartley reflected upon the relative dating of the different form types of Crambeck mortaria (2009: 584-586; cf. Hartley 1995). Hartley has considered the Crambeck Type 5 mortarium as being one of the earliest forms that developed emerging c. AD 370/380 and Swan et al. (2009: 586) have suggested that the Crambeck Type 8 ‘double flanged type’ may have been a very late addition to the repertoire ‘perhaps early in the 5th century AD’. Examples of the Type 8 mortarium were noted from the coastal signal stations, Malton and Birdoswald forts and Langton villa by Birley during her initial survey (Birley in Corder 1937: 35) and examples can be added from a number of other sites including York. If the Type 8 mortarium could be considered to be a very late form, given the volume of pottery from the Phase 5 assemblage it may be that dumping of new pottery on the site had ceased early in 5th century AD. In support of this tentative assertion the low levels of burnished ‘Late Handmade fabrics’ from the Phase 5 assemblage (see below fabric YB18) and the absence of any recognisable Anglian or Anglo-Saxon vessels might suggest that the site had been abandoned by before the earlier 5th century AD, Monaghan’s ‘Ceramic Period 4c: after c. 410’ at York (1997: 866-867).

line decoration externally, internal wear. Area 4, rubble deposit 410, Phase 4, D78 (Figure 113) 14

MOMH2 Mancetter-Hartshill mortaria with fired clay trituration grits (Tomber and Dore 1998: MAH WH) Seven sherds from Mancetter-Hartshill mortaria were recorded, consisting of examples of a hammer-head rim type from Phase 3, a hook-rimmed vessel and a triple ribbed rim type (Darling and Precious 2014: no. 1670). All vessels with surviving grits had the later fired clay type that date to the second half of the 2nd century AD or later. MOOX Oxfordshire white ware mortarium (Tomber and Dore 1998: OXF WH) MOOXF Oxfordshire red-slipped mortaria (Tomber and Dore 1998: OXF RS) MOOXW Oxfordshire white-slipped mortaria (Tomber and Dore 1998: OXF WS) Oxfordshire products included sherds from two white ware vessels including a mortarium from Phase 5 (No. 15), a white-slipped sherd with a bead and flanged rim from Phase 5 and a wall-sided red-slipped vessel also from Phase 5. These vessels were likely to date to the late 3rd or 4th century AD.

11 MOCRF: A Crambeck Type 7 mortarium in fine Parchment ware fabric with painted decoration on the external wall. Area 4, built-up ground 4040, Phase 2, D95 12 MOCRF: A Crambeck Type 7 mortarium in fine Parchment ware fabric with painted lattice and band of roundels externally, worn internally and burnt over breaks. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D55 13

MOCRF: A Crambeck Type 7 mortarium in fine Parchment ware fabric with a painted regular lattice decoration containing hollow painted circles externally. Area 2, subsoil/Roman dump deposit 202 (GS15), Phase 5, D144 (Figure 114)

15 MOOX: An Oxfordshire white ware mortarium with a bead and flange rim and internal wear. Area 2, subsoil/Roman dump deposit 202 (GS15), Phase 5, D136

MOCRF: A Crambeck Type 7 mortarium in fine Parchment ware fabric with painted diagonal 92

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 114: Paint-decorated Crambeck type 7 mortarium, illustrated vessel 14 Figure 113: Paint-decorated Crambeck Type 7 mortarium, illustrated vessel 13

MORH Rhineland mortaria – Soller (Tomber and Dore 1998: SOL WH)

lids or covers and vessel number 10 appears to have a sooting shadow, probably from being used as a lid on a stew pot. Monaghan (2000: 144) has noted that there were few specialist ceramic lids in circulation at the end of the Roman period in this region and, if this was the case, using other materials or repurposing other sherds may have fulfilled needs. Whilst it is possible that some mortaria were purposely utilised over the fire for cooking or warming foodstuffs the sooting patterns and re-working of a few of the sherds from this site raises the possibility that the burning and sooting on some mortaria may been caused by their use as makeshift lids.

A sherd from a Soller type mortarium was retrieved from Phase 5. Examples of vessels in this fabric are known from other nearby sites including York and Lincoln. Mortaria of this type were imported into Britain in the later 2nd to earlier 3rd century AD. MONV Nene Valley mortaria (Tomber and Dore 1998: LNV WH) MONVC Nene Valley colour-coated mortaria (Tomber and Dore 1998: LNV CC) A small quantity of Nene Valley mortaria was recorded including a white ware vessel with a reeded rim and a sherd from a colour-coated mortarium from Phase 5. Although examples of vessels in these fabrics are known from York and Lincoln it would appear likely that the local Crambeck ware industry, active in the later 3rd to 4th century AD, largely fulfilled the needs of the local inhabitants.

Fine Wares I.M. Rowlandson with H.G. Fiske Seven hundred and seventy-three fine ware sherds were retrieved (8.073 kg, 10.41 RE). Non-samian Roman fine wares made up 3.67% of the assemblage by sherd count, 0.25% by weight and 3.41% by RE. Nene Valley type colour-coated wares were the most numerous, particularly beakers with significant quantities of Crambeck Parchment ware also present. Small quantities of other fine wares dating to the 2nd century AD were also present, but it would appear that much of the tableware dating to the late 2nd to early 3rd century AD consisted of samian ware bowls and dishes (J.M. Mills this volume).

Use wear evidence from the mortaria The mortaria from this assemblage, when basal sherds survived that had not been heavily abraded or burnt, nearly all showed signs of heavy use on their internal surfaces for grinding (recorded on a total of 119 sherds). Levels of excoriation caused by abrasion postbreakage appeared to be at broadly similar levels to the rest of the assemblage, but it was noticeable that many of the mortaria showed signs of burning or carbonised deposits (47 sherds). Eleven of these vessels had been roughly hewn and had sooting marks over the broken edges. Two Crambeck mortaria illustrated above (Nos 9 and 10) appear to have been roughly fashioned into

CC1 – Nene Valley type colour-coated ware with a light fired core (Tomber and Dore 1998: LNV CC). Three hundred and seventeen sherds of the Nene Valley colour-coated CC1 type fabric were recorded from a maximum of 279 vessels. Production of this type of 93

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire colour-coated pottery began in the later 2nd century AD and appears to have continued on into the 4th century AD. Other production centres with light-firing clay such as Lincoln (Tomber and Dore 1998: SOC CC) were able to produce similar colour-coated wares although it is likely that the majority of the material in the CC1 category from this site were Nene Valley products. The majority of vessels in this fabric group from Brooklyn House were beakers. Most of the vessels were funnel necked types dating to the 3rd to 4th century AD. Of the beaker types likely to date to the 4th century AD there were 13 beakers with painted decoration, three examples of slit folded beakers (Howe et al. 1980: no. 53), a vessel with rounded indents (Howe et al. 1980: no. 52) and three examples of pentice moulded type beakers (Howe et al. 1980: no. 55). Two examples of hunt cup type beakers with barbotine animals that were recorded may have dated to the late 2nd to 3rd century AD. Fragments from three Castor boxes including a late example with a flat base from Phase 4 (Darling and Precious 2014: no. 245) and four Castor box lids were noted. Other open forms present included a bowl broadly mimicking samian form 36 (Phase 5), two examples of straight-sided bead and flanged bowls. Body sherds from eight flagons are jars were recognised and a further rouletted vessel, perhaps a beaker or a flagon was retrieved from the Phase 2 bustum burial (No. 16). The bias towards beakers from this assemblage may be due to two factors: much of the CC1 from this assemblage may have been produced in the 3rd century AD when beakers were the most widely distributed type and the Crambeck industries appear to have supplied much of the tableware used in Norton during the Late Roman period negating the need for the import of many of the bowl forms produced in the Nene Valley in the 4th century AD. The Crambeck industry, discussed further below, does not appear to have produced many beakers and it would appear that Nene Valley colour-coated types were also used in the 4th century AD to fulfil the requirements of the local inhabitants for drinking vessels.

the funnel necked type. Fewer than ten sherds were retrieved from Phase 1 and 2 deposits with the majority of the material recovered from Phase 4 and Phase 5 deposits. Of the beaker types that could be securely dated to the 4th century AD there were 18 with painted decoration, a vessel with rounded indents (Howe et al. 1980: no. 52) and five examples of pentice moulded type beakers were identified. Body sherds from at least another nine vessels with a zone of rouletting likely to be of the same pentice moulded type were also noted (Howe et al. 1980: no. 55). A single vessel, possibly from a jar or beaker, was noted from Phase 5. The form types and decorative style of the vessels in this fabric group suggest that they were produced in the 4th century AD. It would appear that the inhabitants favoured these vessels rather than the limited number of beakers that were produced in the Crambeck grey ware fabric and represent the beakers in use by the inhabitants CC3 – Colour-coated ware with a light orange fabric and darker colour-coat probably including vessels produced at Brough-on-Humber (Darling 2005; Mills, P. 2018, fabric F13) Twenty-three sherds from seventeen vessels were attributed to this fabric group, a proportion of this material may have been produced at Broughon-Humber in the 2nd century AD although other production sources may also have made similar fabrics. The sherds in this fabric were recorded in small quantities from Phase 2 onward and could almost exclusively be attributed to beakers. The main vessel of note was a bag-shaped vessel decorated with barbotine decoration of stylised greyhound type hunting dog that was too fragmentary for illustration. Examples of vessels with this type of decoration are illustrated by Darling from Brough-on-Humber (2005: figs 26-28). Further sherds from beakers decorated with zones of rouletting were recovered from Phase 4 and 5 deposits (Darling 2005: fig. 2, 9-11). Examples of this fabric have been recorded by Mills from the Norton Wood Street site (Mills, P. 2018) although it appears likely that there was only limited production at the Brough kiln site (Darling 2005).

16 CC1: Colour-coated vessel with scored horizontal groove and zone of rouletting, no colour coat internally. This vessel would appear likely to be similar to a vessel from Water Newton in a group dated to the late 2nd to early mid-3rd century AD (Perrin 1999: fig. 30, 17). Registered Find (RF) 825. Area 4, bustum cremation pit 475, fill 476, Phase 2, D87 (Figures 115, 156)

CC – Miscellaneous colour-coated wares. Twenty-four sherds of colour-coated pottery could not be securely attributed to a fabric, many due to excoriation or burning, and the fragments had a mean sherd weight of (5.42 g). The sherds were mostly from small fragments from beakers and a Castor box recovered from Phases 3-7. It is likely that these sherds are related to fabrics CC1-3.

CC2 – Nene Valley type colour-coated ware with an oxidised red core and a dark colour-coat. Examples of this colour-coated fabric are known to have been produced by the later kilns in the Nene Valley utilising an iron-rich clay, but it is possible that some of this material may have been manufactured elsewhere. One hundred and sixty-four sherds were recorded in this fabric almost exclusively from beakers. Nearly all of the beakers appeared to be variations of

CGBL – Central Gaulish Black ware (Tomber and Dore 1998: CNG BS). Sixteen sherds from Central Gaulish Black ware beakers were recovered, mostly from Phase 5 but with two sherds from Phase 4 and one sherd each from Phases 94

I.M. Rowlandson and H.G. Fiske: The Roman pottery

2 and 3. These sherds were produced from around AD 150 until the early 3rd century AD and occur in small quantities at Lincoln and York (Monaghan 1997; Darling and Precious 2014).

has largely focused on the unusual vessels from this assemblage that are discussed below in more detail. Bowls were the most common form with Crambeck 5b, 5, 10 and 10A the most common types with only a few sherds from beakers, a flagon, a narrow-necked jar and a Smith god pot present. The illustrated vessels are discussed further below.

CRPA – Crambeck Parchment ware (Tomber and Dore 1998: CRA PA). Crambeck Parchment wares have been recognised as a key indicator of the latest phase of pottery production in the north of England (Bidwell 2005; Bidwell and Croom 2010). The mortaria have been discussed above (MOCRF). With the inclusion of a few sherds from Phases 4-7 that were heavily abraded there were a total of 193 sherds from a maximum of 176 vessels from the Brooklyn House assemblage (0.92% of the assemblage by sherd count, 1.07% by weight and 1.89% by RE).

17 CRPA: Everted rimmed bowl in Crambeck Parchment ware (Type 13b) with painted stripes. Area 2, subsoil/Roman dump deposit 202 (GS12), Phase 5, D102 (Figures 115, 116) 18 CRPA: Narrow-necked jar in Crambeck Parchment ware with cordon, stabbed, and painted decoration. Area 2, subsoil/Roman dump deposit 202 (GS12), Phase 5, D56 (Figure 115)

As this fabric was a key indicator of date it is discussed further by Phase below. It is noticeable that amongst the cut features of Phase 1 it is entirely absent. Only 14 CRPA sherds were retrieved from Phase 1 ground layers, mostly from Crambeck Type 5b bowls and a fragment from a Smith god pot (No. 27), that were probably intrusive. Likewise fragments from two vessels found in layers from Phase 2 layers may be explained in a similar fashion. Only with Phase 3, construction cut 3091, was there a single stratified large sherd from a Crambeck Type 5b bowl (No. 23, below).

19 CRPA: Flanged bowl in Crambeck Parchment ware with painted diagonal line decoration, burnt. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS03), Phase 5, D31 (Figure 115) 20 CRPA: Crambeck Type 5b bowl with painted scroll and diagonal line decoration. Area 4, rubble deposit 410, Phase 4, D77 (Figure 117) 21 CRPA: Crambeck Type 13b bowl with unusual painted arc decoration. Area 2, subsoil/Roman dump deposit 202 (GS03), Phase 5, D32 (Figure 115)

As the products of the Crambeck industry have been comprehensively published to a clear typology (Corder 1928; 1937; Evans 1989; Monaghan 1997) illustration Form

Description

BCR5

Crambeck type 5 Hemispherical flanged bowl

7

4.67

335

7.83

0.74 12.82

BCR5B

Crambeck type 5b Hemispherical flanged bowl

19

12.67

725

16.94

0.99 17.16

82

1.92

BCR5A BCR9

BCR10

Sherds Sherd % Weight (g) Weight % Total RE RE %

Crambeck type 5a Hemispherical flanged bowl Crambeck type 9

Crambeck type 10 bowl Crambeck type 13b

Bowl/ dish other

BFB

BKEV BK? F

JNN J

L?

TOTAL

3 2

BCR10A-D Crambeck type 10a bowl and variants BCR13B

4

Crambeck type 11- beaker with everted rim Crambeck type 14 flagon

399

7

4.67

299

24.67

2

1.33

1

Jar?

4

Lid?

2

Face or cult vessel

5

Unknown

34

3.33

37

1

Narrow neck as Corder 1928 Pl. VII.191

1.33

116

10.00

2

Other beaker

2.00

108

15

5

Other flanged bowls

2.67

150

1.33 0.67 0.67 2.67 1.33 3.33

22.67

100.00

141

2.71 9.32 3.29 6.98

771

18.01

24

0.56

14 14 25 69 22

100

1037 4281

Table 7: CRPA Crambeck Parchment ware other forms by type

95

2.52

0.33 0.33

0.12 2.08 0.27 4.68 0.11 1.91 1.18 20.45 0.59 10.23 0.67 11.61 0.6 10.40

0.14 2.43 0 0

0.00 0.00

0.58

0.13 2.25

0.51

0.02 0.35

24.22

0.21 3.64

1.61 2.34 100

0

0.00

0

0.00

5.77

100

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 115: Illustrated fine wares

Figure 116: Crambeck parchment ware bowl, illustrated vessel 17

Figure 117: Crambeck parchment ware bowl type 5b, illustrated vessel 20

22 CRPA: Crambeck Type 5b bowl with painted swastika or crux gammata motif beneath the rim and lines on the flange. The crux gammata is a common device in Greco-Roman art and was also considered as a sacred and auspicious symbol in eastern religions and has been found at Troy and Anatolia (Watts 1991). The term crux gammata, derives mainly from its appearance, which is identical to four Greek gamma letters (Γ) affixed to each other. The crux gammata cross appears to have been used in a number of flowing architectural designs shown on friezes on the Aria Pacis at Rome and there are various examples of this motif on mosaics including sites such as Pompeii. Romano-British mosaics with this meandering motif including the examples from

Leicester, Littlecote, Chedworth and Woodchester amongst others (Neal and Cosh 2002; Cosh and Neal 2005; 2010). Local examples have been found at Aldborough and Rudston villa (Neal and Cosh 2002; Stead 1980). The intricate but replicable nature of this pattern must have made it a popular with mosaicists (Lieu and Toussaint 2010). Other occurrences in Britain include brooches from sites such as Benwell where, as with the wheel iconography discussed below, an association with Tanarus has been suggested (Irby-Massie 1999: 61). It appears that this symbol was adopted by Christians; individual examples of the symbol are rarer and occur on some of the plaques in the Roman catacombs in the 3rd century AD (Watts 1991). A crux gammata motif on a dolphin-loop 96

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 118: Crambeck parchment ware bowl type 5b, with painted crux gammata illustrated vessel 22

Figure 119: Crambeck parchment ware bowl type 5b with paint decoration, illustrated vessel 23

buckle from Tripontium associated with peacocks flanking a fruit tree or ‘tree of life’ was considered by Chadwick Hawkes to represent Christian symbolism and date to the late 4th century AD (1973). It is unclear if this example from Brooklyn House represents a religious symbol, Christian or otherwise. A quick search through the examples of illustrated Crambeck parchment ware failed to uncover a comparative example but repeated ‘lazy S-shaped’ decoration as a frieze was quite common (Evans 1989) that might be considered to be representing a ‘double gamma letter’. It is unclear if the decoration of this bowl was viewed by the by the potter or user of this bowl as religiously symbolic although it would appear possible. Area 2, subsoil/Roman dump deposit 202 (GS17), Phase 5, D139 (Figure 118)

have been noted Braithwaite considered the examples recovered from the Malton vicus excavations likely to have been Crambeck products (Braithwaite in Bidwell and Croom 1997: 103-106; G Braithwaite archive). It is notable that the sherds published by Braithwaite mostly had painted face detail but one sherd had a fragment of an applied nose and eyebrow and a crudely applied hammer that had some similarities to the applied strip (No. 26) and the more accomplised moulding of hammer No. 27. A further Parchment ware vessel appeared to be a fragment from a face neck flagon which had lost the applied face plaque (No. 24). A number of similar vessels from York have been recorded by Monaghan (1997: 926928). Further examples of this form were retrieved from this site in the other fabrics and are discussed further below (Nos 59-63). (Figure 120)

23 CRPA: Crambeck Type 5b bowl with painted ‘eyes’ and intersecting zigzag lines, burnt. Area 3, construction cut 3091, fill 3089, Phase 3, D117 (Figure 119)

24 CRPA: Face-necked flagon in Crambeck Parchment ware, applied face plaque missing revealing reduced core. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D52 (Figures 115, 121)

The Crambeck parchment ware face pots and Smith god pot – H.G. Fiske Almost all British face pots found north of the Wash were made in grey ware (Braithwaite 2007) however three fragments in Crambeck Parchment ware were among the assemblage from Brooklyn House, including a fragment with painted comb-decorated hair and further paint decoration (No. 25), a possible grooved face pot ‘eyebrow’ (No. 26), and an applied hammer from a Smith god pot fragment with applied hammer and painted stripe (No. 27, see discussion of Face pots below). Although production of similar parchement wares vessels at Pakenham and in the Nene Valley

25 CRPA: Face pot fragment in Crambeck Parchment ware with small area of painted combed wavy line ‘hair’. Area 2, interior of Structure A, subsoil/ Roman dump deposit 206 (GS13), Phase 5, D94 (Figure 120-1) 26 CRPA: Possible face pot fragment in Crambeck Parchment ware comprising of an applied strip with a scored line along the centre and paint decoration. Area 2, subsoil/Roman dump deposit 202 (GS17), Phase 5, D138 (Figure 120-2) 97

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 120: Crambeck parchment ware face pot and smith god pot fragments, left to right No. 25-27

27 CRPA: Fragment of ‘Smith god’ pot in Crambeck Parchment ware featuring applied hammer motif. Area 2, Roman ground level 218, Phase 1, D100 (Figure 120-3) The chicken or cockerel figurine – H.G. Fiske 28 CRPA: Partial hollow head from a cockerel figurine in Crambeck Parchment ware. The chicken or cockerel had a special significance to the Romans and there is a known connection to the god Mercury. One of the best-known examples of a bird figurine from Britain is the pigeon or duck from King Harry Lane, Verulamium (Rigby and Stead 1989; Johns and Rigby 2005) but this is clearly different in subject matter, fabric and manufacture. White pipe clay figurines of birds and other animals were also imported from Gaul in the 2nd and 3rd centuries (Fittock 2015; 2016); this example is in Crambeck parchment ware and appears to have no published parallel from the

Figure 121: Crambeck parchment ware face necked flagon with missing plaque, illustrated vessel 24

Figure 122: Images showing the Crambeck parchment ware chicken/cockerel figurine (No. 28)

98

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Crambeck kilns; it is likely to have been part of a finial or lid. This vessel, like the other Crambeck Parchment wares, ought to date to the late 4th century AD (Bidwell 2005). RF 112. Area 2, rubble spread 215, Phase 4, D131 (Figure 122) MOSL – Moselkeramik (Tomber and Dore 1998: MOS BS) Twenty-two sherds of colour-coated Moselkeramik were recorded (78 g, 0.06RE) from a maximum of 11 vessels. These dark colour-coated vessels were produced near Trier and exported to Britain in the late 2nd to mid3rd century AD. Sherds from two vessels were retrieved from Phase 1 with small quantities also identified from Phases 2-5. All of the forms recorded were beakers of the typical funnel necked type with seven vessels showing signs of folded or indented bodies (Tyers 1996: fig. 149. 3).

Figure 123: Stamped Parisian ware sherd (No. 29) and rubbing (rubbing scale 1:1)

Dragonby and Roxby. The two stamps on this vessel do not appear amongst the material known from Market Rasen or Rossington Bridge (Darling forthcoming; Elsdon 1982; Samuels 1983, Buckland et al. 2001). The stamp that appears the closest match was a vessel illustrated by Elsdon from Thealby (1982: fig. 7.92; Rigby and Stead 1976: fig. 91. 23) or perhaps an example from Old Winterton (Rigby and Stead 1976: fig. 90.10). An Antonine date for this sherd would appear most likely. Area 2, subsoil/Roman dump deposit 202 (GS02), Phase 5, D99 (Figure 123)

OXFIN – Miscellaneous fine oxidised wares A single fine oxidised sherd that had been burnt was retrieved from unstratified context 200. OXRC – Oxfordshire red-slipped ware (Tomber and Dore 1998: OXF RS). Five small fragments of Oxfordshire red-slipped ware were recorded: one each from Phases 3-5 and a sherd from an unstratified context. The sherd from Phase 3 had a post firing piercing and, although damaged, appeared to have been fashioned into a spindle whorl. GFIN – Miscellaneous fine reduced ware

Oxidised wares Oxidised wares made up 2.4% of the Brooklyn House assemblage by sherd count. This small quantity of oxidised ware was in contrast to the 11.6% from the Norton Wood Street assemblage and 3.4% from the Malton Jack Berry House site (Mills, P. 2018; Mills and Evans 2017). The low level of oxidised wares present reflects the relative lack of pottery dating to the 2nd century AD from this assemblage. The Norton Wood Street assemblage was shown to have a considerable proportion of pottery dating to the 2nd century AD, when oxidised Ebor wares and local oxidised wares were produced in significant quantities and prior to the expansion of the Norton grey ware industry in the 3rd century AD. Although part of the Malton Jack Berry House assemblage dated to the 3rd century AD there was still a significant proportion of material dating to the late 1st to 2nd century AD. The oxidised pottery from the Brooklyn House assemblage mostly consisted of tableware vessels dating to the 3rd or 4th century AD. Notable exceptions were fresh fragments from at least two white-slipped flagons, one from Phase 3.

Twenty-four small sherds from fine grey ware vessels were recorded. Where a form could be deduced most of the sherds appear to have been from beakers including one vessel with an everted rim and an unusual fine grey ware beaker with clay pellet rough-cast decoration. Sherds in this fabric group were retrieved from Phases 2-5. PART – Miscellaneous Parisian wares (Darling and Precious 2014). A single sherd from a fine grey ware with Parisian ware stamped decoration was retrieved, No. 29, discussed below. 29 PART: A fragment of Parisian ware with rouletting and stamped rosette and fern motifs. The rosette stamp from this vessel was distinctive with a slight notch evident, a type known from Old Winterton and Thealby (Elsdon 1982: fig. 7. 81; Rigby and Stead 1976: fig. 91. 26 and 32). This fragment was probably from a vessel of Elsdon Type 5 (1982: fig. 4) with rouletted decoration as with the Thealby example, the orientation of the sherd is not certain. Fern stamps are common amongst the products of Lincolnshire, particularly from the area around Roxby, and a number of examples have been recorded from the villa at Rudston in a similar ‘blue-grey’ fabric manufactured at

CR Miscellaneous white ware flagon fabric It was noticeable amongst the assemblage that there were no examples of white ware flagons likely to date to the later 1st to 2nd century AD. Examples of these fabrics at Malton and Norton are rare (Bidwell and 99

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire Croom 1997: nos 1 and 32; Mills, P. 2018: 56) and the relative lack of 2nd-century flagons in white ware fabrics was perhaps not surprising considering oxidised and white-slipped flagons were available from local sources such as York. Five sherds were attributed to the miscellaneous white ware category including an unusual cup fragment possibly from a ritual vessel or triple vase type vessel (No. 30).This vessel is discussed further in the catalogue below but it is not certain that this sherd relates to activity on the site in the 2nd century AD and it may be an unusual vessel produced in the 4th century AD at Crambeck.

YE1/NOROW2 Ebor ware 1 (Monaghan 1997: E1; Tomber and Dore 1998 EBO OX) YE2 Ebor ware 2 (Monaghan 1997: E2) YE6 Ebor painted ware (Monaghan 1997: E6) Limited quantities of oxidised Ebor wares from York were recorded at Brooklyn House. Of the wares considered to have originated in York, the Ebor 1 fabric (YE1/ NOROW2) totalled 14 sherds including a bowl with a low carination, a jar or beaker with an everted rim, a segmental flanged bowl (Monaghan 1997: Type BF2) and a hemispherical flanged bowl (Bidwell an Croom 1997: no. 423). Examples of the coarser Ebor 2 fabric were restricted to seven sherds from Phase 5, including rims from two ‘African’ style bowls (Monaghan 1997: Form BA). Sherds of Painted Ebor 6 were retrieved from Phases 3, 4 and 5; the forms present consisted of the common hemispherical bowl, a segmental flanged bowl and beakers, all common forms for this fabric (Monaghan 1997: forms BH1, BF2 and fig. 323. 3141). This decorative tableware was produced at York in the Hadrianic to Antonine period and rare examples are known from Malton (Bidwell and Croom 1997: 62). All of the sherds in Ebor 6 fabric were residual within contexts dating to the 4th century AD or later.

30 CR?: A single cup or bowl element possibly from a triple vase in light fired fabric, burnt, appears repaired and/or reground. The condition of the fabric make it a possibility that this was manufactured in a local Crambeck fine white ware. Examples of triple pots are known from York (Perrin 1990: no. 1242) where the vessels were joined at their girth or via a basal tube (e.g. Darling and Precious 2014: nos 1429-1430); however this example lacks the usual piercing suggesting a possible alternative source. Braithwaite (2007: 191, fig. G7) discusses an apparently unique face vase from Noricum which has three blind spoutcups mounted on top of its handles and this sherd bears a marked similarity to a cut-down version of one of those. While it is not suggested that this is from such a rare vessel Braithwaite does say that the Noricum vessel, in having these blind spoutcups, does bear a similarity to snake pots (without faces) found at Vindonissa and Augst in Switzerland which, it is suspected, were connected with the cult of Sabazius (Braithwaite 2007: 483). RF 427 Area 2, subsoil/Roman dump deposit 202 (GS13), Phase 5, D115 (Figure 124)

OXWS White-slipped ware Ebor type (Monaghan 1997: W1) The most numerous types of pottery produced at York were the white-slipped wares (recorded as OXWS). One hundred and forty-eight sherds, from a maximum of 26 vessels were recorded which could be paralleled to Ebor white-slipped ware (Monaghan 1997: Fabric W1). Single sherds were retrieved from Phase 1 and Phase 2 but the majority of sherds, from a single flagon (No. 31) in fresh condition were retrieved from Phase 3. Although a sherd from a jar with an everted rim was recorded it would appear that the majority of the sherds in this fabric were from flagons. The frequency of flagon sherds would suggest that the inhabitants looked to these white-slipped wares to fulfil their demand for whitecoloured flagons. Monaghan (1997: 876-877) suggested that most of the white-slipped sherds produced at York dated to Ceramic period 2b-3a (broadly AD 160225). This date range would fit with that given for the commencement of Roman activity on the site in Phase 1. The smashed flagon No. 31, found discarded in Phase 3, may represent a vessel that arrived on the site during the initial phase of occupation. The majority of the Ebor wares were retrieved from the later Phase 4 and Phase 5 deposits where they probably represent dumping of material from earlier occupation, not necessarily on the site itself. It was notable that only two Ebor wares sherds were recorded from Phase 1 deposits, one sherd from Phase 2 and sherds from a maximum of two vessels from Phase 3.

Figure 124: Possible Triple vase single element in white ware fabric (No. 30)

100

I.M. Rowlandson and H.G. Fiske: The Roman pottery

31 OXWS This white-slipped vessel appears to be a variant on the Gillam Type 16 (1970) Monaghan’s York form FP (1997; Perrin 1990: nos 117-119) commonly seen with a white-slip at York. Although the rim from this example was a little more complicated than standard pulley wheel forms D Williams suggested it was produced within the province rather than being imported from outside. As examples of this form with a white-slip are not evident amongst the illustrated vessels from Malton (Bidwell and Croom 1997) it is presumed that this vessel is a York product. Area 3, Structure D internal floor surface 3007, Phase 3, D118

group similar to Swanpool type Late Roman oxidised wares with burnished external surfaces, eleven sherds were attributed to this fabric group (SPOXT). Forms in this fabric group included examples of hemispherical flanged bowls mimicking samian form 38, a plain rimmed dish and a beaker base that had been ground down to form an inkwell or counter (No. 33). The majority of sherds in this fabric group were retrieved from Phase 4 and 5 deposits. Monaghan considered the examples of these late red wares to be most common in York amongst the Ceramic Period 4a, late 3rd to mid4th century AD (1997: 907). Although the production source of some of these red wares is not certain it would appear that they “fulfilled the requirement for fine red wares, often in samian forms, which took place in the later 3rd to 4th centuries” that, elsewhere, were satisfied by the products of Oxfordshire, Argonne and elsewhere (Monaghan 1997: 907).

NOROW1 Local oxidised ware (Bidwell and Croom 1997: OW1) One hundred and twenty-four sherds were attributed to the Malton fabric OW1 (recorded as NOROW1, Bidwell and Croom 1997). This fabric appears to represent the late 1st- to 2nd-century AD local oxidised wares. The majority of the sherds in this fabric could not be securely attributed to a form, with only a few fragments from necked jars, a storage jar, a hemispherical bowl mimicking samian form 37, a bowl or dish with a plain rim (No. 36) and three sherds from tazza type vessels being recorded. The majority of sherds attributed to this fabric were retrieved from Phase 5 with only two sherds recorded from Phase 1 deposits.

32 CROX A copy of a samian form 38 hemispherical flanged bowl in Crambeck oxidised ware with paint decoration. Area 2, subsoil/Roman dump deposit 202 (GS09), Phase 5, D19 33 SPOXT: A beaker base in Swanpool type oxidised type ware, appears cut down perhaps to make an inkwell or small vessel, RF 396. Area 4, layer 409, Phase 2, D129 (Figure 125)

36 NOROW1 A plain-rimmed dish in Norton oxidised ware. Area 2, rubble deposit 230 (GS13), Phase 4, D64 NOROW3 Local mica-rich oxidised ware (Bidwell and Croom 1997: OW3) Examples of the mica-rich oxidised Fabric 3 defined by Bidwell and Croom (1997: OW3) were restricted to four sherds from Phases 4 and 5. Five sherds were attributed to the Malton oxidised ware Fabric 4 (Bidwell and Croom 1997: OW4), which was probably a 3rd-century AD or later oxidised version of the Norton grey ware fabric. CROX Crambeck red ware (Evans 1989: 55) SPOXT Swanpool Oxidised type ware (Darling and Precious 2014, SPOX?)

Figure 125: Oxidised beaker base showing signs of being re-ground to form a small container (No. 33)

Fifty sherds of burnished Late Roman oxidised sherds were recorded in the Crambeck red ware group (recorded as CROX, see description in Evans 1989: 55). The sherds were mostly from hemispherical flanged bowls, Crambeck Type 5, a common type in this fabric along with examples of straight-sided bead and flanged bowls (Crambeck Type 1), a funnel necked beaker and an example of a Type 10 bowl with white painted diagonal line decoration (rim form as Corder 1928: pl. III.70 and Bidwell and Croom 1997: no. 361). Late Roman oxidised wares present also included a

OX/OX? Miscellaneous oxidised wares A further 116 sherds, many in a poor condition, were attributed to a miscellaneous oxidised fabric (OX/ OX?) during recording and this group would appear to almost exclusively match the local OW4 fabric (Bidwell and Croom 1997). The forms present included beakers, straight-sided bead and flanged bowls, hemispherical flanged bowls and a bowl with a plain rim (Monaghan 101

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire 1997: Types DF3, BF3 and DD1). A heavily warped vessel, probably a misfired local vessel, was recorded from Phase 5 (No. 34). The majority of this group was recorded from Phases 4 and 5 but a small number of sherds from Phase 1 may represent contemporary oxidised wares dating to the 3rd century AD.

Sherds from a maximum of eight Black Burnished ware 2 type jars were recorded including sherds from a jar, a dish with a grooved rim and a bowl with a triangular rim and burnished lattice decoration. The small number of sherds from this site would conform with the evidence from Malton that suggested the Black Burnished ware 2 industries were not a major supplier to this part of Yorkshire (Bidwell and Croom 1997: 68).

34 OX? A heavily burnt and warped ‘second’ vessel in oxidised fabric with sintered surface and dunting cracking, possibly reused. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D10

NORBB1: Norton Black Burnished ware 1 (Bidwell and Croom 2012, NOR BB1)

DBY Derbyshire ware (Tomber and Dore 1998: DER CO).

One hundred and fifty-one sherds of local Black Burnished ware 1 (2.586 kg, 3.45 RE) were recorded. Small quantities of this ware were found in Phases 1-5. The forms were typical of those illustrated by Bidwell and Croom (2012) with examples of jars mimicking southern BB1 prototypes (Type 1) that were marginally more common than the simple bowls and dishes (Type 2). Lipped dishes (No. 38) and plain rimmed dishes (No. 39) were noted amongst the assemblage and a small number of beakers were recorded including one example with a handle and a fragment from a strainer from Phase 5. A single misfired vessel with dunting cracking was recorded from Phase 5 (No. 39). Bidwell and Croom (2012: 21-22) have considered the local Black Burnished ware 1 production at Norton to date to sometime in the late 2nd to earlier 3rd century AD that would equate to Phase 1 at the Brooklyn House site. The most significant vessel from this assemblage was the nearly complete lipped dish from the Phase 2 bustum burial associated with finds that were probably dating to the 3rd century AD (No. 38).

Six sherds of Derbyshire ware were recorded from the site including a jar with a channelled rim (No. 35), a further three body sherds from Phase 5 and an example of the classic Gillam Type 152 Derbyshire ware jar (Gillam 1970) from Phase 4. This material was an unusual find in northern Yorkshire, Monaghan (1997: 913) recorded no examples from York and, although examples are recorded from sites on Hadrian’s Wall these sherds are an unusual find from this area (Tyers 1996). These sherds, which may represent only two vessels amongst a very large assemblage, may merely represent personal possessions which arrived from Derbyshire in the 3rd or early 4th century AD rather than representing any significant trade contacts. 35 DBY A channel-rimmed jar in Derbyshire ware fabric with a channelled rim, an unusual find from a site in this part of Yorkshire. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS13), Phase 5, D34 Reduced wares Ten thousand six hundred and ninety sherds, 50.76% of the assemblage fell into the reduced ware category (173.240 kg, 170.54 RE). The majority of this pottery could be classified as local products of the Norton or Crambeck industries with the addition of small quantities of wares from further afield.

37 NORBB1?: A jar with a curved everted rim possibly in Norton Black Burnished ware or perhaps a Late Roman burnished ware. Area 2, subsoil/Roman dump deposit 202, (GS06), Phase 5, D13 (Figure 126) 38 NORBB1: A lipped dish in Norton Black Burnished ware, burnt, sherds found in two contexts, RF 826. Area 4, pit 475, fill 473 and cremation pit/ground surface 475/477, fill 490, Phase 2, D88 (Figures 126, 156)

BB1: Black Burnished ware 1 (Tomber and Dore 1998: DOR BB1). Ninety-eight sherds of Black Burnished ware 1 (2.181 kg, 3.25RE) were recorded including a Hadrianic type jar with burnished wavy line decoration under the rim, a lipped bowl and a bowl with a grooved rim all possibly dating to the 2nd century AD (Phase 5). The majority of recognisable forms dated to the Late Roman period and included sherds from up to 13 straight-sided bead and flanged bowls and a jar with a curved ‘cavetto’ type rim from Phase 5.

39 NORBB1?: A plain-rimmed dish probably a Norton Black Burnished ware, crudely made with dunting cracks. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D14 (Figure 126) NORGW1: Norton Grey ware (Bidwell and Croom 1997: GW1) Four thousand five hundred and sixty-seven sherds of the local Norton grey ware were recorded (67.908 kg, 65.69 RE). Further sand-gritted fully reduced wares that could not be closely matched to Norton kiln samples have been discussed further under the miscellaneous GREY code and a proportion of atypical Norton products may have been included in that group. The

A further fragment from a vessel with pre-fired holes, probably a strainer, was also recovered from Phase 5 and appeared likely to be a Dorset BB1 product. BB2: Black Burnished ware 2 (see Darling and Precious 2014). 102

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Form

Sherds Sherd % Weight (g) Weight % Total RE % of RE

Bowl plain rim- type 1a Dish plain rim- type 1b Lipped bowl- type 2a Lipped dish- 2b Dish with a grooved rim (Bidwell & Croom 2012, Fig. 12.8) Bowl with grooved flange (Bidwell & Croom 1997, No. 352) Conical flanged bowl- eg. type 3 Bowl dish other Large jar/ lug handled 4 Jar/ beaker everted- type 5 Jar- everted rim small jar (eg. Bidwell & Croom 1997, No. 40; Bidwell & Croom 2012, No. 26-32) Wide mouthed bowls- type 6 Narrow necked jars- type 7 Jar/Flagon Crambeck 14a Flagon disc neck- Crambeck 14 Necked jar other- eg. Bidwell & Croom 1997, No.60 Jar other Folded beakers- type 9 Carinated vessels Norton type 10 Beaker- funnel types Beakers- other Lid Other- fom uncertain

29 91 18 2 16 7 156 132 182 48

0.63 1.99 0.39 0.04 0.35 0.15 3.41 2.88 3.97 1.05

1197 2306 540 50 320 233 5169 3077 7171 448

1.75 3.37 0.79 0.07 0.47 0.34 7.54 4.49 10.46 0.65

2.55 6.2 1.53 0.2 0.96 0.63 11.96 0.59 5.21 4.89

3.81 9.26 2.28 0.3 1.43 0.94 17.86 0.88 7.78 7.3

33

0.72

511

0.75

3.54

5.29

45 55 12 2 25 79 44 36 30 9 5 3523

0.98 1.2 0.26 0.04 0.55 1.73 0.96 0.79 0.66 0.2 0.11 76.94

2168 1370 227 22 376 1442 1123 578 762 63 80 39295

3.16 2 0.33 0.03 0.55 2.1 1.64 0.84 1.11 0.09 0.12 57.34

3.02 6.49 1.33 0.52 2.23 4.41 1.6 1.81 2.06 0.63 0.24 4.36

4.51 9.69 1.99 0.78 3.33 6.59 2.39 2.7 3.08 0.94 0.36 6.51

Table 8: NORGW1 forms as proportion of fabric (Norton types follow Hayes and Whitley 1950)

pottery production at Norton has been well published already (Hayes and Whitley 1950; Hayes 1988; Bidwell and Croom 2012) so illustration has mostly focused on the significant stratified vessels (e.g. No. 41).

It was noticeable that the conical flanged bowls (Type 3), plain rimmed dishes (Type 1b), plain rimmed bowls (Type 1a) appear to have most commonly had carbonised external residues along with examples of Type 2a and a grooved flange bowl. This would appear to suggest that a proportion of the dishes and lipped/ flanged bowls were utilised for cooking (12 vessels recorded). Examples of external carbonised residues on the jars with everted rims and lug-handled jars were restricted to sherds from a maximum of five vessels.

The types present were almost all similar to those already illustrated from the Norton kilns (Hayes and Whitley 1950; Bidwell and Croom 1997). A range of small jars with everted rims (Bidwell and Croom 2012: Type 4.6-7) were present likely to be of 2nd- to 3rd-century date along with sherds from necked jars perhaps likely to also date to the 2nd century AD (e.g. Bidwell and Croom 1997: no. 60). Dishes with grooved rims (Bidwell and Croom 2012: fig. 12.8), also possibly dating to the 2nd century AD were also recorded. Using the rim equivalent measurement (Table 8) the majority of the forms recorded were small bowls and dishes with the conical bowl with a bead and flanged rim (e.g. No. 46) and the plain rimmed dish (No. 45) particularly common, both types that were typical in the Late Roman period. Utilising the rim equivalents measurement (RE) as perhaps the most useful to calibrate the assemblage narrow necked jar Type 7, large jars including lughandled Type 4 wide-mouthed bowl Type 6 and Type 5 beakers (7.30%) were also well represented.

One vessel of note was a conical flanged bowl with burnished wavy line (No. 47) retrieved from Phase 5 and similar to Crambeck Type 1b vessels (Corder 1937). This type in the Crambeck grey ware fabric was considered to have appeared in the late 4th century AD, a date after the production of Norton grey ware has conventionally considered to have ceased. This vessel may have been manufactured at Crambeck utilising the iron-rich clays that Evans described were used for the production of the late Red wares (1989) thus producing a similar fabric to the sandy Norton Grey ware. The other possibility is illustrated vessel 47 may represent some limited pottery production at Norton itself deep into the 4th century AD. No kiln evidence has yet been published at this stage to suggest that Norton type 103

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire grey wares continued to be produced into the late 4th century AD but as a number of the more generic types published from Norton (Hayes and Whitley 1950: e.g. Types 1b, 3 , 4a, 4c and 5) and Malton (Bidwell and Croom 1997: nos 191, 262) may have been in circulation in the 4th century AD. Also limited numbers of jar/ flagon Crambeck Type 14a forms and a disc necked flagon (Crambeck Type 14) from this site also appeared in a sandy Norton-type grey ware. One of the main features of the Norton grey ware assemblage was the significant proportion of conical flanged/bead and flange bowls 17.86% by RE (No. 46, broadly Norton type 3), dishes with plain rims 9.26% (No. 45) and bowls with plain rims 3.81%. Bidwell and Croom have observed that no conical flanged bowls were retrieved from the Norton Community Primary School kiln site and only up to 2.4% at the Model Farm kiln and 3% at the Grove Bungalow site (Bidwell and Croom 2012: 12; Hayes

and Whitley 1950: fig. 10.3). The conical flanged bowl and the plain rimmed bowl/dish types were common products of industries working in the later 3rd to 4th centuries AD in the north (e.g. Swanpool, Holme-onSpalding Moor and Crambeck). At present this appears to be an anomaly but suggests that this site received much of its Norton grey ware pottery in the Late Roman period and raises the possibility that further kilns with a 4th-century AD flourit may yet be found in the Norton area which perhaps had only a very limited distribution due to the rise of the Crambeck industry. Four warped vessels were recorded: a straight-sided bead and flanged bowl from Phase 4 and a second example from Phase 5 (Hayes and Whitley 1950: Type 3), a dish with a grooved rim (Bidwell and Croom 2012: no. 8) and a jar or beaker with an everted rim. A carinated jar (No. 42) from Phase 5 also signs of blistering during firing.

Figure 126: Norton reduced wares NORBB1, NORGW1 and NORCG

104

I.M. Rowlandson and H.G. Fiske: The Roman pottery

A lug-handled jar that appeared to have lost its handle during firing was noted from Phase 5 (Hayes and Whitley 1950: Type 4b). Three vessels showed signs of dunting cracking during firing: plain rimmed bowls from Phase 1 (Hayes and Whitley 1950: Type 1a) and a bowl with a grooved flange from Phase 4 (Bidwell and Croom 1997: no. 352). The presence of a range of ‘seconds’ amongst the assemblage was not surprising given the proximity of the Brooklyn House site to many of the known kilns. There was no evidence from the excavations that there was pottery production on the site itself. 40 NORGW1: A folded beaker in Norton grey ware with burnished wavy lines, unusual scored break, Unstratified, D103 (Figure 126) 41 NORGW1: A folded beaker in Norton grey ware from bustum pit RF 417. Area 4, bustum cremation pit 475, fill 476, Phase 2, D86 (Figures 126, 127, 156)

Figure 127: Folded beaker in Norton grey ware (No. 41)

42 NORGW1: A carinated jar in Norton grey ware with surface blistering. Area 2, subsoil/Roman dump deposit 202 (GS08), Phase 5, D58 (Figure 126)

with an everted rim, a beaker with a bead rim (Mills P. 2018, R02.4) and a lipped bowl. There were fewer sherds of this fabric type from the Brooklyn House assemblage and the range of forms illustrated from the Malton vicus (Bidwell and Croom 1997) would suggest that production of this ware did not extend beyond the 2nd century AD. Unsurprisingly there were a few sherds from this assemblage from Phase 1 and residual material from Phases 2-5.

43 NORGW1: A narrow-necked jar in a sandy Norton grey ware with cordon decoration. This form is more akin to the products of the Holme-onSpalding Moor and Crambeck industries. Area 2, subsoil/Roman dump deposit 202 (GS08), Phase 5, D59 (Figure 126) 44 NORGW1: A jar with a reeded rim in Norton grey ware. Area 2, rubble deposit 251, Phase 3, D67 (Figure 126)

NORCG: Norton grey ware with calcite (Bidwell and Croom 2012)

45 NORGW1: A plain-rimmed dish in Norton grey ware, forms a matching set with No. 44. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS03), Phase 5, D39 (Figure 126)

This mid-grey kiln fired ware with calcite and sand inclusions was recorded from the Norton Community Primary School site (Bidwell and Croom 2012). The majority of the vessels recorded in this group were jars with everted rims similar to examples from the kiln site (No. 49) and jars with everted rims with external bevels (Hayes and Whitley 1950: Type 8) which were probably also local products. Also a dish from Phase 1 would appear likely to have been produced at the Norton kilns. These vessels were likely to date to the later 2nd century and were almost certainly local products. However an unstratified fragment from a Huntcliff jar in a similar fabric may suggest some of the featureless body sherds attributed to this group may have been produced fairly locally at a later date.

46 NORGW1: A bead and flange rimmed bowl in Norton grey ware, small example, forms a matching set with No. 44. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS03), Phase 5, D38 (Figure 126) 47 NORGW1: A bead and flange bowl in Norton grey ware with an internal burnished wavy line similar to Crambeck types. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D46 (Figure 126) 48 NORGW1: A fragment of Norton grey ware with stamped roundels possibly representing hair and/ or beard from a face pot (see further discussion below). Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D51 (Figures 126, 138.1)

49 NORCG?: A handmade jar with an everted rim, possibly in Norton grey ware with calcite inclusions, with external burnishing. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D12 (Figure 126)

NORSGW: Early sandy grey ware (Bidwell and Croom 1997: SGW) One hundred and twenty-nine sherds of the sandy grey ware (recorded in this report as NORSGW) were recorded (1.895 kg, 1.89 RE). The forms present included small fragments from a rusticated jar, a jar

CRGR: Crambeck grey ware (Tomber and Dore 1998: CRA RE) Four thousand seven hundred and ninety-three sherds of Crambeck grey ware (84.012 kg, 79.91 RE) were 105

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire recorded. The production date for Crambeck grey ware is traditionally dated to the late 3rd to the end of the 4th century AD. The occurrence of this fabric in the stratigraphic sequence and in relation to other reduced wares is discussed further below.

Type 1b bowl with wavy line decoration, considered to be a dating indicator for the late 4th century AD, were retrieved from Phase 5 deposits. The next most common type was the plain rimmed dish Type 2. It is possible that the Crambeck Type 1, 1a, 1b flanged vessels may have formed an interlocking ‘casserole set’ (Gillam 1976: fig. 6) with the Type 2 and 2a dishes as pre-breakage external carbonised deposits were found on a considerable proportion of these vessels and they were absent from all of the over CRGR vessels except from one large jar, one widemouthed bowl (Type 4) and one hemispherical flanged bowl (Type 13). It was noticeable that sherds from only

This group provided a statistically viable group for studying the relative occurrence of forms (see Table 9). It was noticeable that Crambeck Type 1, the straightsided bead and flanged bowls or conical flanged bowl, vessels were the commonest vessel type making up over a third of the recognisable vessels by rim equivalent. With the exception of two intrusive vessels (in Phase 1 layers) the remaining 15 examples of the Crambeck Form BFB DFB

Description Bowl- conical flanged bowl- type 1 Dish- straight sided dish with a flanged rim- type 1a

BFB (BWL) DPR DGR BCR5

Bowl- conical flanged bowl (wavy line burnish)type 1b Dish- plain rim- type 2 Dish- grooved rim type 2a Crambeck type 5 Hemispherical flanged bowl

BCR5B

Crambeck type 5b Hemispherical flanged bowl

BCR5A

BCR13 BCR13A BCR13B BFL

Crambeck type 5a Hemispherical flanged bowl Bowl- Crambeck type 13 and variants Bowl- Crambeck type 13a

BD

JEV

Jar- everted

JL/JLH JDLS J

BWM/ BCR4 FDN

Jar other

0.58

2374

2.83

2.41

3.02

179 70 10

3.73 1.46 0.21

4185 1616 181

4.98 1.92 0.22

11.87 5.00 0.45

14.85 6.26 0.56

0

0.00

0

0.00

0.00

0.00

0 7

75

Bowl large- wide mouthed- type 4

33

Flagon- Type 14

1 2 1

L

Lid

TOTAL

28

1

JCAR

1.11

23

Jar- double lid-seated

Beaker- everted type 11 Beaker- rouletted/pentice moulded/ funnel type 12 Carinated as Norton type 10

CHP

0.89

142

BKEV

ST

0.53

426

Jar- large/lug-handled- 3

Flagon other

BK

443

6

F

0.13

9

JCR14A/ J162 Jar/Flagon- Crambeck 14a Flagon as Corder 1928, No. 186 BKFO

6

4

Bowl- Crambeck type 13b

Lipped bowl- Norton type 2a Bowl dish other

Sherds Sherd % Weight (g) Weight % Total RE % of RE 369 7.70 12642 15.05 28.80 36.04

2

Folded beakers- type 9

Beakers- other Strainer

1

29

0.19 0.13 8.89 2.96 0.48 0.02 1.56 0.69 0.02 0.04 0.02 0.04 0.02 0.61

335 99

585 156

10261 4410 394 89

1060 2132 2

40 40 8

16

218

0.40 0.12 0.70 0.19

12.21 5.25 0.47 0.11 1.26 2.54 0.00 0.05 0.05 0.01 0.02 0.26

0.45 0.44 1.19 0.37 0.83 5.42 2.87 0.12 4.65 3.18 0.00 0.26 0.14 0.08 0.00 3.85

0.00 0.56 0.55 1.49 0.46 1.04 6.78 3.59 0.15 5.82 3.98 0.00 0.33 0.18 0.10 0.00 4.82

0.22

0.88

1.10

17

0.35

228

0.27

0.95

1.19

2

0.04

8

0.01

0.06

0.08

36

3258 4793

0.75 0.04 0.06 0.73

67.97 100.00

395 32 72

722

41084 84012

0.47 0.04 0.09 0.86

48.90 100.00

Table 9: CRGR forms as proportion of fabric (Crambeck types follow Corder 1937)

106

0.00

185

35

Other- fom uncertain TOTAL

0.08

0.00

0.33

3

Ritual vessel

0.15

0

16

2

Cheese-press- eg Corder 1928, No. 188-9

0.00

1.31 0.00 0.00 0.75

2.69 79.91

1.64 0.00 0.00 0.94

3.37 100.00

I.M. Rowlandson and H.G. Fiske: The Roman pottery

two vessels had internal white residues one large jar and a wide-mouthed bowl that may have contained mortar rather than had evidence for mineralisation through heating liquids. Jars and wide-mouthed bowls were only present in small quantities but a number of significant decorated ‘ritual vessels’ were recorded that are discussed further below by Fiske.

51 CRGR: An unusual vessel, possibly a beaker, with burnished lattice decoration in Crambeck grey ware. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D54 (Figure 128) 52 CRGR: A jar in Crambeck grey ware similar to an example illustrated from Catterick (Evans 2002, SS46, no. 52). Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D48 (Figure 128)

50 CRGR: A cup-mouthed flagon in Crambeck grey ware (Crambeck Type 14a). Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D49 (Figure 128)

53 CRGR: A double lid-seated rim jar in Crambeck grey ware, Unstratified, D107 (Figure 128)

Figure 128: Crambeck reduced wares CRGR, CRGR? and CRGRC

107

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire 54 CRGR: A dish with a bead and flange rim in Crambeck grey ware (Crambeck Type 1a). Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D47 (Figure 128)

early military sites, and continue through the Roman period, with a strong connection to the army, and with worship of Bacchus. A sub-group is related to the Smith God (…) The face flagon is the latest arrival, first appearing in the later 3rd century”. (Darling 2011: 95)

55 CRGR: A bowl with bead and flange rim in Crambeck grey ware featuring three burnished wavy lines internally (Crambeck Type 1b). Area 2, rubbish deposit 216 (GS06), Phase 5, D85 (Figure 128)

The assemblage from this site included a range of these specialist vessels that would fit into these categories and are discussed in more detail below.

56 CRGR: A copy of samian bowl form 38 in Crambeck grey ware featuring a zone of rouletted decoration (Crambeck Type 13a). Area 2, subsoil/Roman dump deposit 202 (GS12), Phase 5, D114 (Figure 128)

Mask-mouthed and face-necked vessels 60 CRGR: A crudely modelled head with stabbed cordons representing a headdress or braided hair, the eyes are pellets pushed into depressions with stabbed irises and has a slashed mouth. Its form is perhaps most closely paralleled by Braithwaite’s (2007) “Figurine jugs or cup-necked flagons with a hand-moulded head or face on the neck” (RB Type 41) and more specifically as Type 41F “Cupshaped face, with “hooded” face projecting above the rim”. Parallels can be made with similar vessels from the Rhineland (Dӧvener 2000: abb 258, 262, 263) and this could be a degenerate copy of these but the German examples appear smaller and earlier. Type 41F vessels are known from the Wall area and East Yorkshire and possibly dated 3rd to 4th century. A further example of a facenecked flagon produced at Norton has also been illustrated by Hayes from excavations between St. Peter’s Street and St. Nicholas Street, Norton (1988: pl. 23). RF 426. Area 2, subsoil/Roman dump deposit 202 (GS03), Phase 5, D91 (Figures 128, 130).

57 CRGR: A Crambeck Type 13b bowl in Crambeck grey ware. Area 2, subsoil/Roman dump deposit 202 (GS09), Phase 5, D18 (Figure 128) 58 CRGR: A basal sherd from a bowl or dish in Crambeck grey ware showing substantial edge wear characteristic of reuse as a scraper or rubbing tool. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D53 (Figures 128, 129) 59 CRGR: A large wide-mouthed bowl in Crambeck grey ware with a scored horizontal groove. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS03), Phase 5, D37 (Figure 128) The religious or cultic vessel H.G. Fiske ‘Face’ vessels “There are three main types of Roman vessel which have faces depicted on them, face pots, jars or beakers with face masks on the body (Braithwaite 2007), head pots where the jar takes the form of a human head (Braithwaite 1984) and face flagons, with faces on the neck or rims (Dӧvener 2000). Whilst these rare finds can all be viewed as religious cult vessels, they occur in Britain at different times, and have differing histories and interpretations. The face pots arrive with the Roman army at the time of the conquest, appearing on

Examples are known in both Crambeck and Norton fabrics; Corder (in Wilson 1989: 32-33) describes but does not illustrate “a small human face” found at

Figure 129: CRGR Basal sherd showing evidence of re-use (No. 58)

108

Figure 130: Mask-mouthed flagon in Crambeck grey ware (No. 60)

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 131: Left to Right- No. 61-3 Sherds from face-necked flasks in Crambeck grey ware

apparently identical heads just below the rim and pushed out into a negative stamp”, she confessed she had not come across anything very similar to this before. These examples differ slightly in that the masks appear to have been applied rather than moulded.

Norton as an example of the “crude mouth-masked flagons” made at the Crambeck kilns, it had been attached to the vessel neck and the upper part of the hair (represented by incised circles) rose above the rim, the nose is moulded but the eyes are applied clay discs, the mouth is a “mere excision” and “the whole workmanship is crude in the extreme”. Further face sherds matching this description were illustrated by Hayes (1988: pl. 12.3 and 23), the first described as “From Sutton Cottage, found by F. Ashman in 1933. Described by Corder and Birley (1937: 406)”. The second was “from E.J.W. Hildyard’s excavation between St. Peter’s Street and St. Nicholas Street, Norton”. Either of these could be the vessel Corder described; although the fabric of the former sherd is described as more typical of Norton than Crambeck it is also explained that the presence of the Norton kilns was not suspected in 1937.

The face jars 64 CRGR: Joining fragments from an RB Type 28A face jar (Braithwaite 2007: fig. J13.2) in Crambeck grey ware, this type is described as “Wide-bellied face jars with incised or stamped bosses and large faces on the girth in red, grey or buff fabrics from north-east Britain.” They were mainly produced at Crambeck and earlier at York; those from Crambeck were often also Smith god vessels with applied tool and/or wheel motifs. They range in height between 22-30 cm and date from the mid3rd to 4th century. Evaluation, Roman subsoil dump deposit 502-1, D1E (Figure 128, 132)

61 CRGR: A face-necked flagon or flask in Crambeck grey ware, two small faces from the same vessel with tube-stabbed features and applied noses. Area 2, subsoil/Roman dump deposit, 202 (GS09), Phase 5, D21 (Figures 128, 131.1)

65 CRGR: A single fragment from a face jar in Crambeck grey ware of unusual type, it utilises two different roundel stamps to represent hair, and an ear which was applied afterwards. Apart from the fabric it appears very similar to Braithwaite’s (2007) description of RB Type 29 as a “Red ware face jar with stamps round the face, but no bosses”. A single example of the type is known, from Hibaldstow, and is further described as having at least two rows of rosette stamps with no bosses surrounding the face, and a small applied ear in among the stamps, and “the only face jar found so far which has stamps but no bosses”. Examination of a photo of the Hibaldstow example (Aldhouse-Green 1984: 239-242, fig. 1) shows that it is a close match, the author also hints at a head pot with very similar beard treatment from Castor, Cambridgeshire but the publication containing details of this has proved elusive (Aldhouse-Green 1980). Provisionally dated late 3rd to 4th century AD but a 4th-century date would perhaps be most likely on basis of the

62 CRGR: A face-necked flagon or flask in Crambeck grey ware, two small faces from the same vessel with tube-stabbed features and applied noses, possibly same vessel as 60 but non-joining. Area 2, Roman ground level 231 (GS15), Phase 1, D141 (Figure 131.2) 63 CRGR: Possible face-necked flagon or Smith god pot fragment in Crambeck grey ware with applied rim section and burnished diagonal lines. Area 2, Roman ground level 282 (GS03), Phase 1, D65 (Figure 131.3) Non-joining sherds from a vessel or vessels from two contexts with small tube-stab decorated faces high on the rim, in Crambeck grey ware. A very similar vessel from Malton vicus (Braithwaite in Wenham and Heywood 1997: fig. 40.5), is described as a “most unusual, possible flask neck with two schematic, 109

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 132: Face jar in Crambeck grey ware (No. 64)

Figure 133: Roundel-stamped face jar in Crambeck grey ware (No. 65)

fabric. Area 2, subsoil/Roman dump deposit, 202 (GS15), Phase 5, D133 (Figure 133)

“Roman smith pots only seem to occur in Britain, and apart from one isolated example from Wroxeter, they are only found on the east side of England from Canterbury to Chester-le-Street. They all have applied smith’s tools, either on the shoulder or the girth. For the most part these just consist of a hammer, tongs and a triangular or wedge-shaped anvil, but in some cases, and in particular at Malton, there may also be spoked wheels and crosses.” (Braithwaite 2007)

66 CRGR: Possible face pot sherd in Crambeck grey ware with stabbed hair or beard. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS03), Phase 5, D36 (Figure 134.1) 67 CRGR: Fragment of face pot near to eye in Crambeck grey ware. Area 2, subsoil/Roman dump deposit 202 (GS13), Phase 5, D72 (Figure 134.2) 68 CRGR: Possible face pot fragment in Crambeck grey ware. Area 2, Interior of Structure A, subsoil/ Roman dump deposit 206 (GS03), Phase 5, D35 (Figures 128, 134.3)

The Smith god is one of the oldest gods, originating with iron working in the Hittite areas of Anatolia (Darling 2011). The name of the Celtic Smith god is unknown, but the Romans invoked Vulcan as his successor. The distribution of Roman Smith god and Smith face jars from the Humber north to the northern frontier suggests a military connection but they are not found on the continent so this is largely conjecture, but there is an obvious connection with the dangerous occupation of iron working and the perceived need for protection by invocation (Darling 2011) and they are found on both military and civilian sites from Yorkshire to Hadrian’s Wall (Darling 2011). The vessels are typified by the addition of applied metalworking tools and motifs to often fairly standard jars and Crambeck was a notable centre for their production.

69 CRGR: Small possible face pot fragment with stabbed representation of hair or beard and a possible eyebrow. Area 2, Roman ground level 231 (GS15), Phase 1, D147 (Figure 134.4) The Smith god and Wheel god pots “It is notoriously difficult, and indeed dangerous, to draw hard and fast conclusions in any matters pertaining to Celtic religion. All that can be said with safety in this case, I think, is that there was being worshipped in Britain under the Roman occupation a god, whose characteristics included those of a craftsman, who was identified by the conquerors with their own smith god Vulcan.” (Leach 1962)

In Celtic mythology, Taranis (or Tanarus) was the god of thunder, he was worshipped in Gaul, Britain, and elsewhere and was particularly associated with the 110

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 134: Face pot fragments in Crambeck grey ware (left to right, 66-69)

six- or eight-spoked wheel symbol (Aldhouse-Green 1979; Irby-Massie 1991: 59-61) which represented the sky or sun. He became ‘hybridised’ (Goldberg 2009) and was interchangeable with the Roman sky god Jupiter (Darling 2011). The wheel motif is often found in association with Smith god tools on pots found at Malton and at Norton and appears to be part of the same suite of iconography. A similar eastern cult associated with smithing was also worshiped as Jupiter Optimus Maximus Dolichenus at York and at other sites in the north of Britain associated with craft and metal working (Irby-Massie 199: 63-71). The wheel iconography was also absorbed into Christian symbolism by as a representation of the ichthys or ichthus (ΙΧΘΥΣ) explained as a combined representation of the acronym. Examples of this representation are known from the Eastern Empire at sites such as Ephesus (Rasimus 2012).

the Crambeck kilns were later. Corder (1937) suggested peak production was at the very end of the 4th century), and supplied pottery to a wider geographic area. A single possible face pot fragment in Norton grey ware (No. 48) was found during these excavations but the majority were in Crambeck grey ware, which fits with the general late trend in this assemblage. Face vases are known in Crambeck parchment ware fabrics but fragments are rare (Wood 2016); painted parchment ware vessels from the Nene Valley area showing the Smith god at work (Darling 2011) are potentially more common. It is unclear whether the Smith/Wheel god pots found here represent a shrine or personal observance or more likely a combination of both; a finely sculpted image of the Smith god from York is considered to have come from a shrine (RCHME 1962: 128: pl. 53: 96, cited in Darling 2011) and altars dedicated to Vulcan are known from Vindolanda and Old Carlisle. Darling (2011) considers that observance of worship and ritual in premises where dangerous fire-related occupations were carried out (ceramics, smithing, smelting) would have required some form of shrine either in the workshop or household. Given the evidence for metalworking, as well as pottery production, the frequency of these finds should not perhaps be too surprising (RIB I 712; Hayes and Whitley 1950). It is perhaps also notable that applied features, wheels in particular but also hammers, appear to have been more or less carefully removed from the pots they were applied to, perhaps after breakage of the vessel and possibly for curation as personal talismans.

With the exception of three sherds, one each from Areas 4, 7 and 8, almost all of the sherds from this site came from the Area 2 refuse deposit and the area within Structure A. This pattern of distribution is very similar to the bulk of the Roman pottery found on site and is not in itself thought to be diagnostic. Almost all of the cultic material came from Phase 5 deposits which may consist of dumped rubbish from the nearby settlement. Face pots in Britain were largely associated with the military (Braithwaite 2007) and kilns at both Norton and nearby Crambeck supplied the demand for these, including, and especially in this case, from the fort at Malton. Other users of such vessels, especially the Smith god and closely linked Wheel god pots, seem likely to have included the metalworkers from the workshops in Norton.

Whether the concentration of Crambeck pottery finds featuring cultic symbols on this site was due to the proximity of that industry or whether it was driving the demand for such comparatively niche objects itself is an interesting point. Although only represented by a

Face pots and Smith/Wheel god pots were produced at both the Norton and Crambeck kilns, the former being established chiefly to supply the fort at Malton whereas 111

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 135: Smith god pot sherds in Crambeck grey ware with applied tools and other motifs (Top row left to right 70-72, Bottom row left to right 73-76).

single possible face vase sherd in this assemblage, the Norton pottery industry did also make Smith god pots, fragments of which have been found in some quantity at Norton and Malton over the years (e.g. Mitchelson 1964: fig. 13.145) however the Norton industry appears to have started earlier than the Crambeck one and potters from Norton may have been involved in the early stages of its creation (Wood 2016; Swan 1984, though later refuted by Evans in 1989 and Swan in 2002).

to the inhabitants of Norton, presumably a settlement that retained strong connections with the military (Braithwaite 2009: 39-40), until the very end of the Roman period suggesting a pervasive ritual “vernacular” or “lived religion” (Goldberg 2009; Rüpke 2019). It appears likely that the smith deity and associated iconography that appears on vessels from Norton may have been variously equated and ‘hybridised’ with various layers of named gods during the Late Roman period, but appears to have remained part of the repertoire of the potters until the end of the Roman period.

Evans (1989) makes the case that the later and major transformation of the Crambeck industry may have been due to a military contract. This could presumably have also involved the manufacture of cultic vessels to fulfil the demand. Hayes (1988: fig. 51.1) illustrated a calcitegritted Huntcliff jar with a crude applied set of tongs or ‘trident’ from Eastfields, Norton presumably dating to the end of the 4th century AD or perhaps later. The Crambeck Parchment ware vessels from this site with face pot and smith iconography would also suggest that this practice continued into the late 4th century AD. The specialist ‘cultic’ grey ware vessels appear to have been introduced into Britain with the military in the 2nd century AD and these vessels continued to be important

70 CRGR: Fragments from a Smith god pot in Crambeck grey ware featuring an applied hammer and a further hammer or other motif, Trial trench 5, Roman subsoil/dump deposit 502, D2E (Figures 128, 135.1) 71 CRGR: Body sherd in Crambeck grey ware featuring an applied hammer or wheel hub. Area 2, subsoil/Roman dump deposit 202 (GS05), Phase 5, D62 (Figure 135.2) 72 CRGR: Body sherd in Crambeck grey ware featuring an applied tool or other motif. Area 2, 112

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Rubbish deposit 216 (GS06), Phase 5, D83 (Figure 135.3)

78 CRGR: Body sherd of Crambeck grey ware featuring a stamped 8-spoke wheel and part of another. Area 2, subsoil/Roman dump deposit 202 (GS15), Phase 5, D134 (Figure 136.2)

73 CRGR: Body sherd in Crambeck grey ware featuring an applied hammer or wheel spoke and rim fragment. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS13), Phase 5, D93 (Figure 135.4)

79 CRGR: A body sherd from a large jar in Crambeck grey ware featuring an applied wheel. Area 8, Unstratified, D145 (Figure 136.3)

CRGR: Body sherd in Crambeck grey ware featuring an applied hammer or wheel spoke and rim fragment, Unstratified, D106 (Figure 135.5)

80 CRGR: A possibly burnt fragment of Crambeck grey ware featuring the edge of an applied wheel. Area 4, Unstratified, D142 (Figure 136.4)

75 CRGR: Joining body sherds in Crambeck grey ware featuring an applied pair of tongs from a Smith god pot. Area 2, subsoil/Roman dump deposit 202 (GS09), Phase 5, D220 (Figures 128, 135.6)

81 CRGR: Two non-joining sherds of Crambeck grey ware featuring a partial applied wheel hub and spokes, or a cross, and a partial wheel spoke and rim. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D50 (Figures 128, 136.5)

74

76 CRGR: Body sherd of Crambeck grey ware featuring a small fragment of applied tool or motif. Area 7, Roman rubbish layer 715, Phase 5 D148 (Figure 135.7)

Vessel with applied snake (No. 82) 82 CRGR: Although the head is missing from the applied snake on this vessel in Crambeck grey ware it is possible that it was originally a ramheaded serpent which, as a Gaulish deity, was rare in Britain, but its representation has been found

77 CRGR: Body sherd of Crambeck grey ware featuring an applied wheel, Unstratified, D105 (Figure 136.1)

Figure 136: Applied wheels from Smith god and/or wheel god pots (Top row left to right 77-79; Bottom row left to right 80-81)

113

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 137: Crambeck grey ware sherd with applied snake (No. 82)

in association with Taranis/Tanarus (AldhouseGreen 1979). Alternatively it may be a Mithraic symbol (a Mithraeum is known from York), or from a snake pot connected with the cult of Sabazius (Braithwaite 2007: appendix VI). Whatever the source it appears to be a rare representation and no other published examples have been found in this fabric. Area 2, subsoil/Roman dump deposit 202 (GS15), Phase 5, D132 (Figure 137)

83 CRGR: A sherd of Crambeck grey ware featuring stamped roundels in the Romano-Saxon style. Area 2, Unstratified, D70 (Figure 138.2) 84 CRGR: A sherd of Crambeck grey ware featuring an applied ‘Sunray’ motif or possibly a ‘Smith god’ hand. Area 2, subsoil/Roman dump deposit 202 (GS03), Phase 5, D90 (Figure 138.3) 85 CRGR: A body sherd of Crambeck grey ware featuring part of an applied figure or animal. Area 2, subsoil/Roman dump deposit 202 (GS08), Phase 5, D61 (Figure 138.4)

Other decorated fragments The identifications of the objects depicted on these sherds of Crambeck grey ware are somewhat open to question. The first, (No. 63, above) may be part of the face plaque from a face-necked flagon but could also perhaps be part of a representation of a figure from a Smith god pot. Norton grey ware vessel No. 48 and Crambeck grey ware vessel No. 83 both have stamped roundels, which are commonly seen on the products of various other Late Roman kilns as well as Norton and Crambeck. These stamps may also have been an attempt to depict wheel or sun imagery but were also commonly used as part of the decoration of face pots (e.g. above Nos 64, 65 and 68; Corder 1928: fig. 20). The identification of what was depicted on illustrated vessels Nos 84-86 was uncertain but they appear to match the representation of limbs or perhaps tools as seen on Smith god pots, a similar example has been published from Norton (e.g. Hayes and Whitley 1950: pl. VIb; Bidwell and Croom 2012: pl. 17).

86 CRGR: A body sherd of Crambeck grey ware featuring a partial applied strip motif or decoration. Area 2, Unstratified, D71 (Figure 138.5) The Other Reduced Wares I.M. Rowlandson CRGR?: Possible Crambeck products/Early Crambeck ware (Monaghan 1997: B11?) Eighty-nine sherds (2.721 kg, 3.29 RE) from this site were broadly similar to the typical Crambeck grey ware but with slightly coarser sandier fabric often fired to a pinker core as noted for some of the GW3 fabric from the Malton vicus assemblage (Bidwell and Croom 1997: 70) and perhaps as fabric B11 at York (Monaghan 1997: 903-904). The eighty-nine sherds are included with the CRGR Crambeck ware within the analysis below. 114

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 138: Various stamped and applied partial motifs. (Top row left to right 48 and 83. Bottom row left to right 84-86). Vessel 48 is in a Norton grey ware fabric NORGW1, the remainder in Crambeck grey ware CRGR

The forms recorded were broadly the same as those recorded for the CRGR group with nearly all of the rim fragments from conical flanged bowls, one vessel from Phase 5 had burnished wavy line decoration indicative of a late 4th-century AD product. Other forms included sherds from two Crambeck Type 5 bowls, a beaker with a funnel rim, plain rimmed dishes and a lug-handled jar. This fabric was found in small quantities amongst the Phase 1 ground level layers then mostly from Phases 3 and 5. The almost complete conical flanged bowl (No. 87) was retrieved from a Phase 3 pit and, as a primary deposit must suggest this feature was backfilled after AD 280 and probably in the 4th century AD.

description (Bidwell and Croom 1997: nos 160, 306 and 399). A plain rimmed dish and an unusual necked jar (No. 88) were the only other vessels attributed to this fabric. This coarse fabric might represent a fairly local variant on the standard Late Roman gritty ware jar that was produced at a number of sites in the north of England at the end of the Roman period (e.g. Bidwell and Croom 2010: 31, fig. 4.4.25b). 88 CRGRC: A necked jar in Crambeck coarse grey ware, an unusual form of uncertain date perhaps erroneously categorised in this group. Area 2, subsoil/Roman dump deposit 202 (GS03), Phase 5, D73 (Figure 128)

87 CRGR?: A bowl with a bead and flange rim possibly in a Crambeck grey ware, burnt and spalled. Area 2, pit 275, fill 278, Phase 3, D121 (Figure 128)

89 CRGRC: A jar with a double lid-seated rim with a string-cut base, possibly a Crambeck product? Area 2, rubble deposit 251 (GS16), Phase 3, D66 (Figure 128)

CRGRC: Coarse ?Crambeck grey ware Whilst recording the assemblage it was noticeable that there were a number of jars with light fired cores, similar to Crambeck ware, but with gritty fabrics. Twenty-six sherds (0.564 kg, 5.64 RE) were recorded using the code CRGRC. The forms identified included sherds from at least three lid-seated jars, perhaps similar to the ‘double lid-seated jar’ form (No. 89) produced at the end of the Roman period at Lincoln (Darling and Precious 2014, JDLS). This form appears similar to examples illustrated from the Malton vicus that have a broadly similar fabric

ROXGR?: Roxby grey ware (Rowlandson and Fiske forthcoming; Rigby and Stead 1976) Grey ware pottery was produced at Roxby, North Lincolnshire in the Antonine period and the products of the kiln typically have ‘soft brown flecks’ of argillaceous material within the fabric (Rigby and Stead 1976; Precious et al. 2011). An earlier phase of production at the nearby Dragonby is known and the sherds from this site have a similar if somewhat less distinct fabric. Examples of 115

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire vessels in similar fabrics are known from sites on the north bank of the Humber (Precious et al. 2011; Rowlandson and Fiske 2019b) and a single vessel was noted from the Norton Community Primary School site (Bidwell and Croom 2012). Further examples of vessels in similar forms, and fabrics have been noted at Shiptonthorpe and Rudston (Evans 2006; Rigby 1980). It appears likely that this sherd probably reached Norton by the Antonine period, but the sherd was found within a Phase 6 deposit.

number of the signal stations (cf. Monaghan 2000). All of the sherds were retrieved from Phase 5 and all of the recognisable forms have been illustrated (Nos 9294). The low level of sherds of this fabric, restricted to the final Roman phase, might suggest that activity on this site had decreased by the early 5th century AD. It was also notable that none of the other suite of Late Roman/sub-Roman fabrics characterised by Monaghan at York (viewed by the author in the York type series) or any Anglian/Anglo-Saxon pottery was recognised by the author during the recording process.

HOSM1 Home-on-Spalding Moor grey ware (Tomber and Dore 1998: HSM RE)

92 YB18: A handmade everted rim jar in Late Roman burnished fabric from York with burnished line decoration. Area 3, rubbish deposit 303, Phase 5, D124

Twenty-eight sherds were identified as the hard fired HOSM1 fabric variant seen amongst the products of the Throlam kiln site (0.732 kg, 0.76 RE). A single sherd was noted from each of Phases 1, 3 and 4, with the majority of sherds received from Phase 5. Forms included a conical flanged bowl, a necked jar or flagon (No. 91) along with fragments from other necked jars with burnished decoration. A single fragment from a body sherd cut down to form a spindle whorl was retrieved from Phase 7 (No. 89). Whilst further East Yorkshire products may lurk in the miscellaneous grey sandy wares category good examples of the high fired later fabric were not recognised in any great number and it would appear that local sources provided the bulk of the pottery to this site in the 3rd to 4th century AD (Evans 1988).

93 YB18: A handmade bowl with a plain rim in Late Roman burnished fabric from York. Area 3, rubbish deposit 303, Phase 5, D125 94 YB18: A body sherd in Late Roman burnished fabric from York featuring applied clay pellet dots. The decoration on this vessel was unusual for a vessel in the YB18 fabric but, as applied pellets are known on face pots it may represent the edge of a further figuratively decorated pot (e.g. Braithwaite in Bidwell and Croom 1997: fig. 40.7). An example of a vessel in a similar fabric with applied Smith god iconography have been recorded by Monaghan from the Carr Naze signal station at Filey (2000: no. 29). Area 2, rubbish deposit 216 (GS06), Phase 5, D84 (Figure 140)

90 HOSM1: A body sherd of Holme-on-SpaldingMoor grey ware trimmed to form a spindle whorl, broken in half, RF 449. Area 8, ground or yard surface 803, Phase 7, D127 (Figure 139) 91 HOSM1: A narrow-necked jar in Holme-onSpalding-Moor grey ware, Unstratified, D104

Figure 140: Burnished sherd with applied pellet decoration (No. 94) Figure 139: Holme-on-SpaldingMoor grey ware spindle whorl (No. 90)

GREY – Miscellaneous sandy grey wares Four hundred and seventy-seven sherds were attributed to a miscellaneous sandy grey ware fabric. This group probably includes a range of material that could not be attributed with certainty. Examples considered to be possibly from further afield were rare and have been noted in the archive.

YB18: Late handmade burnished ware (Monaghan 1997: B18) Twelve sherds (0.295 kg, 0.54 RE) of Late Roman burnished ware were recorded in the B18 fabric characterised by Monaghan at York and found at a 116

I.M. Rowlandson and H.G. Fiske: The Roman pottery

The vessels illustrated included a folded beaker in a fairly fine fabric that did not match the typical Norton sandy gritted NORGW1 fabric (No. 95) and a bead and flanged rimmed bowl with burnished wavy line decoration that appeared similar to Crambeck 1b Types but with a wheel made sandy fabric fired black throughout with highly burnished surfaces (No. 96).

century AD. As discussed above (CRGRC fabric, Bidwell and Croom 2010) these types were manufactured at a number of sites in the north during the Late Roman period including into the late 4th century AD. 97 GREYC: A Dales ware JD2 jar form in coarse grey ware. Area 2, subsoil/Roman dump deposit 202 (GS09), Phase 5, D24

Amongst the grey group only two vessels from Phase 4 would appear to be a good match for the dark surfaced grey ware GW2 recorded from Malton vicus, a dish (Bidwell and Croom 1997: no. 25) and a carinated bowl. Although a few other vessels with a broadly similar dark surfaced fabric have been noted in the archive there was little that closely matched the fabric description or the range of early forms that were illustrated from the Malton vicus publication. The relatively low level of pottery of this type from the Brooklyn House site, found as a residual element in Phase 4 was probably due to the predominantly late date of the activity and may represent material dumped on the site.

98 GREYC: A lid-seated jar in coarse grey ware. Area 2, subsoil/Roman dump deposit 202 (GS18), Phase 5, D109 99 GREYC: A double lid-seated jar in coarse grey ware. Area 2, subsoil/Roman dump deposit 202 (GS08), Phase 5, D57 100 GREYC: A double lid-seated jar in coarse grey ware. Area 2, subsoil/Roman dump deposit 202 (GS09), Phase 5, D23 101 GREYC: A double lid-seated jar in coarse grey ware with external carbon deposits. Area 2, subsoil/ Roman dump deposit 202 (GS09), Phase 5, D25

Other examples included sherds from a bowl with a bifurcated rim (Gillam 1970: Type 301) known to have been produced at numerous sites in Lincolnshire and possibly also Brough-on-Humber and a Late Roman wide-mouthed bowl with an under-cut rim with a fabric similar to Swanpool examples (Darling and Precious 2014: no. 1229). The majority of the pottery in this group could be dated to the Late Roman period with most common form was the straight-sided bead and flanged/conical type bowl (BFB), along with examples of narrow necked jars and a dish with a plain rim. These Late Roman vessels may have been the sandy grey ware products of a number of kiln sites including Norton, Holme-on-Spalding Moor or Stamford Bridge, East Yorkshire.

NGGW: North Gaulish grey ware (Tomber and Dore 1998: NOG RE) A single sherd of North Gaulish grey ware from a vessel with rouletted decoration was recorded from Phase 5. GROG: Miscellaneous grog-gritted wares. A single grey ware sherd with rare grog inclusions was retrieved during the excavation. The relative proportions of reduced wares In an attempt to test both the site stratigraphic sequence and the established ceramic chronology some brief analysis was undertaken. The relative volume of reduced wares from the stratified deposits and the mean sherd weight by phase were calculated (Tables 10 and 11) as a method to test if the relative proportion of the main types of grey ware changed through time. Although the results were not straight forward some basic chronological themes appear to be visible.

95 GREY: A folded beaker with tall funnel rim in fine grey ware possibly from Norton. Area 2, subsoil/ Roman dump deposit 202 (GS03), Phase 5, D89 96 GREY?: A bowl with a bead and flange rim, wheel made with highly burnished surfaces and burnished wavy line decoration internally. Area 4, rubble deposit 410, Phase 4, D79

A simple hypothesis would be that the Norton products, predominantly dated to the 2nd to 3rd century AD with some evidence of production into the 4th century AD, ought to decrease in frequency in the later phases and be largely replaced by fresh pottery from the Crambeck industry, a major producer of grey wares supplying the north in the late 3rd to 4th century AD. The picture from the data is complicated by the presence of intrusive material of later date in some of the Phase 1 deposits perhaps explaining the higher than expected proportion of Crambeck grey wares from Phase 1 with also a correspondingly lower mean sherd weight. Phase 2 appears to show a high proportion of Norton products with less Crambeck ware, but it ought to be noted that a total of only 294 sherds were retrieved from this phase. Phase 3, considered to date to the 4th century AD, shows relatively similar quantities of both wares

GREYC: Miscellaneous coarse grey wares Forty-nine sherds (0.953 kg, 2.07 RE) were recorded as coarse grey ‘heavily gritted’ wares. This fabric code has been used for the coarse quartz-gritted Late Roman sherds that could not be attributed to any of the existing fabric groups (this includes sherds recorded as GRQZ during the evaluation). With the exception of a few sherds found within the Phase 1 Roman ground levels the majority of the sherds in this fabric group were retrieved from Phases 4 and 5. The forms included double lid-seated jars, Huntcliff jars, Dales type and other lid-seated jar types. The vessels from this site in this fabric group date to the 4th to early 5th 117

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Phase 1 Phase 2 Phase 3 Phase 4 Phase 5

Norton % % of % of Sherd weight 25.69 18.03 35.03 31.06 26.02 22.10 18.40 13.74 20.45 16.59

Crambeck % Other reduced % % of % of % of % of % of % of RE Sherd weight RE Sherd weight % of RE 25.51 23.68 20.81 30.17 2.01 1.47 2.09 37.89 13.95 10.06 10.35 3.74 2.50 2.93 28.46 13.58 17.93 26.51 2.60 2.61 3.67 21.38 18.28 15.12 24.67 3.48 2.25 6.04 13.15 23.81 22.29 18.42 2.83 2.55 2.01

All reduced % % of % of % of Sherd weight RE 51.37 40.30 57.77 52.72 43.63 51.17 42.20 42.64 58.63 40.16 31.12 52.09 47.09 41.42 33.58

Table 10: Relative proportion of Norton, Crambeck grey wares and other reduced wares as proportion of whole assemblage by phase

Phase 1 Phase 2 Phase 3 Phase 4 Phase 5

Norton reduced Crambeck reduced Other reduced All wares 5.89

4.71

5.66

4.14

6.82

4.39

5.78

5.80

5.53

6.80

6.04

5.46

6.45

5.59

7.33 6.97 5.81

4.90 4.52 5.24

Table 11: Mean sherd weight (g) of reduced wares and calcite-gritted wares by phase

present. By Phase 4 a broadly similar proportion of wares from the two production sources were recorded, presumably by the late 4th century AD much of the Norton reduced ware was residual. By Phase 4 the quantity of Norton grey ware appears to be lower than Crambeck wares and a similar picture is evident in the Phase 5 assemblage. By Phase 5, the late 4th century AD or later, the Norton material that was present had a slightly higher average sherd weight than the Crambeck wares that were present, but this may relate to the presence of chronologically earlier material dumped from elsewhere altering the composition of this group. A similar pattern has been demonstrated by the analysis of contemporary assemblages from the Roman roadside settlement of Navenby, Lincolnshire (Rowlandson et al. 2011; 2015). However, looking at the data by phase it would appear that the quantities of Crambeck grey ware increased in Phases 4 and 5 although the average sherd weight of the Norton and Crambeck wares present were broadly similar from phase to phase. Considering the bias amongst the Norton grey wares towards forms that only became common towards the end of the 3rd century AD it is perhaps unsurprising that there remained a high proportion of Norton grey wares in the Phase 4 and 5 assemblages.

3). The calcite-gritted wares are discussed further below. NTV2 (Bidwell and Croom 1997: NTV2) A handmade sherd gritted with non-soluble rock was retrieved from Trial Trench 5, layer 502-1. Although a number of rock-gritted fabrics were recorded from excavations at Newbridge near Pickering (Rowlandson 2012) and small quantities of such wares were recorded from excavations at the Malton vicus (Bidwell and Croom 1997) these fabrics appear to have been most common in Iron Age and Early Roman deposits. At West Heslerton, were the Roman pottery predominantly dated to the 4th century AD, when the handmade Anglian/AngloSaxon wares were removed from the assemblage, there were few rock-gritted sherds (Darling and Precious forthcoming). The Brooklyn House assemblage had little pottery present that could be securely dated to the Early Roman period so the limited quantity of the NTV2 fabric is unsurprising. IAGR – Transitional grog and fossil shell-gritted wares (Darling and Precious 2014) Six sherds from a maximum of four vessels were recorded that would fit this category. The vessels were gritted with grog, fossil shell and small quantities of quartz sand. One sherd was retrieved from Phase 1 while the remaining sherds were retrieved from Phase 5, including the rim from a typical jar with a wedge-shaped rim (Rigby and Stead 1976: fig. 74.11). These wares, broadly similar to Malcolm Todd’s ‘Trent Valley ware’ category, were commonly produced in Lincolnshire and other parts of the East Midlands in a tradition that developed from local Iron Age precursors (Darling and Precious 2014: 104). Vessels of this type

It was noticeable that the relative proportions of unsourced or imported reduced wares, including Dorset BB 1, stayed broadly between 2-4% of the assemblage and only appears to have made up only a small proportion of the reduced wares used by the inhabitants. Transitional grog-gritted wares and handmade rockgritted sherds Earlier prehistoric material that was initially noted during recording has been discussed by Vyner (Chapter 118

I.M. Rowlandson and H.G. Fiske: The Roman pottery

were produced in the 1st century AD and were in use in Lincoln at least until the middle of the 2nd century AD. Although these wares do not appear to have reached York in significant quantities examples are known from the area on the north bank of the River Humber. Fossiliferous deposits outcropping on both the north and south banks of the Humber may have been utilised to produce such wares (Didsbury and Vince 2011: 192196). Given that these sherds were all found stratified with Roman pottery it appears likely that they arrived at Norton sometime from the late 1st century AD onward and offer scant support to the suggestion of pre-conquest settlement on the site.

date (Rowlandson 2012; Rigby 2004). Sherds of Earlier Prehistoric pottery from Brooklyn House are discussed by Vyner (Chapter 3). The pottery from this site was mostly of later 2nd- to early 5th-century date with little handmade pottery that might be earlier. Although York received very little handmade pottery other forts such as Malton appear to have utilised small quantities of this type of pottery even in the 1st and 2nd centuries AD. These assemblages offer some potential assistance as they were commonly found with a broad range of Roman wheel made pottery and other finds. The published assemblages from Malton, Lease Rigg, Roecliffe, Hayton and, following further evidence published by Wilson, the site excavated by Brewster at Staxton (Bidwell and Croom 1997; Vince and Steane 2009; Bishop 1997; 2005; Johnson 1978; Brewster 1957; Wilson 2017) all offer good assemblages which were found in association with datable Roman pottery. Indeed, it was this form of seriation that brought success in dating Huntcliff and other chronological indicators of the pottery used at the end of the Roman period (Hornsby and Stanton 1912; Hull 1932; Gillam 1970; Tyers 1996: 1-39). Although the handmade material can be more difficult to characterise than wheel made Roman pottery it is feasible to use some of these sites associated with Roman wheel made pottery to offer a marker for sites in the 1st century AD or later, alongside the increasing use and refinement of scientific dating techniques it could be possible to move forward with refining and ‘repairing’ chronologies for calcite-gritted wares in this region in the Iron Age and Roman periods (Hamilton et al. 2015: 8; Jay et al. 2012). Recently published work at Stanwick (Willis 2016a) and the assemblages from Scotch Corner will also no doubt produce further insights.

Calcareous-gritted later Roman wares CG Calcite-gritted wares. Mostly chalk-gritted with sparry mineral calcite crystals present (as Tomber and Dore 1998: HUN CG). The use of this fabric code follows Bidwell and Croom (1997) and includes handmade local wares often characterised as Knapton ware and the hand-built wheel-finished Huntcliff and ‘proto-Huntcliff ’ type wares; with small fragments from the lower walls of vessels the distinction between these two traditions was not always clear or replicable. The material recorded in this category nearly all contained fragments of chalk and sparry mineral calcite suggesting a fairly local production source. A small proportion of this group was vesicular due to soil conditions. Evidence of pre-conquest activity on the site was not clearly demonstrable but could not be ruled out as some of the forms present had affinities with jars with simple everted rims found on sites dated to the late Iron Age in this region. The majority of the vessels present could be paralleled with those illustrated by Bidwell and Croom (1997) from Malton vicus and Beadlam villa (Evans 1996) where there was no suggestion of activity prior to the Roman conquest. The vast majority of the calcite-gritted pottery exhibited consistent surface colours and the ‘S-shaped rim/proto-Huntcliff ’ and later Huntcliff types had signs of wheel finished rims.

In light of the nature of this material, and faced with over 8,000 calcite-gritted sherds to characterise the forms have been divided as discussed below with the relative occurrence of the jars listed by phase (Table 12). It should be noted that selected aspects of the phase groups have been presented in the table with the total number of RE listed for each group – the complete vessel from bustum grave 457 has been removed from the analysis due to the bias it would add to the small assemblage from Phase 2. The Phase 2 assemblage was a small sample and therefore should perhaps be ignored for the purposes of looking at trends in the calcitegritted jars.

The merits and demerits of dating Iron Age and coarsegritted Roman pottery in Eastern Yorkshire by form have been extensively discussed elsewhere with varying levels of optimism (Challis and Harding 1975; Evans 1995; Mackey 2003; Didsbury 2004; Didsbury and Vince 2011: 196; Rigby 2004; Rowlandson 2012; Cumberpatch 2016). One of the main problems facing the researcher is the strong conservatism amongst both potter and consumer on rural sites from the first Millennium BC into the Roman period. There is a change from the more angular forms common in earlier Iron Age assemblages to the later Iron Age repertoire, but this continues with few changes into the Early Roman period. Identifying the Roman transition is difficult on a site with very low levels of wheel made pottery. None of the material from this site could be considered to be of earlier Iron Age

Channel rimmed types – JCH A handmade jar with a slightly channelled rim was retrieved from Phase 1, context 231 (GS15) (stratified with pottery dating to the late 4th century AD). This vessel was similar to an example from Rudston stratified from an early layer and may have been a type that was in circulation in the 1st to 2nd century AD (Rigby 1980: fig. 27.1). A further sherd from a similar residual vessel was also recorded from Phase 5. 119

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Phase 1 layers Phase 2 layers Form code JNK JCH JEV JEVS JEVT JEVEB JEVC JCUR JCURS JHUN JTR JLS JNN JEVFT JL or JS JM

Phase 3- all Phase 4- all Phase 5- all pottery pottery pottery All Phases

Description RE % (8.28) RE % (1.07) RE % (4.31) RE % (5.94) RE % (57.26) RE % (94.24) Necked jar 1.33% 0.00% 6.03% 0.34% 7.04% 5.95% Channel rim 0.00% 0.00% 0.00% 0.00% 0.19% 0.13% Everted rim jar 5.56% 5.61% 17.63% 1.01% 12.09% 11.17% Stubby everted 0.00% 0.00% 0.00% 0.00% 0.00% 0.05% Tall everted rim jar 3.26% 0.00% 8.58% 4.55% 2.08% 2.66% everted - external bevel 26.45% 22.43% 21.11% 12.96% 14.62% 18.06% Out-curved- handmade 3.62% 0.00% 3.94% 3.20% 0.79% 1.59% Curved rim- handmade 19.69% 46.73% 8.35% 34.01% 31.85% 29.07% Curved- ‘S-type’ 19.81% 1.87% 17.17% 16.84% 11.37% 13.74% Huntcliff lid-seated 11.96% 23.36% 9.51% 12.12% 14.32% 14.92% Triangular rim 0.97% 0.00% 0.00% 0.00% 0.12% 0.17% Lid-seated misc. 2.29% 0.00% 0.00% 0.00% 0.38% 0.48% Narrow necked 1.81% 0.00% 0.00% 0.00% 0.00% 0.17% Everted-flat top 0.00% 0.00% 2.78% 0.00% 0.00% 0.14% Large or Storage 3.26% 0.00% 4.87% 3.20% 5.15% 0.89% Miniature 0.00% 0.00% 0.00% 11.78% 0.00% 0.81% TOTAL 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% Table 12: Calcite-gritted jar forms by % of Rim equivalents by phase

Necked jars – JNK These types represent handmade necked jars perhaps in a Late Iron Age to Early Roman style (cf. Rigby 2004: fig. 7 ‘necked jar’). The illustrated vessel from this category (No. 102) would fit with an example illustrated from a Trajanic group from the Malton vicus (Bidwell and Croom 1997: no. 13). Examples of this vessel type occurred in Phase 1 and sporadically throughout the later phases.

Tall everted rim jar – JEVT Other possibly early material present included a small number of handmade jars with tall everted rims (JEVT) that were similar to examples illustrated by Evans from the Hawling Road site (1999, G28-J04, 1st-century context), or an illustrated example from Staxton (Brewster 1957: fig. 12.4). These vessels might perhaps be seen as occurring from the 1st century AD but similar vessels have also been noted from later deposits at Hawling Road (Evans 1999, G03-J0,1 early 2nd century AD) and perhaps represent vessels similar to the possibly later JEVEB ‘Knapton types’ with a thin rounded rim rather than a faceted bevelled rim type. Small numbers of this form were recorded from Phases 1, 3 and 4 and their absence from Phase 2 can perhaps be explained by the small sample size (only 1.07 RE). As such with a total of 2.29 RE attributed to this rim class it might be suggested that their production continued until the 2nd century AD, but it does not appear that they formed a major element of the jars in use.

Everted rimmed jars – JEV Nos 103-105 Everted rimmed jars were a broad category of handmade jars and three vessels have been shown to illustrate the range. There were a total of 6.92 rim equivalents (RE) attributed to this category. Examples of jars with everted rims occur commonly in late Iron Age assemblages (Rowlandson 2012). Examples of jars with higher shoulders appear common in Early Roman groups (No. 103; cf. Bidwell and Croom 1997: no. 7) but it can be difficult to categorise these vessels to types more closely when only rim scraps survive. It is likely that this form was most commonly produced in the Iron Age through to perhaps 3rd century AD. The form appears to have occurred most often in Phases 5, unstratified groups and Phase 3. The stratified occurrences of this broad category cannot help to refine the date of these vessels and a proportion of the 6.92 RE from Phase 5 was probably residual earlier material.

Externally bevelled ‘Knapton’ type jars – JEVEB Nos 106-109 This form of handmade jar with an externally bevelled everted rim was perhaps a development from the JEVT. The majority of the examples from this assemblage appear to have been handmade with a knife trimmed external bevel to form the sharper faceted rim shape. These jars have often been called ‘Knapton Jars’ following Corder and Kirk’s publication of material from Langton villa and Knapton where they considered them to date to the 3rd century AD (1932: fig. 30.1-9). At

Stubby everted rimmed jars – JEVS A single small sherd from a handmade jar with a stubby everted rim (Rowlandson 2012: no. 33) was recorded from an unstratified context. An Early Roman date might be possible for this small rim fragment. 120

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Malton, Bidwell and Croom have similar jars illustrated from groups dated to the mid-2nd century AD+ (1997: nos 63-64) but predominantly from the 3rd century AD or later (1997: nos 218, 222 and 395). The stratified examples from Brooklyn House appear to have been most common in Phases 1-3, broadly fitting with the established trend for the dating of these vessels, and they are present in lower percentages in the later 4thcentury AD deposits were it is likely that many of the examples were residual (Table 12). It was notable that this type of jar with 15.56 RE was almost three times as common as the everted rim type. The complete jar from the Phase 2 bustum burial 457 was removed from this analysis otherwise the JEVEB category would have made up over 60% of the calcite-gritted jars from Phase 2 by rim equivalent (RE).

Black Burnished ware prototypes in the later 2nd to 3rd centuries AD which developed a more pronounced curved rim (cf. Gillam 1976: fig. 1-2). The curved rim jar category includes all of the handmade jars with curved over rims lacking the elements of wheel-finishing and the more pronounced ‘S-type’ rim of the JCURS form. Vessels 115 and 116 may perhaps be earlier variants of the JCUR group. This form group was common (24.01 RE) and probably represents material mostly produced in the 3rd century AD and perhaps into the 4th century AD. This jar type was common amongst the Phase 1, 2, 4 and 5 deposits. Sherds from ten wide mouthed bowl type handmade BCUR vessels were retrieved from Phase 5 (1.02 RE). The JEVC group represents handmade jars perhaps more closely matched to Black Burnished ware prototypes with thinner walls (No. 120, cf. Evans 1996: fig. 46. G1.16). Limited numbers of these vessels were recorded (1.37 RE from a maximum of ten vessels) and these vessels represent a subset of the JCUR category.

Lug-handled jars – JLH Lug-handled jars (0.84 RE, maximum of 11 vessels) were only present in small numbers, mostly from Phase 5 and have been removed from the analysis as they often lacked rims. Only two vessels had associated rims both from Phase 5: illustrated vessel No. 110 and a significant part of a further smashed vessel from dumped deposit 3110 that was not suitable for illustration but could be judged to be similar to an example from Rudston (Rigby 1980: fig. 30.28).

Curved ‘S-rim’ jars and large bowls JCURS and BCURS Nos 121-126 This type represents types made in the 4th century AD with elements of a wheel-finished rim and handmade body. Most examples had the strongly S-shaped rim evident on illustrated vessels 121, 124 and 125. The jars (total 11.84 RE) in this rim category were perhaps intrusive within the Phase 1 deposits but common from Phase 3-4, declining in Phase 5 (Table 13, highlighted). This rim type appears to have been utilised for the production of more open wide-mouthed bowl forms (Nos 124-125) and six examples were recognised but identification was only possible were larger fragments with rim and vessel girth survived.

Everted rim with a flattened top JEVFT A single vessel with a flattened rim (No. 111) was retrieved from Phase 3 (0.12 RE). this type should be considered a variant on the JEVEB type and was probably of a similar date Miscellaneous large/ storage jars – JL/JS A small number of large jars (2.65 RE) that could not be closely attributed to one of the rim types were recorded in this group including large vessels Nos 112 and 114 from Phase 5. The majority of the vessels had simple rims and were most common amongst Phase 3 and 5 deposits.

Huntcliff jars and large bowls JHUN and BHUN Nos 127128 Only two examples of this ubiquitous form were illustrated one standard type (No. 127) and an unusual double grooved type (No. 128). This form was common on the Brooklyn House site and a total of 12.85 RE were recorded from an upper maximum of 135 vessels on the basis of number of rims. A further example from a wide-mouthed bowl type vessel with a Huntcliff rim

Curved rimmed jars – JCUR/JEVC Nos 115-120 Slightly merging with the everted rim category, the curved rim jars perhaps follow the development of Calcite and shell-gritted

All reduced %

% of Sherd % of weight % of RE % of Sherd % of weight

% of RE

Phase 1

42.07

53.46

35.81

51.37

40.3

57.77

Phase 2

32.65

78.55

45.31

52.72

43.63

51.17

Phase 3

37.4

40.7

26.74

42.2

42.64

58.63

Phase 4

48.35

56.92

30.62

40.16

31.12

52.09

Phase 5

43.67

47.68

24.63

47.09

41.42

33.58

Table 13: Relative proportion of calcite-gritted wares and reduced wares as proportion of assemblage by phase

121

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire was also retrieved from Phase 5. The Huntcliff form is seen as one of the key indicators of activity after AD 370 and the chronology of this form and the S-rim type jar (JCURS) has been extensively discussed elsewhere (e.g. Bidwell and Croom 2010; Monaghan 1997; Swan 2002). Looking at Table 13 the nine vessels represented from Phase 1 (0.99 RE) and the two vessels from Phase 2 were all retrieved from the soil layers that probably contained intrusive material. The Phase 2 examples confuse the analysis amongst such a small assemblage and should be ignored. Phases 3-4, considered to represent 4th AD or later activity, show an increasing proportion of Huntcliff jars within the assemblage over time and further examples were recorded from the post-Roman and unstratified deposits. If this is taken into consideration it would appear that this pattern would follow the existing theories with Huntcliff type being more commonly found in the later phases.

(Bidwell and Croom 1997: no. 448). Individual examples were recorded from Phases 3 and 4 but the majority of stratified examples were retrieved from Phase 5. The function of the calcite-gritted jars Sherds from 160 vessels in the CG fabric showed signs of surviving external or internal carbonised cooking residues typically seen on vessels where fats have burnt onto the shoulders of jars whilst being used for cooking or rendering fat (Illustrated vessels 106, 116-120, 127). A further 37 vessels were noted where the carbonised deposits could not be shown to have been present over the breaks and may also have been from cooking. The majority of the vessels with these residues were jars but also a fragment from a lid and a small number of small bowls and dishes were noted. Recent research by Dunne and Evershed (2018; Rowlandson and Fiske 2019a) on an assemblage from Lincolnshire has shown that calcareous-gritted jars were favoured for such practices and it would appear that the calcite-gritted jars from this site were probably used in a similar fashion. Where these residues were recorded on grey wares (see above) they were almost exclusively from small bowls and dishes and seldom from jars. This would suggest that, whilst grey ware conical flanged bowls and plain rimmed dishes may have been used for cooking, it was calcite-gritted jars which were favoured. The practice of using coarsely gritted jars probably reflects their ability to withstand thermal shock when used for cooking. This characteristic sees the continued use of Iron Age tradition coarsely gritted pots throughout much of the Roman period. By the Late Roman period it would appear that these jars were in demand and that medium sized grey jars for cooking were not a key part of the repertoire of the Crambeck grey ware potters.

Triangular rimmed – JTR Examples of jars with triangular rims as (JTR) were rare and restricted to stratified examples from Phase 1 (as Rigby 1980: no. 137) and Phase 5 (two vessels). This was a minor type and a Roman date is presumed. Lid-seated variants – JLS Sherds from three lid-seated vessels that could not be closely grouped (0.22 RE): one from Phase 1 (No. 133) and three from Phase 5 were recorded in this category. Narrow necked – JNN A single small fragment from a narrow necked jar (0.15 RE) was retrieved from Phase 1. Miniature jar JM No. 136 A single vessel discussed further below in the catalogue. Smaller Bowls and dishes Nos 129-133 A range of small bowls and dishes were recorded (7.94 RE) These vessels were mostly variants on plain rimmed dishes and bowls with a few examples of lipped bowls or vessels mimicking the conical flanged bowl/ straight-sided bead and flanged bowl type (BFB). Although a small proportion were retrieved from the Phase 1 layer deposits (0.42 RE), Phase 2 and Phase 3 (0.25 RE), the majority of the stratified small bowls and dishes occurred in Phase 4 (0.49 RE) and amongst the large assemblage from Phase 5 (5.35 RE). This would fit the general pattern of calcite-gritted dishes and small bowls becoming more common in the 4th century AD. Illustrated examples are discussed further in the catalogue below.

Three hundred and sixty-three vessels in the CG fabric showed signs of internal white mineralised deposits typically associated with the residue caused by boiling up liquids in jars. This was almost exclusively from sherds from jars and typically on the basal zones of vessels. Examples of all of the jar categories (Nos 105, 106, 115, 118, 120 and 120) had these residues and a photograph of the thick encrustation on jar No. 120 is shown below. The pattern from the residues from this assemblage would suggest that, although a number of the grey ware small bowls and dishes may have been used for cooking foodstuffs, the calcite-gritted jars appear to have been favoured for heating food and liquids over a fire. Distribution by phase

Lids Nos 134-135

The main jar types considered to be chronologically significant would appear to fit with the phase dating that has been established. Bidwell and Croom in their study of Late Roman groups from the north (2010) and

Sherds from 19 lids were recognised mostly from vessels of the tall domed type with pierced walls that were probably dated to the 4th century AD or later 122

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Monaghan in his study of York (1997) have suggested that deposits dating to the late 4th century or later in the north ought to have a higher proportion of calcitegritted ware to reduced wares. The composition of the Phase 4 and Phase 5 groups (Tables 11, 12 and 14) poses a problem as, despite the presence of all of the key form and fabric indicators, calcite-gritted wares do not convincingly outnumber the reduced wares present in Phases 4 and 5. However, as most of the deposits from Phase 4 and 5 were layers including quantities of earlier pottery, particularly the local NORGW1 fabric, it is unsurprising that a comparison does not fit the established type assemblages due to the presence of high quantities of residual grey ware.

112 CG: A handmade jar with a cavetto rim in calcitegritted fabric. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS03), Phase 5, D27 (Figure 141) 113 CG: A handmade calcite-gritted lid-seated jar with a crudely applied coil to form the lid seat. Area 2, Roman ground level 247 (GS13), Phase 1, D75 (Figure 141) 114 CG: A large calcite-gritted ware, handmade with wheel-finished rim, large example. Area 2, subsoil/Roman dump deposit 202 (GS09), Phase 5, D15 (Figure 141) 115 CG: A handmade everted rim jar in calcite-gritted ware with white internal deposits and external carbon deposits. Area 4, robber cut 4049, fill 423, Phase 4, D81 (Figure 145)

102 CG: A handmade necked jar in calcite-gritted ware. Area 2, interior of Structure A, subsoil/ Roman dump deposit 206 (GS03), Phase 5, D28 (Figure 141)

116 CG: A handmade everted rim jar in calcite-gritted ware with external carbon deposits. Examples of vessels of this type may date to the later Iron Age to 2nd century AD (Evans 1999, G29-J08). Area 2, ditch 287, fill 291 GS13, Phase 1, D122 (Figure 145)

103 CG: An everted rim jar in calcite-gritted ware with carbon deposits. Area 2, subsoil/Roman dump deposit 202 (GS08), Phase 5, D60 (Figure 141) 104 CG: A everted rim jar with a low shoulder in calcite-gritted ware. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D42 (Figure 141)

117 CG: An everted rim jar in calcite-gritted ware with external carbon deposits. Area 2, subsoil/Roman dump deposit 202 (GS12), Phase 5, D112 (Figure 145)

105 CG: A handmade calcite-gritted jar with an everted, externally bevelled rim, white internal deposits and carbon external deposits. Area 2, Interior of Structure A, subsoil/Roman dump deposit 206 (GS03), Phase 5, D26 (Figure 141)

118 CG: A handmade calcite-gritted jar with a curved everted rim, carbon and white deposits internally. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS13), Phase 5, D92 (Figure 145)

106 CG: A calcite-gritted jar with an everted, externally bevelled rim, white internal deposits and carbon external deposits. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D4 (Figure 141)

119 CG: A calcite-gritted jar with a curved rim, external carbon deposits. Area 2, rubble deposit 230, Phase 4, D63 (Figure 145)

107 CG: A calcite-gritted jar with an everted, externally bevelled rim. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D43 (Figure 141)

120 CG: A calcite-gritted jar with a curved rim, external carbon deposits and a thick white internal deposit. Area 2, subsoil/Roman dump deposit 202 (GS12), Phase 5, D113 (Figures 143, 145)

108 CG: A complete calcite-gritted everted rim jar with externally bevelled rim which contained cremated remains. Area 4, bustum cremation pit 475, fill 476, Phase 2, D146 (Figures 141, 142, 156)

121 CG: A calcite-gritted jar with a curved rim, carbon deposits. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D1 (Figure 145)

109 CG: A jar or bowl in calcite-gritted ware with an externally bevelled slightly grooved rim. Area 2, Roman ground level 247 (GS03), Phase 1, D74 (Figure 141)

122 CG: A proto-Huntcliff jar in oxidised calcitegritted ware. Area 2, rubbish deposit 216 (GS06), Phase 5, D82 (Figure 145) 123 CG: A calcite-gritted jar with a curved rim. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D41 (Figure 145)

110 CG: A calcite-gritted ware lug-handled jar, reduced firing, possibly a Norton product. Area 2, subsoil/Roman dump deposit 202 (GS03), Phase 5, D135 (Figure 141)

124 CG: A proto-Huntcliff large bowl in calcite-gritted ware. Area 4, rubble deposit 410, Phase 4, D80 (Figure 145)

111 CG: A handmade jar with an everted flat-topped rim in calcite-gritted ware. Area 9, beaten floor/ bedding layer 930, Phase 3, D68 (Figure 141) 123

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 141: Illustrated calcite-gritted wares

124

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 142: Calcite-gritted jar from bustum cremation pit 475 (No. 108)

Figure 143: Interior of calcite-gritted jar showing thick white deposit (No. 120)

Figure 144: Proto-Huntcliff jar showing method of construction. Left- external, Right- internal surface showing hand building and wheel-finished rim (No. 126)

125 CG: A handmade calcite-gritted bowl with a curved rim, burnished internally and externally. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D2 (Figure 145)

2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D3 (Figure 145) 128 CG: A calcite-gritted double-grooved Huntcliff jar variant. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D40 (Figure 145)

126 CG: A proto-Huntcliff jar in oxidised calcitegritted ware with scored diagonal line decoration. This vessel shows the distinction between the hand-built body below the girth of the vessel and the wheel-finished rim above on this type and the true Huntcliff jars. A faint groove is evident on the inside of the rim. Area 2, external yard or surface 246 (GS16), Phase 3, D123 (Figures 144, 145)

129 CG: A calcite-gritted ware bowl. Area 2, interior of Structure A, subsoil/Roman dump deposit 206 (GS03), Phase 5, D29 (Figure 146) 130 CG: A calcite-gritted ware ‘Dales ware’ dish with an internal bevel. Area 2, subsoil/Roman dump deposit 202 (GS09), Phase 5, D16 (Figure 146) 131 CG: A calcite-gritted ware plain-rimmed dish. Area 2, interior of Structure A, subsoil/Roman

127 CG: A calcite-gritted Huntcliff jar with external carbon deposits and white internal deposit. Area 125

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 145: Illustrated calcite-gritted jars

133 CG: A calcite-gritted ware bowl with a bead and flange rim, burnished internally and externally. A similar vessel is published from Shiptonthorpe (Evans 2006, G01.10). Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D5 (Figure 146)

dump deposit 206 (GS03), Phase 5, D30 (Figure 146) 132 CG: A calcite-gritted ware plain-rimmed bowl with acute lattice decoration. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D8 (Figure 146) 126

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 146: Calcite-gritted other forms

Figure 147: Calcite-gritted ware miniature jar (No. 136)

134 CG: A handmade calcite-gritted ware ‘coffee pot’ lid. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D6 (Figure 146)

137 CG: A calcite-gritted ware sherd cut down to form a spindle whorl, broken in half, RF 319. Area 2, rubble spread 214 (GS03), Phase 4, D128 (Figure 148)

135 CG: A handmade calcite-gritted ware ‘coffee pot’ lid. Area 2, subsoil/Roman dump deposit 202 (GS06), Phase 5, D7 (Figure 146) 136 CG: A miniature calcite-gritted ware jar with an externally bevelled everted rim, 54 mm high. This vessel appears to copy some of the contemporary handmade jars of the late 3rd to 4th century AD. Miniature vessels similar to this have been considered to represent votive or religious offerings by some authorities. Although this vessel may have been a child’s toy or intended to hold unguents or ink, a specialist ritual function may have been possible. RF 107. Area 2, rubble spread 215 (GS06), Phase 4, D130 (Figures 146, 147)

Figure 148: Calcite-gritted ware body sherd cut down to a spindle whorl (No. 137)

127

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire The calcite-gritted oven and baking plate fragments

deposited on the site until the late 4th century AD. The other Yorkshire examples would appear to suggest a later 4th century AD date for the baking plates to be in use. The specialist ovens suggest specific produce and cooking practices, probably including bread production, was popular in this part of Yorkshire in the 4th century AD (see discussions in Darling 2012 and Evans et al. 2018). If the fragments from the Brooklyn House site highlight anything they probably represent some form of centralised food provision for large numbers of people, perhaps seasonally (Evans et al. 2018). Examples of clay ovens from Lincoln and Chesterton, Cambs. (Evans et al. 2018: 60; Perrin 1999: fig. 75; Darling and Precious 2014) were found in areas likely to have been used for mass catering. The presence of these fragments may suggest there was a bakery or food stall near to the Brooklyn House site.

One hundred and sixty fragments from ceramic ovens or baking plates (7.099 kg) were recorded from the site. The number of ovens that these fragments were from was uncertain, but the range of feature sherds would suggest that fragments from more than one oven were represented. A further five fragments from ceramic baking plates (0.427 kg) were recorded. Whilst the fabric and thickness of these ceramic objects was fairly distinctive it is possible that smaller more abraded sherds were not recognised and were recorded along with the other calcite-gritted sherds of uncertain form. The baking plates included a single fragment from Phase 2 (No. 139), sherds from a single example from Phase 5 (No. 138) and two small fragments retrieved from the evaluation trenches. Prompted by Darling’s study in 2012 there has been an increased recognition and interest in Roman fabricated clay oven fragments from sites in Britain (Darling 2012). Evans, Heke and Peachy (2018) have published important recent site finds in the West Midlands, the examples from Holt/Chester and excavations at Soham, Cambs. where storage jars re-used as ovens were found in situ. Evans found that the ovens from The Hive, Worcester could be dated to the late 3rd to 4th century AD (Evans et al. 2018). Although examples of Flavian clay-built ovens have been recognised from Holt and Chester the majority of the examples appear to date to the third to 4th century AD. In light of these recent publications some of the Late Roman calcitegritted vessels from York may also be from similar oven structures (Monaghan 1997: nos 3210, 3211 and 3216). The examples from Brooklyn House consisted of three fragments from Phase 1, Layer 231 (No. 143) which also contained material dating to the late 4th to early 5th century AD that may be intrusive (Phase 1 discussed below). Four fragments were retrieved from Phase 4, including number 141, all from rubble deposits (215, 248, 410 and 818). The remaining material, excluding one fragment from Phase 7 was all retrieved from Phase 5 (see discussion in the catalogue below).

138 CG: Base of a handmade calcite-gritted ware baking plate, features hollow circle possibly bone stabs in a loose circular arrangement. Examples of similar ‘chapatti disc’ or ‘pizza-plate’ are known from near to a stone-built oven at West Heslerton (Darling 2012; Darling and Precious forthcoming). It is presumed that these vessels were used for cooking, presumably bread based products, in the clay oven. Area 2, subsoil/Roman dump deposit 202 (GS15), Phase 5, D143 (Figure 149)

The clay ovens are often found with pottery baking plates or trays and examples are published from The Hive, Worcester (Evans et al. 2018: fig. 4) and Darling notes examples of the baking plates from the stone built oven complex at West Heslerton (Darling and Precious forthcoming). Swan highlighted a broadly similar tray from Langton villa in a context dating to the 4th century AD (2002: fig. 18.248; Corder and Kirk 1932: fig. 27.146) and noted an example from the Signal Station at Scarborough (Hull 1932: pl. 2.10).

Figure 149: Calcite-gritted baking plate (No. 138)

139 CG: A rim fragment from a calcite-gritted ware baking plate. Area 4, built-up ground 4040, Phase 2, D96 (Figure 154) 140 CG: Corner of portable oven, handmade. Area 2, subsoil/Roman dump deposit 202 (GS12), Phase 5, D101 (Figure 150) 141 CG : Base or upper surface and wall of a portable oven, handmade with crudely fashioned corner. Area 2, rubble spread 248, Phase 4, D76 (Figure 151)

The dating evidence from the site sequence would suggest that the ovens and trays from Brooklyn House, although they may have been in use earlier, were not 128

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 150: Fragment of portable oven in calcite-gritted ware (No. 140)

Figure 151: Fragment of portable oven in calcite-gritted ware (No. 141)

142 CG: A handmade calcite-gritted ware oven or perhaps a storage jar featuring a frilled rim. Not illustrated. Area 2, Roman ground level 231 (GS15), Phase 1, D111

to number 145 and example illustrated from West Heslerton (Darling and Precious forthcoming: No. 170). Trial trench 3, Ditch TT307, TT303, FTS 147 CG Fully handmade large jar or perhaps oven rim fragment in calcite-gritted ware, crude lid seat formed by finger stabbing, no smoothing or wheel finishing, Unstratified from Trial Trench 2, D3E (Figures 154, 155)

143 CG: Rim corner of handmade portable oven. Area 2, Roman ground level 231 (GS15), Phase 1, D110 (Figures 152, 154) 144 CG: Aperture rim of handmade calcite-gritted portable oven. Area 2, subsoil/Roman dump deposit 202 (GS15), Phase 5, D137 (Figures 153, 154)

148 CG: A handmade calcite-gritted ware storage jar or oven chimney, large example. Possible handle or other applied detail appears to have been lost from the ‘neck’. Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D44 (Figure 154)

145 CG: Oven aperture rim of handmade calcitegritted portable oven, reduced interior. Area 2, subsoil/Roman dump deposit 202 (GS09), Phase 5, D17 (Figure 154)

149 CG: A calcite-gritted ware storage jar or oven chimney featuring stabbed roundels around rim and internal channel. Similar decoration is known from a vessel illustrated from York (Monaghan

146 CG: Oven rim from a late Roman group from Trial Trench 3, fragment not illustrated, form similar 129

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 152: Fragment of portable oven No. 143 in calcite-gritted ware showing handmade manufacture technique (rim to the top)

Figure 153: Fragment of portable oven in calcite-gritted ware No. 144

1997: no. 3210). Area 2, subsoil/Roman dump deposit 202 (GS14), Phase 5, D45 (Figure 154)

2004; 2006, Leary 2013). Dales ware has been extensively discussed as it is one of the key dating indicators in Lincolnshire and the north (see Darling and Precious 2014, Leary 2013, Darling 2009, Swan 2002), it appears likely that it reached this site sometime in the 3rd or 4th century AD. Lid-seated jars were the only forms recorded here (as Monaghan 1997 forms JD1 and JD2) and no dishes were recognised. Although an attempt was made to scrutinise the handmade calcite-gritted material it appears that there are perhaps too few body sherds attributed to this fabric but sherds from a maximum of 14 vessels were recorded suggesting that vessels in this fabric rarely reached the site. It

DWSHT Dales ware shell-gritted type (Tomber and Dore 1998: DAL SH). This fabric code has been used to separate out the handbuilt wheel-finished Dales ware typically seen in north western Lincolnshire, and the very similar material seen in Lincoln and Eastern Yorkshire. The sherds from Norton appear very similar to material from the Scunthorpe area although it has also been noted that there were probably other sources making shell-gritted wares using the same manufacturing technique (Vince 130

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Figure 154: Calcite-gritted ovens, large jars and a tray

is likely that the local calcite-gritted jars probably fulfilled a similar function to Dales ware and the vessels that reached this site may have been traded for their contents rather than the vessel itself. A body sherd and a rim (Monaghan 1997: JD2) were retrieved from Phase 1 ground level 247 which might be considered as intrusive; the first occurrence of this ware type in larger fragments was from Phases 3 and 4 with the majority of the smaller sherds retrieved from Phase 5. The low levels of Dales ware would appear to suggest it was a minor component to the assemblage. Evans (1996: 73) has recognised a similar pattern at Beadlam, Rudston, Malton and Langton where Dales ware was also only present in small quantities. No examples of the Late Roman, completely wheel made, ‘double lid-seated’ shell-gritted types common in the latest levels of sites in Lincoln and Lincolnshire were recorded (Darling and Precious 2014; Rowlandson and Fiske forthcoming).

Figure 155: Large calcite-gritted jar with internal finger pressed decoration jar (No. 147)

131

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire LOOL Oolith-gritted ware

The Pottery from the Excavated Groups I.M. Rowlandson with samian contribution by J.M. Mills

A range of fabrics gritted with calcareous ooliths are known from eastern Yorkshire north of Brough where potters appear to have exploited deposits derived from the Cave Oolite and made a range of ‘Knapton type’ and Dales ware lid-seated jars (Rowlandson and Fiske 2019b, Evans 2006). A single sherd from a jar with a tall everted rim in this fabric was recorded from Phase 6. It is possible that a small quantity of this material may have been overlooked amongst the material recorded as CG but it would appear that it was generally a rare occurrence on this site.

Phase 1 One thousand, four hundred and ninety-five sherds (36.143 kg, 27.28 RE) were retrieved from Phase 1 deposits. Of this group only twenty-four sherds, including samian, were retrieved from cut features such as gullies or ditches. These sherds mostly consisted of Norton grey ware including a jar, a sherd from a Norton Black Burnished ware jar with burnished lattice decoration, a sherd from a Central Gaulish form 31 samian bowl, a white ware sherd and two possibly intrusive Crambeck grey ware sherds. A large fragment from a calcite-gritted jar (No. 116) which stylistically could represent a 1st- to 2nd-century AD date was also found stratified with Norton grey ware (NORGW1) and Norton Black Burnished ware (NORBB1). The material from these contexts would appear to suggest that activity began on the site in the later 2nd century AD or 3rd century AD.

SHEL/SMSH South Midlands Shell-gritted wares (Tomber and Dore 1998: HAR SH) Two rim sherds from jars with under-cut rims were retrieved from Phase 4 layer 230 and Phase 5 subsoil. One sherd had surviving examples of the Punctate Brachiopod fossil shell typically present in products of the South Midlands shell-gritted industry which included production at Harrold, Bedfordshire (Brown 1994; Tomber and Dore 1997: HAR SH). The other sherd with a similar rim (fabric SHEL) had a similar firing colour and may also have been from a similar source although examples of the diagnostic shell were not noted. Although it is possible that small scraps of this fabric may have been subsumed amongst the CG fabric group it would appear that little pottery arrived to the site from this source and most probably the products of this industry only reached Norton towards the end of the Roman period. Evans (1996: 75-77) has recognised similar shell-gritted jars from Beadlam, Piercebridge and Catterick, all from contexts dating to the late 4th century AD or after. A limited quantity has also been recorded by Monaghan from York (1997: 913). This sparse distribution would suggest that they were rare and probably only reached the site in the later 4th century AD.

The remaining 1,471 sherds of pottery from Phase 1 were mostly retrieved from layers 218, 231, 247, 262 and 319. The layers from Phase 1 contained a broad mix of pottery including material dating to the late 3rd century or later. The majority of the pottery (1,067 sherds) was retrieved from layer 231. Only layers 247 and 319 contained material that might exclusively be dated to the late 3rd century AD. Small quantities of Crambeck Parchment ware and Parchment ware mortaria sherds produced in the late 4th century AD were present, mostly from layers 218 and 231. Huntcliff jars and S-rimmed jars of 4th-century AD date were also retrieved from layer 231. A number of calcite-gritted sherds from bowls and dishes with plain rims along with oven fragments were also retrieved from layer 231 (Nos 142 and 143). The majority of the Crambeck grey ware from Phase 1 and a sherd of the HOSM1 Holme-onSpalding Moor grey ware were also retrieved from layer 231. It appears likely that this material was intrusive from later activity.

Late Roman grog-gritted wares LGROG – Southern Grog-gritted ware? (Tomber and Dore 1998: HAM GT?)

The material present likely to date to the later 2nd to 3rd century AD consists of calcite-gritted jars included externally bevelled rim types (Nos 109 and 113), a sherd from a bowl or dish in a Black Burnished ware 2 fabric and sherds from colour-coated beakers. Norton grey ware was well represented (363 sherds) including straight-sided bead and flanged bowls (Norton Type 3), plain rimmed bowls (Norton Types 1a and 1b), necked jars, carinated jars (Norton Type 10) and a lug-handled jar (Norton Type 4). It was noticeable that examples of the early oxidised fabrics NOROW1, YE1/NORGW2, OXWS and the sandy NORSGW were present but only in very small quantities. Small quantities of the Norton

A single handmade black fired grog-gritted body sherd (18 g) was retrieved from a Phase 5 deposit. This sherd appeared similar in character to material handled by this author from Hampshire and Sussex were such wares were commonly produced in the Late Roman period (Lyne 2015). If this vessel has been correctly identified it would be an extremely unusual occurrence amongst an assemblage from Yorkshire.

132

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Black Burnished ware 1 and the chalk-rich local grey ware (NORCG) were also present suggesting further sherds that may have been produced in the later 2nd century AD. A single transitional grog-gritted sherd was also noted. The assemblages from Phase 1 layers would appear to contain a mix of material and this might be expected if they represent a contemporary ground layer that may have been subject to disturbance by subsequent activity.

AD although the samian may have been manufactured in the late 2nd century AD. Pottery from the Phase 2 layers Two hundred and eleven sherds were retrieved from Phase 2 layers including ground levels (330, 388, 409, 417, 457, 472, 473, 477 and 4040). The material was relatively fresh with a mean sherd weight of 19.32 g. All of the groups contained material dated to the Late Roman period with sherds from a maximum of two Huntcliff jars and a single sherd from an S-type rimmed jar from layer 409, a sherd of Crambeck Parchment ware from layer 330 and a Parchment ware mortarium (No. 11) from layer 4040 and a calcitegritted tray/platter (No. 139) suggesting the presence of some material dating to the late 4th century AD was present. Colour-coated wares present were restricted to a sherd of Central Gaulish Black ware and beakers in the CC1 and CC3 fabrics that might all date to the later 2nd to 3rd century AD. Small quantities of the early NORGW2 and NORSGW fabric were present along with fragments from Norton Black Burnished ware dishes and sherds of the calcite rich NORCG grey ware. The vast majority of the Crambeck grey ware and calcitegritted wares were retrieved from layer 409. Another vessel of note retrieved from Layer 409 was a beaker base in a Swanpool oxidised fabric that had been ground down to form a small container or possibly an ink well (No. 33).

Phase 1 samian: 20 sherds, 272 g, 0.25 EVE With the exception of one sherd, a Central Gaulish Drag 31 from ditch fill (297 (GS13)) the samian from this phase was recovered from the Roman land surface (218, 231, and 247). The Lezoux (Central Gaul) sherds are small, on average ≤ 8 g, with the only identifiable form a Drag 33 cup. In contrast the other sherds, all from Rheinzabern, are relatively large (c. 20 g mean weight). They date from the late 2nd – early 3rd centuries. Cross-joins were recorded with sherds from the Phase 3 rubble layer (251 (GS15)) giving substantial sections of two vessels, a Drag 32 dish and a Drag 31/Lud Sb bowl (Figure 106). These are the only sherds from the site that were deposited in contemporary levels. Phase 2 Two hundred and ninety-four sherds (5.994 kg, 5.12 RE). The mean sherd weight of 20.39 g was inflated by the presence of a number of complete or near complete vessels from bustum burial 475. The majority of the pottery could be dated to the mid- to late 3rd century AD. As discussed for the Phase 1 assemblage above the material from the cut features appeared to date slightly earlier than some of the assemblages retrieved from layers that also included pottery dating to the late 4th century AD.

The most common wheel made fabric from the Phase 2 layers was Norton grey ware (NORGW1, 57 sherds) with forms including a carinated jar (Norton Type 10), a folded beaker (Norton Type 9) and a sherd from a lid. The range of pottery would suggest that the majority of the pottery from the Phase 2 layers could be dated to the later 3rd century AD or earlier with a few deposits such as Layer 409 that contained a range of later material.

Pottery from the Phase 2 ditches

Phase 2 Bustum burial 475 (Figure 156, see also pp. 29-34 and Figures 42-49)

Twenty-one sherds were retrieved from Phase 2 ditches 270, 397, 420 and 3024. The material from these features could all be dated to the 3rd century AD with the exception of a mortarium produced at York (No. 3; Dickinson and Hartley 1971: fig. 19.7-8; Bidwell and Croom 1997: no. 145) and perhaps a colour-coated bagshaped beaker, both from ditch 397. It was noticeable that little calcite-gritted ware was retrieved from these features the only diagnostic form a jar with an externally bevelled everted rim. A sherd from a Moselkeramik beaker was also present.

The fill from cremation pit 475 yielded two complete vessels and large fragments from a further two vessels. The main bustum vessel was an undamaged handmade calcite-gritted jar with an externally bevelled everted rim (No. 108, RF 416) (Figure 46). Numerous examples of this form have been published from Rudston and Malton (e.g. Bidwell and Croom 1997: nos. 63, 119 and 222) and can be dated from the perhaps the mid2nd century AD and probably most commonly in the 3rd century AD (see discussion in the calcite-gritted wares section). Carbonised deposits survive internally and externally on this jar. The use of such vessels as cremation urns has been recognised at the nearby Leefe’s Bungalow site (Hayes 1988: fig. 59.16) which was found with a stone slab on the top and a broadly similar

Pottery from the Phase 2 structural features Five sherds including Central Gaulish samian (see below) and grey ware were retrieved from Phase 2 road core 424. This group probably dated to the 3rd century 133

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 156: Vessels from Bustum burial 475

vessel (Hayes 1988: fig. 58.14) appears to have been found at the feet of Burial IX. Although it would appear that Black Burnished ware 1 jars were often favoured for bustum burials elsewhere in the north (Evans 2004, Wilmott 1993, Charlton and Mitcheson 1984) this may reflect the use of ‘cook pots’ that were commonly available at the time. A calcite-gritted jar used in a bustum grave has also been published from Birdoswald (Wilmott 1993, Burial 1). It is likely that the calcitegritted example from Brooklyn House was favoured as a ‘cooking pot’ being the commonest vessel of this type in use at the time, as the local Black Burnished industry mostly produced bowls and dishes (Bidwell and Croom 2012: 21).

79). At Birdoswald a large colour-coated flagon appears to have been used as a primary container in a bustum burial (Wilmott 1993). A heavily burnt grey ware or Black Burnished ware 1 lipped dish (No. 38, RF 826, broadly as Bidwell and Croom 2012: fig. 12.15) with chamfered base was found within the pit and also in the cremation pit/round surface deposit 477 and the condition of the vessel would suggest it may have been burnt on the pyre. Examples of heavily burnt Black Burnished ware dishes and bowls were also noted at Trentholme Drive in proximity to cremation urned burials (Gillam 1968: fig. 3.2.10, 11 and 13) and from bustum burials at High Rochester (Wilmott 1993). Other sherds from this context included small fragments of Crambeck grey ware and a calcite-gritted jar with an externally bevelled rim. The majority of the material dated to the 3rd century AD with small sherds of Crambeck grey ware suggesting a later 3rd-century date. It is possible that the Crambeck ware represents intrusive material in this group.

Also present was a complete local grey ware beaker (No. 41, RF 417) with a curved rim and folded body (Hayes and Whitley 1950 fig. 10.9) which, although broken was deposited whole in antiquity (Figure 45). Gillam illustrates eight examples of similar grey ware vessels buried in graves as accompanying vessels at the Trentholme Drive cemetery, York (Gillam 1968: fig. 32.205-208) and another large example appears to have been used as a cremation urn at the Leefe’s Bungalow site, Norton (Hayes 1988: fig. 59.17).

Phase 2 samian: 6 sherds, 30 g Just two vessels were from contexts assigned to Phase 2; a shattered Lezoux Drag 45 rim (AD 170-210) from the core of the Roman road (424) and a 3rd-century Trier plain body sherd from the Roman ground surface (388). Potentially from vessels still in use in the 3rd century, however, given the fact that most of the samian recovered was residual, the case for these sherds being from long-lived vessels must be slight.

The basal area from a large rouletted colour-coated flagon (No. 16, RF 825) was also found but it appeared that, despite having lost the upper half of the body of this vessel it had not been burnt on the pyre. Hayes records examples of burials with colour-coated vessels at Langton Road when St. Peter’s Church Institute was built in 1937 (Robinson 1978: 37 no. 309; Hayes 1988: 134

I.M. Rowlandson and H.G. Fiske: The Roman pottery

the fill of construction cut 3087 would also suggest that this structure may have been constructed in the late 4th century AD. Wall 205, amongst a group of 234 Late Roman sherds, also included a fresh rim fragment from a calcite-gritted Huntcliff jar. The presence of small fresh fragments from Parchment ware and Huntcliff jars amongst these structural features suggest the some of the Phase 3 structures may have been constructed or repaired in the late 4th century AD.

Phase 3 One thousand, two hundred and thirty sherds (21.223 kg, 17.43 RE) were retrieved from groups attributed to this Phase. The majority of the pottery present could be dated to the later 3rd to 4th century AD. The few vessels dating to the late 4th century AD have been highlighted below. Phase 3 ditches, pits and robber cuts

Phase 3 samian: 39 sherds, 734 g, 0.55 EVE

One hundred and seventy sherds were retrieved from these Phase 3 features. The pottery from these features all dated to the late 3rd to 4th century AD. The most significant vessel was the large fresh fragment from a straight-sided bead and flanged bowl from pit 275 (No. 87) that dated to the late 3rd to 4th century AD. Crambeck grey ware was present in small quantities with a higher proportion of Norton grey ware from these features

Samian assigned to Phase 3 levels was widely scattered within surfaces, spreads and rubble layers giving the appearance of being incorporated as rubbish/hard core. The first two sherds (13 g) of 1st-century samian from La Graufesenque appear in this Phase, a chip from a Drag 29 bowl (3007, Figure 104, DS1) and a small rim sherd from the late 1st century of a dish transitional between Drag 18 and Drag 18/31 (828). Also of an early date in terms of this site is a small decorated body sherd in an under-fired Lezoux fabric (3007, Figure 104, DS2). These sherds are so much earlier than anything else on the site it is tempting to imagine they derive from rubbish from elsewhere.

Phase 3 layers With the exception of rubble deposit layer 251 which contained a sherd from a double lid-seated jar that may have been of late 4th-century AD date (No. 89) and surface layer 462 that contained fragments from two Huntcliff jars, all of the groups could be dated to the 4th century AD. The most significant vessel present was a large proportion of a white-slipped flagon (No. 31) retrieved from internal surface layer 3007 within Structure D. The flagon was probably produced in the late 2nd to 3rd century AD although it appears to have been deposited in a much later context. A tableware vessel of this type may have been used for many years before it was disposed of. Other illustrated vessels include a handmade flat-topped jar (No. 111) and an unusual jar with a reeded rim (No. 44). Colour-coated beakers were well represented including a fragment from an imported Moselkeramik vessel. Straight sided bead and flanged bowls in the CRGR and NORGW1 fabrics and a fragment from Oxfordshire Red-slipped ware fashioned into a spindle whorl was retrieved from surface layer 463.

The bulk of the Central Gaulish samian (12 sherds, 101 g) dates from the mid-late Antonine period and includes single examples of forms Drag 33, 37, 38, 45 and Walters 79 and two indeterminate sherds of either Wa 79 or 80. A Drag 38 flange was the only incidence of this form identified from the excavation, although one sherd of this form was identified from the assessment (Monteil 2016). As observed for the Phase 1 samian the vessels from East Gaul, both Rheinzabern and Trier, were represented by larger sherds than the Lezoux wares. Much of this material is from Rheinzabern (18 sherds, weighing 267 g), including forms Drag 32, 37, Drag 31R/Lud Sb and an unidentified shallow bowl with a flattish base (264 (GS15)) all of late 2nd- to early 3rd-century date. The small collection of Trier sherds (6, weighing 354 g) includes a considerable portion of a stamped Drag 31R/Lud Sb bowl dated to the second quarter of the 3rd century (Figure 105, SS1; Figure 106, PS5). A joining sherd was recovered from the Phase 5 soil dump (202 (GS13)).

Phase 3 construction cuts/structural features The majority of the Phase 3 structural features (219, 229, 232, 239, 249, 253, 280, 307, 313, 314 and 327) all contained small quantities of local Norton and Crambeck grey wares that could be dated to the late 3rd to 4th century AD. Two vessels were illustrated from construction cut 3091: a Crambeck Parchment ware Type 5b bowl (No. 23) and a Crambeck white ware mortarium Type 6 with a reeded rim (No. 8). The presence of the Parchment ware sherd would suggest a date in the late 4th century AD for the construction of this structure. A large sherd from a Huntcliff jar (122 g) retrieved from

Phase 4 One thousand, seven hundred and eighty-three sherds (39.478 kg, 22.50 RE). Indicators of late 4th century AD activity were present within the majority of the Phase 4 groups and much of the pottery was retrieved from rubble rich layers. There were few groups that might 135

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire be considered as primary deposits and, although the material was fairly fresh, the composition of the assemblages was mixed in date ranging from the later 2nd to late 4th century AD.

in the style of Helenius (Figure 104). This vessel may have been made in a re-used mould, the thickness of the vessel, fabric and finish suggest a 3rd-century date, c. AD 220-50, making it contemporary with the only Trier product from this phase; a rim sherd from a large Drag 36 with a heavily beaded edge (230/GS13, Figure 106. PS2).

Phase 4 robber cuts Thirty-seven sherds were retrieved from the fills of robber cuts 341, 3102 and 4049. Feature 341 could be dated to the late 4th century AD or later by the presence of a sherd from a Crambeck Parchment ware vessel. Feature 3102 contained a fragment from a Crambeck Parchment ware Type 10 bowl and a fragment from a calcite-gritted jar with an S-shaped rim that suggested a similar date. Feature 4049 contained nine sherds including a fragment from a calcite-gritted jar (No. 115) and grey ware that could only be broadly attributed a Late Roman date.

A further two Rheinzabern sherds came from rubble deposits associated with the road; both rim sherds, one a late 2nd- to early 3rd-century Drag 31/Lud Sa (407), the other another late Drag 36 c. AD 220-260 (410, Figure 106. PS1). The sherd size for most of this samian is small due no doubt to the residual nature of the pottery and especially as again, like Phase 3, the deposits seem to be mostly rubble/rubbish dumps and spreads.

Phase 4 layers

Phase 5

A series of rubble spreads or ashy layers were attributed to Phase 4 and a large group of pottery was retrieved from them (1,745 sherds, 38.956 kg, 22.08 RE). The key wares and chronological composition have been discussed in detail above and it would appear that there was a significant proportion of the pottery present relating to activity in the late 2nd to earlier 4th century AD (see further discussion of samian below). Illustrated vessels included: a spindle whorl fashioned from a calcite-gritted sherd (layer 214, No. 137), the Crambeck Parchment ware cockerel (No. 28) and a miniature calcite-gritted jar (No. 136) retrieved from layer 215, an oxidised dish (No. 36) and a calcite-gritted jar (No. 119) from layer 230, and oven fragment (No. 141) from layer 248 and a mortarium (No. 4) from layer 926. The largest group of sherds, from rubble layer 410, included a grey ware straight-sided bead and flanged bowl (No. 96), a large calcite-gritted bowl with an S-rim (No. 124), a Parchment ware mortarium (No. 13) and a decorated bowl (No. 20).

Twelve thousand, eight hundred and three sherds (244.564 kg, 190.42 RE) were retrieved from the Phase 5 layers with a mean sherd weight of 5.24 g. Over half of the site assemblage was retrieved from Phase 5 layers. The nature of the material from Phase 5 was broadly similar to the material within the Phase 4 layers discussed above. Significant proportions of Crambeck Parchment ware, Huntcliff jars and Late Roman Burnished wares were present suggesting that the layers formed, at the earliest, in the later years of the 4th century AD or perhaps later. Amongst the pottery from this Phase there was a broad range of material dating from the late 2nd to 4th-early 5th century AD. This mixed group has formed the main part of the discussion and attempts at statistical analysis elsewhere in the report (see Tables 11-14 and 18). The material from this group would appear to represent material dumped from nearby occupation in the later 4th to early 5th centuries AD perhaps following on from some of the Phase 4 stone robbing activities.

Phase 4 structural features

Phase 5 samian: 149 sherds, 1580 g, 1.84 EVE

A single sherd of Norton grey ware, perhaps of 3rdcentury AD date, was retrieved from wall 355 of Structure I.

By far the greatest amount of samian from the site was recovered from the extensive rubbish dump/deposit layer 202 and equivalent layers – 206 and 216 within Structure A and layer 303 in Area 3. Because of this quantity of material, the greatest variety of forms and fabrics was recorded amongst this material.

Phase 4 samian 18 sherds, 168 g, 0.21 EVE The rubble spread associated with Structure A (204, 209, 214, 230, 233, 235) yielded seven Lezoux sherds (42 g) along with nine sherds (89 g) from East Gaul. Only one Lezoux sherd was assigned to a form, a dish Wa 79 (204 (GS02)), The eight from Rheinzabern, Drag 32, 37, Curle 21 and two Drag 31/Lub Sb bowls are broadly late 2nd- to first half of the 3rd century in date, the latest probably the Drag 37 (DS8, 233 (GS13))

The few mid-late 1st-century dish sherds from La Graufesenque (5, 51 g) add to those recovered from levels assigned to Phase 3, one rim sherd was of the transitional form Drag 18-18/31 suggesting a late Flavian date, but otherwise none were closely datable; one sherd was burnt. 136

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Sherds from Lezoux vessels were the most common samian found in this dump deposit (88, 659 g, 1.16 EVE), but many were too small (37 sherds, 103 g) to assign to a form, emphasising the residual nature of the 2ndcentury samian. Few vessels were dated to the first half of the 2nd century; the only definitively pre-AD 160 forms were single examples of Drag 18/31 and 18/31R. The bulk of the Lezoux sherds were from vessels postdating c. AD 160/170 as noted in the material from preceding phases. Forms include Drag 31 (2), 31R (7), 33 (3), 36 (4), 37 (5), 45 (4), Wa 79, cup O and P pl. LV, 13 and a small sherd from a closed vessel. The Drag 37 bowls include two late 2nd-century to early 3rd-century (?) examples, one in Banuus style (DS3) and a rim sherd with a poorly defined rim and indistinct ovolo moulding barely visible. The fabric and poor finish suggest this is from a late Lezoux bowl (Figure 104). The mortaria, Drag 45, are similarly late forms (AD 170-210).

Phase 6 and 7 post-Roman deposits Seven hundred and seventy-nine sherds (13.043 kg, 9.93 RE) were retrieved from post-Roman Phases 6 and 7. The range of pottery from this group was broadly similar to the material from Phase 5 but with a lower mean sherd weight of 16.74 g. Only two vessels of note are illustrated from this phase, a grey ware sherd in the HOSM1 fabric trimmed to a spindle whorl (No. 90) and the base from a Crambeck white ware mortarium with sooting over the broken edges (No. 10). The pottery present mostly consisted of calcite-gritted ware (266 sherds) including Huntcliff type jars and oven fragments, Norton grey wares (239 sherds) and Crambeck grey wares (158 sherds). The remaining coarse wares included a Black Burnished ware jar with a cavetto rim, sherds from colour-coated beakers in the CC1 and CC2 fabrics. Phase 7 samian from the topsoil: 11 sherds, 68 g

As noted many times previously the mean sherd weight of the samian from East Gaul especially that from Trier, is notably heavier than for the Lezoux sherds. As well as material from Rheinzabern and Trier two sherds from the Argonne kilns, probably Lavoye, were also noted in this dump deposit (Figure 104, DS5, DS6). These are Antonine in date and contemporary with the Lezoux wares. The Rheinzabern and Trier wares are late 2nd to mid-3rd century in date, predominantly dish and bowl forms of the Drag 31 series. The Rheinzabern wares (35 sherds, 482 g, 0.38 EVE) show the greatest range of forms: Drag 32 (3), 37 (4), 31 (2), 31R (10) and one rim sherd from a variant of dish LudTb/Curle 23 (Figure 106). The decorated sherds (Figure 104, DS9-11) are small and are not closely datable, but some of the Drag 31R bowls have the flattened rims characteristic of later bowls. A base sherd with a characteristically late 2ndto early 3rd-century fabric and finish bears the stamp of Dignus iii, a potter newly recorded for Rheinzabern (Figure 105, SS2). The sherd shaped into a disc and the possible disc rough-out were also from this deposit. Trier wares (14 sherds, 344 g, 0.11 EVE) are limited to dishes Drag 31/Lud Sa (3) and bowls Drag 31R/Lud Sb (5). Pale fabrics and heavy vessels, some sherds easily 10 mm thick, suggest several of these are 3rd-century products.

Most of these sherds are small and featureless not unexpected for topsoil finds; they include two Drag 37 sherds, one from Lezoux and one from an unidentified East Gaulish kiln; both with no appreciable decoration surviving. The only substantial sherd recovered was a Rheinzabern Drag 31R/Lud Sb base fragment from 201/ GS01. Un-phased Two thousand, six hundred and seventy-six Roman sherds (40.721 kg, 32,26 RE) were recorded from unphased deposits. This material has been described in detail in the archive. Unstratified Samian: 48 sherds, 587 g, 0.22 EVE The unstratified sherds are generally representative of the collection as a whole both in terms of fabric and form. From Lezoux Drag 18/31, 31, 31R (3), 45 (2) Walters 79 and one closed vessel; from Rheinzabern: Drag 32, 37 (3), 31R/Lud Sb (2) and from Trier Drag 32 or 36, 45, 31/ Lud Sa, 31R/Lud Sb (3) and Curle 21. Sherds from Trier wares are slightly over-represented and include joining sherds from a heavily used Curle 21 base of probable 3rd-century date and a Drag 31R/Lud Sb sherd joining one from the Phase 5 dumping (subsoil/Roman dump deposit 206 (GS02), Figure 106). The only example of a Trier mortarium was recorded in this group of sherds.

As dumped material it is interesting to note that there is little physical abrasion shown on these sherds, and certainly not more so than for earlier deposits. A handful of sherds had been burnt, a couple heavily, at least one of which was probably from South Gaul and thus 1st century in date. The sherd weight is below the mean for the site as a whole and a high proportion of sherds are of indeterminate form type, about 50% by count; these two measures indicate that this is dumped material. However, the low level of early material, as evident for the entire site, suggests the dumped material was probably a primary deposit.

The Pottery from the Evaluation BHN15 Nine hundred and sixty-three sherds (15.826 kg, 8.53 RE). were recorded from the evaluation phase of the project including material retrieved from trenches some distance from the Roman roadside (Buglass 2016; Monteil 2016; Rowlandson and Fiske 2016). A similar range of pottery was retrieved to that discussed from the main 137

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire phase of excavation and the vast majority of sherds could be dated to the 4th century AD. Calcite-gritted coarse wares, followed by Crambeck grey wares were the most common types present and were more numerous than the local grey wares produced at the Norton kilns in the 2nd to 3rd century AD. None of the sherds studied could be attributed to an Anglian or Anglo-Saxon tradition. A few rare imports include a fragment from a Moselkeramic beaker from modern day Germany that was probably manufactured in the 3rd century AD and some sherds of Spanish olive oil amphorae manufactured some time before AD 250. Notable vessels from this group were the fragments of a Crambeck grey ware face pot (No. 64), a fragment from a Smith god pot (No. 70) and a calcitegritted fragment from an oven (No. 147).The Bronze Age sherd from this phase of work has been reported on by Vyner (Chapter 3).

with Lezoux sherds (13) more common than East Gaulish (7), an overall late 2nd- to 3rd-century date range, and a similar range of material including forms. The forms identified from the evaluation include Drag 18/31, 31, 31R, 37, 38, and 45; the single Drag 18/31 dish an outlier from the bulk of the material. The presence of a single decorated sherd attributed to Cinnamus (Monteil 2016) is a welcome addition to the small collection of decorated material from Lezoux. Elsewhere in Norton at Wood Street and at Jack Berry House in Malton (Monteil 2018; 2017) the 1st-century samian from La Graufesenque clearly forms a good portion of each assemblage; both also contain early 2nd-century wares from Les Martres-de-Veyre (Table 15). Dish forms Drag 18/31, two from the excavation and one from the evaluation, and a single example of the rouletted variant 18/31R were identified and are usually dated c. AD 120-160. It is of note that none of the companion cup form, Drag 27, were identified. We can be sure that this absence is real, had Drag 27s been present they would have been readily identifiable as even small sherds are very distinctive. The virtual absence of early material at Brooklyn House may indicate that the site was not occupied at all until the later 2nd century AD, perhaps representing expansion of the settlement as activity south of the river intensified when the Malton fort was re-garrisoned and the accompanying vicus expanded. Much of the samian was re-deposited in layers of dumped rubbish, although very little was battered or abraded. The East Gaulish vessels in particular were quite large sherds suggesting that the material had not travelled far from its point of use.

Discussion The Brooklyn House Samian J.M. Mills To date this is the largest quantified collection of samian from Norton. Recently excavated groups from Malton/Norton reported on by Gwladys Monteil are quantified by weight, sherd count and rim EVE (Table 14). The largest of these from Wood Street (Monteil 2018) comprises 102 sherds (1134 g, 1.29 rim EVE) and although a statistically invalid quantity, like the present assemblage, provides the best comparative data. These collections are clearly small when compared with those excavated around the Malton fort and vicus on the other side of the river. Those excavations took place some decades ago, and comparable quantification is not available; the samian reported on by the late Dr Grace Simpson (1997) includes 20 stamped vessels and illustrates more than 30 of the decorated wares, and the samian published in 1964 (in Mitchelson 1964: 251-253) lists 38 decorated and 35 stamped vessels (Table 15). These figures suggest that the assemblages from which they derive were much larger than those the recent excavations in Malton/Norton have produced (Table 14).

The Norton samian shows that at least some of the residents acquired expensive tableware. Samian forms a very small part of the ceramic assemblage, less than 2% by sherd count, but its presence is important, adding as it does to Wilson’s suggestion that in Norton ‘high status occupation may have existed in fairly close proximity to the industrial activity’ (2006: 44). A settlement like Norton closely linked to a fort and the road network must surely have benefitted from the ready access to traded goods, including samian, afforded by this position.

The bulk of the Brooklyn House samian dates from the second half of the 2nd century through until the mid-3rd century, c. AD 160, or even AD 170 to about AD 250. Very little samian pre-dates the main assemblage; the earliest sherds are probably of Flavian date and from La Graufesenque in South Gaul. Interestingly no Trajanic-Hadrianic wares from Les Martres-de-Veyre (Central Gaul), or other lesser centres of similar date were identified, although a single decorated sherd (DS2, 3007) has been tentatively identified as being from an under-fired, early 2nd-century Lezoux vessel.

Wider comparisons The profile of this samian assemblage is distinctive, but as a small collection it is not statistically valid. The comments and comparisons made here must be read with this in mind and larger assemblages sought that may prove or refute the impressions gleaned from this material. In his survey of samian from British excavations Willis (2005: section 8.2) characterised five site types and described the ‘fingerprint’ of each type in terms of relative proportions of samian vessel forms by

The results from the evaluation (BHN15) and the excavation at Brooklyn House (BHN16) are comparable, 138

I.M. Rowlandson and H.G. Fiske: The Roman pottery

Excavation/publication date

No Sh

1964 Malton Vicus 1997 Malton Vicus sites Jack Berry House (Monteil 2017) Wood Street, Norton (Monteil 2018) BHN15 – assessment BHN16 - excavation

? ? 20 104 20 291

Decorated wares catalogued/illustrated 38 >32 5 11 1 11

Stamped sherds catalogued 35 20 0 1 0 3

Samian date range Nero – 230? 60/70 – 230? 70-200 60/70 – 250 120-230/40 60/70-250/60

Table 14: Summary of selected excavated assemblages, the relative numbers of decorated and stamped vessels illustrating the potential size of those excavated in the twentieth century FABRIC SAMSG SAMMV SAMCG SAMLA SAMRZ SAMTR ?SAMTR SAMEG SAM Total sherds

Jack Berry House 9 2 8 1 2 22

Wood Street 21 5 59 1 14 2 102

BHN15 13 7 20

BHN16 7 0 153 2 88 34 3 2 2 291

Table 15: Quantity (by sherd count) of samian from recent excavations in Norton

vessel function. The Brooklyn House material does not readily fit any of these models (Table 16). A site recently excavated at The Hive in Worcester, however, has a similar profile (Mills, J. 2018). The unusual profile there was driven by the large quantity of samian dumped in the Phase 6 quarry pits, a collection with high numbers of Drag 31R/Lud Sb bowls, gritted mortaria and Drag 38 bowls but few cup forms and a slightly depressed incidence of decorated bowls.

using the data from table 85 (Bird 1986) it can be seen that 17% of the vessels identified were cup forms. Another late group which provides a useful comparison is the samian from excavations in and around the fort at Piercebridge. Although Ward (2008: 177) makes note that within the East Gaulish samian from Piercebridge cup forms were less common than dish/bowl forms of the 31R group (which includes Drag 31 and 31R forms), and cites several very late sites in Germany where cups were scarce or absent, the percentage of cups at Piercebridge was not as low as at either Worcester or Norton. The small numbers of cup forms from Brooklyn House, and the absence of East Gaulish examples may be a consequence of the small size of the assemblage, although this seems unlikely and given the numbers of cup forms found at New Fresh Wharf it can be seen that supply continued at a decent rate well into the 3rd century. The low numbers of cups could perhaps be due to site type or status but is not readily explained.

This assemblage, along with that from The Hive may fit best with the profile for ‘smaller civil centres’ although the low level of cups, and unusually high numbers of mortaria and plain bowls in both groups require some explanation. To address these functional classes separately is the easiest approach. Cups forms were produced throughout the entire span of the samian industry. It should be remembered that the term ‘cup’ describes the shape of the vessel and does not imply function as a drinking vessel. At Norton just seven cups from Lezoux were identified and none from East Gaul. An additional Lezoux cup, from Trial trench 3 (Context 303), is mentioned in the assessment report. The samian from New Fresh Wharf, St. Magnus House, London (Bird 1986) was comprised of large numbers of vessels thought to derive from single shipments of samian dated AD 170-180 and AD 235-245 against a background of earlier material. There were plenty of cups from East Gaul at New Fresh Wharf,

In the 2nd and 3rd centuries AD the most common plain bowl forms were the flanged bowl Drag 38 and the rouletted bowl Drag 31R and its East Gaulish variant, Lud Sb. Of the two, Drag 38 was produced for the longest time, the type having emerged in the Hadrianic period (Webster 1996: 49). It consistently occurs in 2ndcentury levels but wasn’t as common as the decorated bowl Drag 37. At New Fresh Wharf there were also fewer Drag 38 bowls than mortaria (Bird 1986: table 85). Drag 31R emerged around AD 160 and was imported, as 139

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Willis (2005) Roman Site Types Vessel Functional Category Inkwell Decorated Jar/Beaker Decorated Bowl Plain Bowl Cup Dish/Platter Mortaria

Military sites 0.4 1.2 27.1 3.7 25.6 39.8 0.1

Settlement outside Major civil military sites settlements 0.1 1.7 38.4 3.8 18.9 34.3 0.4

0.4 1.1 23.2 4.9 30.1 38.5 0.2

Smaller Rural civil settlements settlements 0.3 1 19.4 11.9 23.7 36.1 2.2

0 1 21 3 25.1 47.4 1

Borough House, Norton

The Hive Worcester

0 1.6 17 31 6.5 32.3 8.9

0 1 18 20 12 37 4

Table 16: The Relative Frequency of Samian Form/Functional Categories at different site types, Brooklyn House, Norton and a comparative site from Worcester (Mills J. 2018) Note: Vessel classes ‘Indeterminate Bowl’ and ‘Bowl/Dish’ have been omitted here as these are of negligible importance to any site type, the greatest value being 4.3% bowl/dish for smaller civil settlements

were flanged bowls, until trade with East Gaul ceased around AD 260. New Fresh Wharf doesn’t provide a good comparison for the Drag 31R/Lud Sb bowls because the unusually high proportion (c. 30% of vessels from both Central and East Gaul) of indeterminate Drag 31 or 31R form. The same applies to the Piercebridge data although it is worth recording that the 31/31R group was noted there as the most common East Gaulish vessel type (Ward 2008: 177). Because Drag 31R/Lud Sb bowls were not introduced into the repertoire of forms until at least AD 160 their presence will appear to be less on sites continuously occupied since the 1st century. Consequently, sites with little 1st- or early 2nd-century material will always show an elevated proportion of plain bowls. This must in part explain their dominance for sites or deposits dating from the late 2nd century.

the Brooklyn House assemblage could be the reason for the low frequency of cups, however, factors such as status or site type need to be considered; perhaps it is the industrial nature of Norton influencing the choice of samian vessels. Larger assemblages will be needed to fully address the differences observed here. Comparison with other assemblages from the area I.M. Rowlandson The assemblage showed an affinity with other urban groups from the Malton/Norton area, but the chronological range was strongly biased towards the later 3rd to 4th century AD. This is at present the largest quantified assemblage of pottery of this period from Norton. As the site probably lay on the edge of the Roman settlement it would not be surprising if settlement did not develop on the site until the late 2nd century AD and perhaps the plot became a site for rubbish disposal as the area of settlement contracted in the later 4th century or beginning of the 5th century AD. Similar contemporary patterns of waste disposal have also been recognised at York (Monaghan 1997: 847).

The proportion of samian mortaria in most assemblages is low, according to Willis’ figures usually less than 3%. The percentage is not only heightened at The Hive and Brooklyn House; at New Fresh Wharf (Bird 1986: table 85) mortaria formed 6% of the Central Gaulish assemblage, and 9% of the East Gaulish, and the figures were even higher at Piercebridge (Ward 2008: 178) at 10% and 13% respectively. Samian mortaria were produced from about AD 170, and like bowl Drag 31R relative quantities will appear reduced on continuously occupied sites. Again, the late start date for this site may explain the apparently high numbers of mortaria.

Seeking comparisons with other assemblages in the area it was clear that wheel made wares were well represented and the assemblage was similar to other groups from towns and forts in the area but starkly different from basic rural assemblages where handmade pottery and coarse-gritted wares dominated assemblages (e.g. Didsbury 2004; 2009; 2013; Rowlandson 2012). The calcite-gritted wares from Brooklyn House made up 46-49% of the assemblage (sherd count and weight respectively) and 35% by Rim Equivalent. If this is contrasted with rural assemblages Iron Age tradition wares or Huntcliff type wares typically make up a far larger proportion of the assemblage. Away from the larger settlements only villa assemblages included a

In conclusion, the late date of these groups may explain why the vessel function profile does not fit with Willis’ average profiles. Vessel forms such as Drag 31R/Lud Sb and mortaria which were introduced in the late 2nd century take a proportionally higher percentage of assemblages of late 2nd- and 3rd-century date. This does not explain the very low occurrence for cup forms at Brooklyn House, where no cups were recorded in the Rheinzabern and Trier samian at all. The small size of 140

I.M. Rowlandson and H.G. Fiske: The Roman pottery

CRPA % of assemblage

MOCRF % of assemblage

% of Sherd % of weight % of RE Phase 1 Phase 2 Phase 3 Phase 4 Phase 5

0.94 1.36 0.08 0.28 1.04

0.84 0.55 0.31 0.8 1.71

1.5 0.39 0.75 1.69 1.52

All parchment % of assemblage

% of Sherd % of weight % of RE 0.33 0.34 0 0.28 0.19

0.58 1.6 0 0.66 0.48

0.77 7.23 0 2.31 0.57

% of Sherd % of weight % of RE 1.27 1.7 0.08 0.56 1.23

1.42 2.15 0.31 1.46 2.19

2.27 7.62 0.75 4 2.09

Table 17: Crambeck Parchment wares as proportion of the assemblage by phase

diverse range pottery with significant proportions of table wares as recorded at the Brooklyn House site (Evans 1996; Rigby 1980). Evans’ tableware assemblages from Beadlam were listed as 51.8% of the late 3rd- to mid-4th-century AD group and 55.3% from all the later 4th-century groups (1996: 71). A crude comparison with the forthcoming report on the Roman pottery from excavations at West Heslerton would appear to highlight this further as utilitarian calcite-gritted wares make up between 69-79% of that assemblage and greatly outnumber wheel made table wares on a site predominantly dated to the late 4th to 5th century AD (Darling and Precious forthcoming). However this should be viewed with some caution as the proportion of calcite-gritted wares generally appears to be higher amongst assemblages from the later 4th century into the early 5th century AD at Beadlam villa and at York (Monaghan 1997: 850 – Ceramic Period 4b-c) and also at the Carr Naze signal station site, also probably a late 4th-century AD foundation (Monaghan 2000).

Superficially the quantities of Parchment ware from Phase 4 would fit with those in 4th-century AD deposits from Huttons Ambo (Evans 1989: table 5). The Phase 4 assemblage was much higher in Parchment wares than many of the sites dated to the late 3rd to 4th century AD from sites located a distance away from the kiln site (e.g. Brough-on-Humber or Greta Bridge) but was much lower in Parchment wares than the assemblage from the York Skeldergate well (mid-4th century AD), the late 4th- to early 5th-century AD assemblages from Catterick, Huntcliff Signal station and a deposit from Malton dating to the mid-4th century AD (Evans 1989: tables 5 and 11). However, an uncritical comparison with the good cohesive groups of pottery used by Evans in his study is not wise: the site formation processes in the Brooklyn House assemblage resulted in a mixed assemblage with a high level of residually evident in the Phase 4 and 5 assemblages with pottery ranging in date from the later 2nd to 4th century AD. The upper fills of the York Skeldergate well contained only 153 fresh sherds, most probably a primary assemblage including many near complete vessels (Perrin 1981: 4950, figs. 26-27). The Brooklyn House Phase 4 assemblage was much larger (1,783 sherds) and many from layers with a considerable proportion of residual pottery that precludes close comparison. Comparisons of relative proportions of Parchment ware from Brooklyn House to the assemblages from the Signal Stations is also flawed as many of these coastal sites were established de novo in the later 4th century AD and the assemblages lack earlier residual material. The low level of Parchment ware from Phase 4 deposits in contrast with the signal stations group reflects the significant quantity of Norton grey ware and other residual material dating to the 3rd century AD present in the Brooklyn House assemblage.

An attempt has been made to make a comparison with Evans’ statistical study of Crambeck ware distribution (1989) with limited success. The Crambeck Parchment wares (including mortaria) have been tabulated as a percentage of the assemblage by phase (Table 17). This has been undertaken to address possible questions of dating the Phase 4 and Phase 5 deposits and also to see if the assemblage from Brooklyn House can be seen as comparable with other broadly contemporary quantified assemblages. Although some of the sites in Evans’ survey were recorded with minimum number of rims only some basic observations appear possible. The material from Brooklyn House Phases 1 and 2 consisted of only a few Parchment ware sherds and, as discussed above, would largely appear to be intrusive. The quantity of Parchment ware from Phase 3 was negligible, a single vessel, so this can be discounted as not a viable sample.

Looking to the relative proportion of Parchment wares amongst the Phase 5 deposits there are similar challenges when considering how site formation processes have biased the pottery assemblage. The Phase 5 assemblage was mixed including a broad range of material dating to the later 2nd to 3rd centuries AD as well as contemporary material likely to date to the late 4th to perhaps earlier 5th century AD. The statistical ranges for Parchment ware was

The proportion of Parchment wares from Phase 4 assemblage was fairly low on the basis of proportion of sherds and weight but had a disproportionally high Rim Equivalent (RE) measurement on the basis of the fragment from the Parchment ware cockerel (No. 28). 141

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire calculated in a range from 1.23-2.09% depending upon the quantification method used. This would be lower than Evans’ mid-4th-century group from Malton, the late 4th-century deposits from Beadlam villa, the Huntcliff Signal station and a number of older extant assemblages from other signal stations (Evans 1989: tables 5 and 11). When the proportion of the Crambeck Parchment wares from the Carr Naze signal station are considered (Monaghan 2000: table 5, sample 1,735 sherds, Parchment ware = 0.86% of assemblage by sherd count, 0.81% by weight, 2.43% by EVE) the Brooklyn House assemblage from Phase 5 superficially appears to have a similar proportion of Parchment ware but this does not take into account the large quantity of earlier wheel made pottery present in the Phase 5 assemblage that deflates relative percentage. From this very basic foray into statistical analysis it would appear that there was a limited potential to use the Phase 4 and Phase 5 assemblages for detailed comparisons with the cohesive dated assemblages studied by Evans (1989) due to the relative percentages of different wares being skewed by the range of pottery present from earlier ceramic periods. As noted above the method of contrasting the relative proportion of calcite-gritted wares to grey wares discussed by Bidwell and Croom (2010) also falls foul of the same problem: if there is a large quantity of residual material amongst the Phase 4 and 5 assemblages close comparisons with primary groups of the late 4th century AD do not work.

terrace, levelling up the ground for building and filling in dangerous quarry pits. In some cases such domestic rubbish may have been a commodity in demand when a settlement was vibrant and expanding. However, in a period of decline and contraction such waste was probably disposed of on disused plots as can be seen from investigations at the Roman roadside settlement at Navenby (Rowlandson et al. 2011; 2015). During the final period of Roman pottery use in this part of Norton it would appear that the inhabitants had access to a good range of pottery produced locally but, perhaps in the absence of large scale construction projects, mounds of rubbish were left in convenient abandoned areas of the settlement. As the settlement never expanded again in the 5th century AD this material may have largely remained where it had been left as the settlement focus perhaps shifted towards the defended enclosure at Malton in the 5th century and there is evidence that some of the Roman walls at Malton were only robbed in the sixteenth century as the town expanded again (Wenham and Heywood 1997: 38). The value of the Brooklyn House pottery is that it offers a large assemblage highlighting the range of material that was in use at Norton, or perhaps Malton; including a diverse range of specialist ritual or cultic pottery and a few highly significant primary groups such as the bustum burial. This is a significant site for understanding the pattern of Roman pottery usage by the inhabitants of Roman Norton, how pottery was disposed and an insight into their foodways, rituals and beliefs. With further investigations and quantified publications of pottery from the area it may be possible to use the statistically large sample from Brooklyn House for further, more nuanced, comparisons, particularly to consider the taphonomic processes and waste management strategies in use in other parts of the Norton settlement.

The conclusions that can be drawn from this comparative process are that the Brooklyn House assemblage showed that the inhabitants of this part of Norton had access to a good range of fine wares from the later 2nd century AD until at least the late 4th century AD. The statistical importance of the assemblage is the volume of pottery dating to the late 3rd to 4th century AD that was retrieved. As the bulk of the pottery was retrieved from Phase 5 deposits, particularly much of the pottery dated to the later 4th century AD, it is a little difficult to characterise the relative quantity of fine wares in use. What is noticeable is that the quantity of sherds of Crambeck Parchment wares from this site exceeds the sample size of many of the groups studied by Evans (1989: tables 5 and 11) suggesting that, amongst the range of other pottery from the site that the inhabitants had access to, there was a good range of local fine wares during the late 4th century AD.

ACKNOWLEDGEMENTS Ian Rowlandson wish to thank Kay Hartley, J.M. Mills, J. Bird and David Williams for their report contributions and to Charlotte Bentley for her pottery illustrations. Thanks also go to Phil Mills for kindly providing copies of his recent work. Many thanks go to John Buglass, Pete Wilson and Janet Phillips for discussing aspects of the site and their patience whilst the report was completed. Thanks go to M.J. Darling, B.J. Precious and Prof. Powlesland for access to information about the West Heslerton assemblage in advance of publication. Great thanks go to Hugh Fiske for his assistance in recording the assemblage, photography, editing and research on the ‘funnies’ though any oversights remain my own.

The broader question posed by the assemblage was why so much pottery was present amongst Phases 4 and 5? One possibility is that the plot became used as a midden towards the end of the 4th century AD. Using studies from Pompeii (Dicus 2014) and Medieval Worcester (Bryant 2011) it is clear that discarded ceramics and the associated detritus had a value within a settlement when construction projects were underway such as re-constructing a rampart, constructing a wharf or

J.M. Mills wishes to thank Joanna Bird for discussing the samian vessels from Argonne.

142

Chapter 5

The post-Roman Pottery C.G. Cumberpatch type ware (early mid-13th to 14th century; contexts 200, 202, 207, 216, 231, 258 and TT300; Mainman and Jenner 2013: 1230-1245). The sherds included rod and strap handles from jugs, one of which may have been purely decorative (context 202).

INTRODUCTION The assemblage consisted of 1316 sherds weighing 12,288.5 g representing a maximum of 1282 vessels from both stratified and unstratified contexts. Full details can be found in the archive report and data tables.

Staxton/Potter Brompton ware (early/mid- 13th to early 14th century; Brewster and Hayfield 1992) was represented by just two small body sherds from contexts 207 and 4040. This was an important regional ware which differed in a number of respects from contemporary types, notably in the fact that vessels were hand-made with turned rims rather than being wheel thrown. Neither of the sherds was identifiable to a specific vessel type.

THE POTTERY The earliest post-Roman sherd was of late Saxon type (context 730), distinguished by its hard black fabric containing abundant, well-sorted, sub-angular quartz grains up to 1 mm in size and its highly burnished, pale brown external surface. It was associated with a handmade sherd of probable early post-Conquest date in a chalk-tempered gritty fabric. containing common quartz grains up to 1 mm in size and sparse grains of soft white chalk up to 3 mm in size. The presence of a small dish in transfer printed Bone China from the same context indicated that the earlier sherds were residual in a mid/ late 19th-century context. Despite this, the presence of these sherds is significant in view of the limited quantities of pottery of a similar date from the region.

Other regional types included a sherd of Tees Valley B/C type ware (context 202), distinguished by its coating of white slip, a characteristic of the later Tees Valley wares (late 13th to 14th century) and apparently intended to disguise the typical orange-red body of the Tees Valley B ware (Didsbury 2010). Slip coatings were also noted on three sherds of an unidentified type; White-slipped Sandy ware (contexts 202, 207 and unstratified). The use of slip is similar to that seen on the Tees Valley B/C ware but the fabrics differ significantly, suggesting that other potters were using a similar technique to appeal to their customer’s preference for buff-white rather than orange or grey wares.

Two sherds of Brown Sandy ware were identified in context 202. Both were dark grey to brown in colour and contained fine quartz inclusions while one, the base of a jar also contained slightly larger rock fragments. Neither the fabrics nor the forms (notably the overhanging rim) were familiar and, while a date range between the 11th and 12th centuries is probable, a 10th-century date cannot be entirely ruled out.

Scarborough ware was represented by four sherds (contexts 202 and 207). The dating of Scarborough ware remains controversial (Watkins 1987) but lies broadly in the 12th- to 14th-century range. The sherd from context 202 was of particular note as it was the rod handle from a jug with bright copper-green glaze on the upper surface. A sherd from context 207 and perhaps one from context 202 appeared to be of Scarborough Gritty ware type which pre-dates the main phase of the Scarborough ware industry (Slowikowski 2000: 74).

The medieval pottery assemblage included a range of local and regional wares which spanned the period between the 12th century and the end of the medieval period. York Glazed ware, dating to the mid/late 12th to mid/ late 13th century (Mainman and Jenner 2013: 1203 – 1224, 1230-1234), was recovered from contexts 200, 201 and 202. With the exception of a jug handle all were body sherds, one of which was decorated with an applied pellet and another with shallow impressed lines. York Glazed ware was probably made in villages lying to the north of York and its presence in Norton is not unexpected. Much the same is true of Brandsby-

Despite the problems of residuality, the range of earlier medieval wares gives a good impression of the diversity of the local and regional pottery industry between the later 11th and 14th centuries. It highlights Norton’s place in the regional economy and the role of its market in drawing potters or pot-vendors to the town. 143

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire From the very late 13th or early 14th century the regional wares were superseded by Reduced Greenware (also known as Late Reduced ware). These wares were characterised by their dark grey reduced bodies and all-over dark green glaze. The earlier types seem to have had a sandy texture and there may have been a general move towards finer fabrics over time so that by the 15th century the fabrics were very fine and almost smooth in texture (Vaughan 2007). Reduced Greenware was relatively common in the present assemblage with examples from contexts 103, 200, 201, 202, 207, 279, 409, TT423 and TT502. One sherd (context 201) was particularly fine and may be of early post-medieval date. Unfortunately, in the absence of a comprehensive study of the industry and with very few manufacturing sites published, this remains a rather impressionistic method of ascribing dates and caution should be exercised when considering the finer divisions within the broad mid-14th- to later 15th-century date range.

contained fine quartz and black grit up to 0.5 mm in size. The glaze varied from light to dark green. Other unidentified fabrics, including Oxidised Sandy ware, Oxidised Gritty ware and Reduced Sandy ware were present in a variety of contexts. Vessel types included jugs, as indicated by the presence of handles, but the majority of sherds were small body sherds and not identifiable to form. The end of the medieval period saw a significant transformation in the pottery industry with the development of different clay bodies, glazes and vessel forms, even while the basic technology remained largely unchanged (Cumberpatch 2003). This rapid change was complex in nature and included developments within the medieval tradition alongside the appearance of recognisably ‘post-medieval’ wares. A substantial group of Late Medieval Sandy wares (TT203, TT606, 201, 202, 207) were amongst those made in potteries on the southern side of the North York Moors, between Thirsk and Pickering. These sites played an important role in the late medieval and early post-medieval pottery industry but they remain poorly known both in terms of what they were producing and the dates between which they were active (Hayes 1988; Mainman and Jenner 2013; Cumberpatch 2014). The use of green glaze links them with the medieval industry but the typical forms, dishes, bowls and pancheons, are quite distinct from typical medieval forms. One sherd (context 202) appeared to belong to a flask or bottle. The fine, even, sandy fabrics varied in colour from dull orange to pale grey and generally contained only quartz although in some cases fine black and white rock fragments were also visible. Two sherds of Green Glazed Sandy ware (contexts TT205 and TT314) differed slightly in appearance and may be slightly later in date.

Seven sherds of Humberware (Hayfield 1992; Hayfield and Grieg 1990; Mayes and Hayfield 1980) were identified in contexts 201 and 202 and it seems that, while the inhabitants of Norton were using Reduced Greenwares, other wares, including Humberware and Osmotherlytype ware (context 200; Mainman and Jenner 2013: 1232) were also readily available, presumably via local markets. In addition to the wide range of named local and regional types, the medieval component of the assemblage included a number of types as yet unidentified to source but which are familiar from other sites across northern Yorkshire. The earliest types (mid-11th to mid-13th century) included Buff Sandy, Buff Gritty, Buff-Orange Sandy and Splash-glazed Sandy wares (contexts 201, 202, 207 and 813). A single sherd of Coarse Sandy ware (context 202) and a sherd of Gritty ware (context 201) were probably of a similar date. A second sherd of Gritty ware from context 3117 was perhaps slightly later in date (13th to 14th century) and was notable for the presence of dark green glaze externally; earlier Gritty wares usually have only small spots and splashes of glaze. Gritty wares, while a very common component of earlier medieval assemblages, become rarer in the later medieval period, probably in response to a rise in the use of metal cooking vessels which replaced them in their principal role.

Cistercian ware, the earliest of the distinctively ‘postmedieval’ wares, dates to the period between c. 1450 and c. 1600. The sherds (from contexts 200, 201, 202 and unstratified) included bases and handles from small cups or tygs with one decorated with an applied pellet externally (context 201). Such decoration often has religious and specifically Roman Catholic connotations (Spavold 2009) and such vessels probably pre-date the Reformation. Late post-medieval wares included Blackware (contexts 200, 201, 202 and 452) and Yellow ware (contexts 101 and 201), both typical of the 17th century and found widely on sites of the period. Known production sites include Wrenthorpe near Wakefield, but it is probable that others also existed, given the large quantities of these types and the earlier Cistercian wares from sites in north-east England.

Contexts 201, 202, 231 and 903 contained sherds of an unusual type designated Coarse Whiteware. While similar to York Glazed ware, the buff to white fabric contained flat, rounded white to grey rock inclusions up to 1 mm in size and sometimes larger, contributing to the distinctive coarse texture of the fabric which also 144

C.G. Cumberpatch: The post-Roman Pottery

A distinctive post-medieval ware, probably local in origin, has been termed ‘Redware type’ (contexts TT430, 101, 200, 201, 202, 207, 279, 403, 409 (?) 803 and unstratified). Typical Redware, dating to the 17th and early 18th centuries, has a soft orange to red fabric and examples were noted in contexts TT416 and 201. The ‘Redware type’ had a much harder, even textured, dark orange fabric with sparse to moderate quantities of fine quartz grains and sparse round red grit, generally less than 0.5 mm in size (although there was some variation in grain size between vessels). The glaze varied from clear (red) to a mottled green or green-brown. The range of vessels included dishes and bowls, often with clubbed rims and jugs or cisterns with strap handles and, in one case, a footed base. On general principles a date range in the 16th to 17th centuries is suggested but it is possible that an earlier (late 15th to 16th century) date may be appropriate for the green glazed vessels.

production (Cumberpatch 2014). This change in both the economic and social structure of the industry, as well as of its products, is represented by the appearance of White Salt Glazed Stoneware (c. 1720 – c. 1780) and by Creamware (c. 1740 – c. 1820), Black Basalt ware (mid- to late 18th century), Pearlware (c. 1780 – c. 1840) and Edged ware (late 18th to early 19th century). The superior practical qualities of these wares saw the end of Tin Glazed Earthenware production by the middle of the 18th century. Of the eight sherds of porcelain (all from topsoil and unstratified contexts), a small number may be Chinese imports but the majority seem to be English and of early modern and recent date. White Salt Glazed Stonewares included an array of plates, cups and bowls (contexts 101 and 201). Two sherds of Black Basalt ware (context 201), were decorated with turned surfaces and sprigged motifs. Creamware (contexts 101, 201, 207, 279 and 303) and Pearlware (contexts 101, 103, 201, 207, TT105, TT412) were also relatively common with both hand-painted and transfer printed sherds amongst the latter type. With one exception (Eton College, context 103), the designs were unidentifiable due to the small size of the sherds. The presence of these wares indicates the extent to which such fashionable wares were available to the provincial market. None of the sherds bore maker’s marks to indicate their origins but manufacture was as well established in Yorkshire as it was in Staffordshire and a regional origin is probable.

The manufacture of Tin Glazed Earthenware (contexts 201 and TT423) spanned the period between the mid16th and mid-18th centuries. Despite its practical limitations and cost, Tin Glazed Earthenware was popular as a tableware and as a container for medicines and cosmetics. The hand-painted designs, in blue, yellow and red on a white or pale blue background, were the closest that European potters could come to producing something that resembled Chinese porcelain in appearance although the very soft fabric and hard but brittle glaze with its poor adhesion to the underlying body meant that it was easily damaged. The dating of individual examples relies on the identification of the painted designs and these rarely survive the processes of excavation, finds processing, bulk-bagging and transport. Its presence in the assemblage, like that of the 18th century wares described below, suggests that at least some of the pottery was derived from well-to-do households of yeoman status or above with aspirations towards gentility. Tin-glaze combined with handpainted decoration was also used to produce wall tiles into the 19th century. Two examples of such tiles were present in context 201 and an unstratified context.

While the rise of the factory-based pottery industry has tended to dominate historical accounts of the period, it is clear from archaeological evidence that local pottery production continued throughout the 18th century with a range of vernacular tablewares and utilitarian wares manufactured by local potters using techniques little different from those of the post-medieval period. There has been no detailed study of the period in northern Yorkshire but it seems likely that the situation was similar to the one in southern Yorkshire (Cumberpatch 2014). Certainly the range of wares was similar with Late Blackware, Slip Coated ware, Slipware and Mottled ware all present amongst the stratified and unstratified components of the assemblage.

European wares were represented by two sherds of Westerwald stoneware (context 201 and unstratified) and two sherds of Frechen-Koln stoneware (context 202), both typical of the types of stoneware imported from the Rhineland in large quantities in the late medieval and post-medieval periods. Two sherds of Brown Salt Glazed Stoneware from context 201 may also have been of German origin and were distinguished from later English stonewares by their fabrics which contained visible grains of quartz.

Late Blackware was the commonest type (contexts 101, 201, 207, 400, TT104, TT420 and TT430) and included bowls, cups/mug/tankards and other hollow wares. Fabrics included fine red examples as well as a smaller quantity of fine buff and orange variants, suggesting a number of sources. This would be consistent with the situation in southern Yorkshire where many small potteries operated within the same region.

The end of the post-medieval period and the start of the early modern period, dates to the second decade of the 18th century and the beginning of industrial pottery

Slip Coated ware vessels, distinguished by the use of red slip on buff fabrics to produce a finish similar to that of 145

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire Late Blackware, included hollow wares and at least one bowl (contexts 103, 200, U/S and Topsoil).

Utilitarian wares included many of the Brown Glazed Coarsewares described above with a substantial quantity of Unglazed Red Earthenware. The latter included conventional flowerpots as well as larger and more elaborate horticultural vessels, including one with elaborate ‘grotesque’ decoration (pit 908). The same context contained a black ceramic finial, probably from a structure such as gazebo or summerhouse.

Five sherds of Mottled ware (contexts 200, 201 and TT416) included the rim of a cup or mug (context 201). The majority of the fabrics were a conventional light buff in colour but one (context 201) was in a fine red fabric of the type more normally associated with Late Blackware. Slipware was well represented with seventeen sherds (contexts 200, 201, 202 and TT502-1). The vessels included press-moulded dishes and bowls with multicoloured banded, swirled and feathered slip patterns and pie-crust rims and these were equalled in number by hollow wares with trailed and feathered decoration externally.

DISCUSSION Most of the medieval and later pottery came from topsoil contexts and contexts created by dumping with only a small proportion from specific features. This discussion focuses on some of the more significant assemblages with the smaller and more ambiguous layers and features omitted for the sake of clarity. Full details can be found in the archive report.

A high proportion of the utilitarian wares (notably Brown Glazed Coarseware but also some of the Yellow Glazed Coarseware) were probably of an 18th-century date although dating these types of pottery is difficult and the assemblage certainly included 19th-century examples. Pancheons and bowls predominated with few hollow wares, a pattern familiar from many sites across Yorkshire.

The rubbish dump contexts 202, 206 and 216 produced mainly medieval pottery with smaller quantities of post-medieval, early modern and recent material. The medieval pottery included a wide range of wares of earlier medieval date (Brown Sandy ware, York Glazed ware, Tees Valley B/C ware, Scarborough ware and Brandsby-type ware) with later medieval types including Humberware, Reduced Greenware and Late Medieval Sandy ware. Post-medieval wares included Cistercian ware, Blackware and Redware type. Early modern and recent wares were comparatively rare; two sherds of Slipware, a sherd of Brown Glazed Coarseware and a small quantity of Unglazed Red Earthenware.

Brown Salt Glazed Stoneware of 18th- and 19thcentury date was well represented in the assemblage. Eighteenth-century types include mugs and tankards (context 201) but from the early 19th century production shifted to cooking wares suitable for use in domestic coal-fired kitchen ranges. Examples of such wares included dishes and possibly loaf pots with grey glaze on the internal surface in the popular ‘French’ style. Bottles and flagons were present in both salt-glazed and lead-glazed forms, the latter being commoner. Fine stonewares of 19th-century date included part of a jug or coffee pot (context 201) while a group of sherds from stratified and unstratified contexts (201, 207, TT314) came from one or more decorative jugs or vases of later 19th-century type with a relief moulded design of reeds, wheat-ears and leaves on a stippled background.

Other contexts often contained less than five sherds of pottery and it is difficult to argue that the occurrence of one or two sherds is definitive of a medieval date for the deposit as it was clear that the activities on the site over the years had involved the disturbance of a significant area of medieval activity resulting in the incorporation of residual medieval pottery into later features and layers. Small quantities of pottery were recovered from cut features. Context 258, the fill of pit 259, contained a single small sherd of Brandsby-type ware, suggested as being intrusive in an earlier feature. A small, abraded sherd of medieval pottery was also recovered from context 285, the fill of infant burial SK286.

From the early 19th century onwards a range of cheap colourful wares were produced in large quantities in pottery factories throughout the country and are ubiquitous on sites of the period. Examples in the present assemblage include Banded wares, Blue Bodied ware, Cane Coloured ware, Slip Banded Cane Coloured ware, Colour Glazed ware, Jackfield ware, Sponged and Sponge-printed ware and Relief-banded ware. Bowls, dishes, jugs and other vessel types were identified although number of sherds was relatively low.

Single sherds of recent pottery were recovered from the fills of pits 337 (context 338), 364 (cut and fill) and 763 (context 761) while contexts 904, 908 and 909 (fills associated with pits 908 and 912) contained a single sherd of Brown Salt Glazed Stoneware and nine sherds of Unglazed Red Earthenware including three with grotesque decoration. Pit 3004 contained substantial parts of two vessels; a flowerpot and a jar of the type used for the retail sale of a wide variety of

Whiteware and Bone China, both plain and transfer printed were relatively common and included plates, dishes, cups and saucers. Identifiable designs included Willow, Albion, Two Temples and perhaps Wild Rose. 146

C.G. Cumberpatch: The post-Roman Pottery

jams, marmalades and other preserves. This vessel bore a maker’s mark identifying it as a product of the Maling company of Newcastle-upon-Tyne. Such jars were produced in huge numbers (in excess of 1.5 million per year for the Keiller Marmalade Company alone) from the early 1850s onwards. The presence of the jar would indicate a date for the filling of the pit after 1853.

and significance for the area. A notable feature was the presence of pottery of late Saxon and peri-Conquest date, neither of which are common in the area and are of considerable significance, despite their lack of a firm context and the small numbers of sherds involved. The wide range of medieval pottery types reflects the town’s position in the local and regional marketing network and the extent to which the presumed marketing areas of different potteries overlapped.

A variety of contexts were identified as earlier ground surfaces, buried topsoil and yard deposits and produced small assemblages of pottery. Contexts 730 and 231, a mid-/late 19th- to early 20th-century yard and a Romano-British ground surface respectively, both contained sherds of medieval pottery. In the first case this consisted of the sherd of local Saxon ware and the sherd of Chalk-tempered Gritty ware mentioned above, alongside the profile of a transfer-printed Bone China dish. In the second case the context contained the base of a Brandsby ware jug and the strap handle from a jug in a Coarse Whiteware fabric. Context 903, a probable Romano-British rubbish deposit, also contained a sherd of Coarse Whiteware.

In general terms, the pottery assemblage reflects the social history of the town and the changes in social attitudes and aspirations over time. Some of these, such as the appearance and rapid adoption of fine stonewares and refined earthenwares during the 18th century are well known and have been extensively described and discussed in the context of the ‘Georgian Order’ and the move towards ‘civility’ that marked the period. It should also be noted that the persistence of traditional forms, represented by the 18th-century vernacular tablewares, has not been accorded such detailed attention and remains absent from most historical narratives, despite the abundant archaeological evidence for the economic and social significance of such wares (Cumberpatch 2014).

Context 803, described as a ‘Victorian yard deposit’ contained two sherds of pottery, both of which predated the period 1837 – 1901 by a considerable margin. One was a sherd of Redware type dating to the 16th or 17th centuries while the second, the rim of a dish or bowl, was a small section from the rim of a bowl in early Yellow Glazed Fineware of 18th-century date.

Other major phases of change, including the move from buff-white and orange wares to reduced wares during the 14th century and the rapid replacement of medieval style wares by the distinctive ‘post-medieval’ tradition during the 15th century (Cumberpatch 2003) have received much less attention and remain to be fully explored and explained.

CONCLUSION Although the assemblage was dominated by unstratified and poorly stratified pottery, it was not without interest

147

Chapter 6

The coins Richard Brickstock The investigations at Brooklyn House, Norton in 2015 and 2016 produced a total of 83 coins. Only one was illegible (and was perhaps not a coin), the remainder being remarkably well-preserved and thus almost all fully-identifiable.

be potentially misleading, given the degree to which the presence or absence of a single coin can affect the results. The vast majority of the finds (77 out of 82) relate to period after AD 260, and appear to indicate continuous activity throughout the later-third and fourth centuries. The majority of the ‘Radiates’ are regular issues of the AD 260s and early 270s, but both fullmodule and small module copies are present, as is the coinage of Carausius (AD 286-93) and that of every subsequent coin issue period with the exception of (the always poorly-represented) AD 378-88. A particular peak, perhaps archaeologically significant, is recorded

A full catalogue is appended (Appendix 1) while Table 18 provides a brief summary of the number of coins in each issue period (following the periods outlined in the author’s English Heritage guidance note; Brickstock 2004). Figure 157 presents the coins in graphical format, though it should be borne in mind that graphical representation and/or statistical analysis of such a small assemblage can 30

Brooklyn House, Norton

Annual loss per 1000 coins

25 20 15 10

copies

Total: 83 coins, comprising: pre‐AD 41: 0 graphed: 82 others: 1

regular

5

41‐54 54‐68 69‐81 81‐96 96‐117 117‐38 138‐61 161‐80 180‐92 193‐222 222‐38 238‐60 260‐75 275‐86 286‐96 296‐317 317‐30 330‐48 348‐64 364‐78 378‐88 388‐402

0

Coin issue periods

Figure 157: Graphical summary of Brooklyn House coins Figure *2: Graphical summary of Brooklyn House coins

Figure 158: Counterfeit denarii of Trajan, Elagabalus and Orbiana (obverses only).

148

Richard Brickstock: The coins

for the second quarter of the fourth century (AD 33048); while a single coin of Gratian (AD 375) and another of Honorius (AD 394-402) indicate continued activity right up to the end of the fourth century.

relatively-short passage of time; and that, together with the presence of substantial numbers of ‘Radiates’ of the AD 260s and early 270s, provides the first definite numismatic indication of contemporary activity.

Of the earlier coins, three are counterfeit denarii cast in base, vaguely-silvery, alloy (rather than being plated copper cores). One counterfeits an issue of Trajan (BHN16, RF 84; AD ‘103-11’); the second, a coin of Elagabalus (BHN15, RF 239; AD ‘218-22’); and the third, an issue in the name of Orbiana, wife of Severus Alexander (BHN15, RF 54; AD ‘225-26’) (Figure 158).

This does not, however, of necessity indicate a lack of occupation in years prior to the mid-third century. In discussing the coinage of Shiptonthorpe, Bryan Sitch (Millett 2006: 99) noted Richard Reece’s very plausible suggestion (Reece 1987: 20 and elsewhere) that on many rural Romano-British sites, but not in the vici adjacent to or attached to forts, large-scale use of coinage began only in the mid-third century (c. AD 260 or later), perhaps in part because the native population had had no real use for the predominantly high-value denominations in circulation up to that time.

Counterfeits could, of course, have been produced at any time after the mint-date of their prototypes – the process would have been potentially profitable at any time – but it seems increasingly likely that many, and probably the majority, of such forgeries belong to the mid-third century or even later (Boon, 1988; Brickstock in prep) and thus, I suggest, should perhaps be viewed as an element of the coinage of that period (i.e. alongside the radiate copies of the AD 270s).

The coin histogram for Brooklyn House (Figure 157) bears strong similarities to that of Shiptonthorpe (Figure 159), which could be taken to suggest that such a situation also pertained at Norton. This similarity may, however, be more apparent than real: the Brooklyn House profile is, if anything, even more closely paralleled by the finds from Peter Wenham’s excavations in the vicus at Malton in 1968-70 (Wenham and Heywood, 1997; Figure 160, Table 19). There the 59 coins include only six coins earlier than AD 260: a worn, but, significantly, not very worn, as of Vespasian (AD 69-79); a very worn sestertius of Antoninus Pius (AD 138-61); three Severan denarii (AD 193-211; one, of Geta, only slightly worn); and a counterfeit antoninianus of Gordian III/Valerian (i.e. AD 253+).

If this theory is correct, the earliest coin from the site becomes a worn Hadrianic as (probably a Britannia type of AD 119). Given that it is substantially worn, it may have circulated for a considerable period of time before its deposition, perhaps in the mid- to late-second century or even the early-third century. After that, the next coin in the assemblage is an antoninianus of Trajan Decius (BHN16 RF 119; AD 249-51): perhaps significantly, it appears relatively little-worn (SW/SW-W according to my classification), arguing deposition after a

30

Shiptonthorpe 1985‐91

Annual loss per 1000 coins

25 20

Total: 274 coins, comprising: pre‐AD 41: 0 graphed: 261 others: 13

15 10

copies

5

378‐88

388‐402

364‐78

348‐64

330‐48

317‐30

286‐96

296‐317

275‐86

260‐75

238‐60

222‐38

180‐92

161‐80

138‐61

117‐38

81‐96

96‐117

69‐81

54‐68

41‐54

193‐222

regular

0

Coin issue periods

159: summary Graphicalofsummary of Shiptonthorpe (1985-91) coins Figure *3:Figure Graphical Shiptonthorpe (1985-91) coins

149

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire The larger coin list from the 1949-52 excavations in the vicus (Mitchelson 1964; Figure 161), however, while presenting a very similar profile to the Wenham finds for the later-third and fourth centuries, provides sufficient coinage of the later-first, second and earlythird centuries to confirm what is inferred from other types of data, namely that the Malton vicus developed, and adopted coin use, very soon after the establishment of the fort in the later-first century (Wilson 2006). What has yet to be tested, through a study of circulation wear exhibited by the fort and vicus coinage, is (a) whether the fort was abandoned during the Hadrianic and early Antonine periods, on the model of e.g. Bainbridge; and (b) whether occupation of the civilian settlement continued unabated in spite of this possible abandonment.

Whatever the truth of these current unknowns regarding Malton, I think it more than likely that the adjacent Norton settlement mirrored the occupation sequence of the Malton vicus rather than that of Shiptonthorpe – but numismatic proof of that claim is not yet available: coin finds from earlier excavations present a very similar picture to the current assemblage (e.g. a total of 43 coins from three sites at Sutton Cottage, Eastfield and Model Farm, Norton; Clark 1935; Hoy 1981); and the same is true of the 50-odd coins currently listed by the Portable Antiquities Scheme for the parish of Nortonon-Derwent.

30

Malton Vicus 1968‐70

Annual loss per 1000 coins

25 20

Total: 59 coins, comprising: pre‐AD 41: 0 graphed: 58 others: 1

15

copies

10

regular

5

378‐88

388‐402

364‐78

348‐64

330‐48

317‐30

286‐96

296‐317

275‐86

260‐75

238‐60

222‐38

180‐92

193‐222

161‐80

138‐61

117‐38

81‐96

96‐117

69‐81

54‐68

41‐54

0

Coin issue periods

Figure *4:Figure Graphical Malton Vicus (1968-70) coins 160: summary Graphicalof summary of Malton vicus (1968-70) coins 30

Malton Vicus 1949‐52

Annual loss per 1000 coins

25 20

Total: 343 coins, comprising: pre‐AD 41: 0 graphed: 315 others: 28

15 10 5

Coin issue periods

378‐88

388‐402

364‐78

348‐64

330‐48

317‐30

296‐317

286‐96

275‐86

260‐75

238‐60

222‐38

193‐222

180‐92

161‐80

138‐61

117‐38

96‐117

81‐96

69‐81

54‐68

41‐54

0

Figure *6: Graphical summary of Malton Vicus (1949-52) coins (regular and copies largely Figure 161: Graphical summary of Malton vicus (1949-52) coins (regular and undifferentiated): copies largely undifferentiated)

150

Richard Brickstock: The coins

EH coin issue periods Number of coins

Pre-AD 41 41-54 54-68 69-81 81-96 96-117 117-38 138-61 161-80 180-92 193-222 222-38 238-60 260-75 275-86 286-96 296-317 317-30 330-48 348-64 364-78 378-88 388-402 Illegible/not coin Total

0 0 0 0 0 1 counterfeit 1 0 0 0 1 counterfeit 1 counterfeit 1 17 + 6 copies 0 3 + 1 copy 1 6 26 + 8 copies 1 + 6 copies 1 0 1 1 83

Table 18: Summary of coin finds from Brooklyn House, Norton (2015-16)

EH coin issue periods Pre-AD 41 41-54 54-68 69-81 81-96 96-117 117-38 138-61 161-80 180-92 193-222 222-38 238-60 260-75 275-86 286-96 296-317 317-30 330-48 348-64 364-78 378-88 388-402 Illegible/not coin Total

Brooklyn House, Norton 0 0 0 0 0 1 counterfeit 1 0 0 0 1 counterfeit 1 counterfeit 1 17 + 6 copies 0 3 + 1 copy 1 6 26 + 8 copies 1 + 6 copies 1 0 1 1 83

Malton Vicus 1968-70 0 0 0 1 0 0 0 1 0 0 3 0 1 counterfeit 10 + 3 copies 0 1 3 3 + 1 copy 17 + 2 copies 3 + 2 copies 6 1 copy 0 1 (uncertain C2/3) 59

Table 19: Summary of coin finds from Brooklyn House and Malton Vicus (1968-70)

151

Coins Appendix 1

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

152

Richard Brickstock: The coins

153

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

154

Richard Brickstock: The coins

155

Chapter 7

7a – The small finds and vessel glass H.E.M. Cool and S.J. Greep INTRODUCTION The excavations produced a large assemblage of small finds and relatively few items of vessel glass. At the assessment stage over 600 items were inspected. Where independently dateable the majority were of Roman date with a small amount of late medieval to Victorian material. The latter generally consisted of dress fittings and horseshoe nails, i.e. material that could be casually lost whilst crossing the site. The approach taken with the report is to present all of the material which can be dated to the Roman period either on the grounds of its context or independent typology. The post Roman finds are noted in the assessment report which is available in the archive. The finds from the bustum burial are considered elsewhere (see p.187)

Function Personal equipment Toilet equipment Textile equipment Household equipment Recreational equipment Writing equipment Transport equipment Building equipment Tools and craft equipment Fastener and fittings Agricultural equipment Military equipment Miscellaneous items Total

An important feature of the assemblage is the excellent preservation of the ironwork. This, combined with the X-radiography, has meant that the iron could be studied with very little investigative conservation. This level of preservation is unusual outside of special circumstances such as water-logged conditions. It has undoubtedly contributed to the recovery of unusually complete items such as the double link snaffle bit No. 58; of the small items such as the needles Nos 36-37; and of large parts of items which do not tend to be commonly recognised such as the trivet No. 49. The recovery of very unusual items such as the complete set of tongs No. 81 has been facilitated by this good preservation but it does also point to craft processes that can rarely be identified in an assemblage such as this.

1

5 1

Site Phase 2 3 4 - 2 3 - - - - 2 2 - -

Later Total 5 22 4 36 2 2 2 4 4 7

-

-

-

-

1

-

1

1 1 -

1

-

-

4 2 4

3 -

8 3 5

1

-

1

1

13

1

17

1 6 - - 1 - 4 10

2 1 9

12 5 7 76

2 1 13

24 1 8 9 125

1 1 1 1 13

Table 20: The catalogued small finds by site phase and functional category. (Note: see text for exclusions. The column labelled ‘Later’ combines the Phase 7, unphased and unstratified material)

The two authors of the report have had the following inputs. Stephen Greep has reported on all of the items made of bone and antler. Hilary Cool has been responsible for all of the items made of the other materials, for the introduction and overview, and for the illustrations.

The catalogued material is summarised in Table 20 by functional category and phase. This table excludes iron nails and hobnails and fragments of vessel glass. Personal equipment and the fasteners and fittings categories are well represented as they are on all Roman sites. The number of items in the tools category is, in part, the result of the good preservation of the iron. A military presence can be seen throughout. A large part of the assemblage came from the late 4th- to 5th-century rubbish dumping on the site in Phase 5. Although some items in that material are older than that date, including a remarkable La Tène I brooch (No. 1), this material casts some interesting sidelights on the nature of the occupation in the vicinity and the level of supply at that late period. This will be discussed further in the overview.

PERSONAL ORNAMENTS AND EQUIPMENT The only item that can be assigned on typological grounds to pre-Roman activity in the area is the La Tène I brooch No. 1. This is extremely corroded and lacks part of the spring and the pin. The end of the foot is dished and may once have had an inlay. It is not now possible to say whether or not the bow was ever decorated. It falls within Hull’s Type 1B and was probably a Type 1Bb (Hull and Hawkes 1987: 95). The Hull corpus considered the type to have been a southern form with no examples in the Midlands or further north. This pattern has changed in the years since the book was published. One, for example, came from Dragonby (Olivier 1996: 231 no. 1, fig. 11.1) and the Portable Antiquities Scheme has recorded seven from Yorkshire and Humberside (as 156

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

of October 2019) with the most northerly coming from Snape with Thorpe in North Yorkshire (Echtenacher 2007). The recovery of one from Norton is thus not surprising. What is surprising is its context. Brooches such as this are normally dated to the 4th and 3rd centuries BC. This example was found within the late Roman Phase 5 rubbish dumping so it would have been in active use more than half a millennia before its deposition date.

with swelling bulging stems, primarily designed to hold hair designs in place. Here the Type A pins were found throughout the sequence. Type B pins all came from the Phase 5 rubbish dumping with exception of No. 13 which came from a Phase 4 rubble spread associated with Structure A. Type A2 pins (Nos 5-7 have pointed or conical heads, typically with a series of grooves beneath forming a series of collars. They are dated to c. AD 43 – 200/250 (for the full typology see Greep 1983: 334-9 and fig. 83). Examples of Type A2 pins with decorated heads are comparatively rare. The two examples from Norton (Nos 8-9) are therefore worthy of particular note and add to evidence for regional production of pin types. Of all the forms A2/3-5 bone hairpins the author has recorded, 40% come from the York area (and a further 20% from Caerwent/Caerleon).

The other brooches from this episode of dumping are less surprising. Another residual brooch (No. 2) is represented only by the lower body and foot. The precise type of bow brooch it came from cannot be identified but it probably belonged to the Headstud family and so may be dated broadly to the century following c. AD 75. The other bow brooch is an early crossbow brooch which would have been in use during the second half of the 3rd century and into the 4th (Mackreth 2011: 200 type 2d). Its presence is probably another aspect of the military activity in the area in the later 2nd and 3rd centuries as demonstrated by the military equipment (see Nos 116-24) and the soldier buried in the bustum burial (see p. 187). In the 3rd century brooches are more closely associated with the military community than they are with the civilian one, and the later crossbow brooches were certainly worn by males in positions of authority (Swift 2000: 3-4).

Although these forms have a widespread distribution throughout Britain, they are particularly common in and around York. For the type represented by No. 8 there are at least eleven examples from York itself, easily the largest site assemblage in Britain. Ten are unpublished in the Yorkshire Museum (Acc. No’s H2030 1, 2, 4-5, 8, 10, 12, 19-20 and 31; a published example from Blake Street (Cool et al. 1995: fig. 720, 6403 from late Roman contexts and presumably residual) and other local examples from Aldborough (Bishop 1996: fig. 11, 70) and Catterick (Wilson 2002: fig. 320, 84). The form was clearly present in the 1st century as evidenced by two examples from Flavian contexts at Shakenoak, Oxfordshire (Brodribb et al. 1971: fig. 50, 24-25).

The presence of the iron penannular brooch No. 4 in the same late episode of rubbish dumping is of some interest in that it is complete and so need not be residual. The form is a simple one (Mackreth 2011: 209 type f1; Fowler 1960: type D) and occurs throughout the Roman period so cannot be dated closely. Iron examples, however, were certainly being worn in the early 5th century as one was found pinning cloth in a grave of that date at Lankhills and there seems to have been a resurgence of interest in penannular brooches generally at that time (Booth et al. 2010: 284-85). No. 4 is probably reflecting this.

No. 9 is another common form in York with two unpublished examples with chevron decoration in the Yorkshire Museum (Acc. Nos H120.7 and H120.13) and one from excavations by the York Archaeological Trust. A further five unpublished examples with trellis decoration in the Yorkshire Museum (Acc. Nos H120.12, H2029.3, H2030.3, H2031.9 and H2035.8 and Dickinson and Wenham 1957: fig. 14, a) For other local examples with trellis decoration see Aldborough (Bishop 1996: fig. 11, 68 and Catterick (Wilson 2002: fig. 320, 84 and fig. 321, 123).

Hairpins are represented by 13 bone examples (Nos 5-17) and one made of shale (No. 18). A further seven bone stem fragments were recovered that could have come from either hairpins or needles.

Type B1 pins with oval or round heads and like head forms and swelling stems are the most common form of later Roman hairpin throughout Britain. While they first occur in mid-2nd-century contexts, they are most numerous in the 3rd and 4th centuries (e.g. Greep 1995a: 1117). Here they are represented by Nos 10-14 which are all from the Period 5 rubbish dumping with the exception of No. 13 from the Phase 4 rubble associated with Structure A. Another late form, Type B4 with a faceted cuboid head and a swelling stem (No. 15) also came from the Phase 5 rubbish dump as did two other

There are a number of typologies in use for hair pins of bone and antler. Here, we follow the most comprehensive, based on a national study of the material; the best overview remains that provided in the Marlowe, Canterbury report (Greep 1995a: 111321). Broadly, bone and antler hairpins may be divided between earlier Roman forms (Type A), with tapering stems and heads rarely larger than the stem, designed to stand proud of the hair. They are, typically, longer than the later Roman hairpins (Type B), which are shorter, 157

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire Type B pins (Nos 16-17). Although the stem of No. 13 is straight, it probably belongs to the later Roman period on the basis of the shape and size of the head which is aligned to Type B hair-pins. A similar one comes from Aldborough (Bishop 1996: fig. 13, 98) but with a small collar beneath the head and acute lines rather than the simple trellis of our example.

patterns often seen on a very common type of 4thcentury copper alloy light bangle and sometimes on jet ones (Cool 1983: 157 Group XXII; Allason-Jones 1996: 30-31 nos 57-68). On the latter the triangular notches are sometimes paired across rather than alternating to produce a running lozenge design on the upper edge and this is the design used on the only fragment of a jet bangle in the assemble (No. 31) which came from a Phase 4 abandonment spread. Excavations in the area of the Malton vicus have regularly recovered shale and jet rings in some quantities. Again they have generally been plain but both those of 1949-52 and 1968-70 have produced fragments of shale or jet ones decorated in the same way as No. 30 though generally the execution has been better (Mitchelson 1966: 253 no. 34, fig. 19; Hooley 1997: 150 no. 5).

The shale hairpin represented by No. 18 is also likely to be of late Roman date as hairpins made in black shiny stones, most frequently jet, tend to be of that date. This fragment was found in an early ground level of Phase 1. Segmented beads in green glass like No. 19 from a Phase 7 context are known as early as the first half of the 2nd century (Cool et al. 1995: 650 no. 5830) but are most frequently found in the later Roman period. These are common beads. The long cylindrical bead No. 20 is much less common as peacock glass does not appear to have been a colour often used for beads. Ones combining the form and colour have been found on bead strings in late 4th- to 5th-century graves at Lankhills (Booth et al. 2010: 293) though long cylindrical beads in general have a similar chronological trajectory to segmented beads, being occasionally found in contexts as early as the 2nd century but being most frequently found in late Roman ones. No. 20 is another example of this as it was found a Phase 4 context.

All fragments of annular shale and jet rings have been described here as bangles but, depending on size, some may have had other functions. Allason-Jones (1996: 35) suggested those with an internal diameter of less than 45 mm may have had other purposes. At Lankhills shale rings were found deposited in graves as part of a pile of bracelets, so their function there seems undoubted. In one case a child aged between six and eleven years had one with an internal diameter of 49 mm, whilst an adolescent aged between 13 and 15 had one of 55 mm (Booth et al. 2010: 130 no. 12, 142 No. 18). An adult female would need one with an internal diameter of c. 70 mm. Of the Norton rings No. 24 with an internal diameter of 45 mm may fall into the other function category. Nos 2429 could have functioned as arm ornaments though Nos 27-28 are of a size either appropriate for a man to wear at the wrist or a woman on the upper arm. The inner diameter of No. 30 is irregular and so a size cannot be suggested. The jet ring No. 31 at 40 mm seems too small even for a very young child. Unlike the small jet rings discussed by Allason-Jones at York, No. 31 is very well made, so only its size separates it from the larger bangles. Perhaps it was a toy for a doll. Roman little girls had dolls with appropriate sized equipment just as modern ones do, even if the role they played was different in the two periods (Martin-Kilcher 2000: 65-67).

Bracelets and bangles were found made of copper alloy (Nos 21-23), shale (Nos 24-30), jet (No. 31) and bone (No. 31). All of the copper alloy and bone examples were found in the Phase 5 rubbish dumping together with most of the stratified shale ones (Nos 11, 13-15). Wearing bracelets was very popular in the 4th century so the number in the assemblage is not unusual. Cable twist bracelets such as Nos 21 and 22 were in use by the early 2nd century and were most popular in the later 3rd to 4th centuries (Cool 1983: 120-29 Group I). They are the commonest type of metal bracelet found in Roman Britain and there is some evidence that they may have been going out of use in the later part of the 4th century (Booth et al. 2010: 303). Light bangles decorated in a variety of ways were the other principle class of metal 4thcentury bracelets (Cool 1983: 152-181). Here the family is represented by No. 10 which has decoration restricted to vestigial nicks along the edges. Such decoration is more normally combined with decoration such as punched dots down the centre. On its own it is unusual.

Antler/bone bracelets like No. 32 are a typical late Roman form, dateable to the last third of the 4th century or the early 5th century (e.g. Booth et al. 2010: 300-01). The closest parallel to the present example, from Portchester (Webster 1975: 44 and fig. 117, 99) is dated in the published account to c. AD 325-345, possibly suggesting that these forms may have been in use earlier in the 4th century, although the Portchester example remains a chronological outlier. Four, more complete bracelets, have been recovered from nearby Malton (Greep 1983: fig. 232, 9-10 and 233, 11-12).

Annular shale rings have a long history stretching back into the Iron Age and were in use throughout the Roman period with decorated examples tending to be of late Roman date (see Lawson 1976: 247-250; Booth et al. 2010: 300). Here Nos 24-29 are plain and No. 30 from a Phase 3 floor level is decorated with small nicks on alternate sides recalling the better executed zig-zag

The unstratified No. 33 is a hoop fragment from a jet finger ring. When complete these are normally types 158

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

which in metal belong to the 3rd to early 5th centuries, but the absence of any bezel details here means that it is not possible to date No. 33 more closely within the late Roman period.

organic around junction of curved and hoop. A coherent grain pattern could not be identified but it appeared similar to wood. Diameter 31 mm. 2016: 202: RF 238: Phase 5. Figure 162

Finally in this section the hobnails from shoe soles can be considered. In total 65 were recovered as site finds and a further two were identified in the bustum deposit (see p. 196, Nos 13-14). The earliest incidences were the three from a Phase 1 early ground level (304). One came from the Phase 4 rubble deposit 406 and 11 from the Phase 5 rubbish dumping. The rest came from modern Phase 7 contexts (302, 402, 502) but as these contained other Roman finds these too can be regarded as Roman. It was possible to measure the head diameters of 41 from the X-radiographs and these ranged from 4 to 11 mm with a median diameter of 8 mm (interquartile range 7 to 9 mm). In only two cases, both from (302), were any corroded together. In these cases in groups of two and three. This suggests a nailed shoe might have been disposed of in its entirety. The others might have been casual losses from a shoe that continued to be worn or, given the late Roman chronological focus of this site, from lightly nailed shoes which would not have had nailing patterns where hobnails were close enough to corrode together (see p. 189 for a consideration of nailing patterns). 1

2

3

4

5

Hairpin. Bone Type A2 with a tapering stem, although the head and tip are both broken, making identification of the exact form uncertain. 99 mm long. 2016: 236: RF 197: Phase 3.

6

Hairpin. Bone Type A2.2 with two collars. 43 mm long, broken. 2016: 247: RF 324: Phase 1.2. Figure 162

7

Hairpin. Bone Type A2.2 with three collars. 37 mm long, broken. 2016: 206 Grid square 3: SF 52: Phase 5. Figure 162

8

Hairpin. Bone Type A2.3 with a pointed, conical head, below which is a band of acute angled lines above a collar. Tapering stem 100 mm long, complete. 2016: 231: RF 199: Phase 1. Figure 163

9

Hairpin. Bone Type A2.4/5 hairpin with a conical head and a chevron/trellis design between two collars. 42 mm long, broken. 2016: 202 Grid square 5: RF 85: Phase 5. Figure 163

10 Hairpin. Bone Type B1. 88 mm long, complete. 2016: 206 Grid square 3: RF 125: Phase 5. Figure 163

La Tène I brooch. Copper alloy; heavily corroded. Arched D-sectioned bow tapering to spring and foot; one loop of spring extant; foot with part of return plate bends back to bow with cup formation at tip. Length 42 mm. 2016: 206 Grid square 3: RF 57: Phase 5. Figure 162

11

Hairpin. Bone Type B1. 57 mm long, broken. 2016: 202 Grid square 8: RF 59: Phase 5. Figure 163

12

Hairpin. Bone Type B1. 77 mm complete – modern break. 2016: 202 Grid square 5: RF 92: Phase 5. Figure 163

13 Hairpin. Bone Type B1. 75 mm long, broken. 2016: 230: RF 154: Phase 4. RF 2. Figure 163

Bow brooch; bow and foot fragment. Copper alloy. Rectangular-sectioned curved bow with narrow groove parallel to edges; circular, three-ribbed projecting foot with flat base. Trapezoidal catch plate with rib extending up back of bow. Present length 36 mm. 2016: 202 Grid square 5: RF 83: Phase 5. Figure 162 Crossbow brooch. Copper alloy with white metal coating preserved in patches on front and back. Flat-backed central knob with concave-sided moulding at base; facetted wings terminating in concave-sided mouldings and flat-ended knobs; slot for hinge at back; X-radiograph shows hingepin missing. Facetted D-sectioned arched bow with half disc above concave-sided moulding at base; slightly facetted, hollow-backed foot; return plate broken. Length 54 mm, width wings 27.5 mm. 2016: 202 Grid square 2: RF 22: Phase 5. Figure 162

14

Hairpin. Bone Type B1. Irregular head and shaft, just possibly unfinished, more likely just a very poorly made example. 2016: 202 Grid square 13: RF 228: Phase 5.

15

Hairpin. Bone Type B4. 26 mm long, broken. 2016: 202 Grid square 18: RF 224: Phase 5. Figure 163

16

Hairpin. Bone Type B. A very small, ‘squared’ head, slender example much as the type B4, but without the obvious cut facets. 19 mm long, broken. 2016: 202 Grid square 6: RF 230: Phase 5. Figure 163

17 Hairpin. Bone Type B. Small conical head, decorated with a small trellis design. 77 mm, broken. 2016: 202 Grid square 6: RF 133: Phase 5. Figure 164 18 Hairpin; shank fragment. Shale. Slightly faceted circular section, tapering to each broken end. Present length 31 mm, maximum section 4 mm. 2016: 311: RF 377: Phase 1.

Penannular brooch; complete. Iron. Terminals bent back along top of hoop; slightly curved pin wrapped around hoop and tapering to curved point. Pin now out of position; traces of minerally preserved

19 Segmented bead. Mid green cloudy glass. Four segments; one end collared, other sharp break. 159

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 162: Personal ornaments. Copper alloy (Nos 1-3) and iron (No. 4) brooches and bone hairpins (Nos 6-7). Scale 1:1

Diameter 4 mm, length 13.5 mm, perforation diameter 2 mm. BHN15: 302-1: RF 95: Phase 7. Figure 164

into side; D-sectioned with one side flattened. Diameter c. 52 mm, section 4 mm. 2016: 202 Grid square 6: RF 97: Phase 5. Figure 164

20

Long cylindrical bead. Translucent peacock glass. Drawn. Diameter 5 mm, length 21 mm, perforation 1.5 mm. 2016: 822: RF 421: Phase 4. Figure 164

21

Cable twist bracelet; approximately three-quarter extant. Copper alloy. Two strand right-hand twist; one end broken, at other one strand forms downward facing hook with second pressed

22 Cable twist bracelet; fragment. Copper alloy. Three strands, left-hand twist; both ends broken. Present length 55 mm, section 8 mm. 2016: 202 Grid square 2: RF 24: Phase 5. 23 Light bangle. Copper alloy. Rectangular-section, widest to wrist; hook and eye terminal; diagonal cross between transverse groove on either side 160

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

Figure 163: Personal ornaments. Bone hairpins. Scale 1:1

behind hook; tiny edge nicks along each edge. Now bent into sinuous strip. Straightened length 188 mm, section 5 x 1 mm. Original diameter c. 60 mm. 2016: 202 Grid square 9: RF 127: Phase 5. Figure 164 24

diameter c. 45 mm (c. 40% circumference extant), section 4 x 4 mm. 2016: 202 Grid square 9: RF 130: Phase 5. Figure 164 25 Bangle; fragment. Shale. Rounded rectangular section, inner face bevelled slightly to central ridge. Inner diameter c. 60 mm (c. 10%

Bangle; fragment. Shale. Rounded square section inner face bevelled slightly to central ridge. Inner 161

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

19

21

20 17 24

25

26

23

27

29

Figure 164: Personal ornaments. Bone hairpin( No. 17), glass beads (Nos 19-20), copper alloy bracelets (Nos 21, 23) and shale bangles (Nos 24-27, 29). Scale 1:1

circumference extant), section 6 x 4.5 mm. 2016: 323: RF 359: Phase 1. Figure 164

extant), section 6 x 5 mm. 2016: 202 Grid square 5: RF 54: Phase 5. Figure 164

26 Bangle; fragment. Shale. Rounded rectangular section with inner face bevelled slightly to central ridge. Inner diameter 70 mm (c. 15% circumference

27

162

Bangle; fragment. Shale. Tall D-section inner face bevelled slightly to central ridge. Inner diameter c. 75 mm (c. 20% circumference extant), section 9

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

x 9 mm. 2016: 202 Grid square 5: RF 61: Phase 5. Figure 164 28

D-shaped section. 73 mm long, but the bracelet has ‘sprung’ so it’s difficult to measure the original diameter. 2016: 202 Grid square 6: RF 314: Rubbish. Phase 5. Figure 165

Bangle; fragment. Shale. Tall D-section inner face bevelled slightly to central ridge; one side flaked off. Inner diameter c. 80 mm (c. 17% circumference extant), section 9 x c. 8 mm. 2016: 202 Grid square 6: RF 340: Phase 5.

33 Finger-ring; fragment. Jet. D-sectioned hoop fragment, inner face lightly faceted. Outer diameter c. 25 mm, section 3.5 x 5 mm. 2016: Unstratified: RF 435.

29 Bangle; fragment. Shale. Tall D-section, inner face slightly irregular and slightly convex with polishing marks and pitting. Inner diameter c. 85 mm (c. 15% circumference extant), section 11 x 13 mm. 2016: 200: RF 235: Unstratified. Figure 164 30

TOILET EQUIPMENT The only toilet equipment was two sets of simple tweezers (Nos 34-35). These are the typical RomanoBritish form and though they were found in Phase 7 contexts can be accepted as being of Roman date.

Bangle; fragment. Shale. Tall D-section inner face slightly convex and irregular. Small notches on alternating side. Outer diameter 70 mm (c. 11% circumference extant), section 6.5 x 3.5 mm. 2016: 432: RF 403: Phase 3. Figure 165

34 Tweezers. Copper alloy. Rectangular-sectioned strip bent in two forming loop, parallel arms with inturned straight-ended jaws. Length 56 mm, section 4 x 1 mm. 2016: 914: RF 400 Phase 7. Figure 165

31 Bangle, fragment. Jet. Tall rectangular section; paired notches producing diamond ridge along upper face, groove on either side. Inner diameter 40 mm (c. 20% circumference extant), section 6 x 3 mm. 2016: 230: RF 158: Phase 4. Figure 165 32

35 Tweezers. Copper alloy. Rectangular-sectioned strip bent in two forming loop and parallel arms, jaws broken off. Present length 40 mm, section 5 x 1 mm. 2015: 502-2: RF 65: Phase 7.

Bracelet, fragmentary with modern break. Bone. Simple acute line decoration, weakly executed.

Figure 165: Personal ornaments and toilet equipment. Shale (No. 30), jet (No. 31) and bone (No. 32) bangles and bracelets and copper alloy tweezers (No. 34). Scale 1:1

163

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 166: Textile equipment. Iron needles (Nos 36-37), bone needle case (No. 38) and shale spindle whorl (No. 39). Scale 1:1

means they can corrode and become unidentifiable very easily. Here the excellent preservation of the iron has meant that it has been possible to identify two. One (No. 36) came from the Phase 4 rubble spread to the west of Structure A and the other (No. 37) from the late Phase 5 dumping. Whilst iron needles would

TEXTILE EQUIPMENT The assemblage contains several items of textile equipment of a type not often recovered. Iron needles must have been common but are rarely identified even with excellent X-radiography given that their size 164

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

no doubt have been useful in textile working, it has been observed that metal needles in both copper alloy and iron have been found in dumps of leatherworking debris so cobblers and other craftsmen who made leather artefacts may have been the people who used them most (Cool 2002: 35). At Norton there is definitely evidence of leatherworking in the form of an awl (No. 86), so Nos 36-37 may be additional evidence for that craft here.

externally been worked flat and decorated with series of ring and dots in groups of two and one or simple pairs, although this pattern breaks down on one side. Length 113 mm, complete. 2016: 248: RF 209: Phase 4. Figure 166 39

Worked sheep metapodials such as No. 38 are a known late Roman form, usually identified as handles (e.g. Greep 1995:1141), They may be plain (e.g. Clarke 1979: fig. 83, 493), lightly (Greep 1995: fig. 502, 996) or more heavily decorated (Bushe-Fox 1928: pl. xix, 31) with ring and dot motives. Undecorated examples are obviously more difficult to interpret but may be identified by the suspension perforations and the sawn proximal end (e.g. Bluer and Brigham 2006: fig. 108).

HOUSEHOLD EQUIPMENT The items discussed in this category include several that would have been used in eating, drinking and food preparation (Nos 40-45), some that can broadly be described as household fittings (Nos 46-48), and one (No. 49) which, whilst having plausible functions within a household, setting might also have been used in industrial and craft processes.

Following work in northern France these pieces are more correctly identifiable as needles cases; a complete example from Moyencourt, France still retained its needles in situ (Theut and Morel 2013). A re-examination of the published example from Winchester noted above that still retained iron in situ demonstrated that this too contained a pair of iron objects, probably needles (Greep in prep). A third example retaining iron in situ in its central cavity, from Carlisle (Howard-Davies 2009: fig. 529 no. 3), was not available for study at the time of writing. The present example shows no signs of iron staining in the interior, but clearly it belongs to this group. There are over 30 examples from Britain of this form – on the available evidence they may be dated to the second half of the 4th to 5th centuries. This example came from the Phase 4 rubble spread to the east of kiln/ drier Structure C.

No. 40 from the late Phase 5 rubbish dumping is most likely to have been the handle of a spoon though it retains none of the diagnostic features associated with the bowl and the junction to that which would make the identification certain. However, the change in cross-section and the fact that the piece has been silvered or tinned would all be consistent with such an identification rather than it being, for example, a hair pin. Spoons with pear- and oval-shaped bowls and ones with the style of pinched bowl variously described as mandolin-, lute- or fig-shaped bowls can have a zone of decoration behind the junction of the type seen here (see for example Wheeler and Wheeler 1932: 85 nos 89 and 93, fig. 19). A later Roman date is most likely but the absence of the bowl and its junction makes a more precise identification impossible.

No. 39 is a shale spindle whorl from the late Phase 5 rubbish dumps. Spindle whorls in this material were a very late Roman development as they start to appear in the second half of the 4th century and continue in use into the 5th century (Booth et al. 2010: 274-275). No. 39 is thus good evidence for the late date of the dumping episode. 36

Spindle whorl; half. Shale. Cylindrical with slightly convex edge; one face sheared off. Upper edge has small nicks around extant face / side junction; two grooves around side; circular perforation with vertical facet-cuts from cutting. Diameter 36 mm, perforation diameter 7 mm, present thickness 8 mm. 2016: 202 Grid square 12: RF 220: Phase 5. Figure 166

Roman vessel glass was a very scarce find in this assemblage. In addition to the fragments catalogued below (Nos 41-44) there were four blue/green body fragments weighing in total 2.1 g, two colourless body fragments weighing in total 0.9 g and two greenish bubbly fragments weighing 0.8 g. With the exception of one colourless fragment, all came from late Phase 5 rubbish dumps. The blue/green glass can be dated in general to the 1st to 3rd centuries and so these fragments were residual. Colourless glass can normally be assigned broadly to the 2nd to 3rd centuries but here the one fragment from the dumps had the type of bubbles normally associated with 4th-century glass production and so was probably contemporary with them.

Needle. Iron. Circular-sectioned shank tapering to missing point; head flattening broken across oval perforation. Present length 50 mm, section 1.5 mm, 2016: 235: RF 333: Phase 4. Figure 166

37 Needle. Iron. Slender circular-sectioned shank tapering to point, upper end slightly flattened and broken. Present length 38 mm, section c. 1 mm. 2016: 206 Grid square 3: RD 318: Phase 5. Figure 166 38 Needle case. Bone. Sheep metatarsal with its proximal end removed, the distal intact, although worn and damaged and with two drilled holes, presumably for suspension. The central shaft has

Of the more diagnostic pieces No. 41 is a rim fragment of a blue/green storage bottle (Price and Cottam 1998: 165

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire 191-200). These were extremely common from the later 1st century and into the 3rd. Frequently it was the only type of vessel that the inhabitants of rural sites found any use for especially in the earlier part of the period. Given the likely dating of the Norton settlement such vessels would have, on the whole, passed from use by the time the Phase 2 occupation started, so it may have been associated originally with the earlier field system occupation.

so the recovery of No. 46 from a northern site with military associations is a useful corrective. It came from the Phase 4 rubble demolition spread around the kiln Structure C so possibly was associated with its use. Other items of furniture are probably implied by the short nails with large heads Nos 47-48, one of which was found in a Phase 1 context and the other in the late Period 5 rubbish dumping. There are normally interpreted being used when the head needed to be decorative and they would have been useful for upholstery and for fastening other coverings to wooden furniture (Manning 1985: 135 Type 7).

None of the other diagnostic pieces (Nos 42-44) can be assigned to precise type though the colour and nature of the glass point to a 4th-century date despite one of them coming from a Phase 1 context (No. 43). They do, however, attest to the use of glass tablewares in the vicinity during the 4th century.

Finally the leg and part of the supporting ring of an iron trivet (No. 49) was found in a primary fill of the Phase 2 ditch 397 together with late 3rd to early 4th-century pottery. Again this is an exceptional find and it is of no surprise that the best comparanda come from a hoard and a limes fort. One with a circular ring and three outsplayed feet was found in the Carlingwark Loch hoard (Piggott 1953: 86 no. C.73, fig. 10). The measurements cited there are clearly incorrect but can be measured from a scaled photograph (Curle 1932: fig. 20) and are very similar to those of No. 49. One with a triangular ring came from the fort at Feldberg and was also of a similar height (Jacobi 1937b: 31 Taf. III no. 32). They were clearly designed to keep a vessel just above the heat source and would have been ideal for certain types of cooking or for keeping food or drink warm. The one from Feldberg was found in the northern tower of the porta dextra of the fort and it is pleasant to think that soldiers on guard duty, in what could be a very cold area, might have had a source of hot food to hand. Equally though, various craft processes benefit from gentle heat and so No. 49 might have had a role in a workshop.

No. 45 consists of large parts of an iron ladle with a long handle. A fragmentary example with a shallower bowl was found in an undated pit during the excavations in the Malton vicus (Ottaway 1997: 131 no. 6, fig. 49) but these large implements rarely survive in an unbroken state unless the circumstances of their deposition are exceptional such as in a hoard or due to deliberate abandonment. No. 45 was found in the Phase 2 thick accumulation over the earlier field system which also included pottery of the 3rd to possibly late 4th century. Such ladles were often combined with a flesh hook at the other end and were multi-purpose cooking implements. This example is not complete and the outer broken end is twisted in a manner that is often associated with flesh-hooks (Manning 1985: 105). There is certainly the possibility that the missing end of this implement could have consisted of hooks. That part of the implement could be up to c. 100 mm long which would make the complete object c. 400 mm long. This would be comparable to the lengths of those found on the limes forts on the German frontier abandoned in the mid-3rd century such as the nine from Zugmantel (Jacobi 1937a: 101 no. 11, Taf. XIV no. 71) and the example from Buch which is explicitly described as being 38 cm long and which may itself not be complete as it only has one hook rather than the normal two or three (Herzog 1929: 15, Taf III no. 29). An example from a mid- to late 2nd-century waterlogged context with excellent preservation at Castle Street, Carlisle lacked most of its bowl, but if allowance is made for that, the complete item would have been in the region of 400 mm long (Padley 1991: 133 no. 333, fig. 107).

40 Spoon handle. Copper alloy with white metal. Circular-sectioned shank tapering to point at one end and becoming square-section near broken junction with bowl. Three transverse ribs at broken end with square-sectioned zone possibly diagonally ribbed or grooved (area obscured). Traces of white metal coating on part closest to broken junction. Present length 80 mm, maximum section 3 mm. 2016: 202 Grid square 13: RF 128: Phase 5. Figure 167 41 Bottle; rim and neck fragment. Blue/green. Rim bent out, up and in; cylindrical neck. Rim diameter 40 mm, present height 20 mm. Weight 5.7g. 2016: 235 Grid square 13: RF 332. Phase 4. Figure 167

The iron candlestick No. 46 is a less uncommon find. It was designed to be fixed into a wall. Eckardt (2002: 254) has termed them spike candlesticks and has noted that they have a predominantly domestic distribution as opposed to a religious one that other candlestick types can have. Where dateable they come from 3rdand 4th-century contexts and her survey revealed a predominantly southern and urban distribution. This though is likely to be due to biases in her sample and

42 Jug; neck fragment (?). Green-tinged colourless; some small bubbles. Straight side broken where bending out to rim or shoulder. Horizontal trail. Neck diameter c. 40 mm, neck thickness 2.5, present height 30 mm. Weight 3.6g. 2016: 202 Grid square 3: RF 327: Phase 5. Figure 167 166

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

Figure 167: Household equipment. Copper alloy spoon (No. 40). Scale 1:1. Glass vessels (Nos 41-43) and iron ladle (No. 45). Scale 1:2

43 Body fragment. Green-tinged colourless, small bubbles. straight side with curved trail. Dimensions 18 x 17 mm, wall thickness 1.5 mm. Weight 0.9g. 2016: 262 Grid square 3: RF 229: Phase 1. Figure 167 44

45 Ladle; seven major and several small fragments. Iron. Two joining cup fragments and five joining handle and edge of cup fragments; several small fragments of cup and one (?) additional handle fragment. Hemispherical cup thinning to base; square-sectioned horizontal tapering handle with one side of probably symmetrical flat plate at junction of handle and cup, outer extant end of handle shows left-handed twist in X-radiograph.

Bottle; shoulder fragment. Colourless with small bubbles. Broken at edge of neck and starting to bend over to shoulder. Dimensions 22 x 17 mm. Weight 1.7g. 2016: 303: RF 363: Phase 5. 167

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 168: Household equipment. Iron candlestick (No. 46), upholstery nail (No. 47). Scale 1:1. Iron trivet (No. 49). Scale 1:2

49

Present length 275 mm, diameter of cup c. 85 mm, depth cup c. 30 mm. 2016: 388: RF 366: Phase 2. Figure 167 46 Candlestick. Iron. Broken socket with solid base; tapering square-sectioned stem bent through 90˚. Present length of candlestick arm 64 mm, length of wall hook 55 mm, maximum current socket diameter 12 mm. 2016: 248 Grid square 12: RF 208: Phase 4. Figure 168

RECREATIONAL EQUIPMENT The only item in this category is a black glass counter (No. 50). These were commonest in the 1st and earlier 2nd century and become much rarer later in the 2nd century (Cool et al. 1995: 1555). This example is thus part of the residual element of the late Phase 5 rubbish dumping.

47 Upholstery nail. Iron. Circular head, chipped on one edge; square-sectioned short shank tapering to point. Length 20 mm, head diameter 22 mm, shank section 6 mm. 2016: 202 Grid square 15: RF 282: Phase 5. Figure 168 48

Trivet; fragment. Iron. Part of circular ring with one leg forged onto exterior of ring, base of leg broken. Present height 165 mm, external diameter c. 150 mm. 2016: 3008: RF 368: Phase 2. Figure 168

50 Counter. ‘Black’ glass. Plano-convex with pitted base. Diameter 15 mm, thickness 6 mm. 2016: 816: RF 422. Phase 5. Figure 169

Upholstery nail. Iron. Large flat oval head; shank tapering to broken tip. Head diameter 23 x 20 mm, present length 32 mm. 2016: 319: RF 779: Phase 1. 168

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

Figure 169: Recreation and writing equipment. Glass counter (No. 50) and iron styli (Nos 51-52). Scale 1:1

Norton styli, however, are all common forms that are covered by the system.

WRITING EQUIPMENT Iron styli for writing on wooden wax tablets were a common find on the site with seven examples being found. Of those from phased contexts four (Nos 53-54, 56-57) came from the late Phase 5 rubbish dumping and one (No. 55) came from a short-lived Phase 1 ground surface. Manning’s provisional typologies provide a basic framework to categorise them (1976: 35 fig. 10; 1985: 85 fig. 24), though as various authors have pointed out there is the potential to do more work on them as individual sites can provide variants not covered by the Manning system (Padley 1991: 132; Major 2002). The

The simple form with a simple block eraser and undifferentiated stem and point (Manning Type 1) is represented by the complete No. 51 and probably by the fragmentary No. 52 which just consists of the eraser. Nos 53-55 are examples of Type 3 with a differentiated point of smaller diameter than the stem and an expanded eraser. No. 55 differs from many examples of this type in having a much longer point that is normal. Nos 56 and 57 both belong to the catch-all Type 4 which is defined as being decorated: No. 56 having three non169

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 170: Writing equipment. Iron styli. Scale 1:1

ferrous strips inlaid whilst No. 57 has been forged to have a rib above the point. The semi-circular eraser and the positioning of the decoration on No. 56 have similarities with Major’s (2002) Group 4 which have been found in 3rd-century contexts. The distinct, and again very long, point sets No. 56 apart from that group.

Styli are intimately connected with literacy and numeracy. It would be tempting to connect Nos 5157 with the presence of the soldiers implied by the military equipment. The Roman army after all required its soldiers to be literate and was a great producer of documentation. The increasing routine scanning of ironwork via X-radiography is greatly increasing the number of iron styli recognised including at rural sites with no obvious military or administrative role. There is, therefore, no reason why their presence at Norton could not equally be the result of use by the non-military inhabitants. That said, there does appear to be a larger number of styli than might have been expected given the size of the area investigated even allowing for the good preservation of the iron and the large quantities of rubbish being dumped in Phase 5. To place the quantities in perspective the excavations at Elms Farm, Heybridge produced one of the larger assemblages of iron styli found from a settlement that had neither a

Major has noted (2002, 4) that the waisted form of eraser seen on Nos 53-55 has a tendency to be found on examples from 4th-century contexts which would agree with the contextual dating here. In general, however, it is not possible to suggest typological dates for the broad Manning types as all are found throughout the Roman period. Styli were a specialist production of the blacksmith with their circular-sectioned stems and, sometimes additional decoration. It would not be surprising if the variations seen are to a great extent due to local working practices rather than major chronological changes. 170

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

military or administrative role (Major 2015). Thirty were found but that was from a site that investigated 18 ha. The excavations at Castle Street, Carlisle took place before X-radiography was routine but the waterlogged conditions meant that iron survived in a recognisable state. There 13 iron styli were recovered in the military phases on a site that measured c. 220 m2 (Padley 1991: 133-137). This is a crude measure but does suggest that it may be the military presence in the area that is driving the number of styli at Norton.

54 Stylus. Iron. Eraser with shoulders, flaring out asymmetrically towards top; circular-sectioned stem expanding to sharp step to point with bevelled edge. Length 90 mm, maximum shaft section 4 mm, width eraser 6 mm. 2016: 202 Grid square 3-6: RF 106: Phase 5. Figure 170 55 Stylus. Iron. Waisted eraser with shoulders; circular-sectioned stem expanding to sharp step to long point. Length 139 mm, maximum shaft section 6.5 mm, width eraser 7 mm. 2016: 231 Grid square 16: RF 200: Phase 1. Figure 170

51 Stylus. Iron. Wedge-shaped eraser; thick circular-sectioned stem with point at end. 2016: Unstratified: RF 462. Figure 169 52

56 Stylus. Iron. Semi-circular eraser; circularsectioned stem expanding to sharp step to long point. Three non-ferrous strips inlaid above step. Length 118 mm, maximum shaft section 4 mm, width eraser 8 mm. 2016: 206 Grid square 3: RF 140: Phase 5. Figure 171

Stylus, eraser only. Iron. Slightly expanded eraser broken at circular-sectioned shank. Present length 16 mm, maximum eraser width 6 mm. 2015: 405: RF 178. Unphased. Figure 169

53 Stylus. Iron. Waisted eraser with shoulders; circular-sectioned stem expanding to sharp step to point. Length 112 mm, maximum shaft section 4.5 mm, width eraser 9 mm. 2016: 202 Grid square 5: RF 86: Phase 5. Figure 169

57

Stylus. Iron. Straight-edged eraser with shoulders; circular-sectioned stem expanding to sharp step to point. Two grooves forming rib on lower part of stem. Blade of eraser damaged. Length 119 mm, maximum shaft section 6 mm. 2016: 202 Grid square 8: RF 63: Phase 5. Figure 171

Figure 171: Writing equipment. Iron styli. Scale 1:1

171

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire during the Roman period (Manning 1985: 66-67). They are normally identified from the diagnostic links and a complete one as here is a rare find. The tubes that would have articulated with the side rings have been deliberately pulled apart to release the rings, perhaps indicating that there was some desire to re-use them. Perhaps originally the bridle had rich trappings and there was a desire to replace the bit. A very interesting aspect of No. 58 is its size. It has been corroded with the two links in a V-shape so measuring the part that

TRANSPORT EQUIPMENT A complete two-link snaffle bit was found in the Phase 5 rubbish dumping (No. 58) and a ring that could well have articulated with the side junctions of a snaffle bit came from a short-lived ground surface assigned to Phase 1 (No. 59). The suggestion of the function of the latter is made on the grounds of the size and the changing section that can be observed. Two-link snaffle bits were the commonest horse bit form in use

Figure 172: Transport equipment. Iron bits. Scale 1:1

172

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

would have been in the horse’s mouth cannot be done accurately but would have been in the vicinity of 8590 mm. This would translate into a modern bit of 3.25 to 3.5 inches. This is the size that is appropriate for a small pony such as a Shetland. It is certainly smaller than the 4.5 inch bit proposed for the ponies likely to have been used in the chariot races at the Colchester circus (Crummy 2020). The pony which would have been bridled with this bit would certainly not have been large enough to be used as a riding animal other than by a child. Perhaps more likely it might have been a carriage animal and raises the intriguing possibility of chariot races in the vicinity. 58

median of 60 mm and an interquartile range of 48 to 72 mm. This is a normal pattern with the interquartile range effectively being that which is very suitable for uses such as fixing timber cladding. The good condition of the ironwork meant that it was apparent that some of the rarer nail types were present as well. These tend to have a head that is no wider than the shank and were used in more specialised joinery. The triangular headed Type 2 was especially useful when it was desired to use nails that could be driven in parallel to the grain and so not be visually obvious (Manning 1985: 135). Type 3 with a T-shaped head and Type 4 with an L-shaped profile must also have been for specialised purposes. Manning cautions rightly that it can be very difficult to be sure that Types 3 and 4 are correctly identified as normally the condition of ironwork even when studied with X-radiography does not allow secure identification. Here the possible candidates for Types 2 to 4 were re-inspected after the assessment stage and only those that could be securely identified to the types have been included in Table 21, with the rest being relegated to Type 1. As can be seen they are very rare but they are present. One of the Type 2 nails (No. 64) was a substantial nail possibly used in structural timberwork. The others (Nos 65-70) tend to be smaller and may have been used in internal joinery and furniture.

Two-link snaffle bit. Iron. Each link consisting of square-sectioned bar with one end curved into ring, other end has tubular loop at 90˚ to plane of ring. Tubular loops pulled open. Two links articulated via rings and corroded into a V-shape. Length of links c. 60 mm, diameter of rings 15 mm, length tubes 24 mm. 2016: 202 Grid square 9: RF 186: Phase 5. Figure 172

59 Ring. Iron. Circular-sectioned ring thickened at one point with rounded rectangular section. Diameter 63 mm, section 5 mm, 8 x 9 mm. 2016: 231 Grid square 15: RF 188: Phase 1. Figure 172 BUILDING – STRUCTURAL IRONWORK

60 Joiners’ dog. Iron. Rectangular-section bar tapering to either square-sectioned bent down at 90˚. Minerally preserved organics at various points which appear incidental. Length 50 mm, depth 27 mm, maximum section 10 x 4 mm. 2016: 206 Grid square 2: RF 313: Phase 5. Figure 173

A relatively small number of items of structural ironwork other than nails were recovered from the Roman contexts. There were two joiners’ dogs used for joining two pieces of timber (Nos 60-61), an angle bracket (No. 62) and an item which may have been a small T-clamp (No. 63). As on any Roman site nails formed a large part of the iron assemblage and the ones from the Roman contexts are summarised according to Manning’s (1985) typology in Table 21. He defined six types that were likely to be used in joinery. By far the commonest always was the basic nail with a flat expanded head used in everyday joinery. This is the case at Norton as well. A little under half of the Type 1 nails (112 examples) were complete and these had a length range of 27 to 100 mm with a Phase 1 2 1 10 2 22 1 3 17 4 32 5 155 2 Total 236 3

3

1 2 3

4

1 1

Total 10 23 18 33 159 243

Table 21. Types of iron nails present according to Roman phases (quantified by heads and according to Manning (1985) types)

61

Joiners’ dog. Iron. Square-sectioned bar tapering to ends which are bent through 90°. One arm broken. Length 30 mm, depth 37 mm, maximum section 5 mm. 2016: 409: RF 393: Phase 2.

62

Angle bracket. Iron. Strap with straight ends and slightly dished face, bent through 90˚. One arm has circular perforation. Maximum length 50 mm, width 34 mm, perforation diameter 8 mm. 2016: 202 Grid square 2: RF 33: Phase 5. Figure 173

63

T-clamp? Iron. Square-sectioned tapering broken shank; short cross bar with broken ends. Present length 78 mm, current width cross bar c. 20 mm. 2016: 202 Grid square 8: RF 557: Phase 5.

64

Nail. Iron. Manning Type 2. Shank broken. Present length 65 mm; head section 21 x 7 mm. 2016: 202 Grid square 6: RF 172: Phase 5. Figure 173

65

Nail. Iron. Manning Type 2. Complete, tip bent to one side. Length 75 mm; head section 7 x 4 mm. 2016: 412: RF 389: Phase 2. Figure 173

66 Nail. Iron. Manning Type 2. Shank tapering to wedge point in same plane as head length. length 173

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 173: Structural ironwork. Joiners dog (No. 60), angle bracket (No. 62) and nails (Nos 64-70). Scale 1:1

42 mm; head section 7 x 5 mm. 2016: 202 Grid square 9: RF 529: Phase 5. Figure 173 67

minerally preserved organics indicating that originally they had wooden handles and No. 71 may well have been in a leather sheath as traces of leather were found on the back of the knife. The other knife from this rubbish dump (No. 73), an example of a Type 13, may also have been disposed of in a virtually complete state as currently it only lacks its tip.

Nail. Iron. Manning Type 3. Short shank tapering to wedge point at 90˚ to plane of long axis of head. Length 25 mm; head section 4 x 6. 2016: 202 Grid square 6: RF 520: Phase 5. Figure 173

68 Nail. Iron. Manning Type 3 variant. Wide head without distinct junction with shank which tapers to wedge point at 90˚ to plane of long axis of head. Length 55 mm; head section 11 x 3 mm. 2016: 202 Grid square 6: RF 520: Phase 5. 69

By contrast the knives and blade fragments from the other contexts are fragmentary. They include two examples of Type 16 (Nos 74-75), and one example of a Type 18b (No. 76) together with a blade fragment (No. 77) which might have come from the same type. No. 74 came from a Phase 1 ditch fill and No. 76 came from a short-lived ground surface of the same phase whilst No. 77 was recovered from Phase 4 layers over Structure G. It does suggest that at least some of the material in the Phase 5 rubbish dumps was not the result of standard disposal of broken and unusable items.

Nail. Iron. Manning Type 3 variant. Lower part of head sloping into shank. Shank tapering to wedge point in same plane as long axis of head. Length 38 mm; head section 9 x 3 mm. 2016: 202 Grid square 14: RF 213: Phase 5. Figure 173

70 Nail. Iron. Manning Type 4. Length 54 mm, head section 7 x 4 mm. 2016: 204 Grid square 2: RF 514: Phase 4. Figure 173

No. 78 also came from the Phase 5 rubbish dumping though from a different part of the site than the knives Nos 71-73. It too is complete and has a tang which is substantial enough to have acted as a handle. The blade has a curved edge but only the tip is sharp, the back part is rounded and blunt. This bladed tool cannot have acted as knife in the normal sense of the word and its shape does not fall within the Manning typology. It is tempting to suggest that this was a creaser used in leatherworking. Such tools are better known from the medieval period but even then are not often recognised (Goodall 2011: 68). They were used to mark guidelines for sewing and to compress edges to make them stronger. They appear even rarer in the Roman period but given the amount of leatherworking at the time, such a tool might be expected. This is a speculative suggestion as was the

CRAFT AND INDUSTRY All of the bladed tools have been put in this section though the knives are types which would have been general purpose ones, useful both in the house and in the workshop. Most are common types easily accommodated within the standard typology of Roman knives (Manning 1985: 108-119) and in use throughout the Roman period. The two examples of Type 12a (Nos 71-72) were both found in the Phase 5 rubbish dumping and are noteworthy because both must have been disposed of when they were still usable. Both preserved traces of 174

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

suggestion that another blunt knife-like tool from a brine pit at Nantwich was also a creaser (Cool 2012: 91 no. 4.9, fig. 7.5). At Norton it is clear that leatherworking was being carried out in the settlement generating the rubbish on this site because of the presence of the awl No. 86 and possibly the iron needle No. 37, so other leatherworking tools would not be surprising.

It is very tempting to reflect on RIB I.712 (Collingwood and Wright 1995). This building stone has an ansate panel on which is inscribed (in translation) ‘Good luck to the Genius of this place. Young slave, use to your good fortune this goldsmith’s shop’. This remarkable and unique stone tells us that we have a shop making and selling gold jewellery. It is normally stated to be from Malton but it was actually found in 1814 in Norton when St Nicholas’ church was being rebuilt. This set of tongs would certainly have been a useful tool for the young slave had gold been refined in the shop as well it might have been.

One high temperature industry can be identified with certainty as blacksmithing is indicated by two iron punches (Nos 79-80) which have the typical battered tops associated with being regularly struck (Manning 1985: 9). No. 80 is small and would probably have been held by a wire loop. The use of the tongs No. 81 in a high temperature industry is also implied as clearly their length (415 mm) indicates the need to keep the operator’s hands away from a heat source. Blacksmiths used a variety of sizes of tongs to hold the hot metal whilst it was being worked (see Manning 1985: 6-8). Two basic forms are regularly encountered, both have two parts hinged together with long handles and relatively short jaws. On the commonest the jaws are bowed and meet on the same plane as the handles with some having straight jaws like pincers. In the second the ends of the slightly bowed jaws are bent up at approximately 90˚ before bending to meet. The two types are usefully illustrated together in the Newstead report where they were found in a pit with other tools and stock of a blacksmith (Curle 1911: 285-286, pl. LXIII). No. 81 clearly does not belong to this range of tools. The two parts are hinged at one end and the jaws with diamondshaped plates are at the other. One of the arms has a semi-circular curve of an appropriate size to allow the insertion of a finger to allow that arm to be raised and the jaws manipulated.

The low temperature crafts represented are carpentry and leatherworking. The former is represented by three chisels (Nos 82-84) and a possible sawblade fragment (No. 85). One of the chisels is complete enough to identify it as a paring chisel of Manning (1985: 21) Type 3. These were finishing tools. Leatherworking is represented by the awl No. 86 which is an example of a Manning (1985: 40) Type 4b, though, as already discussed the needles Nos 36-37 and the bladed tool No. 78 may also have been used by leatherworkers. If the needlecase No. 38 regularly contained iron needles then this too might have been part of leather working equipment. With the exception of No. 72 all of the stratified tools identified were from the late Phase 5 rubbish dumping suggesting these activities were not being practised on the site itself. In addition to the iron knives and tools, there was one fragment from a socketed tool preserving traces of an organic handle (No. 87) and one plain antler handle (No. 88) from a Phase 4 rubble deposit. The latter type is common throughout the Roman period and was used to haft a variety of implements. For an example from nearby Catterick see Wilson, 2002: fig. 323, 39.

One possibility is that No. 81 was used for moving items in and out of a furnace or oven. Rehren (1999) has studied a group of thin-walled small crucibles from a pit at Xanten dated to the beginning of the 1st century AD. He has shown that these would have been used to produce brass by the cementation process. Using them would have needed an indirect heat chamber at careful controlled temperature of at least 900˚C. Given that indirect heat chambers are slow to heat and cool, continuous production would have been sensible and this would have involved placing the crucibles in the chamber and removing them whilst the heat source was running. He noted that the lids, which would have been sealed to the crucibles during the process, had pointed tops which would have enabled this movement as they could have been grasped by pincers. No. 81 would be well adapted for such a use as the expanded jaws would be ideal to grasp crucibles. Brass is not the only metal for which small scale cementation like this is used. Gold too is refined in lidded crucibles in this way (see for example Marsden 1975, 100, fig. 46) and these would also have needed long-handled tongs or pincers for their manipulation.

Four fragments of worked, waste material were recovered, presumably indicative of bone working on or near the site (Nos 89-90 from the Phase 5 rubbish dump and two unphased fragments, RFs 271-272). The quantities are so small that it was clearly not a major industry. Knives and Bladed tools 71 Knife. Iron. Deep blade with straight back continuing line of tang; edge curving up gently to point. Tang slopes up slightly from blade. Minerally preserved organic – possible traces of leather on back of knife; wood fibres from a ring porous hardwood on handle. Length 129 mm, blade length 95 mm, depth 30 mm. 2016: 202 Grid square 13: RF 126: Phase 5. Figure 174 72 Knife. Iron. Deep blade with straight back continuing line of tang and dipping down slightly near tip; edge curving up gently to point. Edge damaged in places. Minerally preserved organics 175

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 174: Iron knives (Nos 71-76), bladed tool (No. 78) and punch (No. 79). Scale 1:2

– a thin layer of cells with wood structure on tang. Length 120 mm, blade length c. 80 mm, depth 28 mm. 2016: 202 Grid square 13: RF 134: Phase 5. Figure 174

Knife. Iron. Wide tang; straight back stepped up from tang; slightly curved, damaged edge dropped down from tang; blade broken and end missing. Present length 106 mm, depth 20 mm. 2015: 430: RF 264: Cleaning. Figure 174

73 Knife. Iron. Blade with straight back continuing line of sloping tang and dropping down to (missing) point; edge stepped down from tang with straight line towards tip. Present length 73 mm, length of blade 34+ mm, depth blade 12 mm. 2016: 202 Grid square 12: RF 217: Phase 5. Figure 174

76

74 Knife. Iron. Wide strip-like tang; straight back stepped up from tang; slightly curved edge dropped down from tang; blade broken and end missing. Present length 78 mm, depth 19 mm. 2015: 304: RF 128: Phase 1. Figure 17475

Knife blade. Iron. Broken at base of rectangularsectioned tang; Blade edge rises slightly to straight back, then drops down to point; edge steps down more deeply, runs parallel with edge and then curves up to point. Present length 104 mm, length blade 98 mm, depth blade 34 mm. 2016: 231 Grid square 15: RF 722: Phase 1. Figure 174

77 Blade fragment. Iron. Very slightly curved back; more pronouncedly curved back. Present length 60 mm, depth 19 mm. 2015: 416: RF 203: Phase 3.

176

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

78 Bladed tool. Iron. Rectangular-sectioned tang/ handle; thick blade stepped down at an angle and then curving up to point; back drops down slightly from line of tang, then continues straight before curving up to point. Edge rounded at back, sharp near point. Point chipped. Length 110 mm, length blade 50 mm, handle section 9 x 7 mm. 2016: 835: RF 448: Phase 5. Figure 174

broken upper end; blade formed by bevel from one face. Present length 54 mm, maximum section 6 mm. 2016: 206 Grid square 3: RF 550: Phase 5. Figure 176 84 Chisel (?). Iron. Rectangular-sectioned bar expanding in thickness then constricting; both ends wedge-shaped and one probably a chisel blade. Length 125 mm, maximum section 12 x 7 mm. 2016: 446: RF 398: Phase 3. Figure 176

High Temperature Industry Tools

85

79 Punch. Iron. Rectangular-sectioned stem with slightly concave wide faces, lower part tapering symmetrically to wedge-shaped blade; head bevelled on one side and showing battering. Length 89 mm, section 17 x 14 mm. 2016: spoil: RF 716: unstratified. Figure 174

Saw blade fragment (?). Iron. Narrow blade; four irregularly cut teeth extant, not set. Present length 22 mm, width 8 mm. 2016: 206 Grid square 13: RF 131: Phase 5. Figure 176

86 Awl. Iron. Square-sectioned tapering tang and stem. Length 108 mm, maximum section 5.5 mm. 2016: 202 Grid square 6: RF 137: Phase 5. Figure 176

80 Punch. Iron. Battered, damaged head with constriction below; square-sectioned tapering stem with wedge-shaped end. Length 45 mm, maximum section 5 mm. 2016: 206 Grid square 3: RF 549: Phase 5. 81 Tongs, complete. Iron. Two square-sectioned bars expanding and thinning at head into semi-circular expansions with are overlain and riveted together, the head maintaining the same width as the bars. Lower parts of bars expand to become rectangular-sectioned with tips formed into slightly expanded diamond shapes. Upper arm has semi-circular curve close to head. Diamond tip of lower arm better formed and slightly larger than that of upper arm. Length 415 mm, upper arm section 6 mm, lower arm section 14 x 5 mm. 2016: 206 grid 13: RF 124: Phase 5. Figure 175 Low Temperature Industry Tools 82 Paring chisel (Manning 1985: Type 3). Iron. Square-sectioned block head expanding slightly and then more sharply to slender broken stem. Present length 97 mm, maximum section 17 mm. 2016: 202 Grid square 5: RF 84: Phase 5. Figure 176

81

83 Chisel. Iron. Square-sectioned stem expanding slightly below

Figure 175: Iron tongs. Scale (drawing) 1:3

177

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 176: Iron chisels (Nos 82-84) saw blade (No. 85), awl (No. 86) and handle socket (No. 87) and antler handle (No. 88). Scale 1:2 (Nos 82, 84, 86, 88), 1:1 (Nos 83, 85, 87)

indicates that this has been used to haft an iron object. Present length 96 mm. 2016: 407: RF 873: Phase 4. Figure 176

Handles 87 Socket. Iron. Open with perforation; broken where widens to head. Minerally preserved organic within the socket. Present length 42 mm, diameter 14 x 12 mm. 2016: 202 Grid square 8: RF 169: Phase 5. Figure 176

Bone Working Waste 89a Irregular ‘rod’, rectangular cross-section knife cut surfaces 106 mm long, complete. Probably an unfinished pin or needle. 2016: 202 Grid square 15: RF 163: Phase 5.

88 Handle, fragment. Antler. Two fragments of adjoining curving tine. The central cavity has been hollowed and ferrous staining probably 178

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

89b Section of red deer antler time, sawn both ends, one at an acute angle. Clearly a waste product. 75 mm long. 2016: 202 Grid square 14: RF 145: Phase 5.

forward to form handle; bit bent through 90˚ with rectangular hole. Length 76 mm, maximum section 15 x 2.5 mm. 2016: 202 Grid square 6: RF 100: Phase 5. Figure 177

FASTENERS AND FITTINGS

94 Padlock key. Iron. Rectangular-sectioned stem expanding from bow to ward; bow rolled forward to form handle; bit bent through 90˚ with rectangular hole and two projecting teeth. Length 75 mm, maximum section 12 x 2 mm. 2016: 202 Grid square 14: RF 142: Phase 5. Figure 177

The assemblage contains five keys which can be identified with certainty. Three (Nos 91-93) are L-shaped lift keys designed to be used with simple tumbler locks (see Manning 1985: 90). The other two (Nos 9495) would have been used with barb-spring padlocks (Manning 1985: 96-97). No. 96 may have been the terminal of a key for a slide or lever lock (Manning 1985: 92-95) but the identification is not secure. The number of keys suggests a concern with keeping possessions and buildings secure. Even though the locks for the lift keys were quite simple and presumably easily picked, if the desire had just been to keep doors closed a simple latch lifter would have been sufficient. Most of the keys are from the Phase 5 rubbish dumping but the desire for security is possibly seen earlier on the site as one of the lift keys (No. 92) came from a Phase 4 rubble spread possibly associated with Structure B, though the pottery would suggest it was broadly contemporary with the Phase 5 dumping.

95

Key terminal (?). Iron. Loop with square outline; rectangular-sectioned broken stem. Present length 44 mm, loop 24 x 22 mm, stem section. 9 x 5 mm. 2015: 502: RF 252: Phase 3. Figure 177

Other Fittings 96 Double-spiked loop. Copper alloy. Circularsectioned bar becoming square-sectioned and tapering to ends; bent to form loop with parallel arms, one arm broken. Length 41 mm, section at loop 2 mm. 2016: 215 Grid square 6: RF 114: Phase 4. Figure 177

The other items in this section include the normal range of items such as box-fittings (Nos 97-100), a stud (No. 101), ferrules (Nos 102-04), spikes (Nos 105-07) and nailed plates and straps (Nos 108-14). Unlike the tools and keys which predominantly came from the Phase 5 rubbish dumping, this material shows a wider range of contexts which might support the idea that the late dumping may be coming from quite a specialised source. Keys

97

Box or chest fitting. Iron. Plate with one expanded rounded end and other an expanded broken oval shape; both expanded ends have perforation. Present length 103 mm, central bar section 10 x 2.5 mm. 2016: 258 Grid square 16: RF 215: Phase 3. Figure 177

98

Box-fitting? Copper alloy. Rectangular sheet strap; one end broken, other rounded and now chipped; one long side broken along 90˚ angle. Perforation in rounded end. Present length 33, width 18 mm. 2016: 202 Grid square 5: RF 56: Phase 5.

99 Box-fitting (?) Iron. Slightly bent square sectioned bar thinning to one disc end that appears perforated; other end broken. Minerally preserved organics – parts of transverse section of a ring porous hardwood at disc end. Present length 215 mm, section 5 mm, disc diameter 18 mm. 2016: 206 Grid square 13: RF 161: Phase 5.

90 L-shaped lift key. Iron. Square-sectioned stem becoming rectangular-sectioned at bow and ward; bow rolled forward in loop with end rolled back at base; ward with two projecting teeth. Length 120 mm, maximum section 7 x 7 mm. 2016: 202 Grid square 14: RF 176: Phase 5. Figure 177 91 L-shaped lift key. Iron. Square-sectioned stem narrowing over lower part and becoming rectangular-sectioned at bow and ward; bow rolled forward in loop; ward with three projecting teeth. Length 130 mm, maximum section 9 x 9 mm. 2016: 215 Grid square 6: RF 132: Phase 4. Figure 177

100 Stud. Copper alloy. Half of sheet circular head retaining perforation. Diameter 26 mm. 2016: 206 Grid square 2: RF 312: Phase 5. 101 Ferrule. Iron. Hollow socket, with oval crosssection and solid square-sectioned point. Length 55 mm, diameter 18 x 14 mm. 2016: 202 Grid square 3: RF 36: Phase 5. Figure 177

92 L-shaped lift key. Iron. Square-sectioned stem becoming rectangular-sectioned at bow and ward; bow rolled forward into open loop; ward lacking teeth. Length 115 mm, maximum section 5 x 5 mm. 2016: 206 Grid square 13: RF 151: Phase 5.

102 Ferrule. Iron. Plate wrapped into circularsectioned hollow cone; traces of perforation on opposite side to junction. Length 80 mm, maximum section 25 mm. 2016: 388: RF 370: Phase 2. Figure 177

93 Padlock key. Iron. Rectangular-sectioned stem expanding slightly from bow to ward; bow rolled 179

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 177: Fittings. Iron key (Nos 90-95), box fitting (No. 97), ferrules (Nos 101-102), nailed socket (No. 103), loop-headed spike (No. 104) and copper alloy double-spiked loop (No. 96). Scale 1:2 (Nos 90, 91, 93-95, 97, 101-104), 1:1 (No. 96)

103 Nailed socket. Iron. Plate wrapped into hollow cone with open junction and rectangular tang curved sidewise; nail shank driven through diagonally on opposite side from junction. Length 64 mm, maximum section 13 mm, length of nail

shank 24 mm. 2016: 202 Grid square 12: RF 216: Phase 5. Figure 177 104 Loop-headed spike. Iron. Square-sectioned bar becoming rectangular-sectioned as it tapers to chipped point; other end curved over to form 180

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

open loop. Length 133 mm, maximum section 7 x 7 mm. 2016: 253 Grid square 15: RF 211: Phase 3. Figure 177

agricultural activity. The only one that can be identified (No. 115) is an iron ox goad of Rees (1979: 76) Type 1 which is the commonest type of Romano-British goad. It was found in a short-lived ground surface assigned to Phase 1.

105 Spike. Iron. Square-sectioned bar tapering to point. Length 165 mm, maximum section 7 x 7 mm. 2016: 258 Grid square 16: RF 214: Phase 3.

Some have argued that these are not agricultural tools but rather nibs for ink pens. In the light of that it is interesting to note that the context that produced the goad also produced a stylus (No. 55). The suggestion that goads were actually pens has not found widespread favour (see discussion Eckardt 2018, 32), and an agricultural function is generally preferred. The ground surface was assigned a number of context numbers and as a whole produced a items that would have served a variety of functions including the possible bridle ring No. 59, a knife No. 73 and an armour scale No. 117 so the association here is probably fortuitous.

106 Spike. Iron. Square-sectioned bar tapering to wedge-shaped end. Length 134 mm, maximum section 9 x 9 mm. 2016: 303: RF 712: Phase 5. 107 Spike. Iron. Square-sectioned bar tapering to point and now bent. Length (straightened) 124 mm, maximum section 5 x 5 mm. 2016: 202 Grid square 6: RF 278: Phase 5. 108 Nailed collar. Twisted copper alloy strip with circular perforation now infilled with iron corrosion and possibly originally nailed. Length (flattened) 150 mm, width 33 mm, perforation diameter 12 mm. 2016: 313: RF 702: Phase 3.

115 Ox goad. Iron. Rectangular-sectioned bar wound to one and a half times to form hollow cylinder; bar stepped in to form square-sectioned point. Length 45 mm, diameter - inner c. 10 mm, outer 15 mm. 2016: 231 Grid square 15: RF 587: Phase 1. Figure 178

109 Riveted plate. Iron. Rectangular strip with one straight and one broken end, retaining long rivet shank through circular perforation. Present length 90 mm, width 20 mm, perforation diameter 2.5 mm, rivet length 15 mm. 2016: 202 Grid square 1: RF 11: Phase 5.

MILITARY EQUIPMENT

110 Nailed strap. Iron. Rectangular strap with two rectangular perforations retaining remains of nail shanks. Minerally preserved organics could not be identified. Present length 92 mm, width 24 mm. 2016: 205 Grid square 13: RF 338: Phase 3.

There is a range of military equipment from the site including items from armour (Nos 116-119), strap fittings (Nos 121-122) and weapons (Nos 124). The bulk of the items came from the Phase 5 rubbish dumping. The other contexts will be noted as the material is discussed.

111 Riveted strap. Iron. Narrow strap retaining one narrow rivet shank projecting from perforation. Present length 66 mm, width 8 mm. 2015: 302-1: RF 87: Phase 3.

Scale armour (lorica squamata) is represented by three armour scales (Nos 116-118) which belong to the type that have holes on each side and which were wired to each other. This was the type that came from excavations in the Malton vicus which included two examples still wired together (Mitchelson 1966: 253, fig. 19 nos 24-25). This type resulted in a more rigid cuirass and was introduced in the Antonine period (Bishop and Coulston 1993: 117; 2006: 139-140). The only example not from the Phase 5 rubbish dumping was No. 117 from a short-lived ground surface assigned to Phase 1. The only possible item that may have come from lorica segmentata is No. 119 which may have been part a buckle plate from a strap fitting (see for example Bishop 2002: 37-41).

112 Nailed strap. Iron. Terminal fragment retaining one large square hole for (missing nail). Present length 35 mm, width 27 mm. 2016: 205 Grid square 3: RF 257: Phase 3. 113 Riveted plate Iron. Fragment retaining rivet with circular shank and square head. Dimensions 25 x 25 mm, head dimensions c. 7 mm. 2016: 206: 13: RF 592: Phase 5. 114 Binding. Iron. Rectangular plate with one long broken edge bent over. Present length 31 mm, maximum depth 13 mm, thickness of bound material 2 mm. 2016: 231 Grid square 16: RF 568: Phase 1.

The small stud No. 120 belongs to a series whose heads are decorated with portrait busts which are similar to those seen on coinage. They were originally studied by Ulbert and Feugère (1985) produced an updated list of the number then known. No. 120 belongs to Type 7a (Feugère 1985: 126-128, 136, fig. 2). The series starts in the Flavian period but the somewhat crude depictions of the 7a studs have their greatest similarity with the radiate coinage of the 3rd century. They are generally thought to

AGRICULTURAL EQUIPMENT On this site with its early history as a field system and where iron survives well, often in the form of large recognisable artefacts, there is an interesting scarcity of the tools that would have been associated with 181

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

115

119 116

117

121

118

120

Figure 178: Agricultural and military equipment. Iron ox goad (No. 115), copper alloy lorica squamata scales (Nos 116-118), lorica segmentata fitting (No. 119), stud (No. 120) and phallic pendant (No. 121). Scale 1:2 (No. 115); 1:1 Nos 116-121; 2:1 details of No. 120

have decorated military belts. No. 120 is a relatively small example with a diameter of 16 mm. Those from London (Wheeler 1930: fig. 37 no. 2) and Canterbury (Blockley et al. 1995: fig. 422 no. 254) which are very similar to the Norton example have diameters of c. 23 mm.

both civilian and military communities, but this type is frequently found on military sites (Oldenstein 1977: 158-160; Allason-Jones and Miket 1984:186 no. 3.586 with references) and so it seems reasonable to regard No. 121 as an item of military equipment. Another phallic pendant, though this time straight with only the glans well-depicted came from the vicus excavations in the vicinity of the rampart and was part of the debris from an Antoine fire (Lloyd-Morgan 1997: 133 no. 2, fig. 50).

The phallic pendant No. 121 from a modern context is typical of the type where a well-depicted phallus with glans and testes is suspended horizontally. Within a Roman milieu this was a common good-luck symbol to ward off the evil eye. As such they were used by 182

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

Nos 122-123 are typical examples of catapult bolt-heads with square-sectioned points and circular-sectioned sockets (Manning 1985: 170-171, Type 1). One example (No. 123) came from Phase 3 rubble spread. The arrowhead No. 124 is a less typical Roman form. It has the form of Roman lance heads but is too small to be identified as such (see Manning 1985: fig. 33). Roman arrowheads, by contrast, tend to be trilobite and tanged (Bishop and Coulston 1993: fig. 43 nos 5-9, fig. 74 nos 8-11; 2006: figs 46 and 81) whereas socketed arrowheads with leaf-shaped blades such as No. 124 are more typical of the medieval period (e.g. Goodall 1990: 1071-1072). This example comes from the Phase 5 rubbish dumping which suggests that it is indeed of Roman date. The recovery of a similar one from the Antonine burning episode in the Malton vicus which also produced the phallic pendant would seem to confirm this (Ottaway 1997: 131 no. 2, fig. 49).

121 Phallic pendant. Copper alloy. Well-moulded phallus with semi-circular suspension loop at back. Length 45 mm, width 17 mm, depth 16 mm. 2016: 904: RF 401: Phase 7. Figure 178 122 Bolt head. Iron. Circular-sectioned socket tapering to square-sectioned projectile point, point and upper end damaged. Present length 100 mm, maximum socket diameter 13 mm. 2016: 202 Grid square 2: RF 283: Phase 5. Figure 179 123 Bolt head. Iron. Circular-sectioned socket tapering to square-sectioned projectile point, point damaged. Present length 100 mm, maximum socket diameter 15 mm. 2016: 253 Grid square 9: RF 339: Phase 3. Figure 179 124 Arrowhead. Iron. Leaf-shaped blade with slight central thickening, point damaged; closed socket with small perforation opposite the join. Length 75 mm, blade - length 41 mm, width 29 mm, diameter of socket 8 mm. 2016: 206 Grid square 3: RF 39: Phase 5. Figure 179

116 Lorica squamata scale. Copper alloy. Rectangular with pointed lower edge; four pairs of holes placed vertically on each edge; lowest one cuts edge of scale. One side broken across junction of holes and bent to one side. Dimensions 30 x 17 mm, thickness 0.5 mm. 2016: 202 Grid square 8: RF 66: Phase 5. Figure 178

MISCELLANEOUS This section contains the items which either do not have an obvious function such as No. 125 or might have had multiple functions such as the sheep metapodia (No. 126) and rings (Nos 127-131). The opportunity has also been taken to illustrate various broken pieces it has not been possible to identify but which come from Roman contexts (Nos 132-136). No. 137 has also been included as, though it appears to be of relatively modern date, it was found in a Phase 1 context.

117 Lorica squamata scale. Copper alloy. Rectangular probably with pointed lower edge; only one long edge intact others broken; remains of four pairs of holes placed vertically on each edge, lower edge perforations placed off-centre. Dimensions 27+ x 18 mm, thickness 0.5 mm. 2016: 247 Grid square 13: RF 331: Phase 1. Figure 178

The shale pendant No. 125 is one of the most unusual pieces found in the Phase 5 rubbish dump. It is a large rectangular block with a triangular vertical section with curved front face and a small block on the upper edge with a small perforation going from side to side. I only know of two similar pieces. One came from South Shields and is closely similar in shape and size (AllasonJones and Miket 1984: 318 no. 7.169). It has the very fine edge nicking seen on No. 125 but the front decoration consists of four vertical grooves dividing it into five panels. This piece has no contextual details. The second came from the excavations at the villa at Rockbourne near Fordingbridge, Hampshire (Morley Hewitt 1963: plate VII (B) and page opposite; 1969: 11 plate XVI (A)). This is the same size and general shape as the Norton and South Shields examples though, as the illustrations do not have a vertical section, it is unclear whether the front face is curved or flat. Decoration in the form of cut triangles in a chip-carved pattern on this piece is concentrated on the upper face and upper edge of the front face with a cross-hatched pattern over the suspension block. It too has no useful contextual details being described as having been found ‘on the wall’ between two rooms. The villa clearly continued in use into the 5th century judged by the late coinage found.

118 Lorica squamata scale. Copper alloy. Rectangular with rounded lower edge; vertical pair of holes on long sides, large hole centrally on upper edge. Upper edge and one side broken. Dimensions 28 x 17 mm, thickness 0.5 mm. 2016: 202 Grid square 5: RF 90: Phase 5. Figure 178 119 Lorica segmentata strap fitting (?). Copper alloy. Rectangular sheet with broken edges, one short edge broken across hinge; Circular perforation punched through from upper face. Present dimensions 28 x 14 mm, thickness 0.5 mm. 2016: 202 Grid square 14: RF 149: Phase 5. Figure 178 120 Stud. Copper alloy. Circular head with repoussé decoration, edge of head bent over; squaresectioned broken shank bent sideways. Repoussé design – head in profile facing right with ribs around the back of the head resembling a radiate crown, beaded rib at base of neck, ribs resembling a palm (?) in front of face; entire design framed by circular rib; rim of stud now retains traces of white metal. Diameter 16 mm, present length of shank 5 mm. 2016: 202 Grid square 18: RF 226: Phase 5. Figure 178 183

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 179: Weapons. Iron bolts (Nos 122-123) and arrowhead (No. 124). Scale 1:2

Quite what these rare items were used for is unclear. The perforation for suspension is narrow and their size and rarity would seem to rule out an identification as a normal everyday item of personal adornment. In this they are like the large jet and shale crescentic pendants such as that from Scarborough (Smith 1927: 182, fig. 8). These seem likely to have been in use during the late 4th century, a date that would be possible for the Norton style pendants given the context of No. 20. In discussing the crescentic pendants in relation to one found at Vine Street, Leicester, I suggested that they might have been an item of regalia (in Morris et al. forthcoming: 197). Jet does seem to have been regarded as being a material especially suited for artefacts related to Bacchic worship and was regularly used for amulets such as the Medusa pendants designed to turn away the evil eye (Allason-Jones 1996: 15-6). Possibly the Nortonstyle pendants had a similar function.

convex upper face centrally on upper edge with transverse circular perforation that cuts into upper face of block leaving small pits on either side of lug. Front has diagonal nicks on upper and side edges bordered with one narrow and one wide groove internally. Surfaces polished smooth apart from lower sides and back of lug which show slight faceting. Large chip missing from one lower corner with slight edge chipping elsewhere. Length 54 mm, maximum section 31 x 14 mm, diameter of perforation 3 mm. 2016: 202 Grid square 6: RF 108: Phase 5. Figure 180 126 Sheep metapodia, worn and damaged at both ends, but showing traces of a central perforation. 81 mm long. 2016: 202 Grid square 6: RF 346: Phase 5. 127 Ring. Iron. Diameter 36 mm. 2016: 202 Grid square 13: RF 129: Phase 5.

The perforated metapodia No. 126 is an example of a well-known, Iron Age (Bullied and Gray 1917: 423-427) – Roman (Greep 1998: fig. 123, 180-185) form normally interpreted as having a function associated with textile production, but other interpretations are possible (MacGregor 1985: 102-03).

128 Ring. Iron. Diameter 34 mm. 2016: 202 Grid square 8: RF 67: Phase 5. 129 Ring. Iron. Bar bent into spiral ring of one and a half turns. Diameter 25 x 19 mm. 2016: 206 Grid square 3: RF 109: Phase 5.

125 Pendant. Shale. Slightly tapering rectangular block with rectangular transverse cross-section and vertical triangular cross-section with slightly convex front face. Integral rectangular lug with

130 Ring; one third extant. Iron. Diameter c. 50 mm. 2016: 202 Grid square 8: RF 98: Phase 5. 131 Ring; half extant. Lead alloy. Diameter 18 mm. 2016: 202 Grid square 1: RF 17: Phase 5. 184

H.E.M. Cool and S.J. Greep: 7a – The small finds and vessel glass

Figure 180: Miscellaneous items. Shale pendant (No. 125), iron rod (No. 132) and toothed strip (No. 136), copper alloy terminal (No. 136) and curved band (No. 137). Scale 1:1 Nos 125, 135; 1:2 Nos 132, 136, 137)

132 Bent rod. Iron. Square-sectioned rod with diagonally stepped bend; tapering slightly to one broken end; other end thickens to form angular triangular wedge. Present length 130 mm, maximum section - bar 8 mm, wedge 10 x 8 mm. 2016: 202 Grid square 6: RF 105: Phase 5. Figure 180

135 Terminal. Copper alloy. Hemispherical knob terminal with collar; rectangular-sectioned broken bar with spurred projection. Present length 38 mm, section - knob 6.5 x 5.5 mm, bar 4 x 1.5 mm. 2016: 202 Grid square 2: RF 28: Phase 5. Figure 180

133 Spatula (?), terminal. Iron. Shallow D-sectioned spatula end. Present length 29 mm, maximum width 9 mm. 2015: 502: RF 236: Phase 3.

136 Toothed strip. Iron. Strip with two long finished edges; short ends broken one probably across circular perforation; one projecting tooth. Present length 40 mm, depth 28 mm. 2016: 202 Grid square 2: RF 284: Phase 5. Figure 180

134 Hooked bar. Iron. Rectangular-section bar; one end thinning and forming a hook. Present length 123 mm, section c. 3 mm. 2016: 466: RF 419: Phase 4.

137 Curved band. Copper alloy. Rectangular-sectioned curved cast strip with shallow rib parallel to each 185

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire edge. Short ends broken, curved edges original apart from slight chipping on inner edge. Outer diameter c. 75 mm, 32% of circumference extant, section 12 x 1.5 mm. 2016: 231 Grid square 16: RF 210: Phase 1. Figure 180

suggested that the quantity of styli on the site might be the result of a military presence and there was also a stylus from a Phase 1 context (No. 55). The military, though no doubt stationed within the fort at Malton, were certainly making use of the area which these excavations covered before Phase 5. This raises the intriguing prospect that possibly the Phase 5 dumps are reflecting organised decommissioning or re-organising of part of the military establishment. Dumping material when garrisons changed or bases were given up is a phenomenon that has been observed not infrequently. The most spectacular instance is probably the pit in the fabrica of the legionary fortress at Inchtuthil, near Perth, which contained getting on for a million iron nails together with iron tyres (Pitts and St. Joseph 1985: 109-113), but other more modest deposits can also be cited such as the contents of an annexe ditch at Carlisle (Caruana 1992).

OVERVIEW This assemblage has many unusual and rarely found individual items which have already been discussed, but perhaps the most interesting aspect is the large quantity of material that came from the Phase 5 rubbish dumping. The pottery indicates a late 4th- to early 5thcentury date and this is supported by items such as the bone bracelet No. 32 and the shale spindle whorl No. 39 which are both types that only come into use at this late date. There is clearly some much older material being dumped as well. If a cut-off period of the mid-3rd century is taken as the point at which individual items can be expected to have gone out of use by, then the brooches Nos 1 and 2, the hairpins Nos 7 and 9 and glass counter No. 50 can be viewed as certainly indicating rubbish that cannot have been being actively generated in the 4th century. This material was concentrated in grid squares 3 and 5 perhaps indicating a mixed source for the rubbish. Most of the items found within the rubbish, however, would be appropriate for a late Roman date and this raises interesting questions about why it was happening and what it tells us about the resources available at the end of the 4th century.

Whether this was the driving force for the dumps or not, their contents are instructive about the supply and availability of items at the end of the 4th century in this area. Swift (2012; 2014) has suggested that the re-use and re-purposing of earlier material may, in part, be the result of the increasing scarcity, and difficulty of acquiring, new things at that point. She further suggests (2012: 204) that the phenomenon of re-using bracelets as smaller and smaller rings may be, again in part, the result of shortages of newly smelted metal. Looking at the quantities of metal items through time in the assemblage (Table 22) this does not seem to be the case at Norton. Approximately three-quarters of both the copper alloy and the iron artefacts are coming from these late dumps and the range and completeness of the items in them does not paint a picture of pressure on resources. Usable things were being disposed in rubbish dumps. In the case of the knife No. 71 not only was it disposed of with its handle but it seems also to have had its sheath. No-one felt the need to scavenge to recycle the metal or find things like the knife No. 71. The objects remained there until the excavations took place one and a half millennia later and found them.

The first issue to confront is the fact that many of these items are effectively complete and could have had a useful life still in front of them. Attention has already been drawn to this phenomenon in the tools and craft section but it can also be noted in other ones as well. In the personal ornaments there is a copper alloy bangle that is complete. In the writing category all of the styli present are complete. In the transport section there is a complete snaffle bit and in the fasteners and fittings section all four of the keys are complete. The question arises why was so much material in this state being discarded? Viewed from a perspective of what these objects would have been used for it paints a picture of a source where literacy was important, security might have been important, a supply of good quality tools dealing with a variety of craft and industrial processes was normal and there were members of the military present. The military presence is probably the important factor and the other activities would have been normal in a military milieu.

Material Copper alloy Iron Total

1

2

3

4

5 Total

2

-

1

1 13

17

6 8

4 4

7 8

3 41 4 54

61 78

Table 22: Distribution of the catalogued metal items through time (excluding nails and hobnails)

ACKNOWLEDGEMENTS

Throughout the sequence there is evidence of military activity on the site. This does not just consist of the bustum burial of a soldier (see p. 187). There is also an armour scale from Phase 1 (No. 117) and a catapult bolt head from Phase 3 (No. 123). It has already been

I would like to thank S. Allen and C. Wilkinson of the York Archaeological Trust Conservation Laboratory for their identifications of the minerally preserved organics. 186

7b – The finds from the cremation burial H.E.M. Cool and S.J. Greep1 The1small finds from the bustum cremation burial pit (475) all showed evidence of having been burnt to a greater or lesser degree and so had accompanied the deceased on the pyre. Amongst them fragments from fretted openwork belt plates (Nos 1-4) were immediately recognisable. These were the types of fittings used on belts worn by soldiers between the mid-2nd and mid-3rd centuries. This means that the individual had been a soldier and that he had been placed on the pyre in what his contemporaries would recognise as military dress. Such an occurrence appears to have been extremely rare as soldiers were generally not burnt or buried with this sort of equipment. Philpott’s survey of grave goods in Roman Britain found only three burials within the 1st to 3rd centuries where military belt fitting could be identified (Philpott 1991: 187-188). The number has increased slightly since then with the analysis of the 3rd-century military cemetery at Brougham (Greep 2004a; Mould 2004) but, given that the province of Britannia was a heavily militarised one, the number is still tiny. A similar picture emerges when the burials from other provinces are considered. Hoss found less than 50 graves with belt fittings in a survey of the western empire (Hoss 2014, unnumbered following Taf. 131). This then is a find of international significance, especially at it contains all of the fittings of a belt decorated with a common type of plate enabling the whole belt to be reconstructed for the first time.

Though a bustum burial should in practice retain all the items placed on the pyre, it is clear that some elements are missing. The belt plates show this well. Two are effectively complete (Nos 3 and 4), one lacks a corner and portions of the interior (No. 2), and the fourth (No. 1) is approximately half extant. During excavation it was observed that the ground surrounding the pit had become heavily burnt and that some of the human bone was scattered on it. It is likely that some of the fragmentary pyre goods could also have fallen onto the surrounding area and have escaped being included in the pit fill and urn when the funerary ritual was being completed. Some elements too will have fragmented and become completely unrecognisable due to being burnt. Given the fragmentary nature of the material it is useful at the outset to review the evidence of what a Roman soldier of the mid-2nd to mid-3rd century can be expected to have worn and what we might hope to find associated with the burial. Armour can be excluded as the deposits produced no evidence for its presence. Here the individual seems to have been placed on the pyre in what various authors have termed camp dress. A good definition is provided by James when he describes it as ‘the usual garb of soldiers not actually armed for combat’ (James 2004: 257). This is what a soldier would have worn in camp and on the streets of the vici and in other civilian areas.

The small finds were found in numerous fragments by both hand excavation on site and within the urn which was excavated under laboratory conditions. A small number of additional pieces were found during the analysis of the human bone and environmental samples. Belt plate No. 2 shows how dispersed fragments could become. It consisted of finds that had been recovered by hand excavation in the pit, in two different spits of the excavation of the urn and from an environmental sample. In some of the catalogue entries the pieces will be quantified by small find numbers as these reflect the pieces as recovered whereas fragment numbers are likely to reflect the inevitable fragmentation that results in the post excavation process such as flotation for environmental sampling. Given the burnt state in which they were found (Figure 181), the drawings have relied to a great extent on the X-radiographs of the pieces to provide clarity and aid reconstruction.

The most visually obvious elements that set a soldier apart from other adult males were his belt and his sword. Only soldiers had the right to wear swords in public and so they did as a sign of their status. Belts became increasingly associated with the military (see Bishop and Coulston 1993: 96; 2006: 106; James 2004: 60) and were initially used to carry side arms such as swords and daggers and were regarded as integral with them (Feugère 2002: 175). Depictions of 1st-century soldiers show them wearing two belts, one supporting a sword on the right-hand side and the other a dagger on the left. During the later 2nd century there was a change in swords typically used. The short gladius was replaced by the long spatha (Bishop and Coulston 1993, 126-35; 2006, 154-63). This meant the strap arrangement for supporting the scabbard had to change. Instead of being supported on a belt, it was now carried by a baldric, a strap that passed over the right shoulder and under the left arm with the sword worn high on the left side. This change seems to be in place by the beginning of the 3rd

The responsibilities of the two authors in this report are the same as those stated in the introduction to the small find report 1

187

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Scroll holder no. 16

Fragments from the belt set (nos. 1-6)

Baldric no. 7

Brooch no. 12

Figure 181: A selection of the pyre goods. Scale 1:1

century. By the 3rd century soldiers are not normally depicted wearing daggers but they clearly continued in use as a 3rd-century deposit in a military workshop at Künzing contained large numbers of dagger blades and scabbards (Herrman 1969: 133 Abb. 3; Bishop and Coulston 1993: 135; 2006: 164).

relatively small number of military graves which are broadly contemporary with the Norton soldier, three are helpful in showing the sort of fittings and their numbers that might be expected. An individual found during building work at the rue des Fantasques, Lyon has long been interpreted as a casualty of the Battle of Lugdunum in AD 197 as a small group of a dozen silver denarii ending with one of AD 194 were found with him (Wuilleumier 1950: 146-148). The circumstances of the discovery mean that there are no records of the location of the accompanying military equipment in the grave, if indeed it was a formal inhumation, other than the description that they surrounded the lower part of the skeleton which was the only part to survive. The best illustration of the fittings is provided by Feugère (2002: figures 151-152) and a reconstruction of what the belt, baldric, sword

A cloak would have been another vital element. Cloaks were so strongly associated with military life that terms such as taking or wearing the cloak were figures of speech associated with going to, and being at, war. Also in the public’s mind soldiers were associated with hobnailed shoes (Spiedel 2012: 9). So we can conclude that the camp dress of the Norton soldier would have included, in addition to tunic, underclothes and nailed shoes, a cloak, a belt, a baldric and its sword and possibly a dagger. Amongst the 188

H.E.M. Cool and S.J. Greep1: 7b – The finds from the cremation burial

and scabbard would have appeared like is provided by Bishop and Coulston (1993: figure 92; 2006: figure 101).

220 indicating an adult or adolescent. Belt fittings were found in a position that suggested the deceased was buried wearing a belt. A mid-2nd-century jar was found at the foot of the grave.

The belt was decorated by plates in the form of letters which spelt out the motto VTERE FELIX (live happily). The positioning of the buckle joined to the letter X and the conjoined VT plate with a bracing strip shows the buckle was worn centrally with letter plates on either side. One of the E plates had a small loop on the bottom arm which James (2004: 61) suggests was an attachment for a purse. This could well explain the small group of coins found with the burial. There were also two narrow strap ends for narrow leather straps. These most likely decorated the free end of the belt strap which would have been split to accommodate them. Depictions of 3rd-century soldiers show that the free end of the belt was long, worn looped over the belt and then hung down on the right thigh. James has drawn attention to the jangling sound they would have produced as the man walked (James 2004, 61). Noise was another thing that people associated with soldiers walking through the streets.

The obvious belt fittings included a buckle with a plate attached to it. The plate had a strip, either integral or riveted to it (the drawing is unclear), bent back along the underside with the loop engaging the buckle. There were three other belt plates with paired perforations at either end matching those on the buckle plate. Allowing for some slight displacement the plates had all been to the left of the buckle (as worn) so the buckle would have been worn on the right side. A single teardrop pendant with sheet strip for fixing to a narrow strap was found at some distance to the left of the individual above waist height. The publication interprets this as a single apron mount and cannot account for it, but it would be appropriate to have terminated the long free end of the belt as discussed in connection with the Lyon find which might account for its position, possibly displaced during deposition.

There were also three fittings from a baldric. Two came from the lower edge of the wide front strap consisting of a rectangular plate which would have bound the edge and a heart-shaped pendant that was hinged to it. The other was a circular phalera which pierced the leather so there was a rectangular loop behind which provided the fixing point for the other end of the strap after it had passed under the left arm. Bishop and Coulston also suggest that a small rosette-headed stud was attached to the baldric, but this seems to have been decorative rather than functional if the reconstruction is correct. The scabbard would have been fastened to the baldric strap with the aid of a scabbard runner through which the strap passed. The scabbard had clearly been present as both a runner and a scabbard chape were found.

There were also two peltate studs with integral washers and a rectangular fitting with a loop through which a strap could have fitted and a plate with rivets for attachment to a strap. One of the studs and the rectangular fitting was positioned above the junction of the two outermost belt plates. These were interpreted as baldric fittings and, though the suggested reconstruction with them attached to the belt does not seem likely, they are in the correct position for a baldric. The short edge binding strip positioned below them may have been an additional baldric fitting rather than the binding for the free belt end as the publication suggests. As already noted, the free ends of belts tended to be long, not short as the reconstruction would suggest. The position of a long ribbed binding is not stated. It was interpreted as the terminal of a wide belt, but the buckle plate rules this out as it was for a narrow belt the same width as the plates. Possibly this longer binding could have been associated with the wide front end of the baldric.

The group also contained a knee brooch so this soldier was in full camp dress of cloak, belt and baldric. The only missing item appears to be the hobnailed shoes but that may well be the result of the recovery method. Third-century shoes were more lightly nailed (Goubitz et al. 2001: 351) and so the hobnails are less likely to corrode together to form a shoe shape and more likely to be recovered as single items. They would have been small rusty lumps, easily overlooked by the builders who recovered the rest of the finds.

The individual was buried with his hobnailed shoes so here the camp dress presented is of belt, baldric (without weapons) and shoes. The third grave is a cremation burial in the St Pancras cemetery, Chichester (Down and Rule 1971: Burial Group 251-81, 117, figure 5.18). The rich ensemble of grave goods was placed in a box burial and included a belt set. The burial is most likely to belong to the second half of the 2nd century as a posthumous coin of Faustina the Elder was placed in the cremation urn and the pottery included Antonine samian. No pyre goods were reported, nor was there any attempt to age or sex the cremated bone, but this is to be expected given the

The second grave was recovered through formal excavation of the Racecourse cemetery at Derby (Wheeler 1986: 228 Grave 220, 269-273, figures 105, 12021). This was one of three inhumation graves in an area between a walled cemetery and a set of stone mausolea lining a road. These three graves were marked out by being longer, wider and deeper than the other ones in the cemetery. Only the leg long bones survived in grave 189

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire date of the excavations (1965) and period during which the post-excavation work was being carried out.

The pattern on Nos 1-4 is a less common one. Though the two rows of arches are not uncommon they are not normally divided by a central bar as here. Hoss’s extensive literature search has produced an additional three examples, all undoubtedly from the same workshop and demonstrating how widely travelled military metalwork could be. One was found at Caerleon, another at Zwammerdam and a frog fitting came from Thamusida in Morocco (Hoss 2014: 136-138 nos B872 and B907; 242 no. C252).

The enamelled belt set was appropriate for a narrow belt and consisted of a buckle and buckle plate hinged together and five rectangular plates. Two of these had an angled stud fitting at one end forming a frog for the attachment of a dagger scabbard. One had a central loop on one long side probably for the attachment of a purse. There were also two long strap ends with a central teardrop-shaped unit between ribs, riveted at one end. These were found with the belt fittings laid over and around a samian dish (Dr. 27). Other items placed in this position were a bronze fitting interpreted as a knife with two small perforations either side of the central rib close to the tang, a long cylindrical binding, a small stud with an enamelled centre similar to the frogs on the belt plate and fragments of unidentified sheet. There is no record of the precise arrangement of the belt fittings.

How belts with plates of this sort were fastened has long been something of a puzzle. Oldenstein proposed that the ones with the frog fittings may have fitted into a rectangular eye fitting on the edge of another one (Oldenstein 1977: 196, Abb. 5) but, though the frog fittings are found from time to time, ones with eyes seem very rare. One plate with a bent under strip from Zugmantel was found with an attached buckle fastened as in the belt from Derby, but plates with these fittings too are rare. Oldenstein (1977: 275-276 nos 986-987) knew of only two and Hoss’s survey did not increase the total. The Norton set, showing a hinged junction with a buckle of Hoss (2014) Typ A4b, is a welcome addition to the corpus. Though no belt plates of this type have been found during excavations in and around the fort at Malton, a fragment of a buckle of this type was recovered from a residual context in the Orchard Cottage area of the vicus (Lloyd-Morgan 1997: 133 no. 1, figure 50).

The report interprets this as the grave of a minor official, possibly a scribe given the ‘knife’. The belt though is clearly military both on the grounds of the provision for the dagger and on the general principle that it was the marker of a military identity. No other aspects of camp dress are present but it cannot be ruled out that the deceased did not go to his pyre dressed in some other elements such as the cloak and shoes given that any pyre goods would not have been recognised or published.

Also welcome is the presence of the two plates with the frog fittings for suspension of a side arm. Given the presence of a baldric fitting (see No. 7) these cannot have been for a sword and so confirm they were for daggers. Whilst the Derby burial shows that the buckle could be at one end with all of the belt plates behind it, here a symmetrical arrangement would be more likely. The arrangement of the belt could thus be suggested to have had the buckle plate centrally with plate No. 2 to the individual’s left side and the two frog plates on the right with a suspended scabbard for a dagger.

Having set the scene of how soldiers dressed and were occasionally buried, it is now possible to turn to the material associated with the Norton bustum. The Belt The belt set consists of four belt plates with an attached buckle (Nos 1-4) and two strap ends (Nos 5-6). The plates with their fretted openwork pattern belong to a widespread class which Oldenstein (1977: 197) considered belonged to the last third of the 2nd century and first half of the 3rd. Hoss who has provided a more systematic typology for military belts divides plates into simple ones (Group B – Gürtelbeschlüge) and those that could have acted as suspension mounts (Group C – Scheidenbefestigungen). Nos 1 and 2 thus belong to her Typ B.11.1a (Hoss 2014: 134-39) and Nos 3 and 4 with the frogs to her Typ C.9.c (Hoss 2014: 242). Reviewing evidence published since Oldenstein wrote on the subject, she considers the contextual evidence suggests the type emerged a decade or so earlier than he suggested and so a date in the second half of the 2nd century and first half of the 3rd is appropriate.

Two strap ends (Nos 5-6) were recovered suggesting that as with the Lyon and Chichester belts, the free end of the strap had been split and the two ends provided with metal strap ends that would have jangled together when the man walked. As already shown the belt fittings from the waist are common and widespread forms, the strap ends by contrast are not. Neither Oldenstein nor Hoss include this ribbed form within their typologies. Allowing for the distortion produced by being burnt, the Norton strap ends are very similar to those that formed part of the Chichester belt (Down and Rule 1971, figure 5.18 nos 251j-k – illustrated upside down). Both have flat oval plates with rivets for attachment, a central area which in the unburnt examples is an elegant baluster shape with ribs at either end and a thinner flat

The openwork designs that are used in these plates normally fall into a relatively small number of variants. 190

H.E.M. Cool and S.J. Greep1: 7b – The finds from the cremation burial

terminal. They only differ in the fact that one of the Norton examples is longer. The close similarity of these strap ends might suggest that it is possible to narrow the century long period when the belt could have been in use to the second half of the 2nd century.

internal expansions formed by six D-shaped cutouts by frame and five triangular cut-outs by bar. Buckle – width 33 mm, depth (including hinge) 35 mm. Buckle plate – width 25 mm, maximum joined length 38 mm, depth (including studs) 5.56 mm., thickness 1.5 mm. Figures 181, 183 RF 837 Cremation urn RF 416 Spit 10 (buckle and back part of plate, now in three fragments) RF 832 Cremation urn RF 416 Spit 11 (outer buckle frame). RF 415 (476) Stud and central part of outer plate frame.

The one aspect of the belt that is missing from the Norton ensemble is, of course, the leather belt itself. The openwork pattern of the belt plates must have been designed to show the colour of the leather behind resulting in a contrast between that and the shining bronze of the metal. Whilst naturally tanned leather would provide a contrast of sorts, leather that had been stained would provide a better one. A survey of the colours sometimes seen on depictions of soldiers in paintings and mosaics found that red occurs regularly (Hoss 2014: 176-177). Interestingly on a wall painting from Dura Europos it is the officer that has a red belt whilst other soldiers have naturally tanned ones (James 2004: 61). Given that Norton soldier was in his late 30s or older (Chapter 13, p. 240), there is the possibility that he had risen through the ranks to a senior position. For this reason the reconstructions shown in Figure 182 show the belt set mounted on a red belt. 1

Buckle and attached belt plate. Reconstructed from four small finds, three joining. Copper alloy. Buckle – lacking half of outer frame; rectangularsectioned, rounded D-shaped frame with internal volutes; bar between volutes with iron pin wrapped around; two projecting perforated plates centrally on lower edge forming hinge with corresponding plate on one short edge of buckle plate, iron hinge bar. Buckle plate, approximately half extant – rectangular plate with integral stud with squared plate on underside centrally at each short end. Open work fretted design with narrow frame around each edge and central bar, on either side of bar symmetrical design of six arches with

2

Belt plate. Copper alloy. Reconstructed from five small finds, all joining. Lacking one corner and small parts of internal fretwork. Rectangular frame with studs and openwork pattern as No. 1; studs have more circular plates. Length 50 mm, width 23 mm, depth (including studs) 5.5 mm, thickness 1.5 mm. Figures 181, 183 RF 828 Urn RF 416 Spit 10 (complete short end). RF 836 Urn RF 416 Spit 9 (part of long side with openwork to central frame). RF 459 (476) (corner, outer edge and stud). RF 724 (476) environmental sample 77 (frame and part of openwork). RF 726 (476) environmental sample 77 (small fragments of openwork).

3

Belt plate. Copper alloy. Found together in five joining fragments. Lacking small part of one long side and parts of three adjoining arches. Rectangular frame with studs and openwork as No. 1. One short edge has angled bar sloping upwards with integral disk (frog). Length (of plate) 46 mm, length including frog 57 mm, width 23 mm, depth (including studs) 5.5 mm, thickness 2 mm. Figures 181, 184 RF 411 (476).

Figure 182: Reconstruction of the belt

191

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 183: Belt set. Copper alloy buckle and plates (Nos 1-2). Scale 1:1

4

Belt plate. Copper alloy. Found together in five joining fragments. Lacking only very small fragment of internal arches at one point. Description as No. 3. Length (of plate) 47 mm, length including frog 57 mm, width 22 mm, depth (including studs) 5.5 mm, thickness 2 mm. Figures 181, 184 RF 412 (476).

5

192

Strap end. Copper alloy. Found in two joining pieces. Flat oval plate centrally perforated retaining shank of rivet; body D-sectioned with units of three transverse ribs above and below a slightly expanding central unit; flat straightedged, slightly expanding terminal. Length 42 mm, width of rivet plate 8.5 mm, width terminal 8 mm. Figures 181, 184 RF 413 (476).

H.E.M. Cool and S.J. Greep1: 7b – The finds from the cremation burial

Figure 184: Belt set and baldric fittings. Copper alloy plates with frogs (Nos 3-4) and baldric fittings (Nos 7-8). Scale 1:1

6

Strap end. Copper alloy. As No. 5 but heavily burnt. Current length 36 mm, width terminal 6 mm. Figures 181, 184 RF 460 (490).

category that Oldenstein has described as Herzförmige durchbrokene Anhänger (heart-shaped openwork pendants) dated to the 2nd to 3rd centuries (Oldenstein 1977: 12736). The class as a whole includes large and small examples and the larger ones are normally identified as fittings for the front of baldrics (Bishop and Coulston 1994: 134 figure 91 no. 15; 2006: figure 100). No. 7 finds no exact parallel in the Oldenstein corpus as it was not a pendant in the strict sense of being either hinged to another plate attached to

The Baldric In addition to the mounts for the belt, a mount for a much wider strap was recovered (No. 7). It would fall within the 193

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire the leather or attached via a slot along the top probably via a sheet strap. This example was clearly riveted directly to the leather and so would probably be better described as a very large strap end. This method of attachment would mean that another mount for attachment at the bottom of the baldric as on the Lyon set would not be needed, and indeed no evidence for one was found. The original width of the piece would certainly be wide enough for the front of the strap. A more elaborately pierced example provided with hinges from a mid-3rd-century context at Vindolanda would have fitted a strap of c. 55 mm width (Bidwell 1985: 119 no. 16, figure 40) compared to the likely 50 mm width of No. 7.

As is often the case with iron that has been burnt to the temperatures required to cremate a human body, very little metallic iron remains in these items. The interiors have become mineralised and quite often show as voids on the broken edges. In all three of these pieces the appropriate triangular cross-section of a blade can be observed. Although they do not join they were found close together in adjoining spits in the urn suggesting they were gathered from the same area of the pyre site. This and the close similarity of the pieces suggest all three came from the same blade. The copper alloy strip on No. 9 and the traces of a similar strip on No. 11 suggest the blade may have been in a decorated scabbard. It would be tempting to identify this as the dagger implied by the frog plates Nos 3 and 4 but the width (currently less than 20 mm) would argue against this. Roman military daggers are normally waisted and wide (see Herrmann 1969: Abb 3; Bishop and Coulston 1994: figure 95; 2006: figure 104 for a selection of 3rd-century daggers). This blade would be more correctly described as a narrow knife. Whether or not despite this it was what was held by the belt plates is, alas, impossible to know.

A baldric would have needed a fitting further up the front of the strap to fasten the other end of the strap as in the Lyon group. There is no obvious item that could serve as the front of this fitting but No. 8 might have come from the ring at the rear of the fitting into which the strap was threaded and then adjusted. A useful group of phalerae showing the rings as well as the decorative fronts have been published by Bishop and Coulston (1994: figure 91 nos 1-3, 6, 8; 2006: figure 100 same numbering). From these it can be seen that these ranged from the rounded rectangular to D-shaped with a full half circle. The width of the loop ranged from 17 to 24 mm. No. 8 is a fragment with a curve which would correspond to a diameter of 20 mm. So, while the identification must remain speculative, it is a possible one. 7

8

Baldric mount. Copper alloy. Lacking one upper corner and heavily burnt. Rectangular-sectioned plate with straight upper edge with projecting ends; curved sides and broken and thickened central ribbed terminal. Small perforation in extant corner; lunate perforation in each side with central circular perforation in space defined by larger perforation. Maximum extant width 43 mm, original width along upper edge c. 50 mm, present length 42 mm, thickness of upper edge 2 mm. Figures 181, 184 RF 829 Cremation urn RF 416 Spit 3.

9

Blade fragment. Iron. Both ends broken. Edge parallel with back. By one broken end remains of copper alloy thick sheet strip wrapped around sides and edge, back of blade retains copper alloy corrosion. Present length 19 mm, width 18 mm. Figure 185 RF 845 Cremation Urn RF 416 Spit 7

10

Blade fragment. Iron. Both ends broken; slightly tapering. Present length 37 mm, width 17-15 mm. Figure 185 RF 830 Cremation Urn RF 416 Spit 7 11 Blade fragment. Iron. Both ends broken; sides tapering. Copper alloy corrosion on one broken edge all around blade. Present length 16 mm, width 16-8 mm. Figure 185 RF 847 Cremation Urn RF 416 Spit 8

Phalera loop? Copper alloy. Fragment of curved square-sectioned bar. Length 15 mm, section c. 2 mm. diameter of curve c. 20 mm. Figure 184 RF 414 (476).

Clothing Evidence suggesting a cloak and nailed shoes was also found. The brooch No. 12 belongs to Mackreth’s Plate brooch type 3.a.1 (Mackreth 2011: 159, pl. 106) and would fall into the larger end of the size range. This is a British type which has rarely been found in usefully dated contexts but such dating as there is suggests it was in existence by the mid-2nd century. The presence of this example in the grave provides a welcome addition to the dating. It would have been an attractive brooch with alternating fields of red and blue enamel surrounding the central setting which is likely to have been an intaglio depicting a bust. Certainly, the brooch is substantial enough to have fastened a cloak on the shoulder. The fact that there is a brooch in this burial

Weapons The finds from the pit produced no evidence that the baldric had supported a scabbard and sword when the deceased was put on the pyre. As can be seen from the Lyon group, had this been the case fragments from the scabbard slide and chape might have been expected. There are, however, three intriguing fragments of iron (Nos 9-11) that might suggest a blade within a scabbard decorated with copper alloy.

194

H.E.M. Cool and S.J. Greep1: 7b – The finds from the cremation burial

10

9

11

12 RF862

RF863

RF864

RF866

15

13 RF865

RF856

RF855

16

Figure 185: Other fittings from pyre. Iron and copper alloy blade (Nos 9-11), copper alloy brooch (No. 12), iron hobnail (No. 13), bone inlay (No. 15) and scroll holder fittings (Nos 16). Scale 1:1

is another indication that the deceased was a soldier. During the second half of the 2nd century many of the civilian population were giving up wearing brooches. Within the military community they continued to have a role very possibly because cloaks were central to the identity of soldiers.

12

The deceased also had nailed shoes on the pyre with him as two nails can be recognised from the X-radiographs. This is a very small number but sufficient to indicate the presence of such shoes and, as noted in connection with the Lyon burial, there was a tendency towards lighter nailing by the end of the 2nd century.

195

Plate brooch found with central setting detached from plate. Leaded bronze (EDXRF using Oxford Instruments WD 2000 at Durham University Dept of Archaeology). Oval plate with single perforated lug centrally on back holding spring of c. two turns with cord running below lug; pin missing; catchplate now lacking return. Raised rim around edge on front with separate oval box for setting originally soldered to it. External field retains traces of red and blue enamel, probably set in alternating blocks; setting missing and box retains only corrosion products. Diameter 30 x 25 mm, thickness of plate 4 mm, height of box setting 4 mm. Figures 181, 185

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire RF 784 (476) – plate RF 785 (476) – box setting. 13

Hobnail. Iron. Flat-topped pyramidal head. Length 14 mm, head width 7 mm RF 738 (476). Figure 185

14

Hobnail. Iron. Head width 8 mm. RF 742 (476).

the same cemetery, a very similar disc and terminal (Barber and Bowsher 2000: 195 and figure 96) may be part of a type of objects recently identified as possible scroll holders (Greep and Rijkelijkhuizen forthcoming). These finds have very characteristic discs (identical to the Norton example) used to block a pair of latheturned cylinders. Few complete assemblages have been recovered (Rijkelijkhuizen 2017), more normal are parts of the assemblage broken and burnt and deposited as secondary goods, as in the London example cited above and also the grave from Caerleon with small triangles of inlay discussed elsewhere (Reynolds 2015: figure 14) as well as here at Norton. They may be dated c. AD 125-250 which fits well with the bustum chronology. The normal association of these forms is with female burials.

Other finds In addition to the finds that may have been directly associated with the body on the pyre. there are a number of other items that indicate other pyre goods were present. Of these the ones made in bone are the most intriguing.

Amongst the bone there was also a series of other, perforated and burnt bone objects (No. 17) none of these form recognisable objects, and do not seem to belong to the two groups described above.

Four small, undecorated, bone triangles (No. 15) represent a part of what must have been a much larger collection of inlay, probably from a small box or casket, rather than a larger item such as a funerary couch, although it is not possible to rule this out.

None of the fragmentary metal items can be identified. The sheet fragments No. 18 clearly decorated an item and the stud/rivets Nos 20 would be the correct size to have fixed the sheet in place as they are clearly too slight to have been part of the substantial bar and stud fitting No. 19 or for the baldric.

Such pieces of inlay are most commonly found in 4thcentury contexts (e.g. Greep 2012), but inlayed items of furniture occur throughout the Roman period. The use of small triangles, such as the Norton examples, probably to inlay grooves on items of furniture have a long history in Roman Britain and have previously been recorded in a number of cremation contexts. At Caerleon, from a cremation in a 2nd/3rd-century burial (Burial I/014, with a mix of 2nd- and 3rd-century pottery; Reynolds 2015: figure 14), a 3rd-century burial at Usk (albeit, decorated with a ring and dot motif; Greep: 1995b) and a later 3rd/4th-century cremation at London (Barber and Bowsher 2000: 249). The recovery of a wooden door panel from a mid-4th-century well at Hayton (Halkon et al. 2015: 304-06), with inlaid bone decoration (including small triangles) demonstrates how they may have been used. Small triangles were part of the furniture makers ‘stock’ found at Gloucester in a deposit of late 4th to early 5th centuries (Hassal and Rhodes 1974: figure 28) demonstrating the longevity of use of these small items. They should not be confused, however, with the thicker and more decorative inlays which occur, often in 3rd-century burials, as, for example at Birdoswald (Wilmott et al. 2009: 283-288) and Brougham (Greep 2004b) which are of a different character and must have been used to decorate far larger objects.

The disc No. 21 has been deliberately fastened to something by the iron shank as the same shank is present on both front and back and appears to pass through it, rather than being fortuitous accretions. Delicate copper items such as this are not normally nailed in place but it does raise the intriguing possibility that it was part of the construction of the pyre or bier. In the 3rd-century biers in northern military sites were being elaborately decorated with bone veneers (Greep 2004b; Wilmott et al. 2009: 283-238), so decoration of such structures might well be possible. However, this is the only piece like this in the burial and if there had been a systematic decoration more might have been expected. All the fragments were found during the human osteology sorting.

No. 16 consists of one or possibly two, small, latheturned ‘terminals’ and a disc, all burnt and broken. The most obvious identification is that they are parts of pyxides (cylindrical boxes with separate lids). Disc and knob-shaped terminals are well represented in known bone pyxis forms such as from a cremation burial in London (Barber and Bowsher 2000: 188, 7 and figure 95). However, associated with another cremation from 196

15

Four fragments of bone inlay. All fragments from (476). Figure 185 RF 862 Very small triangle 14 mm x 6 mm. Burnt RF 863 Very small triangle. 14 mm x 6 mm. Burnt RF 864 A very small triangle in two joining fragments. Burnt. 16 mm x 6.5 mm RF 866. Very small triangle 15 x 6 mm.

16

Fragments from bone terminals and disc. RF 85556 from Cremation Urn RF 416. RF 865 from (490). Figures 181, 185 RF 856 small disc with two grooves and a central perforation. Burnt 16 mm diameter. RF 855 Very small fragment of burnt bone. Possibly part of a ‘knobbed’ terminal. 10 mm long.

H.E.M. Cool and S.J. Greep1: 7b – The finds from the cremation burial

Figure 186: Other fittings from pyre. Bone (No. 17), copper alloy (Nos 18-20) and copper alloy with iron (No. 21). Scale 1:1

RF 865 A very small fragment of a ‘terminal’, possibly similar to 855 above.

RF 750. Section of a rib bone, with a central perforation. Burnt and adjoining to RF 853 37 mm long. Figure 186 RF 857 Small fragment of a burnt rib bone – probably belongs with 853 above RF 858 Burnt D-shaped section of bone (two fragments) 52 mm long max length with a perforation at one end. Figure 186 RF 854 Burnt D sectioned piece of bone with traces of a perforation. Not adjoining to RF 858 but very similar. 25 mm long.

17 Other fragment of worked burnt bone. RF 858 from within Cremation urn RF 416. RF 750 Context (477) sample 78. All other fragments from (476). RF 861 Two burnt fragments form a lozengeshaped object with a central perforation. 20 mm long. Figure 186 RF 853 Part of a rib bone, with a central perforation. Two fragments, not joining. Burnt 197

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire RF 853 two small fragments of a burnt ?rib bone. ?Worked? 18

consistent with the dating of the jar containing some of the cremated bones. The other pottery from the pit suggests a 3rd-century date with some fragments of Crambeck grey ware suggesting a late 3rd-century date. The deceased was a mature man and so may have had his equipment for some decades. This would allow the cremation to have taken place in the earlier part of the 3rd century but certainly not in the later part of it. The small Crambeck sherds are most likely to be intrusive.

Sheet strip. Copper alloy, two straight edges, other sides broken each across a small perforation. Fragment now buckled and slightly bent. Present length 20 mm, width 16 mm, perforation diameter 2 mm. Figure 186 RF 458 (476) Other fragments of similar sheet RF 838 Cremation urn RF 416 Spit 1 RF 843 Cremation urn RF 416 Spit 7 RF 831 Cremation urn RF 416 Spit 10 RF 842 Cremation urn RF 416 Spit 10 RF 840 Cremation urn RF 416 Spit 12

19

This individual had undoubtedly been a soldier. Though the osteological examination could not sex the remains, this was undoubtedly a man as nothing we know about the Roman army leads us to suspect that it was recruiting females. His military identity was clearly important to the people who prepared the funeral and, presumably, to the man himself. He went to his pyre in the full panoply of military dress possibly clasping an important scroll in its holder. Why this should be so when so many soldiers died and were buried without any obvious archaeologically recoverable indicators is an intriguing and unanswerable question. Such age data that is available would allow him to have either been a serving soldier or a veteran.

Bar and stud. Copper alloy. Approximately squaresectioned bar retaining hollow conical-headed stud with square-sectioned shank bent sideways below the bar. Bar – present length 13 mm, section 5 mm; stud – head diameter 5 mm, original length 12 mm, shank section 2 mm. Figure 186 RF 834 Cremation urn RF 416 Spit 4

20 Stud/rivet. Copper alloy. Square-sectioned tapering shank retaining small part of curved head. Present length 5.5 mm, shank section 1 mm. Figure 186 RF 727 (476) Environmental sample 77.

Whether he had served in a legionary or auxiliary unit is another unknown. Belts like this must have been worn by soldiers in both types of regiments as they are found both in legionary fortresses and auxiliary forts. Assuming he was connected with the fort at Malton then an auxiliary unit would be likely be likely. His equipment shows that he could have been contemporary with the unit recorded there on a dedication slab found residually in the vicus (Wenham and Heywood 1997: 39 no. 1; RIB III no. 3207). The lettering and leaf stop on this suggest an Antonine date and it records the presence of the ala Picentiana so our soldier may have been a cavalryman.

Other small fragments from similar stud/rivets RF 725 (476) Environmental sample 77 RF 745 (476) Environmental sample 78 21 Sheet disc. Copper alloy. Disc with damaged edges with a rectangular-sectioned iron shank transfixing it at an angle. Diameter 18 mm, thickness 1.5. section of iron shank 5 x 4 mm. Figure 186 RF 846 Cremation urn RF 416 Spit 8

What rank he might have had is also unknown. As can be seen from Hoss’s useful compendium of images of soldiers, being portrayed as a soldier in camp dress and holding a scroll was a popular choice for military gravestones and does not appear to be limited to any particular rank. So while Aurelius Bitus, a corniculari (secretary) of the Legio II Adiutrix stationed at Aquincum was portrayed with what could be seen as one of the signs of his role, the heirs of other soldiers stationed there also chose to depict them in this way. These included soldier C. Iulius Sabinus from the same legion and an unnamed standard bearer (Hoss 2014: volume 2 69 nos 177-179). So while the scroll would fit a general military identity, it does not advance us very far.

Nails Numerous fragments of nail shanks were recovered from the pit with a small number from the urn (RFs 833, 835, 839). Quantified by heads 13 were recovered with four being complete. These had a range of 54 to 66 mm and so fall within the category of general purpose joinery nails. It is likely they derive from the pyre either being used to help build it, or from old structural timbers used in its construction. CONCLUSIONS The consistent pattern in the dating for all these fittings is that as an ensemble they would have come together between the mid-2nd to mid-3rd centuries with the strap ends suggesting the earlier part of that period might be most appropriate. This dating would be

The obvious thing missing from the ensemble are his side weapons. We have seen that in life they were an integral part of camp dress but it is noticeable that they 198

H.E.M. Cool and S.J. Greep1: 7b – The finds from the cremation burial

are extremely rare in the already rare military graves. The grave at Lyon with a spatha is an exception but the circumstances of the find apparently within a building raise questions about the nature of the interment. Was this the result of a formal funeral or was he a casualty of the battle?

deprived of his sword and other markers of a military identity as it was a normal part of punishment (Spiedel 2012: 9). Feugère (2002: figure 156) was surely correct when he captioned an illustration of the pit as ‘Murder at Canterbury’. In which case the absence of belts but the presence of swords is understandable. The bodies had been stripped before being hidden. Disposing of a body is always the most difficult part of a murder and here the equally incriminating swords were added to the pit. Questions might well have been asked if the murderer(s) had been found with the swords in their possession.

There are even more questions surrounding the recovery of two individuals with spathae found in a pit at Canterbury who are likely to have died around the time the Norton soldier did (Bennett 1982: 44-46, figure 10; Webster 1982). The skeletons gave every appearance of having been thrown into a pit, one on top of the other, with the swords in their scabbards tossed in on top. One individual still had his hobnailed shoes and there were a few other copper alloy fittings in the pit but no belt fittings. The excavator noted that ‘the very unusual nature of the burial could suggest murder, execution or perhaps a ritual killing’ as being possibilities to explain them. The last two can be ruled out as unlikely. Certainly, a disgraced soldier who was executed would have been

So it can be concluded that the lack of side weapons is to be expected and that the Norton soldier was cremated with honour and with all the parts of his uniform and possessions that were thought appropriate. His side arms and perhaps the rest of his battle dress being inherited by his heir or absorbed into the armoury of his unit.

199

Chapter 8

The ceramic building materials J.M. McComish INTRODUCTION AND METHODOLOGY The ceramic building material (CBM) from the site comprised 1267 sherds collectively weighing 109.163 kg, which ranged from Roman to modern in date, though the majority of the collection was Roman. The various forms present are summarised by historical period on Table 23.

Period

Form

Roman

Brick

Chimney Box Flue Imbrex Other

The CBM was recorded by two people, with the Trial Trench material comprising 43 sherds, collectively weighing 5.104 kg being recorded by S Tibbles, while the CBM from the main phase of excavations, comprising 1224 sherds, collectively weighing 104.059 kg, was recorded by J M McComish. The results of the two phases of recording have been combined here.

Parietalis Ridge

Medieval Post-medieval

As the focus of this publication is the Roman remains at the site, the medieval, post-medieval and modern material is not discussed in detail here, though it is fully described in the assessment report for the site (McComish 2019) and it is listed in Table 23.

Modern

Unknown

The collection of CBM was of very poor-quality overall being highly abraded and fragmented. For most sherds the only surviving original dimension was the thickness; breadths and lengths are only mentioned in the text where such dimensions survived. In addition, 0.62% of the total volume comprised sherds that were of such small size, and so badly damaged, that their original form or date was impossible to determine (termed ‘Unknown’).

514

4

56

36508

325

8618

33.44

0.30 7.89

202

24647

22.58

1

400

0.37

2 3

325 579

0.30 0.53

Tegula

57

11051

10.12

Plain

59

2535

2.32

Brick

26

2789

Tessera Ridge Pan

Brick Duct Pan

RESULTS

No. of Weight in % of sherds grams total weight

Roof tile

Unknown

2

28 62

35

1580 4651

0.03 1.45 2.55 4.26

98

11953

10.95

1

275

0.25

1

20

131

200

2020

672

0.18 1.85 0.62

Table 23: Ceramic Building Material by form in relation to period

Roofing tiles The tegulae ranged in thickness from 17-29 mm (43 measurements). Seven of the sherds (16.2% of the total number of measurements) were below 20 mm thick. Brodribb (1987: 13) stated that nationally it is rare to find tegulae below 20 mm thick, though cited examples known from Ickham and Slonk Hill at 14 mm and 18 mm thick respectively. In nearby York, however, 19% of the tegulae sherds are below 20 mm thick (York Archaeological Trust database used for McComish 2012). The range of thicknesses at Brooklyn House is therefore close to the pattern seen in York.

Roman The Roman CBM accounted for 75.6% of the total volume of CBM from the site. The forms present included roof tiles (tegula, imbrex, ridge tile, chimney and two roof tiles of non-standardised form termed ‘Other’), boxflue, parietalis and tesserae, but the majority of the fragments were of indeterminate form (termed ‘Roman brick’). Roman tiles and bricks were made using sanded moulds on a sanded workbench, consequently the sides and basal surfaces are coated with fine sand; this was present on all the surviving sides and basal surfaces of the Roman CBM.

There is considerable variation in the size of tegulae flanges nationally (Brodribb 1987: 13), and the present site fits this pattern with flanges 35-53 mm high (27 measurements) and two abnormally wide flanges (35 mm and 36 mm respectively). 200

J.M. McComish: The ceramic building materials

Various features relating to manufacture were present on the tegulae. The upper surfaces of the tegulae had been smoothed, as is the norm. The thumb and side of the hand between the thumb and index finger were often run along the flange to smooth it, resulting in pronounced lines drawn by the thumb adjacent to the flange. Such lines were present on 18 tegulae from the site. Two knife-cut upper cutaways were present. A single knife-cut lower cutaway was present, which was a type B6, following the classification by Warry (2006: 61); this is a commonly occurring form. A single tegula had a knife-trimmed edge while a second sherd had knife-trimming on the base adjacent to the arris. Five of the tegulae had reduced cores, caused by a reduction in oxygen within the kiln during firing, while one sherd was slightly curved possibly due to overfiring. Two tegulae had features relating to use, one with a sooted base and one with a tar like substance on the upper surface. Four of the tegulae had been burnt postbreakage.

Three additional sherds probably also came from a chimney originally. The first (Figure 189) was a wheelthrown sherd, 100 g in weight and 19 mm thick, with a horizontal band of pie-crust decoration on the exterior (context 202). The second example was a sherd 60 g in weight and 15 mm thick (Figure 189), which formed the base of a hollow cylindrical object with pie-crust decoration at the base (context 236). The third sherd was the uppermost part of a chimney (context 202) which weighed 25 g (Figure 189). This was broken off at both ends but formed an elongated cone shape in excess of 41 mm long, which was 13 mm in diameter at the upper end flaring out to 26 mm in diameter at the basal end. Roman chimneys can take the form of a stack integrated into the ridge tile, or a free-standing pot which would have been fixed onto the ridge line with mortar. One of the best-preserved examples of a Roman chimney from Britain is from Norton, this particular example being integrated into the ridge tile (Brodribb 1987: 32). One of the sherds from Brooklyn House would have originated from a free-standing pot, but for the remaining three sherds the original form is uncertain. Given that Roman chimneys are rare nationally, the sherds from Brooklyn House represent a valuable addition to the number of such items known both from the immediate vicinity and nationally.

The imbrices at the site ranged in thickness from 1429 mm (185 surviving measurements) fitting into the nationally recorded range of 14-30 mm (Brodribb 1987: 26). A possible nail hole was present at the edge of one sherd. Nationally nail holes are rare on imbrices (Brodribb 1987: 26) but they are not unknown. Smoothing lines parallel to the long edge of the tile were present on 66 imbrices, with a further 4 being smoothed parallel to the curving end of the tile. One sherd had reduced surfaces, 35 had reduced cores, two were reduced throughout and one had an oxidised core (caused by an increase in oxygen during firing). One sherd had a sooted top and two sherds were heavily abraded.

Box flue tiles A total of 46 sherds of box flue tile were present, with an additional ten sherds that were probably box flues. The box flue sherds ranged from 13-24 mm in thickness (53 measurements). Such tiles were usually keyed on two opposing plain sides, with the other two sides being plain but pierced by vents. Box flues were used in columns to line the sides of hypocausted rooms,

Three sherds of ridge tile were present which ranged in thickness from 19-25 mm. All three had been smoothed longitudinally on the upper surface. Two unusual sherds were present which were curving in cross-section and pierced by a knife-cut circular vent (Figure 187). These possibly represent ridge tiles designed either to allow fumes to vent directly through the roof, or to accommodate a separate chimney which would have been mortared into place. A single sherd from a chimney was present in context 206. This was a wheel thrown sherd 140 g in weight and 9 mm thick (Figure 188), with two horizontal bands of pie-crust decoration, between which was part of a circular vent 22 mm in diameter. Despite the presence of a vent, which indicated that this was a chimney fragment, there was no sooting on the interior.

Figure 187: Curving tiles with vents from context 758

201

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Figure 188: Chimney sherd from context 206

Figure 189: Chimney or finial fragments from contexts 202 and 236

thereby allowing the hot air to pass from under the floor and through the walls, before venting at the roof line of the building. The vented sides of the columns of box-flues were placed side by side allowing hot air to circulate through the tiles, one keyed side was attached to the wall, and the remaining keyed face formed the interior of the room, which was lined with plaster.

drawn and 0.1 percent relief patterned (McComish 2012: 162).

Thirty-six of the box flue sherds were fragments of plain unkeyed sides, with a scar from a second side visible on most of these sherds. Seven sherds had traces of a rectangular vent, but no complete vent sizes were preserved.

Parietalis

Five of the box-flues had reduced cores. Of six sherds two had sooted interiors and one a sooted exterior resultant from their use within a heated room. One had signs of heat damage post-breakage.

A single example of a parietalis tile was present (Figure 192). Such tiles were used to line the walls of rooms and were held in place by nails. The side facing the interior of the room was usually keyed to enable a lining of plaster to adhere. The sherd was 19 mm thick and had a complex combed pattern on one side, with one vertical line, then a diagonal, then an opposing diagonal, and finally a second vertical line. The comb was 43 mm wide with nine teeth. The sherd was pierced by a circular nail-hole at the very edge of the tile (the original dimensions of the hole being unclear).

Twelve box flues had combed keying. One of the combed examples had keying on a diagonal line, one had a vertical line of combing next to the arris, while one had an arc and one had keying on a diagonal then a vertical line (Figure 190). The remaining sherds were insufficiently preserved to determine the direction of the combing. Three sherds had keying lines drawn with the fingers, of which two had keying in an X-shaped pattern (Figure 191). The proportion of combed to finger keyed flues varies considerably nationally, with Brodribb (1987: 78) recording that 49 percent were combed, 33 percent were incised and 13 percent were relief-patterned, while the figure for York was 93 percent combed, 6 percent knife-cut, 0.9 percent finger

Floor tiles There were three kinds of flooring made of tesserae, opus tessellatum comprising simple geometric patterns using the largest size of tesserae, elaborate patterns called opus musivium, and the finest quality were opus 202

J.M. McComish: The ceramic building materials

Figure 190: Combed box flues from contexts 253 and 309

Figure 191: Finger drawn keying on a box flue from context 4023

Figure 192: Parietalis from context 202

203

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire Signatures and stamps Roman tiles can have marks termed signatures on their upper surfaces. These were drawn while the tile was still wet, usually using the fingers, and they are in the form of simple shapes. The function of these marks is unclear; it has been suggested that they could represent trade-marks, or were to denote grades of differing quality tile, or were designed to indicate which tiles were to be stamped by the overseer (Brodribb 1987: 104; McWhirr and Viner 1978: 364). In the case of the present site there were seven partial signature marks, one tegula from TT311 had a horizontal finger stroke, one sherd had two parallel finger strokes (TTU/S), two non-diagnostic sherds had a single arc (206) and (251), while one tegula in TT314 had two parallel finger drawn arcs, one had three parallel finger drawn arcs (760), and the seventh example from 202 was insufficiently preserved to determine the original form. As the signatures are partial their original form is unclear, but arc-shaped examples are common (Brodribb 1987: 102). No stamps were present on any of the Roman CBM.

Figure 193: Tesserae from context 202

vermiculatum which were made using exceptionally small tesserae that were of sufficient quality to imitate paintings, this was usually only used for the small central panels of floors, the emblemata (Johnson 1995: 8; Adam 1994: 234).

The CBM fabrics The Roman CBM comprised three distinct fabrics which are

Two tesserae made from cutting fragments of CBM were present at Brooklyn House (Figure 193). These were 32 x 27 x 10 mm and 26 x 25 x 15 mm in size respectively, suggesting that they originated from tessellated pavements. The small number of examples present suggests that the tesserae were dumped at the site from elsewhere in the vicinity, rather than originating from the various structures at the site.

1 – Light orange-red fabric. Frequent very fine rounded quartz, very occasional small limestone fragments. Very well sorted. 2 – Light orange fabric. Moderate silty patches up to 2 mm across. Occasional linear voids. Very little quartz. Very well sorted.

Roman material of uncertain form

3 – Beige-light-brown. Frequent very fine rounded quartz. Rare mica. Very well sorted.

The overwhelming bulk of the Roman material (537 sherds) was so poorly preserved that the original form could not be determined, though they were clearly not imbrices as they had flat as opposed to curving surfaces. The sherds ranged in thickness from 14-53 mm. The thinner sherds were probably derived from tegulae originally, while the thicker sherds probably derived from bricks. Two of these sherds had finger drawn keying lines on the upper surface, one had two possible finger lines on the upper surface and eight had smoothing lines parallel to one edge. One sherd had a hoof print (possibly from a deer) on the upper surface, showing that it was laid on the ground to dry prior to firing, as opposed to being placed on a rack in a drying shed. Animal prints are commonly seen on Roman tiles (Brodribb 1987: 126), and the presence of a deer print would imply that the CBM production site was sufficiently rural for a wild animal to wander across it. One sherd was oxidised, 49 sherds were reduced and one sherd had blown through over firing.

The Medieval, Post-medieval and Modern CBM Very little medieval CBM was recovered (3.8% of the total volume) comprising 13th-16th century plain and ridge roof tiles. This material probably arrived on the site as a result of the disposal of waste from nearby settlement. Post-medieval pan tiles and bricks accounted for 6.8% of the total volume, while the modern material accounted for 13.2% of the total volume, comprising a mixture of bricks and pan tiles, together with machine-made roof tiles and ducts. CBM IN RELATION TO THE SITE STRATIGRAPHY The following text relates to the phasing for the main excavations and does not include the Trial Trenching work. The volume of material recovered from each phase is summarised in Table 24. 204

J.M. McComish: The ceramic building materials

No. of sherds

Phase

Forms

1

Roman – Box flue, imbrex, brick, tegula, unknown

2 3 4 5

7

US

Weight in grams

% of total weight

45

3725

3.58

16

1095

1.05

1

50

0.05

72

9250

8.89

12

910

0.87

60

7555

7.26

3

175

0.17

536

45038

43.28

Intrusive – 13-16th century roof tile, post-medieval brick, modern brick and tile

56

3385

3.25

Residual - Roman (Box flue, imbrex, brick, tegula, unknown), medieval 13-16th century roof tile, post-medieval brick Pan tile, modern brick and tile

148

8350

8.02

Roman (Box flue, imbrex, brick, tegula, unknown), medieval 13-16th century roof tile, post-medieval brick, pan tile, modern brick, modern duct, modern tile

121

11685

11.23

154

12841

12.34

Roman – Box flue, imbrex, brick Intrusive – 13-16th century roof tile Roman – Box flue, chimney, imbrex, brick, tegula, unknown Intrusive – 13-16th century roof tile, modern brick Roman – Box flue, imbrex, brick, other, tegula, unknown Intrusive – 13-16th century roof tile Roman – Box flue, chimney, imbrex, brick, other, parietalis, tegula, tessera, unknown

Table 24: Ceramic Building Material in relation to phase

chimney sherd and four tegulae) is low. It is impossible to know whether the CBM in this phase represents the re-use of residual material from the area or the arrival of newly imported waste for use as building material. It should be noted that some intrusive medieval and postmedieval CBM was also present within this phase.

Phase 1. Most of the CBM in this phase was recovered from deposits forming the ground surface or overlying a field system of Iron Age or Romano-British date, but some was recovered from a ditch backfill. Given that there were no buildings in the area at this stage it is clear that this CBM must have been brought to the site as waste, disposed of from nearby settlement. The presence of box flue within this phase indicates that some of this material originated from a hypocausted building.

Phase 4 (4th century). All the Roman CBM occurred in demolition or rubble deposits. There was very little roof tile present (17 sherds of imbrex, two vented roof tiles and five tegulae). The low volume of roofing tiles seen in Phases 3-4 suggests that the various structures of Phase 3 did not have CBM roofs. Given the nature of the deposits in this phase the bulk of the CBM was probably the result of demolition, though some may have been residual. Some intrusive medieval and postmedieval CBM was also present.

Phase 2 (mid-3rd century?). Three sherds were randomly incorporated into a Roman road surface, and the remainder of the CBM was from deposits building up over the road. This could represent residual material from Phase 1, or the dumping of new material. Phase 3 (late 3rd and 4th century). While some of the Roman CBM was from surfaces, build-ups above the Roman road and pit fills, the majority was re-used as building material within the walls of Structures A, C, D, E, and in two possible plot boundary walls. CBM also occurred randomly in the floors of Structure A and in yard surfaces associated with Structures A and D. There is no indication that any of these structures had any form of hypocaust system, the sherds of box flue and chimney from this phase must therefore have originated from structures elsewhere, being re-used as random building materials. The volume of roof tiles recovered from this phase (15 sherds of imbrex, one

Phase 5 (late 4th-5th century). The bulk of the Roman CBM was recovered from rubbish dumps in this phase. It probably represents a mixture of demolition derived material and the use of the area for the disposal of rubbish from nearby settlements. Some intrusive CBM was also present. Phase 6 (medieval and post-medieval). Although no deposits of this date were excavated both medieval roof tiles and post medieval brick/roof tiles were present on the site, occurring intrusively in some of the Roman phases and residually in Phase 7. This material 205

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire probably arrived at the site as a result of dumping from the nearby settlement.

SUMMARY The bulk of the material examined was of Roman date, and it represents either the re-use of CBM sherds as building materials or the disposal of waste from the vicinity. The collection is largely of standard Roman forms, but there are a few unusual pieces, notably the fragments of chimney and the two vented roof tiles.

Phase 7 (Victorian and later). The CBM was from a mixture of backfills, yard areas, subsoils and topsoil, comprising both residual and modern CBM.

206

Chapter 9

The stone building materials J.M. McComish the use of the word tile, which could imply that the item concerned is ceramic rather than stone. The term paving stone is used for stone flooring, whether interior to or external to buildings. Where the identification or the form was uncertain a question mark was placed after the form name, for example ‘quern?’.

INTRODUCTION AND METHODOLOGY The stone building material (SBM) from the site comprised 11358 fragments, collectively weighing 4494.29 kg. The SBM is summarised by form and stone type in Table 25. Within the text the term ‘flag’ is used for stone roofing, avoiding any confusion created by Form

Year recorded

Paving stone Ls paving stone MSS paving stone Non-diagnostic building stone Ools uncertain function Roof flag Fls Roof flag Ls roof flag and probable roof flag

2015 2016 2018 2016 2018 2015

No. of frags. Weight in % of total grams weight 8 8124 0.1808 15 16269 0.3620 7 8195 0.1823 25 15969 0.3553 15 10559 0.2349 188 75706 1.6845

Mss roof flag and probable flag tile Flint Cobble Glacial erratic Quern? Ls building material

2018 2015 2018 2016 2018 2016 2018 2016 2018 2016 2018 2016

1 396 1 2164 197 1955 32 3 1 1 2 1109

748 51578 642 305326 61541 352587 10041 2647 2122 2412 1350 679755

0.0166 1.1476 0.0143 6.7936 1.3693 7.8452 0.2234 0.0589 0.0472 0.0537 0.0300 15.1249

Ools building material

2016

5165

2735910

60.8752

Ls rubble Ools rubble Mss rubble Ls wall facing Ools wall facing Carved block Cap stone? Chalk tessera Mss tessera Coffin?

2016 2016 2016 2018 2018 2018 2018 2018 2018 2018

13 23 1 12 19 1 1 1 1 1

4995 19491 425 34614 81824 525 3670 10 5 7250

0.1111 0.4337 0.0095 0.7702 1.8206 0.0117 0.0817 0.0002 0.0001 0.1613

Table 25: Stone Building Material by form and stone type (ls =limestone, ools = oolitic limestone, fls = fossiliferous limestone, mss = micaceous sandstone)

207

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire The SBM was recorded in three stages. The SBM from the Trial Trench work (592 fragments weighing 135.408 kg) was recorded by S. Tibbles with a selection of stone being retained. The SBM from the main phase of excavation was recorded on site by S. Tibbles, with 332 fragments being retained and the remainder being discarded on site. Twelve of the retained fragments were allocated ’Recorded Find’ (RF) numbers. In October 2018 the retained SBM was re-recorded by J. M. McComish (2019). The results of the three phases of recording have been combined here. RESULTS The SBM from the site comprised a mixture of building stone, roof flags and paving stones. For the most part the building stone was very rough without clear tool marks, and even those stones with tool marks were (with one exception) roughly dressed blocks with coarse diagonal tooling lines. The large number of fragments of roof and flooring flagstones were badly fragmented lacking any surviving original dimensions other than the thicknesses.

Figure 194: Two stone tesserae, RF 6 context 201 and RF 296 context 202

Floor and Roofing Stone A small number of fragments were probably derived from paving stones originally. These comprised flat slabs ranging from 17-60 mm in thickness, each of which had one worn surface, presumed to be the surface of a floor. Two of the retained pieces each had one straight edge surviving, but all other edges were broken off. The paving stones were either micaceous sandstone or limestone (roughly equal volumes of each stone type), and some had been burnt. The use of stone for flooring is known from many Roman sites, including in Norton, such as the floor of a pottery kiln at Norton Community Primary School (Stephens and Ware 2012: 4) or in Building B at the 27 Wood Street site in Norton (Burn et al. 2018: 10).

Figure 195: Stone roof flag of rectangular shape, context 816

Two tesserae were recorded, one of chalk and one of micaceous sandstone (RFs 6 and 296; Figure 194). These were 33 x 19 x 13 mm and 14 x 15 x 18 mm in size respectively. The low number of examples present suggests that these probably represent the dumping of material derived from a Roman tessellated pavement somewhere in the vicinity of the Brooklyn House site, as opposed to originating from any building at the site itself.

The use of roofing flags was widespread in Roman Britain. For example, on sites at Caernarfon, Chester, Cirencester, Exeter, Gloucester, Heslington East near York, Lincoln and York (McComish 2012: 48). Examples of roof flags are also known from sites in the immediate vicinity of Brooklyn House, Wenham refers to their inclusion in a Constantinian rebuilding of the northeastern gateway of Malton fort and their discovery at Orchard Field, Malton (Wenham 1974: 18, 36).

Numerous thin flat slabs were present which either definitely or probably originated from roof flags. These ranged from 3-50 mm in thickness, but the thinner examples at 3-7mm thick had clearly flaked off larger slabs. No original lengths or breadths survived, but the largest fragment was in excess of 740 x 137 mm in size.

The bulk of the roof flags were of limestone or micaceous sandstone (almost equal volumes of each stone type), and there was a single example in fossiliferous limestone containing numerous small scallop-like shells. 208

J.M. McComish: The stone building materials

Figure 196: Stone roof flag of elongated hexagonal shape, context 207

It is clear that roof flags of various shapes were used across Roman Britain. For example, in Dorset roof flags of elongated hexagonal, elongated pentagonal, lozenge and ovate design have all been recorded (Palmer 2019), while in York hexagonal, heptagonal and rectangular examples are known (York Archaeological Trust database, searched 8/7/2019). Closer to the present site, lozenge-shaped stone roof flags were recovered from Orchard Field, Malton (Wenham 1974: 36). The hexagonal and pentagonal shapes would have been laid on the roof to overlap in such a way as to create a lozenge pattern on the roof. Pliny describes tiles, which he termed pavonoceous, i.e. resembling the spots on a peacock’s tail, stating that in parts of the Belgae province these were made of stone (Bostock and Riley 1893: Book XXXVI, 44). Only 15 of the roof flags examined from Brooklyn House had surviving original edges, and there were clearly three different shapes of stone roof flags present, with one rectangular example, one elongated hexagonal example and five with an upper edge that curved in plan (Figures 196-198). One of these curving tiles was re-used as flooring in Structure A, one was in a Phase 3 ground level, two were from rubbish dumps in Phase 5 and one from a backfill in Phase 7. It seems probable that the curving flags did not originate from buildings at this site but were brought in from elsewhere

Figure 197: Stone roof flags with a curving upper edge from contexts 202, 220 and 830

for re-use as general building material, occurring residually at the site thereafter. Thirty-three of the roof flags examined in 2018 had nail-holes (nails being used to attach the flags to a roof). The nail-holes ranged from 6-13mm in size. Some nail-holes were roughly square in shape having been chipped out of the flag, while others were circular suggesting that they were drilled. Five of the nail-holes had traces of iron from the associated nail adhering. 209

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire blocks were retained as representative samples in 2016 (RFs 173, 175, 183, 190, 233 and 334); these were recorded in 2018. Approximately two-thirds of the building stone was oolitic limestone, while a third was limestone. There was a single fragment of micaceous sandstone which may represent a badly preserved portion of a paving stone. The dressed stones were roughly hewn with coarse tooling lines in diagonal drafts. The coarse nature of these stones suggests that they were from utilitarian structures. The rubble from the site could have been from wall-cores or could represent badly broken fragments of wall facing blocks. Many of the stones (both the rubble and wall facing stones) had been burnt.

Figure 198: Roof flags re-used as lids, RF 194, context 235 and RF 308, context 202

Specific items of building stone:

Figure 199: Roof flag with scratched surface, possibly reused as a hone, context 202

1.

One fragment of oolitic limestone from (202) was a flat slab with a bull-nose shaped cross-section along the surviving edge (Figure 200). The precise original function of this item is unclear, though it may represent flooring given that one surface was worn.

2.

A 45 mm thick flat slab of oolitic limestone found in grave (284) presumably represented a disturbed capstone.

Numerous examples of the roof flags had been burnt. Two roof flags had been cut-down to form disc shapes, possibly for use as lids (RFs 194 and 308; Figure 198), while one fragment of roof flag had fine scratches on the surface suggesting re-use as a hone (Figure 199). One of the roof flag from (202) had a tar-like substance on one surface, though this need not have related to the original use of the tile. Building Stone A large quantity of building stone was present at the site which ranged from rubble with no worked faces, to very roughly dressed wall facing stones. Six of the

Figure 200: Limestone slab with bullnose profile, context 202

Figure 201: Carved stone RF 177, context 215

210

J.M. McComish: The stone building materials

Figure 202: Carved block RF 96, context 202

3. RF 177 (Figure 201) (context 215) was a carved block of oolitic limestone which was roughly rectangular in plan and cross-section, with four surviving original faces (F1 = the base, F2-3 = the sides, F4 = the elevation). F4 had three parallel grooves 46 mm, 78 mm and 111 mm from one edge, which ranged in length from 75-80 mm, were 8 mm wide and 4 mm deep. The block stepped in slightly above these grooves. This clearly formed a decorative piece originally but given its poor preservation and small size the original function is unclear.

of Britain viewer 2019). The oolitic limestone outcrop of the Malton area was a major source of building stone in the Roman period not just for the Malton/Norton area, but also for Roman York (Robinson 1978: 1; Buckland 1976: 36-7). Chalk could be sourced some 4.5km east of Norton, slightly to the east of Scagglethorpe (Geology of Britain viewer 2019). THE SBM IN RELATION TO THE SITE PHASING The following text relates to the phasing for the main excavations and does not include the Trial Trenching work. The volume of material recovered from each phase is summarised on Table 26.

4. RF 96 (Figure 202) (Context 202) was a block of carved oolitic limestone, which had a curving outer edge in plan and an L-shaped cross-section, with a ‘rim’ 82 mm wide and up to 70 mm deep around the outer edge of the block. This was very roughly dressed and somewhat eroded. It possibly represents part of a stone coffin. This could have originated from a nearby cemetery area (see Figure 3).

Phase 1. Most of the SBM was from ground surfaces or from soils overlying an Iron Age/Romano British field system, including fragments of limestone and oolitic limestone, limestone and micaceous sandstone roof flags, and a micaceous sandstone paving stone. Material from within the roadside ditch fills included a cobble, a fragment of flint, two fragments of oolitic limestone building material and two limestone roof flag fragments. As there were no structures at the site at this stage this material must have been brought to the site as waste, disposed of from nearby settlement.

Two fragments of flint and a fine-grained sandstone cobble were present, but none were directly associated with the structures at the site. A glacial erratic was also present within a deposit associated with the demolition of Structure F, suggesting that it may originally have been used within the walling of that structure. Two fragments of coarse-grained sandstone were also present which probably formed parts of querns as they each had a curving outer edge in plan.

Phase 2 (mid 3rd century?). Most of the fragments were from the road surface or road repairs and comprised fragments of limestone and oolitic limestone building stone, and limestone and micaceous sandstone roof flags, all of which were re-used within the road surfaces. Various build-up deposits contained limestone roof flags which could represent either residual or freshly dumped material. A block of oolitic limestone from a grave fill 285 probably represented a capping stone laid flat above an infant burial (SK 286).

Geology Micaceous sandstone was widely used in Roman Yorkshire for paving stones and roof flags. It was sourced from the Elland area near Leeds and was traded as far afield as Hibaldstow in North Lincolnshire (Carver et al. 1978: 41). Clearly transporting these stone flags over considerable distances (by water and road) was routine and did not affect the viability of the stone roof flag industry in terms of transport costs.

No SBM was present within the contexts associated with the possible mausoleum of this phase. A fragment of an oolitic limestone facing stone and two limestone roof flag fragments were, however, found within the Phase 3

Both Coralline Oolite Limestone and Scarborough Formation Limestone rocks outcrop at Malton (Geology 211

Life, Death and Rubbish Disposal in Roman Norton, North Yorkshire

Structure

Context no and type

Forms

A

External walls 205, 217, 219, 232

BM – LS, OLS R – LS P - LS BM – OLS R – LS BM – LS P – LS BM – LS, OLS R – LS BM – LS BM – LS, OLS BM – LS R – LS BM – OLS BM – LS BM – LS R – LS BM – OLS BM – LS BM – LS BM – LS R – LS R – LS BM – LS, OLS R – LS

Internal floor 229, 280, 281 External yard 236 C

External walls 249, 250

D

Possible capstone 246 External walls 815 Internal floor 3007, 3011

E

D or E F G Plot wall Plot wall?

Post hole packing 3017 Wall repair 328 External walls 821, 314 Yard surface 3082 Wall 817 External walls 3061 Internal wall 3092 Internal wall 430 Threshold 431 Division wall 3062 Soil 309 possibly from wall 308

No. of fragments examined 1002 40 1 4 20 4 6 29 2 1 2 10 2 1 1 3 2 3 1 1 1 2 49 3 7

Size range in mm 47-391x 21-281x18-331 7-30 24 341-358x 284-298x 61-86 6-18 Not recorded 10-32 110-355x89-214x41-138 18-21 833x175x62 354-391x251-291x113-247 16-67x10-74x8-27 10-11 205x100x23 324x220x98 324x220x98 12-14 91-165x80-155x64-84 384x194x132 306x235x181 295x242x143 8-10 16-23 123-130 x106-115x65-78 7-18

Table 26: Stone Building Material from Phase 3 structures (BM = building material, R = roof flag, P = paving, LS = limestone, OLS = oolitic limestone, sizes for BM are in the form length x breadth x thickness, for R and P the thickness range is given)

robbing trench of this structure, which may give some hint as to the materials used in the construction of the building.

paving stones, re-used burnt stones and in the case of Structures A, C, D and E re-used CBM. The walls of the structures varied slightly in terms of their construction method, ranging from a cut containing stone blocks on either side and a rubble core (external walls of A, D and E), to internal walls one stone wide (Structures F and G). The walls are similar to others in the vicinity with limestone walling known from a pottery kiln at Norton Community School (Stephens and Ware 2012: 5) while oolitic limestone walling is known from 27 Wood Street, Norton (Burn et al. 2018: 13). There was evidence of mortar bonding in the walls of structures D, E, D or E, and G while Structures A, C and F had no apparent bonding. A posthole within Structure D contained a fragment of oolitic limestone acting as post-packing.

Phase 3 (late 3rd and 4th century). While some of the SBM from this phase related to repairs to the road, pit fills and ground levels, the bulk of the stonework examined related to structures, boundary walls and yard surfaces. Details of the stonework examined in relation to the various structures is given in Table 26. Most of the stones within the structure walls lacked any form of tool marks, and even where such marks were present (Table 27), they were in the form of rough diagonal tooling lines. None of the wall facing stones was finely dressed. It is clear that the structure walls at the site were not built from freshly quarried stone, being constructed from whatever materials were readily available (Table 26) including a mixture of stone types (limestone and oolitic limestone blocks), re-used roof flags and

The rough nature of the stonework in all of the Phase 3 structures is suggestive of utilitarian structures. It is possible that the walling of structures A-C and D-H was intended to carry timber superstructures, as was 212

J.M. McComish: The stone building materials

Structure Context no and type A C E D or E F G Plot wall

No of BM No. of re-used burnt fragments frags. of BM or R with tooling present External walls 205, 217, 219 6 10+ Internal floor 229, 280, 281 8 External walls 249, 250 External walls 821, 314 Yard surface 3082 Wall 817 Internal wall 3092 Internal wall 430 Division wall 3062

1 1 1 1

1

3 3

Table 27: Tooling and burning on the Stone Building Material by structure (BM = building material, R = roof flag)

suggested for the buildings at Orchard Field Malton (Wenham 1974: 36).

that this building had a stone roof originally. The volume of roof flags associated with the demolition of the remaining structures was too low to be indicative of stone roofing.

Structure C differs from the other structures at the site, being some form of oven or drying kiln, though the walls were in no way distinctive from the other structures at the site in terms of the building materials used. A single re-used tooled block in Structure C was possibly a corner-block originally. A possible capstone for the structure made of limestone was present in an associated context, surface 246.

Phase 5 (late 4th-5th century). The SBM was in rubbish dumps together with robbing and demolition deposits; it included limestone, oolitic limestone and flint building stone, together with limestone and micaceous sandstone pavers and roof flags. Phase 6 (medieval and post-medieval). No deposits of this date were excavated, and none of the SBM was in forms suggestive of medieval or post-medieval date.

Relatively little evidence for deliberate floors within the structures survived. Structure A had the remains of stone surfaces inside, 229 and 281, but these were badly preserved, while a mortar floor within Structure D incorporated random fragments of a limestone block and limestone roof flag.

Phase 7 (Victorian and later). The SBM came from a mixture of backfills, yard areas, animal grave backfills and topsoil. It comprised residual limestone and oolitic limestone building stone, together with limestone and micaceous sandstone roof flags.

With regards to roofing, Structure A may have had a stone roof (see Phase 4 below) and Structure C had capping stones. Given that relatively few roof flags were present in the demolition contexts of Phase 4/5 associated with Structures B and D-H, it is possible that these structures were not roofed in stone, and as noted above they do not seem to have had CBM roofs either; these structures may therefore have been roofed in thatch.

SUMMARY The roof flags at the site were largely typical for the Roman period, though the flags with a curving upper edge were unusual. Hopefully, future excavations in the Norton area will uncover a complete example of such a roof flag to determine the original form. The building stones largely represented fragmentary, low-grade and re-used blocks, from what seem to be utilitarian structures. The only two carved items being a re-used block decorated with three parallel grooves and a second block which possibly represents part of a coffin. It has proved difficult to find comparative information on the relative volumes of limestone and oolitic limestone used in Roman buildings in the Norton/Malton area. It is hoped that the information given in Table 26 will be of use to archaeologists investigating other Roman remains in the area.

Some of the pavers and buildings stones in the yards associated with Structures A and D had worn upper surfaces, though whether this was the result of re-use of worn stones, or due to wear during the lifespan of the yard surface is unclear. Phase 4 (4th century). Most of the SBM in this phase was from dumps of rubble associated with the demolition of the earlier Phase 3 structures A, B, D-F and H. In the case of the deposits associated with the demolition of Structure A, 413 fragments of roof flags were present (collectively weighing 808.64 kg), which may suggest 213

Chapter 10

Chipped stone lithics Spencer Carter the west in the Vales of Mowbray and York, noting that this area was not itself glaciated inland from the present coast, where boulder clay deposits are evident, at the last glacial maximum (Bridgeland et al. 2011; Young 1984; 2000). The speckled flint is very homogenous in character and consists mostly of grey or light brown items often with a range of different shades and mottling within one piece. Translucent flint is present too in expected proportions and is also derived from the same sources. Three notable examples are the Late

QUANTIFICATION AND CONTEXT A total of 53 chipped stone lithics were provided, comprising 50 humanly modified pieces and three natural, a total of 497 g. The natural items have been excluded from analysis. Fourteen individual lithics were submitted as Registered Finds, and the remaining 39 as individual finds or grouped by context (or unstratified). All the chipped stone finds, irrespective of context and recording method, are residual without any indication of a primary in situ context or related clustering. Table 28 summarises the submitted finds composition.

RAW MATERIALS Flint - Translucent Flint - Speckled Flint - Opaque (non-white) Flint - White Flint - Indeterminate

RAW MATERIAL (Table 29) The entire finds collection is comprised of flint. There are no examples of chert, quartz, jasper or other finegrained stone such as tuff. The majority of the lithics are of speckled flint which is present most especially in coastal beach deposits to the east of the Vale of Pickering as well as in glacial tills and river gravels to COMPOSITION - ALL Formal Tools Non-formal Tools / Utilised Debitage / Unmodified Total COMPOSITION - TOOLS Arrowhead projectiles Scrapers Awls Piercers Utilised & retouched blades Utilised & retouched flakes Gunflints Total COMPOSITION - DEBITAGE Cores / fragments / rejuvenation Blades / fragments Bladelets / fragments Flakes / fragments Angular debitage / indeterminate Chips